From dfe1bfa58072bbff441d94bdf8dcd5a985990b7b Mon Sep 17 00:00:00 2001
From:  <>
Date: Wed, 17 Jul 2024 07:07:39 +0000
Subject: [PATCH] Generate Pelican site [ci skip]

---
 .../auto_examples_python.zip                  | Bin 31143 -> 31143 bytes
 .../auto_examples_jupyter.zip                 | Bin 46873 -> 46873 bytes
 _images/sphx_glr_plot_compare_time_001.png    | Bin 41142 -> 41021 bytes
 _images/sphx_glr_plot_compare_time_thumb.png  | Bin 29312 -> 29280 bytes
 .../sphx_glr_plot_lasso_vs_weighted_001.png   | Bin 17276 -> 17076 bytes
 .../sphx_glr_plot_lasso_vs_weighted_thumb.png | Bin 11681 -> 11750 bytes
 _images/sphx_glr_plot_reweighted_l1_001.png   | Bin 18550 -> 18582 bytes
 _images/sphx_glr_plot_reweighted_l1_thumb.png | Bin 12504 -> 12506 bytes
 .../sphx_glr_plot_survival_analysis_002.png   | Bin 25012 -> 25194 bytes
 _images/sphx_glr_plot_ucurve_001.png          | Bin 28775 -> 29790 bytes
 _images/sphx_glr_plot_ucurve_thumb.png        | Bin 20286 -> 20374 bytes
 .../sphx_glr_plot_zero_weights_lasso_001.png  | Bin 23226 -> 22668 bytes
 ...sphx_glr_plot_zero_weights_lasso_thumb.png | Bin 19010 -> 17954 bytes
 .../auto_examples/plot_compare_time.rst.txt   |   4 +-
 .../plot_group_logistic_regression.rst.txt    |   2 +-
 .../plot_lasso_vs_weighted.rst.txt            |   2 +-
 .../plot_logreg_various_penalties.rst.txt     |   2 +-
 _sources/auto_examples/plot_pen_prox.rst.txt  |   2 +-
 .../auto_examples/plot_reweighted_l1.rst.txt  |   2 +-
 .../plot_sparse_group_lasso.rst.txt           |   4 +-
 .../plot_sparse_recovery.rst.txt              |   2 +-
 .../plot_survival_analysis.rst.txt            |   6 +--
 _sources/auto_examples/plot_ucurve.rst.txt    |   4 +-
 .../plot_zero_weights_lasso.rst.txt           |   2 +-
 .../auto_examples/sg_execution_times.rst.txt  |  28 ++++++------
 _sources/sg_execution_times.rst.txt           |  28 ++++++------
 auto_examples/plot_compare_time.html          |  40 +++++++++---------
 .../plot_group_logistic_regression.html       |   2 +-
 auto_examples/plot_lasso_vs_weighted.html     |   2 +-
 .../plot_logreg_various_penalties.html        |   2 +-
 auto_examples/plot_pen_prox.html              |   2 +-
 auto_examples/plot_reweighted_l1.html         |   2 +-
 auto_examples/plot_sparse_group_lasso.html    |   4 +-
 auto_examples/plot_sparse_recovery.html       |   2 +-
 auto_examples/plot_survival_analysis.html     |   6 +--
 auto_examples/plot_ucurve.html                |  30 ++++++-------
 auto_examples/plot_zero_weights_lasso.html    |   2 +-
 auto_examples/sg_execution_times.html         |  28 ++++++------
 generated/skglm.CoxEstimator.html             |   4 +-
 generated/skglm.ElasticNet.html               |   4 +-
 .../skglm.GeneralizedLinearEstimator.html     |   4 +-
 generated/skglm.GroupLasso.html               |   4 +-
 generated/skglm.Lasso.html                    |   4 +-
 generated/skglm.LinearSVC.html                |   4 +-
 generated/skglm.MCPRegression.html            |   4 +-
 generated/skglm.MultiTaskLasso.html           |   4 +-
 generated/skglm.SparseLogisticRegression.html |   4 +-
 generated/skglm.WeightedLasso.html            |   4 +-
 generated/skglm.datafits.Cox.html             |   4 +-
 generated/skglm.datafits.Gamma.html           |   4 +-
 generated/skglm.datafits.Huber.html           |   4 +-
 generated/skglm.datafits.Logistic.html        |   4 +-
 generated/skglm.datafits.LogisticGroup.html   |   4 +-
 generated/skglm.datafits.Poisson.html         |   4 +-
 generated/skglm.datafits.Quadratic.html       |   4 +-
 generated/skglm.datafits.QuadraticGroup.html  |   4 +-
 generated/skglm.datafits.QuadraticSVC.html    |   4 +-
 .../skglm.datafits.WeightedQuadratic.html     |   4 +-
 ...lm.experimental.IterativeReweightedL1.html |   4 +-
 generated/skglm.experimental.PDCD_WS.html     |   4 +-
 generated/skglm.experimental.Pinball.html     |   4 +-
 generated/skglm.experimental.SqrtLasso.html   |   4 +-
 .../skglm.experimental.SqrtQuadratic.html     |   4 +-
 generated/skglm.penalties.BlockSCAD.html      |   4 +-
 generated/skglm.penalties.IndicatorBox.html   |   4 +-
 generated/skglm.penalties.L0_5.html           |   4 +-
 generated/skglm.penalties.L1.html             |   4 +-
 generated/skglm.penalties.L1_plus_L2.html     |   4 +-
 generated/skglm.penalties.L2.html             |   4 +-
 generated/skglm.penalties.L2_3.html           |   4 +-
 generated/skglm.penalties.LogSumPenalty.html  |   4 +-
 generated/skglm.penalties.MCPenalty.html      |   4 +-
 .../skglm.penalties.PositiveConstraint.html   |   4 +-
 generated/skglm.penalties.SCAD.html           |   4 +-
 generated/skglm.penalties.SLOPE.html          |   4 +-
 .../skglm.penalties.WeightedGroupL2.html      |   4 +-
 generated/skglm.penalties.WeightedL1.html     |   4 +-
 .../skglm.penalties.WeightedMCPenalty.html    |   4 +-
 generated/skglm.solvers.AndersonCD.html       |   4 +-
 generated/skglm.solvers.FISTA.html            |   4 +-
 generated/skglm.solvers.GramCD.html           |   4 +-
 generated/skglm.solvers.GroupBCD.html         |   4 +-
 generated/skglm.solvers.GroupProxNewton.html  |   4 +-
 generated/skglm.solvers.LBFGS.html            |   4 +-
 generated/skglm.solvers.MultiTaskBCD.html     |   4 +-
 generated/skglm.solvers.ProxNewton.html       |   4 +-
 searchindex.js                                |   2 +-
 sg_execution_times.html                       |  28 ++++++------
 88 files changed, 217 insertions(+), 215 deletions(-)

diff --git a/_downloads/07fcc19ba03226cd3d83d4e40ec44385/auto_examples_python.zip b/_downloads/07fcc19ba03226cd3d83d4e40ec44385/auto_examples_python.zip
index b0b35efe7efd0a1aa6f63f90b1c8ea651a3c071c..0f4e8454491aadb459516cd0c8b243af5c903007 100644
GIT binary patch
delta 202
zcmZ4fnQ{4NM&1B#W)=|!5a_q~xRF<tmHDj2$IYg!e-xOf0~sc28(ElV02wc|D%hD9
z0vX9h6IhuS0~z1UbJ>`e02yg^HtfvHfsCcD;T+7XfsFsYd$^g`0U7%uJGhxo02!XC
zp@Pgn8zwI-^8wi~`9qlt$b!k{<<=ncCzqC6f{dTMvD^h@`s5$wt{}rFJ5;!W%%0p)
Z;SDl+^0f*FkjaxJEA2rBPmZqi0RUjeRxbbm

delta 202
zcmZ4fnQ{4NM&1B#W)=|!5D1KYzmZp!mANPO{bp0vKMKr|K!%ChMi%BMAmfEr1v_&R
zkdbUOft5K0$oOWS%f_4vWTe^Iurp@>8B1NmIhb>RjQ_rSxS8{TjD3+E+{_(7hG%N1
zAT!X0$qUPTKsHSNQ04-%V6u6+HOTzQrRA0&<0o${cLA9``A4}c$nePy6|Nw&CwEkM
XgN&Yht-=9h@?^<Mdyv7Cqbq#?d_z;m

diff --git a/_downloads/6f1e7a639e0699d6164445b55e6c116d/auto_examples_jupyter.zip b/_downloads/6f1e7a639e0699d6164445b55e6c116d/auto_examples_jupyter.zip
index 1a86b3ce581fff6741011cd2dd8599212a7868fe..08a4f2725b33dba671f744b0005b7c5b9b1bea13 100644
GIT binary patch
delta 202
zcmbRFj%nsQCf)#VW)=|!5SVW9aU-uEH}h+YkDKGTgG`ub02!O@C$Tdx1Ts9_FY_@k
z1~TNr>NuE}02%Y+k~o={0~stCd-<4G0~trk%(<D@0U7@q5Arde05T3usN!Qj3uM^O
z-yp&av|;k6&Hf-8CcA8L2U#$=Z;Kts{K@CH*no_m%(vAYWcuWYtsWr5CocxlK(i;m
a+v*20db0gCXOPL0TemrZ44!;yn?C?!ty<my

delta 202
zcmbRFj%nsQCf)#VW)=|!5QvO@zmeCEn|WpI`^|COK_<*mK*nbKN$ku?K!%6=Wj^K<
zAVV&!j)OTB$e16O#L1iiWUyrH<zvnPGLDp)b2H}y8UGp&@-cS+8HXlR@iF%R8TRux
zh%f_fnEYw8Kgfp3E?e9|7EJEjVh1vR^7$<`Amb<VZFL8kJ~?8m2gvZri-9!I?8)!8
Z`hkp|Y`@JJWb)+JZB8JAC!gBp4*+~^SZ)9S

diff --git a/_images/sphx_glr_plot_compare_time_001.png b/_images/sphx_glr_plot_compare_time_001.png
index 8f2c23eccd3190eec97f0b463caafda736e2e369..539fd8acb554f83d525a4b7f81dc39a9f615321b 100644
GIT binary patch
literal 41021
zcma&Oby!wy^fmZ^NC*gmbSsF`(hVY^fFL2=NT)Q?QYxvmQi2Fbcej9acOxMo-4e4O
z-`_WLeKT`iGk?6VN<8PB`|Nw~wbx$je)n8K2KNT}4Fm##`}B#VG6I3>i$I{vW1_=X
zcm}4I;12;ODGetTJ5wiDLkAPYGeakPYda@v3nN+=69-2NJ6kR`J~kc}T5~5Sdq+Wb
zcANk62iWW!%-DC6lQrQ(u<V~`IwBBwhRFX=vcxhi5QvoTPbD9zx+QMTxN5{4UbXM8
zE)?FvmlVf&wRnS$<WtPz9iH=STrL$#%{d3vQf|kx5qXsda$Gi9@!2uB7?$y$C}gTy
zS!tO(QXIW~rtOz7Wgg+Ja?D0-RI~S1_4RnouBx7`bG;BuQV7PEz@!Vv;>~5z_5Jr(
z4t6YG<d^VALea$G4~e&y(x|EU64QYUa<Qq(+58?yxfyb|B;ImHA@0MsF_n6rAa4Kr
z5L0@@4didAauFZ`@te`SFHg<R4%Yl5BJf;XT(~WV-yY9;;Dt#W+s|}FF}k1cw(lMt
zwav~Z|J|yI$V~D66dis0f`@|7cIitnb((3<r{OFm|8LJe?;TCqQA{&FeE8wRU75w@
z<-oc+5kEga7TxM$mXsHmN~fCI)d+Zq8?h{!5@F;#Z{NO6l@24fo~=!Q>qLF})Y+ZD
zh4Qnfhk=_rIx&&5Eu1Q(qJl4uP1k6q+L@k_@z?r#XT#NHmvMxMbdFk{+wqKRQAx?K
z;cOM_xq96zJxxtbr?tUAl}v?)rro6E+~$oDVwXP)4O^ZlD6A~U>6Z_W(FSDEs}>W!
z#g~Yl`N$wgD(cDq>eZ{^0{xWSToy`7N}T-DxvR@|*Ez5D!;LZc^t-&gu^7~DzBE4P
z9~hI}(tiByP2dVv*Y}FaQ7>3rSWsHzN%1-&E2!POG3Rq7;*WucPRq_7(VHaXc(P;2
zre7agQc_|eR*0V_`+?1XE`Y-?ZGoMYHSFtGdTQ?zLryL(0hdiuM1FogB3Hj5xx2fY
zU|Xfc^qKgY@~CAQgWmq%?S@}}{`90i@;_L$wzekav!z^HTRT4-)BkqzMB*)zJo2r_
zH^REeU0q(ha&~TQZ4I0(v)bP*=|lb9e#~Ru@B8*GCNZm4KyWa|moHx=rKQP3uPFHK
zX_Ew<7pqp2Te`YJkgqj{S)Dwf$#J!J;OkSu++3V1OuO0Is!M0OfJ!Q4u)T`-7!|cP
zUB!igN4+%TI=4Kg?_=_-ElfT|l&X=ow&xR1$m8IfyATCYF)^)`uYQ?N6f=;p>!C%*
z{)mq!<+Yjo5lQ!$-|090HH@2IWg@6s??q$M{h0ZP4A5iqjnO=v8kaB#-<`cK_Wg0=
zh=Do}*XikL*;-LrT3^>$_h9&xa+|5P47oUsLIb}`XJ%&RYUhnMxcN?>tBZCF8qda+
zp3f7-CfEpXZ|_85_hJo_SBB~(DJl8_UdnWmFX&L7W72JX(8@zk{he4nJ*^{pb?)fD
zp{%0P(bvbvPlAu{8xun$ARvIyY4ENazwm$a25~TUur>4f<iss4E$v`zX?2y$ZdTZG
zIP1}BZ_@U~V=TJw+?wdp@Ua8_c%pAVf2LSiSO}%yyBQ<fiVUN?5Ef*FKQwOJS&C9z
z+}>97_7<J<I;OHL`_xY3L+J}Y6bBE6+<_@6H1FQM^SnG=orbRpAM`&)2w$D=d!8=G
z;ZX2~di_08YVh__%hPH-`ZJZlWftCzAeNZ^fJ^|DGG&${$bC|KG-*}4-z!vERkaHt
zFqo&!C-M04;n;O_bOLhn&Vq)^PFO#CCntIij!3b~6aSl3f{6YKJN@2dk*>wokf*Y;
z%`Gkd5WrkEliU|9Mi>|v<+d}OC%eBitDPQXXJ>QUOh(()p9UGW1ZgfZLmuHQz6mji
zNhis3*BCMi{Ox;!<t@Zmp0>=;w`UfEU$H`nS&N+n{^cv;7t1xi`}_MA18LW_%dMM6
zMq(f`3XR)uT8?}-SsO&=Dms?5&I-iVY~ANrR&C`9TV3RG7wFZE7aIC!bP*swr!f8o
z4Z_&S2ooE-rM(>k78_zdM3<0^?CHhXVNpc|BR_xqOpWV?evRiTr}M_>h**Jk1sfz2
zPMNiTH$tQwi04a2Mur9pLq?{#xS^z^Bq=G$VbXc~XnU?e-<0m*!-4LY|5D{mB2mQA
z`bfy}oKMQ-#oxG-Gi1<ZDWxMb2&-19LDQS-{DOkTZo5VZPft%At#dZr>VSxd2u({8
z<khm?G960vS5;MQPE|x;5#IaV2gwQ<-!5CzSmJ+={s=fMVZzNd!*U>B%G0QOWOZf=
z3gO;049|IAHF#eL!uyrTP>z`kzJ>h2`xt^r;iaYAaIwN-<;hDXXN0`0tVH_2{PMDi
z{cKm4T*HIvvNE0$)1Eu@^k{i%U7zjh-KWSJX}k_@!CPS-tN!<<y8ko>Vk4tXoB_`_
zV+7*h?C6j7cPmEucyUR|fJ=8-*&B!JBfsFMDl03a)!+SFyX8APJVg5X`u<+c%+82|
zogMK!wW!EQ1A9?AIutv*?V9S$q$Db+Cat|Sm6iV$9?6UsPWQC~Q_viQ%FD}3Erx=A
z|2Dr6$WxPZa^j{Eaw#=1`R~1`2il>?D=Mx+UO!uf?EIsxZE<T$q3-n|R18{S;Y7#*
zgY{k>E8Pi?rKQnZg7GR0M*i2Y;`w?UZ!yr*OJ3F5d+NEl@yo@s7|u8P>D0ZBfht2t
zL!<coIld2+yrOq>$Y0_R36p-Tr<a`3l{E_a^?A;S$r)9;2g(@Kgj;u|IjqMy)<<&8
zKB6+G4-c#A=;}Uw`t-v?WbKtiNm7vU^c3#s=!j*LLhx9P{+ODo`N6Ou+sP@djnH>r
zCW1=WT(ey-_v8uYU77H&K3A6v3X%V|%AqpprLc&|D<`MuG6z=4%%Kkn3E6Wfc?wB_
z8D}EttiFK|gndw=H0nJC$isrnlHG&gcKRXTc@$I=L=wZE$8V05PG9)(j?nG(>sc2Y
zT|?2aeBHYbAKvouxiXq*VlYV%>P5Akspf&4zPz{GC7-|<C>26HkgqHB=FJ<8TDOO;
zKO>N9kY#kE{aip$P^*$aDQhQ1%A(zP_7%YsJw0lx(cA?H979O+PQUy8ps+jaE=tXM
z|FuD;^J0^(tvt(rSVC~{kM^tR!<4`QM5WJ_cm3%KMJ&61IAl|(2_{FI6Wms#@;=X?
zL|KgG&&;SZ|J&!?z2+qLH7<Yn-S_DAJ$J9e4p@v*6n%*4gTN){o?pvMVLQI_FHSqB
zppO3@D@Yj_%T!2`ef~UhV*P<4EOTF?0OZX%8=r}ej(3oj+LolS{^6lL<I?LK^z;E%
zWBEivDT#^lEkU?lW%vJmZ~Z<W-x$lbF*hPv#Dh3YT4mEyLE$D`ooOuX-;H&4eSeXI
ziHX_R)a0jCW{KRXoSdBRbt}yJs4d6xdk)ry!e*@o(xkZUW+^>9JU;PQwXLT3*zTNZ
z+tmd&G>C<VhjUwxM<Vxe_b}_GnqH~-EqJ8H*;(r6_ncX@O5Z{n`Sttvlc;;oP5_(8
zAK*s!U)MQU*%&Km9~pUZrIS8@Pet_%lG!u0e~^z!48l!KZ#U?!qkW-L+-u@ZDA=Uj
z=J;Hu-Kh{+Uo$d_Y^KVMx+ZO)-k@AV=P+)=Nrp%#*gkB*6XS5-d%4n=(svC{xY6$#
z1}Tr_Z3P7dzy#OkH#fT>&~|or7j_m}Ldkh<n3<W)dhS2F@PcS;RTR7U0XqP9bmud#
z4LMxQ05)FtXTFFge+<X1X@TjPnXWS;A{7<{@l9{lY}K<x?1}O5=TN8s8anscvA}}9
z{eG8O@o`AU2J7DJRUO6q^gmSU<=3JS@hK^Nz9`oY9H~%IQE~D$OH6T*;R5T5L&@L&
zyY8EOLUXQW2`PsGB2_;8V!Pq$1hBy&*?N-e?0~*LTtm<2!dWd>!w?_g=;(L~*dtrF
zro8DQ7?+|AAQR$brv)!T$d#?eb$ew%hK6;-tS-rDE><mP#OJMu$I<-4!mB>JJIGhA
zbPEarvRcBOVB+FtnDr$K@BSc!GKx}YHO4lYr`--AVJmc(|L$G?p-e?ChovWgUNgi}
zQc~>tJ;;fOTBoZV1-#C<0iFR$Aeqwa?&~YiIJss%@;&CVC@PAOjg9S$b#5zWppue=
z#Oxy<zUu>a?2ig>RBY_`5M^N*%*e_X>QJb(KUedG*kgN8IcLOry7B`gM+km&fHv*0
z*(b`ZNDu&@(*Oss=+(xHdYu^<tL)7+z;zKUI+c<0^YcdAD20xz3g~#$-H>SEaay6S
z5HYLW-gxTl;J~yiV(VoWBb#BDRiloMV!t(Ag@hU7Q&U{7ThmjQi-6tam3=wx?A~X=
zsXCo*!c7#85!Xm}dBky`+H4s1W<M}k7q@GKH2^<NOF3ILV=JhPJ4RLtHs<f)Y{Do;
z#kRh_2*`&BSl=dyowdOXY(<g7+sx|uEDzCMySp#^`qcvXg;dzh{@r#7)c*7H^LYMO
zX_v5stDUQlV?H#AWlxltak=dnZ2tL^AmULzpbu9fKy+#A5fAUY38nA1pxp>^pr>ls
z-=5@WYU<zcHXp_KbcEjSLBqiWFI3e9^2!}S!B4Qm7uVK0@^xz}Umr5{eCAv2|MD0C
zDYMvp-&8_E0=b2l)p8prtKBM-Ai^QlwZLl0(0C!6kfW`ct}qHdWUm5yIZ@0f8J4jh
zGN||8KYU*OrQH6CDy38E)xNB`-i&S+Xy#?IJ6%r#LUSVZc2utTvwvEt(}`j4anhaE
zI#6Cwsah_3_xxq@>u+rzOi_T;TZl<QLSkfO6whUL6R_3E>FGnOQI^+->r$$!(U*dD
zc6PjWv+|IcV68a1Vp*A83RaUnZpz5WnD!=iP}tS_L13Yvpw#YlG9wQDOvNEv69Ai$
zQBjXmM7`p<EeN0=;u*^1QOKxqo-N^s{PgL4P7d>f2M=1&Z9ctM$TS|Q4+*$_W}qMR
z3Ws4D(}62mJU$lj>Y@FC`V)R@w;8WLsqUoQ9ypdf)fJ17i&cdtCMJHgwvy#)GqP;B
zT**VZLud89b?a7!cRD1CVYKBNz6D0TI|DTB8fDveuv?}UjeOz=MOB$h&Qo-8?d>qV
zslIe++l$-`&r@lh57f1P72#l=h?C43!!a(5Q>(>v{I5jDnBpVwDEXg%*Jt`;6q4rj
z#?xxi+N>wPazprHS&W5hAbGOa<(tQ~WAvSQjpZkGsO>%~%?8uFPtMf^F3(7($!`ZV
z)jzMm)gYU>5BNL)KPKJ@WzJ8a+7~5UN7ithujZ+2UxqiUOg#lzXhDr}RUHi?VQk??
zioFw6VtCN`=H~)|F{L(aUc74!3dJ~Ka8n{{sQGT4lh;{Lc5ZePAy5`Sa+2;Hva!wn
zZ0>3ex068|8lI}snNwkBq*Aa^)<Rdz4W2HAkEwjX_3m@<qZa(a4_D)X_-TSs2#wu%
zZp7A1O&nx#hs$$!2S-O-J?*;J&H~YPuSwW+41Pw^(cimgWMk9w^=XvbUI(Mx#orYi
zUj$b61kS~4zIMVvd*?#K-9nerT}S5QW1h$q6%M~868<BKmwGj)pTY=Fj#ckPa<5?>
zEng)ei@XLsLQU8N*H;?DJx`<P4RlM9=`s<;rak0)dwVB<Y$s~nOPYuoiIM;}I<9mt
zKu>85MH3lXFJHdYsAfs8)!BUa`pdPvjS+eV<_I&AKZ^x~>OI;4WmMd2C|;{gR*VMY
zcKGLF_Sx9(WINmBCl>VG)@WkF*Z(&7hg!MT7a(6O6B@vqJgu_DKoUKJgWm)ELPCx^
z^N*1IIyQT)ICQ1)yf&Yzta#ovHi|>u4K66K%g*O>#WwiTJ?*J??k(n%RQWzkt2H<C
z&2I|xg~Pv2f|a%UWL$0sB$Y1fD^^U*Cb<SB?a{;%$3d|<DuC3gJ`y*+Z-2?^o2i)c
zO)*6jF+Mr@@xuo{*^I>zb^Ya>f_hHPriTw71_TB|YVw!OF#0oD)|&q=H1w<Y#Tg_<
zLQ2YRcv;))b<lYQOQ%0@J}IG>(Y7MKX(OkRisw#F#a_VQ=ql8wRenE4Hi2r|>qh7V
zHv6A9k*<w3M88YzT`0~^iMzcJD`kO)gbUsJktrUl9@r3+&DNo|ldCH$(psN5Z>0iS
zqZDum@byJNuieqz&8>;@UyH(N^aB$ehxGa8VOG}tJ9q9t1B}wt)C7!4aDV^va2m1j
z%uGhaqeqXBrRVPqE(P!Wuk_MSx((h*Umgb++b>8&jp)r<=jEMP8{pf9^Lp0oQ$~4T
z#%++fhi;6H@5$|Esm@Qx3kWu#@bxmW`ZHbz><cYSJ{@-_aOuU+HZvc-{waXz7kW}1
zi*^mNT=tOtLG@lqFf*SjXNA3A3!D)W6kPpnxH5(mS_)61bI&hJm1#6r^D*=<2di)V
z{Q>(??XUJ>i06zfZftao7a5a@d3$DWeTHlmK`nv=C69w}9;iHvT}Ne;%^Z4YHuO|9
zL=>yB5MzVuw1`+k^F{w5X^G1G0Wp8oBLPlZvK9?S%=5=F;=lKObNPTrh>?Hs2bG%F
zniNnu!;$o-eQMWV$u6i5EHBSbUz#y*G(St^zePbo;j%g2^eaNFS7^IdC7fpA8*l;e
zX4?a>2oF|sb8~;s)(Hu@Zn2DrI0;h09!LiinQ_Ahn1Q9GrQa!;-@e7x*%PW{tJY4o
z(D2Dne7iiOk#k985a5bSP((LM&-Y#Em?hFWQn{H&oKNXXNHM?ky3!Mo|AVPrf2e8T
zaf1EeS4Qsj8#j4awW^WJNkxk5yS24-)~=?emSJndl>r11H#x4JUXX5(o}M11kV~%W
z@p!$L5Ok@fmLsw$Z4H07gaE6^)?TJZh<-05)CqF76n`e8t#`FcYh}zH^)aS>)qiT<
zzyt3uMjRhAv&po<3@<+CqlxsB;d=}?$+M^lPwnli9I}R=7o>B(`}*|$uq~7J5E48u
z&xiNO026@>MZjZZlrgQ3@k9dDau|4d)qF^?eGr8A<Y^fgq<nNL?VCbJW;Giy){E`U
zwKr{Y#~sd@9@o!G$(lk-#Ux_J?q<Lte8^hsj^6AO7;fCyLgPfve_T;4v%A`~qse^o
zhbQBZgS~4o1~x<19l$hF;^K&#H*XGqe-VOAkCl?sP{FfwYj^>4jiq>8T^#Qo9Q>G=
zNPs#RzWb}4@b>MO2X)m>Yt0Z&?#nUi_KuE7N-GeX1h#k+K-yO>F5h1LY6BvZW_edB
zjB^sT(r!)^SPp^L2lp#(5&~#7=l<lt-rLf0txW5ggv6Vs(|`c9+Pb=~WsyIdn}sD@
z3!=5lv|ke#=OS=Se|!|#6mj)uB;K6wk@-B3mtfXPMWR7ahRf|HDIg){yYb52A$?Gg
zv;y~IUrVx7r{SMe9rbZRjgrN_C1T)dr8516hn^9z5>m7)ACu}<IRrlPM~9M)b9TI~
zsHH_dRc^x^>kMVT8=!IgLu2$lu`6MO*!j*i8qXb+6rW20NRdXZ?}@sBaYM%?pFR=+
ze9#GPS_drHyZ7%GTS@d^0v&g-J(p53=lvN8q+b5S?@bYlNH-R+yQbavn?4e9D>;wl
z4`jO%7pGpNnF);9H}<xl^pUU+kQA4A0-ADxo=96LnQ^S`+ItSHAMxqx%4F4UeCk8d
z4@1H%g&wXi1RZ#!vK{dfRccpq19F&oqL8b8@8efJPR}}*{m+K|e5AwDSNh26`!0rm
zftK%@XleiJGCO^A?M!9eYG@QEOD$U9{z6O=nz1N>Im4k6jD{Zdq5Zu0#Riq2Q!2DH
zXYaC854gW&W)f0RSlLH38=C1>J3UGtD72a4MZhKj@)U=Z!`H6ig4h>$y7l4g7ND=z
z`oFa3dR@Qw8O>p(o6P%cjgg$s)>Az&TJ7~~!Ayl2o#*^)Y~kbM8qmOGs~1p15`%7N
zsN6<-vce9!O2>nqZzH<<f^6raTZZ+QLg6$nNxu$8aM}t)yBR4IIBHeAl5vk(Mh4mj
zO=hXI8O<l<1e^q)oE)#@E@IcruPZH&UnX#_eO2N!F@5!GYktXaZ5;`po&zWDv@se6
zIUb0)wTTi&Md4jETied^96+o{3NKfytTXk|jd`HlY7YkG&{vT+1nA%HZ!(n4K{4oF
zV0UyG@M|sT1poN)!{gN4#E2W=0v`(ebF2qqy2h2Aj*jkZ6`lzJ4^{HAg8=;|<<F5E
z^#oye&fVQzwCmS*p<gBbOscFq;!<wxSzsD^(?W7M8Jh*3wClk{w~kNZPu5Jg?|JNZ
zX)JvUYTQ3XcLmB`pFL)7Z@CrePNiU{riGzs^Cov=XWMc))u71o>J1OZ23n^i0gx&f
z6fzkMy~^9VvOseJel7-#&ucy2SnIwIIR6ejC~B%wx-@c=w}z0w&Ze=p-m0Tp9`%5V
z;qMq&m#dX`lsuPH8j9Ws;c$zZx)->I%B#!2Aw+XPQbpbF2Nu)lVygl=QXJ#9(1#s=
zU%co&;uSu&f80!y+h-S?zhlx(7OwZyq$9k<^SDpoHOBD5X)pzWz5en<`^0@2S4=kv
zfl~6y%8FEHdyyYBL2@23#$6cCTeLMz+r(~@yPyb}$*Hpk;J=8HeX(w<AqUysX?^$|
z^u!OLmxhi-1KQQpZZVCz*E|rA3)z{;M$4TsXE?*97AnAav_PM6csyBR79l0tQpZV-
zfaa<LHkVvtj8Xnm+O$pUVKwCTxguF=nalim!nA@jrEqU|qp9os^$sT$#pvQ`FYel+
z{?uO0yZART<m=a&Wh$OVzShL@XXI^6$cv=L!>Cue$xN~1v9oe#n0k4t8rhL8A{)ek
zfq{d~2|`6hMVCL5t?=G0fXerdL4E<*PXHg~kB*DXAsX7+TrQg$(8)?c-f`b)!aybv
zxN~4H`kPa}Z^8>GGH%}kDw)$_kbccJFF=)v2#+tjr{$UD{@k)Ev8_-`0S0@?em9Rx
za#c`+I&QK@C7stbFQ%)24{17{sXZLaKl}VjeC850Juh&6Tx3vg%%L9#$>JYuV|gbv
zRvYzQ62ov|TSrId?Ynn|y@>)}zkOR=U2TWP|My4^2^74u4amiSPPt&24R_|7ShUMS
zk?aHPYh?dlTvFn`o}EbwRUi8N@_EK;A!#!dAt52i7jo&+Vcohafyy7uP~t|Gmssj2
zsN45uNS^d59VxgJRhWuz{B^rmNqwpJqSk|ZEt&ugFuS9Wxc18nNeae{knx60PM7{x
z39boPT7s<a+D$WI)FQ9vVI%{jso8*_JTJCeqc4i}L~+I>M&tJ%ncuDv{w+0-AG^mg
z=asyDCD7hVp~*7wDwaX0cW|^X?9rc%fy@_4-$$4<ex?^WI8ovXSV$c)&t^Fh(Of+f
zK-6Blp@s1{<|%{Pi=DqnWN<GbA%R7s&=+9duq97d9J?9Nw?GRF3}(b63=lWsiJtfh
zIIi@=9s)S<{uwG6pKYS%jndNZ!>De|d1~$DDo<IaoWq~V5Du8f_z>xA=8?_i2=t+i
zmDmTGrj1CPe|b~nzp*eVn)=g+_~usU9_QtkPVd(6EonOjm@htZwn6Fc?7YLm5(;yh
z$fTraaDDgHBv)W%d^6;j9kEjf5N<Uu@?J0zne`?<G@Kw(e}U`!bWLDW$kqR1)wXbZ
z8uzBIJe|6u<Jswjv#ti;ks-F4`6IHY*MFh&HLjWeRH!=r)G4d!R{d0w0snP*q=_k9
zSHT(Amnfm5+V|JuVrPZ2;ln9KhKHM<XO=A9oB2?by_fkkypj<qZC)xk9AjZ6pz><Z
zH}z5HmkxE}WYH?U1WnJw`-yYelmkPw_V@=$>_Q)qJ&@{>U7qF7N!h@#$@jGcv3ngu
zzu$_+J~=3o99Cnh)!oeASZVa)p~~7=@l||=u}&<UuRZj|GMtD-)&}=b=z&ax>q1-A
z=U?cM`24#bMPXbtMB@YMOdI6*jT-*;?R?$wY)ePS(}ugj4^zC)EUHM<Z>f&3&2I|h
zsw)>)Io%EE>d@4Y?_C@WV_{=3a}W^t%=|!;p&Wnqdu&vmD9&h}CO-ZQW!CabIY4sL
zls3@fHEG$H1Sy)B%J>A*1pD#QZ23^0`7qEu$5djy?ZP)&picu68o9W;&)XOQC>?E8
z`$E$!4vh9?QC9V)%`o?leVU(4M^3p_sbK!MW%~yLQ6YTAm5GTx8B5)7BC`ToLv?V+
zI322DKXiQwdB8B68Cg=tV;f#R>{L2{(eml06g5e$+5tI<$Hx0Q4d;Chw{>j|vYRsA
zs7b`=EFa<Vv-WN<2AP_g$|7=UVW8=^IMqvL>_qVPx#|dUqesTj+LNBn)kW4XI}WnM
zaUV;nK75dKFlEc0ICZ_Cwqz#CzsY-MpwZGx&M3qaUOvh{@%>dP^NEGkGh87n%0I*W
z6I#E_bJ{w3cQVuSoKM5dgpY)BDZl43kEhWH*JBPJ*Smg&viedzUx!XWfDAGmB^4F(
z(N&?kA&f}6yX6T839~gz?m4dyKX7@LV2Tn!`tVJk=wBQ&vzAsrE>2ETAs3bq5_Y!W
z2&LU8h+4i=a@Gf$?(}BM9}*L~!b0yva5+Bh?N$lTeKW=hf9t%<HFltlJ2|EC0Hx?_
zs=0mrlYW*5_LG{wtnHrRYFp*rc)YwTX1I*sza+6=zw^Tg%MZr1&oYN@!R*1op%|v}
z;sDNIrq1^8ZQKLI_Yuo{ovJsPnVC(e0Qw=*b@Vm-jZNy5L`^k$xoQzTAD^I(f7|hC
z-#}!-XT`QWF-M-4D@#!yJ}a8$2Zbhn7UOc@IXlGddK~tK<2T%c+0njfC&qYW`>nX%
z-2E$r494F#V>VfaW?^y*#(zIIn3?IeG7iF!jyp^`oksK?OG{fDdfDX397Bc{`aKxz
zfZRYK^DZlEj{)r$7%%`cz)pN4kb}pAg6D|WvQ}ry#_Iua(fQJHrs3ay_{Jw_+t=e1
zWSXsRjLFOQbufxBrZ7LSNA*$As%wtpEvXVaWM|uY9aH3P+)>l=(nGr}1|q$%qHUsA
z`f5CbmS1p=fKqoeFI&kz2_jEHO8YLz5<-V94{ch4v*O=h^r*r>9I6ge%njsZqXFa%
zNQOZ3J>jBWwkp$R?#Q@i&MU`ao(V6<J)W;7t5L~GidrtB78CAYmzSbqsebB{Yz*-n
zy`X-vZJ0Nu?zzm9utZ6O{PyZ?y_ABv2m$r<5yfv++Dt^lg{%a3{$NXdP2yp@*(q;L
z;<0S@G`OS=?FnQLh4;=VS1@`1?Ic+3@~0EXu7H}qJvIf_Ha5Ha`!I)W-wf1x?}QRk
z<6ue1;)BjD&|k_YTy;Aa_ga=(&Rr3{h9{BTD$RzyJ?(4g3!DuaZ{@psnMWK%T+@^v
z%uKYQrpYxaEm|FozHm&SLAFDk+~_+%iJqLD8JWDOyE2GF+~1naHrLq2Su*qYVRy59
zpfQzjZd1I^eMT(h^_GR;zDmjdla9{2Jj@OAnR~T6KRauVvG^3qExn$zJP;=L*jl7D
zrUe43Hs%2Wb-MDEKg@$a0SkoMpC*MJghylB&!jQ(6mk1_>A+M&^a+;VGYP|a*JT3K
zZ7JJ&PL&#t`{V&Xx@(OJGF~ok=(BIO)sr0Ys!2aKHmoQMY3V77RB#zRYv;LP`l@S>
zPy*9XCjr$!ULMEL(6F+)dJwQR0sxSVg9B&7<r#B?@ZQ@~rV3gt-^Jvli5p$XLFX*y
zgM`UvH?6<7t*pU5w3s+WXMHafF>JK3Yg#DZ)zN-CCz3eaCszp{^HA)Mt*C1`GiLaa
zR7*Opf{SJ1AM1lvo4g#?h+zCQOC}Wr4kl)*aeKJuZY#;zga}O0;`Cm>12)|rq#64!
zY#)LC<!Rz6b}X){gu$;Zj-6m8Ley;-4%}sBY|w~vDpLEIE9^z$D&5<I#l0qQdQX3`
zUCqXXl=e>uMhCf6^H;|vMaG)X5mw_Jxz==&t;*jJzvt)0t*u#}edgui_c=d#1#`Ba
zot;cq@0k%;zRfQejpl<hqw&YHD1_^BCxaV3z6`DX>>XNfF?bE(D&YRrjM+-!8;N|@
zQ+5JUTxt?`J!<R-p^5JyrLvQ^OC`P^b{*fw{>XsVOYSl{uBrW$qym32Iogt{Q6f*f
zV&(ThdMoPuK)MVT4i3wlY+%-Va$!_|p8ylZrZsbohgq~}ouZ#yA7>~p8054ysXen<
z^`KDC9#}tM4PBGcCM&_#qB=Bu7;!-PtjUVc?B&KaGZ)d*>v<o^s>nGVc<3V?70V{F
zUYxy(kZKndBy!K)?q&Nu$vgU9({_sP1|1VsI*L7nRPFg*SE8Wvd!(*l2U=HrGO`X3
zI&=dKbud;AL;iE-IvKz;FgZ(PO!54vD0=#1aQW@}_ofyOH=^^Mr5>z1{;_hHrG?F2
znYvV4Hh)DS*(lVPySbKqh;2`c7`h>0B5l~n5^oTdIj~kCU#lKCMMLhSPJ-6)>1O%$
zIX@mADr{;Yrn#p*Z{G3q7EeduzX%#&n-^leN>k46Uvq*MgPVNFH6?x$#wh~M>$mIc
z>rHO>4qj*JY^|c&xgd;;L&2?S9O`Ns9T<G^?uKN}$OG1p3hXGUX5*lGiu0Az=FzTc
z&%CFWOcr@RKV7HWHPt9(b*T>ueRsP#t#kM~a=cUC+&KjUXQa9Zq92$Pct}hXZ@lIb
zu_ipl`y~N!cma=;2v7r)V`9D=$i#VA#h{t2ehQ`RVU046?%*L4Ait3Rvm|e__sl^p
z|I&Sdp0cv)a3N!<I1Hsi*-!3~TpGa*n?u%+K*J*YwUkfSzZ7vinVkAb`h>32e*p1&
zb=52u6**WCaJ&clf?u66%mAo^fX*b-0hL4tFo*9VA{apg;=aE!Q1A6uxuJcdprIGO
zK2c;G{4R%Nk^010Rbu`)f|*b<R^)o0q+ecC1~cIc`30A^Tdq}9Y0^1EeVqzy=Lrcp
z0j=^a`+-(6-)^IIYznh#QMt~9*;v^)e=am<)*Y3Y&Y0R#_BlJSfYG6GA^t%ZdxJax
zb9ps!7~n%c910~FCYHB7M3|V~Bh^YQ8shf$902g?KnwP(J>u8e8kRT=?oUr!Nxse$
z8ea1_+H9Jeqw(?a!SYR8%pOBY*3AoOESZYVFrr6$e?!_Pqq(Q!DF0}vw|9s%sJ%PQ
z@L|Oh4Q`ys4o!D}j1K!{I?uWmEVU*Whi<!Q)=U%7+Krj~SxoI7k7ct`8X=>)R_T55
z8nh5{a8=-2Pz_)aDTUo4^?fd4paF=lCyn{8fo%YH2~@E$u!{)uGA&OpFET(AaKjD1
zL_k8(2(>dpEAO6&NK$RBu%5>so_mVPI*W(@c*N;*e`K`<83oz3`<-KfB0Cp|pHHth
zDC~607UmePeGe!kg~*rA4_a!DH199gsAk)!an5Ah(AK8N=+aVkGSR}=@x!2|g(U)^
z+xYxB5C#*lHh+(|`P}^~&>Ml&1T3UkW*PJ7J`NrpqkupH&_`AB$NI%Dv1Gz2tNzeO
zW~*d!+Dw)jtui3qJGj1L_Z-v5@;jtfjSc;Pe_t|zXy(g`ZFTN;Yo6|CMLF-NQ|-j0
z#?Sz9#habq*6exVY3I=v1V16?XjVCx6z$)Y4mE%dZdvW666hq%xyfs5YjJu#+}&So
zju$z&_kr{U|JJQ9K6>M*_a&=p3RrP8a{?Oog`#24`=B8k?t79MMtSAmopeW;h|cCU
zhKgwK$L@@m%@=7mxtn=VXwE536Q!dRLR$;EkU>zY`R^AM6`d8i#`M6DbjvO5pYY_k
zJS!L9^-n!7hZC;n6ix58YV3P=mzV%?y6zvd(qVOB<ce{{e+E@1**2ncbuI$pn`1iw
z23MyMn&g|ro_W1svM;J=>QP$zp0CM~JkYbdGT!znTuPh9<}m6dJ-?cr^tE|V|G0rv
z3wkdc=)q`HLs1e%y{aE!Yr60K-vu#tCBw7<6C*a)9*;G<Ud%U(b)tP}%4yUKqQz8F
z0tj;v&U{=I^T!q7%UOSXVoZBj==7bu2)=9SO~39m6>*r)KC@@B)@8woU_kG<!y3tU
zG^)|Hxz5yNj!1+FUvF<Ok_&{zMa;uw6@)YYdcgN3Wn~#qiX>%ZHrf%EWm;+J>114{
zp_-QX&3D$zgT9EhRKU%!nY{T~BIZ|@1DjhSK0c|8T+xgtx!1(Xxw~hKIG{nt_BK{f
zU|tv>C$9fe7!JezZ>vibXzh+Bxnr%fT5~*qw7v>W(moNKm|uUsRO=sx)_$W62*$L*
z!9j$*3$GWS>OaCQ2{rm1BQ!-1fSIbSsZjxL5fpWpczCOj4Zec54LQp|4#OZ)jEt{!
z`ER1#VOe)z@%-x(nS|p!)f_nj*%JSlUNL<n_y@EuYOT3x5so2FO3aTq8Mq=!6Folb
z{;{^xHF9+-qFu`hkl~ZT3je13O|vl&S^xp>bAANi?sSNpd#skAHVg;|xGh8j%o2zn
z5b<I@_1<*9Q&VYIdXwbr?F*l7aDbq;u#y=g20z~U@fx+a3lINrsnZK!aPKLmGw?7+
zP*jAP)GB=)Df&w5UeLA2Kc&e$(_ay8!$S9j*d`%5u+a`Mh`gNKuZ%c7qtm@*^RmuT
zw=%6B&|ko0>BmUf>rx5OMiDi~Jl?U<XFxkE4WhH6M}txtc-~Chu6?h0`lqU?8lzbZ
zK0Wjxb?x&SmH3sPRhk6DD!Nl$SUI~U^HN6!Mw-yIH@=flcHKr`0Lcg9j(9#hYJfRt
z1|U&HmSbT1ii}!s{0DiY0&n?{#vTRcZp`nCKf+}BW$p-2lYWm!OSxn}hPd@jQ7GSg
zCry%|c5VNcmM+w;)YGPkQF&)8dqUxSE$N?su|68?7&V~Vc~6PFWKv?HMuWEqp2$ID
z(<zX(yT8*TfOl8ABvMBZAfy5bd?YQfx@*y$-$YL|BVV9If&KRS$Le(KcHUWXdb@bQ
z|C8TC*c((zJ*v&)%De|daa~3&hrHbGc9!Bh0pTgTx#A46)}CjV&9|EG)NTjwI2Ur`
zz$nG|T5#@l8CS>|0r*637G_5>=G1l5M?S=QIkFfpEX6I|$qQi+QRq;9V+@^A_cXHQ
zj>*=O(dJ1Jxj=|$MPL0GO#^-lDJ_MfQx5}`i%%ZFj@ItBQOJMd?y$aogU)tg1hv_R
z71IwRhPdfV!xPNg5^p_H%aw^rdt+=c$nw-=v|Q?ohZUa6?cUZrs4;hw@92E2jZaBs
z-Pu62qlT{Ah$p<`+jMV~Q85`b%)v0)N3n-_@8;&FIi#<u&Feb1o{|)16%s@+(Q*7@
za=U(i16^>W7y=EGPWDj_JWhJ(tK2`<!RdJeyj319?QHYSt+9dlF}`N|pJ$q|g6+!D
zWRZ7uqpwpo(5mUqMC|kLfvGuM7AUugW=9AM4<0^IMD)etUt;sSYg9AAe1}N#sphxE
zUQJ?&XayUe42HrW68g)WQp|7qW&bYA#F38D-`RPiOv~l~T5H;}3&U}{=PeWjS5Zi|
zW3fCmoL{HJG^}YKs@hFT{>H%!xddJtttygx0r&4l+>^!BIV05(`F@UbMP%uwRn3XK
zrhGy>f(?i}m|ue8e0A@8wgl+t(gv$s$YX`kE#jKevk6d>dTQ#UK`->ah29t;<n&wI
z$*Hs>1k%!zJdom1RSpM!{Us}BpjDfZ17aB^zvu@DQxlViu3I|mFe%nM#H5pdM1#8R
z|CR7M>zjYkc0V*6#mUi;OXIc{i0ooxgVNLQf>OH<x<`rWYrsQd>X4(4DKvi+?X)(G
z6O7H52c9SpS^t4F_eYqUm`Fi^$GUklU(Exhi&R``ZH6CPu9<dOxRRtC=j^`)HC0xQ
z{ScVCEsmO+MPQq#QthYG%gpM(u+jJ;$N}ch&d<L+r3EP#Bgm91F5O?gq=!kZ@P0QJ
z;N|r593qMM@XslOd+H_BI%4Tg7@0c%y?&M-<_OPUVSG=XJ-a>+gDjAz_Sd-DEiNvC
z9)&~OPXd$h)xOFoR_+=G-{K94c_}${VC<SvA*L%U`L#;TQ5vB~mDJE6rSU$Y`|#ld
zas+0|{2bFzf@k9aU4PzX4}QF7)8Gpj`f+f<J0Bg51;PXY+CIs?WDyj5SJ$x$9LPss
zKGk2RD}2<1_CYk#bN0WVznhM-0|o$?MZbaBZv<#<ku*L+A(6ky>v+bj9Jcy$<6-@0
zI*!vs0#v)il7FGq4UYsW8qv|wQ5bW)XRkkzXgJvib$t^|F)hncpV7fkCNdHqZvGJr
z0%7<a%k~tr%P|kT#Ya6h=&5KyH$f~i2uCE!G|<0+vauG10>bAzEuh$@0xeB$URb`L
zjd-lztrvRhCJg7gVz~MB@M<g0V)JE@bEx>2?DL%~5CW}jZ9%~_-lWglIAVFvv6NUV
zPwvk&m61R0yngZxPbcEP8}9tjdhLH&*prhJpdK`8Tx^Q0?%chbt=}MebaaG<fl&gQ
zQ{}m<rH&}16{4_^L#x8J2R_=oc|9=ej?Vh+fW~`~uZvQpc0&J)TLl(s1TasSI5_m2
zoW(!J($mu5jT&@DKiHgWNP*EDOopDy%Ma_raP!B-@tpo<C8(HBPfx9<%EO?iN=y=R
zEq?V&n%8#PJU@a_a(Q7hFF+jc>NDNWq{Y7=NTXEEdg#j(^B!ba{RR4B7gh7~1|X4c
zKR?|^>9;HcN`Vu2|F#GkpLP!rfJFfgEC$4j09g^%eQe!0WUd1J)5SihUIEsc;G|}W
z4id)f#5qwWdPP6<D<jwZmrGpN^4k!}(C>j?0(Ae^eFgT7Y;D;9EvLh(oH@cg+i80?
z32r!cof>u$=u<y{cO-Sh5(dm)p!bk2xPp>l@9aE61AW->xd5=VmTpX63S79J)a0_<
zigN1wFW~4P;NS`1)2IOUzdqjT0knW16>y+0s5=5SNbY~rKZn5EQl4p<ocx3*b{;81
z?H&NLWu&u#+GF!!`j}!ohab|%@@nDD=11D^U&xd3>u*}}1(z~kr`!4Z?=2bL0gwmM
zkoO>mS?x<{9~e;b@dhdJ`RQtk>4RQX$mFdHEx}0THo%%Sn}V{Ek`{1Z%!4ZKyOOpA
z)`2zF+OBNH&IsXup@#)CehOY|1CaW+x?Di^$6D=8Fe}o)!@~n+9}5c$eU*@aVCU$F
zR}NP~)k=XH|FWwGzfx5K)b*;@rLP%CO9wsb(*FVyQDZYR^?WI!;{OZa^;!JGcWdKC
z>45LjCLQ;Dcl4V!yq_pOul6sXuK2h6w)_VENh0_ERLrfX<ft;0YO#W%la`+T$II(q
z{|#8>fWAnzo+758=sMl+TseRdB2rWc(>8~L-%5I(yHX&81=V|Hhb6d;{8U&O&|YJR
zOQ#W_5=Q(_M0EV-<bdwti*|)A$Ur`U<ZcK?c9$2&`uX||y+8n24FBZrlCYksF8d))
z_CEm3M1K(!fdv2Qvte5^@biz;Z~HJ76D8tHz0?b~yTV(<^D{Q=GQj>9;ETMVJFC#e
z;9y_~Jl$VKDu@P$hAcqs40P5v7}A+CPY#09=~YPg#*VI7rNeV9KlUwzS%J{yCm5@A
z-FHOX2L9*1(kQOR_=^d){5e!i@|SDxK{`mP^er$IVze88w1B(v=p})~baNA~C65~Q
zMrmxG6eRQnGF)G{&I*jz(=CR+m0FF31Erih9Kw#&^g#XED93)ZJy8-2^wR*OY)BWo
zhlfZ}DSC-NU?BL2Wry}U-;0Y2D4gVs3W)|Vegl#Ei&%qO61VAE-U3NiT*THR1rse(
zaI{EDNqqo^6o{iWXPvR=n1o|8u;F7~#X+toWK;l|t5~7usRNXT-;-s`kcrJ;C<WQ5
zC6VW!g8`Ee29Te5tfXP|N-FZ28|nhLo?dZp9DW*F@kc1Rs)7*hAg2Jk#ybJ4e;xyT
zQv+b}1H|Zl*kr%NK?b232>n2%lF0uG6@nJLInQTS@J=8e<2+u$GePY+z(JkgoG1aw
z#i&@lF4A!p7>WYmq_d3{5fmJ}L4!xZ+dW-|{8<T1E)GHT=inTOkJs2dFS8m0@%^$^
zG=3UqBsD5v(%evV@uUwKR1OXfg=J-|6>pHgLnnF4RF2aPtSHL!Y*i3a|1=md0|`H{
zjDYUB^aEk>pN%3fDVwQ*NEGqlO)W*RX%sf5){H@UJ3HJcF=#^Fe*q-~K1}|*1|l>h
z<QnXN0GN{gT3InY0tI-CZ!Q5X{1H2lBwlVkVbm-5@23M2NR!a^PJTsjd7T{?t(tYm
zHz8k0J}AzaK2(xE38`=*f)DL%G)II&a&z6DZS6e<*GPPPyaTH&rV`e$W$w9f@eR{C
zJh8Aha(`L0qaaq4ACe=6hKA&HwvZKRczD=>wR70CFZmhBJ&`UMP*ZVL{|LiR!>Lvb
zWQ_B*<1pW!-;vcACn7%dy_@qW%53chgcPYEHv%0E4WEeU#pd~YB4%??QCAV0$f_6U
znE*Y3G@gJ}q6CyLfWBzt;us{C%`$>wFASH^Nz1Nl;i559R8~@`;Rg4i8ct<KKKL7i
zO(!V@bDW=`KSV09k=i+sK_Q@>EP+t-KHrroAu_)KRq3Uf*|o^HxL}xJ1Tqt(z0aN^
zNB6ugG8<SI5_9*EKvL1a!BjjallH`C{)PMHsuIi|I6(slyjZSoO)ThBcL(LzcPFh1
zUV*s)2|hqC2Henp#PR<})op-&uJ-b5V>m}0kA{Y(+I3suAG6eepN45>hv`nCCxCiU
zkJwg=Ou7Cyu9<|&&_bX8lF_^@HeVI)g9q*d0dMmQr~;R@{{PZZ8`^tFX1A}8{ip^q
zI?^Qo7Nb{3o0{MuJc!buc3=A7aG?W9oc9R-00-^n8%H67F%tVtzPEqn1g0q@{*pvn
zZR)b<=E0^CVw$lrm4td*VmZPD<Vg$AaN#vILO|oRgN_LtAD@71LKbq;QVw(sj3H1T
zgOVT5S1qS?Y%C6@^PJx2CqlFPz_PxCz7`CAs6d^`gK`c05T@_OR+pAQ4xH|j4AS3%
z0(Ox8R#yJhSCyD9qgXZQlNi5ZeVKWG)jDax&y4QGo#8el434V*%r3aap}?Gi25ez_
zyAMnc3BV?TvJR1}SN93%hWL#6|Accg$6GUp|GRJwKm!no7$6n~8VDR@vM@Fd)4Ku{
z{=)kDcQN?(_NgN#9fD8n_o=8PWuRyYXu}+aqDPym@(d@2>`ikvS;0GCYQT+<mXt*L
za1@e--__OCU3zZM*7cO%G*DJ1AP+-lXJ_}{2Z`K|_Bh*x31FR(5SiL-9tC_GQV<Hi
zrz#2tHjsWMS=slOmG%qGB3|fpKUl4kQfC}@2HD(>7|o}z^*6l}o1eP!XQvMElA}fH
z(QeR5#+W$a%)?L>X*vi=0{@t8c8j{yZW9I#=p<;O*R?=n0DW9%)g3xCHMRcA8VCBM
z09}2FP0lChc-~%Qyw3u8TVe`-Tb|J|LbsAcW`j~%Qo{VIX%$lTbI8OFtluNSl?fUd
z(3|=10|apM@c5j8P8UGt#_cDXBL2fimhV;1>qzHomzTF!&>ACJ7Mg9>s6jA62w4L_
zRu|G<q?cU$sR6XCfbD7zN4147PksSFfounR2}~q;Vwr%p!o$BoF*415glQ;!lV<@~
zB{pJoZ#-{2KZR>o`9TF9_l3eJwXjaB{%s5<9$@o9x(T2FxYws@YQ%uxkSaP*>&8_J
zm2pF{eHRu6f<R(l=rxRtjRkxzg;PVNKoLkM7ss{?!-(3G#Sjn<b9kQEM~#5<)*yGd
zUl=Fo3Z32BKh^A!4MyB?%es<m&b&Ss?Zx*DW+KShYvo;yh;x9vl!8ug4-T9FN!xJT
z9lMPTpeIkB6zzjXj775;%@=Au7X<m@uU}X<ZWsZA1swjl^ro@wCqdSLLf?a&&ZS5Z
zY`QuFjkl9GQIi91UR~Qpgv2sJC;H;4=1Mq9VKq3TC9g-NU|;10`qV$yLc=~4`HI%`
ztuBUG63UbK1Jrq_GVYKq!A@ufSE#7d1m6X?S~DiJ7FndgcuN$mfI)J6|ItoEN=(t4
zRC!v9D`8;{Nir2=gECEZ%>Yz3;~>TY?_L^+{NMw?W(2sm_G~?;J5iv-lNmYvl2Ao`
z5NFVLCnJP^quQaj?dgOi+sSp4*dj;{D)+WgXDeUb?GxU|Q7=%N@f!!wk|nuU91YGg
z;`<v|#Fo=QPT3CQsxozrz|#>M>>=X8e+xc<Rv^}_i|;U#BQGH#NcE!??(~F4yott<
z`cgK;V4l`Q2E<xf=$({^^z?LaS@41@_j)RTa-`<_UJ91Z2bcw>eXQ|Dck#N8QLlbP
zjb7*)IQkeKmS;n=ToinWW(+1|=^F@SOD}x(I~~GfA8Fr(|3J4P0Xr3_E^h4sfGUcR
z_X;9nbfu}2@`d{p{HXTlOT^!jqgr6O)<?jj(t(?6yJlMydXhenJ>vArNK5Z61k*UI
z4KkG6Hn$iA%9tp9q`3pPa08pa<b}Xk7PD}TRvX3*y6;~(%MkY%8TX-+AQqY{F;i4l
zR_-j2h5J*4jd`F%dd3wxnags|>ophU<-lMJG`thjF6{G=RBsDi?e~cx{l+m?N-(k{
zcAgSn<%@3j2@ejnip`V31j@MjcT`%4ScUn(8>HF>W^IE-#xe+G9WA$;>$NJVGo58Z
zPF00ZI4L#rSSU&ssD57sA`Lxn9n`Jqek)&mlx+og+nJa3Ua}2wLkU-B99RDs%}I95
zzJ)C|{|O+GZ(muA`%y_@bo8MafK;02)Z-dtdOkTMa<gC@=l{nv+~<>SM-a=(%HDbK
zpo0<#qn4dM<MfK!-@SSH$#%^Wq$oNuq==}fC^7(f)YQ~&V737(J|r0s$$>s%qMHHI
zJK1=Rc#2|l`_!t$xnbAnU6B0#JR2nh&toR!2%T!DcOb7Cs&#inT>jk_L;Cf>qbcOF
z+1S(5v)KxnvC0z6Ho;8;`U`QdcBbzYDr+N~L5?aTlpSz}w?gUx2>NSz`Nd`#<j=jW
znr%*xBkK~P2awQ*tnR_km2EqC2K!6P6^rf%@UQC>vCRVj1eZg~ZIBnw&dzqnKTQ8|
zhytrBYoC2>CX*>5eJ*ebo0t_|Nay1sgxhi_kyg32#R17y!<9Frfq&k<!YtmFriyDz
zCXIFlk2-`*;Z`$VupnN$x!FT;PH(^`AqlRj@o*od55SK#F}ZG-;unLPsDpQvSLP%I
zFJHM8)xg~RKBQ-QKE8_aRv6R(5t#i$<PGvhcCFFzWbQ}!_z)J`YLrvM1Jn5@i6hwb
zYM%glhtmR#!Oj3iN1P;ub{L7Ap6~ZTu{4BD0PVKcchi!ON#X%nZF`RMAqw_?=u2+~
z*E?>X3qfgUijFpUUxt_kP1z<i<KWikFl<Hxuqp$N$plbLB0a{C?umHM54Y!N`1tse
z-4@>&6+kG4!_f=KISBmDOTTpht>I-eJ^d9v4-zk?ih&Rdcz4?hg=M`hi~<bt1itX6
z`bm)w)G`J*h5*{SVn7`ir|0LMK)E!5-x@k`<j??Ecz>v%GrP&K?*it>;tb~qvPRG@
zkNL*OwX5&I&MH$bLl7^6a{=zlVy@mR5YGnz0==e*iK_8aBpp<2A_uIT9OO)-jt_8B
zp*Wy(+3dgESqm9X*RL+JGjiaB3ZOLWe4jDgyB7q^3HX@-)&ndoALjwW3QlIA0edA_
z2PWz~DxS{%1sm&dp0@f`Ez)05?bJbkOe~>#?IzC`^`^_lDbc_m>Hi{RD`o-}+GXTy
z5#Uhl$!k^B+tA4(9k3UFj|oXhpTJv&M0{oiIPRaUlq!y0pLeVs@{l^~nV%1P|H=d1
zX05z`d$>l;nz?dsFh`vodVZ6GLuj6&LB0AA3L64Om5rc0ZUgH)co~EdFpP3NoD~%n
zMQ8N}S#}x>HW>7f{su5}|8rBP9;9H?eXk16Y%FS+9~RqDr!U2-KnC9{nLx5OgUR4t
zRKyJUNe}uQ9UPo2%O>X&6*7l#sRUKWD(vPsXFQ-Ss0OGvJ+A{IuY4%U)u3KlhiQw`
z<>h7d*RNj*e0+dO5){8FuoU?i`M-2oDsU9_uTqsjh$;MoP2a{ZTmHvaFnqiQZU;Ej
zfgYTV$TL2WXI+F-2_daVFfzu(!U7h0Z{@Qcc@l~+L?tkO2YqE)tjKHuoJyAJ_nN=H
z%lH2qWwPm1(w0PTfAEa^#xR$;_S4lua8LJBIQ&y(b}r+etOyri`bE`^TZm^S012h_
zH>N7sVU_BBi2#-=EGgmQR|nU+2W#GR72s4K-+OoFGAufd|IZJ)JD0H*z3RHQzHYfQ
zZ5<_U+y4L1?X#-J{A(3QROwJt^YU0>z)1)8W*>!|5dtvPwue)(Y9##k6UOCi^MEse
zFv#$sCLxgl`U5&#=q4gyBmf=(3-Gu6nW_L~qX1x;-)7fCJ_4kDXwn&NG&7pExm3f=
zdMf%%re1F7!}g#P{F6$~Gp=IhMsM{0Kgg0U{EslOY_Ue3O#yADDh`nmve|}zaM1j}
z9P(no0E&wL{g4BSI+um!Kw%(3!LGUhI}*^iorjmRO6+!V7Wfr`wFK+5{wcA>yXeCj
zc?AVi$WuUWU;utSzqmLBxYeU-)2cqY?Ff$v=R4d6Ng)7q_rnqOj^18gl`<lg|64Qm
zhswtcdJW)r$l*gkS)ZE{eZTY4D<2vI7>@(EwE;oz@A0C$zz*;!77<O)oY!cYL#hD9
zG19Uz_zCQ2_W#?N3yKCHIQ#MiTyL>%7eMEefin}1h%kgH7Ua)CsOoU2!poIT=Hz1U
zxIpB26-dDXFdx?6WSQ&XYjC~-JeYrpSsyZ~fhnpNjxEU6uE<YdE{N|u^Fodvfa2hS
zBOb!3MLOa9l4p<?VE2~84`=|d^*@BH>R5Yg>j1FXU=Mv876zUYy#+WZC8}RJHFwSo
zRzl?U!90xi%^>T+6cL}8xDBRmomGSH3v%a7zAP=h47y1rTVXpRuOF?TY7XHu2BKjt
zq^J8gBseEHgFz|VTTDz$e|2*+SG(HD!aYXAV!FpQAtj~219}T*I51-niuT6k_SIz{
zPK=tT$pu@($s)Ecc#)jpF;|$1lpkyygMr0d7fk=Sy;72rzXhGwRr8(I?)lmR6H-yP
z(!)15TxLBnI6n;&M`u{sWev?`)lb_c^iWOuA;<R|Tz_F?a>@o{!z>R_^g5}0lOKSC
zLOzw=v_Q#i1WV#Xg&hr~^0AbX;^H*mbD=nus2>09?xtsF$8&W(-eW3JuAJK5Ll*s^
zp(xPke@gF`U9he1()T_Mge1ZVcL*8)WQsinl4`)5dxfc{x%%_=K6FqoVg1f*7rD%O
zci@<tA?VvmhIL-RHHPLrFu=9UdYDqMcmy1!g@wh)<#tNVAv}^jsJ_7$uFQ7KH>Ch7
z`i~z!_M3A9&LK-Y-rHLQ6HrEVJJQw?5usQlw)FjT%_$pP-LC%f%nI6+p67<kulVMc
z0ga7=q8tnpEAS?fhfN^87E8;^F;g_QF|9M6DM-i-A_n9!EFk#X+uuhX4W$03#<z<}
zRJ8ubwg?YT%zvi_v8ZZ}>;QC$Oiu1qQ4}#*Q)Ws67c{ixZg5sbP?__oBo8cTv*_iC
zF__hxAiEb<R8%+##K?l*2OMvPuzlU!+%&2aWfj312X<wE?^58mCxv5r0F1nZ<3(cG
z@{}j5otcrN9Ha>d=m9f0nd2Nh?@04qEZZ|oak5N7IJ8AAU*{cAF&SOis^E@6J_h93
zmbtlx=~)>q1T)9~gKw_?v+WJg%M->!){~|0AgRZgCB#T*Hgo}sa|Gbj+}wPy3hw~{
zV;f2_??kW*#MliG4bPl=Xl5!<{daE0nIka0jH05+KnZ{%va4+6{ibbwAK0cJg8L}h
zV@m)50`;Fif6fDeZnpU2orY$|)l^(iMTP6TxtfrWkkZQ;f5=ul(~f<+;LTI}JDXVp
z6T7zAIu8N+c?1Ahl;{Q@pf%twtj7!ez|uXt|MzDK0a5(_YmN`QoDBt6=yu)l2_Sq(
z<aq+Jc{`nFy1Z7Sp}>WKJs?L6BtAJQtiLxlta|sDqiar&i-C^<{hBOI8Vby^9)b4f
zyYq>E&0kS2F0PQk+GhK_s2LHtJiy85b3%;(r|kZ&=2X@#cg1l59)LqH6u@`QqN|Xx
zw796ext(%#vF(FAG6$SS0Arr*OW#XUz9Vl7-U~s<oXAtCoWRqTvIUDWHj=BULPA4R
zFtp}WP4_tHH?%?^(1f)ZotT_lhxd|E<FNex@`;>W$*@co2RzEGtSQQz>ge_!^!|cy
zZo>?8PEPF+G;sJ%g~ICES~k)!$9i~+DZ1D@1-ZQigoHNWCi8_HS3+$en^eQQTuXkv
z{y@$v;07Hl>oeE+MwA&yXHLFbLyBS}RQ&ehU|bPM0uT~LBc{g(@ABdh%HaSU7_vU=
z_6v6gpsq7uHNi5iyuYx}|LbbtSo80loYW#ABeU2!{wIAFl?YUVdsSP0h^>KYz`*f;
z6!#|ZT(*0=_uo7ZDYHzKN=SytkP;1uNHS#1lsO?J$y|sM8KWdJt4JlG$(VVFh>%1j
zbB4Uf<-VWi+3VSBujhI9XYco2pS9NAecy)be_iKwp1<KZzCYN^Ou=>W01)OG`U2xb
zMz!LIN<$~H`K2xW{yi38_3q~O;|=ECRQmDwPBs!#-0bs@P^d}B%NybYXE}Bk-T49Z
zQuD~@X!6r<S98Ta*&#4;femc!vw<JriN?}1Xg`mfA9~9)&1rU-aBI9)S*3sZd~(vw
z6z5CCiZ+lUD47^^6<HqWll$a%dhXed9NaTpOvU%vjt|7~rSC(zFHU>w3KelUvvyn|
zD%<!|_CFHU@Q-r1|H;9<cBX@o^w&9mH|hGn2DAP9kI_By9kDq2b2m~_oW}GJ;jbaz
zvhNWdR^S~7oLk)*1`NgVzka;d06Ovz)sdq(MhFc6;zEqr`Le=u+yFvqa2SSh#|EHy
zcm=K-Lg}-(TUPVEa1KT9UEBsk9biVSPtVyQy)C$O5&x2qM37|cn@MpRA*eK=#6R-K
z2qJ{E5~2?gUMC?WbQ!fbL>gQ$WFeLN0!f<1aLsj?c~DCz{93qzx_R-#ho7OJ-in?L
zgh-O0JD_@-`O<d!OXCdc<S17W_$Byn7Z7d|jSQq!8Hf!DdmpwzqQim0coec(5`M+*
z`FETsY=~lMKJq}vLN`%){^Pl2q^x{<{SuJ8j5V?F$b=I-C46Fn;(_zrzQ24AVsse5
z$t58hF!TRy!z{430U=Pdb{5ggA+Rm@7K%6xs(mPXiJO$(LyV36hAaS5)nK`P?;sSs
zzI7G4&NYb#XP4h&Fa-STvUb)GkWg^RAo)9_r+4EI&aW-_oJcpu3kA?G+q+CgL}(;9
z3#kLJ8K9b3E8V2@3Lj)<zOJSw9Vr>id}xd|E(wLOp19^g#vt@snSfRxiS$D#M*MG!
z8M-9|WsHQqU)no=avAPA62Z?Qp)z*xlQdkps~3<=$PvrpwemKns0(u+?4m+Tv3vYN
z*PATeAXl^ezMG8p;0k)ZM~)pkcBQK`TX)Dz164(0*`D1A!c7$7a0sShDCK=oAW7;!
zf=VjM?d;j7-&aU03;U4%?SN_#+%}M(rtxH_u~(eM<F+a^f8krAHSgGVp9zT{LGuw(
zR$eKcM>0bm6OtanY9k`P29qNWL-bZup`gUiz6+NhI{2x$U60PRiJ(zv_U!@~S;h7(
z!UPQiNe?$rTd1;+U}@w1r9!XYe`WUMKDr1Jmc}qxL2uDMK?-?M5^v?{zdk<52381$
z%aPU)C2OK}j&;$3%oG>^voA}Y;059hpX$HxbJwLmekf|fmmval7^j){{!6AP(Gtf8
zVYVmK8^-zZc=WhO;fgmr|0bPruiXYLzW?AW=BSfSJ#pEH04o!RAJL+M^p^<j4fW#$
zD8RhYL3q~<N*&%af`bLA7S}`<N+U##SpPRi<^PmC-}2uC3?JUJZ`su}m7vApKnAa8
zf2LkBULdP3<hXi(;b5zSP{kG>c(DuUK`nSmXPX$h+SZ6^*#)A0q=uRrK}dda2(y_-
zU%g6*H?|+YUj!pAIRynBR_s48&mf31d*R!Hyf>rp(&AVa*g#j&@unz32aVvLaO^Z|
zfaVm6sa^KJo_mb3A<~ueeF@nBT6X_qj`;BBeIlU|-H3Ms5f8FsLnEUx9Rr@VP<3$>
zWfq(s&Tcqk#t=H^ASEm${*1@xf%V;KTJEiTD?a}I^!$7mUMlX3(YWfbq<_nA4_p)V
zMBRCR+IdkwhJo*!w`{r7Vm~=GB_;oJWaJGRB&4F!Qd?BB5v=3=@|kXwb5*49t`4Sy
zq6TjMd-sk(jfm7WAR<EZw;$`;t#-fhSGxq_%EfbW3Z!@gM5|-HPr#xe8vXCz7j*kz
z4gR&!XFtBmY?C|rFO$(X1`jCybu#+q#NhJ)v;v_?zxl`K)G0n}RDVk)@uVP?%!0t~
zvFoSX(459xzkVl;oq#LR3NfM^4f%!L>+p31ClK#T;v*4?{{W!{9{lfs(8Td@z$WC0
zdfS?*@90p%<LEdpC4SP*W9I;VUlHU`*ZB6R{sYP^yAJ3?FemX$)I}~6Yn%#6!&CPW
zdI<apC>k(WM4oA4(rZ;1chd&VYE>d`t-Y~D%)Ejx>fkG42Mk}g$sSAfw_hOr8pD*<
z@^Eg4`_%9es3NdCzQw7$TeFB2mj_`9+L`JZEzRgjCPyf9mgSNRw}H1wyx33reIuYp
zLySp4G7;YuwyrjBS!11s!XkPlBH~q36AQ>WFv|r%<^+%UA=Kv)iOYHC-xE9c?uej|
zj}IY|N82nE%dYj#0h)u?_-f%$_F4M=<iq1L>#+^(NYg>R6mq-q&Ug{aJvJ+LXuSS_
zY=`(RM1iMmY~H{>OH|X5hP9cuy}Wb@vX%4m9VoowX<=XG#6kNG=78UyBX*YIwMp}v
zy0RajsOEGsl$Jc6a{jd!^oKI>9z-iNvSKeXI~1xK40L=zFYrbN1pBeeQauP>ept*R
zI|*eIn5F{gW<p*Ij%I%MZ=|6-yPXV)?lk_L{vo7|wmt$7(ctjjA~BH=8~Y07zq-GX
zXj4SWjC<$i=9Z45g3#<lMn+n7eTRqvmyeFON5(u#gzFIGzeCUfxYsX)ti0uaI!#>B
z9gK#q4%88Alp>Ktjlmm4Fwd@6F`9wRh~53`r@x?RXIg`i8op8cw2#iTQr)!znj=8<
zW}9Srm|$8)z3-DX74c#2P9X@k9x<{##7e{K`QX+)M85QdX&FEQ0&D>GA#$s?`MIC}
zui(yfb2}ENkv1Xe;6|q=7Cn(&2v@LXi2AY7%ez)rsVb|hqDA90%>9G_6hT1HgthbI
zcT`Bo@V!^FZ5E^Ca|IToj1Ihi39@d~*R?fUO<|4BU-}vP+5W=2xbLJbq%Q!9T2zu&
zPK5yo5Cq<y5#bv*!i|R8R>rpw8=WMGP+-51w#A-aMLE7=c}nwQs4w#MnJ!dy(on`o
zMcTwgN$CA6R_07_d^@|X0?)b2zFCm8cX{#vG|lxm%6(>94r$IBxn^}g$?R~L6|jQH
z4{7Je>Ev+w&4aXJ%(6;$-$`>48eD_{^C!7ifja;5k5k=Le0^0NoOuzhip0v-SNmTO
z3|vSPJtYx1qKo13>3vW89;qIB#*ut`ZlGE^)-*ag$~$xWpt?Gt*F=_d6uNMz2j1+C
zpPZiF!SRQLqdD`^#PD!%L<9qhv*(vun7E?qfq)e8q0p_epS}~yIec%wS=FDY-$pUF
zsmADRUZ!a<x%eN;JzLe1Jz%%qt$^a;;cQ$Y1M_Z&KNsX?-gCycUAPr*oNo&i1yRz*
z5+$+@5$%*3q>m)~_Uv^8IQ7@(H@H$7-sBq>Y7qpG8KR2`yw00`3PMB8$gh-u$Z#DG
z4}s1CENok6?6=`z{m&OZF7{&ud^&c%5W1KD)Gb_jgSavA9?8;nRv@rTXIr$aj|OYQ
zk&uaY6aFj?m9$vf1>=J^*n|tybg#woGJdLmyY3?|J6%XrRmPR6k5!F;Wg(z{3V*==
zmw>fg@{*|5sB!OGK}S3l4nHUF(q`w&a@|{T)_^XIE-Y@y;97}>i>!w$X*cuyOfj*c
zokj?vj{XqL$FETXLeO+}_S=ahIoJ1h4up{=XJ)Js-s9wz#A-zK<8q1TFE>S$DUdpc
z;dT-U2hRTquwJ0Kf;(hCkcjeVCgVP*<A)Iij>YsdXQ<Qe+dnCUxl&Q@LFJo6dw7(r
z7Lr6*TvB$hF-HWyeq0~pb99fy#IxD^b!DnS2Co9s3Mwp4EZX{GGx=-)_8x+3NxS!a
zII&>XJdGEPNI9TJUhzb`iD2}ix=5rX=z$nam^CTS^WpR&-hG7p6kE4$#pdXZQXmO}
zf}MpGi=sqB4I)YK_as3^faJ{KzK&yx#`*DbJ4-=S!%v<(i7w<-rk)Mww{JBaldAA)
zE$WaQ+c12OYgM^Kyi884)uDUX_OQp)G4cmEP~309TYu_QO4T#Z>W41#P51t42(DYF
z`%P|frcLMQu>w-FgQwr4A+eyIC4;12>*rq=P`97S<WnE>(8x{S%X+)~gM;@9qrLmU
zw_PJ;aY4KbN+i$T03FBdUY?W}jnh7OT;UpT1)cfi-O}}6>i4|iP#+WP=YO~Ka^uL-
zrT4ooSg`J|kQFz*VUw+|p?N!&rhVLEA<2MZJ7+|stdfuPHWBlHJYF~7NA4Tk=X3+p
z_j2teG0QeCUMGI=)fz#P%ZG<=^gQ()8^_BAzFX;>(xZ;PR2+2HG+9#N-nvc)BUSaL
z9E+#iaw_&88z)Q3oyvSSNPG{{q#2m<&y17&F2(np@m*(MdFKn={oBOsx6HUum0tU@
zB8C_zIHZ=VQc<#tt8%#ERKL-MV9&4u(&WAuNe$dw!RGllx~SRC$HX%K3ZqvY^xv|f
zU*2R{A&y7MWM3bROQ>oS$3w-d&?|;UlI3~VDNs+AlDVNRmv`#v4e+R9lan1MdjKk|
zzq5SxT}tde)R<J>nc24XY=VIp?sMCz)8bs+?@vDKJ1R8&jE!>V#$$nlBP&~W`a3(b
z^pk}S(e?4x-t9#EzK`7E%{iUeeOHZ(t*D@zAz0?vbDB}SCaOuQs^rLUc7D>aRQ;!n
zH*r&gD@f2+__H;87hTFrrv#sWE|B)0wY(d>W$IS|$)Tg@k;j#|<P1fb9|e5rS_(&A
zoMNS)|HK%!XFGw>+wkxlRa#~8<suhew+B0<qoE7a*Faz0q{QJ3;A)jnWk8zuWuPhu
zxqJYUoXaR0OX6ib*X0WX{Zk~!<dCMne)kS%A1|)h(xW%d#K?i0E6U*AU1z%m>sM!|
zWE7Mfx_3=qF$k8KmN8-KxM7gnATPl~U&t4{Jn^|p!H4JC)|oR*+9eI<kI>sqd<^cK
z$T+!|yZdJ(pqa2B65JOQsMuIpD^WZeL!ysgzI|2aI5W}^Bjipf*knHIZyhK7xv(eX
zWQNtA>**7F<mJow*DuKw^4Gu7-;ktqSjmBle*GVc)cXr1o`IDr8<J-9Z8z3ecff9c
z7?0t2C&}L97@jRJ%Eu&h4R3+LONs#r*4FlZE&a7CSA>wQqZx%84qcnL*HP&55k?cy
z4aCx3Kl?A9{8<9MoD9#=2d~W4n_J(-nZH$f9j?tpTCnW$;ZI(YquIIHq_pUpnsY{;
zQMWDEOe(|b5^sM6<tDl641oic$Lp-xNhhra3&)DGdHq@`4PKJMPHwMnCjuPy;1=6^
zgf1D8CUGN>Dc3@<(LfN)A^1Q*EIsFF6My~sezZjhUmXOUZ_NHq+Nc;fSb?NNEdgbQ
z5#(UrzM_41{t^E+L{W)P!>!=4je%1_#_0POE7uHLcSD9}yQha|1z6ngyFE>CIqLl&
zLX7kDy4}nJy=HspGq+aPM!C5MJ><9iA}jw`b&tNm059WF-+i&iCLOj9(Vz<+$o)56
zZTG>%Y5p%3j~N!wr-lS@2;@A2TpI*T#6c@^G-(oawE-w97p?wxC~!}QN^mSJjuKSN
z#YKaJ{DB+N=Mu~2XkW66akf1(J8E0+zKcCtMCZfk!Z5Xb@qVMMqHT5pTNoF^RD<5v
z{n&=L@dyFmfV$=AZieo@KGW6iy1JZp%leytcQ!yVlg!Ti_qGPC!9@B*l$CcalldW7
zV)qRsKpbCKd;@+oDkqlDna`!^-0!mMO<?N`%<8Gb1%&oce^a@~2eISuv<8>b({=Bc
zGsW~?TzcT1?OatDE9FvjL}$|8=`Z%v+4B>MT3TADlWc;%sb)Yj48(^u%$phY=!^35
z@aX8B9m*)WX`?TNHMwQmws2U|P@r;}S%ls$=EjY?D#%mabkgo9!N$C(;wEsZd`hy4
z>0@m}$7H8Yn9tzzi>EZ3o=!6b|L(}6`F=ad?(RVTBguuW_N(M&9W+NJQa+H6Z>?P3
zEjdrNFeEO|o<HFZ0B-d?1pWVrUW<i`YvLJBb)uPva!G;8JW6X&gIz&wh&Uc$bKSaa
zTg}%CL=}aA1fkA{=N)?M7AIC5(M?9NU}$1O3$fR46tU}uGI(qrQ@Igcv#_vpkdwQ7
z-(o<(*S0hS<OT5ieieFb?EIkc36WkGxaU1NpO(C-{-CGDV5xLNk6Q}U>XxEazlUO+
z$6G#hABg=?cPF`E!p_f~BE^tp)N5OP^Mr9a-|lM~oW+-PAn9{3xJ{=;%*FU0cx%zQ
z75^-aFc2Fj0)YR3Ci{^57Y1qeE(}l*69TYDzJ{VULDrE24Ck`%7Z54jM2$EWg_`vM
z|3>Vrs6kJMNO90i+>QRDE$E%Fx$qx4H0`A0AVa!_S}aTpL=OtxFtOH>uY-e^VGI_k
za(fc`7i+H3Oh^4*sot-b@1wlSu7vqNO&HXin{*2E)tcxdr!g@Ky{CHoL1o63zmP#1
zoWE*kHO`;yVVD0E=1%f16-zkWdOT78Zv039_7*iW5x=je6X*Sxq`Ko>#FpWdfLfK%
zgo7|k#>&cS;fCbYkqesxYEf2jH7rgeN;{2}$tbXwF96BuBOr$n3IjTZaG#hc5YXDT
z0RPYsn7**S&dd|pD>2Ohl(ubsb`^&XuFlEB3#+m+vp$N~#6{`Ge>7!OcELlB&U+)e
z6JiD@)%D9xHO#0)|7x6j<hR=7<CXI4t(}g_zRHUQ@e(>W<5Q(&E@=U#7OvuVZerj~
zU(7<GyHi@mgm!CLE*mxPd;m|0xakWVZQlPZ6AAWrAICdvh+_m(LZlLI(z}7gVCFlv
zCFmT{$;an>a<K`^eM}&?!-1GOz^cFy_6!1){`x~Z*27$kN-V72MIR|F{&KFiszgUb
zU=j$&`Sa%yB^**yqeQXwv2$-cnil9@@?CuSRMHx|tA?Rw(1;baUvaA3f#h9Ncc{_g
zqP(u5o~hR<kk>}f?_>I*eqqKoGZzTPnrF<b1-i`ob=7Tu?r1PrrqWpdv=R#T`5-SU
zem36sr`~^`{(Sn}&yhZNmN*R?w}PTuC1M9U0huP1p>#?>4&hHhhx1bLXEfg(Z;*p%
z3?6;>A0K=QqGk*=WiZ-3?ugl8PS*m^9>)mbad{&W8lKGWqL(~ds8nsDve9bgC7s{o
zl5T0LhE-?m<#w+Ez9C+9I=()k0lJZ9hnuEHv<pre(-!jSdu@AKE>P-Esa+ChP{pyD
z^_W$IYlG@dxwxG6zM}a1P$K{nO+XB%hHJ4Ni-Dls$L(M0JL^DT9MM6216si>cQ$e^
z)GES|Mtu8O0`COdr;NMO|G-hJJ0b|l(e~YwQw9jfuS<NA${Dk-547toa+>$)Uzt_0
z(ms{hWw*<3%39Dg39f66CJm;F?>95X6=QtW&#)^!CDV4j9GUCpmgwH@z_`&YxYDk&
zQz<9{2i;wiJ^oLE)OIsOZd8o7OS>44Xa{9TrfYRl6lKqVm1AU{8|o@%X=&(9yf<?y
z^x~ti^PT**BmE4^USH`=qPISrpYZ(iJf-fpV<Pr;Aw`{Y?~X}`a@|sB_N8T9X6{#_
z7<{QZ8Br%+<(M__q|3sKA>5y(RpUwU4ZI`jf2BTzDjHf30^CLHW9Yg`mA&unO-4C~
zm{bE)c;vwh0<NY7XB1IVfB<7XC+9nu>P1Jax(}WwN8l)eX*lQ?CO*GUpLwt2(T)s&
ztpry5Z)Pes#eMU2+`)dW)WM<KXvo?wDkD`^b^`Dw(WX@5FGj~-mDV_Vby6306+L>}
z(wRgMOnMD#g*)P7VyTPk2hy4b-x`gi#4%AS(QPy*{~6Hx2Vm5j_i}zvy36^_e7%e^
z9mb}ls*g*4&=#biX0$*pu46CBfE`St`zxPTJkETpO`^DUxc>d7O(zc=)ZWXz=l8u)
z`<jNvt<opY^fh*UJKXqwtL*ay{`&K;G^hZo>F&*bniL-t@j}*|$yqZjv_#AOl4WqJ
zH08^5G1ZF`a^2i?oPVBvQ2v-M?ZHP8^}JqwKL)J?p>#;_DIg$&ZW{$ZqF+H6@y1%~
z8th3k^YeBnEE5uDN=nKN$&^wqAIsc#YA2W-mTu~mV9Rg$MqJXSLkAAJSjy|)w)?`d
zSAsTj&u{Bt(P)Ms>w{H4;jiZ`FQC-OXKsBsePjpiO`x)zmUr$aPk(Co;EuZaNLkrQ
zy1>;H#<Q2UhlQzlc+-)mQS;Jh5<?Z(Z=gFz2pZ7e6I4`8Cgw7sHA4(A(7a_}d+>RH
z>yw-kd=1?8BL@2FdMAJ`SBJ|uF>BD<Pc-_HuHUL{G*4;!S<lMej@t?abNEyGwOr-?
zg1Ln<g@F-=u+$PUTT=<JO$&Bfu0Dy0%}F_zN9BA-R6O*(RGS&CuGHU<Y$q8KeF){g
zwwgBk#23FP(_EFKvL0;&T5EXtmdlF_W9?EZvKBHOPQ?_b4*Blv;+@N9jt|>a?V;h7
zWiB=7{xtQ>xKmY3LXn->k7d=5Z2M^3IR_J3Z;cPBuX96e#@DOr%d?So@Y4l|h>#We
zL89Se{pIYIEfEVhog}iX*Qech6L-_8V6)x&w4}z7;XJ*t*R2mrN=4aNJE+Khec7`l
zOxD_B5S5WHb!VhAO<O&!7o4%Iw+BsY-!-3^8thi+-XkEvVbqnQdvXiy?r#R)UhGC!
zE}63O>ga8e%V-=-qG>cZJbceQ=0gI?w(YMoBfhBtuJgk>;(4sN_(hHW9P}^%5qpR0
z_<TIhEg>OC?09TymJntk7~GD}`zYKLp;YmE+BO`Zg&WC*Tej?I8cwV2?4i$MQ$0P{
zV1LMDwlPiJR&$4t9^R*^w%YeeJacu7j*ouy-z|l8=jKqgi%rTvi}}sIL)xyPJ&iX7
zxh6Ms*9a`A(puLR(cjBk6+>MLyL(N2eVYnL^$nytF50DKdp@*QkNz+*FEQS@=XXzF
zrJ?HKh8MHNxFo-aQ!>`W4l}VvQuO||U$#BUly;x4ZD@(f)g`yJEGxf$CduuA-hOb-
zQoiS(Eq$_m;ek?t?qql7S06>*AV*XBg6Fv|uAk=W&JIm9o(()1UA6PHWt)t>iU#+N
zZA&Z+y5j8GjExNqudDWpTsuLDHo?AM&ZF452NYu@EY5Cy`rFv%i^uh%1S3h{0FE_m
zGqlb6%6Miaa$LB}c6c~8YggZ;c<s_}R%ERMS*h2j=Q<Y+v}leh<WPL7QehCBz3YCu
zWFsYWFmp!Lqr}>8?5g_A{{V)w$v?@^g2J+`?Tv(l!-#><SJy6f1Zc|*-4j2wN}*uq
zO=+sBu|O7s`3hxxdk?-r&pshpJL|EbX#JPrv0F{s{ZfpppFZY$bq=rQaFW~dJq@aB
zyB$)TPt553{4%zjb|z=fgG(!hQ5If4Ov3#c2lmUIr#ros;jIS8ACJenr%gY+J4&w`
zHRhoJ;qnn}?NX2J<k?`<p%Ar&`j-^dK2RbMvQ5s&oRXH<zMT$Z0H&soKHBG|zLo5l
zzQLZ~JjBlMatl}d22lnI>z1yFf+53{YrAvSE7`3m-Z8iUb=ajyzK+GIlKXNm?fEy~
zQapbEz$os>h<2LScZarU?sCyvZR_MQaioh=O^7|1(ZV9IKHl>5=^$fgzuL!o-zkai
zmv+{HHP5`QWUN|73!%~CWFY)?@7^0nnKH@p%TD}Q<TuVQun@avDOcKR(-%6EvehWr
zFaGV>Pi+%*O){h8X@(bTt(W4Q-2bTfKN(k7^^dZ8S=b_adDm8RO>KXPE@`dV2UQ32
zMLAe!r7Q9~x<459EuYDF`@V->aa=f>!N2FzRNH&w#9c0S8E^X+zv!HKv!Nr^IKRSq
zftCGuF~`0gxxGI><g5YVc+dtGoBu3I4)DR<n-5*K#j1N<q<e7d=DE%}R{6spZ0?n+
zxIKzkvW+R1wdnuob8z+1*9DDDg~<==pNq8{<lLY*O*h!ut+O(8ckuB*)?V-asgDNY
zOO6ae8d>_Lm6uj#IRO+zWJzqOgFk*0Tx#s><ba9~lD-C$r&NShB2va@@~3h49+m>W
za-WJL*$*SNf)3T$C|4;(v57`ey`E5tYW-|DZ09xDq$IxQcir#mahDaNnc~dqFZmkm
z60UTyZ>vZ0Rb0>OJd)7a99(8+!$aW>^$o<B#W`PiUu%*!e3lsVaB~TJczfX2r^|sV
z4&~%G1-(~?ESw!%*w)}>Cy(Q8Zo83MsO!(c5F;1#C0`|i`gjJ}<AHv5%RXw}k~;fr
zu~TY&Rh>NV0;#$-I=f2b{jfLjF8F#{_S$aX+fX~FOH@rbmUy20>DeJ_5RYWnd?9Gj
zBRTk(exfLc7%E72N-^I_mUD?>c)Q~X-B-MUX<-oG7cVX+y6;2*&4<UdIOYtih$ii=
z4?aAv+;Ucp+$S5v?~F*BJN6IzNFU~*)(C30&ka79RoxXqSySCLGfICTD_B#P32?f%
zuU|t}8H)NaTAoxENEu;}n}*(Vs%BGkLZLeYF*T0P#w_fjrp5fp7%%lbc0;uLoAymc
z477CL9ibbVlI^iF9b8N|YECsP(~NFqFBHnPXpNpSJYaj<U1HPdm(webR16Q3gI+|%
zv9UT|rBS0XDcNhU(8KQhrabdat3gyyV`KA}ebz^vIbKe7U4ZPsBS5Qu7};KP(+?;a
z3Be#y>OiY8`L8Qd+f&mNa;a<E5akfVs+jKnHZzU+BJ-^u&X@7DI>dE0-n}EaNzQy#
z{!z_Pp3;a<i=%V|)%&|I9dEmxv2(U*R`7YGKGkV(WIRy0HqPbTQee7KYsGNLqwbAP
z?cLcYq<$Z|X_QEw9u!1@#xOKx?E}k@v%>!*ykkf1@{R%$L=U>xJQU=gGkN(lZ~ar|
zQ=u)0&_#SH=$S_^58YWsEZ_;H{QZZ;c(z599_A_4;LsM)<L9M*ANZ+O`ct=7NnFlh
zi^tEnxiq;2naJ&~e=2@CQvBH3Ks?i!rsx<1zl7ox<_qNRsMOUz!BD-~biM750(q6l
zk6?eJlQ?e6Qpiumo3(F}p;?5F?;|{p*{}6Q<JjDsI@+avy>K<{*|wuFn4`X$qfY6S
z#LtPqv?o7~7btaccsv+AE>C$J02i~O4r?FctQOj5%uW4PxJp-G8$-cU(@HTVS37Y{
zZlfFE-lB3(3@O5tNX;<jjid$OtivNChf!y^k7nJXBTAk&w;g`&x3v{Pj5`W>Cg$($
z-9G#C2tzunZTwlLIq}dw1|jABSUHsWVf(tySHAx}_j0Ei`*Hc-TXQWUZ49nMSi0|v
zgJ#8_-7nKp?fOTe^hJxFNbs;l&}f||Yb^?C%jup~FD$<Prl-D>$}elg9OtV-Dn|c>
z96Wuyn0?*~Do1Du>B0DdPC%df_a3+5tv1|=!JF!yQb$(~oUph!|2!F-m-`)};)#pt
zhGP2tJi<}IS{Zlr=@OsvOj92nuXdTX8f?z-=JY9+ICok+p^oivTxM^%mT4s#W41f;
z%2SnmZa|H=4k~2mZ^cZnvEG%*`+oC7)#adpc9Dg%T(93$>NHiwYF28L-e#-5nV09{
z6?-eoFX|zs?vcSMGZWc}y%H0<n-Cz^*KW-Z^hmz7LG=0V>DOI-#iz!fBa{3;0FzU_
zsTfC}G!u|@3-r6n7IUO21H~VtjvIz+)EGOfGnKH*PDCDzxp2RYU98iLl}FR1O2+TD
zbDsL6n|)uNQaz^xa5C;`|I$Z3QEd{XJjN^m4!6sYb`i6)?xVXXvO^H{CXh1@Msihw
zn}78MKRY`_(Uc4vVuH}~K*7L@@rQU_Fg<`$5OuZX<<GVL5I`csq(Ny0xe>Gix{0zA
zj1(NA;+K~y$g+BxOMEAeNCa==jng<$lzE5eWS#4l%Xj;}oc|%U^Fhe*#IMKSb=>*r
zAoKPs7g1{^Fb8PU6WJ{ZLXlCF*_Ri`3!zDWHS;E0|BW#z=zlQ-V|nAxJ+M|(tp2$G
z(E~9-YGi~kM1wRw%vJ~&5jstWR+au87pJCPe12Eng%du@Up>A~s;pVK*iR*jTvnQ_
zqzDcTokF38H$?!36aqK_50S{u$xc#2a&iRP^h8yJHUz0Zv86(=d28$srsNHxDETCn
zQVLQ7Bn<`=<rpg|8uQY&I(|Ox9R<h!+^ym!giX(43)Q}i;}<cmqYmE@kt(puiV_-A
z^WP_o)c&(0_uZ}f3)=rq+Nch3_Yr_5&3o9>el>8)`}^v^u*lk_$3?d63V_su$S+Yi
z=1Pf<it31dk?RBngx5RAHD5^V7LnlWW_?lbnbS4p)Q1eJ>Vza`I47O$#_FIJ(39jL
z9g5<4gOAe$X}32-^kg=SMEm&)P&t}sVTH@?aA0F&V|>4L^Jd_?o#T0sE!~V`pIIV*
z=mHA0;oN4%)%8lrMg*=Xdot3COeaa^3JPJwFa(A3KeW&YP6W9I@!|-^kL}Vt`%lLh
z%7EPwPT$Z2Z!L{yuju~CC9)W?SG&=J)Q&7A<>-r=U7S5Z{O@l(+6spj9ycL4URC<n
z^9Uf?vRIfL6e&~(N#KJ*)qsHIuZ$QEzKU^{O`Sop56=6+K%4XGtJr`ln&*&d^94Yp
zp^M^eEJqQ9pXlveO@;QGP~VUm(p1b}vU&h+_&00go+sBPlPWshBlk`P7akCr6<&B8
z3V=vK=o$+pVn>^lB-$v%kw++Z@(ql}jPnhoF{S}%3n+kO%||#klN}4!82(0Q$VYF~
zb;puc>5<-;2r=;@PpiIC3DQhL9`MP3DGJIIjJF~BJgB{oU<{BXVGz}I0B34awCHT0
zw24I{Lwd)W<1yO}w@qQso?-|^hzfDgG&G?@aFK7>wCR-?m2uKzytQ<x{fwA0)T3lC
zNAKHo^CA!%9R`WUB*%`N6IODCh$38ETt-f2q63(F(A2W>Rsj8+Snx|*L&C_Gee6FN
zU_l$@5Rv<UxrE7CHwZ>l_;at-q2^q{1?LskIkAUM%DB+WTw)Z{l@u)LqOC^>of1(K
zL_vi_bgeP;lE8SpB?C68VqjEAn(j{$^UD{^%s<>#Noz_F4m(!T((Jh19Vx8a6pus_
z_mQG`N5d5*`Lg{PpEY&pz>9K!=Sj*{Or#W6V^8m>%vm@2$=D`zvdH|7%U@~|TU`L9
za3wZ|4=gKBHWx9#@yMbh>PQX4qpXb#$=Jx~jzMk|KfdIFVQ*FC&#F<fn)3S-yeaOr
zYcw5ou095Dq(9V6(^FIOP}jq=ZG&MPBW4L`*`=gDdU$N+^RY574Z6#(fRO$RvSxeY
z%dP;FT{eNzNq{8F9deElV2~5XrhIHZ4Lq)QGgpMuwp(Eh1hb7(gqxxp8Y*B1kD8_6
z#~y9;m9&=a?h1~$U@>y6d3f#1B$K76D&eCbxO-@N)#nnvN|vTCk_(55yaJ9VTYT3{
z?=W7omoj9WCiQ*%$jA}WIp%?Rhl5|g)}Yqm8vmK_z5Xr@yZotYDr<03!!3_0+JEyB
zHH>yj@UdeU12q9OyNZe)COL^+fU-8{nuv1xC(a_@4HJvxDM|y^yTnal8oRjzUynJn
zwqge`hBmw7cwMU`oWPAgw+j_DO0GS>5n{JjSzV5Ox%$UsRY^+f23j<?eZFv`&|^tl
zO0|<hIW5b}`$}f>qRaCUHsdfkx?9VTDxjf;9GaZK4@1xQwePtcj&M{G67nIal}s*(
zP4|@l`gpSKaCk&9;Rw`~_%BIbLJbEq1K-&mn3ilFuAZ_RaL-$NUt>`wb#`{f1Sv`j
zbaIJV5F`|@FV+(jmpuj!Tt&3sR2ZM0nOqri`~LW^Zvh}i052G>jL^JxzqlX=E(X~E
z7`eoJSOOa+)bRc-|7|ns{(M1cx{7g{2PegddzKK{LM(m1y!?EyM%gjQGsOSbC?UJX
z5WRLMsw7oN86iu14F5DlPViZYnchHs_0jpuc9+QKL+d)veqsp_OC)U#&gQ8r=XA4a
zhxEMv(&|NijL3!*_!{&;ot+0yR_+qSqtk5!fz5nAc!qNFJLy1YKz;^s{o#7jwPuxZ
zjd&;#QBh2QK1rrWT8cR@07L!p+p9;pRqxp)mML;W>no}tf@u0oCi!1GPR~F1c?s_~
zF@6{jHFT`Mz-=aDIg355YRTVQgg);(=EY6c8-9>8#G7WMM6$KEW`(;HR>3QUht$>K
zIjw?yq5I_}6|^dFFr(^_{Za(s2a@TR-DiJK*}z!zq12g|u=HT!lr?%~#BhgG0y?z1
zCl)+|0esP5qPg|ze(jn~{D*M}4@EX8WSBjI*`RG{49v{;p_0LVE<nmRhydRMv7il5
zu;<HhU@0alc#6V-M?(zl40zb=c?8Pbn%deH6`i9;K@T|zGrG`*;(Ckk5p_d9+6L~N
z6f?etU9#3RicAS)WPIz^9Xl9E=f1pwxQt+TLtOlyKp%Li4p~_V%ej2SSTB$^51_&F
zPe30zG3j>cqf2!|M~_>qo?kQ8{|Ln7i0?hJfVN!2%+wbb=D*aj+c4!JjGaR8lbBid
z&)TC{W`vath?PSat;v2gNoKJ49lx~i3Wj$bK60c*0LqH8tR?^_0{QtEQq2tbR~4AA
zt+;&f#V!$vHwxY=c1K6YwRuQGjayiV(KxvMXW~C=WkOy`m<ON~n1rIde~v(Wz{tqR
z48{BXB^g>@gAa1v9d7c^`&!wtQ9pq4_CF)NQg_T(la!Q{Kr`CBc{5jv&f&wAy}i9+
zvQTqD+(Em<m^2r<>}2>y=Cwd6f#uQ_{eQ*wQl-sk@0@&Xl|3&we6RXE#ei?h+6$ma
zyuf&22npcIIpiRJHRGYYO2`X<#TDi^5Id-DI`{rf>mdD0wpaLb+aXd=e9%df^=3e8
zg~SKnza;n))*jr%?P<oSR04AcxdKGdJK?a~2{ED_dQ377?KPXJ{`+*VyZWtc7H4&A
z#^no1Tz`1GpRaB9+jiA|^x9YnMjJ7;7C3?<Iyza<7Dn<bC!rskkZ-R^6d|^ijN1*2
zAIsd*Tp0T>BqizvaVzxC*)RvFg3oa`vVbpe3xTU^<PX|<OY}4qCJU$%Tv5Dc#H#@)
zrwqYS)E(<MlNm{;-&<AeS@eTFX0-{na!C-eAei|LSekv$<L!Xg4LxewyJPlqJ-#8o
z<I>^8jd>3PDL(C=k<mN}pa8YYbJiK)4&=f7##$DUl43*R=Q}rXMp6+T5jDCbmHT<?
zX(#8Mx_5mdtodr|P~KU9c6k!5RQIpVLQtCcbe^!R;o65V`NP~;BS)n_gP?s3afIkA
zVxY5Llos4@a(|z}VD#Mfo2qpeCrCz-VP+u>jz#D8E-Udsuc^{bobx$*GE6H8x)==M
z3rkCF4~}mooUaIrj~qU10T0@jJkXd#cgE5Cmeb_lrcR|c2z~YY60xn{O_J^*@6I!l
z#{2yax4{PnfM3wm+nM6idm0*5tu}^13+ApYc7Z;0mCQ_B+%ys2ghd!FprQ>8UxQHu
z4jWSkUT=3gzb+Pu3B5Tc{-<Ogm&J%L<dMP{%|RW2FB&@6-M&KJ#1X?^i1@_6W9_<y
zf=FL6^cx8U385QA@Vv^<n0=(8kpr*CEuW@K=M_uz0k533$1M%YDw^(ZKYTD+HO$mA
zy0;BcUF#`7iy9upq9>VE+bz7m93u8;%iRw-CpXq8z{25v?l(!yhta+5j*q!bWO`iB
z+V4$}`c9Q+kPG;tnAabC^6ZW&8qm=`N~h5bjM=1@hGe`1TA~(b2<;=J98t22UxtS*
z934$usIVQBv=&KYCdsj?3RjDI8e+rNW!hjuZ3}n|333e}8q{ElnJGLY!vi$CeP>?c
zZ&8rSathw8R-I0(es)sGRz%)+U)U-U8L6=wh>_rO<~?Ri;uJHqXzs(NrHM`7sL0~-
z+V_1+|4v&|Ms3m`rt(RzG}3YQ=GyJh1;hd;E(u32OfNiuMLgmpCi%2(Dm-AzHepQ4
zO$%Fi)$GY;SYN~|6nut&o!ZoM&|LxmLtB(R*q`$IK{AEzU7Qs^$c}7Xd#qPM8v+*K
zEa94&SSDPR@QbbyH?m9f^3t2ON={y|Y5BUNcyeS@k?Vs7R*to(fD5MqWOfm$KLLQi
zd`*JxrmBCT@zubajm-n=YU+c{-yx5S@#!OwEIFhMGD&6S7q3Um;3z$OH2+jDN8@ef
zWC$`_F`fc?Q9F+-xZvz##G`F$zQ|)}^a==)2yND_tLr6#%jB?+qDTi*`J4BJ7;%dG
zebZ6*z6V;{#3lmBy}aLoB(|s_Z1=?;Mp(!S1_qPOJAg3B?%-;bc-2}_be4X2)cBc#
zklL0LP8@D`2>UAZ*@RpfMr*S0Yu9Xna$bhz$#zHc1{~fn$Wsf#4UNGw48$&AGdxWy
zXQe>UxbmrbZjYPo4y8Y`-ZG^go`m+&p1}f<1f)G9e>;xL=XxAZm3v=@Tc`z`&RBL<
z+%tR$JTh1@odu>Q6G=5ShaoSIe(^#Xn&gT87th=03o#H8)=&kwz%X|t4C3y#Y~4+<
z9hg^UdwQG_vsjOR3cT2uyk2eY($@8AxGby3RIKTsoZ;QiI>mrwk}v`h6Ye1_v--DK
zC6J1N77ij%01&IVLpE_y(+oKxq0;t0k79|uynK;Mp8{581WFnHZ1tO>mGlbIG{Iy!
z8nGJm%-OP$ftYnFMXl`|T-0<k@fG}JH2;49&dNT%%0lS3zRlrg3Ax-$mL5M0rKP9{
zTx1~<d^IYfzmc1pyFY$c50xLbUetdQ_g++hnq%{xJ=_DUz0a4jv%f8~4E!Xswiwbq
zMD25IJtI3iYtY6vKS0~HPPRQ1h;aJWQl#I*)=AT9;BhhM^c{}--M_aNJ#)z4o=7>q
zeudUes?G0n?KVk0)?@o^fBcMY`Cw17wEn`}u;lqZWyo)|KXHBqz_)i7e$JaglS@K6
zaFFhlxp_2vEI7CA--(Hdv9hrRq5xv>0(iZxcrZc8nPL1ALRf}1(kSuL>$@@BnhpZD
zZ{Lo7u~AKp+R~Da(VV>8o$>28s*e&>H@1?!J4u$aWR<l^czvf61zyS^j9erpLKGDj
zw_FPg3E}>(5WvpP4)RJ&{Ycc?5%)cHb#=zG{Fj!81W;EDMbr)rJQsi=2}(ZhkLrO(
zxRaBkv4j~3i>I9_)?Kf9O5-3BPPxoNB9kxSUXb6gFq_m}8(yO(dx3cXN~s{k<HzhU
zY4Jrz<E^#e75eM89GgtN?f3K^TwcF!U5WIAjgx-D1Ih;HTE3=DUH-l)plCK=wLYo0
zVck0A<_Z3N5A=J6snT|uS40x7M49#jQi+W5@$s<#xB_Kq#tgl6hP9yZw_|=;_CEw$
z9p&9#axBR3bj$IUUC^AcR<l-3awg%xk0d~ke=auqY-YPc3Fnd_1SF_<+1_-6uN(vj
znCU{jwDJu$ZQ4XZA5y)IJ3L&0?+}aNl=Q&UxFp5!Q}UEfyv!R?ITX3qC5GGbR^rvY
zbN@aybQWb7<|W`5BKmb08vJC@&A|~oM_jq;1ATH5Jp;oXeNkOKJ@S3$B&=|>L1G#O
zU#=D0{n`2XnE0*#+y)=OLrm=QI#;Cx!B(2Gl}#~9>uL-!6Gg3k%Ep2;fNewm{0}|{
z2M0;(nwW#q?!5=#Qau>WpD)eN&3y%>6;dTeI|o2?oSv7*2qX;QG=%fKSN71kq$Gu-
zYVCL0J4lCak=LzuD4l1ay#KvfCRD9`Otzz5=^>fSc_-49?6!qJf|EbhD_@9K_At}-
zk``XP#s#FFI{rc{!>AHWjw%dtv_Q8I?%2l*wy1eb&d$pA+?E#>_SLF5zIuV?`scRF
zJ5-f%J1-ykNnWZ=ti4SRk7)=>f*M?0fEtEk_{+mL?Bn0Qxxtl9Fx?PHOL`wGc4HA-
z=c(Yp?)J@8P@0TYKOiIP3$yTx&?~8t;f(2hEM%8{kK7_OEdKbb7sDBcZW5br*6eBi
z$-<KXCLsZ@Ri)2Z59?JvjKB)a-dHkb;FeFIdQi72H?3Z`?)r5fO7dwPa{ou9{Y)v{
zqzOKjk7BHNnB0aDOawFt6-Y`Lr-78Pgqum;f4OYV)5OGtFmetET?wz*cW<t;-MMlm
zHK2Xnw#JT+`_&T3-L47hW_37Gl%dcTDWJj+Fti~&JUD^CrK7=IX(2c(6S(o+&D7cy
z94X|!q5+FM1I>PogM-iGthR0vrm&ym+lTTT##&eb2NCsl{rdISV`2hYTlF$1H}E!*
zZ5DoX*MMxR9EFjifEMf6Z;L?MpA~BDx+P<Lm34I*_^oDkN-{DVPMkOqs(D0B?Q?Mv
z<-rQN<q?L;2b{-FkZ%qnGtVc}XTr0%>(%=CMF0!GQnvITq&ZgjnAA+DmJ3h&@g!WL
z=UV<8L?=9c*>rt?i8`+Bu2=dwp|4@^AO!%JCWxJfKE$?b6)K@90`nCm*DvX!AuKGl
zmzQP$QwT4<BM-YpQu#va4W(EwN9Hg6M*EOj^!`!R=6vHL!u$7M_BkaIFiK)uM^<)%
zkuQPf)T;5cUo%6}!REKRA5W>x94_Z9;>SZ>r=0FU(sm*LGtJUSxi0H=WttaB*+=u3
z@bUXyA?Z6rMUUy~T28vDsaby(%MZ}iy!`T>qe!@ZJn8vcibTgf`-8KC&O9W8n_La<
z6l*E&VfQ`wFaE$GaVi{lY}295!osnixo0c_YqT0Eudn8{HEgIjA?76|?64*EK&6~(
zf3WJEeP_;WCtTt%oe7XYkEDWAg%f)%p><42O;x9Ja#EtyQfm)9w-nHHI^bDvKs#AN
z88_E*VE2RYuX4to%mZLUUB|R>=w%N`+>-MRdiYQZ$LnVxWnKZ_=vY21Y(;TcZHD&A
zRemkj-ztsMRnnJ>qLPeWt8IQ}kV@f=;|VnZ-K^gfBy4hd<}WL1$0^v>T~BoX!PKyq
zJn|~VW{%3IG8Rj21HRXYMbE9kG=Q{%(cJu#?MZ?mg6*7<2a|OaxHNZI=!>z&6(nvv
zWyFZ3Z@m(<UM*ZJMw?<=Z0#|9{q_m^kdP2~?{kHuVLdt2A#EtV-@2KxG1KZHTfp#@
z6hT#MD-WJZc$+N2CC)A^>@q4VC}4(L$9{^x3e_x<3}@KwcuK7m+M?2~Q%b(Ls^=@j
zDx&caxl9C3gzEI3-gcA)iHL_bFo4(t(Ob;U%{_p<K}i|<KPUYZN>#NPE*B1*E6WQ5
zew9@_$t`*1_XpR68_UwY2J#6QCIU+f<S_va&`MQcM_Tj>y&pFW8j5xP14j-$01;Yg
zdre16(4$41L<5$<8&GM~4j)pp4t=(lPHiLAL5Yi!`!};MxKX;T@D|6eODy4Vr8+;}
zDSqmyI~9rO6iki&v9Yna^XwT9V9am-tgI@xH#axq7f}mJORtAX2#3aNG#!u<KRHyf
zbsj6XB=)mYV}E51+diuYi84~3nJTOGa0I0qC%aOC`4fqXA%=1!`c2c2V15q-)%3kL
z(2!{JLs-C!ngdjLWWZFd7ZO@`oqIf!^d*gwTDwb&wf0m-qM)9UFxKd!>6!~<Pi{f=
zAgH9ohb$@P@Mh?7W8O=l{7)1Y5k$iz5SfxP_cXK}3MEQZ0KcEZ|J{s_FRj0@ew{9F
zw*pz47wNc4yB6#ATI$lXfwb3)x%0_TVhKr06F@O1*-u$qoC%>jr0Ik^7`{_Umyhz8
zTX5icBx*LOon%ihU=N8b;Z{?dDcj6e`BS~Ie_wRg<!7P;=7l60PLgK_wj3!2gSU(6
z(6C!T9oJY<;WKU9kg7;bcO?Ztl{<vS%{$bqFq@r$bWAkP_hM;svPp)re`aPTEj@kf
z_>1uH>{O)`!Kt*_%G#*PCl_QdP-cB6DTQ85y_S7k+L)sI#^Rzc+GZFapsi?YX?Yn_
z1mH?gnOmZvq8bDqWt+Ho2)bX$f1%|zka_B4YP#{;w{MS|a&vM>_S)&#3@Es;C$mv5
zFL|ctkq+}QDDhHk<8Tk8)m8NFezok?UwB03?>I?eULd}HWC3_e<CVQgv=-n}(={<+
zzy=P|G7EhOIl%Yba17&QW-2j`f;l<&19v3?|2H`&ay{N2oo$s<Q%h}q7DqC;ni`mW
z{6RqhF`9z3{tT(N9&O6xi_TETBHWtV^l6iX1T!8%&nQgU*p$>jmJ+;vxMxe&gQpjl
zcH@0HyRztM_U%5Cq1;N5%c5o9%&TW(VV-T#4?2&sQtF>P`Kqz;sG=<!Q@WN&P-f<q
zy83$Fs9U#gz3S-b$Tq^Ocu((4#x7XgGHh9ixc<mROGWz5qbhmH{{Q|HPha)CP&WEg
zD5H_w<qC-SX#ZnRyR-3lZU=IS0M$FJySPzL(j7hhSUEseRu<ffoH-v$OQD+@TQDr~
zD(WJ_A|llI9^&;63=AaXb9YKh4HS1IVD>?HPgn~NVPHjl$9QRZ(8mj3d01QfEp&>%
zV2iQ9WaVu;b_8L`L6q((5_)VYRwEOOvTM)GQSsCN_K8`>?+f6<H$HF??`K&1xe5ZY
z@-#8TZ-@+I62ui=Pqf5li$lY<o0N3r^=q}6AAP*Lckf2ta}a*&NA)6Gwpc<7=H}%U
z3ZnzMX2z!~q#ZT;{CA-JijWZ$on21*)^~J1J7HiTtf0UH8d!k7=o{#kLr@sw+MYjV
z@7RY=rwnv;B_u*PEDSMl3sI}xy~m#w&wi$aEE||M>MLr2u!t$Ur=tDA$Sp;w_h17Q
zVEEL#y|s_Gbv^@ohlFabI%=+i0|TTG)ytTdQ7vzIG;;5!r(sA*nRe{hfwB&-n7K(2
znG6LQbQqz$9};flrGulRVc4Vxkv1QdNQ;XL0RV>%db^tU??aJEs2(_=Yi#T=Jp9?g
zv5#L=RJ7WBb@3#aG1Z5U9Mv=#H$WBwPsiQtY+3?D4X1d_&6~syvS};lNt5x?^FPfH
zsR-}g%g?zJ>}7#Vk#J-$e|Nnk?e6}oy|c3#UVH3A`NkDRMXcDPUZaGB=(pJWmlYoA
zyNRU4($4SbEfCXJ!4(KZ#)C)CgRZpvg}J=Tu}QZ}JAd54DZTu=>C)SCP0DuZ*1oLS
zZ4C7EY&bbenofwu!E7QzpdwF<%gM;lBIK&ZJ!WEW**tdqNoXjwCHgqk*zJRnKMz9C
zE~l#3*4p_Df}*2e+p?mMnVY-jg^N7M_l&r9jRUr&HhVD<k%Qh}673k1+Vsg-nfaBv
zKTEm3n2jG8AMdqyLJ7UP?CcO@Lc`;d`z}YhG5@#-_JKK;_REhmy^uL>09Z)^-Z=(;
ze}C%64&#p$-L)NEU7JuHz!uT7i-CawSsi4#&rke3JNka($CsA`P`j7a*NV+>?K3De
zV**fuK<~Z{KjS|0pch(kYCQC>U+rr~+!^`gSpkSfj`sGdxWdJ6yFtXpc6WwlJ?DlE
z18Bw5GB8MMe7Uv_KVCPA%mQ0<{vYwp#O4E1;-kgnNqHD`LjnWIafJ~ZGuOqIaY}M)
ztE*E%o%5cm1W6w}FJT5R-T|<J)p>vV)rUmfNc6i0A*u=0dmy8#+ICN;kDi`>vg4%r
zAe0w`;KU8p6@q76c4@VucEy6{>RwOG8Q{Q*BEZi$-SH&Qo%oc;O!njih%T|RsKf03
z(f@)0G4=Huo?>wQsvSMbWZRG+<h3F8Jh-k{Sf4RWrRbqletg6Q(gSq_2L~J+MDajq
zVYM$FO31@U!aFPoZKPD9)`a}VtH$}E=o0N+{JM#fn%WAF%COXl8%`<-FPYeQtbUkm
zP?BaxC4=a860RP?(_(cfL)!@zDns<G{$Qb1zj>oRAun<?Q_m6!D)CPVG*<D@SIK;@
zn?0%j)Am{N#uI&yaMYy-VZ!5rX+-FBQVLc!N{8D-MOD4M%ZOBxf`lm44IE<vQ@8su
zUGwNtXT~8GkFi!NWUxKOe^!<<kgUzl&H_}j+urtgx=tTql?H+mFAEKcuukKhvg$AQ
zAwHIv*uv`4r*od3&JW)*Gc$i&>Y|5BE<xIk866rsgG0eHCrnMFu$iVPU*Lchi!%N9
zb3s~i;u<EpccFTB@Uh==$k$;U^Xs^{pqgU*o?x7R5UI}wdi6)+`Q|VAf0j=(W24>W
z@%7kgHQDXX`6d8S5NzN55A+=b0JrOWzk6n8M#k2!<?Hmt3wURUb>RGpkpzw^&OGIY
zyT2)Ba(e;S@klBUJNmd%(Qo&WJI7>lc9Hy$w~7DZxN%TJgB~N`U}_7zvVQB;jadf(
zpJ@L5ToS!RY@JA(g~9mx7$B^pDzphRkl0ZwfPE9=NZj^g#M$kE;o;3U<KpN71!=Lb
zt2#O5dHu$dphFs8jXN5GBA?Mxc0mD$<9fB0&SxvPpS~w|aqI&<nG#G{g1J|XZ7mp~
zAp>&v>QAeG&KK|AzX6}s&#&CV)paLWs<c#8RMprAiG<J+dxVv9V9SY<CmG;IVMi5g
z2Z?pJYrh!ov%xkA2@cBrtmv(T;IVz5BEC`J&m_$tn1vgF2;K&+B|V<Ds;}=e!(_p$
zrJc^eF9gQL#l<8fbbobKwXw;>KTpE8H?MVuLy~u&^e_1B7+}ELK*VDR83&<%?5x=B
zK73e6N=k339L*|nQtBaG?XRGSJaNw@<#Wy?syp933Ah_VKe5A^dc<IZYFNFHa8Pf`
z;xIS)=C%6g;vjk&9B`sh;&rNOYdh`@v{VRePuwbHhcUuCc{$(LFbmtJB@AH-<2Yi_
zf9+HS*e)3R6}R$}?jGYOc=t=o%J;O4D$bkSZzpL`UVcYUoU;vdM6QZ-P#UEK;_n%G
z_u+4arC(7K=WtC!(SQA~TVFwz1<2}#mc1&%!sOWZ@{Q--aGJkHp@uO9c??@WU2yo<
z^PAa@L=<A~@@Z&VSSiVHe!r!g<%PyX9M3A~^JlB>4<D4>!HRDh`e2u)j&KA6F{5H)
z!nHQyqV9g$zGFw*A4NpCu;jePVt9NSwTyg+az-0e+KL`Nti}sv^5Q)NzfmZeZBkI^
z2|#+4F(71tZ-(rpV=sQ5(DlPSQOHKd_{jNX9X1eSJ|P_E0yqfOw8NN6<m@btb&eIq
zg2E7=vhU*-yHuD1VLuJPv`XZpxI4&YTby>IYX^V$CMBivmIWAaS|BLv?(W_vaqBdO
z21S`6x;*6I@NI}QEL?$VcYvDOcWcB3F}2oc=~e+l9tvOZ0Q5AKaP7pTrA0S4AHyy=
zfVi3zK$Hl$rK!ot$Z8rJ)nRNy6u1kfDO;5xfGe_da;m`{DLD;FjQ7fHS24_3;49NX
zo8%2qJa0Y@nP7mO+B-V9ZHKO8A0UFLc5S2s1e9&&doL$ymnbJ*))IMzA4gm`%=K)<
zm}mTU8q&f*5DRRF0oaCT#=BT>72QwIZN*YU3w%UGC8Mm23-1O1V=dR>Zr&WYvh{W)
z5I@a8Pr*eq0A`%)<J)j@e;Xg?-PkW8JpBf1DVqRu7%sVLYHE6ncQOGE(|WCT#GO=0
zAkpka&Ozkw(6k*B7Y<<WB9TzOcbt+Y4jxnz?eErs6iT422<`{AkM2iy(Ox@;4nWjJ
zN|)AF`*H=*qORf?4#SL5wk!8qx?kNtO6=sQys`50GiU03_<9y+RqBB<mRf5md}bPK
z-#xpqvkZ@pZrZYi3SnU-YOF*YhI{{A3XwY{30$2QrloP=3cHENC(JEnO-qRWh!_;1
zUlhNN(l51G;)SCip^M#Y{}E&O5IY!R#3Jrv*xkFr1O=|At0_$di-D!aS}MqyWt}G<
z@v5o8VHuSy=lU7*CaZ~}E3%D2sFnA@Osp?@B6b|Vd?RS1Rq#AKJ{`s)@D`nwbW|iy
zor(n7juP1gD#|@J#GDPRoFLcr>pU6Pe(-0wdQs`Y=xEI2-GCcijgkhuFO&=sUoPDq
zZAc8lE@hvlxM$BGP_qO9aB==;G}1caKETUQ3P90%5bG;J@q9tmlPlp0)Mm6v;n_h?
zQ80ykJ|;dMfzGd*kte{V0?A5HTn{2g-$+Iz)1H!eC<{>U8kbwlxWa5q_4DOw#pUCu
oPE;%7E~=^7hF<vBTg@tqylbajyB}9E3I92wuB}#}Y7zMV0OKzyasU7T

literal 41142
zcma&Oby$_{*DbnG5fl(a0SQ3_LAoTRL=gm}q#LBWyOdN)q?HmxK)SmHq)Q3u2I+1%
z^YOR$zRup?cj6B(-xpcWx}Q7d9COSu?su}%VmQ}HuA@*WoEPHH<WZ<g-Y68>6y{a<
z31{Ef9Q?&=FREhy%F58*N!Qi@^-|Z~+T6<C+(hr5qk*lRiIpV>GdD9Q<2_@0duuyB
z78Z;D^$W~awni*liHWN4kZacBs&*(8t}gNiEkh{X1cgebdhzUuf^+=pxQj~U-f`>J
zL1#YIy+H4W?@tItER7A{lQJ*Mb#yY?<h(M$viMX{l2vi4&>5lh=UZfE0;BRTpUjE~
zv)uBpm+0@GaB<WLk>9zad;7NG)pG5jJq?bXP7aQPV1uOEEQ!3Km75}%l)iY1{(t(>
zk^d0E-1&%li2N<rERMR3{9}e5HR?L$4-b`EcV7Ca_gF;CDwVe>eIIgO^Zw6ggRcl9
zpA4jbekmC~NcoI_3QZQ1vb~<A?o_1C^K5f(uX$o3(I1C&y8(kD`L}6E?qj<JY#IKo
zD}sW8JUl$I>C$%-J&#Q+Cu=Rr4LhQl_>Y%T74ozaOn-mxD>ivG;c=judhkMAT&FXZ
z&DqUuc4|skN9P`cQZ{Rc6M=|=J6ANbFeW9v&wNMBs~k1X#-^rBr5uvcQj6Dre`QkB
z(sH=%o1+Sgimu_{JYZl5x^|0ZYIe3Mia}YW&QsvUix=e;6*i}bHsoHXHfklN%_kdm
zVKp@ZMm>oJL%Ev%J?%e#CYufA#tXPbZ_T#8s`K>lJYJ@K9!mB?ULHTM_V|06=T*vQ
z=wCi22>art>_)R_cW20@28M(*PfjKU2V;}ny&IBybWnHERW|7<si${;e}DhSj~|<x
zO}IM)GD2H(5i*1)wL6PFi+e-bi^qE_&aSS6Ze15=lNVcm%<{IzohI4b53NzQwzf<C
zX;@-m<nw=jXO)+iN3IbOiYU<cqci&A86;aDO1^k8-4a4-zdjbXyK8@beqx0h8yiCn
z7n|H#US4jlmD&6I8!b&zHGP1<t!tz1A~Y-vx5*!eBfhAlWIjq+i;|Nw>f&s}%U~<Y
z7cVLnS=Wb`Tf-=fdXt~P98LfIE7A3d(`A3e2n~Do<nK_PjI%QjtXI>(Ktyk^?EU-q
zgIg|e?(#)SC-AX3uBuem)VLfkryp&#QiL>4H3t!yjFq#QmTYNHcYWsO=4bK6bNaVc
z$Yo(7H7o6>^XkrL&JHFQhqS#6;JKagytnEfR(Hj51W<{EZmw5td<+k7V92W$fqy(&
zO!7jdJz(SG`;?<zwmoLoV>9>jrvHt*5<Q6mK~fKIiUgiKKsF+IZ|L2}<t{rk9WgA6
zzm>J;VL0h9iRD#QL~b6!!l;d!1M<GUzSXgc2nOZceEl~3rXACgPwV4V98T*R((ybE
zViX=fN-ai*Yds#LU=C+H<2Xbs19z%Et^CeXmY!6=lv71cOCZMAK4a8F{&eQYk3Ylf
zl4{1p$4gmS=Jr*1d!v4ta;mDTqNr(T2wg6clkcILI6Y4GA22hA{`@K4#92{M!KhLG
z0on6uk}((;@*;r}FQebUYhdjDqzao_Uzb)<xszADeH~7rRoz+iV<OezQVY`A)-a==
zA1EK0^!tvDY1E$Wt3M*aK@JxRR#*_;Yj<TR%ysW?_=gYmGtEJnuhMUX2_8R0^$1<?
zUmW##-Ql&R#=s^20ecl*rpRoFd7&qXg~Qtq?|}N>^vQ;ha3TlAs(R}xGF9vA>-;CH
zMfMvL39#r*4sYJP@%8uLUdgMooo^?IkB`U1#@2^zWxv#m=8bmws`YdudhPLF2FtPX
zH|goLu+F+I!NmTmKSD!8SBDFH-rb@t)N8)}F(Sg`Pcfag*BK`)cJU9Hd%hVu`gi|l
zY-xcY1W`tTP4PY|SGz9p+sn`W<u<RQnbZxJZIOUQ$@snB32kv<p+7|wtJG>T^z&y@
zi$BG>2kYa%GUS^2)2<<(S9z9Zg*H}ZMZu(A`qgpmk3yDGh@i)DzRQlGWXuEPIl630
z6sXkH)aA)Kq1xKo3<dhZ9ChWnPl17$r)Ot2*4FoUcw%AUh!d+^_vlN^hPe21HLJLk
zb2Zq+mj6XWrMAD9b=1^IU}By<dv^Q8U0<JidU{&wGgrvdfE&cos^a3}tafStewV=K
zH%2`SU~@Wvg59_`SuZK+u@m!bOUP>&bg^;Y+Xs@-Q*(1al(jw3z4P+&<boz4Zb3-p
zlwAAwfWZtAmMaMf30)077;%qGK2FE0<$QzPy*g5a;k~~$YPUMfma9?m7G{H#%Uq{D
zl8$DQgMpzVe;ohe=W<#a8u`@cI12QNM;^rgqPT7H&<cdahM^&%5;c>PtX{ND2r0%T
zYH*lCy4~E|NV!DI3JP?)<9U63eUk^4oln-vHux?dM#u;v>pHJihAf6}g=VdL9G}Bq
z6ahaf`WNhjJFGgw?{X6!nc%}*__vl>O@6Z+uhgh@CstM-+NPI7(oc#8MC;kbMWbRF
zXTKZXiHV8Ek4R=_XG6MMTU%#mXYI5%OsuRRID3Te<mTq$qx|5z<a7zh+YR?3gx6cT
zyF<Aw{@@W181SSIEC;@SpYpmjlrUQgd3K|#st1;isVP0gbx!`!kdOy7G*@wPasLkH
zNJ1FX)zvMw7>$M*<M29n@9*!w!)Y3co<)efOTHFt30yKBVR7-mX`A!I>im3GB+B=G
zd0OFl#`h{+8ac&qq`54|RP^-%kqKx|RLXilOWWFSpkPv3T5A00lYC9foL2q$`FSw8
zz^8C3G5Ipfadx`=`x4U9EnQt9qh(edcaYtFKTIo+7M}O@E45_IeQyZ$@aC^pKRPrU
z5Jxb1=hf_!Adw8>BrP?yFWc)DR%GAMivOEH(>7c$EiEk`Ufwpf3c?gvOBosNRd1y%
zr5ya4Zhb7BTG)aG?nkytxf<`3vgYHA5_NZ%`id>bj~lXbQD4y^VEVFt`)$hWar_7#
z<fN&t9NPoo8q#m_iwGLgN}r$(_Jerzu}TLsLPnK(h#7HQ79<dmQ#Gp`ub`uEL5Q^p
zJ4HZ5pahLqXHrB&gdfArqwNbu2|n_j#0X)?V=7gSOvp)BFE#%H0}3YQ@b@}D3UHeA
zh(UsR6Rr@y%5P-qGpxUk4!K6|_cZFG`0bbQkPktSewL%eq4xuGq3g3)kzp715H`u9
zKse3gUM0wf@ETX89MQ_mhC(@vx;G(F*g;?tJXz&P-NcZE^bKHR*_K8A--K0%knw(e
za~b24yXW!lGK@biCvjW95PkFXXy;d!@<uL;qod;j1UKfLlz(C7*%8~wr91fejW8)e
z@iGz;DNlT^&Z|Wti~dY8G9ketx`&Mz#iCITv(98I2!I0y9BC%^Z)EcZkUy?mxiYo1
zBo1+PsV^17#Kc52ceKt+h|_FPf1xvuX3}h=(BP5P1Q`Yf#wSj*7Kqv*iC&8dPR9LT
zuSrWwKQiw1rhE16PS>xl4JEBobK<>~q1J{5G+7!4M)lHQ7|!(E+%t$FM;l%j>d)(h
z*);3@Cfx_Vy@VuAOm0@9mXeaPd2ql|bP0JQxvk|rS9aU?Tx4-RsX5<%6zSw?R<-2o
zHpOz95&B%kx&PpSU&3Q2Le)V_fK=Gz{IRM<hFhmQJwzv3lU^4Bu>B4CQbhSrH=D==
z-D7QLTc%f6r65<Kn3<UwwQI?=wY5XZ1=`DPX02^(rq{|QxvG0cij0n6Vc<z2eI%Y7
zhMhWt*{Yq10-yb`h}Z$_Fl$u@1_X$ic%|Hbf2IDkA&fq%UDjPtU0of5a2zb_Wv_f>
z#_7X?$(WeFSr+~@N^%yuO<U!bYmV^{jg^}_1_C1K0w2uv9nB)gHMQr#cNz)}I_~h<
zXS5xBf-P&e(5bsQ)!@y}yZ__et<y!J3pV|Xg9*1D>zcYcA*<RW8aOoU*S$wZM%;J0
zxPN7<4lJFCg_5<w)SwOzXTwzUb<j@s*HWD}CLSE;I=z}3>R-0^XDRTy@Z=W|pg4T`
zZ@*+e0jxS+<wSe$o;T!6$i&SMD&JLAJw~}5tQ(vi?`3P%9OrVp9q8}paat#P-4d)j
zR&L|J1AyRpO?r5EIHO9wx7}iQ_@_@2mX?;`3C$1^l&MeAicNpFTThORKmb*4inhXS
z*Vlls#uR~zr!?12nX!Q0t)=tYXxS8op>RL6ltj9k!nB}=nJ(mc6Qs@I8aM9G+*XFu
zjeb&z{BbpI2bX&Db)TC4%4mROBMQJhzodj|-$VY@D>Ri#d%A{(22sJwIsmS)DFow*
zISlbgNZvzMfk~q8kNCT~s<CP`b~FO()GK?C9DHf&EjpI-`OzX|Or3_e=uRHf0Nmth
z&@nmM+Pu>!(ZlWvB!A*AcV;L%3J40r$IqG{=PQ4cz02&2H$x5y>0OmB)5|pZzRXv&
zkVby|%6yfnoJS5g9K9akCY+Vt5;NsQ*S~b0r(3OhPm{~axFKD~3wb3@Hw6%#@K!l(
zeAB9Ng?AgVUWX$zzmq1>wj^UZQSAaRfSfIe9!A@u^idL*bVMUuNHxr!n({B%Qc}tB
z<|b92teBHU-cw<~Q@in&rlukH@ZuFxD`k&Q)VRHa#G;(1<?P~8;IyFyXheEvp^Mjc
z4ig0t(|oiv<iiI%D0f=ED`w>v(2-agb;cs770JOc%`^}ZE@RxVUKwP9N(F($G0d8J
zzrQOML%^E!Jn6!F#MHGkdV768ig`++r5U49##<VLxGeHsED1*DwGSVi+!klEM?ZYN
z_97;7b}mn67W1XZ^p}R-4BFwg=Tpd`xB)vNB{elp$V;%TtqsCh7_Z$zZ@LV{eJU!P
zp`*jYe2Y=8O2@VK&FQ8iKpyX_hlUGY+t}Lb?XQkh?+vQ+Jbru&53k-VujXn!<ZkD)
zy&+;DPkz9!2%3c8!|HKt&*7mypm6+nQ2>hyz6B|w;56L#;p&jcRcbw^8d@Ti#d=H{
zA<6Mc+mY|<1L18&o+9+nyf^AZ*AvSe{J(ibMmIbcs2^<5O9O1jn$+1TrT6-E$Y9Q`
zXooVPh?&A*IV>G#XXkf9#_MAh2C}&*6|XsxJkb_byV=R$wK4DX(W7Z|EA-iitMT>S
z6k}&!Ca8Kn=rw}gyf&LU<SX~M7Mi1U<1vx(Ad%V)e}77y*H$Bc5X7MSQLdc_g3LQD
zh5lwj(YB?&*?GO5lF>N#{PQbVWeJqosX|l~Tdm_G?+sVJD@>i_+wQtk<dXT$nnh6^
zT_<~|aKnw;_`JNIeK?uH_<hoof55Th1M`#{s~8chuwiI)NkHCa(nZ2-W0$>f+@N5#
zhh~doZkUxrtL@1u#<zh6p_3@)nc0=d0^cLD@h>B@?XgcTT7_SuGku2jBM-DkQ%~x5
zM3Y2SW7whByQ?K6hP2M8RL?TD=Fwc%NmXLooc?oKC3%^|r6SS|2lrXJkEC0qQp2sf
zhp#brcz=5@T#k%>av}CabKBuF9DfzuYu<zDQkSP7(H%nhQ(?c14`J<jPmXHQH?npX
zw-^@fg^I;QDoRQ;Qc}`yQgMwRD4$B4y4qpXySvsTCKy?69);Si@_OBpb=7ez*<5JO
z4Z+;bvMg0w|16C~<V5%Ss?0f2zZD_>;wn1R)rYUIt~h>R1q?tCctv>Sl_8on?C;<^
zY;Pd~a#>AC<!ROM0KnMV+Cnlnlzo+_buWp_9oI<UN05YX5)p}g=CXj_DQ;6N+tANM
zsSHm`mZ@yl*_rc^Tk#APJi*ag{qvzaM1Y|sR?31|Bd{CGr%uFBzQOLdd-A++nnAv_
zo2+`Dlrg{oQE^nUpaCID9g2G_mg@Qrmb&1eAS|!5Jt9b*i+)5}$x1nDKrW=hD0q?b
z1)}od(NWZrs~M*EQ8&%k{5h{hRj0%26@2qyIca$eERGi9Ndcx`<dnVdVJ3II*?Amk
zITNuGwRd%=Vo>R=P5@CuPH3|P5GcBzkO}yMGA_)VAFHr~<AaNO{P-~dOxa9@>ht3j
z_pN3k10`#F`$uOd2dx0Rqa>Ou9ai3J+`4)54J4u)H*Q?Pz$mhstUWo-_QAkS`E4pm
zcAou`iiJp!6kmLIlrnf-fkI3j2et;sc|enWQjt&833GT<R*8Mw!M6gx76IKUJ?9l{
z>>yvIr%`G8AM@%HMf3A<%oEV!+M;un5R_#vbye8O$w^*b9+d&bOl#=f%lGK$-tF%0
zvNPOcVhX9R7l!Koy@wz>`^U{qYbd^20Jp(RQ9pVV>FfKp4gJZJC;Yp;Pentsl=Db(
zUr0&m{|LNkK%&LeTvkc->ZDBP;D=@PpBlE`3FZ~n90@PXLi=479M5;^?7r$Rs8n`-
zEGP*0R-?N~v^l7H=UCv^%GXul?N$wyP1aTQ@5LlY@s}bMyL!oUA<oEnw3G>o2=~nf
z45VcCI$O~;A1U;MIxVJy8V~00Pl*{Zk9G32hP5^8+Mklw08da$_I#(Or?ITMmu}HY
zGxnVr8W{8y8OcC#(`wfvp#1$)b|%(vWO25J(pp_-VZvgC<KtT>Wen2i_Pa*<v|jvK
zius~(d#*esXr)qqpN%5zI7yy_5%*>%v6DWBDQs34WS#<4|1mOB!QARA@8ql}f3NgC
zKPI)}z)q<LQjmv;0lheGOmy9n;csx;s8J5p{vzz?$OXl>e*f1O{&v7QkPEoxLw$l@
z6QGr6H1{*yeq}J{*)$7ejO=X2rrA0Dm5sOdrDeLEse3NDPX#>RRWyEtC^NSAh|of)
zqd#Gh_)mA3O#<oEqTG}HHO0(GO;M5q-=_>ZC$!c(jL&WaRNnnE6d;?Pla_^$d-RME
zXuVBMP1f2uN-^anIgI1wHkWghg38Lu`dqB6tbjt=z^`eA`Zh2)_^V=;lK-P~z;F0?
zc(SW|N$%S({1T@0j8YB7)rAs`H3Pb6XO7%kRnLzXxh7XDejd27u-)s8@p3W|e?`Hd
zEy<eHLii`d&pTab=%OJJ(;%e5U_5I6b)^h?*l~0&0>?9~*DEHTi-=r;4LJKVoa(ul
z=+;tdSPIl~@7}SIeQTts|N0`L+~cHly4?M!kT{Ep)U@RN1nyD(`n>chD|0NwgTBAJ
z*#ho<y+?h)s<la%iSU^;4cmCjMGoj#r|(_C!BA+=r3(7f(Bec{bIF4UwiRKS>z?VQ
zOPAmXetjN70@aXDadGkep|#1nB!J)Sb_-HX6}UpDcq}X|8X6jR9y>C^4x)be@aLS7
z!`h!Hpc6NNrkh@04+RV&8N>Vl4vbvt^XZuxk<{lQ!vLyW+!Z6FuGYip-R^ygE0?9j
z1U#dttn7zr=fmlMF8O%iMMy&Y4-cJFcOU>B9v((^82uR?o&EdQ)LuaW2p=&L*X^}f
z9nZTL56>={evyCjYw{X@g+8u~S3!X9oRnmYs;y;7R9w@=#5Eid;h4)_&+Fgb8?M2#
zEm}W9YS{L7TDT)7(Yz7`nIphi+~G36lPVUrFxM7=P%r?M?G_^gEk;X`BO!RaL^a{E
zco$WBaqfovk70Lw8~jS9dKq&PlxYG&cFpezfTS$6niQmwimeY~sRah=sq<X87!TBF
zHLiOzkW3f`UX&i_+2V@>{>35XnoDq+Xc!%>`l;6hB@n}jo|M!jlgU~EAe65}zPRVG
zGH7smxMjE;or@VEEHQh!r%)Rx*wYhmPAQg^9$H1k@VqQtx63q|IcK-q(y!|k<#hRW
zxvS=qAP%{!2+eyazIf>!8=p9mxUZVUy`8AJ?)YGIY8p!PVv_-X_>$Xg>_@kF8XFt$
zu<2caa`p96?-&2Y$M#FYP}Shvdi7-4MC#>BAO%{M1hxYbU}Xt%lsdW-9;5Y|mH@b6
zbKcacJz5||&CJYXC}t75?XM!G>%rcw$Pqs8#hxT1s9+#c&u(nUzzhV37Jj@)oxiXy
zC@283*Lu1C0Se%rDj;bfuBTw)s}E;_{0(2px#E~A%?{A26>wH#?Fl3tEh=4=@vQya
z6l1AcS=?{->!}G2Sw5zQ>688khc4KRYgdUJRV(>m<0?I5W&HpwC4$so`Vo*`W=U|$
zZM`v3gH6ot3nby-!&9_)K8N=}Ymg>X?bsmvwn2GbpD_^B{(aWPrxiE+V#*n;XlUOc
zd|uu8aS$AVC^IyS8)WyWsef$p@bi;-|9{XO4-;HD<5kZ#LW8wxlj_bZx;P{c3J*sH
zeIM-=+MA@M4hpX<PJL;}H)31;gsY6(uwC0*OtBk}mKegMnFh6qXQ{M=WPF-nl`k-t
zzz86eo%uv{98wCpZ0ql@k7Ee~pO`6^dR5zFj{tIW)8OEKLr2xD0340Jic>A61nGnN
zBt?0e%TrL|VPav;0-XJqj^0xM?G$$#YK5WuXX{QR0AhhKWr^syg>I@)v>0G>*LmyY
z>9eHT?qLFq*rR6*iRQ1+%5;kC{F$wud}bZqx<;mfmhR~@K>B6v;S&4J@4t)fqo8JX
z$y9i-?qBjKN<uVUI$`No<}E0L)1cnS)~q6wU7*NTE%Jx3x&7;v%+_p}5DB|1l;0yz
z<WoHlX00ANKRaRCzDpk2hAKDeAqVC{aW(sPl!Tv>I$9~&t%EsptEL!T3m2okn}cMv
zjh}_in)f0Z&&SiAgxe$$h;XtvyOUqJAT5=Gap_#JA2Q&Rdl2wdV>=S)3p|6JVPFYc
z;F+jsUzRdUb#?V{l@rTQo_4y+&Vs+9H1Y%BjVUQ9u}QgtA*k!Ve%<7I?baP`OEQQ6
z^WkFTl#u3ugaz9FGL~0XKL_ts>qg&e2<xj+<%p9YZDcQgn5ALPX$lDBCsr8U1*&K4
zkekC(;EAhat{;ONit}?LN7g-afnV3(uHGKr|5DhxIlb|O_5vFVIFdQ&BLLh14mAxJ
zrD`a-flm(r67-JdBJ^c|Xy<X=eb^L$E0*;e;R0x0e){p*s`drs!gnDd29OpBY`Jeo
z6nSfEXh2^0qTA%ZklXfMp;Wd0pwN8LMeuZu8pXWN#ozbsMvr5LBvsz^LT7s$dp@3h
zj%PKvWl?RS&O%pe1ik&o8Pz|-1{98RU?@L9LZ7?4`{Dkm6_c7+T-POXerHx-<ya<7
z{eBaVR-h%}T7`Jw(8?Fus@9%d5?FY?vrFU#B?<jV#R6%vn&B@~W6mor8v%(MJ~gC2
zD159i8p?bq*JINdAlJ&FeN9hC`U#;FLft@{bU|1eu5@4k-Vs9DFR4&8b6|HOXkN~Q
z37y}&cQ5%T^#^4wm$b2<ThV>A$%J2+Jcb?CuTHL4{2YwfI&~2Y7wVyvA81Y#BN<GF
z@aQm>fL&Jm`)+<&pGxI!Rgr_YbY5dq&y!vA;j-J^q74-m?Kh~#X?(K$2J`U&mhLIg
z6VzRtKc?{5y%`r5_sHT;B*gYO03a%r#qt~@=o)pk%K}P@^aK#vSjtkeUns6zV(u;(
zCC(mmiUWXYBhb8SNjj0c?<b;*jv4rFQFjHKr$E}pd=@h-SXm^+1ATp)P$bZnMzek}
zy;!rJ=5n&EIEwK@>hFS`PKWLXU$=U4?3pLqr;^l{Pwg0dc2}!pzRPCS16+hCxRo~@
zT$j!P|FsGKl@xHR^pl&Vq^2kHhELZ<gMk6g!?816^V>XbYPmE((CXo%a%4;{C$}-{
zP_FA|$qnShagTo@P8wR-8<Xf^^KE#mKI<LpWw{tf`aEwaL-Ci5lfqkF9NJTXolAN+
zMt{p6>%E$o?#&Yjp14ALfkis$r%p7L9=<uo*ZZ^XK7y1eQZFj~0s?lRg*mmlYCas^
z6j@=n7;UDTG6O)~nP+r*+FUG)?{^SCbX6<DKXv(iHu*6jKHC^lTz#^vQ6B#Md1!PH
z_dta+N5aMUmKtfvbs=2Zr>xl<%Tk)SwAGq-GnVzj0wV63IuK_oXUIPZ@_J)*OJ(pf
z^Qte?tK<V(-17<8MK4~^rtCdjOY)OL@w*?g^78WfpNK@FLl}BV<YrnqJkUMaw8UGZ
zB_kpcO?jtfq7y5SyB`2|NVdk^0`ADEpP4J76Y8l-)2dyepP2WZD|lp!K5g@A<?f>&
zXh!BQ87seK3)Je;om$+SwcA|OGr(!6|8n3S2dCu0(XFQg0|Pep_Qlqmmm-C*9#*c!
z)3IOOB-h7Io?o(eo%zU7VO?|&uLWM)R*#QczNkDd`lZSFa*m;uPfU<N;_p=(oBW2h
z34hallK3@q>(a6>pK5GZ%=5nd&P4wuRB<JtB{56p^sPlgfK6#xLz~hoE~Zd;=>}ky
zLx2v4$YpJ7`#rWwOz{#_rO@NV$H&jms^Kd!{Y`KCGLz;FLakq%)Lk?+H#>9bLl5Q_
zjnq_DUM+n<o~{QqN-d8pC-g_Ld2YIX_=k^~V>CuK0-OvR8qw1go$>uS+8NJVV@ivz
z^_?~gxyl?HdHiIG6!J29{8CF+p)sW>!N!@9S(YSeVCt{2q=VS#HpL`Vxbx5QhHgQ*
zDlh*rP4cENU>iVibX<Wk^eEFuB#;CAV`5_Zj!3yJq3Qh8(h_}wCi^xuN(rf}e`RYt
zeOMj-NmNaVIa5`wztDDPeeq*r&I8U38~D+2C-wdQi61gvA>)<N{4oMInGC8TLxOy8
zpop?@Dom}6xV!Fu(o*Js((TAIYBG0G?1!iHH0w8sf7rDSVDCgz+0G`@@w+YoX$;#w
z^<QU)?UO|IZ_^`a9|ZY|P-e$Tq%}4-5E5aill;n7sw8UV_BIf3re5SpCm-#~>E;mc
z>F&aM|3_qcYvN7eW(ACR*Hyf&^T7j-n2vYE518AOQrUe)XM0twC%f>-qRhTO9#M15
zyi55kEv?UZokY;rfgT?PHDRbknh3CY0N?Mkz=45cyNR*+ny$?RlTI(}HB9q;59uzx
zoG-s$;hvTP-!V$e3n!O`PT;e;$<aa7sQNN5HGylq*XO7#DzU6ZSFHm_lk(wF+3^t`
zMx~CGXve~imG8)=k$yE7Z#UsV`<V0*`&B3^=&SE*)?b`VUi7Z)K{*uyQH~fl?5*l&
z;PrtS!_}VrUZfdqkA@{Y!ZsIHp)-qm&Gj@#I4CkaJf{mQ(8;&C_-(HWkMM*<4<#>0
zx<Z7(dUJ!41!kasek}KwmbUH|<KV9m2{zu0%sPC3Bw|YOL?YAWL(r<7x9)p-dS?7q
ztQr-+l#CKi_oMJ{9ATPLB1)V@>*i$CRD5@*jEnF})7RbAx6FUdnU-B<?++Ah@TZRr
z{cSgDKt;WO>uqf0^XkhB4!<TnPB$FseRQ;)Zo>ArtltZ2EinA>8sAFP5R5^`<Z!!#
zh1X@98U^L14s;Hm=e&(dxFW3k!3L`q^No6e>;;>%qeavR9c<Q!oZrMP0^bZ6Uf;vf
zW!v!lKFt-g%V;s-;G)Z3r8A~~Ua_s0B}ZKHobX1dQ>pAi&tUF)hEySJxF&(TERB|z
z*FkC_k6ypDWPE(E5eB3tpa%4*sRkeF2M@NOQ?fn$IxH^F77LUl!9S8@9j~{&@Cluz
z+@+FKR?uCxcV)LLfmww~r4}|EzsXvJ)o6TPlBT)7sI}et(seH6^R6}XBaTs3VlkIR
zyBnqb#}9q;nbWR&qoJYMuMUSp<9dv>3b4U;za&e6+kquC7ffs0nP8G1&K&X`(n(ba
zITII?==Iq?d!vyPx=XfpH|JvuT9LSS=gNXKEMtKz_tDwD^WzbdV6~e&Mz4vAnF%`w
zyX#nH@A=`)XDXoe_VtCw#@>N0fvNpDFwek9g@?;+8!*tbU-M2;QBQB8*r{(}vyvBQ
zkW{8_{~(=C%gYh)qq>c=tiKQ^!YY_Nef~~D-K5ROeJ;~kYvUA$zEs)3$US)c1H}{v
z)J}=z_@@Y38NzGcv@(g*ti9|U_x$on#=i-V$?;SXDB<MEMu_~V$^8Ajdhgbn#3=~`
zbX!B_7t15m-v~PUyX%r-%jUwSZ6+*AmBf^CL(Au+Aoxn2x*V-MapfWKefofr(A4Eo
zrBy(dPl)ex$}83}sKa)Kyv@H^dA9qjs(kiKsCqaEFP=TSlH_?z2VD&|jnVH>du5)0
zQU<DB?EJBb8;z1Yptt@^G<~rCR>xFOuJy{LY;yFT+)VZek!N!BQP;V})W^syGL@Su
zp5^te?0q#b!zy7VwA54}C%V!~_SCz2a_=6esbV<#4}th7YZa<IhrXtfL+g*x&nf92
zL}=C@sGW<Z$$uGIk;_ytE}7^4oR^KXB8rT9W)>HfzoQ;%*3ab5gf$f=wtxK~MkAbu
z%cL2eaf!XKgnGg6*{hr&|J>zj%fGuP@eUS7sV%9ptdFh-{f;CQcQR5p<r(c*HKEK_
z$gK2!{rYv6l8OtWI`X<W;|88>5$c0zW-$et2_{{gE2od}9a((Up5SeMP0PZMicd}O
z|Kk|?D|cOGW=Ma*7>f{><yWOM1Jfwpom)E_=>~Qjd1-7Xf|co-cg@%F33s}A8Nvj<
zk4)>aXby<Per`jl<Y@^M-Nw6nw+q_vyw02VKu7c%h$w6_p57%yNnqHM_)<fI3={$g
zfd$oB%pu2w`_4=Fc^6c49p$c+-qAi&f3)GnwOf3QZpGccpB?IxbCrS{N6_-dLG7J1
zH$Ez{jxL2+YiA=BMxJzE$r|#H9Ck0dJLem_cJV!;L(~)t^No%Dildrq{hWoJ?&gCo
zh!?aSR>gR`kP;=Jr0U4_E-rp_>GEY_0oO;sPB3r(53K^DYN7Acln(s1jIr@UsBjuP
zI<UN<!)<G47gX&C?T%GoM*x`jetmHZj>M_`=SxGO^%V~T$!_#q;KqONe!HBqEteVD
zT%5pfzPV#f-)FY%l)y$nxT8*Jm1acx*1)wiT0?WWl|!khtYLhF5ZB0!f7?)+h=_0p
zR~bY7S7BHQ-c_1bl;xk|cOW9cyM6l^3<EV>Vn&4Uwos`u2)OMtp6Drjzt7|i%_*dN
zA{s=929ku$ot@WEuSmu7Y$a~+n)FlGo~%EH+A=dUb8c>KvxU@()#2|iy}X|G_WKN3
zQ!HQR=JbIZ1Q=6Zw);$lBt`1FNiTu_1!IJWZ+So`nZ>&%J-rF%mJ?M=TCYpNCz@)6
z8)S<QgIt5!QxsNEg3$Ry6fn!_NhAqLJT~{Ct+oKfgMrpm0vhY%$E38hv`Tqe7PqfA
zH8zS%NX*1&d(!<rVbQLQ2YJG{1OCt5<$*Z1*U!Uu@KK`geo*v`(ox?fcR_dc&{=-(
zo1tLOCqNfhLH*4rA~rhBw^6T2@5u+LHu^tL74V_0mnyq79XT|!WFGW6;;v<AX*u8U
z!!j3+df1!=Md2@48!IDC@qe~O&>}jlpm+*cYS5aol-<XwvYEZAmO}_cU+T`p#Kb9J
zg8A(@s6p=Y^N(%v)S=4;;%`}AyLc}f?Ke_`GDUCyibvW0K`p1TtBsJGR{gJAw0T9s
zs)lC4?41p+S~jt);5XGDBO9W=_FK45>11*FKAE*v!6VcRNi36Fe>@`J)&0EkyPs+-
zwm0<eKnUdz>?>(P7;t;RHCe`>sB`HtTU}c#C@W*R(0ew&*Q;Y7|1<Xp4#B6F4Yc+B
zir3@}Mn*EW93ARMOXhCr$QRkZ^lc`?A1EwO2ya>oCi}=?8#UTVCeI!f-O@>RH;z~J
z2W^bI)N8?#>l$Q{vc|XX?_5fe3$*`<8^h(_+Hb)m(XNjogTfb4LqVal2*od2B%3!h
zm>-!^GRXWF1<LDa65n!`@pO@~n>=MXI_Y?DFrWD)%V%2K<Tp2enG16DVZTPh$q_Tx
zd1@W|8Z*Gjzxi}Ag2N8ypvcx?Y4=X!yn#gB{P&uW%3(q=?x#A%Y<+wtK|wynf=Um7
zzKEoI1x<29T_g4B(KQsb&cDedbwl$|FM4Kmb=7#_8;+cuoIjg4UqWa`#>18{3afu3
zGR5KFN-KON-Y=>mK;oYl{en_NLGP2u_`m^&^+OIZNn1lt>&}00rY%?mjbq?2UJF;X
zqEfymqdQ%mt|)8#{YoM;8R1J`aW~SE>_wF~vGG^rRU{;?BTak*E#d#@X3!(w(LyKs
z0X6lN|Dl`#X%keiNb~#))bpK5LP-=}r{PdoR{b2V-@f<wO&|uo6L#%trquQ2UHV=3
z$$o-sl=PGd(lVxw%w+=`nEozL3X(b5%Z3NrI?2W_59y@KErwVX)BT!>9TL@)3VJ=w
zx;&#3^jwXi<7K8H+M=Ym_!MMh)FZROU|?sXP2?Ust^!r{_2<!lXsR3aDWXBxSWhs}
z(F=4MF1_{f=_tPOiW>DE!_hp1o$kt1Ti(<cT1NFSJdrU<24yqj7mUn|G?S?eCxS`~
zbdAkg#2riWQbP{m(PFf-JE5h8MtkFt^r&xNziuLU!**W6o^SpF8^Q^&e-G~8zr@YW
zy`yOP1deXpFgt<c8O1PPLu?m2-7`!%lchb^{=yO`d&SAe&lV=@COy7m75&b@W(v)v
zX%U(rzQ2=XXOo|~d!h{0!VM5k0_jm?JuSlLxQbbyDB$M2_gnee_3KD;cp+=-ExZK#
zmw`qnnfjwO*_uSD>mUx~=T*}TzsA-aV)N+6NA#hE=qEd=MF|U`ZS2iq4vhGWcP(s6
zi_QmI<<W(WjEq1X^&>S*(A3^@u{&XRef%>p-P?J?*Fn8#KqF9~BQM@$35UFdapaya
z-hm{5#l#$@KLQ?kBSS8f?0bSl*I1-dmEvM;EcEcZ?n)lTzMkHcXS$*ZJhxtz%lO!s
zwUZM|9lc*hc0B*o@`kRejDbsR#=#e(K20O&Lq-NewJ|br)yR$1w1PMKc&|e3E?pct
zT0wcfb6M^xSq^!QbWhGMnPuAI*pQQT@$Y==;ig?<x$|4<SRBNtNw*CF;82f%p}{d*
zZ*Feh8dSGpGibkEMDXxg+RQD~h1wONFge2axda)EWK4(-$o!G%*iPv&0}~)9!_eGA
z2a1804kkpMo!#0z{8VB=H91GE_ejv|11LO)Gbm1gi24OxB<R0gLKs_!MTo$QrF!$N
z3j;w)A*QqZ2%Y}HrNeoHP4i2FxT=-(z_z6!$G>=B$~c{Fi<!f3o`uZwF%^bPOg~01
zhQV#r4v{*cNMzCaKxBeQ_lEb+X;#ase@jRIb+d6gS~kTmIg35+%Ki42uJ_&h&_6P_
zL6n8M@>@CMs|uppn~k4%UFrvfyhBpi>GVf2qH{s%^^8E59rgV;6qq4xPch}LZqxdq
zPT6on^GeJGfUWp`1=z8nT+VY`Icez|0k{-~K-hQ2Qc5ag${Ait^fhD4ccaX9dl7eg
zX1pJwDs*0cNC;frgju);)StHJQChWWAwBz5*;qd=qFPeyf{V*eUJ&Y_3D+6fcBdZp
zDyZxf7n-lhrJ0!P{It!`BC^!<PHLqeZAEmJ^T`TmtW;D!=vaiZRgP=h-^BvIceGvh
z_e|oZqrP71VGMZ<C<S>)@Re0nq)gXy{-Zi<YkRj*d(sV+q1^m0Q^ovOKPl{@5=7K)
z6x%BzEDUi3nl%VGo}<a;D7`8&d<m!9fJKM0{Z)O0^&ac0s)Dh+IF_U`^y5?T!~K#s
zp3uC(3}XFsTb?3u@<!;kkQO{dB2OVg7?gTHAyVl&lij@eL{=8Bhkpkf^mz0arhXai
zkGMAHhM-QfY(^3u(5a9l+*fkS#HFRLPJv|0Hy~i3%t{-?3fFLPxz*o@D5!RG!gh~*
zL1|RhxF5!-8IFkGK(ZDb5|Sp0WiZwU`9<pD{Oky1L+=6ubpU%AGROwf=P2F7R>>Hf
z<i(Iy2-)KXaY7^uwB$g?2;>H;)N)(`I_0R%%}v%~?CX@B$-nWIc(f|&_C>{cOh#Oh
zDRhRU0Vi4lB#ad9oB#Ak+I60&ctMX(Ae)KJ5vIJ-`}*oHFKpX>o!=ZEC}*wVOU8s?
z9-%2ZJUpBa%`d@=Q@f^_9?*v-gO<sVA)oO_OAAmdB0+>qn3$N*0PgC4a7~Vngw)x#
zf}M_fi~p2gn#K<rcI8%zdc?8Jnw5xtwZFp72>;eC8wZExmX<mEz~}nJcXSnVTk4g+
zfLJAA(s%vUznQ^gP^NK1Q+@=>^47t@XlSuqZokv=(O-x%JJd?{%L~e+hIbcg#bhw^
z1QE@bb^d>}#_{VeKoph%s(%3Y>Cj7Z2T_oLN(~-i?_av5%|ctRwK@ZN5M~C-7dfM(
zBqgb!brKjDsFxHL98As6A8#`7&1R{W5}+_*5}>4Puzm@<0#{&HHhdQ3R&xsrg?5Y5
zge+R1QF9wEG{AyB1>4ZjtrT}qhMYcLe0iPt2JwskoK18+#n&hxPC=_=FjfD?-|qhf
zomK$-k7Ri0sH0*9+_;h%?KR_}*IQUvh#dyX$*rB82sZD=<FDi1ER~IG9lhm5*y`w}
zAyB6PiEFTKhCKyRx)D(L6B84oVmVk0jg0Izr=Ga(F6r{5D`hQ%zz{lBbbW3JKKAv!
z0-9St>2xYk&4c{ubM-lB(HgqCx~xI&Ah3A=3S1ty#@>zk7PM}2|GP;lmG_ZV@NGjm
z?<~B8hdu#xxHMYE0tL-w#EwzD)k11WxBkCNgn<rIPC3XgD42%U$3j6Xc`9@n9$AjX
zwo-VF&Fvv1H<Ud7v)l6>xVv|KA1c>A>GgVJE96>=#6?WG`>>~hM*uM!4oIUKL0n|*
z3EevLvGP8~=BOy>6j<*r$$_fB1r!w<t~p@V05$OQ06Zlo>xcHvFDmNRMa2(C3-@4$
z=DWhf;8o*z?F_>;g-Ez9J3*P>QNxOok&%HU4rJYP=sL{8-)xdNA_|yC5h8`Jb{wz2
zD6|)aXptS;hKh}jHvF|6!@|aP)6MO~mWe(oA>l4)%PHyU+pVS32W}D&h(gD@OWp^M
z;B5<`KO@Ev-v{R#gZ#LB3IAEBH0b`qJ_i}$<Q?)X1$ukFyP#XW53r^4y7<4Yqd)Ks
z!Xu?j0E(f7jD|-}-rWyYjRHM%U<l@#iL|Nz(^ZaAg+X5LPTaIV|N7x2f-)$5U(m$7
z;dNZS11dvg?;{E{M0XFC7C4U3?j?fuD`UhZ$JJpu^JHu+tk*}|^WKFlTTq<73ksro
z_%IOSG#(k522GL|OCdpghV{7`x|z1@I_}wxPX}fHQVQeuaz?u$2NdW~x&UAu!ur6#
z)8Ol>Pq#aMeOK%-7gYzXiyWmzET)?fMwMyrxr^~14AZi3hf98VDc9*8P`3dip{I4d
zfyRhm0{UV+|CbjJA1&kHRK4-w?D$O!`Py8907c4a`lF~vumiMdAdV|g#VVnJ&-q1t
zPRqt}cD)=wa$U6yD{)}>$ysY8>+n}@g&6c6_<}-uMN~Yu6}iyq=4Fsne0}+u3u#zk
z-{mWAXR5o8CaZA$4_SFJhO~w6tGcEfmD$Cb%gH8I#{C%X6jw`{%dQc@&U?O|g>?wk
znES=qLAFN4$KhcW5fKq*aJ&$+>G`U@FMR~XZhzm(bc+cHL8PsGJKUaV8@I)pPP9)n
zzC023$<<jjFOY{2%U_N`e)$kQD%&$bEVkfGDOR(nATWIi7*2v^jV69P<w|3b+=m4D
z&G?YZzcxM&@ePw|crYoaLyS7HokX*qto4Ap68G8j=O4jD0=YxA#8k0@$Q%crK>vW7
zlzF|~?Cu!qMNDu~9zllvE$%g(lDEQXh2n8Yj>}a*+w9Jrgwu{#hpyt1OTnj&_LE_O
z43_skSj%5~o6glJbn-^TOSAk}`)9>53=$BXi3LT`!!i^DS}aj}37Oj&<TlRW;n~{X
z=Q1l86akBui+o;FT$jW>{wf^X{_BZFHopFpE=BL3Hl&%pB)lbpX{PiY<qSGz`kGTo
zK`@EkrlP&Yd@9$=7ke7EwNslD^@94PtB;JgJpmS_FM69u@4*vz>Ou0*(^Iwwn28f8
z<r^eI8){-E1m=G*P9OW^_!#^JZw>w5L`19NGmtJbyS$9(XTMjSYP;#_=^ZJXqk~6H
zPJ(XB+Asniv&X_`EcorBc|ug~%r}o@n}-TnCKY_JMczMT?ijvZs{;rE+*5bfCz-%h
z@~l9k`pX{X{LeP;;iYQT=Fj)$=}&94cx9g?luB^h3<ysB7O8v>T&vVqtV41T$Op2g
z3^kS1MK-6)Sa_SUwe%?GspoShT%>}&Fq|FPiFCj32!qHLdd{2M+w6Q^5IUiCp-*W{
z`NQW$?f2sAi4&NMPQsRk4R(&qy&0bie@M-0O=&dg8NhHvoqCw38l^a~L7=2-WTY%e
zaGmn|4P|ZP#}d4S15tR+e0f<soaEA{DP;7Vzf83Am2JAd^(%q}3^zj@r2(2`6zE03
zK1IxQ3`GCC(9mz7vR=rN#gsc5;eAq6+x}&sp`;1D=>2SH<GfJR>lJ(JgPzflueiV5
zt~@G@&eZ~(6%U*_NaHZ3KsXE47DmYP*UsJD-8a0vye7OZJOFlesb~6$6XoL(I-VrG
zweczFOYUNJ?w-IQaZqm5NhP0}(?zDdQ6kC|TSr_R(`+!iVSIc%TU3n{wuDM}wgOL`
zrpaCYqHF!RMEvRf!s9n!L;?nsZj`<#Rm|gBs7M+mmhIb929B!2smZuKYoDjSWnCIV
zdqfk!4_)Lm*ky-5Zprl4dblBCmDC*wsCsZrPIm`n=Cun5KA;{xL$_Jcd!oGMCnOy6
zYBN59mcNR`ouX9315zg?xG739gAk(+^mxN}JdSrM`S>bNzr$8LnbyL6tL=SrWom7Q
zRUiGzjaP}1b!c0y8^^YQlqDdR|N8YyFUGB)q~rlFZ-qt>y%{iKsW9h=V+Xpn&2#wN
z2mov#P`jbvNQIVf<7;@vsP++*n71O(SV0N?^XJd|i^@4_z*sT0mt+*$pB+08GO7Il
znGG*AzafCl0c0$;R~5mu8?KY8u;9ExmbLH81&+MVxr`-5feU&RNGW1~T2bJ1(VK6N
zBCtHwwzOo9i)&}T44pwyNA8;<3L<bSX6o9{9A7i32f+Xk4+(%Iqwz|IcyKhvXt}=Q
zo^bX->=)2TV7H#C2gx761CS`puB-$}r0KzLisk}pt)HBv0NFGw{s^GocawF7U7xCt
z7CuA8L<B?yf)4(1ik_6dQGCTBahUSyAOI+~=Y;82T(4wC%%Dp{dG4)UU(?cV-nsJ=
zpfoYRvvv3)6%`dAX`{A}4?sYJ_-{m>KIL`z8~Nn{3nVA-Fy7A_uXf=8!&o+0zd+0b
z_6g(IA4sDiYUKc5MCAh)(GB;UF)N|&!V*Y`1eIneDG3QZXvQIeaPCR>TfkR9(c%jV
zEjX6U5tqR1@ia-$LrPExdh&<}5EL;el|ub%b!Ypeh|B~uURUmc8fb3C3*4S0oTh;K
z3!e(ZI~?xx2)(>o|36V0!yjKFVA%gUiC27Qyy&0Nj?M)^V0LI5^Aw_?0D=_JPeHDI
zApuHrFi--s)5h3r@c#tV0Er=84R#KWKGPCJodwwN(|-YKBC#XqY3&v(HxX5{8pQ1D
zaoFA<iVOxj3$RyHz~1sq!ZH=*Ty~yq!6{Z_0*3GTVogmAq<n@xH*n{@sr5M7h7>dK
zU6Fx=%lsz@y`{8-fG$OhxIE0-wWq?9PcU^2hh*s7u4et}b&O2Q0(h3CUx?+595f((
z=xl32yz=0N_TUT@A;=&4z7nvv4v$I(y}=_GJZFg-rK$Pis@X9JwWCtIK2|cAw+dz_
zrIQ3-!c=sWWVrvg92*{ixDx$oiF^f%WZgf6b9oBUp+j13Ym?g_P{>r!sI-6QG~p6J
zo9Nug*bHjszHcu{!PtVm2;d4hGH4a`F>*pI`;aOmqQvg}$onege^w^C4aH;BO$Nq>
zt?g}4#QQ+~(wiYi1>!!5u>>`Ipou}Tb;oeK1*)K1x0=9?i85fE`7dy+i<LnIlM>at
zqN_d8O|jPU_RpNQ6@ZBOII_!23kxk^vVd|=hzE=T@JO&&{V{%SsmfzLg`%WHzep!1
zC%+*KHW@H;83DT0sPjw!b0$8tlA-c|b}=J68&F8in#Z^BLm5uauL`_ItQ?KsGk0BU
zCV=NF0@fOi2B;yq0R2KY15>yCZ=dz5k@F$%;n00aWh$8Q)N;nxzb=&n5tit6?m|XJ
zrjoDIpgEZhI<R8%k&m!VwzJ?RBvLkO>k4-Qpfrq-`v;?di2_g28;}f*@A-l5pPHI_
z^LQl>tkkQvOwd+c|J%!5xRD11>svXb%w^iIrGuJi8VqIO;2_T@vtTCc96eBu_5_j&
z-zC|^7IB?|e^g8LH*AG8Cpa&)DXS-=w#mq@3cNm@i^fV%{~k(&c<@g2gK84-V?zmS
z>O2AMIPj#Dj0J&^CSre)>FoZ=Ii$*Wu;qh@SiV8g3i%vqZf(P#8ITtNc!Ge%Y7_#?
zQN5B*x$pI#v3_hb&;d_)6BvHL&rscW1icu+^TXHRkkpI5#Dw^58{gkOC--!@ra-cE
zBxvtWtU&*dPf$vOnJh$S1ywXSK!|uJL0SdpMAK6d2J)hW>`E7gw-(lC+9uqmlxyGq
zq5dqVp&DReG}Y%$!3y69TR|&V^coC%n7858q;OqDazOycXPp<c<(mO%41k=nl=LZ=
z+@XbwRP=gKQNSL)=`D0|8UX?bD736%%Ot=kA)5Bq@pV1R9i_0hr0Ke%Ue2o)OWRx-
zoYA>p-pdZ{yHgJy@|ac8|0Ugd7mRj0<x-op&*Y?k>_%kk9$&*$+LY30p#yL==maqa
z`Ycm!|MI2%oS4c5pMLt=R1EA>vQLC}2LM+nJ^^LL9MUY{wj6tSFoTdl#je-20+mYr
zI>ddWp6bq0|LiKXH-I=#@E*q-?4w9c00PmOwY3n4xNyRQDTES1!NCy&iiv|S&!65Y
zLg%ag%71Rh=01XDnOJ152_Hs(Z;u9!)c%6kVz3~MNC7ZW&5NRyj-LW%s=lWu3@V)2
z^>xeew_PtQz5Oz-NLzS2u!yuU<!4<D>j_YL%BUHIClV|At_4-Q|Hljza0L!4N{DlH
z;vP6EL7vxF<K~3;e*V)h#*6x~#<01gQeLd?a=Ka6YeYEAt6s_sN!I8A$>nm(aXxSj
z7Tur5t#54&02D<~B?wm>sBo73!P{rBkHgHE+H~3!RStY?xZDm$a}}&=Q1tf|8c0U4
zI4^w(2!v2OlmhdmVQhX+|0DhtO;b9vtITUxPwK9y;yuBX^SgZwak*-gTQ|&M0iU$B
zx%sf?6MVq^_XhUG`DK!WOBRHugg1mjy@RPGm1%tO4CSa%{{H@mFcNMw2-1j;ifV?Y
z!TWJ4IL7q1jb!c+UaJwjRKgU0%EGve9dn>w$qw)1BueuJEPF!WaU}#3ua@VL0kG%(
zT6v40L}k>fj)iP?gN(;JDJdz<l#2oA$Y%uB^Y^mJamtvzB`%Mr*Hh)aJx-KFoI1Z5
z`lvquxdcjea<yXP2GDfe;jz94Rzbb(xi$m>!FE-(I3#^!jE9)TkLDso6V<^%4kcL{
z*aN|&*m>#qHOg$IHLlG@Y3HSDy=J;A_J`-!#9!tkyGV|Cz4D)|fd@LaD6mNu9v65a
zG(R{I2^f?R%BtxI%<jmS1Cd~3XO{|(UY)4<3~737M<5MAI{mY}fTl_MlWU8H-bF;%
z=a)$QRgjq|w@RD1-aS5^03RFx?3jh?g@WM%nVk0X2wE7(;+YQRz643p(m*;cwDL-Y
z{orVRlYg>ubMy4ZrYV}qwavAA!+e*Ed82csL0TO=a<%-KnAkv_mzTST2lTSD)yr~!
za^Ado^YG*ZxvT<Q2;huji0G;PKOEz95eIOAgNRm_&KQ?5$MxPO3OijFA@NFmWIKV~
zfeH2PA5szrA9+EK*z@Q3Kns9y5K&qG$9DhmgDmS+S^!A_<gOV%ZT9{u+IirB#!`0$
z^K7x{!@SPM^_YXkk7Cn5yqGjW#l_D`YUd#fzz1F~tZfmx03G1V1h4aLlw>sHG+a<I
z3%y~znIG>^kORktj82Y^NlFeiD{TE>!id}~l#~b|8~pQcUyVGs115TI=x*ZCT~8O{
z+~)eS9{TYg<KN!q_@rgs5(#u`?_u>z_kj19c=_Ku9WXpE;JN?Rq~5G`(ng!a3TDXr
zi#ZMpMR<+)_qT5V`sM+TinX2H<!oxy#6Powy8eak|FkO9$%K&${G%o$B~#%N3Zxc5
zVkmf;pdft;bqKW0;2lTju%|$JdXt>o(m8uzSzJmAQT}Hrd_OsPOQ2zE%MQXFaHgvo
zio|?8b;Nz{S$9HPH|`TYxpCfRb)eHedoTcUhp%6!R#wb-j0n*3v(3jVf6v)nzKX2_
zez^m8TsULUJP7Gl%+or1BAIntLsN5Jd*Ouo|Mrx*gDS7>QGwXP^RsRB^KI2;_JjWl
zg}!_Tk{1p(Z5F0C2$m2nJGA|pcoU?>6cldXee77Q0|y2;x~Q;`O7B;%Uh%r^XHVOh
zn=`Vpv4M9^w#-r!S|-=9uu?&#Q*1Svc(k(^0en_00V888T+0H)C&4la(E-hW9oZeb
zGf$`S>$!|;075@}NExN`^YcTz<0zz3`9H}c^U&>oL;a3jgf0vNf_Q#6TMzr+l`y`n
z6#w0AF{pdcK7*1HU<u}p8-|dyY4YHD2s&^Zw~UY91w#_l6gERE3|0N7j($06CD*~*
zQ~&02N0~5#@}Ef-$i>h~j{__NJ3t0##Y5;|cPZpJ(5+c<Llpt~o%gA!RIr$c(MsRY
z@KdJR%FVlXUjlmF+S>!|NdXWfZncJxY5|^+z@k9_(T}ZGA&A;PP%kpX0g1s^sB+3H
zE0LrG-820}y`0Xo)1nwFNbEqLgDydVrUZ&r;A#Gw!}$SxG7z|QzyxAGS(^Zln_65H
z15Ur(cK&@FLs@nIx#JPDMtMD=Z-&FzrI6`BzqS#Rknj?WwiVXXmq0qV02Mf+W@RL}
znm$+9b}+129e_ce)n?`y;=@83`=I`M7aTkX7sYVe%}2*XM?VM8#m$=!MtgX<xS~DJ
zj&A~)FC~SI+#LXtgPcOjZ^d=mz^lW6<_;#k7D2oPw>k7e1iL{YsL>szmV^xK<HwJP
zSpt$SOO@VkO1s%)ZO!?7j5ZhlplOLKlzeBjN$&C~LvZQFU(jOjg0bOY9)8Z?*{LI#
zmygBBU7ppMkFa~FsKlOYUmV8})0ar6ot+`=1`o&<y8zl3==)5g_RHLk=41At&g9@D
zA}7z?fETV9G+9HP!7FdLc8yxq<yegrru9{T<Kpkrr%yWy3g3UBuG6-G4DclhloKOF
z3U4A8%1FW7eO_}gF(C7q>m|jVJ1R6uEKJ^t8vltsU*TGe|J`LJ=IY7^g)*FzNaZ-C
zWlygvm<$T=`M^aMuS409H}OO+#91CLpaKZT4kZ-gDnObGa3h9Ylm)^1VwoFadvV@O
z{_^<{?hH5u%hc}L=nx~PvOZvCw&NT$C}t&3oxZYugGlq@MW*dVSGhKFj{)Kt`BQ9S
za9nFRdhMA?I$(ad9SLFD_%_zox<DkZnR4I^zJIvxgK3*c-@r+zZXB9c4Y29KU>8hG
zUu=1>bCLiRE}U(I=YzpIbtejRNV3#C1I=gL#k_GBGD0BnAj)4w5ICq#A#<L?ODRKv
z+W#I9%6=LAE)5mcWic@^EJY<Y51>20e*1=a7gQ^6XE`l&64OY>{{*t3pnz`lCk>PL
z%o}&$1fhs{55@Ehl$y)7*=isb=n*&|1NOQZWWxRP3gUbTH()Mn4`;C8HJd;Y6UXnu
z4r!*q-bBd?&Kxq22=;(RdVygViRZ~0x8UK_Wyu(t?;j{U58r^vd=pUf%>29oe8H#x
zaa{$h4P18fwR2Ki4VA_7Us(K#6*D2Y>Ot^q_Z8AV1ABUVkyMt#pGoi#+X2ve1H=Yt
z_p&`7%G}(X$Kmg7gdCqt$@vMVS{a;K<`x!x#*h5%;Cct-77oyQgYKfe#L&zJW*?B2
z^LoW1;sHZ}91(a7<RSzP0As&QO>kM8TU*nIC;$BZt*CY(p^I2v?>>$Ofn8?%-ktHV
z$jI9OvFgE|j(Dk)Pr0Aw4S_VR3GnFhV9uyVUEL|E_WwgAJGTO1s2PL8y&kv<lhM)~
zUWpREw6itjeu%zQv3H0)4zdqxC*$tU4u{bZp_9<XDu|rkftZfuEl2(GKzb%LVWO^i
zS$oxKy1KfCe~6*@x>rcGr0D>K@>1DVXliw6UWDFmftx|vfDBQouzdhln<9H&g11>&
zX5klxry!#J2KfSj0W=Fvbtq(MUcGqcR|?{5pcQI7p{sBVtjQ7x3tc>bdfBtFv2m@@
z%JYI5+l7C4q;UZfslRKhtGf`@2DPgkyR0FJs{Cb8kuveBdaSsoS`Xm?s`0N*Ha2}>
z6hc~Fe}B=zAXfIkSbq#2yk2?j20e+BQGk0dH)dv1v@2}W&f(M@!)mQS|1!htZ)Q5<
z9swiGg9oBshnu=*Fxz|Hp0M`EP|>ciXjNx`C*T<F9QvE&w$9tM2ZwOzG4KD?+ndL8
z-LC83UrEXkQf7(nQl`v=(jXO431y5-8Ol@{GDS**D48-Rl_VnbOp&=nh>W44kc^S%
zxU|-K?!A9&|JHu~dG6O<>$P94Zob3yIj?g#j`z9!OSk*5=mKKtB+@qbK-I+VT*t(u
zV&8qZ2!7#_o|Ki9okOvhxM&HLtKL6fa_E4F32rt>{wL9uYHkH7$ll885r!m1ztO*w
zR%%UBXD9Km*AbqupSy0G;Yk?W)=7t8Rd7s)$;aLEi+2eE?69Ao-y|9*b1N9RJ}bj>
z=Lu|7WJ>nxq}gxarTj$KyH(Dv?Z{|PPyF=6#JiH8P3m(7IcFnF9Oyfc#y<DHI<ROA
zDb5@|zO`-`UIuqHG-!6i!LtCG)|sjN)P2_8CHNmZC+D7g^7`E_{8WTGUM(#=+nZe0
z0Z&x^MOgZS)!~1_@$c?zl8yX>@P-uiI;r*XNB@l$hd<Jm`1e2NaXhiWV#I=OSjM@n
zDEybOaoR_~Yp}$lP5!N<%>+F4uODyGQxhr-Wj`0vJS0`t?HQZUpbPDtMfwBj1u?YG
z!X^3`WGnnbd~7)MY;GlBEi@Il*rP`R*T8+mrI8O=wPW-6-EJ7>Qy!lz*$ELVQVJN&
zY8DD0Yu?GN&JN+rIwKDcC0>PdyOOjdyyjhP-aoBIOM*knk^!U0LKK!)Y?TpS0Q4f1
z5Ri;4Vf`ai4E#o}XYDXop_`n5GYaGrW0rnBz5orhY1{Ig1;2IVSOGdh#`EAfN!oWQ
z0Tn{^7&iR8l49oj%e25^DM~`0HsHD)dVl0L;o0GS7ikGc0~7382v!c?KFGim2O-o-
z5~Dydf}fm1SUkwwYY+_FP<cY^PAok*O`w&mfY*nD`7r7DAkTdc*cFA~aJq2jNzpky
zE%%EQB8E#OIi_7{trI8}Nj>w^eFtiOah`;PgwA5mC<JQGXM6x7@y}C1P0Hy5At5z0
zT4Wf$lSvUMYY%&i;BEihRip^?7(CY#UI)>v-)L02XXQ8ncWDCkcG6><2LIXjycJy~
zfX!C#!qp!bTdV91CI$uCH){07BB~n;iP87`o|TTuO9(>h*Y<%IaW`pP$%b9uvG3(7
ziZqwT^VUy6mDS!Q7=yDS85_w$*u5Bh2O~IwBUB1NgoI<$S2raI*Z=0US|*=q(p;h-
zBOjgf@-g+m-?|%y4hcYO@o!JN_FWU40htE#)DyRd(oU8bKbf|L>W(;AF_qjXBQyDI
zbZBIRVFUecdVVzxeVo@mDo<R8B*DW~i1@r8O?}i9pK#-$E(k@uTvHORpK1xL4-Vf|
z^aUktnzx|Jtj8H>15^(EyLmL%eAcj0Z9W`Z2Pj}*YN{x@^m5Dhl7`RHJK2Eak6Nh;
zc)~`!sS|~+5;8W;@t8V*lb6{%+6o>D@EsN8lE8!((^*_cSMiFBU3+ON;AhS1_hZ^=
zf+jl#=`1K9?ChQb@%M5^wSv!5388b7d<uB*9xovummcE(8*MF!s;<612@?#wb6%+6
z$R!W)a^TePpP!;8<VoLy9`k8Du*ef0B2UNLuS&n2zZ67t%qOYD|MQ?Bt~U?2g;v3Z
zGZioezaY_u|A|lvlMlE_%U7&e!@_cOWCD~NwCAPHe2_+yUj@l84lK@AM4r}CaJydi
zmEHF{%6UAO60`G21URAjglKIU`8ejGC-Z>P4til0B(;fbB>*;&xl+hzcxdQ8UEQ$W
z=a5WDNZiI___~-7I^a~tqp@}C)|t7v`?@KNg2D{DN&p$E=tm!=NHtMH;SsI`cs<`R
zafsu28lFU&IfrNO%i`=1pQoDI^<vbYbIAB2=5OLk{X(_<Q}Eu^<iBpIok!`}9_;e8
zm}cxOf#&I^yz1!a_*{ri8j-J4a1{~5Bp@^Bbi?F3Wn@J;d{6oszzh}r7-k(q&4=#&
zP2AeE9xO{AfD>lOH{;^?Am+T{yiVcoM7FG)9Mpluuie0m`4jLTr_(W6WX>Jj;6eti
zs)_tNcXoY-Gg1k-PPQj%Br=!h>wKo1wz49S4h)|pqu+fwNBpKJlIV5Or9$yn^t6~>
z9zaQwfUK`t>ZJQ!y-Ewgz$WNzpw;}4eMTK^5dx{fS9wLRZ0Qh|LvQhUbhMGohCn97
zUzf4cP(|MK5BS>h!2jSzCg-Cj3<xzjIrdat+b>#Aqyts`0s}RY!&0R3*a_E5uh?Td
z($9_m1)rvE`7emg8MMFf9zIn!Gz_8wfhEEqe=QZ5YQ(vZZ1B3$al<e$z9w)-1OPgk
z2~j8pEJQ;FQ~g7{=79;ur+>Jp<q?{X2rm*KJKQ7JK<cv#GvP-Tj>Hn-@D}0{T4FRe
zb`0gEQ6jbp3SNagWuifGG&m$=FW{5}_=to0q}luPp9|G2GBOhKEG<|Jz*-bBx<zSv
z2Up>^z#j@&+2_2zjJh3ym!Nb{$B8WdJ!hM3*S_cKYF=`IEzVCcL18C`)L{TU)h%)M
zm=d79b>2!nBt5$6H_>L}ul7OSVTp<2;}#pj%_?-8IQ{-<WbZw&#oofw_A8B0L(plB
zsm^VCW~7LaT>+^>1H#$NNM_^Ba=u;2+6eozw&=jp0sJBB^y$Eg)jX$v%pw&<pf=BT
zuL-GUE5!|R4E1n9aa((PBmu$H-GH(Q|IR=aXmz;b<m|gdoOYtr$6fYMiki;R3m3BW
zc2=2F3Za&LPOperxg%b#ljzdvyk7{U);%CML0%=U@g43zp<zpXeg4<<=p*FOh?nm^
zT?7I0CQ@(Gpu^}G;q>|2iP6G|)f)Gt%b-KVuM>x~6v-Du<B*H+HZ(0yuSml<6j2<|
znw1ds6ETrDtYwmOG}pf@6Pq8z(@+jE)!2!YmNo0v`InS<N3KIYfy?|bq_nrqB6)Z7
zoRQgNeCIok6>(pqrz~;$%?9+eNOIuv?0OEsOi*jo!9PymSvNe>K=F*lTqxu8TY>dn
zGp4A|>rl%%T)41VM1+CD&dv@J;jOf^ZMfYmzzrjV5X3X?t|2sN=24vBC#}HvUS8R(
zl7Z9^_+!0X2YzJ3Lm2Lp(sxQw=>7pp9pC(8tz?E*Lx8~&a4m>YN@KJ)XxDl#YvxLS
zPRmU@plg=a`9x;gs|Y9N=oLw)7cV06s|Dgb`xcHwk9^O|;Ei8iisBNl-!5X9nVH$-
zTs__=lr#V2EdXtYOM(P1%(<y+KDD7L048a^DmbAF+TMyu&=JW*8I@(t77C<nL@;#(
zI?u#uJ_G|ym#KSD5L$vKHqwz3gh5RB=YlSc+b7i@z2hu?iOO#L=TDCwCYFJolq0Xz
z#C~tjI%6y@vYBU4S!`;yU`P038<pJ%U99ut)n~mn*zP96R|+(LUy!b#oP&v@31gjD
zfK56-exxF{`Qg?yE0i)WF)YzxVU=+Fl7|RrJ*nu%`g)nKI|6UsG{>unR&=@5+^kL9
zF?J$Hah@?Vh$p<1Ur<e;dvqr_*rDX)$AIcw9&T=f&3qebu9=-rO4=qz?<BQ5P1sd4
zoVtB`aEwi?zFkaN3a6_KZ5kWRewl5LUlzQ?!?302bV~2bUkPl#M(#A7mL{+p)C`%(
z<JiOuU!#NHj+aV7ibC3$<SHYWAp?VsQ{93xI<xj87+tPfy;|%CCj~7D)0weeDEIt;
zVnS+s0=l5vvD#KvL#LGc8ya<!wXa;sC$eRBB$UVkgZ%uc>Dfebhv`qc4TMT=-PQT1
z^%j@%`$3`I;>yf6nRMTWEY~UgSZ|woGp(_{xPQ_`No;CTFW8IKsi~)C_rR9Aw|iIl
zc8SuJo7P*Zxvs;>Nk*e}jZ208J+-9!peCF{m5KvS0^0j$M)W>lp+GH?fI*7VWb?6k
zj8I=)*@fH(z4h(hzm;%sw@o>th|oYq6`|yo(c4J`eM9P@cizS|BT{kf^pHm#P!}cF
z_~`}!5zBYkCO)UqkV#Lo>}qTo+|#%HfIQVILHe_-jeAuE4CnW5r*lq?X;o`0?8&w~
zmVRE^o{&|akYpYY@jiT+Q1P4sP=QQn5ipz~=A!%?h{7@H&IWv}kblr18FHWgu7O@T
zI2t|R5)s8PXtgN6UIGety7d_^uI(uP)B5+`=4|c!CTvuCUqxD0%{8Tct^R5u`M40O
z*2C-Q{N2jLXVuCaFQ}JI3dy)iZ}|{7@^;rO1ES-72ArxjFqq*wJE*BiNO$D1f;5tu
zlhZmEywt1c3B5&1c!i0t0hRNSPeqltcyEM+u;M}<OwH*kN4!OdG`{#vhO&ZzqZARp
zVP?F)x~Aqn1jL6TCH)boqrP<9d)U;yeN>MFzvUm8Qg+T4TfgsBU1PXtu<G*oZ0eON
z{(jx@R@;%%%;Yr;MFwkMaOBTc&+LKtGWg#x-QGc~{5PiCn=d~_x;-yF>iDG-G0r+&
z|6HN@TRhJ!T*ew(nzbVM7#4La*(>>UZk1Q8|7KWPuxUo=L8m+Ima;dUMm_kVRA|ia
z%95_-efH3E&-WdbtaUlrvoI+*W7$YEsI*Qc%KdwoSaqScki)p{tt<RAL+3$;+nSan
zuDmw(GWR2XMH;ciWF8v(<?AQQe(?{CRTrchAJ?Zx{SdGR6hma(qB{FS-@ojj;X6(%
z)_KV4^_`D}ZCU(j7wmmwG}fIoIGnn|#e3lN`<$LaS=zJFmuAdrSX>zTg7(@(OU^<o
z9WC<0=vI-fv9M`vp-Dum{^{fMFQ9UQP7KQ4X_IFXJ~{84BYqD%zF)69rFZsdZD`(J
zYU{y`{j1eQCf$v<C>b-aPK(nDd$Dy-(joEjv&qZ%j21pKw-!aka`BD?z-}+hD5G2i
zP+JOi#B0bEUN^}84*HBDp(ep&*zEhk4JCYl&t~n<3yt-QQLX1Lx+k9POf7A*YSZ+-
zsKj%eReV=-6KQ(IWr}Bp8B_iJ^MjMJP=Vhh+%zWXRS(h;$|S&BMJKZzIvhNBeaPfr
zR#i#i@;bIHQ{mt8<3x<Y4ti4CynWvtb5NOY)7R1EJKM+P3=UUHSsh?LZ!52tUwD1L
znZcWdj<3nS6Ib4s7qV64n$?VadTO5gV*hrAANus3>$NsVeLB4JcI0b()k0gKUBVck
z55%?}^r0(Xz49l>8Z?L|%IUWqL3U<Y7h(Ecf3z0$MnF&y8I0xT%9Fwbhgsunfs>%R
zC^dZWv*V=}U5k^;t=sjZN2Ci|(`YI_{Hp#d`p2CpMud&Pts>IJ;`J|!Z%mzi)u@sk
z5-)c7apwb?H*<IO*wZsJrHFJLcO>e8bhE0}fJe#0e2pTuANSqB;+Y!4{KVB!)!3M(
z(rVz&TUvYVOX2vtlOs1YmT!1qtzo(IE#0=`v+NCaTC6t>=q9SzzB*kuPI>O<9<;09
zxaMf$?H6m%kKiTPz86$)K1m<@pXhAz|G9$d_Vj#@`xa`VO{=}%mo`>h)9F_b&$D+P
zk+#W>%hybsefw0|sii4Nhqt9L-c-)jd>O-Urovo$K$I4{X#9$gG{~)_8oF0=Ib=f7
z<Xu7fV^NWnzTMrvQU-TF9wq6cu!$2xk6F+TB6)#fzyvg?!nv<0M2L(m>kT3)G-ZD<
za;ib=>oIM(p~T0@c@U-rw1fEp)W%ebm5Zwe_W*+JvJHP%eSzvppy`r0e_>)q`j%Ny
zH}q3ab(%4LZhEmfZGWZ1yp3ePw|w1T|9IhIT9~#=M&`L|IzfRkO*j_J;@aCEwN}(;
zs<kO*a*BQTHDEXCX0I{rr~HnY(5Jm9WXpz$b~+c7T7PZbn%q{o#ku(Ec5vo24;}hT
zN{SWr;2*8_DL{#cgdMaRCYR}qlTS}zdW8I%<h1haJUoXWmOb&}Gz$eyZ)Ig=g2jT!
zMEz$mrlmhe+Cz}!kAqSFapD6iM3CH;9H&n$g>y#?d2eLL@Sm#M{z72~a4L1~lEGpc
zHv0L}j*p&d2L$St_e6>Y6X~oB7yIQh_aL9Zq9sAB)|013rR%m+!rDblpl+-1%+fdN
zOV<g5EPz;!F%FY_LV6xZ_=d#NM8NeX1r-iyfz=H443-W&b@#my6L@dTSYws-;45zB
z->+u5tuK3K1z9OJ4}2`H7hcqLT6fJ;#)4|t(lhjZ+2s&54y6K}NDGgNn@;nbRM`!0
z;#1bF;GSNdqp&0!RoQf*F?BPHBk~ZV6HmE99oa|lH+c9TX8<wrqGHFGoItaEeSN|E
zd0ag#2+j}sV~kZOsrWuC-VMJxoIlYa9K$&qpiJpraJyE#m=og1&MImUu_acGrc1Qv
zZL7&Ach_G^N|TIA^CH?4x>HI!C%K<`GY)25e#L8As|{sMMpyzuxn$=5Kae+Bc3vgd
zV3equu}al3MK94{Ow~2`{wmZTI|SCoXddqy{zeS0X&gq5w=Qr!WIBB?Xt1j=Yc8!Y
z_0!bYs2HPL>Q%c<RxLc8@vlnm?tSDplSpIqYH_5o<JZjckvU}x$FEoT)m$7QvF$Va
z@)dI9Ek)~{YR=v{o!q7n>iWtbd6R@)n>b1J<9GXE;AD!q;=YUqV3jxkG(xG<hxQh2
z586mWL}oPJ_ljg-B*H@!_qDa;SHyRSEJzUn=;Ti5<J>@~)B~^^BfCC~7%d>5R+{c~
zh5&Dspx}haCQC^Q5DfR=1|LIEhUVwY?W`fRH8DncS;am2ftimlC@5%tzj$|I;Ewj0
z4`y{UqmG1Y)yF&3=Ji#<Igajun`mv2!P@C~U7$AxTt`F?9KP(_L6?8yoqhSWlXt1_
z4*v*`O#Zl{nXC3a)<(o%E(!7!ooF9579GE(fl2JfKfGsDk~=7f#y6jl*NgUx2`Sci
z6`CEqfH@|Vad7#lVE)!75K<ePuSUvJFanWh`iMIh>FMHSqcs%W>q&-`w|KUbx>;>1
zhA$WnnBVhmUcYCp$cQ*>&vmo^xBGF3;q6lUgEWh8v{(cJiV&$N9S|#8=;re6sHU;{
z;K>^o#t%6d&8=v@ud&0WUGv#}haWyZf}wjK`KL&@rTRtne&x@d3U^XzT(RHfblZ-!
z^XZCU8|AwrHrYVwL7Lh~)^PuRX-Q!}ymR(}snamZA3dmQ&>fS#cbCoqqY`8HFPjpb
zNsk#!JC}tzv`IO?+oJQ({MdA@!o=15O@BqSxyJLL$VdG!gmPPGni6sEUJaV3j6s-1
zVUAP=jXnI27-o}w6G$SH{uq0^-`I;?|H^vgM)vRZhxDB<rwX#300s-$R{WJr-W~TY
zvd#s|1xf#WS*EKWD6~9l#n%@{Zv`K$x!-J+cvsw2j85`w+4Gi~OKR3Eb#Z!z!UwF4
z&?a<FKd+_VCo<W-39B=RuZoNTAf_f*1WaQx!B?h0n@U*eGtl-yKFkN2K9g}5?`VQn
z^38^U>5MST*B|^OIVs_y2C8_)XzcVSTwO1uH4njP%>y5xmYLVAIpxapzmQw51%s)v
ziRDcDl*_W=Aj6Gc&6VWdWn5Eoh<r^Q?wIlJO_sIB9)r0zS;_2YZ!!I5$`KE-F=m!N
zwEab1jceKbbHDNZy){$=ZR5`~#Lp+LuUY?^u|O~1iC@LykydCf_xmO^;$sxfld9)X
zj6x__;N%L~#p#ue@h9OiA3MC`6>_-KuXka3U;&u|n7QR+#^_t*ymoOt0xJ`-pCPM%
zo~cy>v?7lF4Ti$PcSzimYIE-5$A65v_^v2VZo5A3#YVx5SrwkBpeokz=<3n^{WA-v
zd9u0I@p3iCA6Msd&l;;V$c{*nmzF(}mgw{|Vny}wPzgUP*@`^g;0wIJ^^<ut-1mO2
zI+B=>MAjhO_<$M{y4s&#Tj(*>#RQ@j9VZC%@=QZn@dt=6s%Y1sMmIwT9?}SL)AHq{
z8IGgqh7makq9C_z1iZeqy!**di0BD86Wt<7bXly=r1StMxBkpXHDW}hMEs`x;0O+~
zI)e7w*nV02LTp}HgOprE#&?e0UM*kR%cE3{TgIGS?%%q)@!9I8Jtz2c=k^S<)ZC6@
z3}4jVe2d3$^&>eUR&|A9hL~>c_gjd6<aJ9+C}iV=X#s@~ih47&tw=wftY0~g-X=*`
z(A*-(Ij@;sT2eS6EpbB+qPni`Q}UzOQ+NZ@Mf*}WMRu-wiQE&GEiT!oA4FHYN{uy&
zj4p_+31c0{E3WEQKlsYHfAF+Augm^T%=wB!(>tOchfKG%8?(h_+q538sC2nYS=pmS
z@9q_UNb-5<h!?-gtK)hu%kuc3$Ut6wm9QJ{pTFC#xdB)XOSDOKc2aaNFxu)Snugne
z{JeL~9jSH?f!G~G_CF4vQV6%V2#yG=2La)%)h8rG8P(I)U%!=?TdjY1E$0Sj3i~Fl
z%vDMSD>N5Ya|`Z<cvy;UwZ|$?%iXEE(_Ty{M$qHT#NrzS7({L2e1aDM17SAxa9tY^
zr;$uVU~`A}{5UQ8B+wET{-_ZiFpd`Aw(U4=?6)H%Wsr(iudNw~GvEZxEe826fx6Y^
zH04D{lwGr>xiqHjL8t!P<<|2c+4y@!j12?<b|yXNsDU}-K}Xt9x_gty!fv(~F<)pn
znw|LTKK%>MRUaR2scl62L;PTp);R8~iMRXtZbwEQ1WJ)C4}iDvYLdFJ3`jCA1}*wY
zT8<b6I|<j^)1DIR`L)h@`?X6f1>fCb0`>hvVm5DEE407Y=S1q*cEx^Y|Fn8#tCZUo
z8Z|G!d7DhEl`mFx8#%(8UqpLGb!&(^5cFaX=-WuP?*Ns+tgevIi_Kjv9KTL`hAptv
zR-8`>iAmWgDjeP{B5JLdXZ~qR=1b@E?tl0zq9Ki0nMc`|A1F{<W^VlWT($C0eDpRC
zjjTOGDd&bC81QoZka@vPT~flhs;YNduD6HovlQ#RqFz1<X6Av{kjNi7K;N?fj_m`B
zp*vxrwrd%VWqx-OGuR#;>n)}j{z61F8NYKVY@shg7R?)z6U-ZK*akNDn6Fg`t(tAs
zX^YW)k>@aG+cw2-zIp2q@Tien`gA+)V*}39kQWsFI9eMj1}3H<OoK`x!AakcxfM_c
z(f+1=7Wa>)IGj%x&D><3;e7sq)zH;_-*|URD(?~+Iwlobwbeg%o5xPkg81{2-oEO3
z4(U4jg7kY5Dw)^cIA`Bs{r&Tx?9L(cM-nGpaw=DR2=WW_Gwq<Q$&je>x^j#=euDdr
zqf%SV$+0~F_a87;FD>=hLRr7lBO*GQr^1u}seeo$mm;?!2x2?NE`Ir;FF4_6re^8&
zxK6?*^iXx3DBhkeoxTkHp3~1XXT!y$_V2gUlnq`em08wi-7G(F{yczA4mI@G)Jp3L
zL_Laghv<ub$A&mX+wNMNg?XV#)MNJWCg%0O+4d-3dgy81&cVO@NP3nYG(a$xusu`t
z`C&!1FDC}8oqR@j=li_+2zo<Cmck~UuD4uzGmh>TcAJQNH*~y|q4Zd%M|%siLi+&C
zOGV%NA*uJS+S@0*o28MrI7!=MnOZgTXhB=rSc3rW@)JeDWO*!X)0)HFcM15+zk%%X
z&7O!JB8-JtO)!tK)Zq~>?0wI+iq+zRCV1Y(-=m8TjuOVLq9~q7xWsOi(2Cadofrk=
z4Y;^GvJmd)4?H_jx$mKsFqNdJV))Oeif1Rvt=toaCj<xXhDmW2+D*`#$&{ErYPyry
zIu@EONAFB~`#P&lXvUG;f)L)8m>>4v99%M9easu%J|MeT?3eoRrne0C_h1l>xhZfp
zn9?sJkUd1mS1ZQh4E(#b4b~D&hF;f;M(M@8*IX9=m;qG`=XM$#nhjjEuYH3fV^d}5
z?uuEnyUBiW{xHBp>B%ovi2I~eum3Ue^VWqE`#>&V|7wMb!KPgbLm$}`hQ69?6lFj_
z$Lxk<=VlS0UfBXGe=eQ~Kg}6|stIE^3{N1XNsRnUKQ#;@siFf2?<eE}^(`X4d3b-I
zI$ZumBi&IzhP6?u_3SDsTizbED$bWWwwhY*vN=u8xuLNdTJEcJn(kKBxD6O;sovth
zX?(DzM@>sLxp?pS{RTLHeZ6*85f1$(DMu?0e6V(q?)=bWZZH!a*jsFtFmd6YP~qBO
zW($?Xx1BN`D~Gk=2}~{ibus|9J1CaK-F|#>+Sb5zfIgnt<#oNUGJD0M<z7v$c;EMp
zyQi&tM%-G%zn*gkMRauKjrLBpW4@<OxhK4d2s6w5?p2UJx({$h^Naf(>Pk*-*`n~_
z0|!mroy&*z8wB1A!rSy;bDnAuYh556dY5V-NAgi6CA#dZDEbo71H?G;Ud|GE?eePX
z<EFLJ@^gtl+PWC`bopN}tM!n(UB{E!^yI{283`HT@S8x+wd+g%C`fgFVLlk6-ndmh
z&fCa6JL7=ct@ldK-8^<J^6l*$<>tNMLOznXe0ZSvTCm=hn}a3ci7_h_x!c$6UwKqA
ziHq8Gyl)@)Nk>vrrcO%YXy#tManYp0fZ0Xc%`h`BBkF*kSXk<?nA3~=J!MB$`P1+p
zRab3@DSIn#&Q_>-{4evc<F*4acFj-rm_FZNuFz(KTY>xRw^3Ka&Q!ltu8Z97a;CDs
zJ}Gj(#k|fidD({2{~374uLjKCUf~U2sb3$^ZpfNp|2fC}q&3U#j3vZ()padwXWT^A
zGEY#O&&8BUDIXf-h^zIEzMU%CYT^6+RhOy}^ZGL_@@;F2L^Z|=TQ43QTYKLQ>=)HP
z;%;=y33;9Za~^sH*C62wty+Jv&Lh+y7;|(r{v9mY-VC(G%Zpaq@$|lJO?W44B7Eaj
znE!}j%o?s6FSLaCrG$?I*M@oh{hY|{H4`pfmXY#tk%_N--QLCjRrXF^nvL<n8l`<Y
z@HTQVGw(twJBtYpW9CIn7q=DtT!Z9z5QEV_EEqaRzP?*!S6q7f_-f2@@|ovYq(`?E
zT+`lw=fg0jxXmPjPbBmDs(TOZM^{Gbu2=eI*DU`j%Tl3p|Jb<~^`oa+Ir2YDj4F1G
zx0UINW;X51+p0r3ckUd)GC^0$6{8Go9H{>0=65}2uUtPwJh}(duJ+sFx`cBl+V;&l
z_r~REmTaOLCvpmC4j&T@Poj(6tiMB)S4l_6A^qs)S^2id8Y}kyDy+KY@0$#G*;Pg3
z<A$%LJ9CbjnLS&m^nVq4nzz&g-zMRZ?_1V`gntzSgGjfUKl8dcH5#TJR2uFT4SDGo
zsiM27;*^(_ZQIp)t*6KGi;(|1tz_*s8~k5$Zs)HblA(z*yz-+TEMgNkW*#|y(Xy-m
zDUZb=tKYLX=WQpJS*a$xNp4NN$s+SkMZh3XZB@{P%H0`l!ERRSmcO$u_%Hixq`y5$
z{o`DO9o0>wsX*LYf*-cDz>c_Yh)D>w;XOI~v$@$Qln0Zi2+j9zoIKsW)`NK7Ezs_@
zb7#_Ac&K#Y*{ULCe;+|ngK%cH2d%@^Z-rQ|N@mjgnpt)&PpI3I8WQQUyxP|>#;tX*
zNOtkOp8p=pUWxdyQ!B5!rAE)1NoZU-RFOq9t^T3LS~xXIRjq}_BW>U$GJN-<^@J2{
z8@p)W#tewyAR0XI3?8JcEx*%X=O4)0JIFcUShBt5!zJK%u?LMlzi@Qrt;lx$MBPJ)
zJtq#STke)C7<cyfdy-_CYTGQdSN=ugsBVr-i+*xL!PNKE=&4uBHLD8WYuKLrS)^@L
z3WB+%@YBMW?ao&Kc8bE3jp+5&!#Aq5yw9a=?q&oJF^*-dgF9YpZ^zrPXNLmR4#$8b
zm>A=heYSQ~YulRDH>?MwIVEqbN|}}Fk{UfdQ1__i%fb4ZO)|A2Cs`G_vmE6`n)Iy&
z%m)LjM&8#&<;z6Z9kW{>LY0n~X1SHvQBk21t_AY*N-|T#c89mbVL?_#p~H(A9vNz)
z^4&IWMJ9;k&r#xrJ5IvRd&=xl9i48nNu$7t#r!qPI&Qvz%OpS&oM2(I?YuRMnQ+v-
zeGuQLzkQTTQX{Zc`fZJ#>;43E=GyZV&UO#Yp8%?Nc$E(YofI?Vn5y)<OVHkwwrfkT
zqdV{iCIa1+l<QcJz&~9YcQ&aLO*{TJrixCe%`q{etfZq8kdw=KZHu9H9X?0MSXW`(
zOhj}tfaUq$pH#50?>%{vCRDZkII-m+@#WNSt=k<XcWmLTQO`72NF<-6ZC<2zjCP0_
zMGiD46uJ1Rq@tui-gbku!%1L5nHc6NA*6GgtrQ9%d}39Q1X136Ifs)Xfx--eUf0Iy
zXyo+Wp3}V8JwT?HPEM}<fpQTF=WU+1T3J}!4%K;M35@}SBXPt*R`_|cCZe=RowAO>
zkYRV@2?9TQEUhN9>J#8Eo3RMwu<}Cl=Xta}$&dhz6QPM^qYPw!7Nir;(SJ)k!okL-
z3ISNj{15fyE^i1%d7zGe7P5;VP|%*a9vu$X_V%V>(~6GtQTkmpeqp<M-i1BGRA0!y
zZ&`eGaxrKa=9pgYYk<?^9>&LaPlozq?EUwx6-ZlZ2BiMpO{^VozRc|ZSQ`ef7rfM;
zO}~}&z7G2HV@2^fV%!KlHA5h`oU8&LE2|mD#p7?`qx})v+SbuwW^aG;_bUuTlC|GV
zOIaN1DN8f`Mql%iSx*onlA36kNzZo56L)1poO0vFAf;tn%qUh3(P(bx>Rq4s`4f2p
zI$-wakUy>G<KxS0-f&rbV;0w3JxvkYYENq0%5@`KAUC;RhotY=!$V~#EHQf*hYy#6
zk!R#iPX%Z|uIpXT#1)UO0Ax0tm6WgN6ySB`fQM1f20DyLea4{0xTZC(Gm0~A(M@aG
zoPvp?FhH%soC2>bu|JUiKy6Fu>C>7jkbL;?ApwC_^n-8#gaQad@vuT*Ov#FG2o*co
z%m8VI;6I=vZlY-NC+O~7^so6895HG-$zgFVf{j{)Sc<giT#RU$H}Lac!Xc!3^Hfg?
zLKnS<G(7SSKF}yZwe@}AKWTRE-#MCMTDs?rb?<AT6FcaxJoOn2S~9LZ9y=sP9dhSR
zIV9~bAp?E;<?c6~|H59&WV6|tK|ll#-w9><2K`=X!|nRj;Pm@e!nIQIKl;SN1cT37
zG|OfY*OAW(;O<7}!nvoEiYVhU*NI+!+}F&-F_S0Bq7qIItKBhfQQ{-iKE2t10>KWU
z>p(jH`0Xr=DMV2Ufn%Lkd=iQHx_hGNZ7a*iv*MroC3No%F<v@VPmPoHDYoG7+R3GX
zj|&sr!0{yJJ!|)yQTv4xvbQX^z}1S#)IUe?uMny+m=-nyx_G)O5Yde2GO>dR$Ui8p
z0xDO;&aAmNeSA;P6RRFOc(B(%G+XtU9-Bk>R1qIymEe-Yl!4greV}qBUKl78Y5~Tw
z{R_Iw6A7Jm7P@2I!)29JS5kei@K1Zi;LYs@03p%B<OiXXtxT~rqdn}!sO(>s+-*AW
zZmcfoP5h0e6GT_`=`ZGlV76WSPfwg>Q`t#tC3`7wBg5jVuP)hy$z`g8000Qh)7Zcr
z3=3>r)4rVyGre1L(`%QwVlBY7wb;mT1y+7Sb*o!eX{^VSf0XI<!K<TJOJD8y^J(3y
zPVK?y-}a|G<WHFIXJQ>J26;!n`xCcpyTX>?WzRgg8tku*#$C4ZzLL+obnY;K38I?E
zhd??@1gV*J@9FhFPS@(!^p_M%8)#w;=)Jxzc|uQUN8Y3wd6PhW^k~NSFu3L070TZU
zC;T=Gj#<{tc*T45QZ!3f0A?6MS+a2tF*E!)U~1zgBhh!yQ!Oa&)s^&6T=|S=2}NeA
z17iO<7Z*AGwfd;w?%eUkWF4V2M(;WvyI-e<7#dXo>MKwBsJ4=4oPK7f-I7C8D1wJN
z7Y?dK%46M>WWCJIKtt)+Y%}4c2CO11^vR)>c`s;ro(%byZdh_1Tk05wEpg0{eXTRp
zrSq6xZrGJ_xitM`f-ca*;G9_=dvb6CbcCPKHWXL76%asXXI8OpSD^3;CF?McvW*y-
z%-B4Bw^}l!C(3(wr^M2~$7Uf(DaNKZEL^1~G$okCzTo!6U5_9T1#H-5SCiGxYVv)u
zI(DfFct}zD@IWHqr*i;c0aFnyV3e_csDz1w8tb88a`J1ZL#s`)w`WD;v!k1RA9P<{
zN4=q!P<7Ej6z8Ns`+wl@VOkTcav*&~3g%XOYazhZ-g2uY^i_f_&7L{N6L0y}{Qh&c
zE}-Fw8O8qvZgzRgv0LAP=^wMEv2Fz%W#Prr@5)3!m|ax=G6+({K!w*SC-S=xN@&P3
z&SWpE-2;Oh7GMysD%$FJo<)rAFqhwnR=T^%w2*p(D3z|#YPp7`Ga()&^qk=kI8-x0
zC4A}3lR}Ri+6(BWofz#lW-#1j4n6Uqz7TZxBlWbZ3#(<g^vUP6<(NeVyTd+bK$zet
zA^%?dO;Zu;8P+}8P7N{}MoM~GmTC2y%=SA>L@zzkPFCX-j{ojW6KI9=vuAXasBmyp
zAbLuBIqD5V4pbznXb)R|%$I>9$dOZb+vUe&(K}3EdM8*)0Yvd)<>uHN@_ZhBCbFJ_
z)skdamw>?iwtuoVFcas2dJ&7g(DTD`S8+p#@?P#`+M)d^7cN!!x5z$hC~F`WSq}kW
zIYzs6lamm~@6Nhcvzk9vgMXGzWRxif>)^@TmD813apZq95*?k5B_bs7H|##VSmlGf
zj=E+~>nrF^1hAHP45<Pq=BMO=&<S1>C&&LuI=99D_HbEI?9#_X>vYnL0&iR`Y7{_J
zD5cSE64<tlE@@NZex{1;z!o7hXu~d#Yh7d`EmXi{(;|kr(4%9iD)?hqT(t^|G_O=N
zHjZ8@xypRs7t#>@Wwz439pQJyBhIxi-5imKT=rds2l3O$-c<}i1`rY`WQYV_3gGqs
zZC4hu4z-`_q0J?(CZ(x1Gpy&lj{f%q1WZwdUfA#j1$~EuVX=t!Ed9}>+?g%>ZxX5g
zGV&c*=$TSp3WPP-2Qvd5I57o5#R1oXWHg_`xh;6SrPUoRk`Bl1BxV3O_#|AucNQ4i
zVi$5sO3GWLVW$i|`R#V(wqAX3qh_S1<M>k0yGwA17^V<?`0HnU$ZrP8B1~frqOtm~
z`=%n2u)==bfUcrFL%ee5bsbAu!IN1_=VvoI^?*bthLw2W<Rc*pw>>)hHx#Y&9}+E)
zrKdrnW!8x>;7!_;?rwBI>CxrVyLT9sfef+5i0}fo(!(xMIfl5M@a_v)E{iA7LYJUr
z;Z=o5o>G&22LB(KEBZjojA4Jx^>xBTSpg5305rlkxn!QyA>COS^5oZ6{x|hZXJd8p
z{f_mGE?^x}`l<g+K?|$Da1aUS7TF_L{y158wTHVxGmmCI3r;^amLg{t3u!E2Z>Q5-
zNrijI>m$>h#iQ@(dGgjjaa+QD=01YH)~J;V<^);j{QEFO6M)PaBCiS9FpoillDY1T
zHPPy>xc0mTo1TPg8u(3zL-#!Z4(WQR0)-AL>j1@0rSMip51sbC@*~GB6u0B`YbEKg
z%*TB4P~6TK5L^hl5%fA4uEFz*bGyn{%?zFlzVL48XeJ|5Bz8oCszFAWgkwJ6y|{4l
z8v99XTv4lMURK2aJ7UeUcdSc!%-uF4cHb`S9wO>-oMQs5#UqC4lZCV*s*ERpHCEPe
z?6TSxpe)1>OjN>M8-^a3xT?usFRZk~{;L7K@Rnc+yzT<E8H08y`ELr??87o#`u5|}
zEx$XtlEQ|a>pO{>48qx$P}^`ynjgqx6GFhaa$L$nARt5#i#hFS_p07mdX5Vap>dwc
z8XWR?VJg;#CIC)q2)3Z14QkQAh0iPBP3CI)qVF-rT=$j0kVfV#y`cyYXilW+zJu~O
z)@_gz>zu8@c&Imu(h0`?7nN(Tc>QthY-IWR!q2MJ$8Wv}ALLW46AethaX3a{ACjYb
z-Dq4ot~;s3``BkJDf5|p#olXT7dS_ZmQo82iFO*1-ov1r>`m3h?c(6zS9`Jd7k_Z5
z%@-K5%L9&5evXax;OVC178LaN_L}~{sWkPq4{lm>dUQqjR4LQp-52gHT}9N1Vm+X!
zU>n@2+%&CZ5iM+r)^;ssstCv$tKXHIOYg-&*QnwflevTdX{VntHeVBfR`I+urI>fQ
zUUg45XY4G?2$bHTpy+LFQSiZ&3$w1k5TX;*`CEb8Wxg+&{7yG=<*DP0SQn#Hu|cqc
zd@oPr!4lQ3Fo6yc-tht5VEkqUuW$5wdS-dkwf4j_muFlTq0R&xhJwBVS>X+f#+4y7
zn@Fs*eP8DR_Lz#rksNA<lZ@RvsD^SXp7f6WWHK2xA=!-1k=qhJ$fW;C39EWs(E5Kc
z!d_#ixk_Yyy1E{pxF9dr<mH6!gDcnWxtoITqxvAK27V$?cA0)sl$?hyno8u2kI7G&
zBBH2_MNP)oYXq8bg{?-4X}&ZAn@2*J_ZXju5o3c4fA4Y+mSu`VzM8g6F9(!LLIe}X
zVrPLNE9f7v6s;y}3!>xCc|QN}-<K_e{KRPQTb8zZZWu2lAGbCYW?O$&QHeSg<2VB2
zfbfQ2xk>(_kQs&6YxM=On)i-z%5=2hej;y$E^pb7f5XPc=6P*V+lLQuRTw_o7n+6d
zmvAEpkAk4Oovv|XXr&lR=No&8-DP*tettnlFHfEJ?SF0RY6AE|XzTxH2$*SI(fZ2^
zbVWOt2l7U@FTbH!dhH}Rs`u9jWvn(S{`WjEf<GWMSI|rEL`JUK1r{)Tlx)uc5Wb$E
zTd|Qf-qhv9FQu2f{(Sp)^@w8(#nUNbl)!xdj(GdBHrT%q)T(enyM0LtC!GO;GbB5!
zFj5!ET!sb1U%E?_A<ufOErh>${1e%1fc4zi_`Cs<E<~EpnFZ9=)x_Km-4QJJM31!8
zHu=hqG;!$TvdXi54fZ~2jNYl2Uj>!UF6x2Np|Z&|gN8XCDpbgr(+uhVw;h(f>c$`^
zA#>ShTmWe_#Q#JdvDn>Jv|mSOXpzs`={EH$f$Qs+A7mYs8(X8wuCcf9uUkf{g50V`
zk;f9O4{_eyY!*0pTGmsvE<@{=S5cr>%iO)lgOU%MI0PcCyO>4kZysHHWVh+!Zxf}G
zl}A>WZKrJ*e*Rvsc47O)xQP#=uM~<p!W$cEYOpP#)`?Xe3EJ?v5)~D%buLV~_H5@k
zv@S@Gd9W2Kuler!#IEpmcoFOM13PpFRNheB+}sG-40736&nZ19H&ro~vdoN(kB|Sr
z)ZbT*P*clmIs*9v@D+;oh2twINl8hBmWJthUw^;-TqJVpjR2%Pc<>i71NxY9;36q%
zn!IuS`W7Lf43U$CuD8(&(pWAkFm+6u(qUO%Y++F=s!%~~zuP3j$n4sI1(Oz)HxMkx
zBf*&eX76nVVf&|n0V?7kgs97I4ZH>9TV;X5nA7?E`AtAY1B7SesS3aea=KpB=XdYk
z4Zd?{4^}liF<c}P&DPe|A9Sl%|8DukFYqHJqL6=1hNED0Vg1;<oUsPau|}FBTH_;E
z176o_w^N|!BoZ2cU9UVETJ+W&00HQXERzYCcAGYBil`rS>c>Zd+{4DRlGmlP4F5Bc
z!(e2NEblUID&0IU4AR)e4ATp-@}b~l7L$NL)<d(kN|AI%s(Zf3w2dD>A-ndsXL(n(
z%7fVCA6%ulM(^Zwxu=x79yxL(D{9+*B7y=d<XGx?jm@l^Ly~W?UJGbG(myCZ=P5tN
z)b@kNV{wy69`}7?-{PRqdf_bsPOLiXsiY$hg=)@tppHTeIQl5z*~Ij8RegP6bo7sq
zguzdrh#4Y;K|o-2)hn7kRi(!+_GJHdm|WxWQ}D#BxZ(WUt}iMN)<4dm;(<bsBs<tH
zPE&LC4cU<l8?h2N#K9WnZmzCmhxeV35K%LH`Ka8mBG`24$JXY^y!4>#9lBH?$kV;r
zUbVE0aY}Qu01CJj4p2U=UAy)T;uB0^b?6tKMsEkrsFs~E`n)tWH0I{!!MATym7@Dz
z{rYte6cM}HvjWgaumCii2jf{l9`2Nv?}4Q8h%Kjz%2(0j$8-A+M*KW{dfkff#)`RB
zV<oJ$zeNSKu5oKpQ>O#3N9<a=cT++4Ddw{O>=`0L0|<FtWfeGkfc{dEU;+>s6<ULr
zk%+<C0txT&l<iNn+Ed@*Q%e3MH1E9AqnP>Tnz2=vkKW*a&}aWNKt8C(Ju*txr$Li3
zKA>5#`%&4IgAU)7U1+Z^vIs0LTc6V=6Bf``rNLbaPof0+XApDgp~PcUuIuQCLVK%y
z<Un-MW2&{c-IH^fcgE11EIyXqSxBuk!?t|#xxE<6^E~|YoR1f{k!zw<i0(LB@a+np
zuD13|5(MFI#cSUEadXZ_86N?yG~VvjQKu3%a2s}-l>W@HVknTwaFJwrep_FeW*$Cw
zNf=1I;DFMn|8lySa{WM&%NN6TX^HOnq3=DVU1L2RJ>6x+c`7eAKXqSgc)rtf`Xfox
zb<3{>9N^7M;+((0sq|CKb*`N0OZunvGp!v*^le7Z37n#pv!j3uO-wboJCO&7Ll=wG
z>#SRMZS4){XWzRWU|~t7*|A>i;$6dnOAg;8J?44l8~q%5Zx)4I3kcL{r-*^M0Fnm6
zMg>S4kqYO9(1a$o5LnyWB@+Vzn%TNH%q*<n$fk4_ZE)V-{^{+>v>fv05`@yIU}=>B
zEg747@;Y77%!PJfppJ8BMp@N}dW!&~?)T>zUru=)*k<s3YC~s9`pDhq0n!}0M${4E
z;Z<<N3g6tnZ=WL$Wxo1W0fFq^ZmRViE2cN!{m4IGzv2WBRaJJkhXG4+>)9ER9#ky4
z9E&}EOu&$+#9V>bliSgD`E+H)%fq#UbRw^>gff*4=4RwA-}hj*)8cG-qo6mORl-OB
z_z?+l8&L>+f*8)@e*YBgc(|73cK*6%+TXdXcdRN(!ghTq2#(Z}H99n0Jf$fpkoB}&
z<xR_n*RTBR>gp8e2M(OWIg^iJRqXLQk?Xp`(6Dy%t@UGtn}gVrAN>^R;W~o1tNG=J
z8wm<U&+PJawkVCsz8JCVHWWCO#ig-4CMrrDb!N;}Jf^gL1GG#<)J)rie@zFSEHhZ(
zKmQy*$eJQ_hBEQvM?hd8JXd^~?Kj))Dj_V#vltQ{Zq*+kAaF|VBcDi=QoiS3ch{B)
zt{Uquxl}kh5_3q!%b)I%Pa_DRSmK0!b1%d(C>q5mq^!q~iiL-Vu*$IuCsW<T;N4ae
z7M4=+4c~UO$+A8(UUrO?qnP4;i;A87;v?4E+>hVO_&`;&ar^d2%uvi^qiQ|LYHG;6
zI$gRw=Rpg}F>+GW!W;e}wYZatdvRsAw_bLq#a}mF8M)+gY(-d-^|PK)b)r*>X0@9C
zAQM(t`anH(TtPWLAVB`oAiub&FvGo7&%Qg%RAlV0=3RJ+82plM9aky(7q{_n>g4T)
z#2S{%XYY;(2=u{7mayn7F2SjP?uswZ2^p^6_CjmdwJF8~Xz%r3@dlsK)2B~`<VbZ`
zfA7oo+X4cb+r$|*CzbL(x2XLUR&l4G_h-onhnmWh#68sV=C7I2P}~h5g?}h6MAu$_
zb|FvIbiA?zRb%mUVmE&&pKju6h`VgD*YuZNuI_lrMu9#Kq0Z$$KolaZ|3jNu(HbM3
z_X=@&EM6qo!N-pSP#6P;`Yp8m;W^fTfCDl{c@#cnrjw~+2CF|Gm34cOm*VMsO+L@&
z<MO8b^R^&Pe!=lvwQ3bA+3U3HU45Vzz-rp#IHBLt>$|C$lmKAkgImYJf0`&%agBlu
zNex%xE#9VU0n!e!#bc#wI2Lxg%J_|>^?svDA3CqGUbN5T?97&2wCl-Sj`AVLesrw$
zchZ9g!EteosGv`jcokX9qj_--+p57#0JTaT1TgeSnChdtJw7bsrCQ69d}*F%#(|FK
zSL4UBGvBWqy-}7HNM1XC3Lni7m_?B4gD$$dd!DKjyUVemTAQOAf<UBo;D@h-Jqv_{
zN@xH_f{&=YJUA0!I-(!&d)w(|4$W`Aa-+Tr!nK=2*;oqlzhwH@lC8?%5^ok4XNG?f
zR3&ZG7^E#Q1m8eo0?tu0xUI5&d?YAuOrXOCpc02x@0f6$sDOZugX5*0{k_#j%pSCB
zhbm|fmwo8on-D9p?bKmeTM9irJpuSb%HK2%ailtyY@us`VJ&eNpdOG19?K#I<A{7H
zOM555QEIc1Q&D-TU;m-(vZ%vJfm~T%d%21}#hwmP+@O6gI?qh*Bbr7uH+6~^31Slc
z94SoDC}ARc<q&nh#K*B8=4b*~keT~lT3YI~e;%)PFI-N6JJnTGY)V@gzhCp4x@{43
zNAcpy_e#r(yqIuRC-)5~3TV05y{B{?aCDSFeX|jJ2P6l9uf$dz5JOAmCa{~GX!-GY
zm$kNrqs1G+ZtR-SbpBsSU9JPwmI5Cw<d{j7XH4s8)19eCJh_W!?BxroRp;3|KQtn>
z;Ce8;qporE*ZF`I;t)y$>qOVcCmU7o@#|Yo%kOCDAB*R{3e(ZklXfCWBhes$--LJz
zLEG9YDr$jlolX}OppM%J0F<u(mcnFUz<~u9jy3m3c=tRgt8HwoYHekA2wst%4H`gE
z%`}!dRik-eKgWsVj{b^Xn7X?9c9-uwU{+KCwx{c)a0d_38bm2Xm`B!UpX{E@xb@Y7
zx20C$N%(o~ZNt(J+&Qo8_VU^Z{d1Va)zg#(lorm;8wCX)b-JRIfe=x){*+{-c;umh
zO0uf&9m5o8yCLq2v}e`-{D{ZK?~fKeO?20Ll+Zjpce`ZR?gIN*AikJ?lSfsVqlGQN
z7*v0Y!nSrTf#=u$_>&UF!=>BlKAehoA%Rif+|)O9JPm}U1s+;hf$u)_Xdhf?drF19
z5@yJA!Ul$0+i`W>f{#JUvW^2;fFR`f`w&n8Xu<WFRUXK%Re+F!qq$D=qWS>d?Y~~?
z{>cnYsCyXWXhi<JpuL6&yZ9$Nam5k*rkB;oU@-Z1l(eD5?85yEi8g>sA)tIc2MATD
zl@Qb*&u0Sm48vX|>uwd3+yb^=C!Gm|8`DMfm|D*4qBT2H^keDfNy2RracDC-c(2jq
zwY&TN_tVr?+q5E8gdezZ1W?P7fdZfiZCg@3zo9J*87(I6h^D5vG~_-4w&RcanwexI
zNiD&%O`C|mRsO;KOk4UT7?0lF-$#E>+IkOJHt6x=+L>|%pL_G3Mmc~ScYgc`G+j!j
z$9WA#6Y3m@p^4@U)|(wWV&G3DTL_VQ2^%TnHvlJg8PO~F-aR$c8@QUHw>wu>R)4Gb
z;BEv?!z;2gV8?W)S>l6RWgpzvv9n_yZ-PA<2!A_ZMlt=;l!QHw?eck^Ck)ca1<A*E
z_|z!@Y{wZSt_JWi=4TtdvmKp&Z4nXq3*&4|OKnoUe|$Rly=2h~KA!`ssw?n*%>YN#
zeAkR0f^ncEeh2A}19L$bnCP2~beELe3Q<^LJAB^A*m%A7Oa)&q9QR+hL<Vv+#rt2q
z%8iR8Y~aqFHP}J&#uebRl&l5V)fg}vHbcgJe%;%I<_E?IUQ*u+Cp7zGT)HGs{aMWq
z%9T9dtyN`mrig3`B)x4_N5|)>0t#zwhu3j)+rpIoxbuXeVYFkbkxPrU=KkdzXdD;z
zDu5j%2SvZ`Svw5QD-i_0ibt9~(JT^rhpWd7?awT}*JXOg(j?3(ImqHOFpmS@ym_Ng
zuM>W5c{-J_m>5Cx0dS{Z31|vHbAQRmWa-B_VnNa~<03UUO1=s`taxl)WDKT@ezsjT
zoM4tWB337Eo}9zX1R&lGva)QM?T2u`M=H9IS}yzg89ABcR=l@rd~za>u1QHr!Ky)z
z;a|?d!MV-&1%{qf`1-ZJbt_sO*Vk9`@+v+y-__a-QS$4$y1yhO-gX`JlDX>2mfjo)
z&$*hKn$V$#XJn(ua;TI^PEs`?Zmjdt=DvTg8J-xULKK83b=o#0m?a-90$qC++H=>R
zZ}*em3qRiw1m1mE^}^UkJE>};mG<1M>P}QtEsjQ@DF@MQ6411@CY6D(qGCbQoH5F2
zob-OM$sU<cR-o;(2Ne%Qo7paFuYx%#f$W-yw;Wy47QcnuqNs(!I$VXqzAe{rBRG#!
zxB4A^OiUmGkb?<OXosR^(Gk2+wBg(km(f9G0{zkyNF*#{eA)GuM+<;8<k}W6>=3m!
zit?S{L=nr!#r+{GEFrolq8IP^em-!1nz|f$G!GtJg;L%E-!Oq;cRAYZTd+4xQiI7p
zH!tsYNXW6XN*K8M`1o+2{e|tI;-C?slF-C-WmxiIMS8X|$h#Yag>T?-uRURHy_JGN
zHY(K1!NI}#jA!=C%>J@zd9n`H3<5ond=JzWWQbeE#40N)5;dhdj%j?Ige{ZsfuISY
zf^dx%0TcA|Ye;)~zCKyJx|oMwt~T5OMmq08!6hsOBveE6l7`k=U9=HUR*pl0H8tAZ
zh}Q6>jP8pS;@B@nye&BEuN$_eCY}shQ3t;YY|p{b(RQ4|qM`uuM`E#$P)-!MB1AKX
zIcO4iT-gC%nV+T}hk|Do6BZP@>4mtWtn;WP<mKDK#?L(4?0wSZ<9T+RgdFJ_3i?ze
zEHdWL`rT_8hJSwDy~5%8%!Eg-y38QXJS(OE>-L4wdo=nZ&P=4Z=@O_Bw5A*vo~tdm
zh3HDA4=ar?M_-!|#u1#`74J`uc(7h?evY@vu{osrM6pL6**Ro}(U2<QN+5_J72*sV
z!JLP6HRq`^z7E{~6(oHd2lYVTcy@Mn>H|$wMz(RuCp#A^1A{a?+T1AREAjaR!otV_
zT&*mgK<oL*H1Didv9`NelWK<!NkgK06Y^qiphHNGb`!D;v!9>ek{?#n*eHij*dAf4
zn{pGB?G`ltN*p0rBCiGHmw;>6o<EI?jI{1_suittZrovoo<JLP6q3+G-h_DXgDtyV
z%?I~X!@MSZJ95Y_Y8JYN1_s*y{P}{GX9@iY#&}lOg~`@H1N)*Dvd#+Q=0G$$%71jB
zA#oGtNWHL*7uIH!b;fuivYE~bzHc55k4~kMW*8pM!UQrA-M}l==90)_tzdMo%DwRI
zUf{8gAeqU0_sJ*Y;U)90R<2qlfr>v6O|E`C%aV}e*#GLe0E6x=HSeHjt*0Rqs>E*-
zvV+FO5=j6DLcSDkJj;pBsZ(c)V1jqRO_`6oX~EFbo?^tXcZdZSu1fT^XaBl!$z30P
zlWP|)U>3pu^Fsa~|L5dO<NZ}qpgt9NdP8c2=Q#C!e#x(oqbG75OnC*@+@4c8yo`lq
zmWG`@oVz$KBI31FqmrK~2OwU}vI$;;1j_p+es{FOc{+Rh(WqzM(86K+aw9Snv{kcq
z?_qH~XG7Vi;>EA1*zsc&_lczKdl4AePH|35O+C}?*EtT!R{*xAz;K&9B}=-9{Z&DN
zva;E$L#`e0OgL65AW%ZuRBp3#bGU@H&Z22jP`<SX>1{Y^FN3WFjjwoAJwcKFX$ScP
z1u>yq9Kzugn>K<h)*M-~>&1(xAMq{nL+MsyhhR?{o1Z^_cIqcZ6A)sq?(U~#quiih
zsV(9I!T>?h%hOTi!ERDeYpLr#J#nd|MET<McZ-_4TM0dSek!MA07x8?6@Y#Vy;APE
zFKbB86mcK`62aSGb0BH};A4tot>aa7iJCSS@KiJ_vu>hQHjYO4#6R}%5GW)Xv$bbH
z8f`$<vQ<dvNmLB`ZTEes_78}ndaywMkE}<)wk(_{(9woKn0tg5UUyIC2I0Lx*+lUn
zOEd7RF^M%)aQS`+m<U(U?Xg#w7O^Asry}sf{?Puw|NP=+3~+C_!==!w!Q6c>`MZRT
zbB3LR1F64b-n$NgzCtvK0-EojWY_=wZE#Q(b6a9?v8c0?rFMW+D}KZhjnNAyo^Aa(
zHT4h&h!QRDc)w{BauP}ODLVs&FayU>T$U{Qf-?yiN9*l1t})@^?7RFOSsXRV@vu1v
zmQamoL2m9{?t-otp<!Xm(D&Sh5eTZ60b>Y`Qyl1~6}-K@fvQkJdSOuqGZJ>uVSlX#
zP?}_F3$zxx!g}vXGEer6KMwU=cY96)pI$%=St*Gj?-c}@CS2@JX?WMi^EF`0qYyoy
z94-(|6gMbUhfrCVTUd;@TgB;hDWk?gBthq+q+x2m%2y^d*aR-D;pVOdam5wpu^Chz
z+a3Ez{s<eyp?IWbO=r+*-Hk#K3$DT~?1*Pj{{^H#4EW$*!7Owb7Z####?uK{g;=qy
zaOl@x0+W8^(N%=X*c{0Sn$_51xDONl2-!1S_~>4>WPObr0vs{}M5kIO1m*U0Q)>8L
z2|bYPL`D!LvU>`_)8n&t#i)PSBTnUk*Bg~L0s@nD$oRiBC$l2oz>u}?;cyA0sCJ#X
zw~?c_Z-@zKrd3#mBi`uh<~EMIqT8YTa~x!1ESMp?^FAU$U_ZVj5@BxPFrYiZNj?a)
oHdM){8xRn3bg2~RGylyeIC)ic_azk;3jR5`PgCXD9uwdH2MT3b00000

diff --git a/_images/sphx_glr_plot_compare_time_thumb.png b/_images/sphx_glr_plot_compare_time_thumb.png
index d1dd5390fa7c4eca1fb2ecfec3663f76b0f2067a..ff4943d50f338c74c66c72000ca8253938fdbd51 100644
GIT binary patch
literal 29280
zcmZU5cRZK>7d9m!B%;VFduQ)mcJ>y^O329Gn~=y1k&&IfcgQ9?gpBOHN7i%h@9%m3
zdS0&<`FuY2{eGYKIoG+a>)avAiqbe(Bv>dYC^)h*PgGG*P$S{@eGGK?&fDjyuJ9+F
z>=SWym(QEi&MxYszrN4S93&k4!Jt0C^_BVf=$V)@-W?S@re{xz{wOKm-N2G5OH0>A
z$z-^O^_=yYc;3MudGpnVXQInzf2pt6_yj1tz#re<f6ti^$Bm&g;a=2GDHmtjovOfw
zk0y@CNUfn`@RfC&9{wF5=r2R*HygCr6TjcX*5o39Exv>CQ4FmaNBRTl+;`vWDY%k&
zLK5q%C}?RM!xj}Qhi>0AnI<dWP%$useEM{MtMSYlWy4AE0k1VR4Nc~cizHr~rSsGM
zCsI;FS<fCExC!55Gcf6iV?jeh8^~3oE7#raSIZgyF{3PQ8?BkYc(VJK*`WSD1qFrK
z&(B?IO2Tc?N+Mk`3~J+zo`N_JcubDBrhjyFa5`^2_qsUL+Nxb=t+bku@JGj?r>CFI
z%}z;I$&h+pZR;D$q{ZR5rmmjHQSb(<h2<gsJId-V>#E^uJ7Zd<6a{N*Rwm7FFII;#
zg-#cuHzvx0t*2|*t;Qd7k>H@{Ras|xo*mWibu+Cv2%xloAd?RxV)-QC$hf+?Dzx+c
zI;n`qry!gMZhIM+Eu#3Auwb+Wnmx__H&@4sD5#`AEW{gnJrotC$R6&JXKVPeHkw~^
zb$ON`;wgX|?0?VD<KuLlv;Eb@@%~&eg;9mBtehNj*(d=40eA!it*gV?!slIRb6-O4
zz(tE^sy^)DnDN-hnXYv-Ug-SzO}oUeU-XJJsZR>l{q<xe3!UQU+12bXS>pca4qH<R
zU%W39%3k&FZIty7)w|law6^klUkafF1_n~o(Gd_6ug)paVO%Lt<B8k)_nEcM*(&F2
zvRt0;D>(>|!J?$5rlQnM|9nr(shmGr*mhK0{22B^*2#$n6%~~)Sy4h#l9<ny(*Ne|
z_MV=wza0^3nR0~l^YcO%yJ&M1xjz&g#Odn_^s44<`OHW22;oX-nV7<U73f7LCU*PZ
z7U_h4PSm?{qQD|#yY4Sp%`_&9UY!#Jw>AqOPf4B~ZK96X*wZm-6;jjFzw4mz{_Nv}
za@&hz4p%M9B3v#>jP-f-_pRx=*JFib;^N{dxYBqKi*@$834EluOT7t_u($im{nAm?
zO2O0^ezqOCwwMbGUD5T&f7}%kIBq2Mp?^={eC4v7?EHz(?$N}=#Mi8>`M-Z#eNe9#
zYcnq_EV%GlASXd7P@Xc1TBXeF*{^Rp$VypP?O&d4s}&gtdw6)9tYkz~d;BhVuV(8<
ziQ$*T_?%EPS->$zvw%cCjv03CV@yoIgw;|>2RW6XVA67mJ8s6{y9q0^%k$Iw;Zo-p
z7y3p0Jw1Pmnl4)`D%^I|F#IUhtr%lNVR`B_SPkazF=i8MRddzv5)m0azW>}Xf7HT6
z$7~=~d}FF6Iz`w$UDJ?6hCf-cw5-gIkMw-6hmFf>Jo4Y+`d?c<p*<bsV3Q;(b|i9J
z+6p8etdAA-ZmT^cc-Q<wtUxnkcohQ^^C27C?fg+WadA||Yy$sgF&P;QjXd?5`3l%*
zq05uMC=f5h6_#3wyf$9>qHvcIfmxdQ^5=W{M1-uq)YR0E-H)t}HYYpae7daVW}6SD
z-!N`}pO%{HQ>b4zzrL=QcU(HTX~Zax7qA{>dRl3qq>-zx#kEvEVddiPUU%4NO&2-i
zbrudmv_I>ATQyse%y|YE4!>@tl`I5;oj@}D9H+$yJsfhs<`=1^J#p6Fx5e?$l<}oC
zHHmKAxKZV}Y^)Qvdw9rl+?NY)N5rBV85Pw6SG03@xR~O;m1f-WfrhBk<_X-Lu-ifP
z{H6O=O~<tI4RJg<dS+&7QPD5&-eFW_+cPr8$^}ka({XbX!bT06Gglou3Hm;H`V>hE
zzsp~@{s<u;7Ib0`f55wx7EGS*eRa8bCxa5B<(Z05c*fYeFj@Dr%t7<VB;n&$m7%d>
zM;jBZ&NH6D)7IIq`co8=1qhS++HX>Lh;9FA=;79e2-{oA&CTD~u;Sq0$k8dwT5p7~
zg`Ii#?%j@z0&n*wT2w-U>=zNwm8lv|iwctb(NKe;<$|g|wVP$Nuo*g=zbnF!<?9zY
z>vfz_61{k0ZcZ;5guVO1By-v68vok`49FSweH31Yqgv&bV;}0aYI`<kAu_D#Aa(qW
z5WS=n68a3cKSo{!86P>w3bX{e1(iSSEl2a7HhEnX>lU1!@Ap5icllWPW^yP^^5)Oa
z{H;He1-m*sQ=K+{o4k5uT%s-W?3pP9W<XHTpOOB*n%8uyZ6AtW91u)a*<@Rc=E1pC
zOcr<wchb!u1oy4?*GdX?Qt?%~N*!<JXn{DN7wt6Radvh#=gSr}1Fuul5ex2`ni`be
zB;E)}5}ZyO+G7bEP=H=;P1QgoM_gSVH{HE^cL$CjJ~_Ffl@+t-`Ofv1FJD5yA~|($
zt8N?q<L7?AXusfgz5lfJGTf}$`N^)eXWY6SlkZ5rR>av>U4($+TJ_R~k+E?|csNVM
zII=Pck|8w;x{lb<g&U*VrP#adzFN;HMMOl@3v^>0*MAKb=<!~=c1^p{ga7sG*S(+l
z36d3qsb!%~R9a7WLkSpMKcuFmP1nr-NF8$hCN2_;gBjBNj%#<lyu8X^4+lXAME3S7
ztc?}*Q$+|#$jM=4NJq3=S7i+A6U)0Y*k6-XRHS|M$nRB(YY<eS8lDX@A(zO=NWAm&
z^KKTIi0%fA{B~d6>mqZrWV#NRs9zp0{+_+9D{TI>*@seHQ}}`OP}y#upw#$_$2+?n
z6!tr_;<V2apM;Zen@^O|NM(E!y*%-U{CT}Q1t(ciI$yKkKCd+?WB^DONDRRR!I4>9
zj5lnzs7P1MeV&9GR_D-J`W`Apg=cL-_h+%a{r!>&tA(YdQsWNXF~1sW*(uNC=|!T{
zj$AIe^iFxdaw^-GH^=vybyJ^D-V)2WewC+uL7p=~Z1mgjPXk$sALW$6(9qDyT7J>v
z``9O^r*}w5LaAlz4~FF5lM9phH~UCsG#i!JLWwMM-8XBGpoq}S&xETQDKh-RZTai5
zaXa=03egWz8Fn;bMr~E5I_;|te-8T9mk2Q5@I1K3`W(q|W9#xJwL3euV-k^co;%Gq
z9ae`Hg9vmhCw@{$tF27%;b&;tucGQ)tGhwEBtKRu@%WX*o;cqTMW>u{lw!Dh>YwX)
zgf|;`x2s6R@w!==m~QFB*>vS*DP~6xXDda+(X6l-#lXOzgzVngDFcW2aZ=wW1zHGR
zUtiya#YMY~ads%FaJ_(7f+5GjNr9}}+|h9}TnedDA3uJCqk&IAkgHq4=v0>+6LU8`
zBSYV`yX^sH*sI{ivn|1E*RQWmR<XtGK55~&iV?M6!rpGWx+Ki3jiTO#KsTrud};bo
z-`sf5J7TxvqdM)OWoUdOeMrto>ai;8Rlbz8`Tn)9UwrKK=2qTn3*akU;tis$9s7+t
z&Fb;5)!h)MHzWxU`dwieU)^uQfr>{;>e06E>FHUNW7eQsmov}6T%airw}~>aeh3Gk
z;`#g~ov9+F<Jt&(xEHd>gXYDPn}sT!eb37o%qnz|jP0tw7+jNN<lk_*)a4R%Rh{(F
z-kE)KkndW9TunC3*UGoAFR2%3TDu%=-?Droxm8MFv=x-7pZcp|mO_84<Q4=6BYXy9
zEEYtkfdM(^tDh#zKa<C;cmSJy@%p#=Y+Ckp)no#<B@t|NDjX4q^<M>3P11iEQG4C8
zZcussZX4hiJT-b@-*RDX`*Di*TB2~_^S$BOa~{6(yOkcemNObzRj>*c&XU>}Z?CYS
zvhhYwPPckpp1uM!Yf+I1=lJjTw&7Z9(q~1wS3f_$yt=&j)Mqw!0h!!%tWaN@IWE6Z
zF@Kc9W#@&|`}@sHeM!?77-Q;mQ}O)$l`Tt=E@r=Qe}26ebJ6D^5o<EIE$CuKBXrJ!
z*50zDG?(4D^G0j@0F#;m#nRf<6M@}H?VGV%BO~ep&RYsl*bdi5shFAH!#C11ibRE@
zV`JMs^V@6JIxqmZ<A=O&3=rDm@^l#><BP}lKXJXrKRe!j)tAU^UZD#Cs3f$D2H1$%
zu<<j`o8Q%EcP>?8#Rx=&rlY)B#`}gvt(NmU{=UXrGz%$Jx<9g({(QsCv~S_SZM+GO
zN6|NR4Uh14MIB=XL6KN@bvq4)#EnH;K4hOABP~lNH~Er(VohibY@!X77)!&akqJ62
z%v@boS!SEOM!Ru?9r87t2KafOQl@<bn7SWtVUr7e+MKK!E`P1|h>A*o6b-T=vu-&?
z_OL9hr+j@jb5G$Nm0kB*rg7>1_=t}E-|3!0hpkSHZuR`mZE<R&K|jC7T>jd?PX41z
zG^g~<H{e<E`IT6yq{4%oKe9Tam^R8HGg^3P5x*<nJOJ#t(0W_+ykoiGc}-Mn@NG`B
z-cNdWbyjPV#X{vxk`?KY1*W7-9ig#nlV<EL-H^it99Xy&qmi$<*b;cF9J+xgaXd+V
zR#Vk?3=IwA*bEh+Z0kqR(9v<4cHiwa;}mq-px^x@c>YX4#Pu7o&6U%;hh@%tyBT&4
zdG|C~w>Q(dC7u{B4sFdu+LPlSx0Dd>9LE>i;kyKiZCZD)(3Gc?epJB=kbB6_pKx(@
z{C9a-T3p=MGVyXGR}J>i$jE5C!JQW{dq7~|9U`Kj*17LSB`=^?kddK;&Z2Hw(ZfTi
z&S|s7oD2Wnz2%Snf1R$?xa>X}&U!}0!h%zw8$P^h(h`7~nb^XZ1xdW`ei-i%3KF-t
zlThQ4bEGlD#K95O6-}3+nePgz7E-$DiMrOSk=G~1Hpb`AbPYaNC27@;W!RK4p14m3
z<Z;GgUCU#LM{j6Oye)e+d#=e>iRE;WN*z&*uAF^S-6ne@X6zkqr=K$IFQHDQ=FQc4
zpFFJ#=TAvVJ)hW(&HvQ%3nYJlgp7@T9pHyCVD<W=-<IECizdom4W~c71;?lZP!>B$
zIMh=PPEO*-FHj<c_TE`iy@dPHY4Q@zkP6@HM7ORY=&X0$N3eEL<3Gk!g@=IwF<Vte
zvQaW}a*t?e0wF;Y+`A|J{5dH!ZY@?k_VYh54P52|AAHE=Bc6UQa-~}l%DZJFR^1fB
zb}7_#oTNB>RhV~RqjK1uUv3@HeA8%>G5#onavl4#ZjplH#p%Ouy()RNM!&`T6uaMi
zR6t8R_`n+c)PH*Y(Zim{M$sICJEGLGkLc)vp;xJOTyKXS4!K+g28MdT8Be68k*G27
z*nJA&*)!J~{E>RKv??y-!-u7cF#}^*4TU70aOe#0h>D8BiB;z(klA1AgT}DMx@u<U
zltME<77{;sgF%0L_ybH#%$Gmj-SOD#qU%lM4uhBN21#*L4b3`@u8wj=wbikfnvdNu
zvk{ixuBIn%$Y*q<^04tj@g1EjT(ceHRQ{BvZbPP*``pRvmOzobC6h|uj}!v&>J+!z
z-+Rs&;_q~hZsB6QXn#-q4VFpEz%>wX<1Ikt%tT?94Zl_$YO~g_oFkH$d(dv)tTVF;
z<+V&wpw(h)(d*kpLts~uO=H)5TNofAnW!HS&+-+whD<nimy&D@?7W{oXRb|rY)IAb
z_^r#BCBNiIhNG<N^J10oz1Mn70olKQ;A?Sz-~2U@IZya)BcN#T4ymS$DDHAq>mcFI
zmqEFr!tZO!RX5eQQGD6cwc;ty+5C&%*2oSG59`!9vD~?H2YM^`;xz;tB!I^@^}Fac
zO&2`Lg3e3pMNLb8+_z1ku{c;Az6o%;&rCATbj&Gu-tmrdUU_%^ixnxz5OCOkG8NUg
zKn8hEzN;>WY9R2hO=~gDY|x4?u6?5bQ|R=<X+H1YaFr|9G)cf;%lgy0ebfOGR+b-C
z_px0Wy3yMqIsFwpvA-5D62R!p5VotK=Qnuu>jy1{^J7>?=pOm(=F$3v_OZDwMrMJ1
zC^sMU1C)s5dEi)fkB-`{c+{XPhm!);_mO}=;`i@jLANPB$YrJj>K5AlL4X1|D)HZ;
zaM_)SO*Eu)a@rUdk7d*VKqV9P1PGBry-5$jp&Si|LJFx4w=~k~Pd#jc{rkI#e?1zv
z&>9k-4DyO3u05<e;q{)eHQz~nTaxtf5aY8p&Z{cMx_J%8Olhq{T)7%sna%aaOsJ4Q
z&AIlVNsFsyV`Riqj=P3}^x<ie!GL-YG6#qi=xIYjLQqrHvK4(e_6f+zah2y44w0xZ
z;v(U9+PFXCeI=}0Km@eNq)n5jG;O4j1yp&drw*x@6#j7y%l3S+PJS;8TMLvZ<A)Pc
z<?I(9@yBw^-4cw=E$sUO^{5FWXPPYWAoxZ9Mcdr~K^1N7`?rO5P=Hl&J6x+Fq0v+a
z<bWG|T_<k!SHT0=bg4K~EzY*kYU4f4DANR;r2d`BKTk$12>+ntxP>Rq92uC5O)%%(
za8scfc;b*6!stuu5iamvfLp2nz2Vp97Dr~rFZu%YYsP7IqQwhhLk6aK>pR=U%hD^N
zH`CiDj{I!%fdNb8d9$|qSgxQKb73G&5*m|8$avp$%AONyLqCs6KwBMT=KLsDuI(h4
zWM8hT_Un?a?9YnGD9bsCxP+MJx6WA2_B%@M4()wgNF+PCUlH){KE^IP-a<iZM^5Wl
z4$XJ@;4dn{O{+s~MJvC@a-L>rrm+^Onr1X{gljbTmxkC)^c259!TwbAEgqxgjZCI$
z?7BfoXfGVpO1_wQ4!THKoQJabus0qjbHRmCWg{aaSfU%XgykgMb6!#{7w@^U79Leh
zGrTO3*WOUvW5H!z`c(5tqw|!Rv#W*(uIsZR+xI0~Bm2(rWPF}%f;!3erI=zt8ZL`i
zrW&|lmsKto;&PYuu)?s$BM$SOU$5EMM(>xCVCz=A9n^QKGhZ+s$sM%hHg_}_a28;8
z$Mhn{xLA@HB_jV9eK#PfV{$TSYU;VFsw(4;hc3MGcA^D)2D<e7Up_UzKzT7yYB~jd
z^b;w@>@61UVjpPVgHKLQ_Cj`ez7FRMz8*OtffXakA$L0rShtn?6owx@vGv2M&LP>d
zBwE9qE_*<#aa7!QR6MOu_QQm@QDxw@B!7u~e~%=%4+*K)_cA1OGE{QKo(@*zt%$_T
zbaIhEc#t(FiVH=$`u^ERFD~p-SN!*bASZr@(_)u%ZHN2y?vUl+YxTnZuVXoEo&_3d
zEXy-4XlX*L4g%@v>A*x*uPz6>M=55vPEI}um$JX0n?b2k-GYOTpw=XwH*lg!f?LJk
zzU6@OCga-XEIn*7azaEy;w)3TL9%<7j;hBrq-#&q@SZl=f*$F@qzP_-mrQgF1J{Wq
zTF<pV^VPY;ME0nsnu|3SgoP>!T_3~I{S^aM@|#oD{$`HNTaLi5Ib&t(_)l?6axwzp
zB;mS&QkfF^4xR)gSQvg`n)z)rhB2k_fqJ3|z->qZyN4tiF0Q*3y3e3hwC*}>kyy>u
zV5~n#l$YNBLs@rd>5y+Q7nTuSEcEyuLl$RLtM4?!M;p7SS0&Llj*+h_qHUZaUum$O
z5((B2-Mm3`^HlRD@u?;=E@u98$mqjZ855#90^*$^%N_2O9nROg-1M?-9jR?a=F=Tj
z^rtQlVxBA*56l`5%)Nm>)l>3^->F1UPGMQoL_&MAKP1mqV$y~0b-H-(y%ZH>_F@^z
zD&pj3G+|M6oxt5Oy4gXBSbI4iF*cu}{`}uzzih2=F_Y9Z3&|i}pTOo<>LkSV@Jx2G
zmPcDwM(dmZb<h>zHxmrxp^KuX>gw*4i({LPGl@ja&<y_~EAJy4+$6hAvU-x}a6(dk
zazB>Zgpq`0HSCuUwyr6bVvQ2<)7RWaCsY_?s;qr?d9B938%GFBDJtSpc<i78RIw-9
zVx6KbM|YVu)@H{d62gu6sqvGpM!e0ECR5nx(xTUiE>M#$(CMaGQnbWan|Ru&Y|5zO
zhY9#4{$WDO%r~5q_(pg6^TC79ZFB1LD*1!v>MJ9Ra9(vQunMGf>KGb%%tu`wlo#gm
z8fHWVcpTK?8=jF=xU9$CiPX$c&6*$ScBdc9(;NORg=c-Iia<tJm*hG+b}IDHC>t|P
zUs_ruoM$T4LyuEpoL?u};*bbWiE7j|yr2}~DN>OAp--EP9<3<(m^URbRGQf-`$^g}
zoo$9<0s3No`eFfw>J5>xzdVbT6I%;LmBe9vvM+btR0rRbO^38CJ&6-9(tsFzF=Y9e
zsLcZQdx+3epDTF<!ZLFVHh3O3xL|BvU!8wBV~las0UhU$o9UNG)tyxQS87;<L6ex&
zjEtdLhMuwH!fugqaqWK^+y~9MfL!8mSQ$tfQYHItR<`>~&t|aljUHq+N43UaTS5}D
zDbc5J1F0Ozgp)tZ*r!o|&6B!8QnA-6)Opq(6-nL0M?)N}*3NKGUmhP<w*6t(r#Fgw
z%(#06PK&kXBhjyF%2{J<>>}e{wHZ0jM^bMnE_%isM}}I<7oL4CeQ%)8g~h15yvuHZ
z`FL&Tn6XBwb263X<L7#8l6nWZM;vS2RoA5PrW7f0NVser`+?^Ky5UAxn_omkzoX&0
zg8;wt7OAVN>qw#gq(+9v(S{C!QpSttxScC(W*W(|hh2y9V|H$HxzU8=aPRCB>~xam
z{P{DiGK~*4vO?;q>jiV139;p%fbwod-yE~um$tdD-_sTP7WHCGVk}jDy~F3g>V)W4
zGtVB_Vh$~#|CxSAe)mndE<+Z^&Je+DCe?R&rBqp23E!v3r%4W<91_zV9}>D1p!-Yd
z5Y(FL;|IvKLPuBpQj%JB9wDV6DT{di!^6W<&Kol`HD!PGlQfQ1zkB(Mw<7c%<E5sz
zkY4oR!-w~=(Scz4ad~m3Q)Y$(^&R?l=;<b)5wo+i13o>55C8bBz%<L6k2~GaO4_oC
zDM44i(syG~hF~2`-X{gBHCYQ&i!mmq;T>)hogI1V&5Jr@3pz0>Dup_^9|XVXbiOR}
zf1aOQpMLfY7c*HZB-iENvjiu#d!7YtV>t|K?^3m}2<yw`O8iu$ed0@5O$J^M<iv1O
zsC@7|dn@RB4lJ0*ejhJvNFa3725ze~BEW8|Y8U~3{C}I(=NB8x(0mLli9Q2`iCD<x
zF~VKG9?k;cigCY*kq|mD0JC`-`M}_$88&$phKCaq6EkZUBU%Y8gls&k{?x_0mWrH;
z0plpue_Y<FrNv&s@1HO}3fEms>}JG~jr&M#OZ2RVpN2SWI`=N+-QQg$<6(L88w=W2
z-Hs}D5^{y36<fxaUmBi5czzw5H7wumD9YiE_>#M_-Xm8U%TOcl?(Y%0b<8u>Sl41h
zXMJV(8*~)OU|h$GRD|N?_x$IOi~mlyK&n79{=PVHvvt1-w1Bs~b&qubx)P?!*`|MM
z`Gq&m<`+=qs5-KqB|YllxQR>p9iYP5oj~vOUlRSY-7+R~lq;fo{Nr*tJ29(#0|u4M
zhErs7ZYIRvaI^E#>Z(a6*J9|)1-jTLI~j}b#Trj`@=PM37nP_~Iq_klFvi%>Ac&<$
zjw9NXxI*srD8Wi5)qtLoPp<jzKD`8|TRXqzEC&d~Lr3l(CgNw%s%$!0Rws*rg!D<!
znf3Ja)WpOj^3x}m){X%pib0IjLG$NyZ*-XByf-!`zsg24ZS5Z&W$6o=GMYlBrTKZ6
z$Qr+b$jt+dk^6ivrj%AS!#_I%-}+H3*8AF0w%#7GmnxZhX?=5yZLEb}Jr(p+U*<2w
zCd?jSCUm<gIjNkH6Q&LmRrNE}{>mG)jHyVP5BlnB(5<d#_Xz7ib8Da7eDHOc?)Kf-
zH}{++@Z>t+jO1&v)z{Y}CxA9O1fWSgF!rhqno-nIQBg9ovUAW~{Ag>tbMIb9`qMCg
z$DjG^{O%dJr1h4B40f=5{Us3*(!CkzAG5~Hj(?|6{Dw<iR$ig#JOgb;jL{?+K9o<W
zOf^KNt*<}m=*)C6{L{_#E3<#Z)@1&sQ_*)+XC$j*F0{ESr?kMONXaf!&#o?cJ}c9!
zaPoKhq;5Hh<Ei71!hmkm^g%!$&~DT>H0ZyXmT$GF*t<BM0a|pT+Kvj&?fd)eeo7*z
z0dQ(}=fl|k?3hdKP{yKfCFPtuoSb5=@IJ6`JaSNPdGob;E?YeNp?WsMs6&!ulpSAK
zcV<|3)*IZA`t+c$I)OZ<i3KZ8n!d?^WwTY2pXwLK7Gk!M-Q9S)eWO~6|B^P-hcY#1
zU)-na?POIomD6q``9)^oM3tmY#U!Q);5+8$)seyEZts8qHSj#t$Wlz!&D$3Pn!={>
zbcyizUi^()v5~shr|D2gv=6w1_O6fI4aD|aHXmgyx4>y<vba?^<JYkjQ!-Aby%Cka
zo1-d3$&fi>_F&@1Q`t@{x{UAbUF9C7STV18WTRB_HE6nX^-V`fxJBG>e4+Lv_UTDS
z%GTf>sBZ0ZTj+<&9TpM^b~`@3e+p5l_(h}-K>g}~c*4Hovu8Bi+=m}Tl-E9F|7!94
z8}{o-N|3af<TKbuMv6uAYh;b4LMGf-rTA;|Rbha+HW!ElI!owHN9feP(PS+^x$%)-
z94WL$w@?sZ$x3HMm;9}j?e*1=-)n7%37CJ-QAYDLOd#97`OT%6A|xv%g$`=8>3H$W
zMdA!);@pvei|b;HzpeNlSWK_4=Ag~R?J+Pu`*6_Yg&xJql+C&Y7|f17<8!)^pM1E~
z8xulC*>(okEz)QXI=uYRZs1vV_xCg2Ojdr)%mi6kA9`jLO{S@4b&HjwmxD43#4E?b
zBU@|m&S?_0y7AY&*~kWY2R8p{!qp)x_22ezDp1TozcA5}${0-5%!l{uwUx{986jCY
z@-6+|xT5p%3ECXb{1Hhf=-xW@F6_DmaSID3Ky&NVyrltwks;zFVA59(XZnwt_0=BN
z!5QkB!@jwUA7ynv9Ih%G!+%Gcw|rXXUI2Xs%~UnnC)+v|@iFJvq_7m<;3ogtV^<z(
zalEN{r)xlvJYr|R18f&4tvhqARfls2eh<_wkQ!R-f^m?H>FDkvzzF<(=O$XI`>{1*
z8UPc5fq?-w3D-N2e`2JqS~g`egWP9OM|RfnTu$`h^3$TAY?Rx~qyQB&yS?2HHv$3?
z;VTD>G3RYPO8K4jvVmj09OtHNDtxp#bl?mS>X?Mf+z0es5Q|U#{oDH~xc%;_!`#0u
zPx;mE<8|egGn^6c@56U6{cNvKi5M|+bI1DmUHet2-vhnyGhl9I6%?2w#semk<#3K0
zahxLpXj<nw-zrH{W3=Gd3i?jDZ#VRQV!sZ8E&Om-e?Q&h$HhWMXVSop8~;041*UR0
z%bRVEc3p@_H^Qe)a;}p@Nty~nf}ut@aJ1A^zVY806EF|J<Riw0le4pUX{)gXTRv?N
z>mUvY)l93x6pdro+gfMShzDlH``f87S{!5L1&_f2Ft+)C*b_=QqPzgn&|J6ODBl`{
z)79H+0__7>JqELtC;*H-b#oJd3$jmsv9iA2ldG1ELgPDa-!H7CU(F`+RW1ijoe}ni
ztJopn?H8Fxe#&XC2Xx;QvonAp2FedQKonpK(SaMTheCyrTVNjm6|@Tod47i#JkS7}
z&C2FLtcNo>QT*~cRExsn^2yE5!tQaPo<@KC_>i%MksuaJpaJ)dp&w0ub6RG~2N}Eo
z@7W`aCsEWy%-Y{i_m^Gg0`HC2IkUk^w)XU_*A^Y_?LA^;{V*_~Ja+pZ=y7}h){E4t
zto7G^74$TEo~=yR9TU3C=V=xFoT{<M#KK}6tY_!sj0S3nD6Fjoq|y6)wyAX#D%GO`
z^;x8~4Ga8YK|!}&1s;jx6}CX=IjpK4ZO`<>`2(#UXn9g^Z*LNA3-z_*NYjZ}F=|<f
zY&Au%!BvOx&7=Irr4YK5S*Ri-KqCBLcsDmUR}cOX6opTZQIq;Ke?`!`s0NS4FcxI?
zGj)AcG5jw^u58`0);m8^9c)g<gAwOH9hWG7)J<|b4d`D?Q0oH#Da<(tAPh1s?NVM5
zoLnkaR$N2F8Oz2ZK7s~(wcOO!`~k1K=OXyHmusu<(8NzD7E1{(z=N>>#vFwdp@gqr
zsVyMD^TrbU%tS5@)Gq(6vjGp}GiqVl8Ts%%g=h+JJjrlghQH}>I!oSVEOR$NU67~x
z*nS~I_K_0fDInzIDf`=(mzO{;Vu5Q0^jS$UF|k|&%g6U!4hE%HXPQJoVgnu(j3%cC
z!%AIUT@QJ9h{Ch>D1_LmS5s7cuzL)HzOUhtZ6lYMN{ZiEYTCo+x<?Idp>3`-c^gO)
zqxo7*^Ovxl@FvqZ&#IHBE`&A`Zmo7axppT23w8vbf<no{f-#y-Nh*8z4txO$kGw`8
z4MrxdY$`kbt#xo?U|3&Sz7GvG+L>*BPbNqNIsjOKdfLM7^-%GMSHH)WNx_xC3+Sm9
zzBU1|WWUt=5JCd@*)q_)^A+VW{my=5DU0XGz&!&)1z|J<wurrV(s-8%e4DCxjPHS(
z0j}(2iFTxELSGUuqIE(fL{Wpch?b~1aE}h15{^g=zwiHKMM$I|G-~~o2$xbzuXslp
z!^RGG6Pn(ObWB)2fw%4r)!4rR2@y53q6;n)Hkj!P#*OBu*y3nukL)aQ5FHI5s}D$O
z%HU!l#l6{d`Hu;DLYATvM1{iYE<k&`Ho^cR7$U6|7vuQ=svk}j_kWA2h*wDYd;gY1
z;GE_)SJ<c5)3w~_{=Q%yKw&kge>ZJ?Z=F(tU(I&`CAe+a7sF5a$7Xndd_1dZe~Pfr
zP~1g+QRDo0iLv!fCt-lyg>sMQa8rLtykP!zh4HK5w$3XgJf?Q1{@gUGn(XT7xgLWS
z6&>yG5p8)$hI<p~G|>G~{hPnPyF>M(%fCu68I$}U3GO+zyO$psXWCECfpgh+gzk}O
z%>=O8w?$4UZ5j@V^eU~uhV=#V7eu61-q`O=Cxj&hg&QO*hve)=$92u0UqqF%6_X2V
zzI+AWN@^;l6;F~3xii)_p7}+66qn*9YMqx)V1%SLmH96A$QC-f$^{)pRp=z<S0G|N
zptrF4Q=gENLk|uj`OkPKN_%^I;}w?ofW4;z`w}jhK(W;X2e?w6K)NL6G{Xk6{>IJQ
z<R)#QgrM`3>#~4Jfcnv+Ye4*~s;QOe7HH|)-zk{2QFEg(AJD6LK7OOl+vn<oN>){(
z2zoe0ppUtdGy_JJF!b~Q4k*0O*};_NvD<bJaZ8+SH#MmX4*L7~xxjZ2(?Qf%uI@8{
zeZ(Rjd~bf2zXSuB)pTv_2QtApK5a%QLI(#9<Sw(QAjW{j$(|+>Fj$t-mk(`5FF|4i
zWwmcQf?nNZaJQT0jFeOx&D|^D4QJ|>C*BqIleQ4tMnBtVKuJjXLktb@wzhnvkZWwg
zsqw(VblyQge^}Y9ECEy>l$Ms3&dyFTpeOG?d-CMT03!JGe^G*dOP<v7!QHz*A|FbE
z<xD2Zp^tMka8f~x+x*IV;O%`<ovd%sy<0f!1!<?l%GnRh<#;)W(9#@WqoFGs$x*p(
zV`BrctfQmD`@BVO3jc;!_VCA~q+4Q&ia#~_g|u?j4>l$c%ap<HC&g_0<P37qxrgr^
z1Y)(%Ubnj*lVQWQp-waz--K=O&~~g$`O0f2e}0wz#$4sB(>27>2U)@lj?QO6=MU&O
zB+3vjstv#9Y>Qv@lY<T6C-hP8LPFBuD4*}N-u^7)8U<-mx7=K1X*@K$EkH0*>L$B0
zQ+)YfTiZKUr=UQiJGx7p596`&)MNPx!fmj#vwOgiA{BDMgWefJNB%+=NOD+Bb2w$7
z{lHuAE<`J(53eG^0U~ICy1NxbKeZ~cE|$%kdMLwJe*fqa&3##-fOWtMZ)D?|g^KM%
z4BkB8q@gH)hYU_rYZQ(AbLsI*NUG*sBv7r}z;L%8vHcP<FKEd{)!yHIP)%{<@e2DE
zZaWKSs9PjYVoknknV;+aCmYtRyzRAvb>tA^r~DdHQb2Gp7|k$l-(CiXn|<sSoGno7
z5=6a4!28GQC|g81#UVg2vhnS?DlQXM)FHu(zv|{EksOSeg^RGZX)S;#YdkDgdlNXJ
z%Avy+0_>9qZuDJ5gfq)~NHm0hP--KBE7@9`+vJFM3MaY`G|xGe3o1VS!4I&ck2lap
zOp(A^wwQ3RgGCGgraUPJ4F+L~(*1HT+@MR;2GQ&GJiVJYn}l$Fc_>NKVGu@5N`hjJ
zhJKq|q2mJ?YUTqQ3yZw<?+oHO>gDJ;Gx9pOYl;boa8((r1RCa6JJHfw{^!RrgC+&K
z(#Ll5Pr&ac{XbN`ShJkM)PMiR+%L7<Gz<BaBh+s$PZ=)4L-!rjk+kjlA=hR2g^LSc
zgWI756d8Pci3oDxA3BsRn;C<xc*E$L5!bW(h6lk`r@C}U#=(IL5i`N~^v$p-<^0?O
za+b5?nHXGznq}&``;Moxsa?aOGE(gcgY@#J;0}yo!_L8zxG`N9Z}+#I-F468v*Ih)
zHad)yd%E|2Ur>Def{`#!$s|s{^y|?-aQFNGki0rw$D4Z4=C+y{&j%LS+{pwCiOc~D
zK2m*j%Nbv3Gk-omDZHr%zToePThUltTzm?P+}YJ-C-_)AN4Xq*<>=w(=eMN2jvcuC
zlpj{diXk8g9IvuQMvwUU`2mQ21P`n-yl5{}36JqZ9!u<pM$12bE!YSiC0~a60cJt)
z-Ys;|eK7&wm<O11upjWehc<dIv>(ym!3C5EO$uuF$4VsifAKsbhN7`}d4V>U_q*vD
z3ClVE;P6KbzmA#;^!d%rJ`fpYWo40E4-Ab#)p0Cy5BNnO`Gh2|L5{eO9b`!^2`T1D
zH0?7!FvCKKE`gp(%gl`J1D`XPDSyvdB$P7d0|QEJ<85d&f^JJ#2_h>^uEyGz2NqiW
zYzzaW4>$$SQjS4`0Fro1nfIriXG**-+yHGs6{X?fiL0<2GvY}C@7MUaHt0ffpvZN0
zcVD9v7ET7dmkI|MNKB2Uc&QBW?BUoB>AL~)Yezm=5#K3nnunRh-t0L%@w091&d{1;
z*>(aSbbr&;g-zqdksb<I=-S}!P5_xAjg+D1At8n5!JWp&MnKLcI_)N^m%bJix>^$8
zKwMZ}24~LB(UDVwH`qJfk2X+eVdq_3T|qL_-pXtv2}ci?^6%4h<&7uCM*pUnpWSL3
zCN1JGhx0+Rc+S1JKFwh#w)^N<zG&>Pg8<W;RrlT3>9H3xzI|r-<=OE>bk~AgRo2Gf
zlfmXF3HtT)*&pkMT5Jgk2{|rZA<utYB_$=`1sTlKVA3oy2&a}U2T=559S$jyB8b8Q
zPflFG<=F)&7ZwWwrZ(kOiS~Gv4FyySsc@2)6}l6Rp6)f5ivt!F2!R4A5ERk~Ip?if
zxlF)Fh!g-BV_j|}HbgEHWb!hXT_bYWg*zh_hXohd7C2lNcvu_7o`SZ2XY{g%OWrQZ
z0dkcE?~KjLO|hrIDplLgU8{56hH57gyETJKk7+yC5{THlVVN{)y>Eh90YDk>C9en5
zpXOds<ft=(y7Mk9%zpjM{p(CEwQP>M+(S?Bo&q<$2g))kkflC~*?&j13~zkz@9#gX
zpPik>$6PdnI}ro}?lO4l8dqo5&5*(j6l*Oky1ayh-}<(=sY%(y%6JJCi(ch{1HI|n
z^AkI$$ZEM)WJoy~GjJ~$s(gX<pLr}8SlOw)ZO<Npchs`l%tTf^4|tNo0rLM$5taw<
z5D=GiLPF<xZ)P0@$kNi%XtwHv>|7*vM|4dW90Ycck2}oDG=57q!Rr8pgO|sW$WYxO
z_H-MM{wl>2TqR^ddSEpK=VvUBFQ}Ul$qio_|3(({{;dZC{$t=6kq?3`cLh%MbhBzk
zyULmjh%nwiF8**p;tw}%h|4D4PgoU}zpCshRRHAyt|F8Ch9*~4>#NS*-XA6$8o!@_
z?;UAk`KtHGZ+uQnl#`K>$rNs)=jFXup}T4Qh;3LjRWC&hMhTqi$RJd~|I*WQD^|{N
zn!^D6zjIl+q?k|e(8P0WyJN%{mp^S1*$X6loNZPK?e}tm*|y?I*F3bJ_W^NHNxh#f
z>leu{HZ2i!-ZBB_#QD16)jU|jwS-c@la5Roz;T}QT`6Sa<&Ag#;~M<xi)S1F=y(@}
zI8!1vgI*}1uImL=GV=2K$}v~W0P`W+?t;O-)M}y>KO!?TbN6V{hSTH37MNw@UbC{`
z^!iFj1svuBl)z@WwHYH$5HP!76LSOr6m_Sp9YPK<U_TS+-D$SsO%uQh&;E8n2BD_`
zq|T*bTup#Y{l2gp7x;U>St?CkB?3JPW+v&VuSw()GL)Icc7LUKtfxBlY#Q3&N5LMJ
zyUyNVfV48M;BM!0*`Wj{)WP%AF=cJ-POwBnlS-*y>oAuQA;PG#MJJ12X4VHOyQbxG
z=K?IV0KV#>p+kTqoU~f<NocUa9y*`!D&-|~cjRn^)o}n}XQ8vVENW<ncI!L%T=XO0
zl-$;dOW=7!0@ey+z`5k_57wYpR)lf~7com1;dwOiO1;P+`Rc6c3LN6ZKp41EwyfuC
z<P(7_f15BV!c-?qvp^Sy4uZTc{;`8u2J6->;k%0!rza;n<HivmqoUr95-9>#{Q|68
zKweEpmD+w3lY2SS=vn?|G7bW7xWV0dOzNo*<jDU!_At19NW;Uec^B3`XVl_%m5maB
zI0kxpaBNN<o&Xp^P#ZY<2%T`&W}E$yqYv^-5{MbaHZvk)#+JXn&B5%-Ip8G(=tDhp
zdpu+b#;^NO(bK^qhS*VI3#-X3!C8yQ>kxn7KOy0<T4|OkDm3d$<g%VR2@R!y*@hbX
z<?8LmGxLE|RB)+$=2v6yf0nCX#{*}oetx?~w;&75?clK_gf>%wRxOaZ7Xt@}Mnr@h
z`g11op1wXv9$r9O7*O)@oy4h|g`lZofk91UK6`kzFmXckY=hI!&u?WUcens|yVi<F
zFU0#X_()N~tKGpn<Kf?BVgzfo0haOldLgAiw`Ypk!y4A#hF9g@ry^3+NX{K7j*sBY
zLm3J~#o`v3L|PZUBVf8<cX#(86iU#m0W4)pTu5t8va|JYL{Zn|eJa;|6Zq%usZ!k~
z0Y3gGhya*O@dzw@A8mRXigw&~&oJ`KmwpHmM4AXObg&_UJ%esO4#N1l?aIKzJk0_*
z@OWzG69V2HGxUB2&}zHkIHhB)<VqIY>`=VFW@kr#{`?+1)`xCy?d^B>_R_VA48RT<
z1SlK02jz#87SK-E&VEM)<qD91dcI~jxaz58(_gj*0Tq`%w$7fQK*-@0EB6%M5}L0c
zs}66MdWoS^?)v!?EFutjPXW~52fG;<a&n68cNk;45B+;q9iozx!$B(_8;ggY`9;+v
zFtjX=HbP3(=bT{O9XovrUMz(5fC8o&8lK#SDM&|m6U=#FJA>dt2Bu8R%%VWd7&gY!
z(9VoZPVTF=`};hJf({fj<bDCffw+lH$`=V~eRgy6S&=~lfc}bX$;K9t7CH>QF9KWV
zW`Pibit?ZFy>R*hMJ-Dq4e+}V^c%=o!=(`E1%ihCJ~k}G4gdsBw*&J(9{bP0ehhU+
zfHoQvA`rT7Fq|DZbZQ~g7i=~FXO#k7?#qj9Z{!8E3ZD=^58}#x7a9sKO@D}+Xh1D4
z0t@&&pmBpH+Xg@2vQg5pGMI4_@!TWG%=_ws_kRt#FA%Da-40f;_{nOWHg&;t*$nnV
zWLgNUIA+iiLSl}EDJ8qz`KJO-8}FK$L?IB0Y}JNfDr^C<=mXA11{IJQsT;S55QF~^
z5J_SQr9j+{okEVaF|i4w|J#rN>7fSdH(01)hk+psiI2Yrv=QigTAY>8ZGv%Hm~2ZU
zS=gNjEX;ty3Ty9hKyfM142zGCznM{A`Tjk2X=y3!5&*G}adA>FUQhzT0=OG6ivE>1
zV;*$I5dA>>KFt^e9a_rP7LZjCd;;>_ckgl{15pwZ5}++e9kuvqngG-c78VxBgQ&J-
z>&oX7jOp+R&=SG+0&;lt=n?or;Ou?^E*vf%Y_t&1VDn_0T82^!{N*q1@CU@iAuwmu
z(b2)(XCTCZjfGYF(UBC!?|_@(xC5;KGCUy-gIh51qyS77eBJ%z|Nk<F5(x-stc1ye
zC}VWnCV5yeVn?8!A%kXsNI}z#ij55%wP-Ud`v^XE$YyY7K5tonLKveU844ESSrQLB
zOxYnA7LYtZ^PPo`I|1wXP<6YTyJMM7VJrl|1RV=Y`0n41I$h&(tBEK8X(~-#o&c#7
zy}U#q(?0}u9vEvFLpcTS7@zGd3gp{G;DU^xb9v+o%^EUpm<6j2O>{dP4P<1@e);Dj
zD6xyaH_64oV-3Xs2&o9+qx+Cz&rg@X0J`Ob$s@34-xfZ=gPQ&o6!foOze4jZ>hg~Y
zQzG{r+9M3>?%v)7h<T5HtJ$#BVA+KM5y;^~z&mK^MFG<TEipnW2jh~xCHrSwHOb^~
ztSep(DGIImnOj2&u!kZiv)A*~bNzG++M&sW<ruR!(Q&GnJ%S_zY%1dI9N%<;QwD;<
z`LgI$ZsaRvjFyW3J{}7PCll<%9#Hy({;gy{oV?&j7%=Ce`M}3&eKiAoInaFV(927y
ztKX$lN?CwDE*o|SVHx0cgW!}dL4Ty~lKx%`Z|duy`6nS)4%mr#U}H@{_=Kqml?FFQ
z-3p6EHv_nZo#W$d7$SrTw?L@J$bjVAw{IUOC~&<V3h=#tlLNj14D@!h5;aC9RqZYo
znBJW=dH+E7!wy~T2Xf)HRb_~$EtyyjU$D9Z$=lj+{HF(~Rk_a*EHIolMFK-#8-E&-
z5O4&1yfIG_sy29B5gQ}a{J)SA;2!ay$%3mJw5UMJ<^LO`%Wug8@rbQ{=h}AtJ}#V7
zkUknu=5GO6Dh(_wq_a8rFi<FKEnuq})(@dUAfQvMUR%>(Bs6`#43S>VLKq)5mZQ>I
z@Hd+2`87-&oHmd*8w?_J3uqY_FkyTRKDG<WByeCI=3L-7Bo_4&f{7g#J9H-~F+K5Y
z9TpX(FI&-J+$R^D!%+YHzsSK!njctoKxXRTp1?Q!g`)`>{PS54Zuqay7+I6&d;`M>
zP@e!q%0rI}3LVq#Z-@>9?+Zs1TE)*Y;C2X!h-h+bhT#SEYd{2xM^RV($&jGMlLg}m
z<X&+wDnbK`j8T=Ba{)gMT^j%es5QG0+d!ifPgqsaBSVMX8A72EzzD%a0~Mf57}Kfa
zj@R6P-c2EqD+GE>gy}#&7af<h6F^uHa}PF|z{n`R9v)h=eiZfeQImO|I^%(xXH8YN
z9gMF@NtHq-*8HX|4$@r>fUVb{*Cfmo0X9+i%y$RY4@TaAsDySE$;gN=Y-VOU5~kx&
z7OQrE4D`WQTdaDJJsbk-EF?r)P<c+9GUb&GJlJEH)<O~uIJOh`?P<Z?zl@muA)<KC
zz0m#V@`Qc-{QakKG+>Pb;Dp05ffHEeNF@(GZAHbb;^N}xW}%^=o*^@3;OFLp@fPr&
z%_E~nbGAEh((T9eoq?U|rc#8pNdgkZmD*LiKoia~Dk^H>8E~zDSz$yle*IWAe{@k`
zkACS$5Nty+j2d~5jecMNB6@xtLBfp{F7<b75*aOt1FfcgZ?Pvq$hCC65%lFCn5;^1
zTVuE_rRfIqgOJ`~24h2x9E7JVP{yqb{E^uIfLtO-8_$%L@c{Vn@NZ96-3HET>{npV
zmj@3X$is$#f8OeMc?i4%3`|0|?g;Gy=<L2wN05<+q$2ybZ>wPk0F3Y+?gMXv%-kY^
z6!f*ZI%Tw=5g?;kF#j2^<pIaD>IwM30^VO6u~vLF8n)AxQ}_0W28NB0Mhh8UgAW2}
zA2g$HdQ~5Q{O7Qq;!|fNgtirwFl6px2m+Oe^*Q4GA`@_!1?373m&m^rDg+uJH-vBt
zLAYe0U%pU4HZwihnM0`Ch0Z&O3STACa|jm+srP>#AFXG*x7g~%;J&(Ps^!*OSjQjV
zIS_fV)37mrQ>sb63a9Y}7TUJ*07*<3B?fuPoC^sX+jy9);Zd%$PSE0&h5Y7Y0J2Oe
zLooFU?FBg`>|(b=3rNBVFj)<JCd?LtG>i1f5bj_tN9KiLdUa)@j1ICf98G)ZiXq_M
zM?~BPO!}UbKMG(TqRD~4h@_FCh9eT(;J;blJnu5?3ob7#*!Bw_<G~yN;8x3$DeEfZ
z3f=9E1`+|Plc#<KC|JV}kFSlh%$75b;WF`s7H$Tnj!KGQqqR#YJ1$mwbDd}f`sAhV
z?;X(L2TU6DOWi5=-v3id_TDZ5h3!IP`xxd+!BTE^@$c~c1Ktl%zf@s<5_o8+QZ&CJ
z3pAOKNm1|~b*wtz1H0{a<JLIXb*}!6HSs|UK@@fakp9>@bht2G2>prH&D+k-iH1rE
z1f8&%U>gDd4yX5}hnA$IY?KRh79pXbj8&9O$vZ&<>t*tWJsf;*p58mlwCu{Fj`<uG
zX-7Krp?RXf*5rnaldXVQ+;=)Oi8V2qHyDLDUK-y8(dzx0>5tklBwQb_aNRHDI8;YQ
z@U+X#@4(W-5rVaO4V53690FzxsommnrT|n{e^)ShUmV`E94ma9sX7S$FdzpydwPta
zmqO;?5!ebuTL*1V>Bia%)}FT^e`9RoIKIfx%i@DN3pd6^@+{;-l~hKFgjF|!0#=4H
zZvnKe%9gBQcj8>2xu+jR$%M+G$Qb_(E%Niz;TIy^G1_A9qh9z~O3JX%)Z9N8&)7ep
z1Ko~=2^qqZKR@y=19}e`{D*W54HPvP<iqZ<4Skg*K*j>wjGCb|0PIsxz=o?dhO-ID
zR=3Lf8IZB?MMiOPBq%mO=0N60v}^FLLzsgGo%uA32q+a$ZJu~|)UUs?gQx<VL&fa{
z5WJ}#KaK<uPl8y9nt2}sQAF$ae+ts*=K6D7OMpr8!NEaOPs<4TUY`Bk-+zUsY2$)a
z$WL$#U#Vu5`sEp(ei}yS4~cXn;pnHwq*)ds)x)geOrzA-QNzr=mFD;`{mv6E4g9bp
z+!W2bXt#&g4;?nfOMZB^J@{HwY5g3A@0b$XoM418bU82}0HvFD1G*N}Szu?O*XX9&
z@K#?!Q_YaVfjA7(Eod~NQZ%*UP1nfpWG-g6U7C5nrWx;Bke`-IU#HIPh7qd4%FR|c
zV}Ma9QEVc)hsxJ<T^09xuITsAn2b12xPln_9@g8b_~oThGa=ZM-u@ai(=ZDS`mqj7
zp#kRlVN?RR*A;AMP>Nal`Q$;vMBtycwhZ9O!uvOBn0os=-;3||{)zg&pdmfsz>&;L
zk)!1B;)X%*FIR?p+jtm-H*hFax7cL-Ki}#;7z@<5HNx3Q$xyI)jV<>m#5(QyHRG7U
zus<DTU!KCGRP*M)Lg<EKp}yGdEy&h7Z~ug0cl~VCKg=*Y33I~HpFVX$!E^<@*K^^%
z!etRuUL?Lw()?b}Ez*m#a!;`+XliVwTST8Q(K+XBhy|!=DqBpw0p%HyZfL5wExQ`U
zr}mj{y`m)7j2@O^KIxFuzW(@z_)}tEOh4t9h!^NHUo$);-~<X2F_$?W=uQ87pn~i4
zm~Me4lFdO(?1x7j$UwPAxq`L|wDws9)&cj~S^7Mk^jY@sV3W5u@?&5Ihmpan92G+E
zC7d^HuhtcJ_F5W*Jkqf;YR71HNHX1G{j~6Yl&i3&e#_T>`*0DDdP_Cy`-gLMy=?yl
zADCSX|IqDdV%%N*hZbD_?O>gOhdn@Le3r=d5;8d`JRoyC&l@8G?FUu{hP`Eg{KUL<
z%NS_RZ~Ap{AWH7EKd||4n3j~kC6tf}s8)CxQV}4<LJ?VRzDZHPbQZ5g^C{k-=M(wi
zYkjH!=XkP(!J;yk?dG8JLc1gfj9)9UxkV&oF}HlU4HGWq1g}YJeg9zda=1fo!gkX=
zZK(lIlfr{(;++B><2%q@ivfZG!Fw3duLtfUcy`4LVC1UkONZ<QeGgU!TOI45LxfO^
zwKzb<yMMnEP*)GA(E6s|^R<eWKy%4}hzF@G61dU<csc>{43M#PYFQ(&3xTV}uw;fB
zo<#z;dmo$`NXZ6)ar)vNE7^L`P>$RKk$>AN=>5N&i`^|A*Gu;0SF+?BV2gCq<Oqyx
z<}we_Z3gdqGpQN4FEP}uNUn=*ZpqE>hT+E=`0d<=#a-_I(qY2!5!!gj(I}?ImDI9m
z=;%?vPe3;{GeeOvC<9h{r~pvRfUICo>Vsh%P<D{ly?9)0W@F6_!Y$1H0`oQx+&830
zXk#GRGC1O+E*3}6GqUE7uH|*%;^NW@3Z8{7Pg?Pe*&~`0Br^R7phMWrdhY{(g!Y{=
z-uMCAy?NsoZ=15pbBSAmnQ+TiO$4?@$KSbqTpTlQk{EIoulUuSxhdbh)WNy_@FjuO
zDYj;Wnp{wRg-TXkx}HztwR>b3Emu*WyB!Ma%F&he`OjYRB)R?#=Y`A&Lly&&?g3-)
z12CiTI(Z1?loW-4Rsos(KL=5{{&Xn`$nuw5B(n1Ih!zDTe<)%ufF969X9?I3MB3pp
zANYPTZW9S!KL8z&Z`ogbM**Nx${fF9V@N_0-acB?_Tq6QbfSo}skgU&&7C;B-4b`1
zizJ+>25OU9dwA5?p>i1`SNPF#!(y<j)e?`vra?r9uBKz%9lgvs!yEJDTC7XVruQy(
zV{M%XHj-}Vq_zupt82xY$(eqlR5uBZ`)<=;Hzptx_p+13)Oh;}d0q^(EbmFUL*rQV
zN`df09@qe+FfqFk8o-PCot9e&TMHS5^Tl@`{2Q{R%o(E=#4sxl4Ew9Kkz8iI%J;yo
zBm$qS^%D*M&K+@h5CPnQWd7)*p0~OEsL<n|AC4J<-yU3;6d->itO+cS30NY478-)y
zilf7+EK0hBBu&I)e?h<Z_vh~?vC`vj8I$dfc_~}2aFg2EZy1my%P7i-qK3H7j(fk@
z6zoo{v$;F6g;lW?IGdP$BTRYsM>!(f{qAFU-IQNt2QtA+I19*gVh~G6t)nGO<HDdi
zvTHu2tfXXQ;kpG*gm{EZDV2M>yHG2Wpo>h?%m*%O4jSgIw&zlSgkk6}j4s8+4@gVo
z$tBdX2&)d}0#hC6w?crWg^?oQKj3)`2oT1{$G<P?<q<d&lQWh>?4*RFt9g4NyzS=7
z)uE=y%J93R3d!x>ZmkORmNka>YSixb_Wj9+FF8r|ag#*VNS^%K#_9W_F2&)eeCOn;
z{(i8xE2w73JS7xd<T*aLq<mt)(3!*2ZrGa6IUoip!L9{Pu4v#FVQjPpt`eFkXvRSa
zGb*9%;lK@U9bNA>E1NzjAQ%Iypbj7%ND6!~YzvQuQD@Z1%`OJG34?OA^OrsMmYBk-
z^L%?LoOq+P$QDm)>RzjHetJP641*AC4Mc4I=xoVE(FXpqMa?-dVKZzrlFz<c{>i~Y
z!Hq>KnIu}#{*1Zz!Sz%p-o<$r$=3~UP54Pqh@E-^f;`hE{&tHHidTojbPxcGAKsVy
z6txbk63~zFd!7F$4o6WVu?MIZ&XdhJ_~yY70Ok;YwjZD&gNFwI+`sJu1k>1W8=z8b
zWZ8=Hp}7OL7_o1F++c~dZl*-0p?*nAMse_3;Y>H}n_!|`%BZaBhrnz3(Gp45>(oxm
zKB9sL-PH6hZ5O5e#2fH$YR<hH^2I)U<e9K1Gu++On9faqQ6sS~(bH_}Q4#gT>zpvJ
z!Y{6svNBm!lNqCm7@|1E`|=e37+6w_?-gV)yi5!r7O)dE9Z-y6<ZsK(U{|FI{y!CR
zJPHk?x(y^eXe&+O4w1tE9SA^VFjN$d^xd5vZBRj*E-E$Fk2C#$gF-JE8d)3rx0q{L
zRLq_C4*#iH2}XDQ*`F{W8qXzMdiJ#`gx;G#1yxGI1LwXYOqg+#RA5!Ws7IzmLDoFo
z17j4f)!X4`%?aVf`+_Hn#F?wc=bYVl<9B(AtOx?GR6H7k!K(&H5&R480$;rsz~M2R
zeH(EM0U^M>M=$bRj$SrX_2tm($FXnS@~o?Gdy^@3Jd}Oj%VRH4;6sL)V6rsXNniZ1
z3kLM)2IPZPvSHxGcmT}fBllsb!g5i>Uo(M+b1mjIhrecciFuwvf|**@ulfTR6d{3d
zp=}^3&<&T%{Nbgh8YSoX`nvHR>i9)Q=}zNok{>vJSrLrOUocw8zQeJE$rpEl_p0EO
zKzJ%h{=lh}f=Y>xgVO=b$&-Nbw2eaAK~$~~TwBl07?W?fbVPLlqjS_$Gg7J_V6>3w
zhUQHeyDHOfv2r%5<bOPwyzuFb_6ELFc7SBmM-$>~hs8D(2Rk_q-0T3gI2Ba%0(1y<
zn+8o;RyUSEZH5n6JvU))=f9yHqJp`*1&eoSzRo`}9kCq9EBYAvh(xkqc3}7F#kNN#
zsY{1ew}9*vWoB40WeioU++xX%6}<wN@LHi*eqWe?2ATuT^gKMH>utCIFr_<t_zya*
z(Zh8!nLF<-uE}^4gy+RH$~{9TRisH6PkuZ2hPaOUrL)X9DU5Kvhv7s=xleC!u?iHi
zzTxhL{j!6<u&fkee21gV>aH%?!ek7nfQlxvQkkl0tmG9rF{|bCmK9hqGxf)f#Y>9C
zi<ZnwimWkfzUc7(6n5U>T({x-m&z=&sYs!$Xc8GAiInxRM-*ixGs-F>Av2>030WD5
ztW-n;DTIc-C9`3L-+9;b{qOfZzQ^(W)A1bmeBR@}uj{;C=XuY@t(jG1syRHn{U*K5
zp7Ev7ta9&Ty4kxOe{46W5r1tp`7$-;gH6=$g*Y2NKDk+1OgHb|-3ql>rDE}3fkq|;
z-3KtS(LWK>ny9*sOv>R+gW|1%hJ%U(UU73@MI9ca<|kjOk?g#=)yl5ZDgyVy|H*yx
zvrd|NQTFwU@b)FF(;MqIOz1ZxCco5v{X+?Z++FzZ8~>HimD#D(E*GA*Of*-O#?$`J
z=}BC<_;ydrpjy(x=lkVbvSp}pY6Io?Qbglo_KzHN8$A&kN*=MbUrz7{_->fEZ*|yL
z$9LYFtz)1~km7ny`8Vea(>*Lm%Yjr`>Y9^r97O~MQNzvy?kbsvHr&f|@4bot0wX+_
z>^ZQKBG5(!W8OL;ndHbRk+@erO?kd=3pAq#qbzupGmi=-st6SdDerpSNt5o$uPmXI
zdHK@zM@rArwK|#q47kf?wk(+b&XP}V39Yf_6KxB6r+*;JQSy<5xotx)YwJMDZhB>!
zga?)vRPr3w^R3I<Qra4=VsyH!>t0LAc-zJt*^!{!U$Mrc2Na841P2VXrsu@>XEH9=
z6KSv0EZ(~=(2%~mt1AhgRRd1xEQ-bkN%x<_a@c_LeyGTm98IGRQSql)|L481ylG(G
z(`#ZZZ8Vah_ZzPHf0W3liqsXNwJh>vqmEwWWvg{gaEy4gk&ickw~s&Z*)4Nn`18ap
zC=^o{O;e{0+OGv{PHrgwp;>gd0~XS2H5!z?uAjeXI$VCD85c5Y6Eu;MVOu`TG*_??
zA(x|*+?=u3I+~g;MaNezT}061LrD;K(6(}C!!eaBHHxf~Az79NY9Tr)f*dX}PoGIS
z7(1O*Vaa`&v#JS(Hj)njN|Y!Cot>Rnq+%E3L%5A9p9>alSv}<&{N<fD?cpK0U+(u>
z)IVw)-ftUj?7H9eoioy}V|#gP?DtcdmKT)u&pNK_a-HkT{d(#lQ)$M{uYpAw$K^|7
zjAjN5$MnXUv`<uCU>Ycm6fS2gC@U!}w+zW}%Nefys&ZzpWpLJ{Nc?Xhk-L*kLjLSE
z&At!rFQ{HwD;c7BELS7*Lq@cDM(Ov34X?S|z1l_kj4{MUDSIC#1<>J>dESe`B#r9q
zn7{wwn*uCxYkf+t)jCuh_P$!WxP2ef_!XLOGi7|c2aBu^8c5$>a;0-&f2+7Gz|{fY
zs|s&TTTS2_Th|TRLjpWm^5yJx7RB$x&TBYoj@JnkZmd|+jC35+DibbUF<dBklf$Cg
z{6ps6bsZvHikk~#pS9YwC^q}6(9)c39P1UV=v9&Kv)~bI%XpY^z&J-lvg!M|=orSd
zLF8g#3V|`D6sIIVEo;TERzu{!1+CiAZXkH@<4EoSTh%XlwyH)4H;AcfvMKw@h!?$&
z*p(5ql%l!hOJi}RwtVXov-Pp=8}pP(MHELVHECTnT37wW0<!I^Y@IYMD;|dT7VHT6
ze!svmUZL!}C!a2T=?@Nz!G`+{YYjpeQhqmGiB@*Cm=>^GP1(0sU?C}}Ez4O~KcLBm
zqMK`I)w14OE0-3UsC8TJU#tnVHbyL%pK0PL2WhL98z8YL3N;!s{4sO^P*dtoF38C9
zrD!Tpw*Fk&TF~&~nScHT8rwTeC%ODX^8yv}ROE6rlUv){x7|H8V%>ZtcHHUY!up)k
z`kawA>jui6NbLQ}dHX`_Z&{0tUsR-EAFLh>59i$+r1JJ!o#6MN?6H#X1_NcS8PQb%
zy&+F_CUB&f7#8+C=|9$f;<Syi|7Mz;FGV}k1PqdzVs5KYPR|>pWKrnrz*uu|sDK}U
zB?@{tT)eM=zhmI~01wM@2^Xwq(1$QZg`Q;O@(J=d%pQOG_|E}f*n0R<eGBt3mRuG+
zPRY<~tvpeWxt3;G97U^qn-!nuFdWh6DRl`^7*kZva9bRn4j!8>+NJb%NIzFc-OMRc
z!?Y%Hcl7OfA(JFQuUrL9wx+Z**XtNKJ_<S;9Q@L|@KG_$CfY@yLgx<K0V`3y2iYM{
z<6f^~>+QQUz72&h%47}D(N8XRg%%eV+X!YKAYdbA+u#ArKx>de7g6H9{<xuc!+O&f
zt!vT)Imq_wlXotY6s(zZ7(zfe8<iEa6Nc@f<mKkY?Mqg4HuoNeJ-p0V7WVvGrbE?O
z_AG5lSh9zg$e)WVKMLhpd}rGHxUcEUtr`)OKK!F<ih&vVhkDk9#RavcZe_cCj#mHF
zmP^X7#>0(x3VJo--ff?1e0fsvA~h~6Go~e}AQDJ{VA;3>Lx~&3#WOIT4h_9VmW4dV
z>bSsE1W-Va?Tf-<zw@v<2105WK0$MrN0gIJHn{6PG^Mtm@~?rJ1`#5o%jt^Q3DN&B
zv_a_{duUvz<fZln`o(9yrxU5QSe)Hj;Ax50n_3rJ=C=3S&PemLKTqApMy6d-s6^j~
zJm@&}cV{8-fJDfp%6mG^0Pz=K!%|sZ%ZM`=vE;T<V=ZIwdx**EgPsMci_rYgF}Cya
zrdHdKQ8{)~0E5Dh=->RC$pZ(n6$_3$lbmW`9Q7*#Mu_!By92JkZ&4zGYS6@R$}0vh
z5$3CU5E-D%kuiK!6y<@JP>t*`hDwpCQ2@M7ef(ylw%_Q&Yxg5Zs7QBk?`P|CEsivL
z@q0ZA>bF;Co9IR!@AD!K5<xyR?szmG=?N1X4x9@>(TzMG<)+Wxc!hBmrfoo<;n=tZ
z?uW!lK05o_4GArIbMvDiB6dTaVm|Q(QjCMDvx|y*X5|v>{{B)<6vrlc>5v{FBlIE!
z2o9ah^bjZwq~AtE1`iA*&&!~>Nkk6;$PsvzhUo(6Upn2krS<v!FM7@Pn95YQkHHR+
ztckvW5b3b4Hp28na#8Y4SD@Y&Y1v$Xt)Cc4a;*gQbqua~<J0AS^|5cxpx93wHf^|m
zI`R6h4GL}w%(Q%T(E(pwZi1ihf#A|dGHRy#-vV=s3=dbIrxG#jE0Hq0(Z9W&%{s*K
zhIUK2X%Bn2ODmoP`7{l%tPnVet<-c+KA=2wS?CinS)j~1d8|%el|X+*Ex_NmduW$g
z_xo<4B__tIUs0D=MMCZ;gevjKNG8-L$tJBA(t!ZPFE&XZ9VWms@5x-S>N%4}>oNFJ
zlq{B%Ph&wOrlze-0ojx%ZYf=|(VYzP{sZhxVM}yQ^Di!nzIeG`u}c^QoE`Ief^lKh
zzj-WWJK{(YH$kjmZ7Z~V)H9LFmEzQ$?uC4r$rrEhZQ!w%E8!?!wvd|b_H~}<<U}L_
zto^6G=J>?obEM9&uX=yOk~*y3XD&>@v#{*huK9EHRoW$sd7u)}+x+F#6yPzy>Nda{
z&+T?fb&K2V;ZJkYw)540X<v#+MgCwzLNKB@vs*}i4thzK>k+-r`Id18W{<4zIx`HK
zlJg8c-RN6KIx7xCw?mRdJf<dWTebwBN*6*@VdalFsr!Xea}%ScYp!1RB=!l)CahlM
zqOr64B1;z&RUSC$(b{wejhBMjF5)oxC~lJqi|pfDaIX0=RE`}CM=l*t-9D0-6(Zhw
z4?=)=A-ge)pg+@D?r{gb1Diy~fB)lr6OIsGb61i_6;LM2FNFeNebXdWYmJed^#^3e
zR!sL{K$AO?=402Fd}-tCWtP-!Gr(E99R><$h&%#S6qPif91YB|r*DkKf!Dz_n;0qh
zm~OH8#jB+Fn<NSB;%L}L=j-MFqKVTOQ?L|A0lSO%b3i?)@OrUVJabkO-hlK+u(xCy
z3QcJx@gD&zBZM6$x$Np+o~5UzhJz}^=nMX#t+HNs0Tl#F@2bj4+h3{Bbyatd@D+|}
zEoodex0_0?O2|>gj85Xp&kUQ6ek|TY;E0Ihj$ru!kem?Z2Oje!Mrb7463%ZblC+z4
zBDSnSE8s)`<+H7r!@Vebr`y!s5!+sv`SZQMcJjMB+qmjLQwqthB9X+Hl%RK8jWtO{
zyA}(x5@aC#BJ1eUXNdHRCum%X@~xc@O)BC@P}<}3YH&mwz5j<2ld)H&DSxB85c0wV
zv>up>#eaTumbSuuMkFkZpN?-%v6APpnVszH;_1l`Yd%ET_3x%DW-M0X_Wm<Pi-EKm
zQfD#MC!xxqBJGByd6gX#mV4|E83j+G&V)a}WOJdxyR%`^G%vv6guXR*l-{s;^IPa)
zb4Zp2GG2VUm(4CQ{EJH6t^Hd}2R4N#kOj;x(@{d*diX;!P&lXqB-#kGq4MMzfDXB=
z!H|)meA03Y2#my?52j>l`eWxQZ#5ComD<4?ph+)oOxZsy&iGr@Pxh6&B!^r{T6<!Z
zbyF(+^Tpp=p!f75ZD0_e0Z3O4C9lXJhq#x3H{6Pi4ZsuLsI+YW?nw*m!H`SfxahIF
z;bZTleUaJoxkdvkAcu0Lzk327M(Vg$$Ff@@6B8TYt8kkeh&|!CBl%bg;i$+Y96@v-
zfUx%<1SJZL4jy)aT4(>-<jzI=<#l>XK?Q~tpScMVE}rm}z8_0T^|{9cxcBh%y!bud
zrl_6R9NW7xG<bka_1jtc*xS@dGY#76X+$0UHuIP3*20zqmRujeEHZ+VF@wj9Cjz4-
z!vmCc<EpM<l=*Xuzn?m)+Wmvi#Y4UZ4<k?lLo=Cfz%Ar2+IJsmWIR#w*u`TdhctC?
zqJOGqVkkmBIB>I}5%&>F@ZbFO2RP)8pE^6?MaVp8-?eJb`O2zYW*>H>_3N_O?Z;Aq
zA_k^}DqLUUXbii(DtWq@<K9fV3iaF^{iecIe{zK-Hu}{*gue`g1*kUYn^@FfPv9b4
zLm;G4l9HI{z#u<xcG5TCeAQE>p4jxBDO@#*k9|C*MxZT8Zrwk!+aPCJ2jDr#ak~`^
z@@a#d42#}=+ZL+w6nY{g?ui>OyZETn!&>q&OPld3p=W`^!oRoT6a4Ck&cH}GR?W{x
z{qvmewbx#L63mqa-j%-1(E=iJ8U5{9RX~VM!I%*a#|&Y81nIygdab^`p7;>#n76y!
zTedw^O4IsuZNYc!m6aFlcHakX7e``RAtPe&j+YIU20}1i{{4|{jjnRfCwhj`k4&uQ
zJWU(&NtBObf<ziDVU2Lwkql=oJ;*X=*ps=9PP}XQFeEg)xca!n8tI+#rVo)q;J44g
zou>!N7<pOm6<Q>L;Q{gIbU1%B!lJZVP7|y0M{lY`1De&DlZ&4A{jTHzvmnmqlP6t@
zPU>d20Cqb4YfacRuQD4_3{7jJ1i<Gq>(}pMj6QkvtefLz=v~19LPLRij?$db-0dvI
z;%t{DIupWV=5N4;5gr|_hzbsgodPX32tLFk!XqjghHAW=>AIx*Z(+E{kQ+Uj)#1jc
z+>7;%(6rcSf}CXiYTtIRHSY4gEryNJ-Y&bTQQTmw88?Lqj}GX4aJ$mDWJ*v_-MX*Q
zs&qt)V`IgrARlH|ndLuDq~eFPJ_}F`g-oUCo;_Agyo5oL+;%jj@dM^yn81E|?k$VB
zR031VMAZcrZA#Y6a7x;@1N&O7SH!72g=|)Cuf1%7hfpZNTutEL)YLryt>G#-i5!y2
zA3w^+ZD4rYglJ~0%g`;TNQ~hBH_4;y7ze;2@UJ9f_<`yjc{8)*(FHhr@=YtLF^Gq@
zxeI)>!we5Y)w>~GMTxVXL&f>yp2t5u6wY4=e&#4&GXP)vzf=bdGB7LoT;?JKXh@_(
zTsIrg)0_XIB*{iaNMuJW_%MgS!TT@A@fb+N49zqK?7HdTeff?_A&?i39I>O~f7wgc
zNVLfFC2PtBCr|cIgvZi#xdiakksE7t>nl$<8FIrF0RRzyD>tG#L_L#|>r{V`V=Zt3
z^<x~mhAV*^O8xq%rRymQRsKtnR)guMLMwKVLH1HU-5MGWDs1}8tioIjQ)&w7k4z`0
zoXvgIeekP6%}nB=JpxmHEQ||&`b$L;rFhRv5Go0P`mT&;9Ev7vNdy@H1Kkbm2PqfP
z5@LJjSp;|ciaIY-am|?s{7Cpk#XCMlEKuav;=2(p<zIexHzI8n^4em*75yK1h7rTV
zmT(XI)n=b!rn6a<Z4w~CEXWfPqaZi%d&0}7`w5)01jvPl;J*lU^L>ajASD`<Xy_*<
z#dq8OCvtEZmhxJT7zh^BSyeobO<%j_dr9%yYBD{X-uT>zK2dw{^jWY;k_i}d90%W1
zSy_pthJD>SzYe?ez14vh2F$FhdIUcjSRUM(IJ?%d;pI$6_ppkZ5?^~ytrRh_y@d{i
zu{80eqv%3i*NxHF8Q&!CsN#%6Zr@yy1o&rOxX?{U!q<3WY~&3rfnV7(w}{!A!o7&X
zI3GS#M3qaR&$NP#Ut4^J9!CoJChhZYnb~q^c1!*E>Oglq1opJZB_MhESb(O0=}u!l
z%YwoRt{VwIOo3&yx)^DeH+t+ay>#)D3_GvJW&U<a3$7g?Ngms@i9qp#xA3K{7|9s~
zm<x!NAO1BUkIR;B-*?R}*yzpJHBv9QS)4EDZa25!x(1*xAJQg#`5*~NTZs8B3me;x
z2f3?ybTl)Qf6=*}p*K&^FsR;kdSE2O@9>d3-c4=jfuO0Qs)Svic(TBjVfa<MUj~uu
zLS_Dr4R4bA%o_K9I78>`c2YQ>Q9Q&@!5=U;0g-RT#|P)-iK1u&*A|6Ia5Q0|+UZhc
zf6mfn73&&QCv{NIsQJHnkQl03vbMD7h73?paL3gkDPMuHnFVp9UvcFiIMctbX8vHr
zTJ9IB6w=FV`Z864y4OwtOE~@<m&c?W?mHs-FrL>yf+GC45Ie!t3hvp!wWxgOmB9kj
zpK0NKK6gn$BwFZDgTOh1BO^eFV7dvu7Imf3Lti?1#pK4y4xXO3HQ^4_&qjA{_~h@{
z2bbI6o%0CpSP7rR9)i^%YlRn>!mgb=A3W`KLciOCvn_gHEuq5dM;x?ZwfGGl&{h4S
zl%CKx&FkRw@>1Cx@#$S+s0ubXgQLuhjD3h*?uJ(ZeLgZD6XIK?+f4Uh+J$cG9!7E?
zMts3U;hX@h^CXj7d|-j0>cY!)R5NPVKRuc3+FMJ#uyPxEJT;}A6ivH}p8)>0yR;dn
zc`bphr~#Iv1i*`|rBvj6jJJ_|Shq-(4U+DV$Lt=l__fNBBmxPI!U3Dsn?RgN4&)Ie
zt>VSQenZgLw3hs^CM}zR>^-Ai!_#8bvh;!>={)R$cMQEUj)0yVY02IQVwJiZ<_eFQ
z4p-v#gWf=ZP%7f5C1DTnu3(CO)sZc^($v1=$=9*Je!dPk=iuO#J`LMFy0#Zd$qR=0
zR@4vOR@5uwDwX)FxZ5?$mJS%c0xJ-T2odn)RKz`oM-VPW#ut9{>?mANiAT8^qc=d|
zeHgk38C%2kF^S3-sub>(-{@uRCI6*6VIL#mU3gtz^w_r-yw_7r(<~K<<lSC3y*n&f
z7?W%iHR;XznP)`CmxSva(}E0hbleAjszx&$pgDfSH*J4+dI*a+OCe2op65Dehx!`t
zhTiw{g94$YvzFu2^N$g!N=%bzh1zpwrjEals9UBrk?}zZ-S&rZWMF_~f*KbqC@5S;
z$G*Zs*JCdhDl`?CFCVz@?ejY=TzZsC<g@??6hbv;p2H9afD+~LUU2<@LDObeBsY)f
z6`0dwyTM9NiIW{g1f0xVKRba*DT2}Gmmc=T%PnhgOS~|v7;lDk3|&Jq#C6mItvt`~
zcJRYJ0Tl`74eSQ4;IMck7v?`svf^UpOM@M1Anp%#2g~-7M}`wU5vv^MB(_6$LF2Om
zVkK4^z~V|NrH&%4=wwm(Mwd)XS3Au8z}3}Vl8e>&hfS;(e9-h?AW078d>M8fP=t9Z
zzs_HH{F%gG<Kdik<`*YH1_W#d7~w7%rCC>7dj@k+I4digliR~_EpXUOypY$D_Nrp3
zYsBx4_2{wdgb4v3iNOjEoi`}*ASsz+1um9MjTc0kism00Zwd4PAo`Hf(Pnfy>jtWI
zp^muN=8;sBiYtjseb*H+$Ql|N8rfr~H$}N@p$P;VSx{1(ue393XJ6*j%d$Rx{20ZX
zVtO;?<1j%|VSVpLvu9~(sozy5FKD8aVapjbcF8?Wh0ET``lmI|r=W-Xmh=yAWy1Pw
z{cCsLOM(6RF+v@piuQF&59<!J*u&|E6Gqf~q6Dr{a!&^f9ny~>xjN7pQ&7UU!+(+c
zn&cFrE_nla6(}cEiw@t2n5gjZ7g&9+!ux|3jnvO>Qdp0PHiRkFmM=e?_A5tL<1Hll
z6*wH@9-3r6t}grog$9WCsrvozDYq3GLV*YU%RM2#2ZoC>uEQ8eKL0b#NChrm&;Rz9
z0{H!&qkdRCfXp{OE?E$cBoG5+@pgsl$k3U%b7c87%p-bGqcC=oLI?=ou{1(b4Ae#g
zmf^ZGoXk-4#GW+T@8XrvieWK%`pwb`uoASpeSmqHkLqa-Z1&cnCS@9Vb4m0o>1{Ex
z%6$#}%@Q6a&-tG{q$LEB4QC7+j@&>uuLdG*a%##9o4GFK3h{FhI08heJAbXvBD4##
zwILs18sxF#&tZ+Wvjj_m)a(51V<0V;8&ygp3|*q6)&TWF3OB4q<mG8Qi`9cjYUE&t
z|7v0k28Voh?@(q#9f8S$Bho|I!Ul$h$k{Q?5PRUeMEZ6n9D-#qZ35zRQH<7uTLOqK
ziD!hH=$ZUM?`VrVXe0AUlE~6>7X)5>(H?-AzmGWUveBn*y?_OSK)@ghKES<9a#-5V
zq7P2NzDERV<OO5KDz#~Z4;u~I`+Z<;z-(6onZVxaApOj)LlW`}V*GQ}*X!fBCgPzW
zUl$+f5pqb{&Z6bsfyo2T1$m6*JO=g|SPT&^{Byh`o>+K52@7vvq=Hif--1K|!(WmK
zW#3<~4*A{AV_%}Nq8d#NxgmNRk)5~$Wc1sMy=(osFhv*!$k=`7t&yPDB=a6k5joEX
z1^{{9Z)Z^Dr{LoNKr))rL!uC##&dwg)}Kq=#dcgmP*~Vm`vIQ9Gh(MdKOD#1=EKWd
zRcL}{(_UL7v&FiN$y+_L4(=rdWLV&3B$jf$oVKoA_CnaZ+kS@S<y{Q)^MDHp+9WaE
z7YOgSX%%Kx)Yn7^`zURkPPWTfslbvX9$&mo4zrVHZLIX`Yg*WRQ5T1dBbGUmS35&`
z`*PGKD1OA#fC0@jtHp)dZw3xTSH--WN0}`%Gc)(esYG6UXpppE!B<?xd&jn6gU)?F
z*UqPxiMJQ=z?9o6;oaS)pxMd!LqkKDl%w3ji5tx$BnlHosC<jM4Tvmc22v)?x$j=J
zv4{O<=%gs%>=*S(!XMK8$Vn(z1cj<)4_t*g0k!0cG!Bg%o$@$86Zj7B0SQP${HbKH
zd+E}X^DCd-#1~ym*xmV^c;n$6rotzNS>DkqOv7XijY8mWr5bNpG)9;+?Z78NaJl=w
z>g;`+*5p3lVlOZb4^0~f$!5Z53ETB@)zS4jrIHA!AQuFS@E{2df*%DBPSbzXBQ*0(
zVz!&m@!^>o$l9{~;z9iwV$?tzil>xJKtWc$M8^!N`{w9ncmcWao6dnP<GT4f5*p4v
zG?8RBhCg{{p9KSuQ<6LcNQ_w6NCpYqNSBdjVfDEnEnVT2UGWbvHQ3eh%`n=a6%`eQ
z{PPM?BW?rHx^zrTI5|WU?pOp1N-YInM`%839uKHPSS^2`b4964wml3rg<%#Tj6R4P
zlJE|xY_QQSm=n?>Yyu>l7Wv!_SRUTP*7pvYP902al~^@N@(N^6a_kGPms-G@7f|4C
zOS_99mC+o_DLz#I!N_3VVMS#23w*p;eB|2_PUkLr*=W2nWF_HNXvv#^DWhX1Wf=TH
zIEf$t%Y|9-SBMtaPg!xW*NFWJi#Ja%d=<k3N|Y>E_<sJxBA5#kXuw<ZEcFFvgv!0e
zn`hPfa5G%l1fxW01l10P3^I5liBseagbANCG!qkQB!(Vl4P@hh#gBwcg5L!K0*U4K
z<!!lc@dlzL2!|XeTnDcSICsAm>oiDec+c>d5oZBdx<sZnKIoRXdu0Ywq-?CNK1yyC
zj1TzTmx?s7%xMQ&A5v3603L8nKtgTr15u*mMvrks;4GDhyLVNgg4g{p3`T+J!yTo3
zhwX;m%Kyy)$G=)XhPcadQUcs(SX{A86PE<>6fG{1f0$IvXARvZlM?KCrHj(5lAi{e
z==rEK;F}kBUI|7OO(<|wS%Eii_M@i|jkU&YZ*_nN0i*{U-OzpN6Q$!(`Dkan^Y`J#
z33E|Ca3SJHc#x|D<uVaZ&6<?PH`sznfsCFPI|ptU3?L+2j_*VaNnygd1gmkN4hfck
z2mT&5XJTQ`J?UB?jLAC3^);J)gzM@P4#z_GMCtJwo`VKR-|w((;}9lpUf$}L-fTYh
zfw!yJ{I;ljn0QPbCt%}0+iX30;Y|{byu@`k)8&00!#}duVe3QtycG*LK{^q-(hDm{
z-l#MjDmcq24TD7dIPs7j3*Gt_rznNJHLv=2Y|WW7_tRyba2rK$x`tdj{!VdikDL~3
zjqDb99KkLS<R0ThcpZtO2-Z2Qyc<a#5+VxVv1h$_af0pLG}cVOlz)N4WY>vK)g+k&
zJY4(C->~SO2n}iqP-O`U#BCNIo_^&M$!i=z5%F#xamHCh<irOJ4GqF&k_Vd}C&bY~
zb%tXTtywREXvnrf+6mmLsSgUwh}MFYwM;Qdokc}S)r`JsEsAPD73B}|VGzZA0KLk}
z!a{`rva3RJc;(Oji^3sCXOY+dyw0KyH2yGW<Bfy4haU9=wt8}Q9F~p;7{zUsaP-4#
ztt-kam%)Ej2*)U3)?M0=9>e=KS~Vy=IcP#Flfq}`>v%f?jmxitf-!gR)?*)Uz($BM
z7wQO<k1#Qj2?{<K3WQO^1v&r2fy2lmi@wug=bFIoD1p?s<S(g=#eDzc1%0;;4lO^!
cq2(3xO6#uvoZ)Z8$&pmbil>g|%A5N9KdiO#-v9sr

literal 29312
zcmZTwby!zh(?z;Ox}>C~RHURqL`u55K|s1iN)V9-5di@KkuCvgP`Z>B3F+=oI=(sg
z{qwueb5%rs=bXJ~&#YN%%?VdketaF58W#lx<+{9_j5-R+wOII{5*ri##?|nn5B&OA
zUPemO`@_~BFYg4cvu}G&o<r_C*D5!k;C02`!seAhe}s-fN+zZ1J*p!^dS0s<yM%%v
z_)$qsO3H+)2lJZL;Vaaf=cSu$%@XMXmBG3x?<$?e#}YZa6FhQ5rXH+Jt<Z!HuYIEF
z`Ntqdh@PpS{&q#G=S{RUA-Yrz|7&zcY)9P+>+3WAXT`a#gy=DZI5LD&&YQpQ5zq=R
zuH8ld&NPxvAW2woze+z!yRhB7yv%-#Cq6#@b#(Od;iMDRH|vUWj}2{gb#(>WnWh&a
z6nBi785x7&r$3I`)w-_{i+FB+o0-WF*=ZoD(my{UEfCnlAww6;AGK3X7wtZp35cn#
z7R-2YNYxs4V`cC$z14DG^5%47>gOk^Uz}#d^{X9!%(q9JY}fD9Y<->lU2cK(B!w^Y
z<Hx|c*0A9sz5B$(#K%iXChz&}$#Eamb_ih1z54GNEGx!sryCyJx9(#Q6ilkJANyiA
z%0c(yuiD|xj7opH7!|)ADMuo1V48^6P_3JdvG*JXAN6&V1lA{6Pf`W?QurhO{&hb+
z-0B_7e*AQzGM{ZGg~2d^5{>A;dnce5@B0>n=Chn^a}3MS-QV9i=e)aF(f`HqkMMTw
z3f=P0w1_{BHd8e&o0HWs&$nyCs%QLfMzl+-s}ui7;YVW-5_-S7YWp)?Y-7%g6?=?T
zpCNb_GgLZgcx`c5MIy(hKV9^o?|zX{{ksQl%cfQOYro5xXhl31NAfj94}V+mJsk>h
z{8LwM-j4e-L!zHVMO3O#yBHe}Zvj>dg_N8;{=)|)8JUo<nKTB&R5|3s3rFAlam+Pt
z^gX{k+p*=N{#9b6C?$1GAT{dbU?p3V!|3H!W@e_};kb>Zjm^T&PE6|@12c2Tt%tAf
znzvy$G&BhL>@%yWskODY`<!m!wD8|<Q*xJLZ-x~#ubzB5RzizM!NbhS8F9EZnUIpA
zaCLbxUhBp-S?z@NQK!_n(hja)?{#pt(`->aS-TJ|vN~0JALaScHf*2&t-E()GarV{
z@9gY+_hOwg@6jq&dqyHBEsc71xRtGutBj6`=_^blMHsTa`Ae^-&9S<|c97B4%}uY)
z{dPRFk1#WQ%Ro6%ks((lWBG^Rj3va8Zm~AX&oq(Fjnd|yl~0Gee*R?T=O=S=a-#D;
zHtVy#|KPy`x@dhu^toReOT@34<wuGQlKyRsSybtlk2^H{`(?mqH-ZiSeEP+ak%DJ^
zVg;Tu92;ki#KIQ4fsfa!>_+Fq==~)#Rb$`24e7R=&kUnu6cI^<h{+z?xR<QNWmHGn
zXdIJ~k+HGo<KyEn{skkVeg3=GE$m>`s@tTv6V*;T+O^(r(}9H>3Yn^^y1K8GbKQ(|
zI1+`@lwV!H*)iStd~G=QhQiX)tGBQnA|f~I|NMkVQXKyL<mpt+J<s|-OgoR@To(-5
z@)emh(}lHuhq&y$5Y^%hT6GtGtH44X^=fw3bbo)pOqYw4lyufvXkg799Se(ujLf6m
z;^N|hO4ywprMtWP@9N-7?PC2utJMr}e>t}kd(<VGdb4|2Vvmo#$$1RBg3z$uKX7A(
zgPIFRMCpT2TyQX|@88v+m>8Ah^j_<V@DeU5!pp?rwTApo4KmzzGCVi^V(m!TY=iy{
zPgK1|Uw$@~^iKg-mwdJZ-~QybPEMwNeKv_4KN<oo+3dKzh0Y3#P6Bsf+GmqrJ<j;i
zr3fXJY4r5;;J&9uIO#>wl>PktIQ(1-CUHK&Eqwa)>8{(d68v2a5>69U_W40u8$>iT
zeSf@WUbOFtY}d5@QN@rVRJhH>#VjF_vAgSBUGG-F0q5+8H50@F?2d71Y1xXWNRW)I
zEU(=NyRiG}H;9$@d{2pgqzbOUQL!2LC`Cq;&44YP!@k`yXjhd^kfWZj$??}$OuMkk
zjwSZu?C7AK!o;*<d=>Vp!um&xT@|lY59#T@4V}#IxWNpXd@3GBRyH;XadA?A>O71(
zd5jxg%Vysa6`f+bXLP5UM9$ip_2`cmJ$&<FzNUG(?top@`OMYD%#(~!xc#1ueu*o~
zf9s=u$8%vN7BhYa*KggrC3>=$Ko?oGIi<|d8BLoF8KdN~)#qS&Fkgesv3^US`XveE
z_~gxrst8E$QDtR!TUuL<+@^jN={?lcB*lDrtxDf$y$L?nV)k2*E*H6Wk?!{##iUip
zG)8sqb5pf$n}6z)A|vtF$4XT^Jq4^Ql1D}~AuMI(<m7Ui=&Oml_Lq83mQ$U6H$E5S
zxBmsF(rvkq3UYN5F0Klks}7+IwQnUA$$4+LJ)Zt^AbwRU<<`CjiS@H`T8~@5SeSY~
zv!EbFAdBL=GD!BUPf`{qG?pRBa~U_hhj80(zPgBkAj7;tr=+V(dGFpmzkj1TA+KIR
zcI%sTYW{K4sX+?9p(1g?SK53j47u!R`gx4kv`=U8(Z17^0z(uvf0ozIG!k&5HWVZz
z50#XZ9<{Mz2RpaY$oE;Dma^To?J<d3a+ls%4J>3Ogwt;KtEjsp>gL*b1>5E2WszRR
ztIA3~@w3hQP#@x%GY4&dLskpBK`&c48vp+NYxR7YcxI?~G71W~oiX$)eaZKaPfv$3
zA7Y&z?Wn-UM%@%#{*Fb@%)$~%Vd6)W`s0<Xn-v#X&J+&wg9phFl5_#*E(A0}vngd>
z$Gc|1FJJEc`&V4XP?&fiVz!~Q8->23x=X<Mf)_nd7R&BsE)yAgW{CI?rq$tw{jWs7
z*4!K5nQ6rR-b1E<L`ggG^;v{*=?Y|tY)HF-qBak+2M0=wL|9eQBU|Srvj=w}J;uhz
zw|tUM@P=9id4Zi2cg*tJyq53>Wj07DRht@C9A2nMRkN##PfhJj7kh4qwj9swa<)~i
zqZ>FHEqlXQO7%T~-kxc9z2BE*G=<%m;0dElsXMQ?vF*}qEXdI_hjMaq?m111!f9@5
zY6=MnVUQ>87eC{=({#)zoJNqS5IXL#oc?^<3=(|9!B3IAQM)?NEro@SD7|KXv4fTV
zmBsFaJWURrFYMZW<7@8o&O#g)Vu{ZmOBaqRKwVJCAuiT7;1~`K4Xwot7Jb_D_O|=Z
z^hF(se)@RFX?DJQ-c6aBjGI<n1UG-<ez)cjCX+T}Wu^~akdRt_FyP%C{z=W`^=phY
zuM-m$!jO&l#6-YJ4Y%r3glPz%EGxp!L4t1q?6SG(z{tqh)z$UQlGp4mr7A}vIy!o6
zQc`C)3EO}zA5;wr2pfPi;o}ahLPC`Ag(e4CZEdZLjErgZq`Z>SC+i<6za0bsCFQAQ
zzbaI!^=344*%^Rvkdl-{LBqrkblsNI-F7;0Tlu;7%fQJ<t$<UW*s1yAP>Vvih-YY@
z%dcUL?|y*T&p>T{@4RQ+cIa=jd*ySkCDN~632-)h+SFV;m3x4@^7Y#w{p>g1+R4I0
z+JV3R*v}Z_!x^Sz;C7&{-SOx%jcwKm4h}A<X^x5MCmSu&G2khv!BZ^O)+p5GMyb^2
zhErS=TT=?b`AO;}CUV-?6BXdc1yX;E;^y~BrJ9pRV1%^>B_-*EN_eFaIyK!K4$?Zi
zv-hLqfa`Jr^<sgM>X&$mZzkDs4&5&q@hbZ3PuXUV`<;DhHx1CQ-BWGG;#<AW?}(PO
zV&ERkfX$M4W+0!b`f<><tG9QdqF>x%XSxyUHe8b4Z&xbk=cmeX2QJ;Rz>pAhVXti^
zKyIAp{O9=8ue@H*1^wdJ^<qwbB+`G-dWr7KhqDrQIN0{<Z|QFe{HN?0jsf~~A`j}Q
zl`lBhNOn|hhv?Tm!}7EllcBt^)HM43enZBI1b?D}Y$Rn+MuthUpp&Z*%?A;$=$BYG
z%z2XyJNp4!WdvNp#x|`ljM(ur$wr#?y}#!$)AZ+0weyo{VSY6IQPc6f!#=;e??Sjp
zTAL?+<UM*m?je9yulZezz%K2BT%A&2=M`2=QSkE1#b4d~X1nqfzeQ>H<2esKAh(w!
zOiEa3XlShcD!vJo4iLxdfdN%@4UK%Y?2cw}0l-s_BFM`DyaJ$kn5&$IKw!wSw;<5c
zlso^f4L|VSW!M-ieO#;!AkJqo?lCq#z9nEx_tk;#f2vhkeWTvcJ!x4RWd1;#JA7jI
zSd@u;`&Nl8N~?&>aIqS@ofbyva{iE{4BpAeNpox&#m{q0tfbZqyv$Zz+kfn~e+-6A
z2t#i6J62m%j2kX6T(wzO%sC4cEY<}(vc;Ad)n{pPtW8vLLWydy|7@uTK#lk55Hl|?
zZ`rs*peW}-c_$r`h$1QP1q1#gxp`9&aA{%6x3RILL{1$z)~f)R@>kr2(@MzI_T029
z3H}AwHX7dzFzPcOj6R4o@d>Q^K8?Ra`Ecj++$w*8f{&5i6l=zW7{>Faz-K$%MutYL
zk_l(CtkeCHgpLCBo?G((DCf4SXPTBT{7;rBqUpqX+QLbkR4uJmJW+4+#CZtOAjmEt
z;EFe?jm#+@fRa)7EfJB)Dq$;FIC(j_k1x&+bG}q|TbghDDn33wF0=kY4dr~LjvG>T
z@_ic!j^7nFD*2Nf@u;?X_XFR(L=#H>(3zFC*U)%KNq^y=(kVQ$ylp*)PR;rBo_u4*
z6ywu-*q%Q_`TKiua(dBg-zXA>aNLFugGig-*Z?3NRG4}RwFA;%R(AGywG%xgHmG#O
z)YJgmI~@cp%5`7I#IQ+7(5L>8g}XXA=$9DW_=lkJ#fzFWNgkuRw$1Shc|}El5U{H6
zg+1(&>&o5$&Kxt3Y$th&(r103mHG_xWzY@r<Jp3KG9seykyQ5~ogq+HxA>Q9L*d3S
z!&KPqBSD)*S`Xb58`K>qp>7+4pSzYT!uZeVjhaw5zU&rWYs{7EeQer`uPcDQ{x6W{
zPLI{MV1~D8EI0Nw*jL`Z?mf2;UhkogfeeU%M8K3#G@&(lm^)k!-6K*;jeXZxkgWk&
zDVHknO!tc=4j|fxrDNHU2>SXK=fdcbn(6+8IWj8B3ceLdFEPzTQu9ZdLGc||CVZsQ
z`@7d->tmr>_WsqCkC#(WV1xo&Vbf1cLGjqokcLyIq|K7oWqss!M>Os8KljMkvTrv2
z3~-N)qwp129GtY!J@3SrzC59%sK8AgD2$uLt2ENk)$n${_F)n~^>b@dBf~aP%>~tN
zVbg&N`rIqwCQR*UXD6k>wUB$044($c1m)vpl$8kpBK+MPPk;-5c{<Dr8Qyy#N)WO<
z=6#!=meBsWZ~l(UejLLnuO6Q3sh%x+9eH%mt%-KnD=I-p8mNs9lV6K{S!3a$;-HWC
zVO_s{A;CqCXX1atqF?;3*r1vi5bgVW7T#AEM|x$Z*Puk!m%LNbBAU-$sI6ZgX6#8?
z(V=0eAL!V)NHNbVDvvX-s+TODTq2$=u9(~PawUE$cHYw+348P_?~bTfEbh}F!>zJr
z38cu5nFK6_QW(#Eopb~|ezRhHle!enN6GeSbH9t5a2h%WMy;>I8#eo!_?Tn*A%ZEB
zILusJkvb+X5_g)f=!&%?nB}2cE>x;uxQ>Zd+H|^3)>$}js^;TvPt&o|>SQq$e-UZ$
zC4&mhfKXL$fw5gv-}aJWWC!PrqSu1?jN;+S%3_dblx~|EVTr;DQwE2;udnb;0S5+A
zQJO|$75Ji9ziK`Ri*7bi$9S%4%sAVi+R<;nn{CQt%m^V0&bI4sM6?Hw+O@B_*L@Zt
zjfn^}%+b=rczD&Zh5iN3Ez;W4dCvyDdA;&^n3Xm55!1we-~0>wQ%klOuYxBT)<nn8
zwe}reG-in2E?>JSv!^@fB|7AZ#hTdgD4iN3N1wx;o}Tu8{@e|M3BhXXBl)e+GCv*u
zd>sKZkhSrtc+s;5Uw2z(L3<6^WPP;oF(6NDZ0y<m+;nN^?3(>w(D^J9p+L*00{bVn
zpANA7IyIju6f~Jfk1~hXVwuAM|DM}FJ1G7n_5Magx@~W+K|in8!fo;2&hh@`YVJi8
zGM}1*2<d6B87GeL?nMQrJ4yFumU)mDd6TYB-M&%zxJ~<~J#Mfw1}q$O--@&mk&)<&
znJ7r{jHY=I+--@#ndKj;3o9#+-QD?604TFD9G!=RhF*I+p>ZMWCK9yz&sX__(09n7
z;rhiLKx$0bmxiP73ld#o8rN6fwx}Dix+vGwzVwMRiRr57enx<krvFGYFi1ZL(+=HR
zjdLKvSwbvGmiwBq{*hwvL+X}5mAP|*-{(AB4LUq%O>XfGkHZ<T=_shF6@V`)pWK2>
z+*PG&uBEk&4m5&`a2j%)fZCXYq#SWuyPYaQ>rq~5dghkhp7Uf1vt^20e|4;yDGPtU
zlJvXqkRoXTI>kBluQ;iH!qswH@-!?)@^OwvHAH^PZJ9E8JZ!)@8a5j!*c&Psc<Vv#
z=Z@vQh$Xal|6Ew!SlxeD(A+_WumGi4ThYPc-x?LW7wQAoMQjWVjK$^UFQ(s7buivZ
zXPUx>mwZlk_D~27)|WCK!L2*BYw-+|u}crdOAa;c<5-k_6|OkWdT7!&ztSqtyrQeh
z!7!zwQj?KR=tO&IOdZ_lf&Sbchj$(O?=ND95u(lFhoxFsTs-#L>`QMF20F^hZk!*I
z|1FqmOu)$GP-iC<v@Frpf==B~gN8$)mP9UFHVVR{txaay|LmDgx1ODyU80kx7M5LN
zyj%cU+$H9On@=f9p;3#ao2qGf!dr`FL*HvQ!?<!?9E+9;;!sID%oECf)9P*#Tc?H?
z#qSQ<V++{r{VKd?w?rP5ltjV*BHgO%tB!mYdB7|2!0MrB<3gES?X0o9f&9eP5stsq
zEW%N^jr!uHf8#BsYOY1fGC^{O_FFUj$u5WDp07TcSf%QA^Lu4Br`A*8uLkSAoERAw
z8$*(~4LWGOrlg=?ak5Hhk{0WSYZtfmf4ns`Be}6Uv_Rq6Eg#gb+|<4;+|NIbY;C#|
zOTPot3?tje%1y(c1&(aIcBIb)ujmU$$$wHBWV`6Au`im~;7aQG>0MKJ^D_|_pPCHT
z7&S^5HHsRwp%R_43Y|P^{CIuT)<|AX{ORbvhg2Re^lou+alj*Jo?gzSe;TyCOC6P`
zNp{xSU;MV-c}|*~OZOXa5lG_ptWR59w50Rf{r={SN{zBz4ywdUOo>*Hw>TmTvz;{W
zj16B~bxA6{x~rru!kS83q2;*uz9?FKB~)9amQ`e&mGeQUz|R~C*Kcp1M0YDkcgv>h
zl_-wcYkEIjB9F*yt4-35%8-g`P9hJubG^woO7xe8dnyvSS_?TE76S#UoMh}|xNmXa
z{3K&fG~9ki65^^}IC^{%9{zP{P8)S?AQNqTe0-!(Cnh$Q5cnOB^^w|JXLxvIa+!g2
z8nda_R{bw;)P*I@I@L5uV|TB$-7*p%XJM;ly*VDL{RQ`#B?Y;WfL=F8S)9d}xOAr_
z^UpX5KZ3@Wzu6^~34c;w5eOp*MSZYW6Cd-z5?`OpI6!1s%YFQ#R^}c5$mM;jM(t23
zCa-?7na2i$y}NiV+%3iE>X6;iUmWWBAJ1VmT^t$!Gjc<8H&BjYG){+MN`jvJZ|HR=
z>C)SV{qY0}p{O|uQr($E5-(q+1dUC*{=W5~{5OjO!HK1pP`Ui06@mvr?~B@vTnTi>
z43WsFT79oacY|a5OT`2Wjj~~4kTRQ4T&ZGvXIv7ut8iMLmI8jtyq)05RDIt5k36k8
zn}IGPH(Bl`-zVWRpEwfR(C4%@ZbjU7oNh=Cqw{_J;RDTQrQ{Z<zDB9_c%yQ1OsiM-
z{6q|me^f6GC<MqSss|-$V`y$tl>NA6R<5Y)kQ9OQ1`m%*0Bf1Jnv9%`g;aNwRG)@a
zpO#d2>(g?|?aGPnd*j_4mGSmpUk*cX%VfuQW>vEM{HL!q_<=Ucep|WQM-=vXqntVs
z)nl-B^%Y40Cs026`qVch&hHx4dt%Jaet!v+ZQWU=mHIheMV8{Q<J0TD(qrMf&Ec-n
z`D$Vjw-)a_WtpyflV23~L3xuPoTeh}gLJCsp~aU34O^Ca)$RD+d*faFlX5)*ZoDiR
zq7`X=5l_SgMNoEs6~bZP9n6QrzBib^wBgI|k9qwP^SZw=^?76EG1bU1)tk=OBe0Bd
z4j~-vTUSPKP5@j50g00e#K-07#vN$JfNzSVdGKZEoo2OuF+xqzBV_dS^PS?iV>MPD
zlvuaJ=FZ}FkM(!Mcim;^ay8ip3baDJ3$gcp867mm4~G}$;r%ooxo%UXxcL5=C7EtF
zS7p!2qon&#?LH;n)1&=;DS&C|A@_Ten4J{Pu4_(e(r3+H)^B;O{U19X2h?u-s6YOs
z6Y+^jlNuHeXnMoK6%JC|I<KARafgn{$ukrB4QHVm=Q&hhT1QGvX0)>XF3)|2hKGT@
zUz7N0Hv=3MLD<5pof8*DRra}Us@;DVWxh1{m->C0%}7|L>AO6tbtN6UyeT~Q#!fIn
zOMEE>?v$J*Eo*rEFX~~NO`Mi(JL?0rT8@!Lw@~#ahk-n7K0BPlUqrSe#F~7Ad|m8G
zMmIg|eEehdLzOi&U5nMUre$qgp0sYE4^=+ZV3s%UeaB51B?|<Z`1!>}@891=MO@_E
z`fUYTh41g%5Tprvm;;puT|gL6wm_T==c=qgY#_ug=r6L_nU7=W5I(P_M#$CGRpD8<
zLA<O+H|O+pAO7W|m+LQPkE%0QNL3!Y4{HP#zKPuU5Mf~J5o?fjlN(ZMCv*R`9-e|%
zy*!mJ9Q+xLd;3dV3D*-wgS2zn6m>N5KXMjWa&mC&-tQ_-lPv8GMQ?X|;Rg_f1A`-n
z7><QPKB4>y`GP0v3A9$ne}`4prW-}_^YcZ|HY*Sc%Br9Z7dp?OPx6+QmNmaBOaRo!
zh#gWk`u;O-drgF3wco#Wfk}D8T5+45iwkHo|BHV*fZ(ZwoIl!|Tgb{{!uqsYSL}gC
zFx*(LTh}lpO)cGATek0lZI4-g!$+8Lj-He)-MQI5$%G=wU^1eo_=$E{p;dy`NYy(v
z2CphBGJW~}YG?5Zl!SJv%pYo+mZrG_%aZx3U2Z=3`aV+~A6<TI>>Ulw=8nzU$F3Y7
zQ!l#EWow6)+w!Alp1PfxxBwdkfj;!_k(gNvI`BXSpV6VgRpUs^(I<qwTe1`m42eAO
zPg|4K&E~k5Avh$4Xov0HXp5mV78VwOUucBgvBO&DwhbDp_#1B4HJ~>)piBMx*zqhq
zTF$w3rnk7!!%o_-MC`N0odR1<y;|#XdCSxy%SD53u2>56A6?b)YNm4zzT!|!1i}_M
z(%vz%5i2v}d=OIOAT1$}F7*D!z((Cn%$BVjrJOBeGSc{YEKk8un{h%keaRCwJR-tk
zZ@ztTWd&iI&CTaMoE{o9JYot|$Z_4Ov=T@yt^MK@LC#$wWX(stHj+<*GWqoxXH{;Q
zsxobdR1eXyKdI<fOuu&N<Z9w7g>e0xHt#J_x6pb1$$f6m{U4sV5gl&0&-SO%3@f4x
zH3qvKn8dToM#`mA(J+!Y*ZDX?H@?)Zeo-!NGsykkUnoRHrc?S<!AQ_{G%pc%gagCQ
zN{y;YohOI9r+*9Gb*J%Yerc&anpWg4C4R$R$IXziu%xx}cxGsX5Iz++Z5AG$Xb5-h
z62ov%SU~-Xq!sx(RUZE9F@{<}W<S2}?;HN7H}HvHuD3LE8e=dVn-u)IR8Bv>ty8k3
zx4D?S=(=i{VASOJwU#*6kb}E_uDh#R-L^aPn?7G@?&_3z(m)>RBgKV0o;$1sT1M$Y
z`tC(Fg~c@^6P>E_UyP%Sv*YndWm>WF0?|?TOZ|b`13AN9@8}B<M}8+uNw!t`UmgD_
zI6L39=!l%VIGVBf9*hxrQ*gV^;J!4=J@jvZ{ibFnJ`W4(jgQ=ATAnCmqF~_G)+-$B
zSPi$+#Ea^5@^o8`9A=SS@AJg+VHJ76T4+zF5-p!jsV^GFcDpmOowf9psmEJG94XgC
zOOX%Cf9=S`efgVwulr-duiQ<T5B9ni#9Xs(8*6udR3jr}?^4_$EyePxa25$VheOI`
z()0n2Q^wo3q^wHszX6@yQ%V2&yXowxj!EOZw~JnxYI}%4us?lE4-w%Ir8(AfVMZzE
zIL_~yhBhh&xAzM$qI*|Zw7J>2IC6Q?CG|yp_!__Uj8~p#;R=a(%1f&?*m(@LWW0%I
z`SplXFX@Aj{+IeE;;c@MA361xBoC^!Zd7VnKC-VS=^C%Fp?cuT4AR22#S2MbxY}Q-
z3l_e5F&j-Wx3WiQNRid{vCjE`ecJOIEqT3YhlZ&OY4Mn+4|g@HqS}$cV2x_}xUznF
zXvI^E{a<+2A}7{2Psl}-*>Z5P{<bIaKS~;PJ(^5NvmX{}sSdtG9m@rRou5C^B;X<u
z(yVoV`VSz;ySutfSAM2NQu7xuT96h{k174joh{G!&ek_fgd?4psZluKQ;2<fR67zV
z8l=f#+&!@-!@*I(Q`Tj`L`p%mNd7EQb>B6pJ=$o0N}P2`nSnea6x!$opl~2_1NqXr
z<~{`dzPeRa%M-br+bQ|e&9<s9i1ON-YZ=~#p&F9hQ{<TDE8*buLe1B!U`=3CVFh8T
zuTKf6chIAA)Wb_lZBjH^u3O><V>Jb09DVzoZ?0-Y=qTkF6I5hS9gTF|Rr*g{U3tU9
z!_W77??C&gCF3WpMLtBheSy++^P1U{%-Y}K>V%FgK+s)7!@O(y4FzS>Qv}3;oR1&B
zZG8)|%Ni6LSZgjjY$nvD$kK<rDn;l>2z3?IY4}uNM3q=cpw=|z3+7UkWE}+D+7de7
zB^5noI)1Y>gE6N)>xqg`R5LG*X!VQ9@7%eAs6sC>uX8X8WY(z`%%zX-Gq+Y<g-YQ`
zX&vI^fCQ!j_X1k%Fz6!Z=jXcBj<lfWn8!pL#C<nO&%YcM+gJMMsH03shCa7E4`N}t
zIR*&*-z|B|p$!9t&->!vSUbaMF?(gXXx?QjyN`(By*wUlLH_EM*4)pZ)m2r$!993y
zRgL8tRP(??kP)S-((>0*2us|sHwkRJNRO5!3{hD0M6KN@sedmKAPzfM=dlr=k)a|f
zDfy_)^9coyq3QNit;3(Xsv#4_8s)U~u)05s$?GTi|NGf_PYP{FmcSCz0U-b<A7K{Y
zVK^AS23S^4>NWX25cB;@Xk?`or;c{i7yRhL?8m|GEUM389Rs#tBCg)hJw85CSS%D&
zR8+&%hZm^;MiFfplYpixj_EN9FkhDYi`@uQn&E$n69_^XXz860!R_B)V%2)@G1~n6
zV0Ll(Z;b0Ek3Eev%Pwy4+4A>Lqy3M_6=-ApTkZI$egCb8g1E@Sf)NNT1w>{q)F@rx
zjaK?I*zVs?0R7f&v5N?j?kgZoK!b%#fAKl6-aj}n>VH}>E);Mnir6@qm>z0|_^5lk
zt)9_s9upFj{B`Gpjl5f3Uj|5K{EL-ggLhoHMdv<HNP}Oj6m=pN^}!|ZIwFFJnYr!X
zqk!{c7i0euD{27;xz<<LA2`pVz^&JtJIZfOROvN%-wO%~${}~^1G*2WOkg-GfE;{<
zPc^@7Ksq(}iT=|ig}=`=>2)cY54lXl=$XMlaso$&EX4!3MQGDNte6ER0~FVua~kE`
z;Y3MhdFkx9vtEAc>jJ4;1u+O4F7*76%VXmgcNLyFD1}aq|Hx6`Ahg?6y*nz#AzO;x
zi}l6#r$lQFK(#3Q?>eY8e?mRL0fUK8DG4OWe>!PYyLoNNiYhy42|7Z?!HPKx(X@4T
zlF-t=80`eT4_L=AkW?hZOMjCzzJU`keI%6`l97(lIT}C;dPAdTm(pL5C{y6T^ICM^
zgM|j-)G?XPo;Ue;_T`P=U##$OaBx6{P50ZEpY~ay^*`TxY&TZ&1R&9$>uPzp_}84T
z{Gi$isDwEth|<O}U^}{s1Wmbg(jK3j?19B0OS{kq4$0op(Vbxn4`EthmaRd$7NQ{l
ztO=hz_wt4~0~=e|`}dTZ1*tY@rv}Ebw;~4HJw2j$Q9D6hzX+ZGX@6$G)=HhkS<KDw
zp9gV4boU#Ym7@=-21nvZiP71c1!x37%d5GWqDsc!NI!!4yBga+4IP&Zy{)UaWo!%g
z^4`b%iTW~}U<M$QDjol%fMjgl_dW=~E9?S3HD8{RmTn7oftk~Z=KE_aUef2<zYjti
zD$JbG^2Q0Mr?1CxOWFT*uMB25M}vf9X~_(l%SQkLeW`*m@(HYi<ra@Yk2WFN!x6k+
z%3wRCBR;Ma>q@D7iM!(UEvWtDzffVLkIFi)w^78+Gl-<}6o6<8xcEl_n1>OfAK0%0
zSS;fz0|mkod5b{mXNw7~xOyvd=bl1Mm;wzvI*y?NuVi0jYCrc8G0K^wH<2QonmxRd
z1mi{EHMCVgAdlLZu=lMszyDj(*yv;s#NoNXvGIRNOW&oKeLNPj8omEpUXE!%XZzGX
zn87fA3TF`j{rex*kl?`!V*>zFE<S`=9!ih`elQ+cry#0dL|q*Xx!iBO+<*iF_UBEp
zPNk!Bu3kwxvAIm-AqLX5Bv7y=^EF|I!(YF4<~Gbz&sX5)X2d2_$Vo7u8>Id2`04|?
zP~$b_;;<;EZ~Y`6tqp7k6t@ZR?AO;`W6XU9ZjG3V3g;Wh^@z^IZ#ROnnDYb_gOJiv
zr*;{3>?@h#u#dj``G0J?ny|aEK(H(OCl*_@|Lu9>wrVp4`WRYcb8a?yNa`yC3kI<4
zSi$i}bTZi}YP0G|#GUn?Pri?rT=ZM>XFt0U3J2?&8bux=@v3W{{#+`$k#^VrJ@4VB
zER|Gw%!NIgV({OZK(K>*3pxRGQ;1V=G<}i|2DW2(b!jhF*eJ7CF_!EG+ZC3LU~B)K
zV9T&o<^9EWSlz%z2JEr+Vsu{bT~nY_W0Z55fGZAEOyV)C7;h7Lt?H3ixsTpHi#aeS
z5$i_8v-L7t;``x^xug796Ey0hmza%9#Ab+3WwLOuM8E0{*f-Gjzib{X|5yN($Pw$%
z2^SaFFwr#F+34I(I^pu}t{de)5&UCPJFn2<`(|5s=2fOr>`c`wK_KaZuLk8o(Bk6a
z`2u+bS-WF3(Ec>gxe2F5-?Qj=orV><KX}%?)I3PLXjejA9ey#1kr@dQ2^U?ty4<-!
z)V!0E6A*fafP3N<`vrbFn|?ab5x~yEdF|RYl+CHygb0c|t}AI?mkWOJ;D$g&Lo3%W
zX7s_{7OBVK?f?GL#7CG|Do+4D7F6Dq#-nKq2r}@q3_*QU7eO~xG&ZISxIAG4=>+jv
z!M^VTM_2B;q>v);43Wl8R<mP-{EqM1kCi+E8Bx2Ciq{f*Zf*_|TYO5&8%TcZ?;~_^
zjoC?ZNjw7x{ymp8rB!*%ggyES)Bq-?mx$>Cik0~3D$_eIy%w>fDUHv`tB~IT2}i?<
zfb@?dae2(7BKD7zR@i+u*(Ti#P?$Ruf{_k&=pjGl<_=donLsZbGHU0+0)8q|#6dUi
zU=64W2;t#B)$CK7>M_#u8O~R1>Cr&ik^!Uy0oN8>P|puHzdKfw(9`#a;!%{?-zI_f
zYyA~(vJygU69Ur=oh5E)mZm+UsOSfQXA@=aYwA!Xwd@QgzkQQL5-#v)@yr(s>FdF>
zu8Yri30_X5iP45LXmSE-X=#xL0{G>iE!bNU5)(cCur#zQW}$$C4(w;wy1Kh-*V9*U
zLxaU%oDP8bjMH)0mapW+$jjhhju(<9dCOe2&ZZCC{j_<Gz_`%|n}mi%VA&B#g_vu=
zF*oo@o`jmuR`qk|#Fi(R6Z^Xp*v7v;yE!;GIOVgFhDh)q#9yR*`ouf|9K!bRH}&o4
zbM&+qYIJB%-D7f@h(hc-x6p}cX@3A?(+jRhaDsu?uAnalWE`Nm!T>^mb|~bwj2|O@
z$_{ogD1Upv>Jw42ze^?L2yey2de>ubbt3C4Of8k?*^?#-&_LGA>=xML&UagHlyEyV
zpDf=55(u{Wju%uM{tS@v_Mo{iaXAO2j>kziSk+S=i@`<W?}zy{IrL0@aF5{N;kCl1
zkdt>jydk#8@uzCs!P46L3oJbxT9f5?W2u_CLw_9p32ARPmB)<Oqv}9(cbDXV1)|*l
z!WSrJ@I`!l7)Ag^WP<-L54M)_p^wkLX-sw7ubN<0WwI~izo1Ad0KEXsBj|(aZEbDp
zsY-z2-A)DB<r0&4h|n_)r!LpXATVANF(C(vp|m44C1riWS%I+N6Ls-Y<4BS?`U{(f
z#gk%CQ8h&X(1Q~W>6?JVxch8^4}3Y25IrJZ+m_&uBcsY@&}`Tx-|f1wyJ4w&&fxxn
zYC?>qt^$2d_-%bmzW0zRU<_oL{^z>eZaQ-;r+BR@M)}b68{586*Rt2#Wallx!C^HV
z=_<AD3<F5TrY#{jh};13AwaOFBYA@!KhTImIOe`~-pax#@*!xdyLsAl!(->KyJoQJ
zQZ0Yai}^Mj#KMR`1caXhXHOp}xd2JX$TAqFDAkJ1j#X55cyz@{1yr!=+mWj!E^CMF
z2+sJmYN&`kYWBZC@i8j0$$0S&PUl;n2P9H?ynip9`?;V7UkVcOY=($eFvxN3@nGvL
zXybuLt9nFV%h=d>q(CbIKx;J`33?0&+H?kJU!`>iu%b)<xyyj5qBc<y&|X(3eSng=
zIU~dmVCh7Q&S*7&Xh5bDiZWH=;GY>Qv(ChX{3{LOpZV|3=Cx6&#QlUo?-o5>&Bnhe
z5P4hU^Ih=bfnSZ2Lp%x8D`UUCM<AMFbQw(i1a=IF(^#;v;0B0EN!o|dIkjA5w0vLM
z{eiZ>tP#wB;uswSEE{-k9s!2{*>$8y@4#fN3D{zAmhe$O$xmK~i><oX1x&pXfP@HX
z$9Zr55yWLbeZaYttt~qWxCpYK&TTi4Ngdlnok{Cz=$@f+t>QK2c5{<R)wFAM73G7Y
zZj2}R${JjI2si=z;Q<h$pj(~HN8I^c`IH2tc0epdBqRWTZ{AH-l8cv7Qo=7UFNc;b
zE-?{964Xtg@1?Q_ofo^}3Y*?3<Va@^YPK1WI(h_E&Bg7Qt(=wlqcUINcs1cu2dF$|
z#a>b;vL~C~eOe8&@bbw@#tJ}ZtXTjU=cgN`h%Ab9gvi~10g{5Ao*r}$<y?}N-x;t9
zwd;w4jY|;&0eG(lZQaEJ!2Nr6w1YJZ1`KZ>A2nTF-5iC8Dt&Ix`o@W8-daWr+U%sz
z=5Ub<_PUZOs8(S#KQ<lI{JksU?t>HgeYWVhSw6IV>zZ8^b*Qx3*4MY%`I_vBJ$4j=
zhBH?xQL?R<7iVBw&Ne9ZX(f?!c6QDWB?q}v5!L`oL#6##G$6uT^2AUDIRDjvy7i97
z=*_spXZYL7XOr)0-BuXEGA4Qux#17iL&%z@yveq|im0G*10#t}a`CfKJAR}E1uB}E
zkFRpVkE(Fg4rvjdRsWII)YME==o-sXWrzPv;TrgKqQ5VitEI2?nA-B=(P#_#k_#E!
zDu3rsg#jNwTBvj1c7TBt_fg^r{?$#m`=%FXUH$zl(IVS<0wx`HRVhr5qro;OeYE|j
z@X3_(#SAAT{jM&#`T6;&=SJ0ZM#|t0RAomoT)AQe-!z;x5|n^|0N4Q%RwWd$yNr1>
zfBg7S9VdsQ4vg_vC;b7h-@Ms(XCGtX=8j@Z_YQ`d<RkocLtR%F><qYf8g_$oAJ+1K
zAx}$G{m=RNK|eDa+d&(NiZ}eDeIRv^HV)hk8*8d125{a$R|cVwT5yx!$s09}8K+`A
z5*j?aDx@|*C)nN7V-EKYHPOg}$~#K-A-Fj6Em%k%d(3ZbMg4I+a;F^oF*#X3Q7NkM
zMZb7-zXFV?1y)lGUKm=PKNd-GF?<`Rht1wbur=wBE;*1V)pIvO+X!{K7d{>Fr-S;>
zQK(aD4L3rGk7?|)L<S9v(R6MpIneNX2M2|oXy<EdYk(@W>+fKOwMxm#zACgUtpjcI
z@87==H<JGT;sjx65)u+BVXj8FczA$wj?d1X!}BgLA1RF21z%o%yqPBHAaGM4^*V$x
zIAeNy@j|8dO062eO(va_PX_}G=u&yEJ&BTS=@p5aV25wmZNURhHy)16ySis5`%6F}
zzH?dVD1UM0dE=3{hrGOe!|{9s;-@_aGr!K%wzH7;T29h?5w8-gtopAIk9&&=D*Ho<
zcj8h~^w+;43F-M3Kh*Hni7GAos6y}}--CSB5ks${t{%TpGrVRAYk*wFCf(;1x4~EM
zyFEkyYnt(CgkqO^lYnT3*4fa7UetR4jp(uhlosQF3lBQq)muD<HLB}h!BDc_!&!RQ
z_pdFub<mVp6T-3@)WMXf@ma|X;4)x>(}JV%2A})T=ds?uU!iV6g!3u>N4i)lq_ex8
zo4Ux4feXv9dbGqS6^Dd%_qRpNJ^No=09i#TpFRa1NgRQTkVI^N_0aJnDmAQ)Ct*{5
zdzw{<OH4wN3>*o>AS&Xz-TjVLG6|f}PSY9B^_v2V;LE)O1o)BZr<|PJTm<u``z%u-
zXc7oUL#ukwU=S4pAn;wkV`t<}inI&En3}i*1l~<L)<+?i5q1<v^#RxO+GO<|m;;dj
zu>(ke=j=V4Ywp6}os9w0-F$?0_TXLTS;=Ydh3nDu62>|H4gkDhE@Ok5>8u_wf4Sd@
z>z6NII9?>P#Yu9+joK{@<tReq8U|VeNDB|1AMSq$0%cbn430n}0aVZ_7C%6T;Pdkn
zMO<EB<A8LWc<Oo#Z6Tr`!5J!WMO#O>XP9IGC-5+c+2Gh-Em%=eRt6qX43S_Uzb1h}
z{JGFXFAQL)n3^&`UjMs38Ua9UEU6eYb?~rTtqtY)!6UlE?>(fZz^hRUIw}ETDLFp3
zwzf8hv5uh}Jdgm}tt+%n@62z^0gng`KMp>A!~EA2e*2*UEzXvf7I2y7uyQ_6iHW)4
z;^G3993z+1AJ!IwS&xv435|KZv~{Y9;^=Wlojh4m1NjwG!RR_?k0}Z4BCi!L@R23F
z^h3>^#e=qnOF_V#022xm9{rG@ANZdOCGi;T>lOvrfFE4`NxtU06hHVQ;GnsjBx7^X
zbNmi|2x@({&+>jD^zGjxLz;S+BY_DVur9;B)c~SP<V0(pXsen(F=kQGD#xE%bB9)?
zGDhN(5=szE2)Mk=KK0Uxcs^R7MI+|>7A)n;8R8JC&pAB4e34^F>@vM;dY4kjW1SRo
zI}ir2aos<EVlG}DLl#XSVv-$z8G(LGgJ2jz>g)`I=mE36oRShV;CXQ8GRtSFWQc=;
z>I&|2Xq^f)#|*&X4(Iy!mM4&oz|6#K;0ZA>1VV<*gwYJ@C@nirFBuv1r%#{0S7M!C
zSTJ%U0`MjO<jE3rWRY}YDG<6)l;JN6r-%02O{}#G6Q`yO{a(Cq5u*9@`7;>MBmX#-
zez9t7FJD<&GAq{w^yRnLPEkVdop_<x)S%GuOdvJiq*=WBkLOLS8+6EE3=k&@7m*Ph
zMG9YC7G#@_=kdUC3t^ID((I2|p!?E9lGrl*;ukt&hO-~vXlQKoS<lZc=xZy7)W!Qh
zFEzJ*r6~-CH5^XZ_kDOk0)hkzqM{*=GW*iT<vH6s6_Z==odiJnX1vMo@0y`PQ^I=h
zUfk9BiUi`Mg=7VOJRmYifC)Tu?+QG(zX0K4Wgzn*LO+3yKQuHX<T8(rxV0hMKp#LQ
zkFZrda9^doEITkwz-{8UhXEE4WE2U+)(r5tBNGpZ_>X{66j<l>GP71unk_3pfqjsa
z!0I`a^#~7+Q|CKw1KXJ<F+fK^<qembNWi&-$%&ZB4Oy+X=jVRY0aq{H-7!&y6#CzQ
z6+oPPozirBFJ0hZpiA}!7dr4D?7hQpc&7^rY9s)xO~6h=8o(wXU<Fepv}{l!0E4ft
zu5$cS%oS<^aK<E0oSdAD_}S`oA(g@GM$bh77hUV#-rg-fKIFoip7r*wx-RWcz;Y%^
zxs=E(&nPZV2ZSvE(cA3o-=LHd5(c);&7Pkez*&YSOUtUGSMu@$sBQJ+0r<GMEtb5s
zr>l2hykmH+$Th&J{?GPlD)e;_&}!_Y;0%M0<mSE(ezoB>fGmFtELhT2&-H*<g!>0X
ztD3I~22dSAs36Su?~{N-^~>b|GS7xC#ZGbW2J(B2iwr6eOvdy)VkSf;`V9a7mmwVU
z0O^oPu5^Njggr&;O2UL8cH58`VF-bUjO-CaKDgFkw~MqnVNDPV7JuuMb%hJio+v4N
zPlNS~t%2!ArZ6Cl!gD_I_O4-@iDo$GjFUvqMEwtBCf~<`LJhIv4b%i!Q{mU?Oo-+W
z58PhCHsAN!HiRRUEabunY&={G=?vc9)(~=EB?KngX1c*g6xvJ1AZ)F4WIpXPYycwC
zfa@QbB?HQ183@adPBTrgBGx}Ygtwc|Ap=$r$uj|0Y4ESXJo*|o0~yf)(>Olm-9VrX
zz<|i9(##^oP6BuO{RuqX2O*bWaJ+)aE27(y5q28RF|^VP`ZHXB{lV%3*0>g6biLuM
z#WLfh{?K019wQ6+^4}QYr%^kgPR(Gr0CDMQa@?cDhgJFQL7zLkrtd%5+Z%zfDIGVR
zYJN!_?C;O7fLpxNsvT@x#K<HiDKW7$MCUQI(Nv<|cd4kT02o0hO&-zi(L8U}vI9;4
zfNbqx7z~e$T!gwc1!FOADVMnx4CGv!TUfMHzE)id=QWmcPEg1(2hf3de-J7e)*7i7
z0BvEx(l-1-S^M7K9|L{QJ(!C6>NN8czz?MH@1}Qa-u*CX@ODX@lrSVLkuB=r#13Yd
z(l6ZGhDrtn&;_V^l`uw_icu_hhXcM@qyyis+u-p(U1M_)NX>mn41u?#S&NR$WTt!1
z;~<XtZc9{Cu;z+h9xniy`3ABBB02xB62hJVJ0eC?#em&XoZF*iHd>^YfQSXqrYtxM
z0VJQzjuAHpwF8nVFWm05-@y|u@`%D$#)YGGbWQ@P4?Nc27M2)is%ox(-Un<q`oyKy
z8ypgb2XA|jp)_zm0Y?DpQ;xbk$W}nOL8!ttPoMz;Uo`j_5f&YS4ni~Lzxg|WCYTZJ
zEx-UnKq4MupvG@Bal+tW0gT9$dF>d(Q+<DN^halVCEpbX-TB2hCn+vS3i%CF!=F{z
zA;XLAwUIEhvI3YWR*$(3oI1=dVwyrUA;WXOiVbFJX2H{65CoR}Tk`P+{c0>EFjokC
z?;l4th&&+gAY5d?`UBG|P#VD~h1j>258(6{X>VMA1H*`!r+KgEnSk7aGt}MJ_X(`?
zj{kDVM123UBX$sA>kzI3{F%uxnYjy?6dF#@RX^FwJ3u<{l8<*Ez8cU(8n+2t6ApwL
z0-rbZpSty)+{i>*N&PmXfWX(Oew}JZBbc&%1vS;Wf{2n5>M1@z5g-7;9=hV#1#2F7
z&M68y(nn35lONy({iw+Stj@#YlA36?>WqBoZI#l569HZD0#kz6x?w`2&mu-l7D&6&
zvlDmWKU!lbK&HA1YQiF`LpA_z3aHNM5j;JGco~o|p$#V+($MS!FTnQH1B5$-t`F()
z`o)fD0H?t0tMlZuYOu$0Fe<}mC_v>BjTZ%XG;0D5G*@Hg4Gp3|-pOW{OPeU>Yiir}
zX1|~8_l9`R%F3d=Z;f=CB?}Omk8{*(&;G(Q3Fg2<mV+-CREas^!?oNDMA6xuYZar^
z8?Q}6glXW&Q#pOO3~OScRMqhlAO{0^HkyF*JuLV@x%kDDMZei^;DgMzUzhGa>u75m
zXnOG+vMFRum@7<#;2QiWg$BtH@$8F<rM-TQJ+`<3)CqE`@G&3$Z+s3g_|Gp^y+VM&
zpveIC_yC652BGmp$PzG}cX5>wXQ%}Z{n##&ofBW>*Ewv0+1vx@3kCp_!kq$=0XQt+
zIE75&C5!s-;)bF^>)IYcaRQ=}E#%SOq&sZTUP9P8LrMX>fk1eqQ3L*U0m1&jJU~On
z>$)hH?z?^)pwI))O-krkoi`fKZ&nk(Za4qmsL4vfZ8{N8E-(b;#m*KK@n3~bp{s64
z<zOTdjOaYb|Eu#r+fOgxWQq{m*W_gXiD)bvw$}J^!c)=e%{%xL`(BMAX=Te2wE6Vw
zZtc(7jZUFOn1ZtgCekvPnT}6LNQMz^SX?-iaF;L?fYdOk*3feGz~}`a)*Il4hqlOd
ze^DMrI^MwCI6{H|mkgsDsZikWL$3sD;sdQ)+2mFt@1a`JRR4#BgjU3W3L^o)GnagK
zh50++vujuEjzeCyq7sM*6`A&4%Qb4^xwx=I3wtPjb@q(lY-z!tpP?qm54Q*}$OC%6
z_qj903{%9Xe0aGBg<gqH7Lml@p3@QAURY=a#MT1m=sm3d|EL{ML}2ENi~p5FxvnkD
zpaX~5KIcqGN!bN^{Tqhlk=7ppSk@I(kOamS0IPtY42WV3mu~wf-<2THkWh61nSQCL
zfC1-xd#Q)=@rbtlIx?~dCIdh4T#azK2QvpPqSy?|yBfmHx8Bnr|3nwM_>i#qdi164
zTW^Pt{L;1-;F~ggvM94ai`k*7`i2ohDvuqUynnRn(m50$4Z)NI^WD3I;9YMw&xG?V
zy5GfwOoPDO22t*?JdnVkXq!Su;_$odTBx)Xq<a`nWVv%Ert$fa>)%y1#L7lP5N1DV
zi4?@|fc)Q0DA#^@3uAin@j#zE2Gt70s$%cLv5msEPalBU8)$s)<=A}T2`&DMzFJob
zJ!MqO522L{vA*cM{XOyzSFn5~b_;RLp4@7vopQZ*{PHK(TP@d6;HHsA&sjJv$Nxfy
zyl}J!Xllq=pP+t?<f*-cUJ&{ry#I~ib=dMDVi~-opd}O!6ciG0#DgXR*~<b@C3FfP
zhk;Fv86YC`Ybk<G(SR<XrXsWdAk_gPkOs%H(M&tU3{;Xa(O7W$2L}eCfGCWzG!<7i
zHNyw9%!~iVOzu>@74G7Wg}YFL&T(J7s*}Y!)!V66{?!cK!wU8reSTwS!A&?J0zs!N
z(FS<8Hn={eh5Su$m3~k~=NW*xOtTvi5|~v|8jA_`8)@obomff(BJ>->c2)n$Bjxir
zLpnDFOSB4qkwvt7!i3|SGYl*&w0-7HK)u&ymzEyES>a&Zv<J>TUn7?QPKa#2rtv+Z
z+`2o0S?U~Jebu~+2S%a+I%srH>7(?QPRPmhOS4LftdnBKl}x+C&MoC0pvnhGt=|)B
zHd5yJg}Qq^Y`~guTFEj0c<imgozwM;9cJucn?#WyAOjhhn4kkH0^JNEz+50zt#A}^
zi$Xbm6Inr;JKPP^Q6TAkg2m9O44xueF0{F1D%NY*5Z!qu)DHRJe5Ub9K8ay@#Y4`J
z4u*_|hl?dezrz+6mT9^aUrl|=+Dm-tylZgRsw>e#fuAyhj)aJ_+Oy%|cTPe+<MbCQ
z;6T80^m-QnT{Y5%LrEMi)VTwq&f>)NvM->>f!y&J$S~OL1(-Z5+#MUpdn@{@TwrSh
z^I#$-%*T0J*?4>Sbx?7k{Md&k+Yij2ZjEgDo_I?MTc!s^-ggz$oxlMpbzgmPaqi}|
zCN;t#Ic1*AC*K6sc^;4h4F3+b3K;|e{v8>ZhAyY8rw6$1<kzoX2fQePI07<b1(4U@
zotX?6<wLnVYQ7=@2?A8HhcI0R4#AwBL0|>JE;A4R49d&^GD{fM^5Xb3+|#daiL10&
zyyY`Ve#21iy|~Lnk=BYK%9l$TPijj^bJ0ETlUjD6{p$|9^-t)-A9SpD(o!j+pVe2C
zUvD!={4CtZqJcAh3Mv!yFwl)4M*_*~fH^@&XCJn}!^4B*v<O_sXa9>+D8^raKoM~G
zEeTQNy_BefIAfrETL7vA42S&K<LEXpiNIj^P;<alfZyet0`%d6#K<k}m_v_X?+Z|0
zg#y?TcAobA`EeyCukC}IGi9~3!x9t4x5`aloxHSN(m*OyPo;c;`#2zWrBbFcQ1(bt
z(;mCzj#oGih^&X3<E?<ns3iQ)`FE<#S}~Q=gxg`<$?<>ldbKd%W0aK2_f#4rnPh!>
z6c9xb7)ArknHFXP)DGDgIykn^5CIe3Hu4dCOH-bc&q_Mw<J&;32i~#Xya|Db;5MWK
zrWb_BmoTPNdq!^L7=dHD)^*r-u|;~v?|C41?hw^X;Vr3-)LlHyY^*15r_#tqafK!B
z=)7Q|4II8{){9?eusCU7753`Iq7R3$`S=T|V>Z@=aA;P{?@}V~m$-H7S}dEjV!RB@
z^#b&Uwg(JC0~`OKY$w35IK0CF=G_yzLH|c`7W`b+@S0(x;D2LP@SYM3JaXs5O+7#~
z1Ys{>y!hRN@W6&~c^9&zc5|0eODrgdX=yLoFN(DbH^i|BXg-36q-0Y1;X|a&Qpkyi
zDyh@;HHzhAb+lnB_g>mY*zWpzEB=S-G{G~$T(^r^+!tva`QDy>!)Hs5rHW^ZB_((w
zrdNXd?-hT1B^<_CcTe2m#2yV@jMaJPoTSfIy?NDP$-4)D115BL4-Y>>ngeMCd8r06
zeh(G;8q67TAUq<lJ&32U9r##CwoqFTcRNJU0=!3n&vA+pQUko;<2}$(;OlacINSg$
zmR_YT5wJ8!9|$ApACztK|84b315*2CKe*-&a-lk4RNnjdOHU1b8!zco-8d3^5)A?N
zFMVJ?DHR^Qv0Q(T_I>R(dUo@4@4(?8ezVxCyg)l)0oTS6Umqprx|in}Ja?#m?gWxF
ze=G4R+k1jt1WttRBOIq&D(<8q1?ycm%1)mi6%eK=d%HFo@>~hR><^8E{|9&##qif>
z1|a`^1ClKHo&_fOI}ry9Lh2$M4Mfl$bahD_a>x+Es2!rgZU3o9K#au2Uy=;JeH^GD
zh3Pm5M+KG}{jUzlKo!(SA9}*nA02=!g<x<O>OP!7SIA}u<2D)P0ayO;xK9B^)>;%6
zEvj)1<SnJ0E&b~JdzChuix$p@e>9iAOjKD&A`^x3a+348HXaA}@mRb{`HgP0{k6`e
z@&tI&43qAw+pQMx1$DKQ<X^3T9Gu`z8LvPrin?Z**qzZN7hD1YYY?@O*UJDIJq*|t
znBv`|qg?+NJ{`_b_Cc2f!$}?P>o;Ux20qW9Tkuf>2SW!T2F`RC5SugZTSB~a@J=3Y
zFj)b8eg~plF1`&hnDUYQS#ytn`?d{UMH0sx-q)uD4!75Xw#?$<=`eDZtFNTcP_Ku+
zwSiu~;ErjaoOAE_g?y1`xGYhVdjY*tKdadY?q6vQGFGN(><)&zp5Uo*Ug6yDJBg^D
zog&x0Drg|Ig7HTFcE)*8+G2(-Kc=g_-eyKiE2}s_d*ISR-Ubv#Ks*0aWaksekD^CY
zZb+vBK4OrV6hLGEBTg>776Rlzz+NDbfmAZQ2Ks*euR4%E62(2{GZ42~iD5UeVpM##
zg#Sr9JyW4Mqjpq~Ey+3+3Ho#>V90IzDbaT^ZpJRb$?@mn3D-N-PxeY>mM=Y1FwLg<
zqW-e9^h?&1rmLEWnTZRxoc}$h;!7uJb`9l=RVk)R96M3?&F$sJy^2>%A4agSC?6T@
zOak@+ng^~SKqMqXKn*wexFK&!0#^r2T5irXXXK{)zJV8ByoEM~h?X`=zxXo%0f<1L
z3)lV_l*C7ZMJVb5ZgTfk(UJ#0I2eFPj!#dI?zVLGFw^oJQ`1^0))}rcV>i0_3Z_<E
zg!wamunh^NyzMoiaJ#EVE5yteofT53eY>VnnJ{7dX>iib=R}%#z9xBI6eHWHk9EF(
z_F6g+VU>)KGH=Op9^OAZT!uD^#{*u+<fE50zlIH&4~C{q;XNj{d~o*PBTDJTo9en^
zz$YH-U5_|eI5nO4s!3i0&otK(io6&H*co_V60jSizW3VSEEbLp6lkLq+hNp1FC}kT
zlIqI!S3oLba*#;-xi;c=S@%Aw8vO6YZHk2*2BjSaxgB}+e&Ys(@iA-s`yOeyjk2K%
z`I5z&AJsV&4V(N5Ts5hYvC}Pr06#{rXSBMwUdqAmqtm062}QHK17Lu(1-KU+yuyjI
zYtRECs1{KmTIZarMBfM|5tsaYT-x!tl$EusOXNt<!pV;uv-r)X^vx3(0V_mH7>({`
zjkDl2`}zQuZ@<>4#lge$OZV6_0yd5u)JEBG<LnmjW~Xs3_3EFzmx3>}%tl}&G5?;O
z7G=cGMt6>e%Jr;-g4ItQFlVK9TU9c@otN`1%ubzrF{EyI%E3j4{R69qoA@dv@gOBw
z^$}H+EHaJ0cDJVp>}B8`1w@OyrvsWmMQ&SoPm?7goil$+HR{#BR>?Ied?JDjf;}i#
z{F&ugIy6@7iG@Kfazc&;2I>=Q5e)4U(HV2Leh0Q0bEX*wHkN*{otwD{l`9f^hMkMH
zC(){}EhvP3cpv$wm_cXAg(W?UTz$~xPBAq+8%#XqYG-P2E+u?!TOuA))XEDL2$sT|
z+w?uYc&0mmCM=5O(ki)ObaF-5!VRx_@OD5smF1+%E|G&S2hz;9|6Dcg(105iM*~69
z(9mXALiQ+M^`qwPysoIB#Wp504>{e<c)!}!x=|i8H~6z_AzL&1gKqi%YwWz^v3~oo
zuR=!2x+ofykxPik_$gUsl$DiDNXW_z4Uy5(#g$45Wp8C?C#z`OlHym&l_)|3C7$E#
zzF)8BpXZO~zk7&loZs{Oe2&liIL<D(26qlrCm+j9->Ofs+3-qS@}16^<uO{9>@%AB
zqWm{#|9eZP$Sl(;MbAWS^<Hg72hn@+osN5R`VQunonmMAUt-uj-LUqDO@w0lK|x6=
zxEy(lY+`Du!7tb?ZZ$g*T(YpRP{CLu*9Y@Y%K}P4-}viOcQcd!Im4gQzlZ~X2tNq&
zVj^N>W1|K-Knh=m$*{Qr3u_U_3d-+Ik7V_)710L!EdSHyUYSzu6XaBB-{9bJ+^L-3
zvGQ=aie7h5t81)A{U-OiO1?E8<rtN>Z@Z-CseYrJ)l+`ip1bU9?WbKH$!im8|Gl85
z1m_mawXkq4`|IC=*>O%-aZX6FxJ)!FbB>uKt7RgxG-g~s=LL^|z+Mm=F(o#pbhMUO
z5YIfQmG55^$tRMn1rot=C>Vrn!X$~nSYNiXYWnJWX7T#g^ZG{4YgiiP3?4%)Lz$ZG
zX8p)%gTWr{)STyE7gEKP!)|{6XiU2(H!vULDrnt7Eu|eRr+Jn0A?>GZ#BRUk${5QU
ziW}9ynqwrB)lu6&B$Xp_+S>HlTGMG2n&usy5{nG0f`m!Qkh(FYfZ0LD=lsmWU73fe
zW%GxX5S<12?iSY_n*ixEKwF<-S0R3C40J?#hY2(r!z?*?f}g`givM)xr`EjaYdze3
ztR>pgx%ZAq1U-M**=JMqmEmct+C3D&wJo^je2+^~-AKTJoeG?VuBqPq)-dl)=t`;%
zxC+`x-B;UM+EFA~pshtiQ~_TqUrxm2Hnp5j<`F|u?IosCcu2Q<GVTV}sAyi)JZOG5
zJ&99j<nzn~7Rtr^IWeo}c@_rq4@TVR%PVGP7zWaQ+|j3qOY?teZ6vn|od`-MOv)av
zw8;$>s#oVJJ8ArCT6AG*)qm{Rh9tCIyUjvmHWtq2oldEsT!b`sfrlcQ+nsX$QIa?B
z$IP-VH7u^}uJr8#Uy7OO))A>i)z5p*bW)u|^wJaS1TA!rpG{8a9y&&av$Ej##LS;$
z@kzXL!-|COLNzQ|(zS&za-RxTB`bDizNDAyNE^mm-qE>Zo*7s#Tio4df3nL=-rs9m
z{}wy?8@C6I3p1RY0^1013#f7G_R<1@ZJr+P=HygXz3unu$0haHk*X@^=K7~6R00>)
zgjz*BiSR$)WX(<=6Sq~etQ#5k+;M1wwKvzcHui#C7KW!OUC%+>)>izyC`S=j!2V#X
z<@eRZymeW4En~m=_A`FJ5?2+VYn|C!w5Px<(^Q&qoQ_A(G|(Zie#EW+55Gqr^}eU>
zn8AZgootKFm^Eq@97XxLhNU_oPMyau8jMj}_J}^p%ij#a6y;^;udy65Qe|42J+-vY
z(=%UBM2nH&&8~g5`B~r8@B7{N20onYNa8P8Hs5W!cutW&D8%CK`)RS*7x7}MR}zYs
z{*<>$Ifm$^B+&*w`c>q@_9>AX&G3!knfW987TaS=-QpMCo%$`JlwFPHYFzuCd8vuz
z6zm@o9ZqNyRb>?K&(hj>oBB+P>Z89d!N#bTw|(`E?jo5;Q!hiKC=Nv`#l~WcIajLo
z30>qoZ+3R>56OjpfH?z<CzT^cQ5dL~L5aqSZ%4~T4cw8_cpQ^O1b^N4cov0%!0Pp9
zHj8U!tDkI4ta>x_N6I^N1D9;<?j0h9c{z8!RpZT2IH=X{iXVTW@zrTlZ}^(O59qAX
z<6^gsf8!m@e!f-8YGS}}e`W>cb~pQD@8gMw%V{@1H+j<y1}B?7Ch}Q#^3J@!dSS-5
z=*D{;hw>4-Ec0|zDLWQ{;`baAZzaB5-TGkecO+~`RIGWeGX~-z^Y#tJfkC_S`7I*)
z;=JHH!km)?PGJ%YvIE_mKa7}AvXH1Xs?@Fx4jN1r6mJYu?So#S;D#;R?jGg%)H_X5
zS=y}GTOAX|0{iQX_fJR3=JzXyhqJgv4LWxnkY$?M|Fc|8tgw>yEzKb02mQ<Ph2J-~
z)>w#Fl@<o8Ub%jZmqlRZbr#FVmEA+}DYl-FIM(;@es#q13Z!UxB=lOOlz5pNgsnF`
z6o}Dwk!F!jiK$@4eOXD#ddtaiFMH|JPMdlQ7l(tF6K@<c90)Rl$Jffr@(}tFh47!=
zFH0PI(bJ$vmytBO`3_UB=3&35FhCGH2r1Ptj<*vwhK1k?X*D<QI$?Kw_|f|V`O&sI
zdfH(-f9K1WeivOkBNMIs4Mbn8$NAbU>WzE3-MwSXc^<0_100H<Cf#&ON_bRkCnKp^
zwouRFCL1<zaz4Oc*`>I1(lc*h{S1|tOMkcA#>Nw~X9W6`&yczht=Y3uXKj*+SQ8))
zrLNUOaq;uv$jHdB;TS>0R{=g^I*lFaCb9OS`Im?0%e7IRF!s<TkQxJ{V<%FW0F8RX
zEe-AB#O&lm{)e9K?!%A(A<W{Ztv%~(hXvA=S|YzW_Z=+#?i_YgC-4Y6MbFec%`$M}
z+ynP9#)Z3PhQjmVN0<JvQ{&+!z);M6ZfY16*2>d@4xwkrd>d#o)DrkqAU_saejXn;
zFfhnOe?>sF5{F&j0}g=LXXBEsADzQLBt$HR1z8fyRM&H1Eg?7p@~Ej<g!J<ho%$oc
za}sPecZ+_1<U1?HM#=g<r`LBVpzXg{EM^}V8na>&i0LQYtDgQ|?gRn=#q>|#T$S8)
z$|uU<*rPfe-H>TnMk;>z(qo=SH$D>@UJRT1qi3zK_S{KM(TnOE=puvWgOeib(F~x0
zbzWSUk3b#^u**jkvM(E<>frTY)&U+EEb>A0t*F48@CZZbeW;|IkWwJMSQF#FqvY6@
z<TE`hCzqJDgj#8_B7g3}bP}DK@P=Kdpzk@PEI68bQ0yC`d+Z)O=86$MT4rnvWOR=h
zQw5(v6(W#e4zzB&6BFckKV_X^%k0^+D;uNV{VE-C*3)2&#W4?O4+jgrgcx>xSNKj$
z3dJ#maO=Jbtl@&f#Iw(EUE2EAGsd&gy`uFdBlmtcO~2Ha!W-w7TkhsET-lmP;WqF6
z3#yI=bGkwkB3ML<!-@EHJKVW3*TmF*V8vGXN2UtSaU-JUTkGXk&8=GS>>4_BeU$i9
z%)=QzX8^Vae0nMhA|6BAT@dim3}VJmu&}$KMW#!dd!rrG+ugx6YSJQ)S~c0VeppNj
z=y`flNyH(B#0XU-tTst1-LvQPwI)e(`*TcT<ryK@?9#+-Hb<HGFcz~*(3b+<2W(tt
z;C5c}mwmx^OKx?euWME`Mo}(zB(07uW?HhjbqUDthN@H!@;am~3-@H10m0Pu@sR~V
znC5j>N;8*sK{LS6HO!QEUHS5bhes^Ai(`>9d1`49SPv(rj7a+d9x4xTV6zKfbeG$z
zGqD1#<rx9IVWD|?o1^?T3{HB{#MtYGE<JJA#;*q}c=F^)U8D`td!^Y}Pi^@cTK(O*
zLmvkDD3fF69X8g`fA;3Q$gnIp3LfI3P`vn-XwQdhSDh7TE*A~(a{UxA#Wjm90?=>T
z9v~IEi2p)v1<@y<C@9_-FWYicq<Sv8Rol>MjN4j)`P{amWSIb#VRu;7UScAmikTJ3
zBqVvLWYn6IW2NIA<#%{Ky1Tn^OJ~uQ1y|#JVr&)&^%=q39igKnl<6<H*Own;zvCH8
zhQTx!-F9=Z^fqBu;)B!U&fkXf3IatKR(}m}DO9|u6q1kbkP1P><-HR>JQ0RyK2{<=
z=n}DoF}^MJknA^^$A90mxJAcWz0gn<#LQw0j=8rk_8n=s=fw1RVRBP~j3ifaJz2zj
z?UK!mnWQ0Wr_7buhAYyQ2wlxfl=?u%D>^z%CV05+zb-#kW)0IHGW43~ra!2L7&(}=
zxRh-_wHQGXmVQnSadvb!<lMcRh$vd9QD{kTs)x}4+{dX(YB8qI(~jh&9(S7GzLooM
zc$5ikU;ldO+W7T;m>u?4cVfUwr0i;#1w*0E=o|kXy>7F)ABW2M9yC)0g%Q~p62XgT
zw*HgjB692)T7y?~(UxkCY%rag;@#SjDCn%{M%<OT3RKs=Ry>AR5)+#+_1XsmIRJk*
zD&xnGA6bBx2)<i<n9LWK{(L=k^jRhgp8vx*d$9M#LbvFT1UVuhaSp>)D24__^>id)
z0<lBNEMfeyn_CcU*ln47K*-DI&803X2sZdP*MQr{-=}KN$=?&opaLI)0Kq>XU<e?M
zkF~e7X=440n*f+9vH2<rAtd#If=3n&F&S*9QA|P7w5tLu0&DZC-HUUF(Q-=GR!1lB
zSPP$v;0)6b;FWhHZsDnFjsX)-w4p@i1CRtE0>?-l{Bn8b#RBIsr-Pl4Swu=|Y}2ej
zd_qDdaHrz36k39<n9D(Gh8BxXx6GBBoHLW>K6;boxRC*)B{qL+1jZ7>V_O0RwBrq%
zimpn<$~G5gc#jakmxjqN68S(segq(N1QirC5W;o%#6_9-FaO-s+U+#$WIZpYK1cZ!
zLH6&^XgtpN+xR)~sIViifhKQvJw(hr&m0V0Dh^Nt+aa_m9nuPFm_$S<W1;I_FE;tX
ztejYI@9L_kueTCs1)iM<SEItd0F}LQY4IPRVb~JieK;Q3_z<s%I3y_TsY+gZLA?*I
zKEO>i_KOVL?>fDrG=N!ss>UOpt!jN$`yO-zuj)eNpieQ#^NABSGs3c3iMc<i?w;AD
zrrA49)Jl179~lTZ2!`4Yyo#j%V8*O{ZWWK0Gkevgr;lcuPRuyZom6~PTqwjJol1XW
zZw(jgJ7}|v|4vVl!1=(yu?v2g<kIY!YaJ-2e&@i8kUl*_HkL8G1r#wPjhgK0Yu8%y
z$YU4gdu{E&adP_XRO{O6Q@8Xsw;U*enOaH;aLis%0w~|SU*Bifr=h%q4mvz4=|0X;
zb*<EWreWN0V^rVajHP|)@(aH|FlXGrREnTe$gcy^0ShRCZO!3xt{ZZ2UdEwS!OJ7L
zrcf@D?c{n%NuG)xa@=Ggg^6g5Oh^GEP9ae-G3L{l^wLsN_L(GF|FWsp)qO;fVObJc
zP!ag`WLd?vUa`T5!Tgql%a@NJ`hbAa3mD81e+>rR2{4$Fae97EoOP&&)Xf2Aja{6<
z=gFzdG5S*E7YY%AB;h@gIiWr+?&3wlgc2u(xua<jBW-oc>MM2kem!QcFUh%sfAj%%
z1%!u#`fu&+?d^-P05_kVHOMl2)#P50y2Q*7oApz{$?IH7Z^$ai<@PZa8-y7ECZ%N~
zZUaXiUu%Z2Kgjjvtx8&tnXo>r85SRQhAs%Q<*#&nyT;c2w7vHZ5%W{{+Cf70mybbn
z#SEcfa@h0=5MtCFgziIP?z=!W=^sCsn~ybIfkmx4oONMZDq;uI^GoE<`kjBn#~uy}
zkC=kM-H;g-rnaoBg_6h0Fk8Srhy>Gpcxvp0wbL|C-GY4!`0k0mzEE-f_ZXUz+&a{0
z=OZ<u$KhnX+L~?|KF1KB%U=^1$zro+CH>P+O8zDs1=y#;z=sZ0`)?u1?BKQB9O%mb
zU$K?xUZ=R}25sq@XEE0vLC^_uRS}qK|INgJ_prPco!{J8_S=xsrkxF3M|Ns`&6{~y
z6J>G{!YuGT{MjuC{3Bv0sz#D|91*!x{~FWmM9i^~VKZ#g9MVY>CxjxwasiQ+hR^zd
zry>^03Pr?!gFqn4F!?wE4EBF^b9F8C9#DWHOrC033_@mQIJDXh<|7F+Zj4#ua;#jq
zc5hY2{EAzb_H%bVo{&!j)rcra(E1p)l^q{7B7q^5f>u;aMRAx~1z;qUjWdD?4wNKN
z<EvO#I6FH-@JFhn0R#=CsT1xKZih<q^ZQErNjN6n81~4PVZZ9%0xTIbI**#~6l5?=
z=EgKi$B%B+BQ1l81om_S5<#{BC%d=EGQ~@dJ4S2xb-P`bne)!nWi?s;74qb>g{+3*
z39_|OGcGGV_GYJ;e(hu^(30mu8E*<qMtd?}>dAdlS}la{-!;nb0DdtsEj7UKg<OlS
z^KuPHwF{6rKI1E<{~pH&a1An`LX$y)Mi^ama{2%@<q;ILntY}6Sv%L#MXmY6QW9Th
z)#a8Y2iD`o<e@n=63i(E_8Z=||0R55dW-<&t7By#e&re$kmcJAeu^D)2}B(}SX%r|
z>Rmwkm?N(a(dLgHZ+*9B_AxD~qeM=7VQ2LbiT`+yj&&fL4=N%)hE|s{lE{mx3J7m}
z^yIw4j1~04Tz`PW+2*E#l7<s)(`2NkyiK!ORdG_<c#oa(x`uUZ8<tj7U8Z!ma#|SF
z)Rp2N30-8x1PdISrluwU6rfRYkjQYfxcCNY`o$Wns86X5{*hgfcf1vGE#ejStSG}8
zTM_ZPt9VXdRd@+vs6)~TfqSA95|fx?*V2s>Uas`IX-T5%^U<Ji%4>C5<Nf^1YXW}r
zRl`C-d;_4-|7S8ooB8}f*tY_sNj~O`^--U@>ox?3@th1_adk~hVW{yw@xk}lwd+E@
zx?_Y$X76AsxHoPj(daDE5^P#hRGGoWRjV9Vu&kTXl~b9MiK}?WyhbX7k6<zyOACBi
zW_T_P;YY&2#Q)d#qeLrK*o&0(1P@W;-Ge)AZ$*TDRQ@!r78)C-rZ6mU5H&eGa2g9!
zE2N~PP`T~&{TE1YIAB`=d#fmu0Y%no<oab>zCRf@Gn^NMQoFh>`FCmqRz8qsK}`P0
z4%FK#i2~;7)29g+5V-7>7r9b>nsZ^1t1xtxjRvc|WJ21S7~z(pY;r{cx>R-G#TX*s
z5Pl4S8$%as-8M<KmPUEzJ-64HR0%IAEbH;1Np9L3G8O3Ba-gS3^E=$1#3oyg8WaoR
z5poHhklh_6A^E3ThY-heYmU-#qv`#b(lE7GGUSEE!?Ol6{-?K%8u;FUxL|JXfpC9-
zALIyN>u*)&WsTs7bZ@!U%a93d8_~|7RnzwXgE&;-NdduwcL!z-n)}_gD|wiW&9*W+
zV=pglF8N;fZey6Redpv@M|A21IG(55HyuHrLLBu3vmxX@*_ePRCnWTWV=n_Ya2wUY
zGHih){sQEdpkX#!OCNJzjA-|}o|NILD%7hT>Z4=1(W!B^XFrka8ek74+23GViHY=R
z!RUA(N%6+?D-%x-c{%~e<9{IcEWq4Sk_@!f^FT&a@SP0z@UYW=x*eW=kcr#yg}lns
zZ}@fsNd&QOd;M#pTeko~)bVSZkQB~D_(8G+@J9)N2;(e)u=+f8X=0$8^DGJk;B$c)
zcm+wDgeMB4pY{I)?+2t@eYHacsf(``MHd#L?DG0Ptez4WdZINyG}ZTtgsr37?FN6#
zRE>ZO!j+RyP4d>hKikJdYIYJ>fnao4QkI2DYNZ*C)_UI?D77i_P`m-p+}=C>Lyp`&
zVK#RL(|(Kyb*L^g{MB6JxN8Ibov7PT)tHUXT=Fz}AzV5!k0})saT&nUqUimK`!skN
zSj<q~JLa_`u6+8#Weo(0>;bXsfeM_oBt|(&`!<F>3~GdD0c+|UHUgOW?5h3FfTcW)
z9IRzJ&=L%?W;(o5_D#QRp@|I_vVN8(p}To%ZH4C4U3IqEHY?|=S|@woy*c^RQzd%m
zV!LKA-N0gEu>qMCLHXnA(@RF+PjQXPIj#-kFTaJlKDk0$3Nxq_F{k!vEZzwZ3zK*4
z;O=Jtc_lZP?72m8-2eBmIRC_b3s_Etw~v(Y7ePhBiR1eIVSMqdCH<(P5b5e6;T*sz
zU$Ag;2Cx%2eM|y-aMTt=$j4}9h`_u}6d~-3Cw_h{h>Q8>D+~XO9nWYV-wm@oSL}Ht
zjO5+Bcb?Ph9Iz(A=aHSwVsQspXw}GHir7zf;1?{czNn0#CSV+OKh|&57ZIOjM}9Rc
zAU3E*&~h%`zP*V`A5Un9l!i(8p>gZZho&<pf%E3irFS=v1(w5%4&`HV%%sa3&_Lk$
zf8x;}!VRZ@#toSh8C?shOQ{eKz&h%|ucE%e;ChjDSrqP-NlH%_F^)Bp0QVuTpiurg
zLIN8a=wb7U_^P|YxAp{24t7GXn+Ng|il-$E-47z+qyTGWmX;=B(Es*RA{G9(96T=|
zb<Ga#+PQ><VN@Zc%)!Q%J0p`t<L8JesH?2?cvb3L@p`>1Y(YT5jAKpR)zT*OHO>{R
z&{a*n%kXJ;2jv03{N6j7G%*(!7bd`;?TEPs95u5$h>k017X1?&spSLSx#szvNmID>
z!YJq7>q4dpOj+Jc^iA%76i{%07yr3<x_x{PO$=yVjs_p&S9|^N|Mj)yxA#_hk|->6
zqil|jj`*Nr@cD`?McLvj+IML^E|$2tZ&M^c0NGwvPM8zE0rX5l;y+*|52WEyW?^+R
z#<>5;lvA&<I#Pa#&zl4dK)MV-Q<Pl$2l^7Cqt;`0`#YJlHHJHO6Lgbw;N@5o%j+G!
z)*b_PU+7RTgdB2=$@A>%DpziyPax+Fj44?Qj%E45@R*VyV-VNs-=Au4lrfXKiWV5$
z6I$g3kWp<OinTuf;AR=3%V5h#hJU@m1>GA@+zDL;_F^5*aErPfB#9lTBDt_2f8k?9
z-<{i;Lg36simLwIEGvua1ll0Q*-@n1=<b01Ni5T$d+uQ>HUsz|NKP|usrg8LmROl5
zgq3|<;RF2grm9z)=V~RK*8p~Pkw-g^HhdntfL)cigapZHA)a6W1yc~<obHyRAP4EW
zx5{S-;32uECur#El`GBoDhQ=S=sWmj@^~pmRC)p)Rw%`%+KwR}N)nG5@d<!OsFwI~
zg4Q~mj*cBYR6HsMvns~O2scuz;mRiT6}IjfljM|viC}O&@$gmP*INz3EJuQea3z2+
zOiLT%rhdgSn`x4ZM8x#TkwR?Kc>1e%KSi;J=zfyOgI&0IF7qd|K8<+W*1}^Q`RhAL
zfM|`CD|vka#5Jk$&?p-I@%(ugEf(QWUSt^x5)Bf&on6&2yhLKMgB<|>{o_EGfG%8k
zMsP^)3jgW)>1hUp=eJ7lO8PuWm=zgcWM@CG_CLI4pS0IdtrCj71{d0Om^46wkXMNN
zY>4j+n9?qOh8q+R;@W^NEzRo(g6q>zoibd7cw~S`jbv<<`v&3&n8uRzwNlqkUYiO}
zT@2F-+EiToE|=cGw;ve!EZ-a{kZlw`He-mQQ7oKml{}LrwYGt!hTjMC&(rEv*>f7#
zT8EvGpRefq?@f%DF_?f%JMCJ1dO=y%=Mbul<X>8P`ufoEVHK4_Vni1#>|gS>+mj>^
zdn!x@v8|@mGvK5HFf)b5e@OIRlyq0$$IYaNTTy?azQG_ubj6a|vaeV?5>#pe@d7mE
z7EgH6&k|85RoR!IQ>1!Ehf1B+{$!)GJdYHZ(9T5kX$3K`rP*)#f}3`-jc@m$><;ZK
zJ12agppWoj-4F4x)}CyG^o{sLa{UPW2k)`%7V<Xk)YjIvPo<L<4uXCETV(|W8oE4E
z5F#WKiBjKPb5F1Lo&Hw^&Rd|YR&G!6=|lZlML=+Db8a{tz`+=w`<&V5X=t%sHO&ar
zAFRH8$LQCu>pMB6=V)!y{W~16XN-y`b>c%Ns~?(0p_#=YLS$VyFKJw0PQ_ZD!xdO6
zNKi%-Q!v9Ps!OqCfvQH5{Rp9umUaN{N%l2sj^I(l<9ZI`EOPM#mYWf}A^0L^LJy6g
z4C=-B%YH==F#%vr6UsK^&vQrIQwBMP6ymp5Q4($a$c-<AJcDwHp5gB~hW!F>X#=hm
zQHAyR2@12dNHPf-3QF1cL1x=FayF7L5%VZwVgYJ}J5I<QKg0!_1B?dW2JS6`XELE(
zgAd;L(Qdg$SG*aLV}k#C8n0vrybchgyU8s~c+sH$@Z6tx5RTiVuKPC3pMz`e0(@=X
zI<{P%!jFH0>((EWE!6nEh+c@npaC1&RH(DAzvdINH$g`T!_<--ruZx?p0hULGOuPx
zT3>(epZ+Y;6}syQZe2rf3G5AzqbOj0*a*lS0aypPj02JkBvsU2-{(?TIvm%+*fj{f
z4<Z+d@gPB|->Bq`K6&zE?$iSpA`ijhNS*|E*U68-42BSUz?I|*9#8ZXZjboU-whxy
z70oL7K@@a|k3-RfhM^2U1T`P8B#)O*<pb;j0NIfMfJizNx8$x9taD5hNN@(?4RDqt
zR{qV4VF3QEI9*T`g3-msgq10#rbg1CdTy$I_(gdOr|V_k;88yn(m}!90VU?rrQo0~
z{e_2K1T`Wyi<d&N$j>jN#pBXCMA;kG@Wzqr*#6&6xUDOIuLJagtOEuU&J&1GPC&n{
z(o7@j@l-FCc^}7Kf<mr&qMtiDLE>c#Z?pn>qQBX;2>K)l)_>zl*W!LG9v&VajU_m8
zRB=+_`>TFm3CRG=M+X*Pfez!A2tOYV=&mLh!XPZ72<XD;@C3Jm#o;O*^q%C50%Qbk
z3jzPFPTPs0mBb|rDS1A`*MVh73~?k;Ojx)Zd&uOC$jLypL%0dYc=x;*s!*aj!>^8+
zX&7QBYySLvLvB8Up$u1fh0n=}vZch!qOs*_`@NsPikWQE`1DabEnsD<<0iSA*dQ>y
zxB|WW3QE`GXyTy=M<Zbe1Pg8;xJY#8;?c*bw%}jE%#75DIHrjMlT6iNd_x5O9kV>U
z0#-IH3sKxx=eHqY^Z?~HG8e$BA`1bF2~h(FpnBp461b$!<>qcB*Yja5<FJoFHrd*X
zkIa&$u$F+_VW5K4Ol&@f;XWrfuYs!MO*4u$$byU?I<kwiMS}+)Y4O2MmPrHI5vvG?
znP6Nzqm_}4vl`(>Gbmy38Xc%-m|0noU~mA?H)J1DWspOKyiZiS0hm%id4##{f#z2e
z?ykLo&g?Bg5T+4>P*YoqAU@~?m^s$$WdAlh8zLTuViL=nTqs2-Xre{H4pFQtpy{xx
z6tXZrsnxI^psylXYwR2xnaHa_pK}5I5CmPHx!;s%FCN3|Y%M`FvCaysvts(iu4aR)
zEh6@3m?YP(nvqQuy7)06s07c$JUYU4@-ECsJiy(2aEwFgLktBXbM|6s`hF3n_=vm6
zY}3r`fh6Tw@hid1zM#)0>nA6Vj{$6x@G>zr3moLf6X0`mX=Rn7ugF5@8AIGG$)vAq
Lq*JWr6!m`q1fu5Q

diff --git a/_images/sphx_glr_plot_lasso_vs_weighted_001.png b/_images/sphx_glr_plot_lasso_vs_weighted_001.png
index ccb95f872a62ea7a7d62a2a0b0085018bed59898..9c254059514bd67960c6712a505eb8cfe5f7f1fc 100644
GIT binary patch
literal 17076
zcmcJ$1yt2(+b_KB1S}BAQ9vXF=>`kw?jE{9V$<PRsFae@NH-$gjuJ|DZ9+i0V}tZ}
z-ON1i^StMs@2vB!^&OWp?itztxZ}Ehb-$LE6(cxHb{2!d5J-qWRK#G8O~dEq(<kBI
z!7Z|Z@CToxsJf$)t%;+vzP&L<M&I$Vm93+dxxsHv#`X^8wl*9rJS<$yznM8YK6c<|
zWwrjVA7HVyH)UOkjn#mQoOvv+;ef#q>7##+q=}@OV=(uxOFX=%>=L~&<gOiJynnE=
zd6(S<6B6>wC!~wLgI3{uAf@=n)y%@svN^TnZtjk%2xIeVm4e_1oViwK4`EsVpUj0>
z{+taPnIDo&-lTSu5E1`iAWVq=g5P<@8cf4+hvMW(7a{6E*RH`{u1)Ptp{DvBj$W)@
zcrB^nad;*SW;R4khv)DU=572j^x?-v@wOxA!!m{zjRXecg1O{)_(2eJhX2Xgg^_B#
z2QSYn+`apRnZ#E$SNFEW_G0-`SFWB2m+i1NiLWTG*fzJ*=_sb{f<Lv0ZVW#*W7Sft
zz)T^ZAR~!RD(ss<-t><@pJ4|G<Ar~_cZybnnIxQQt~WP#E~V$S-FTh(R7;H8_Dm{%
z181A*yfn=3ycoS$Hpdbp=&HzC^($f6YsVTp-<OvYb(T{+`D(FNkyZL!3g(vxwa4-w
z=#jEDRZ_A{TVp?|GjDIt<u-=sIQ1LP_U7tYbLJ_^#tCQl=bQG<6t>Lw=Nsp#JMR8k
z;kP^{-^-FdcM=o(W1)DY$bCQ1d-uUyPc~EjUPfk4Mze|Qa}Lv%TXb}y2TKQgA~Lap
z{Y9Qj!<GDw({D*wb+gV~zS*UJ?g%E$^VrGrZ)vUjKHe)H@hZ1=x3S6T=;(0#P8V)A
zUguvn`jXtrSS)~|{T+=Mr$twqWM%JIt*?r<wl)WW^Vs@KXC#Md0PPGswey_&k;ehY
zDc^FMG?A81MR8kqlws!@tg8EFI#au|G&r2Ovs)+yoTP)gJ{e1uFNg(FVM{16PwL{W
zaYbj(o$Gu_Kt5lTnw0d7G_icKEK^!zX07Aru!>A1XS%N8kuLJ{cBix*_wc=$?wgAP
zW!TJ<1Y}7j&5<)J_0-|Y><_&*=KGCxJJXfuVG@UJw49uZQluji2r2o$SrleSg|kHR
z*v2+i`+4tgO_2$@eB9mL{kgQ1E9|*FJ>)!6n3-=-dsg((qYU%T6pNv<!aBceiE-Zh
zrA<{Ei$j?LOBG6g{PD*s9<Ks}i2qI>*Vh0OW=&osY(2mU3j)VRKqaUi<k+j@{M{h0
z&&#^CsY%(xqhg@gmNm0GqpeMzfLb_urv<~q@4k@*kDdJOoA`<2$J1QbCdqgo$A29y
zT&g~(x<Nx@&6Wg1<F+*sZao1Dt23I<p{vkRZEJgbwnK`w`^k}G1vW$c<$E6ZZZ-5x
zc&(hU+gcXCmPZHjO#^X(#LKiWr)`@f-np;@`iiUvOBB@fJPThEQfA0Rarf`-RN0L`
zf@j7rjf@mo_kS`j^3FeY>=>6t*DXpu`<g{hilJv-V>M^?HoXsuclY+XVZQvyd1OU)
zm!H9M4P(|&Ovx5G*v)_BNBV)}wq_b!)$QjyF=b_CS_yua<@kXDa~cV>Fcx~M&gEJ&
zmYLitwCrUWavArx7$``GG09Dhb<23<ejpbsm@Y=;`cTMiUEY`EHkyzR4}4~QsJ(6}
zZHB7pws))PWy5Htt7NLq4wV-VaVhEQ#-e$Zjp6@pQOJ1fmN?IdM}Cnt&ZL-HEjRN?
z<Xs|u`7`vg(QR!X9-vq9zsB+C&p-bRb=!k=$mO#9``+$a>vm_{eo`Y#x%7zFPPR$Y
z`wxD8=PTDcq$aAoz6a5jO~tx9!th$cBCzQHlw5VN?@mWYH?vuNP?;?Yv%0-LlUV97
z#mK}Y#ag}liQ8r{4Q6k+XS@<#6NXi9VW6-RHb-}btApqApQp&!44w;bPty!}?JSgy
z2MC}YAhh20fK14(z@~J9q|mJWK0KpEZ%$0C>*Te))`Q(ZxQC&@)#{z0r2@F9(DLVF
zvt8-I0aQX6*twpW<<YMyu|H1U-LiLZz{BpdT%T@_<h4s!U0q$TBe%W9&8?cAp8n?5
ztGBe-TE()7Vu3v}JVQya+9ZNtqa?z7&i;9dIz)N(>)>Fi1DF1n<D#OX3_Lt)P@URg
zI33rTd8C4>c2}hcDEXDc^}I?=T4S?eK~2N7C>-qX6<XuCtLDr&Po5`@hti=H#baBv
zCtnFm3qRK@x3`T~l9dfWrKFjwx32+8L5dV>rE<{vdZ(P_WK%@h&SE*bE2;>&6{~Mx
zHydr*Zq&Z<_2shayOW%f!g&4qy`eH}XFaud7F1cjEInoAq`p3N`~ptRd#h1zeIi_M
zc55n54R%TB+qd!D)&pjvUtUOxi+|kPVg24?qn2|mDTsQ;wq*2VjEMKJ>e7<;L3t#%
zwWd<4_Q6)GcYzz;VhG!-1I0&4RW(f4eSr&Y;&SIDRa}uZUFaamvdF=jNZDG80RG1_
zqKx?Xy8=%0ZIjJW*1gi3L(Xx+yUP!tvTBTuj`G`$J?FOSix+a=NK?zz4bra8Icni|
zM7}ObBDkqmFYwQ2r^qEt+7m>LXQkL*zkY38<UHTk?nlnkvat)*gtIw{hv#5_vwC4x
zCQf+B#oayXs}GTYwGvdZm9;ev&z%h})0XHwCMFgZ?mAM3jd_jBSFRY9IWO6$O?AqN
za0NAGY2;hsY|3VCb|#7uDap(KUL5DOw;hqAQ$A3XoSF)E-{0MGTH3<*>cQvQH^fEO
z)%)9-U&}BVQ`yDC>_JV=Bc4-G7Pr<?K7MSo!R1!(TReUCtXXJt&=9u)13>2O?d`C?
z>zcaMoo_1VFqd5%hKqgm>NI*1*=SyU+{RZT5mHuGRt~3mbv42>W*3|m28?MV=;J&(
zNDZ4{!$_}9HaB-9OZl_a9Y1v{>clw`QbWJ*&COiH!^0+HHLvSP4GF0P%_=G@o8hIb
zi(6v^oFk=JN^5L(O6<m^>jNmQTSZ_W(R{Cf&01s)OLh`gJ7-XnmTjf+rl!Q?Yx>xa
z?s1m6l|2t1JsOCm7I12s!TV+Qo>xkdW(#UUbBBXHWO4N9Q9s(?*x1Eo!7xc_X}Ch3
zVSRKD!>wC);G#~rj?Pa1Zi_;Jy3wybQJJ|_WrLgNPM^LLd{f3=Q#WFcF+Sx^4Q1MK
zDDA&~{bH>4g7HMfShLU~dAvRlN5#g*#>mLnQN536P0iIQwuym_$?3W62xWXA6l$+T
z2h?oG^=XA6?}I%jtW70e6}GiZoaaVFQqZWro+mzZv1-R=Vq&6yzz%A0ZC#xOY&sRP
zvhD(NrLAAT%wUIY!(1}$xnWes7Zz?)Q&UUO30cz1#ddA2O?8($+m?TbyscHF?B(SZ
zF*<t{6*8EQ<h`9u0p`|k-`X}7a9m!yE+<Z&94OLrXyz*0nvBA?dhh4I3^=>&p{9ck
zWvN_O#8pn9y1%-%W}RQK2|Jsd$0kJC-quY);oZXiS}2oRMoEb}EFGxL_r%3tl{qcU
zE|Y{RN=QgpbR@mLDI0Buo1B<PmI%Jd=h^JJIb^!&@#FI!+i)YLQ8v}d$kDii-PD}&
zMSd9fe%+B28S=eRBJZ?iPi^Bu&4~$>pf6X5iDznFU$$sZc#%`F+Q7cDm>TERfmW-d
z9=u=|Oe(&SwVIdN8DDlbO@em%E*+H27N?8a#5<cyYKn?M(2xu_U%h_)ZXH)+8L<0n
zJTNF}aj0A%+hMxxx<m(5G>w!jtzw&^)nlg!0e?hrn1-2485kJ6u(qqSD6|-^aNUNk
zK}<oBb?@29nbN6PB@+{q<joWD_F*hK8J0PvbxtEL0D0E?P2(8o>FZUe_IH+giPf|}
zKs(ET-HTdENJxlLaU6ZD%l-cT{<5(*OkJ>3p?AtaRrweie$_OmY?hHy#7lc?Ys-dT
z@^ijA^ys<XR;S#pI)8HFD_4?Objrj($+bxxXhp^^0j#Yf#X_fPs;+h4m}l4=8GYbG
z$iT^2(Dbk{dC8qkPfrggG?@%*a-t(}CbZOq7r-CPkERKAz?i5C@ztwywO93&dvkPP
zq`S+omWAe>kI=LDItL1Eyw@x+`;eA)3+|u-%l?z+$>dn~MZV<Z<aD^rn>TMd0FEwr
z6+Av1^ucF@%J+?zJ}}03BrDt5<=IqinFwu9Clb!-+0_!O*k~Dt$o8y$qX~r0)T2*n
zdhkuns3C-bjN9rxAo6U0$@ihRbU`bEzN<GgV40h3A9aMc*s7l$VRD*Ber*NjomZ}0
zp|!QOZH^J>-58IhJ#pfM`Oe0|_Lwj0Iy8d0wN{Z{C<Mu{l`UpEl7o`aWI_ArLM;Rk
z$SmxaUSHO#G*MrYyrdgHJ|i$?(oD86Sla1y;F=Eg(j1n`He9Pv|LaNUaH^>BB?Wo=
z-X`%CKG=0|b$1W%w*{nPk>ww-vBc=C%b{G$(JF{m%q$E~d$62v`sdGgU0q$x;q{z?
z+jX$rDgb!mg!k5!5Qaz$6z+g}lA)1r(jOLFj=JFEA9u4g3neXkbGo7U0w)p7)-F|q
zLC!8TJj(L$DPh+Wpa={C0yrBSFoDo_@AQka9j03Nv9lrB^%d;EH<%?PrKB9TiS%du
z=3qoR{dtDX)QW6Jgtc5}YVX1HWpt6K4QwxeC7K)cq3+q=+abPq@xxS{w{BzD?aY<U
zqrZ9n8ucmLU2QD1pOD1eW(~5c)VjL5veN!h|0^f-W3*V%l7`Bf0`EFtAoct2@nhrT
z#MIQeN~yBuZSnUah1`p7O27a8swqp6^OA0%&5$N^HRsNEM`NeH;M;frBrzzZ^fds+
zh4ohGwdZC%T%l1sXz!by1#iP%cZ-=ht3#v>8l!@q_Z~K4LYIhwSAJr=<c^be8hh_~
zfC`Q-?tY0b3M$lIzdqPNU$0!LBX`2(=jWs4TW$^a>;3kQhRgn^c#M#{s`H?oPpsFr
zdGmX=uiQpLoSe!4imMl^54<Ra+>|Tb@Ey=)vSEjodFY(zqJ3F%j4K|V#CVbVU}xk&
zH|5kG@I+Nf$v0KM#(dd#zO}UUKxvJss(*X^!Oy9wROn~`3O?sFk<v&Ybo%t^(*m>h
z8^k0e0}X(W9TxiY#fpdBGLwRI3+8gGb#p5>G$W`lCXMt6=Vo=L%7vS20d|>~nlkOB
z!f<64IA?K^h<89a3k?b~P+NkU=hRwRQDK$O8SW*<n$wvedJa3&QL9QtMwS76_qI-1
z*P_q&f&$KLo$?-*BS*-d4OhB1_V%hRRjdVg?QfZuQ#nG1(yeqe|MTgIhSt_%t%KbF
zfyKAQ#l--@3^#{DC94{*dS7L|LQb9uZ*IOaR$F9UTU)F7@%MY7fr*^$)g@m)z|f;v
z0(_O!)umj%>OAb4+OSaezPQ*qVy)1i_6@tkl(xsdhlj@%Dk>I3zm^z*)P#f^<=0mf
zzvlsz>>*jwt^_!ANoe{m!TUS<PhnJr_ZFS2JDYh%vS2Fj>3OV(M(Ek6%0#8e32!|t
z9&(DR3#85kiUfmW31fs-E_V6HpX|8|6TlI!>Uk(!z9}sqC+xY7W(u}X2a(8bJdCL&
zY<$#sU=&k-{c=Ho4#02G?Nc4V3=!ro+Pkg(@h&hV`9LZmPP?&(I$m274Gj&7Fwb7w
zGxR=luT_7J9PAe(EHwlT>nssbFI~8H0#FbJ0Q`xM{4>_~F@5d;kQ)a%SR{J&yen06
zw9|36)jDXl0XhJlzwyGyXRa$<smKi;wC#H^J$Brm2Xz9(z;#_Gh(#YhoCc~MM2mh9
z5WJd^QPR<4#|n)bNpS2BlP6o_5EWwN=hyV!9jCT#%mTL9mmPJU^Q5q_@N8#l;6BhZ
zHPuX&ckkXgO#m9`g!0EK;dI6g#tYx2Bxern1+g(;Y1c(WL;wR?c4w%RZ4SHRp$_wU
zv|glnbc>Nu5(e%gAq8)|%~074l&^MpRyw{V0#>=s6d7VFs+@_&@VPfkIr6aK0dM9&
z-z@iNUY<@2oCW1URa-ln%47K+j@=TrXF*hB!VT7n<@*skbFgsNq0!D2;cPlD%L#sf
z$w0(Fz<IF?*tFr?3@ew>w+k<xKTq>M*gs21*b^({zFu6tCxgZi$azRiOg2E2xfX#X
z<<?lCZdm-DK#GxcV-=@noRj&4Oi8V?{#a~NKf7g1*c`kHJ8E0zK;F%JvLasl`*WQl
zq_~ye6v<Zib7CUm8i&byXrQD3lschWbb|h~7GAwu8x3{qOOfa5cls2)YOfUF|8*lH
zBM(JIBVednb6Qp>8k-AybuF`0GtZnq&*}Z^D|M5c$eyx^p2+^M#A^ahS!d3iVFQ_h
z4SF#LTDx1YTCK5+`Aq{W<MqwJ0Hlj;h7k6Z0S#&fSDd>AXqr)w1=AABl0P|b_iI^y
zt%;-ceE_8ZS!TCI4wQ+tnbf#2jmQ~5)SMt*vE93OudcqnUK@m=NjdNB;P2nQee8Cf
zO;43pRt_--Xa|ieeWoWn$`ypD#+p>1r5)?Q?@4??Rxxo}8rIQDJ-&R4!=&j>kaV@z
zUPB1<m=1Av0EdL(yt^=)P0$PvpAkJLJZ}?XE>)g8R#=vO7ILLe%mYB5+{d9<0d{Ws
zwKCoepv0ec4Y~rZvIvN!@o=%DME*I?orU6??Cc$Yi$gT*n?o&68v6C;8TwoU&1D6T
z=R$)7n>@)C_5)*KQ)HqKR3o+&qliLa{e910zKl$fiHd{{6%`V4g#-9H`Uh8q*cpde
z+iwmt$>C^_py^q`51QE9+xxNAT_GWfgo%rQHADt$y%TtMCvcRt+05K9^YF}GmAf(@
zNgdirFCtwK-!j{siG|0s{Q8n$cBI-{hv(E?;=klXd=qrHdDU<R0Pj%c3t%@Io0zD&
zO<cpo{^5><#wTu*f1ILfgCDy>NtylS%YD?Wk&J+tdyqpsRpKd1xJjDnu;-?Z(82B+
z|KlG|&YnHnl`M6Yj+VAhX8Lse{`GsO(35;V_K}bueZWvXIgY*sTI=1Xgy_#Fr+z2d
ze3Sy%x_Woj7xsfup89MV-`zwLKLwCL28yhiSaY}6+!^eSFQ2cufazAz8yA1@=;`n`
zXyUB1Ab#lZ9Q%ACfPyapc2O^uL;DfM5y5k8$M0ghi+MZVzKiEUlOH5m^?ZMSA3I1E
zZ`at|tl~pRX*JTg{gn5p%SnUNPi8Z6?UJt3+_N>3^a<M9*;xy`@#Fd|?t<Ld=^q-e
zu46JXt7_l9^gB*ze1(`iIVG;Ez2fMJtu01IX8B=VlJJHh1td#j{EwU|i`t#jJLmv_
zQ;3~YH8wIbgJJ}_$ivcHjcTvT6xo;zopL7xEw|U=4zgh5;JP(F{{9E5gH27+`1yV%
zBoibgBv@n(Ntj=Y%r_}!w?AuzyMN3hl$TlCZs<Yd?QrFBWB6b>--qt{Irv?CxrCQj
zmD|RAy8B|O7~lc17x>Q3PAqWO*{YqzPN+nQKYtp@yuWkr!-o%V{QW~|Z|hZQ1EPU?
z<J(nitA|t}`2XFo699cGFL2W*8{g9J=Z^3#GM&7FJ`D-nl7>oL_H}h9%-W%(bvA^K
z8JF_H(Cu9k5~$6tFxS2JUAx_N$G28dpJ>}KP$;EUct_gY+`Pydnu#J%{Ggy9S_$Y4
zDyX5M8mRc4osN><vF#X<U_8iH<y~QegWAvlZUMJgtjbL=NIbGu^=YoQ>V3E<V@_=~
ziEq+djGnd2{G*rUiTCU`Jl7rZg+q5ADCt%E9)76VY+w7>7vm!?A)z2G?Ke~eLP~N{
zl6X){tWYK>L<8L`tE-CeH`xI+FJ8Q8Xl^ca=M9SV`ZV}cJ7eglc936^U-VI!;jAH^
zlfTz@^B%<nTzG)_jHR8Xy1~Jr1nQL&4%C9ssHjXN5<z_c{%fyJ&%x0-;{_cl^T`to
z?rVQpqS#u4T+3V#eKCRAd2N3h%#YJ-$1&PsPlrDPOnbn`$5&JjS~Eq@xz)AoiLla^
zVBAU0lR0%4F^`{<9>wGs2JsSIymTocD~q|@T0=u)b#qe<8r1|aU|<G5*!%bIM=$nT
z6gF5Cs{Yzn(EsBs!Lo{p?7awyACj&bw1z(mDpEp+pHUH$s7o!K{(Mnw{4|F5{`n)A
z-u7pU*Ao&GL!+auimgF-dJ_;3Mhi^)78J7_S>QB5!NJzpjIX#x1FhBhv6(*RgU$E6
zcbd#g6AyA<f2z?RmA-&>b-coF_nd<fC<Y`xP~};k6lB~5SUd+59Vy0Zi(H_NXzc;;
zmqQ}dI+R`u(3IVP_uxU1oBrTK^{G1oJP0JyCrgD}K=T7lOCHuS;BCvKW?0~bkEK~p
zY~c<XC@=lUgq3DqcY9Q}@v8q!vM85!-6c!3`pGW8EGhqvREDnvJbkfpG5jh)oHm2J
zw>3UJymswck&#Z&moHxsVV!|nb^w9SLhTkPQgQYlED|(g0bRhOI4ygap#U=8xN#pC
zz46U}%z%=`53?i5g>Fr{OcI)xM#PSl7v<p%)RY$k*wKB>&WpDV{X~GqRjP?q<5SJ3
zCFrL+_L~B(t45%fm6kO#^}cv*r>y>a#KY1Gd*=}3-rP0(e8Cxg9fM4p!w>&Z62W~z
zfA#s;zr3F#|6A|p|NJtCRtknJ7B*JbdN6;aTKsP!dZI5*G~gN)dz+7fxnu_B(i5`K
zkgyMYmQGAGsh^8(bADEP{q+lRuTTM@7L>Uk{iI`rm^3yhPLk?H%`AN3Nlnp5Q7uR+
z@Z-0*1~q+(240}+72E|b(#U&@X^7jj*@l={AOnp7xK?-l#brb^Ql@tAKT>=o`-ftu
zT#s%d{MGEm!@uN>oD$^Al@x+Y)xCs<Ay4yk*Tb@a3XR_b=o@Jl|61^&`L$ilj1kR3
zjLaXoqfTHj^0sKiES?vKi^3z*rAr6tZVTF}^Ui&M62*HDeU<IM)|J*mk<am87NhTV
zvz|dOk`IuYEjxmF@{%scJ#E|BJew-NS28FJD$3^^>uEliku;1D8XRYRcGy<p_VAq_
zNB?cXl|>=I4X#>HZf@@O#zH2Lh*T`7i*6dsVeElG*t=Ct!P7Es7#|z67?+3FwGdUv
zQ)mdJUtigX3tb&QLrff1W^)PCb$0Qdb0XkTOOV1ro_L3@7Ov|ik(ZZ;I5zOFhAzEQ
z`(!9;mhQ;HKyAZ$2pg_bSyIAViSN-Wno`Z&Gm?gT;8hegUruKAXG2}X6}4gX=m^i5
zCP-iisD!3%*1w!%H3LENO<<s!i%SWZ9EVmD@D)dOkvn(fuV24jR1cbV7pSi6XUae+
z%<Roj$sFLNfGZbEN;b-W`DE?%y_DAYuO2~{Y}}>Lh>3T48(;n%I6UfswKX+;MIaO~
z+`RdSfq{W#0RYi(Tbe>*7oeB&1uz)Z<mKOj;w1^NVH(Jw?52XcdKxr>@>3Y)lXlR_
zOUNF)dE*Z+aa-(Jzf^IvMt;=jB8H6fk|PhF*KP@jG*>7ntcrzrdDSB#B8aPmxwurI
z??eIz8`^*o*&C-mkhi!0gvgR@)z%~EzAB*Wecfw{5l}mF<VcYi;DHP;c?=od7=VGt
zTrZPS;+Q)fyspzQ(4^i?pP_H`tj-v&Zt+~g2p-vtcTeLPc6orz96+T*8w!b!`uD5B
z%_^IWwCpN%FwNGh*5w&;REpSZf?5wHvQVhCxp`uY>~8V}F^sZLCfiT5*tMU7-EJzC
z-Z4AFM_kUjwGAbKGP=64p$Z!P&<50f;0z%<1G~E(l~7l<;T6<JceH`TSpf0}BJ<<p
zAe4WA@QX(<q4y|`$W!Ljy@l-@zA=@htFojwRppF{eSUmM1`*b4*V3W?<BG<{4Cv_T
znfUq3277bV^X@|rB;zoVezCL$5&=>l&k{`Z>cD33hN{2_JD-Gpf+Td^LhQwfFow+8
zy}7|n!q1&ByY0aF3SCwVfSfCXDyMM&{!`FSK_+39AOorZUX#Kk*0y5hB{I*z5$Uqo
zNmokc1Q*eIpg`gj5tTB`$4esW9H1(ol@+N5oPmNzSy9O>NkPIQVH~vHF$G(AVnK!b
zMK0Wz3p*l!>AHe82DHuDzC6Q7F3T`>4h|&nBqh<}YV3T7sKdg;_iSmAG6Zy>0AyhH
zowbdP9v~bAjx&nJrly%-!jiEYy^8f%mFT$-dm68+y(xd2qEqU1+`OWy%?E3D@e`ZR
z)bRWG$dV;WjKw8DbjbS%rk8V4F9PL^9fy|cwQ(0x73jx!P!_w4)?haXO2UQkOC6d*
zCgF~%%&(Uc|Ioql&hxRVbnk$di=Lhnloxo(MD6!}Xgs5$!+X*j_U5`v;YNdh90Fs!
zCd^zeT4z594{iRmP01xaqSfNKaC+L$YUk?L?cEOCRP;_^7C%EVf;My^!~Vbxt9uH=
z<xhp;0Gy<umEM(QxyPPFf}{rj10eSQwSMwnZ7IA9i$LI$hFgr>k}@LciLNKkuCA#l
ztH;BPVuzS9Y21gEvWxtmG40$(V~o5fSl>|ME8jx*d2WoG4|0%bXktJm1}aTjV%U+p
z&4+LDq@Ib98v$d#XP-K(9eU<f74j0vF!CaY?eg<Dz_)*IUM-?MO?UsJPskn}jz*v}
zXRE8d*9Sw_7dAF$FeOBXU06`}ZCtJte&CnJ(BLJ^W9q|ab*V`Ec=yH0Uk&{R79^of
zlT`hAD%RZRFcw$-`cCXWbeXi?!y%Dpt-%s?e_^Y<^3zttPw20iGoW<D;Hh9e{ZpR^
zXV$pkTWhenj4!MnX&_Am-4rBlWsuEP0IhB;jVvvGg#lhZ4ebj*(e-E1l+@Uc*;hlJ
zpy*>}7XXTwZLLm30-U;Z`7*LBQAdDw!SA-7*4g`OXNQS~=IaI$#&xzFT7^=<u*pT8
zh1+&G3#uX$_^2mN<TP7B-bQ{zhu04}YTJbVf@tj{29emddYrc^v?+BC$|v=uNjc<g
zZ!4F-oI^e=Ks7oE(1|-t<2+Nqp;iKE%J~<WZ*9Wcw-<bj>qrr=IAl6tFi8dNuLOGE
zf-0S<mfHgzV9?<=cqhE}UmeQ2ILyuLN$w?|7o9tVpAFldE#8FeL3dBK7V+iFhKKIT
z`}d@P=6OAbpq2n^6T?oo-2=D@(tf#>@yyIjLsL@$&J@%K=$LbkasNX2n7hV@b(g;8
zZBoH`fGQ^8(Nc;gvVU9(vsO~ne3V^$219Wj&@6abnQ3X_hbaFUp!|r4nPo0DUDxDc
zk2M+3-IaRHBC80nZo5HL*#n0+$jDvCc!SS*5s{WYDxflHiyw}d8NH&xWF+PoiYw%`
zM{|`=nB8%oDFP_OF-ZAD?6-THzqPlwqbhmh#tlwj{fJpZRcr^*S#`p*<v-|w#m8C<
z*)i0U(&k7JO_6ThdT~fmtcBJs7r4%1qJL01@*ph;%*4>pP~&1DCZ@(2p~<V3APyph
zg_8I2-5{&nMo|9DrfoMRjg5^NZru1XDhxUg{Qk<0d3*x(*e>jOzR0P8hh_DvXVE4|
zBL#c_)0+tDB(w}%5lG`qe0&=4=qw9~Fi4<q70`|A>+7FAb0!aG3V|9}O!KS40P)Cz
ze-jyb!z5ubUQhH6GmDT0Cw5jDg)1uEH&vl&>blL)f#!)b!Gd*}HBe}YL}tmLSeLOg
zIsjy5JCc1r<>zyG?|Z<Qii5QHR+;_lP6&h2Iw*eWU?HF+4QLqMMb=vWq-+T}gqNYR
z8vfn1BKS@aE<Gp~KYM0b#i(AjaEKFg>)G)k|3?87>4x<IAe(8xN>P6J@HsG2tKbSs
zVGng(-D(1ZUqB1EKr00*05C`Ojh|l^c>bBo%a65U#Hc-$!LdTZVkl^KM~<Dm#lRp@
zRh6^He;$Krd+*||5dsALOk>ga8~t=M$(gD2gy)=HZ(}i!=>e!j3A%o|DHl8c(F!z7
zr&c5dSAyI=vDgZ+E3k9hKs;1Xl7aY_S=(5d4}ztVyGG%k?7ho-1%E!pw#mf_>G|9~
z{^ZhOUocySAIk6!@0n$D&!#n(@SrA`A9xw-fT4OJF$SBe?bcsi2Fk-S@8LK8m%teO
z%s<L8|Mz>v|4&czSIfDjqHGls>7^*PcJ|88sZ6NYuI?JMWhmKK_!PE=7HBJAAkl(J
zhyoQ57LlD}(XCJi15`OvHM=WAg$V#^NV*7E#Nbz#i-FKvPIL@2$ii^@!Ldot<aS@4
znLjs+E!W3n$nGpLA%u(VgsWSeI8#Rcw}_%*V2}U?Knlomj8I8Y<V)QfrWxg4s`j@n
zp=PB4=0w2b(5`T^v$JaeWm9O_eW?<G>Bmuj)uj{w>E#Qse1Tc{(}GgQd6xEnkJ1lf
zoQK=@Hb-*7C_&CT@QpGkP>9{DwL)-+mbPXCNl?}lgqRYb8>k)`Q-p-tL_z<2^M<k;
zUd4<TQCYA#QQQpNU@KfjMa7lNmlHvm4BxQZUN?t)f-L01oNy4<%Z-o!O~-X26v%qp
z>&x^O1q&GYC}415Gkn!=--WzWiU(?0*n11m0RLxNI~)7mUv|XA#0<Q=eKzioAAf|L
z*<6P(mHRv!^r;?}?}5I)os`alA3lD(vj;-D8&p{?$7y*Ovu`+Dd6O{gbQN`VGxyl2
zD6LwWqiyP?0s$1w4zbZ{$yAuuFNLQuB>_~9MELpq)>#+`#7}<C%~|4vAr!UO>%AB5
zwm$s<Jb-CHxZU8Ur-D2$FxUzfEb)a4zq>n5w|xi;z}$T>y2yL!=aiORvSin>9~vbj
zAeA6p4Z?;Xx@03AW&v~<D0U2+g@?HuxS<88ya@C+WV?fAp9vFyRM4-hpur;vYl9aY
zx(UMLcN3@J;>09rK)?93q?>OraK;;b=fdELdT(HU@I!T2Sy@p4V`S*KZSv(RQ86(A
z&mq*NAyR@824K25|0hHzz^dCoXkcl{b`v1iV`qw$jZe4i)JteXB|4<ypzN3mImXA1
zX&wqri;s}Z2WiGafWXMA0~sGYGBwC|(V9bCGm_2lW%hoB>l&xSq;w^OBT)VdfHgEf
z=(4gBK@dSp2GJRQ%OIZx<RLDPUys3|69s^QwfDowU;evbon0*+uqHaQe29E{hgJgL
z!s;S86I@f;4t6DEL)n@mxuk%iL10TB@(R{tCD6Sa8|R8g#7S7R)ACJQXTb+4z{$ae
zf*cl-x)%EK9zd#51>{2zgEOx`csUJ0qL6M2Z-sLhyv%=r?p%;G{{r1@4gh<}2>!F?
z&X2twg2avbVin|~GEi9cr6y!br@j~oGBUsY32c^=@eq7M6Z!upx)}EWfVmGe-eOyV
zvYAQ#Z5(UkM^F(?`T(B)u@cn5&I!_o@y{!LMF1rkVUl31dh{Dn1tEMecmzXr8ww3H
zjefeRppXcqcmTd3Az=z)w*gP5X`ose2qg|(<k~6YWE^7h2?k<Z@cr4V08^jT1Kj=x
zC1vnAx_@54B%fSMrXmr%7pTO?FoLRY<81wV;M&VZhu0?l_o_??>K{z}a?~;(6;B%^
zmMp7-x(#|kcoMsW+sqnBe_1|gY>%TGi}}t`Wq>Hh@NLAR_}>t>MCHG08miD7OS<W>
zLvT^XOT9wMd*ubHf&G?r5(vQl-pjon{WbgVVJZ(j{+G-hP5lx*0uetZUbp1VwZOCI
zX!(4!p!R|gUBuqF``_gT48b)d%$&?k>$}CZmlQBbeCu0lFUy~-x##g>GF}0IVu2)t
zCH^`8h;EI#Nh4p2$H!;6_$Ve#0mk6T$?OXKh9J@}gP5*$`E@8>#%?XFMzk<z=xPZ9
zdYYj(k70gXK(j@l>iRWgNvlX^lkIuwu-wzhfDhARYti&mD(R;%Pco-L_u>>MCQx!p
zVI$DbSqhFXwY=r__mF>9S%(=QOM3H?A$;n~^(KN)CZX2Y;d0%G57VI+|DFe_ncIw8
z*J(h|QrtO#iNEpJS0YKh2IYP3ly|HNdQQkT?;91fI|vm!59vH;uRLVTIk0Su-tE2e
z?MM=`Q+Ihrr30B~a`a^j6>EG$PtEf9OZMkmV-8g_BJ_;&t}VizEuh>9F710b6N(1@
z&zW3(Pt5>(_Z%8nO{J(!!Ei#QmL`T<o3RdJ%#0E>WM3|rh)q=(civPPP_rJ4(~Fqs
zVBGf0+~KRk5~&=pWOo~GXi9Vh{SqeYD<lw_0=Wp8L0xY;++$f~lO)}ZuLjH{-%~3{
z&)dv2d;iI9`5?J9d6qzbBc!l9CT4A(V=ALU&Y6WxwV(R2Hsu=@2_PbA9aDDtm1Wx$
zM=<`U{w7juQ>|(yCMn4K0H7|apr9Z-mkY`UirGLKu^TeQ5SNfU%wG(WCf3wEybbLi
zl1bo*RimN<+@Ys#wCnXMi({tdWX@iXQ$%W^&-ik5O;cwHKW<~d=cJ13K|qW9%$<oE
zIVT3D;x<(g1ju4SYq3Px9V7QeB}}aMJVTvZEbohMV%0Xve>4#B!V(MVPi2Z*dE%<4
zGP9O;F1Cj%<j06`Kw?!$sUl(Pqo<3H8pBz8k&Xg3IszxGmFqzUhK7ZZXMxC1X!-#d
z&lV`~D&U-BQ@+4m31T{m*dnL__Ow(G3M%d4bG^GEYY1r!lvqLd1CpI#xwc<R)K>3L
z=uAsyC~82JES@heYUCwsWz3xBCEMA{$Z096#JA;!4m*oZRVL^+v@~R_I&bL-J-}~z
zCkx21FHv4ek@NbN7FwM<+hnlxW_58c$4bl6po3p5ZoF(LV&dlBgeMhtC5G;$%gr27
zUP3twOKiY1UPTdgKepIH6Uk#8U@PHgI>o$qa3i!5khTD~?irQKk0TKMtd!mZ8HZ6s
zL@&;3JKY)w@zhil_0TL3uOmf)B;?jpx&FK;p$PF>FkA|azF+bu<I4ELJux}S_}g!H
z!9f8}#94i6y?^a&9R3BZ`R3(3i%EyqjT2lNlUY0yWBPuf48ck6r(W<?JjrSCvMiS#
z;}uMoW<21f<%?Es2wh5~&yb|hC>D?+s;aL1;@MigMkcfK*k=LT{rYl;?NlZAn4C)K
z+|u0mf=Ibz%dtpqg_aW!h9B%^GxF+XMR!g$ROY$^1LvQAXNTt~l?R5*>9&$`7B`!A
zJ9&q~QIx#AwcYKtb42V~!(x6ybxekRSNX(z{c(fSy*FYtZ`x7BD&OSFu>@lgfGQv4
z<%?{FG5|XKr(8;?7aSCn0F8LS+869OaCGS;{vx`{$i)(bkwAtTggT_olrj0SeTrqh
zc2nJByWmYL?jYhyoX`>&K@E_U=ot(hX`I@%{Yq@vAW}8V6nAnt|4RL+1_h<qhvGoL
z+OZ9ZLc#W0JKD7GwylkwDH3-6wP$>jN|~yAWny+!JgUCYz;J|a3Y9MNvhMEK<D;wB
z_0;C7oQouaba$4-@k@i6dJB6Cr-;Qiv8`2Gj2dTK(tag)5)+e$iv*v;xy|?GvJ@6p
zcs>$o8(DOa-s4lFm;3ombH9<v%GA|cKqb${IVE;$O1JzS=1(tGViMxUsfxWpOwENH
zQ9(+0&=L}J(rMu6dR|+i0J}O7F`J$?!)+j9HsZ175A;)ek_%EJnq`jZj{>QzaU0-^
zTeL*;LF}ai{W!SobC7gX03uxNg)%R25=FDh!!^Q_goGsH<8raLI^@i|A&+A3O(zf<
zlS$Ky&+f20_N)07E>?Avt~&YDl1xR5jbirpY9^Q6@ig+93kIj<bqMGJIB#Y>18eU;
z^kha48JY7V$K+K-gDqSc!AQRLmyuleuI~Bt)J-rW!ypmx>C>nAB2(bx0Mth<!PRTX
z*4rLIXp?Fk{tf9F-oYuqYaH`0o5S}lbf}~{Q_AddJ@#`YxXJZkvIA9%n4%+TIrnX;
z%cmw}GA6!Tum|vx4U>;8W!Bryq(ptCtbb&x?SENnEU~Nj(%kR8y}#B5yv1haY{R41
zvb_WH=y<tYTUZ^p=UqjG@ySCwrEZHPdmq9Zzs~AlSEjU;t*)$n;b~~VRy=Z3*>GhO
zn~GP~9pw`hieXOwNgCziJ`+nfqB6Vp`nt9+rn_5yOx*DP)5V)G3qYQc#*1P(=|({%
z$bN>?Iir?;d5>8rBUc22(8B#tn3-7`oCZl**&@21rKP0<6|VW0E?uHn>_q-%_8_GT
z99{}DX=cPnMv`Dvluqc9#WzJ~=&AFK8^7~Q-+CQuK@JO~K_H+*qA-E+9j#kXq#sRQ
zC99fr?ng3>{p18*eSsyl!eBMJ4ZX`NsVVfq4iigEs|`hr6|~t&Uil`g+to&SH(N_}
zqFDHB$+<PQy>dKm9%+s=+nzdQl}mY`8)^M=G`vnMJi@cDTVjr7Kf02(xnx|Ej+tre
zdOUlaPo}p(K+7M>><YSL^yTvbdC(V~zDhGaBUR)nAxS5Nrz2Y}6gT{G!I=k%bpPmi
z(n)|eQ5f}b5$sWAK*zD(`>qQ^<!XS6d%)kcI#XF|(Ehgsa+qL(02(q`5sliQ0RppN
zoOhefZYB9$?e~LnMU!`&cE79pFa2Hi9VU1|S>s@6JV0cjY%Q6Ls$I4{hBbTD)s9Y9
zd7?s%eqWt-&s_ff{iE~lw*ZIXQ}5{)@rMqXVh^e;W~0@T5*$ovjpKw(i*-xaai2_s
z3P$tD@xwK9>#J#}D|uG<t)snyV=EJVIz1&@0;?VU*E2k*RQ6iJTDCn`*i7R1rhXDn
z(7f!T%n@y19p@Fy^n=p(3`!pvpM4TY^eIU51dsjV#fy@Xl8QhTY0>$Q)z$3<o*+=(
zXW;x&H>9Xdh4z1q&khu3LZa4M>zEE>gb@og63}IGk!JM!XAA^TxxR6=NwH=%tfz$J
z6DBdCxCC~=INQRbx=^Ri&P7#2wEoY5#6<;#=(JhaT-n{%0D;8DbL{&Jb-JsuYWr2P
zAh#qNW@jXfjcl6%y&dQYbaq;NdXP&mxZ~LWu+TAkt>fBD*67dT8N*W)fNP=QD{rr*
zHH^r*&XuawBm}jes0w$%jRf31pRT0s`<NoOr)QSplwFe7Vd6!&rEA)(vadt$Amcw!
zM_S45ef-MKEmxCrCWm25i)s~5Ip;mkQo~9cQ%Rr4dVNw9ZtW+2)LP%>e&p>^Ie+%%
z)pD1YqvuM;9dcTPS%s%WD88p}xGlDLNk6W+T<x2a7Pv6^FVNxjVf&Y-)wFmU`L(L&
zPo~mxN9wl+7GJXD{)qmU%U$1-sMCpV<gxNtKWHW`d=i=;?&cp%t&NE{I0X}Q3UM0b
zEV6CA3oZ6Dq*~n820TCa#;AF;RsJ)90cQpvKa^7-R^IiixGK}WLoIgo<zqcTY537o
z+A(}3!mW4aEpS(NtHLWPn7M?<j$z~v*d&04;os*=e3zuxl>)&(Zxg7ExwGDQ?Y}}9
z%t|DOFnNSM_-g2<v?TgcbfH-`-QKMZ&4>abxC*iW<mFVl5^smDo9Xd>(8B(u6)3mr
z<S)7h)jjsZU^Z_65{-||kPUQFQFeY`mT%m8|ENdIaQWhQme}uAX?>g}a167HW%1uX
z7D6}*==X7-rLBU*4b9;@_|-6>OgYA~rRwVxG?@4e#I5Azx;JVplOL2^4zmsx{Up`F
z#PJ-5YaDZADBS+sFTIhxxfi|>(mRE&O@OQ*_(E-srN8HCXX}woB8-_fq7U#~=~N31
z7EKRIm>fn#5dLmBd|oQ1h4xqwZFI&+$6jLFSd~Y7($!S{{w8uA(x%890-dZg7k2gm
zAD=@e1;Jgzh4vh4F1BFs(p&DJ5q{z;xwI3mY+^6$JdAEiacEdDe#mPzA_C{*(4tbl
zw7CZo*r6pAf3{?a8YAC{On><G<W&5r%6E`sYv~^x9OSn-^Q2=*e`fPj)o%fi93wp%
zUp8w=Vg3!)SVTpIpoAhAtFt9^7)dzc!SLH}@i|eW?EfuO)nxFiG)aMhg{23gd^aG|
zror4o{|U~@poo^^&u<qYjn!*#Zthe|phI-js*Z~n%1nlbrveCWXw2Lnw0lUAF5Dus
z#VgWs@qzrCwQ-866c2*y0h~tb5dG5C)wLL@s%*sX?d_ehn)tEHTVKkcH?x%<Sgm?l
z%z_YpHkx|s)%9DoTXq2sN>$qwVr1-IH&IOCzbgTwZ@Vp!%$KEVsl!uCW9Jn;J{XeA
z786(Jm=#)E%0N9;Us~sAL)Mw6dSa@c@9=?Q{ST#B;My3(#NgMDe++W)7X1F_QS59N
zh{;v65Svs~QkpB1z>eu1Y=?u7nSx%`AkPdx+wC+_M~WmA2yvkkhM?6IK&rIN-0F4|
za0EC?M*|~>jtkVhx`>hjIx$b59)-xJ1f&O0lHkJ#<dFdlWh9dH#=barbvDv6Coh|o
z3=)2=A_s%2!=x?e*piO3)$kdJZ@r?ad8b%&aq1NfjO;AyCC89Fk1cx{2-v{EF3@74
z5a)t?6@<(Sa-vk{!2d-Droi3$tpERRGMy+K>=@XDgn?N{k_4PW2nK&A6=kx&y}gd2
z+ug1Z4ML|CvVp`QSO<>x)giljN^IZkKzo&^yU@=3p-@?5)t^9aTcPl`|3k~QyLSKl
z7Qz2`cCHOl#oTK@ZU8ehfm;?$!2z5Uuz^h$2TQSi8D;d}vos4EHY!R=`kp_3{vML=
zxFT?h$sv4+<XL_*{UPL{y)&xXnj|-r{$2_X{#ql?;6*mXFWljDRZx(kT5}c$gtOUE
z+9Q#dwdO+otLxC}l>K<`Q(O-S3FK9*s=M-Fmcy-0f9fA~))0QHh$#s<IYecLO3$7?
zEk9D_Nj_29@csKcS~$8S8HBh23{!th911XhYGkogo=OE91|6qG&O0UL(2v%0Y&F=L
z3+^=+VA5(Xn!cjhctsPyRYa;dqNCVL@Wk|gO-sNL<LpFL*Az$|O4z&PNH&uBLP9_$
zf?X7jPpNG#4$*~1Lynlb(5k;9$HtIP1WX@gN5=w`^Ms%>I?e%y0wMh6{!AF&y7%G(
zuLtMt&|9aUw)@Fb^X?Ow<@xj_C^5=7$k~R3g{2UXb9Z5t1nb~1Qu+d}n2DZV44ra-
zkS&U4q~t`MaG^DLWmt3ZGZw!7>N@`NgF*e&zoHQ=9KME-jsTQAo2C+*zV5-f;lKrP
zD^CamqZ8$jUUbK8fb)n>MF7Qj!jm^{A!#+7MF&M7&_M>U6w#q$@?F>KFG~VQ6`zuh
zQP?HfdGl!BqhPp4VV9Lx_jVi3cRT%aY=Tg0(iQS#S1w&LaLxg;PzZ_OM|L@WY?0va
z!10yj+^TI!FvXlEAg_)B`a@g&un`HVg$Hl7w4!65AZGXdYo-jujW3G5qPa>Fa6ybQ
zF|X#WTW$b-`m<AR2YZ`vE{{tpjF}DCJU7BNDE<7ssVQqNrMnDD>l$d?QRp<TI6EZK
zpKd`uxfzHHgoFiQgQOTYhE0wKig3d@r#rAic;J!XC<F&~(0)@&T>K7%UQEECxOwOb
zdu|tGx1b|yEamg;Yj7?p0*;%#2X1wll2XSfJBJcXUGR3-{9%FepbMh^Ch@J!0_B9X
za~W10Of@VySn}{5k{89IX##kA`wYcB!E^UPhX5S<jl-|OI=W0uYz*fslzU||F2mWE
z!gtD(Y>?H1UA1}>1RG8;Uf~>CWRy+W_jxFUHsm@|6fsvNCgI@br*1xnNxxEv!pMMc
zBb{&3+ywH74LdvgY)*UOPmn^`Aeqkz!~^2e+><F$HlOyMnMsesZhU*`e>|Po`Sfb$
uI@0lBVK-ElG4dBtvxi^*NAmCIgHz&6E!pm*O#SFP5~8vXbMF8C^#1`ONzcLn

literal 17276
zcmb`v2UJvBx;0vuI4X)5$%r5b2uKuAKt)24B}zVk<QxhlV*&*MC5T9toI!H-NRph3
zoJ3L;NubE#eVb$7zOTFQ{l|ObcZ_aJsM@vnT3?vooO7Kwit>_YPEnphp-^YgQV*3-
zsH3y+dijqN@c$ug@<H&IpriO>M`ashM`wL|Bb1!JqphWlqotX_RVO2R2QwRM9*)}_
zd^fL}Iy%}q2yt>+{nr~fZ0t=qf5pUTz(r2lN@+NtP^9|E|A<mWQ_N7P2Xp9$_f?)o
zEsVHnhZ^l4{Mxw3ef5Z66U!qsI|Iq3SIh5c6nPyAop?qF8C~X7W@9a147@T-nYr)u
z7iE8-Dadrh^nT%Qx`EG#pJoXv#r>c^e)sy-Q<M$|`?e9(5zi)W-TI2(dfJjOU6oxr
zU1nn<G8|i)vr{JPqfdSW?g@p$G$mx;IedkZ>L5a1*q&xY7664RAv*7P_+l7!Qt0r7
zwf8Zad#~0PhEj-*p15)K>OH0qG@E*MD5J;0-UhBWBi!A?!-zYgz-p*SXn`3Pw6j0Y
zu(!%%hqby$Bj1>BC|+7K_s?gtOd)4!?><i1L*b5|cf1$e_{1KUXWaU!Y|%WLoX7aB
z(|mutNvwz3QsqXB>)LF*{}oPLjN4-8Tz3kgi7jWga${7v+-U(<mT~WeimQFOS$A^T
z;wkTg?FP|(`jE!NOKfVPEdGIk9~z%@_33%(Bxk}6!R@JAIp!&~Mc+w*3-!&Fwyn)|
z!`(Wr{A?7$A64wjPRE}`Esn$0^Ii5A%a_avxL%_oJ+D1N;Q)KF?f$Ndk+CuEV6XZh
z<;?|V&EkX2$si%e*>}_Ju{m~=^}YI}M5t5`kL|Vhj9EJ6YR)5=aEzO$XO+|9a9T!o
znKT}o)tyy5YR-Pg{^#pJYQfKB3=%51O87XZf1MH!jq~_$slz8OLm$Su4_elCb8bYo
zZ;#Qu<2CQO5jlhJ%jtthXtu2GDs`CYDKJx3cgkvu7O;B|+*9G=5V_D0L|0r+gYvGw
zdh@0%85x;oxl>l_dP|Eeg-^WOVrh!3#@uSx&ruDzPrPZmCPY0{BzDA6j?L`FH&|Jv
zBX8SaTW^nLCd)<Q+P(Jk6J^AHrYOYfPV+4c75Bj=S`|eLI!J^t$tgX3TGC>iFXS|z
z8o^^CwAZ`|JC*FYx82y@uCTwqj~mEO5!oC+8Yv_b&Y>zR`_|pvy(e9jEr60wS~7^1
zy&D@Ox}Q3>x9EI|ge0`MxER@B_))*6h{v`gK9|6voc#R>x5h@AUbSa%BHXzoQ%K{s
zg`2#5U~q77PqG|UM62;DA0PACj`(5c6D1w5PG8}*8sx33twkcTs;UY%RGc$1Q3z3^
z2H(%;<2ctPo>R4*Uby?C&L3G{Aq<Yu#>S@5azF`os}q*hYH$mek<(Lbqqnuao$j(a
zO)2D<x;{TJx#Dqr7#%C(QGmg#MT>fkmRXIIVag_asCZ2~u4d^~k4`V`lwt5sxFdSX
zoo(<%;{KO44vdLUoG`C_O(q2U+`7|WPQ2HXro0_{utm4Evy*{b9fI6*bAmWRmWlmw
zZk()M5Up^!evMBWB#HS*Ss|_4JXcs~s8ledNh&r=OD5R3<)eOO;nIpjq+c6_^ul1F
zO!0_QB(k=!Ai}FHJbbQ%1BhqW`MidnB$+FG7Fb!>4N6|qcko5N+!#H#1-=p6nlt8u
z1!<ZErt!4GuKiIwmo8p3?Mg(Ok5yH|rapiEyr<gB%WB|DT{4YHdxrO?ua_L-&dWy&
zCEb5<!VI2>lFvN!`HL59gl&hJHe`o{T-Q?f*HT2&Aj7sHXF`L6KhMuww#EpT5BG6C
z&b_xW;v8#TGI1g)F)<aMe6S3Y(be5e*cm$*)nsO2K?}RBt7vL|vUhMGzz)E}W01EH
z{5G<6%8@12@gVfNt~4@F;qe(>yOVU3{FWd0x7xjM+2xB}BhF)4)2)&C_mZ@o-`_J}
z8fctTVF^#2J&WyC*UOrTag#sTsXoYv5q8rIs<ZLh(Lz>^RCw*(V6jckWJ8bzq_~j#
zrfnE=36Wye?yh^~)^sGLpnYO}5M9pm=SM0DSS@K;*`E3Us@_5ib#xcR3ZMN?sS(ee
zg-?PG$&6<0F~XX~*3q<{oBHT*&yO-PGTkM1lNYI|GT>@fg9T`<BFjey`+L{XdY)TJ
zC(ckNT4a|>bf+lvZLQ7_1}$>ZoR>xizZjh!!|+lf(df2RnR14T|FKu9E?n1@fJ@S?
zbY<b>RJh~3*b{Jj)RY}RT$-<0WXXbdT^}&v!{AiT(1~bI)(3P$Ra%29&P_+!6m3O%
zT=%d0OfTkR3cr$;msgu<izakR(wVh<<Z{A?hlbwZ;(BCfXXpB>?h0WlTo0ENuB+sG
zmL@bRDicD^VsW?>K2@~87ye|Sm}iEhD~WE-rjnqxx6;gsb+xJZ<pa@Yj(dj}J=h9T
z%hpaqQWfe>Q)_E4KBqbxN`l$O;>fAfr$Zd)io6f@#twRyJU8wL?=IJ{va)tT7b(=Q
zIe{1KK1zBw&Zzl=*^k;cesAAGs-zS+Eod$b<V%i}JBR<=b-AQyNP;qw{dnuIk`;Vs
zj-IGkFqh3}1vlY3MOs>#QBko)UmMal2wU9>qE?&R+mr|U8`V}g;jNz`J}D_FcALxk
zz8*pj({CqcW=v}9>L_N~V~bhxAK!#FSP%8%ZCu<{23FRl<TmD5cLOPID1*i$PCE<5
zR61pjTz-CjmV<N9-8$2hnej!LnK$A4EwDEuH+Hx2IcL!IT}d)oL!SFvCgm6+mG(PJ
z6^2sWRR{ZS<x|zWEB=qPwJq}(W@5egA#(zD7R%eq7AAb@_;>gA0w5;pCfvj&C8=Hz
zlX4V`>XbY2K^DGMPLbCr-dnCA_3`!f<#Js!HE}v=dd_KK&`63q#&g@O<`&5$6f>Wk
zoSZz{i3jjc1_6Qok31%U`s6<EV`FWey1G(AeBUOixdk=E6!~Yn$WEqQ<Q=E^TiIne
z*2o3J`nOzzg%&0r`%v+_-#uxG5gsbbN=&RRStuGJLDwVCv8su#S5^=<S`~rB<S{TX
z$ZS`L5|~(auOD#mL5DC=1iwXMwU|cKbNG)x{<y_Rp`oGS6l9pz)1x95Z1<z~q%Q`R
z?`?4K`A94HmCt}>(Lh7Im@nn+VOz1%r7P#o{fX?WzKc;Cl$&qooa4?zA%a4mzPp_y
z6P|7}R%PD#`IUUkT{90xM#j{{#D`;^J6Sb-*}9dQ$B!RBEDZ|_R?5cQ5sc_aeygru
zlMRIN7o>wd6nCHI3OO%kDv0i>bR|vb>v<5ud`RiCA+0S$k(yRp`w+TYj<i@YMDow6
zsU#?xv#?LuYj`4+0EnB4iV8-ZvM)~CYi3B)<2n<3TNPWO;AG_H=C;@pn)z%-wE>89
z0m7O8U}9ooF;rwldB-X9$kC%&Ly#g8grzaPgYCX*bJHBuhVvewF9QM{>(=ZA#sAXf
z%bF%;OAuF(Zp?FCm2=mxU&rpWhjZ$2K;}-=_zp!~c=_^WDg=X*%_Ekun3!xPxlg9M
zo4>GDw6wIJZs|X->ASo9^(cS)&iX>e!S0OcrA0%dHi~^hj@Mjrq_vWzWhQ};rB#fj
z+h6uQHP)bEoUcJk8?N*b_9I#dBi;ImuN%{+qs}zCx7Mw&y}z?WQ`8G+!m2;_enTKl
zzSiPABwvVB$1~}YZCpkc^IxiYHe(`5DJdBkJeEJ`<;!E}YWWJ|_Gd?l8PSg(Jvxi0
z7IKsdb{xU<g|Q%Ys?Z&3A2*2@JOg`ocu!ByshuCM2Q0BI5j=zelUNo3fycMU-1Bvx
zP8^^5`BSBg9Tk=j0X_0`G61_%uT^ZFg>0XTiwjnWRAe(QyK-Ig3?;7=6xL3UjggeK
zHOEM&7*`w%f=>E$-LlFY=RO01FogxduB`lMA@u3xgfJ_l=jL{ZyPYh#0_YIRsm2nY
z9u!xk-{l!fqKi(t&UwjPSo7FfZ@fi<!S$rIjQxPd6uRTQ*%2>x>dYB-o<~-5JGzZS
zLqo1XE@_w9H3bNUuCUZ+X@n9Xmcq<xbSqL>VcXPoUBqE&g9j|tGSsqq$}kpIV^t$e
zqr$&xFPLL9BMb};d!dZTLC|m8DB@mS(d__4dWn`cyTAXjT&&2*wB+f-)>|t)d+)l+
zI&FGaqq?-l&C5tA0Zu9z*@ow|3fobmND3*uyWVdI{i#5!p_6XQPxccpTFi%}vnh;?
zur(8F0S_V#{YYKUBguNCOtrKfMXgm}`o2DpCIi5RSyL#>sZ*zdA<(<Jx-vx<(MpPn
z?~1I3lG{ZOv|~NiBSd!Qc`8@i?obN57F_1gzJV@-?uZ%xO60NDDY^Z_pK}d54K7Sw
zx3j`!^|2CJZh}Irh-RUA7yx+v%7Y4(9NkJ@qo#{A!mcV7ec7syAAf`?gc2RTe>qZ?
z$pgCA%G%n^ZtHyGR`blD^+D$kZ0!{3KT2m}x}ghT#ymFW=2BwS{`@oU?%lh*W?hVt
zjkDR+`>qQk<?6BSD{l}W0kuon>tJV$4<I1k-oD0{RDG<<qtKx4lm|4nPka_1jN78p
zrkcYoI^ymZ7Abt4f@r4f_$+=JLZ;q$2&#W*bhK4|mYDBZA-DBRf}e|#<7~(M-~xvk
zd0($TWK6Wj%AIqp%NO_~zf`(yXg13zoc+XWrj%4A(A_6V=b;iLvYm>IL;zTj9@_O;
z+~0h+5+&9AfxVlrPdfp=Tg%4f{gcWvj36f`=NSs_hp<WS>kK+ykcy-t^9dwJVf}}%
zU++NyU;#{I(jL=`pm7*eEQU+$`0b{pMfW!oU}dlC>FHfrU>W!1&?EJLk_<NzTzHd-
zNdgkv0-M?CbB4+hK+xl6m`kqK+`33c_Gpf7cBs%IC8v77ti?WOXJaX6#Z#O14u^V#
zgAX*6flA?e2;Fq3OLJH)n_exON~8QeNKf<L^srWb_`VT9BxM_;ZkUO0$44OP5Y!PC
z9Be@Ig_e>sotVyD5y{LzI??RKk#amFb#L7L7f~WPN6gQQ4Z~7HwddXls{U0^Wt1I9
z3Q#2?B4T@YwY|B(Zn}j6DXeG548M94t*ovp0}@2KZ4*h&J4V86r`Y}b_dngXiShII
z&zeFa5+HfPUoVJd${a5(z>L%hKtp$0^4MHoVLKa?8waCpY`hrp*4Ea*8u7hW)=D8Y
z%F~QyH9lvWT3QOyoN+xYkuy*_u^OLbsuR?)W{{9D$(XW}o;jo1m#tF>r6-b($r%PV
z&Z^CCOhiPKF90Zi9~#nHsy?U+q!l)II6h*@pW(K#_z9XVfEz2UXf^coskyl{3LfKr
zcjqLTu_?A{|8N<AjdlK4_`LSr@%&I21RbVTTtuZW?6GkQs7yyq1m0Qf*e2i$?ItB6
zS2o7Hs+;!8AQw#{CJSL;;dYqOcG-mnbdiFBrRJ97uOH$t5cHG03Fla_fIQ+G6?D@*
z-ouz3ca-l>h!NNEXT)nmWBY1*I~#qZqPuaB!xb%@rOL^&oegwenb6FY4I2WHfedcC
z(6HewyH=4bF)8iM)lW(=D7a3Asq<P73lQc9l&cT+mU<yU5c~lRGzwBb6=^)vEfLbv
z(n{5yd%OUF@%C4d*^nXVu7Q<JTa+NKJ0%FIzR-X1pDeN;SzBjw^Y9P=kqzi56kJ0P
z83quQ#pdz^g7Q?=b&_CCzj6DvI)lX9R7W?IW#!)bkkKMVc%qhdsd#fZXP=JqND}Ne
z8ylO<>U68}lP8I7F~SyEx8XjcMLaZ-323muECDiiuCPy688#3DEmkJ@8Uq6c8Ehdh
zo;`a8ZwbLrd{%-g(*==vk&Z4$bpKZn)(Cs#$Pr7d+zHB~ke0}UA8dFXFy{|se*8!t
zxdLrs4$?W8G4fa?e1q$Bgb6a-08TUS&(%kQRZZ=KTIS>Uouz8MAmR0oBOBB(n?QXh
z%q9exL)*ha{}b8YG(_$-crs&R>esJdkKiFb+|>ANkyEJ#_XfePq@ohqJwt+&_EEPb
z;WN|%im-pVn216oJ6nt%y${#(EN$5%K-_Mx#vUAQrGhX4ZYH=VL!EP>KTiw^PJ*;o
z8p6sjGCqVdrDkiJizG|+-hvG!0HNL8T}Q~aLbGm(>Yb6Li!?M@K)B2`b5{32?Ah#X
zq_-e28)v@#^lP-Brg3YhJq=W&`U?to^$W~<n@nMBeE;#I>HJuAbqluz;AQ|{64KIs
zozOsQnwy(Vp<Hyr?M&8P(c|UNE>YAUL!LcBJxA9kC@5%RY00LxzMhiKb2DBp`nt>q
zas*HU)FUG&H{G2e^}6M`J(t35MI&JQl^og)+4=JZQ1NRK*bxB1+x@u6GNMr4=!3=`
zmz6{={cpz-U;;((FBw2c1kxXDHi)tgbD{#nIP#}w5y(FYjf4k5xF1`M697uo1GrZ*
zHcN(5M{)E7$t`IAB(U=dFuK;q?yoU3+_^KPITK+L!v&id05FeAL_|aa#-;$EIErN5
zx1m%8j8?j78Fs#-7kdlCNgtM;Xc%(BkG^ph=%t7+Q0u9JNkm0O%VntdVY7I**X9h7
zfdpuY!z-P<z(fwkqYzuVQf^5e{y}_aZ^_GMb8Cx&h=`~jh?W2VR)DIZatu{q`}&Go
zc0O>JIh8I|?X<mq^M(s5LetIC=@!8kFu6e234k$+Vsv!09_C%EyMW^Ap@6HV8DZyz
z?+Oc_sVOOF8B7QgF@NdfLy0`i(r9G>a6$<%x=TPJXIfS6Z{zEs*;p1>4s_j$Y?Xi!
z@GT%Ha==`$%Mx~Wd0f!K8H9uemseMLfZz+5oi&|Y@;Y#f@I0(R$3pbAxg?GOL3pn+
z-+oFOu%8Y>KDllz>hz5eDIPzh6YmK-0D%_SYXr;1`{!2_G8zjC3Wn*robYxVr$vay
zxEPn&cwa|&TU|p#MoGzIqy{*Td8EJ`ANZ3xwV~wsS$JCy51sq+GhNrY>kDJmdg9{Z
zxc<C^;E)h2CdXrn2k+vJB2?x0iO0k=$P3i%dxxJwc~f1zPmKKF>zP~rZVZNbj~(ax
zxu8`vKH=`gID1R9-XZ}y<OZ;3w6uho4W%D{jMKiRN2O;})xY<7d#v)kpws-vZ#6ae
zGAa~@!_V(zkW9(h<rBQ{M%};PSoilEf32><%qugyw4~QiNkmx;Ou#E&U#;qJ!g0G}
zwxqS^P`#>plTu<5&qu$9p1r=hu`v^LZJY@=bWY)4u7x~;byBT?sUg;khVzEfyVzm_
zh9gsR_m2?8O)wzO<2@ub7Li4f;ly?)=X)7D?^mrfv1@#{SPRa&oj_$bsCaO!EEJEK
ze}6}hOyWR@0yu2Ws&?P9M*t<v1|O=ctA9*RB7kwICql%d4A|LLB)K?-Ps8O0L#lYR
z8vTp<8E1k+L(~(0ife^d@stO;o<J48I5r{<SG@$q1iG7m>cNmrwJH=o7(e!85K0l)
z&UA!<0OpgQfUrPDCMH^R00lAP8^AT2tG=o0c&l1&)X%~;!f)7X9rk#>7v=5qjDC@W
z8~8Oq`rqfa^G(`|T=s_@+UFpUyV6xNda|@`<JS>Jm7boSuU8J%s5qAk?-;>06VAhE
zG;_%eR;$k0RKtztsJG_B=ZncVFZ%=1y$Gcy-viJgvJ${N!c&KR_>c^USI9C2!mX~Z
z4m$$@!GY)T_IV?4Bu*$QfWF~0HS>&W8ZnAMCN@0M<Ih=@3O8g9T?}gl><L0=0}SPJ
zUeq=+GD?GSmG#CA+kSZfGSFQzgO-<<2WLgRPrUqnx%-}@@`W>&!l=|B;NWc;g8YF4
z%zB*n05FO}`aYnkIHZBXlPW!U@Vv&=q%G<ZOqgp>dhkJBy9pyg>z;*YNlDdpb!jbv
zj$m$zit0h-us-l$3`8xa#n-(%?;kJ1Q>`Sg`+j7nLoWEik&43gH$#u@&!o~5t_Sim
zk=Qx&05ysQmZ|`NhJA<w+I+O46q&3bpT_iQ@&|6_4`_Ff9P4;}P7<Z!o6l9-U6a^D
z|6W?RK%A+O3C$_jpzQlFIO?cr2l<ah36b4pF=!QYHKd|F<A1$i%k)~G*MK^HFFro=
zB*-eb>SvGQKix?$f4*yVVf!7*rj@2Q8q-0<`&9eB^Ltp#LG6LIrlu~~2iT!h_>1-W
zb#dT|d&`%q*yg%&*SQSqXQ95SYHLT?+uJjup^%;W<BxZY!Z7_%azA+uH9iHv4GSNi
zYD`QF(05N3maanVo+f#FN_8nPtH)^jN^s(Fv~X{+vPCxB_r;Rui`RfGP%5?_$&^Wy
zi@B=-?WnuJteX)%Ffag&U~b_)l}$xF@H65Um}C`AO;ZWK01j#4cVSRQ2tuGAg%I8i
zU{nvlxLzQE2S)@U6t9%sr3ec6gHe#U!_@Zph`4p>)a9_-2cyrApSFz2(a2B!#(#xV
zSG!gBP;42#r2boVIr{Ia%Lddkb8X0x`pj-2UB$$!qBcw}Aluj;0X-&8B5A1w3)vGE
zt^Q=>GUa#{Y*D5@2@0in)~5tMD7)Qed7X36AS6l#iH6`SR*$yZ=7n$+urfPH?dW+8
zu>zH&dGts~(vWfHVAgs6$?G@Zz9@M>`MDR+{}EHx61Lh*5&UJ<uK+S^0R#IPM@xVB
z;mPBHA*<;<P;N~v5$TBU+V}Qy2|Rq9mmi+Ro05!*EVTn=Bd96cnkbWB+8}WZYg`N~
zggbM1_2so&EEuUo#xQi3Ol->8H&29-CO~|6DdRs?Ql|rP*;l>tTi5>?GALA&ovMyA
z3e}POKWr@urQ*$oe7Um=H!T&Jls$Be@nio<$Aa(rkIf;sg|g{J9{J!$jAp}Wzl%R-
zbc|0-uQcgALi9jX0MW@Ze|NyK$YjZHGnxrx_wV#&+W`J3v-4+TNTEd^2NZlfr`MK0
z4BHmUki~+?2OxS8<OKt~TwY!tA9iUidf-!@c`+<+=Fo$q6=6b*N=FkOG_!}ROd7!p
z3(GO&OTR|Hk1Lnb*N^|4kZ?E^1JRzH>k0Vt1SxGMG40cP2ms{f=0*@aNF%Jk>;@M+
zX*`bvHiDcXag5tzvJk=xMrRnRdKVr3j*e@SehY#w;z(+5j8Vtmu4?_Fwg2@qyVFpn
zyvE9BcF5a1Xhi(;mUsr$9^tP+3J}!s8m(~gd;PjYhO?3pT?QBe(5hzTN)sFQ*dCC{
z*v(&)Ft0rEXibUrOfMeu)KU=HdQ!NpWx&O2*t^KxSj>no+B#6@7p$%rj=_aSIL<Y>
z9(+INQ;u|R)8&thZE{bHjaljG|1#DxF!&6U#C7x)4($#=WmdSSPoLg^-)J_({P<xo
zH8V4GcYfd_X3+R`sqlvtZu2X*em8-+K@UR|e^dYi>Q7RxW2h*}=Zj3p7En`L=CcFy
z1hrnsYrg`>#WWzq$}#Xe>Y_+!3Eti)Jj30)!`79P$ETQQFpbeWCwNb^)svxeeL**E
zW4LcKuB4<C8Xj)qK`m%sYh8I&bV+kDVs*IG=APlWB{dPs4G_Wj-NpOP7obT2Ctf)4
zAu^K9%gbwPbMt$5w=xteCp@rw0+xN2{LPm9>YMv{WY12WSyqEwt#PXnmJQYq^;&u=
z$BjO9vZUZArFYrnKd+k|SN>*m^QqC3C;gUMm9B-*%C4hzb#?vTyva)gXcPMWy?&N5
zO$;p)KG;NEzQK5V$<Tf558tWv{c5HrCEthK_o<<dXuhdaoPMUfZ94SGh*lsecrfN6
zF)IBT+boAY)NrI7Y~O|93s_BMAYlfJtTIAF57``mquGpe&xmO|#`f253-51EmcftK
zph(Zr9c=oF?5!zVPyz3T{m5n5lORc2S6^=q;|S}`n+0t+T^$H*K4%r@B+m+&(ycn|
z0)2QEdNj-GwC_}~dBxl7SO~ZLE@xwl`t%Y03u8X^@<d3{aegls$&Zi}g(9(cUA{0P
zN(Rz@zvSm*0Pb73Jbd^P`DWpj?+mVlUiDs(#t=ORdmYO-2eVZ#=p%?y2t;ElU=V2q
z1x_A2ZOnhFBzV<%zrRyn@jc`yq(@Rq*3;jK;vk7(MiY;(OLge{1QO!Cz5uOK;h%pU
zDbE=Iy03IWK1$%x1!nn7>y0Liivcwi*<+?h967_6j3Bh9Sn|k`j305aC($XxSi=+I
z=qSDko<J4X3nQ|7Y9GIGbP)yC^SYRllkR)?M|Pz=(Y;HDxHeArj_$aCV2(842e!k0
zk=plsx5A=;7S(g<u*mgT{2fIc`IZ(eAsnsKPy`=GjvVopK17T6DyQf4zi^DW8U^ca
z-^EBVUZ^Fqb^H5D8z&@QnJ=a;o?5v*WJ!8*+4v0~ig2Pi6AF6VeRA?k>7z7SY0Yb#
zjOf$BuIV-@SZzUKMH}1W2u3|qKh^lu^5NA+L2uVr$Nop){a@hxA6A2ZHa#o?X742^
z8gN0QDG0oBk>z`Pq)|Vrgy!(Mc%RPQ$>#%z2=Gp$_|=K0Y|uL3hOzqXsBv|Wvf;xo
zKKzeRwdn8*JN^uNOw#(eP|GL!OP3Xo+Jf4fI^8p<RGq^GFwKR<=5UzWwf9M!uefH_
zBBbLMedO<kgzET{hSN4GN<E1dRdOC$Kx+h#B+wt>2#oykqX0-8UyKcO`Uu%%?jYbS
zwD50f?i9&xeSLP+!cZMn2p0_{f-;L3m0f#tk>h_Y&PF&+y((>}cDW*=7+fe)^xK6u
z`g^DZY~w~<W@zvZ`2+K1(MYf~z%+0PT7Oi<wepSlNXA#kv^DL|pm3Dx9XhlBtiy@l
zze}}7-d2G8=>?2#RqP4r{NTZZ%pjOrGhmD{I`5c2V5yt~@^xX}9n`q}^~+)r#PV|q
zR+3>Z;}=8tiBkC$NJF4-NSB4a)yrOU!M&5hiV=@rMv&C+!i<oDBhwQT@gNI;Q$na0
znMISp<nX<(ug`b~8sNo?7ZZUypDh})?&fK~oA7~M3u}5n6jzid!G%(%fEX)HC}Sf`
z>NK@-7Fm<h*Q0EC7@-)g{6s@Hxd9TFu^SK@p1Z#qPU%&{qAv`WDuXEbp>-V;2pqIU
zHK+k+sRiOdlx#=DZ)hAq&1ob@L#McO<zU<QhNx(@pWCsxyz~G8r80?MVSUXqeK%sD
z%`tb}6Q1Ijxdz3lx7?5o6S4n_(d@pG0%4)&6Wa`G_!NTwktgENDZ9bXk2T(TO~$Ym
zIkqbSB4KB)e$84VbL=dzL4EL~^F<;5>?0MIT}k8U?J(5(0S0mU4k-@rn?ru}zrGjj
zfU@?5u8I)53JMBS)6;{N?X9h<Qc_Z-IqWxY_JG7k08dNaw~f;2PZo%c0;$Q?F2Fdf
ztgK`PF-QgqR9`?{?Q0;z-Z(+Y-zJXw>H+|t3}_9+55d4726IISBghBPDojCOovL!0
z>q=tf;J~InN?>)5jrH6{r^rVunWpZ=#~SBHXp9o6_-9Q1VR7;38LF9U(88eAbO5G7
zcuaPnb==nHAKBY~fe{LJgb|HM1FekqNu#cKc2EK{TWW0(_BxbBxxjPJmFDhKMG6WE
z#7u!0MrvzoH3yhcsb-mv0H)fWM1u4`*5zV(YG<hEE>HXSCB8T>=^}O-J@;SY5$ix|
z6ngA9A$Nj|M!<xgAR&(fknu?`nW#tW_k|%r<)Q>SC*C=^T_^_@5yMTgbAcW@@9b*3
zXinoI(Y<Ikodf}w&dU*uk&MxNenhByh9Ewo{{H#*F(+O{Dev1dQ2V?kUI3m%6@R6y
zgF2Wuc}fWN)hZXl6=mA;?^W#o3()*uisApY%Up>BmKwYcoM3l>@zxBi0B7+WV9Nkc
z&GZDKn>O4cfvO-EA0NL0>?F*bKq}4PMxeNW0gU`)*=cr-FAt%UtC^dp7g-K`2J+`R
z6B7>b-C4cKrKKFezxYYu`1z+eE0&+{BU&dw!@*qr7eEz-ef*gE`SUgPB|zc#UZ@&y
zg}RJ~iVQlbKy68@HR!>Fqj(>v_hN`CzW~Xp0fM!<l2S0E8C+o-<}AYckPRM^s$ho_
zbh~%&&XpVa=$itL*lyO_{|vf<bnvLhhR$(fVvDJ!FiI|imljUIL&Xuv{DTr%>N`yv
zPc$k`b3P4*P)t7$D&P!6-Oq2YP7hYN<exu(K641PrO<cp;(!(N&A9^#5Hd|JFWa7>
z5mJH4`^yKI>sgqY8<JZA6~Y>Z4Kcye%p+d`T!>|`1G=e^O$nf~^jL(Br0~%Wascb!
z8$*~?m6Se@kN+9D@Yi2|H8nQo?R>&#m&@B!?-hw`&oZ!U7D!$87q4_Bbb-Rjh{pD*
zCb#|r&Z6SDUo)W6YoZOMelsLCx7_B6U{>Q9rn#2|P>2DzWF+^>oEFj%r^VLF6k;6G
z1+yA!am0TsJty2q;<&B`qEgQ~2hvmJ%ZHcRdkX${Xkpp|I3`gwLk+Yz6(H?f*N2C-
zeK$VprGdVW!)8UUFO6~jeR}}@rlzJe=p+NF)8?~Lrayn*<o91kW|#&rMYJfDE0B6m
zU0hgzdWXcvTEQd2-+n<s&glVwu@SO5K0cg@`OWLsic(Uq<st!tBf(O`-)fu>jtDDS
zlp;kf8LD@pjf)ZH$sf{Tsl=$>I_R0aE-Oz0D0n1<UDp&r*hU7LZ{Hr^GZzyigAVOJ
z1qHn5E^sR(MOv4kY3?}nfI*EHd>S+$Nkia=IGXuvz%$JO!}|zo9*AS{;57kEDhow2
zBVz<YCpG-v8tPLv{%g5K<tO6aU%T%d4_cTyvwTzsPz9kOAqk-Qu?h>12sI9YTt7AO
z3xCxo)@%PTkioBp_^d!n0L{!9B##2D3-m1zK~oXM0I6Q!D+`T@;jpps95@c&P4ah#
zW&K^2M%Km-lFz;gP4R7P%OHxIhOp|deRF}&?#IJ&BbYSvT=o$&B*?EU#lnyuDKW0n
zpbMzTVaOl{Iw}9Po9fdzqA+0Kgy;<z{F<srSnQY9N%o>gYrgAI_PA3x-6ZI(*}!NA
zzc*duP^dR#ID&2?>N)v;tr)?DQK<iym-O%_|AC<W|5+Lr4IHLh9)a{fwXl#?QX*hA
zT+*pO0zRvZ-!>9tLty6OCV;tt4am?{a)4AB&VMt;27omFfy=;F1fDE^fL=%mgSs@7
zRc+i#89*(l1b)y$Koy8$4|2=wY-a*L;lx?$)Wbpu99U++$RV>+at?}T<?c!|rKo2G
zXf+_@rGWu14bYY=@COT6FYwGFPL*_+MG&cI)O}U%??&C<!xPzh>g35h5xC(;REIM$
z%9j4T<CRj|iRTFG00;}uPfAL9ossbd4^MX*KjLZyR*&`8t&aafT1dHRY-|L#V`vRO
zNT^}p3tB~aHyu_3gFjFLL4ZAe55G`D0ra0es|I1KY;_mR3ee&6(JRF8WI(~ILlPr?
zx2mmaD#U*Qe4o>hl#Gn>y?fpSGFn<)LiCZ28)QHi68|2wlViO0UD$Oi9<!-sz5|qu
z`9Gi~kk%?H?&2rG##dNCjG{dC;}37#RUef1Dq_p!RmsvUM4SrX6SOLtnVx24WK10H
ztJ?hG@4CM=bLd~Ir?R<Oyof07YFV0QKtq>-p4c23*=iioI^DE{LhW9$J9ZCqP7E*-
zFps}6Qoro7n$U!nyL&}VU*+<*)2<*xAx;~ReN{F^MEBN05UWr}M+fEHZ6sXdtl#|~
z0TECGnFL7iG$=1PVBggsB*UOuL`6|Y3%jWWJHl`+7r`x#czR)02z&oN9-3J?P?+H5
z$PDp-LIg4giBCy+2bhsFV3q{=%?x_TBak@Z=B%Ry7MCY#R+cvjw}ZhAjDvQzJ+kC!
z(k%~NI$0r>1JEScPIOLR;XsPu#mkqIkyr&Yh4C!;;l&*w*!iK7ATY^#AnZapK(*ZA
zA}~U&fe_HA*p9plwfK^L@$kiewt^r2>VHdLM1xV}41G)j!&KJfaq&|tt7?`H<L(N}
zVn*0u$fpzoZN}E^e;(~XBRDHIOoXy#_b@2}1~U)ccAbSeM+yXq8y+T3Gz7qramMd~
zCmn`a?%%Y8Tr1|)uPhz~3#r_sQR%rNfI@Wse0>KH95UTaH*HOkpa)#U5w7nSjpkrh
z&GJ*APxX+QcDyL?ihz7ubXvpV&tj31G_3Y6);SE{+k^bbd+-5r^ICt;+aD6@dSuIZ
zwS}Qs$Z8n*G(b%$0dqU(Gb{hXa~71-x?+4AZeNy-BL*h+&tL(Rw-(jER_nAsKNsUj
zf^2Snyw|?>-&e4lSEzQBR1sv7H&DyR*I@`wcWzL_aqpc*cRf#x!40KhO5e!LEpPrF
z3LRM*i*MhpD#M8ds|r0E4QohtJW>gBmO2~SxsEH9w4qRF^RHtlL?%nU!l6(bg>I4d
zZD8gcycY5#<h2^_-6|pEqn*cGnRGtg%kg?56&9#~34E5oSRq2)hqEP+GM$kVJIFh6
zZ&DntVVWC7L8}+SyiK^?nEr+^MU&RXWgGhPfJsQj#yxUsy`eB&d+!BWth>W2lc=Kv
z>1t7e4oIH{Ugn07kY-Nh`X>-zX|Q@_j%msA(di($r-MZXsXvP)lV>eT^~t5><a&3v
zRxu!H6Bb9bK&?`Rj&5$~jC<(9+^)**pjuw2Gi|-6FItj5`F4<YTYGeQ(@%blZB^s7
zS>LxbU?1E58g!qX(cE*_*(!^2pJA$RZoHGzax^7sF=M|xNEOXXbA26R!*y@yX8B&m
zn@P>+c0;*Bd=s54C8dccs%&cz%eP;(yI?4alZoJ&vbFhd?c+aku0}X700M}10Cp?K
znP_Jm_<uTJZY(frz6>%ZIv8=mA>vst<nyK9=me<Ta_6NF#m+U6v#Av!CTLYwh!tG0
zlSTR4J=>fwe(NtXgu+XHkPCJvS5Ex86g@lU<?vEupMGer{Xv|~g=mh7+1%J9G*-@x
zLP+Q7UupanKUR&FFwd&wsIk?<-LZ>lDteM@Bsof-HAXM}^i!EB`Jk@WIg^)e*Q{Hf
zDKJiLB1!4~>w7ETc2$&rF5O~JA8wv_<{IGzSmm(FcJHCjBHjtFouM&CG{C)yi3wMb
zao2u^D11osf+HN@y;DToT9_fZ+arm9n2-ShEP?v^`mW$3BlH^vp@U(Z1zI-q`02~0
zU^Bn8&hode)zsV^0{Y~L&<w}!=kI2*GQm!wVX4Ww>Vi7DAx44nF*z0C1>8o#jt8w|
zPQKD=DI<d%A^oJztRYQfq-ZDKaiLWYM@iKHhAEGf>l!)ky(&>RtQ`JSF^%Z^l(V7|
z8=)5#DLbe<Y7(EWP1p*i*0ah?Pxn;FUc<f4rW_=g<5wXfc<uMM&DYzu_5~bT*eLz<
z!m1Sl2Qf|uqSRg?t)x@|>^=f+QI{DBJ^jvA#V%GZ3m??$OJ7fNzuB`*;ZC!Q_UC<q
zC|yvDO(8I((dazu)#YVHn6&^$EL1V^@F>H8RW$Gcphg&*x^(cDFJBOF2~FO@1sW_6
zOra4Gb1Kgl!Oi<*f2ynU(+7u{CtG#vh6bg}{xvI~Sj(UOrEJz*+=LI;TcXHfS@PUV
zsG$7Xt~BrKmRC$y7o*E9n=IjSFdJr<_<5-Stu(lj<>!^^98mYBi{Hj(iL;sWI)Bc{
zQ(nCkg{ekuu^qhml!pSxSUa1qfY1d=onHxJ;+ig+p{01{=o>ebEJSUlrmtumcnm*F
zFk}_<(q`+n4*Juk(865Fx@y7iD12X+m*MLSHELV+^)_+&sLg@2_Qu8wCIVN|uEQ-&
zXea8ApDOdGzT+ed=ETCPoy9vK`D+z}%WZNSJcR{dPvp+|wX)I$nSl81@kNL!2|2rg
z925ZCKjNAPO5x)JF_?G|PZ4*7tmc;|=m>s)e}7f&T|C`~zg73}9kzi1o-HAJW$mza
zs;@adBU_n%Ioo{4Gjr6dUo+erA$OX|FIyfXSBlMXClm87>sD+Nr@RnUwusVZV`rBG
zaTO~Gg2StvQhTX8-k%=nEzw;ML}8d2zm|`-D=Cl=4NH0mUon$??a2Y+G@Fa7E9>ps
zea3DuN&&+*U&VwVmF=X1R?q-SpT|T-W<UxITJc=JUI%uuxg6{i*Vg828GB@N)}%Sr
z7kP{MJ5=}r(tDUoZmt7&NmnvrF9>2EQroTHsl3XfW+#7kGD|)Z=(&@eUdqN&c5)}z
z2wx&{Ge4jFY5om=a3Mjo(s}Nq&A!ZYC+gOmu3u?yHg`I*q6-{X$sz}Syx41D<;a_D
zDP0Mw2@_kCAvPH8E3Xt8d2iUAoDqZ<D|YG%mh;6_jQ3?!^Zklp|HR+f9!}$X4je*T
zfo;H4e+J1=&es$B=J_>t(_ZbozJkn+nnxBZ+D{0J)WQ`{BN=yML)0Z5RS%?gW;DX;
zO8=mI?qko<^$}3h>+tY!u#kf`5fA%|F5^XOOGx-I$w!%P@WD9S^!<BBB03~<NG(U#
zisl*}6hIjHfwy{mVFfu$0iI+)vra?2k*(a3Y|i0HjZcCBO-olMC)epNC4Bur)YO=<
z6=Fgf@`Sw5tKCS>P;^+vW#`e@5?$6C0i{moLD#q0)ta>OOboqd-C{~jBO{>|rv8ba
zc6P(6W^OB67iz4qCF`6fL=rGzhM#z1b3Wp~skryH$q*v>Rp$wwszM^YZ>p25m`MGU
zR^o0%8hPIIi5IOXy<e8T+uju>YOO7@vz#E-%weS-J*heBd5qWUbNK$!a*`7#-oYu4
zb=cSMr*^K@<g%a`>CK^p(gb(jTdXR-Zf~k(csMJE%v^gToZV?@L+6{X!i*i4I@%`A
zTJKpB?&A*2cdK^f6_Y|tCJZwc6W?A9=9>h8<pmig6B8NHT`+}MF}-(#lRgM`4cB!=
zBZg5xg4sasXk2iG@$mQLd*}v);n;L-wiAK)`+syOK6U+=DTCH_Nw>CNK4=S#U<_EI
zdNjF!b)wjr&^2nCm&CpM?&@_ZQz&XnUVWWhwYc@SqW|ECWe(HJ6HoC9w<RARSS=K4
zb-KCgOvP`WW#=ov_xde$DCw%Zs0bzt?rfC}gNn^R#OC0hPUR&tw6=C@fciy<Npz&C
z1#TQqnAnU&ite`UuAJshtrdUsRBw+Rdrs!s&!W#Egr)+mn68Zo5Kg3T5`@KVE8zx^
zx1AMb=6mJvex<pebc1zLJ8|Ob?-?EJo1ftnM>5iGK)wR4M;(Y`#yNn^O>J#OSj2XZ
z98;?VTdH4RAohz<a{+jl@RYM%N$6=>*ID`zn9q>IY+yOeSX)~=QF@{*A!G*8qN#dy
zR?`*v5dmwMOyj@S>C%FvN3~Z$a_B~MHjXRZIx<jxEn~a%=8|v*Wp~t;J6h>m^Vs9!
zAdeY0Lc~@Wqw3tIN=i=Pprg}cBPs77x%0Z6jfrggYIlP>(!RxS2KRjHbDA}ESx_rj
z-<`6kO)ya3e>|C<_`-o~y(W6QsZ?t6>&?XgS&yi+yjp`hiCp~GI<|dF(v7msi;J3v
zvqa}5>!!jpg2p7%w)e7IxYC>X+02X6mo&O#b9&1iShY`A0hG(y=+)KoF!ofPc(+ri
zFszp3RsLyL-s{WWM}o-AN(J{_O?n~``NGH$y+p|mD#O!Z7Pa9idG+tgzK)5kCsc(B
zDtnf0Z+ckHN_9|CRK4`xNzEB+*dAQ+N#&iZ+|?a!R=qu!Wd;C$VRXZ5E9m5+?JES5
z2Sim@eREakP*)kU!uBm0cr4uRIQPrLQ4~>f;$i@h59-@<l#b<^sSi(&pZXhPJ3XZ<
zMTbwbrnWqh$u_-!>R7(z4@XxCq51fMvhK2CAnr5n{;Vnrl|TCd61z3gPKm(Gm)Xcq
zggQft*ep_0qN1O+j$kXZvQAc|*U&KY9YgWDYlHF>7sE=!uQQ=JLlT*NUR`@3hm_$D
zI2-E=2;1A$Pro}`-L-IRWC-3)a(IWdhW)&tdJ!L9b%vb(4C-DD0-*05{hABMdU2wN
z0|WkypC|#37kPaW%S&*K;!BuO($&66W6j&o1&)*u8^DO0o@19t4|?(wFeovi0X$)b
zFQAeTx6-+DihU!tl;9ZSJamkW{g4tf=H4t#e3bY2Cer)B)A@v3Fy<~u8hY?S<g_6$
zXcl?Cgrf}&T!C_19HJtgo}OSxQUsp-?p<6vjb!b6++=@f(BvdJ*Rdo4kw81kroMA<
z;rdH!67@ATVqnX&E32xm7JZgQO-<Jzz(2(%EJFCzXZ}N~b>d$>s2&R*w)%F)f1eQN
zxg%QvqK6GwX5+I<t_KDgQb(@vB~NyLTrd0r*PPrGz<vyrGaT$oAKw(H&tsu$XsF~u
zgCq@qtPukT^5;JfolgHIz15d3Dn)MsSIqb1`?*)IUV)*{4ESLFDIkksZieF<Vc<Xo
zuD1Yapt3njb~vP`4CnU1IgkPezd*K-MJ$<V$|@>};Ll72e<*TN!xIoKwDRIWS8ccH
zw&)D-19g^FnF3RQoQtq<F%;maeJAspK9EeTNs^jAF!WlWfCRPOTXH3@_i+ht!Q*NQ
zCyS9u1!OM7_5t<_I44z+U9N_Q%KO_j*U#=X@#@N2U-iB!fWvMu@c)F9j_}q?moD|%
zD_D%d@t6d-E7QGgf+i^a9AMUmSw0mGv}n#C=N`cj2R<$(kJZ-ONKh;Sm=c=DESPpD
z2Ot!T)o6tVaxBWKcqAD}26JE#f+_dy=zVtm$bXP?CA!y={g7h$@!GJKtbx*vB(0AG
z(k(Bq{^iLFne8&t(r<HfZ!xg5XPtFZ-C(2uGACE0`0-5%;BJv4aQGq~lQweW{DDm{
zN8#gqF?bLLd>AJqZK}9PCcE<BaOs;oxH}8rJ=V#C-?p^0l;Mn{&EQDaFTi>Hev=PK
z@`A{fL=2?rUNa!Kvh|u_%0iYaQ3RA0OYIWd``{x~1E`FNgp+Wvhk%4Ne1#oE;V>CO
zstc+CKm%^eBD8@O&PWD>gX{(xq4r^&OdZ^j^Iu>n<4>Pd1~wi{wQzDv5wOOY{gQKH
zQIaHi;v{x&1s+h_y%aF;7ckH>^rt6`sA5KlO2Uy-LvY#*$b0^w7cX8ACWAz=k8*~U
zfgocX5LQbv3U~!aIH0W}r&N*gt<2FK+_{H`vC9IcHs^urhHrwSj)?QT01l*xOGqHS
z=*;%t-@SE-t5)R4A9jh?YTxUTNj8~hawS%QO0Q@Ijx*#?5Mlr?Tl5<$SuDfDk$QgM
z-pvNS7$Jv;frpDMTf98};Whm_ux>N(@oPRk3M6*E0YH8=c6aBu+zc!&Ed>%WEkU<5
zrE9XYKd2YjV*;X4!l?;0h>6T0u?G(v;Zz555+>QxP*4OMB#4<70cjv1fYeh7_UnJF
zRq0Cz7`sF}uiH`gsY_wO^b7;&2{b2#EdL|npHmZ48vFY7AFp1%?5EE78^FasqGqeP
z#S6R>;+aAY<XyXV&9ceRz@R4)eIar(AGAIi5%-cu)Uj+pg<iUH<@P2-h#f80OL8R*
z2#@b!5GC)MOmA`}j%G}vapOLhIdv_;-);EU^W*h!{v-gn^)XZMGrkWAQSQ^pz<^D?
z9?sVVfYX&S_${bR1O0Hm*lxCi9yvn^T-R+lU}EhKgXp218d8K34zbn)&cc_>WQRvZ
zSspufj2!58F0h$o1p(8bfk)2VNx)79z^U@4emGu0Mn`8;+tARbpJUMwq99`C4#_A1
z_N+XRDkovZdBIZ)r;+r?eL9zePILWKuRYiBntqBQ@SLPHnIttspa8btP|8E%`Pp(^
zIH!#`(tQ9dlTlGwWa?E{WtXckSM6_Gh$%%hfT4y2T-_Coa6&2o<fR%*ek$-+mFK}g
zRJP#Ap69I1NU>D08d%r{&kz8H(hgI|wodpD226ryksF=@pWR*f6UxM)Hy-wm0@jA#
zx_ZwEPxr9Kl)B*+x&2A5S}p@57i-@uL1B2!07VYR83Kn(Epj>!JDazkE$Ki1qY#Jn
f`hQ~r_eDQmTzQ+G_53mNR<yYM!|VrtKL39JPre5q

diff --git a/_images/sphx_glr_plot_lasso_vs_weighted_thumb.png b/_images/sphx_glr_plot_lasso_vs_weighted_thumb.png
index 12599495b16af06d120b13d327cd166eed0a3c00..8cb001256463d2108be30fd6993e5d2ff548866b 100644
GIT binary patch
literal 11750
zcmeHtWmHt*yDx|W0+LEeg9u7WH%Lo~l1eMmT|>+Oq6{!J(xtS3fOI1zND6~=*U;TC
zaNqggbJsc_@8^3DYt0vC_Ut{+`_wN&Uur7fBcLU~!os?zs-o};3ky3Myr}VU!GBzh
zzItO}asN<Nkkf&rZO?c^s8+t6-kkc4`BlV`GRflKH7Ap!oRI5;$}8CZ^sz1Kn!IQC
z1ASa>;-Cd>+;U^c%*Ys>+dcZ84=o|!xo`9m+p?&K#drJ`n?eF9(`B>&Sn?cWU@6n$
zGd8xs!Oi+w754K1=wP4Jd0(HDTa>vWB?}9SX<~NKQzjM`skmDT3JUhx0U;qF+yta#
zWIq+L!Q0gwvI+_vly|^?$Fs1=$jIK_2OqXT{LhE~ABqOguZ@itAt5S**#iIj;4S36
zy~P;)ulV2AI9ej}H11PUdio^$p6;xW_@Ch@Dk@&BYx!IBCUB>ltgaN)*QeCi*O!$1
zw#+Up44a&MgGJ0Jt^W4y!x)uMSy@?bFt7FDTtQ*s3;}e6$gHTH4|wa_%QeE@nQ}|Q
z%hTN`n5nk5oV<L{OlPe<r>woTt?lC4Q1;bHO^an$B>nvd4~G8!m4!m392^{EeOZMy
zD6C1+Mn^~8R$sq;tEQ{_04vYtm7bmw8+qt;V{vh~ER(D#-D;^n`l*<hnEhmB=<x8X
z;Cm01+}s2#EiFH_FCfZB65%Voi4~4B^=qe9RaH9R>J}$kTXDPe#Q1n^Q&SRIKI)v}
z;w=lkny#rTJ1X9$Bl0|(gKpNS=+sm-ZEZ5%<};SR&AVx{`L3?6A)%qIf*MbrJh8U3
zW2Gama9L97>FLR8y*X{UN%xwACMG0M&Lh!lm0@872L}fnJUo$_Z_D4`ovL;~z~&|=
zU)b2#2!^yR@a5*_3W|u}ywulMR#faXysVwu-7Wq%;!sy#e@klU?>BiGad!?7F;v4G
znrBezl#UU0#thCUyR0laKf@=i<!Xg@+GEUkYq~D3t?l`nH*X+U`&~4m&P-TEdOs}A
z4_1h11w-JBu%wpj<I<7Q(b(wdU$3>KI-iy{J_kn*4(xKLAu0%uBsL~SPF3~p@$vE2
zL<RhG&OdXu$ydN?fUcx=F^tFLX(DbrA0H8pH3?$O#5W=~mYiKHqp#Lw39|5+YHg~9
z*9|qW(oQ0iQ(5^XEsbuc@lf*`J$I9?!uI^b2lgnLYf)l&<Y04O2pQ|yd=QDK>oPv8
zawLd&cuESbQY77<EJcK<`x+^%CG&9x<fCqp-cYHj%r(e}p~=afp=@REi}ieL$a-#~
zwYBy9eitL^S0K*v>gqGUGv{iDDJAget!B}gH&>{}lS#WYN$*$?j@7|T#vkzZg%C%t
zfPY2$75+Ccb>y;@%q0zP<K8W-cU_UWPIJv%CXGqg=*z>IU09-JZ^BQfIccm6*oDo-
z$+j=qnAzAX!WOI;mD??ITJPSmMoFDegZLvdaPe+o_w@D}HbVF(CME<01<ODZU_Jjv
z3W{V&q~-jcmejG%QTM9M&7lmxM%?MFZxnchxVX6Xg?9M&A3s(m?Mhc-lo=LQQ=L8x
zq=^yHiADzp<L0R+w}E4Ca1}Bt_mgI9hm%;%{|ambmxIGwJiSbQW=R0<mnQ1+5r)2S
zL28J^fc(6<MxQS+-gLJH;e9VCSlie*#iya6`ShGw*na#Mcg-~BsPTyDnE3f)a@7)J
z<KpBM74Lv#WF^ISp8thSOG|qOe!a4$Ch?CSKSa|tVMp1F{JtmlCtDNk%YWj$*S^73
z)YO*Jy=IvoKZY+ZT8tJOwttg{L$1$J$>JUxR3@uEfBrB_NTfD4N=7nBz0lCO2TD;r
zQ#!q{@CjG1#cE$N_*&1c@m9v0Gv(jmH0_gBcD##|n?3~5gk~x136?R@(X!8<-%7Al
zd}s--POw}!Jg4TjxP6EkL?yg+JRkO^4JJ!Rb;q!tukiZ+0hyPTzq#O?G&-smOu{(8
z3)?4@zF1?W<S`nbK=@3QaXs}}ErX$bMoUcxALtv&ffls%J7OvHafaFUWL4wEM$y%(
zNDK4RrwFSI@8J4+@pPB*d1p}-EiEfhHR|c&@vEyJJa=dN!D;OKG=)?1f5arFU1?K-
z86RZSN^AEiLlpeer`ywxBH*D<DUgEMmqVD+`gE5f>#&TAu=nB!cSNAZ_w{QEOu88w
z8g~5tz340oSxvSo<3gCCH}t3d4*UI35_1dj`X=W6DMFB@*@^)Zp2xQPR7}#+4Dfb2
zBO|&zo7j|;u(8dRWUGvCIXR#D`qVMyHS3L{PvA9WXo78cHY7{=3J*$Oa<$x`ch~1y
zWcK&>CC``RMCph>^rv)ogi?54Zr2!A***xAWqOk^w)sd}IwM;-s{H*Q@++83a&oe9
zBRdcM&!0ar#o&Kg<NppMd1Y1AKBzQgkv=a00RaSkJa+f<^Y<S<97OqF1*J=P#b#s-
zfZn4FYEMi0oQ+i}ytbI<olyZ9s}d;@?)q3M<M;31JG#0AT$fc!8g}Z}r2h^Ne;h<D
z^*k-j(JwXmPrsjQhSttFE-o$xoS=J__W3M2@DsQUpMw+teTLhq1ugtkzl2RpEE#mH
zSWL^CY4B2(!6T)j>gwff=>tvQ4kV+zk`nGiVS6SzVv0}&$V$9Erk8*RViFO-)bxU9
zL)}JU5w(OTmzL!IZOmNDOLg2R{5SrHXtR$xp)|LuDn2QR8pML1pC3f+V=EKrglueV
z^T$3eqI6(fSWx4FUt>j|3+51cr~)VlgeTF-Iub7*dAHK5uWrWi)$7;Umgzf<9(6`K
zy1L+&s;a8%u`4b2?*z%&+S-OFuz>Hw+O|rO&LpXs((v%`a8DNz5*ivAQOJ~rI@cKu
z4d&(Nd+onE;#zQ~2JObrPZIPrXHj)~d#=B;3)}AEsu~)(KYm1vj~jp@>-iJ+*s#)?
zOiKnW2<FL|U-OzFd~7qfygatHRun{ZU@}T`@bHKRF}005$t#vQ%Sb^^9ug6eTT&7U
z5?7Leq`$wvvbq|aqbA4wEHKr<b>tKjfhSNQ5b&g=q!PnWV+F7kO7--_Kb>fXO2n~i
zES^j`?QU!sjs9mZ?0w|&&<SMtZsQ>_Pa<J&IUiO#)}fA|xF!_3dW3|UB>cN}>m5$w
zG5u_5c(ged_0Xwt|0Al(YLEd8M&~-7&=2E3yqipR>QEVkJ8Lp7i?<p5_OKbE@`G&I
zQ53}RY$(@^Zq1yV22!$U=xu_>06j^3d^`v+zt@ftoxyEZ%D|VOmJLL{g($|!D3pql
z1V)G6j2z|3`kFrfIHm}-!!Ml*?Xn>ie!yy)17i2!sh+z^z}9S2DrlAhkUcYZmT7OI
z_W6JMrYn&WJHJqe#(x`$>BT=yZGV`ov>Af0<t)(zxBh(BHQwY~b8ESXgM(wflgf0t
zOXiv%v}e3Y_(KCwSo%gTcPc*ghe>lCf%%r<fr=U1wEPu-O~fFXl$9cUy}gar5k9YT
zdXD?~LH{!4n1F){!6OfcFCrJo;yEI|S|d1o?5xTC%7yFns1>i%a2AQ<HqSCOV<yrN
zPae|(a5Ci;6&_R*iqR=4lg!NY6cqAcu!6zey!YY{m8&S-4D5Z7BLx#^yWtE{;^I7!
zmWIW~4GHIms0!zW=c6UYYNaMkwO}|J*10~CxmqCk$%O#j`rz;|d(iaCW?N2go3|C^
zdC*sW@j9`9h&W+%fscp?=Y*lObKP17G2vQEp5OkJ`#v-GY$xFy?U?1}*T*2IsjlD!
z-@XMxFHd4JGY7%2kSO?YpvdE%3bGzEJ|gKQs6{j{yE~m%%ap)LdJ}hSdn!962Wc^q
zZy@uvjuPs>e7S#dwC*-qE1cP7PA#fwyy{Agn_n6rMwhASMT?8OZQ4Lb%rn>(prEj|
zDQv~0tE+22Qy;&vVM|AhC!HC!wPkPOx38Fznp(G#@YVqAS$HH2@B5PYWqrHDX#|2t
zMs&a&#QODa4E6HmsDg%Dz^-vzXU5X8y{Jw<I_|>lYh_?qFjh^(M%Uy=0`hm$>H{y#
z*LTR-iFP`{HSQ}<YVFw`J>q}&o9Km_ns0}8LO|D&`EZUZ*erSJ*Y$W9XV_)aG#?+X
z_P1|)B5RyXEi5c*Z>}#VntVmTu+GWLJJ@#ePyPK{sic0RutM5*l@w3@lAX_KQTC&G
z9Akf<xt;bs>8LnVi_3$X)1~a&?d&oc$?|VG+GgHey7m7RjMS^yW1MRYJ7PQbH&Uln
ztVL-<x|8cQ0&;S4_&hd7-14;2MENg{^|xygNt7k8p>zw~`KFm-Gv&qBG9fM-nIR#5
z9y56)IoXD$D|VZQ`G}D2#D@1c004?7{x@VdlSg69<J4vQ2I8xKcm0p_YUIO!uH1JI
zV|FyLS&Vzu`K<4zT&?{xwtrgKYD29fatE@7D;38&JhRTHRsW@(ufUpMPt}`Z--IgI
zyllz9L0?}@q50OG5clsPHqZFPf;WTn$s!FpeA1>pbZ<&Z5<>j$5Rk6F%Dcmm-wJZ7
zw^ptRvi>b}TEg7yKbL-VMB<SX!wu`*Isg6pC%N|vhwbs>*E%{KDMEGyB_$trrt7xm
zS~9`(X4g!EjD~#8pJV2c>P_DDRt<59qeYn?{mS7fj`5?(N7B%@aMAHkxK^TFmqxbk
z8)?!<nK;(>Q_j^mF@l65Vw6q;i#t_7csNlrdfN#(J*^J<Z-eJH3pLSjzE<WM3T*fs
zPo50-I{1*b%ngEuPyn{q;1C<bqy$BdtnG1nMHVv6G||>jipN<_hkh_!=EG}W0nF>h
zSNEHeTP=}CD%SVHLW{lm)^J*(u!{>HEF&H&aC4AcF)=Z%SLcT`B93%~lw2Xri7YZQ
znPAk}nKd)yeE%N0JyipKw|i@;K|F$$nyB3O)Dcv!3pH^YWn)b#5oR|JdT`leEp?nQ
zDE~m?%6;DvRu2(@KS>EVF%O!Z_Wrng4VAb#9y0%EPS0UPP~ZQk7%UQ)C1PW9Q&mmv
z-R^7?rdYs|D&R0Fhy|+2MLd0|416zW*)d>gRffAI?(lMRM|dBuV$;Ud<nNvP^<_w<
z0-QiW@Y$J`5BBnza#2cX66F%!qoqU9yNc?t;{n*U;Y(~;!=TL0xy<LYe{f;fN(>II
z_`ADLv%MM<oizH^7IR|>|2f=xf3~*(YBmVKNw6%4OP?=OV3jmq*eXTP-jN0SD~Qw8
z9$ZGUXl|)*VLBJ}5^^M&4|fv@u-au%^v=$VSr3(TLGsZB2g+Eynx+1=X*(y=`w%CI
zQbz=D504cjo#46rmwz=L3$Ij^*M+ff@j)++*2XB<)kQb`1Cj--Otzkcc6;Y^cz5s6
zcNO!aRYUFk<E)Fj{+yQnS)5ooucgDCYhfo$Y&Ub&d*T(o(=~bZg55ILb?^+^*vNMs
zf98p-LXK5v(x@1kW-cT4&q>xBM}~>ShToMk$T-ra#X4W5p}*DeJ2&Dn9n@a>I{M;C
zK75Kf?cPSXO%#n*8gAl5i;bF}yB=XCF8x6H<e;XsM1T3D(?o(=(0?mZz}2ji{_Dq%
z)Dgl_$ka<zOWkQBk0D(2q_yWQT>0t0l~wYnBYW(<xgl;(v|1*(a!Y_nozP^$H9T9F
zyQ7xAvA!ya*=ApqJ}g=-oyR;h1e)*}$w!^E)sEFhhA{cbKn``}nsM;>v#Ez&?tS<n
zf%`2beM%01`eUa4=@=xV!q$daJ^=Q4^yrZlbl~qD{TRgJ;Yn494gXCrog-uVfG$ql
zIa;JN%hZA>W^frkpu!^sx+DQ1;Y6(qYeYmuvB6Ik@$DcIX(nc7>~=nbGBfNZpJR)K
zw$DTS-nRm=01y$MT$fWvcI=+o5mhjBlyiaY4Ayd!mZ`{vV9QKRZbv4GxG=?)zq*^s
z&I50)&KB}>GUplO8O$%_{7PEa;hriT4WJ6hAIwUA2Ie*PF91Jl0pOiz^cK*}rzjok
z{ZTX9GmtI;_9tde&d~U$`sH4`Cb;)#M>^G*OmzWD0&wmfY9M_YvJ_!#On*)=IXMYA
z>=&5;9Kqb#ZxQKJuK~YA1k(KF=sb$VjzG*{RyybMPrcrq?nil3w^iP}p-mNW`i(jy
zV`S6<SsWS~YUI{0gaN1k?E$#>`Sa)7L2_VVfI<^<TjlfIG7xr}o&JYlGS&9+5tlh@
zB{67(xE!g!s;>ciU-M!5ZghP7alFaw2Ws312j-18L1)vNAIai|FNA#_Lx#|XiII+O
zckPbWhH`RqwG9lqw`=A!VKC{Hl@%_d+USlB1;_ezcE^ToVSpY0@N@%YMkq|riFGTv
za>gg4w7$PBuQOVDUsCiDMLSrOTqFgde8J{s>ocDd!yVml7hUVJcxi4?^L>TZqPiG-
zxx^i=&i2C@Wv;WvG!&ZId)v%^@jA79Oz%9t+R4z)`Rxm_^0Lt%Qx|_(BZfm3qt?5A
z`Nm?J_|=jWil~G`3X`DX|4<QqfMWQ+WF6g$usC^m*m-$*FQ9<qk?`F5T2O!=00@OL
zhbdvr49N*y-7knO6<vYd?Wfu0`_tNHiaB;3so&Hw{kiQkVQ5rTK3I1jI@SgPhUUeK
z7nluh?qam&Y-;ZMcQmv7+F0qFVdgPJ(pbl?j+AGNN~FcO)5GsHmtIHkqRv*9r4;5w
zfR43j-V+$En|7?h{q>~)#O4(U_xR6ub(q-?<_ea4Fp)V3F{H@^EZYJf`j|-ZF;Dmv
zO@9Yq5K90^!3t4Q-0XJ`{P_XqW*LQAau*-G^tB^=!s%emI=aMGniFW;waH__y`m1#
z0GI<6@4ASaeraTZ@!!$esST(jz=#1XT3sy+lJl(B^tuhKVSoDimO&p~1QY|ah={(j
zTw>sUP^{`+J^m++i$}ltik+V8>#HgZ{B__lj{tptdfG^@*q{}VKj(YxBmlaAtEdut
zJIl=54O(DQfR<YVOvP`oyU8S`r1SuKtsZiBxd_D00>Jq%2?+oc4P5$)L~Zq!S{b))
zN@(u2?9$6?W}YDo)WqWc&%$I*->9lnb2l2KGVCBO)69@E*qqOHgaZo@AI~}Nw{3e0
zIL%6O^@^j);U*E20?_#0$ZR~eL$Fo66Nr^)d47<4pTRW`EV}v7lZ!%g>LE>6W8bE$
zPb_&#t`FAp>NOa_YV+NZ*txQ>uBeEaCEvY4fHY>%wSjJ#p=hNG;i)}FzS#Ob=<JxD
z)kV}@>B(@sJ1)yi3U*Rn=g0mo(Kr8|+i~<geH$jGhaCQ=OHsGQDVy`Q0t?=5Cd<o+
zOB!kbCWK0ei<ieg1s>n}Pmg9yIx3T71mGXb*?S&dU7eJ~<UE2j)W4znH+e<iNng6O
z`t;(NbcH@Rr`s<BR7lKHC_|-11mOARZzz#&KAFddGhLhDf%$}4leM0t@Np-lq<`$4
z5YQ}p`VhUvQSwfq<#upR`eku1_`Yk81>(tDWF}K@`i2>`#sv6>xjAHWy6_eI-13@|
zePc6Q>Jpth-kX&gJfTJ3ZX<)aqF-h-9}<XXr$Y5z=s{Ld%M?(7L-Lh7eB!is$a~}U
z^zsGmdLniENBH9mx8R?f5GgR8_xX8x9-{qT4dr0PILcwO%=D~VLEqdmUu-PlZ9>|5
zn{a{;{bYFjc=L^m7#($RYZ%n>>S)!x!+Stxbad<iellH=&BC`k5*$D5SUE@g>9dhG
zgi5szI;PG1u=z2Vp3T_4_`wU>ki%LI8|$U^pm@~n?tc9v^Py^sC#koLgqV`|gWT_`
z)x#YhFmZEppP^wGc#Q<R3O`X$P7a$oG6&$?X6WU{w>GtCb6F0H{i~j!#)g_NoD-Zz
zGL3^&!QCTUI-|>p6k$oAK|Uu#yc~b;s@H=GR<4jyw~>En*tF#AAE9PcYR|oz8->_|
zYizvv*wIysuc$uTEfs2McvbQBBeTuj@P}XgX1zzMZ+p90T^8Z5YEY}V7Q?l1V!Kr$
zbBl+xEW#0Bn9`NlvkM3bziHV|Zh|-taF|T#_4&`pWXwu6-Y7Th<Yt3e05}Z{iwro|
zi3%$L(CEX$!Z2I~7UXOJZdl_;-`UxTp=<FN{rG5EGU+GTEJXP<rJ5YQ-{Ch*dDVuX
zko3G;Cl=mQa9k4KMgtST_!TYBNaY$qH}IWL_kBtpCj8!?)YBZPKKP+Z15U4_I!pXR
zMvBR+v7WvKXg9YbBd6!{54@uf^qzgy=L&jdVR)medldCzFwIn8ibS8^%bj+vRy@1C
zS|p%-KUrqpgf~@7_>l2o7(X}1xS^)vFEcY<My$Lp1ae_ME|}Ec;P5W)-uOK+$#J@G
zjje-&jto6+yASvZ{mGY}#hnjiNj}r$Mg`035~(8I^~TT=^rR;7=ZFOPcD)*+p078z
z0XWz^g>o*9FWR3{$F(sp9&San8r&NC<%~PS*!N4}DwZQWu<N0GNBLeY9kn?Hmo>kJ
zEz5Q<68Fkmo?jwtqt;GcoFsf3fpfOvo4Gexy(CJH+e<t;EHb3#Qy!JSL(}dLg(Tsz
zTOHguNVk6Lxs{h$XsX(jsa*Ppg_}F{8tMx*hm54oyZ<VedQd$6w&zIhuHQ|Hy<|z`
zCZVf%x>bZJrK{*qi^<dM!);UM8mgu6Ih*!*-iKj-xJ%xxmPw;3kpeYAGbd9NsG91p
zvE_wI%_0te#n65L>fVh8sF>e=oDYDL?t_(H8bCk2t+rnPbS5(p4SbGmZSCyNk4ydg
zD+Z<20Eq$y@fV;*=;(a1f%!ls^LK9lcqshp?O#SGZ&v+`g8XV1QXGOO*ah5M?eSAl
zOPpJZ4w;ZVdub)_bET6lQMoXgzWvviZ^Q0fMx4A=rQk9+t&WrJV0&btnBVob>@b|Q
zFg-s{h`Q`%L^ATBlou@KWyr<Z^;NO6pcs^#`rm2WT-U<sB_}U$etG#pF=?`&8BQ~l
z+ppp)c@0ud)PnzOpR&&@qADSrb^kK*eDkt8u>hCKJgB$yPXE176Lq9TK~<`R7d@)5
zhi#PmJ4;DnGA_Io_BHWN)h`uzt2JCsbTDh8GGqyPyi{g(yvc)UC>n?#c7EY4((r`e
zSYBI_;h#U-JO?a}X1m$ZFNIP6SBgybG0SfJ7P|ZqE(`=OXho~5l5~(PPIZW65#mR)
zn_0V)T96UOwsk&y`f@%E#GBqW+JxUh4$RLAE$UcGzx?E}s)X$)bS6}}7ztF0=zjqL
zl^{iZ&qxokS@e(q-DJ34Z<JDJ-a|}GXB10`>*n43ZJ#dR^U)4Ml4~Ge*w}DB)=c~A
ze{(&wzNEF&Rbh6o3g<Po^WNcUscLGNwT+|yo%Ru_+~stm!JiJDR#i0uQ|pFPtsgG8
zLk|xxgO>A&m-b_EN3V}))1>C*Ccx-&zW!Mv?No10ZIGY{HZyN;@1u?5b&rLI!KdGH
zvq%;~{*+1IH5%fLF&lrsx+jhGsh#~)i%y}VGQ@x~$|O|-X2Z%I$j9%9xA6KMZ(TaB
zIkL^BGI5MHY+jJ4k*CWL*-l1r`d#LeIkIEh6^z)t<8?`&jVsB3vF)!y)ZYrlLWah?
zzNIGT<*`>g&1q_DYYWQK1-G&yEAVyoo-sX&jDQ!?2r-3zpZK0+0LX|j5~sPm8><;I
z$@~Fvt&wWiG2QjrG~!!<e245GUjDoPHESs9bR_0<eI-0e?}!n%Hto0AX8r`@?-fC3
zaiSC)vg`T<j+*eTE-fe8o*wARk?N6aZ#S$q9XHR?<+9-DMD1B{tIXk3XNjq)LYp^`
z8BO)x#og4fHTqpkAI&Fr#5!PT662)t;I)>i#>w>IyvB?3FK2+8lj5gR=Wn3ezSfJ$
zf_g7{7JFAo9dtK9)*^$CS38fD_eAOM>((gLWD_E&Gvh)mwe{vaTc-#o@wjFHIdgt~
zPDIVeB#sVL4YHn?`$QvlbA23S2PY+}VVsw8br5C*^Mjw;VtU}P8i|}*cs}W*m6|!m
zw^yI8Q-y0o{K@4gzzzy<jZzdtKZ3VeMNSSIXlC#KOF_^#g+2&?Gx&sfY>w_vyAARK
zx^Z!73FO>=LF!DQv<AB~{Mg{Avo|Xw6<>G$NfqDZi2K0Rv^@y*PqcChIc0U-0;ZIY
zmtN`;-m%mOr~ND4RQ;&C09%!=(4H~d#yS&nk|-EWj$-C+?F}4OD);F+9zUlqu>JNm
zCsIs|Q~l2i=JzIN*2Mtf^recBi25p+%oR86VJ`)=nya|CqNI3|r1Qe>>{4Ov9ntGY
zKVwE0XLjAn*pla(a7}9tH5n#v+~T!)4{3qq5v#5kcsvDj6XIk_Q^hT(l=(&`0ds&b
zdr2b=MfreT(^N%Ox9DFva_>*0R<J27=$xDJxhLH}%8oBLhtgbK=hD-CxBcz?gcLxG
zd+o!JJ&9zHN75VxW<%8!M`3P<{g=PVWJCjpWLJYNPUuRlRNh>D*FVRFIBzgLb7G{0
zZq&GMsv2qHmEy8obu9#eZoeRiD6_Fhq2zQi=~Maihtp(-MrK~J3he{bvkI%hJg?nZ
ztBCH+WMTTj%5_sa6Mf;}b7XHFh~9^zfE>5U_2j8<dFRM&9MdNBNyoW^sFA#eu>gq{
zZTGKb>afkhrp>Yjz!m@rriVgHl5AT6sKzYJ%z>F#%a4aNkNnRQ{m-X0k0PlDr2Yil
zAu&<}Cx?!kx>5KtTH{qMT>07BnvkE3J7ZP$tau&ZwP&xFIP9$bt-L@d?G)l>yO?G>
zpH3XPn%2bo@q=E({rWuE%?s<h>ECCJ$<YHp<YTe3(5#v0t6fLNo)%4&$LoD7>IY<u
zhwI^tslw6hHxrvyU?u}PfOhIR(Tj*Bc<Ijjxfs`y9f>3p)4%J@>&`-i(vm91YSCtH
zPj{tVg;pY98HSz#VBjvg(Ae{GtoYCJe`)>hGM$#*ksWstx8@c{IXCMyk8_=AClJ_Y
zjiyC^bIZ8r^7*5n$Y+BX{YbK-C3n%iIyxei{UP3=-qIa9)#5SLxen^ZCi(`1SBLYC
zJh?HKuX`I+HD)%ZpG?J}5}~%fZ>7vgZQILzo(r{%13`D9-s35W<esd){z~i^(0J8&
zn>qjk#tSh)P;y)JeBpd=W#uK?FFt$Y*Aj43vBQ^e_po&EybblDxrQ&%>-5XKhs1sz
zeKGVv95?^>J}Qd%u&*2^DWP+@SSR7itmh6%`@wJ?3)j4}rNB+oH9i)J`oOY(KfOkA
z*Lhu{A5e;A4gu1kc=*s)&a%|Bg&|-Q;H8hNtfnQ8uZ%G~Gs}#s8dG0b<1;QouAmdt
z3rE@))83J5cx;#Wsg!ygZ2`;Fi;~+dr)xqJ0OfKPIW1-E<M4<|+NK2g=@=;Lz6<Db
z>S2D&ol$1vvilEb@U2M9Ko;fW12I2_$ABVxZ7_>p*^)VPdWYnZG;HZ5){BiP2;^dZ
z&A^x;F<XA;qxkj>dY%+8W$o*H5KnJuJD4Z^UH-|4T)TpmP(Nb0`QFw{cCD@1Gpth+
zsXfBoFx80DJ(->6a^$A2>qg0w#0vCIaUr|RU4Gh8h5F!I3d@qO7A5^|CU^dUpMKQ3
z96yrj^HYyK==D;4{R>x`2DV!r<;89pN{Itx6mAP7<QeNnmpdy($r;F05||4`go8OQ
z8|--IG%f$yp`1%?YKvvxr{N}AC{ZB*Xn3AG(-YTwkv0Gb0$l)LUY=2n5BW|`?hU*H
zh)AwpK%|<CD1As_2h@(BkdR0B6-x}hQ`0de^rBKGfD-?<vv@kyw%OPGYKr`JYyY|g
zn|;eF=^4#11VZ9{Q@s%n&@g)Um%8{h4zp@}+xd79p5RtlQC#j;i5(7&V66tO1mLL%
zHh-z3qu}Vs1Ne9eyO6j6pt+6%(g|w}FfWRn6+i^SSQwNUp+{MQ8v<jS*-_Fyqj5K$
zLPA2?#>NqrhCl!`@ke{Gt0&3o>FMS3B3DK=8T_wYE{-?7S`P_t-vV$~sb-?nB%}97
zdSALkVr}h?Z}LH`C=Q@Z!3NK9fnN8W0V0Io(8k~Kafo&=5xiYJMKH9uxES$@!LQy#
zf<+RGoSa-mMI{6#W8QYVr178!;W^?VRuvbCPh#Ct?D`?KVZ0Y7Q44G+*>B-^Ml4oL
zGl7<dRM3#~5{}nA;nfqLk<Fvfqi`=6Oar)FxD3lb0nJa7*A6JR#fBBYt|1D<Gz1lI
z9uNtTJra;TCba}EW*HgAcD`}IUST2+P9ptP#@sFneL9DEJqFCO)nJAuXK7tsa?Pv{
zLdy5Vea%ls3v4NBi9C<!h>;^B9bkC10>gyFkH+%yXS05X8rAj_<V|elkbMhai4p(N
zmzBkWiQmc2j2g%N#Tl<T%gxar^d_2MIs)|%7bu_f(z^nmpEJ+;-$20z5@}Aoyt1+n
z_ytb`|7p}jwF5>a#0BOUfGmM|rF-LFk;L{hJpiD<Q&_eK-&0^A18RS^CU+^Tkt?=E
z3OgXpnE-*8XbPignVYk*G)(vm^tY9s_}jOzO<?DiKF1rKfcTMk-}m*9M%bPj^q9r@
z)}UgeI&wfhmcLick52>?JV-A?z(m@Om27`GpyV|PZ(lG2!~9EPqSsoZOz+bE-jFim
z(D-;a@Bst^NgmLLS$TPRuX_mL?Vw*B4*Flmp)XG>tcO^D%L06rTt3<b5NsHq0dQU%
z46E@DjgJ1V_t*pi^-DO$c{I3L+7da4LNwRpSAcyG1PqX(Uz?jVyk@=eYNq<Zoknd|
zyd@bfPIm{vPr%5qz{~Nqwl*nB`l8@<kGps}xOIgMk{y^IFcuJrs>Vp?KX*drb|2p-
zaG7$b0a~=gF@7f>5hkMnBMHbRkm8O_$9iAUA@u?nZ^PLc1o&OlJUoE%D~53t0U+Y>
z!aLt-N(xwYK%NlMJj;3AQ&do32IO`ON*x{^CQd)KT3KEOeB|f3T@0ZHaBW;h#;naI
zA1iARumJ(qE7%-J=yluGQ~0-UrF*U>0j=5GoEZzSk$ovb2<>bo0Cz^rErFmr)Zpa=
z+zN{F^78qf*1nPqd79}jfF%T^!Ta~``G6gzu&78PSF)Xt8mJ_I3uBfKc0O9u{?M1y
zdOT)27{jV8p4+#){2qMldr?s~;0M-``Chi(^b4_sq@`mBn-n@p2+(f)TZ6iw$v`v$
z@GI~{JCKl&@Gfr9nIF~lZ1`XWu>VkU8T@KGoyFK{#6S%-c<o9T3DmQmqhZ5HWZnMn
zhZu+is_JzQ0)fB?Cz$*&c#J*=GzC!l_b$<Ae$tri2gDpFCujOA6GKCqL75v#uW1)N
zj9W-Z$YpO{j;nWL4`r)a1k6CAZ=E_fkIC_Gk2iSn0<H})`)r0RHa524#}CW@I3#v7
zcUghrl~-Mz0Hk1GJOZKC%HOuaD$+0g0ICZU%*skZpb1U1`1`|4+66#2<9D1E#rWWW
zZwb)M>2FMwYnN|zth+x0(m96g1Ad-~buWI5V-A=ffF6(G#tobO01w@_UMWL#=gu8A
zcJ^}MqQNDm>jvEUIPmYy0pE;7ZX<BGCGZ%(0NxLooAbdx0|Pr>4p#e9ySf<t`wFzP
z#nFz5KwSm`Eyhs@bUHu4aex-MUM#ZhkprApASeIy_dfy#BviUrxkV?zPSdH0QOQ6Z
zSvl~wbt0$)z{6+w9jF2kS-u5^rk3(?hsO`)Ycb|&Wex!MP~C3xrMP`_=3@Y*JvQ|~
z;|Gon^)%5KuxxwS)`4dSb{ovIgZz+VBmV}@+b8~8c6?%D1!z=Y4L|SXg#la8+x&zv
zGIy#3;CFTXlO7ln(%rf6@vxaySwUgB5m1BwV~PF$*dTmxgTo(R@lS-<qz3r6u~Zc`
K6-u6)ef(dk&ZH~=

literal 11681
zcmeHt^;gu<*EJ=m42^=Mpp>L^gVIVU1|gl&-7z5D2nv#efFRx7-8qs&cMaY7-g&;y
z`|I-uyzj$WEY_OO@cDf1J@=fm&)z%aow6L>LyCuJXlQr}^3tkkXz0=4m;639_{3f-
z%?%BW{i}lXtB<b9$XPef$0xt{4u(W6KHv98L&s2nJTJZfJ0D{g_XWLb<~V%W-fL>$
z&u5jWlhko{53!=T$}=j9p-nLzk1wxY7{8a6jFx%j9L>%~_npB%I%3DjEmagTBY_c>
zc3;@Yjks}dqA$KjD)nNZ-^Qeor)j?51_I|LA|krTvZNNnq^BPhVv&@VZYRDM92~5k
zjs~v0C4EXyZxDhZEiKLV-Uocl@Q{#*s6-l_o}NCA`G3Fs|IThKq4h_mJ|H66^mM%U
zpWw1Fv$7f<tqoYtH>R+Esf>of=GTnZR#x_H?oMq;NlEvEB(8m}5S0~f32GU=y>HCT
z&4u0h3n?Bwdh}6E4b%2`<Gm;iL5p7Ww{PD#Z;&wzZv`FJ`s2lQ>s=4-!8}B1uD$OV
z`erMwC~*jAAa^=CI?CGGDPtYg<}8wIe~XIN&kvTbx73=J;3^V-($aitYlY1%EDl^n
zCHrKl9tQZtHJh26UyeHF=H=!5{fqXL^x=C=%{N@+kUI}EGnRpY0W1vkIoD-8Zf@?$
zQe#{O1_sCzZvVhQ>^E=UKET327XMQVTJ4K3({Bsh-y0hnOBVMMzV`6)y1BPqUta!W
zdRi0BQ=InCpFc>Ux|F$8F%Q?H0f7eruSIXqmv^Qs!oGdO(1G4Kr5Vo)RjC*o84*5u
zB*RRaUSCfY)g|?mP!FD^^dl}V(C@AYdc(TcbSEAW5z)MUH6paSy1Fyp1Z!<=-8aQ0
z<1p%qq+Xm)N=#&CVR?Y_{wZNvW+v)2(ZGf>b8<H6(Vn?CSXCs_$7N)grtwgSx^ns9
z5Uee<_=z5NKdV3NW%ay1=+c7ToUE*`AGP9g?(5#2=+1cTHSEkaBwb&7QczO5ttFU}
zbLpX>(R!cL^~H0VE;Rd!o=zIpA9PZJV3E+$4&idvH`luzfuKJlha9Htc?Q+ji>j)s
zGIMZbRaVA%U7eQLufB0V+3E;?!c%=ZZ6V@X;D{HfuSbqOO#IlU>4v|fqa&PFB6VlS
zQpW?ycX_<o)}O$0alRbI!oouL?3tg(ZcWhq^_l>d-z#i9igunMQNMw#qq4r9-rl8*
zDDkDq5<>{J@Q=|^)!f`%a$!g2qxGTn?(ffB8+1~=PBp}@_oXgRk$tecYtzKUL_zOs
zLC(f=dNiBH3q~g=C()yR9x7qSVC_2R?+FP#k<=pK%{$kR4z7=ej!qFa4qN|N4v&u9
z4tto|j`Yi>T2Jtc-A~N%X+-6jNtd!EZdi&MPASkbWFyj5VB&1*#XjCQmj~U?qKYl2
zxbdk3(VD|J8|7i*Uek7X&TLkTYEQ`_ODFC^4r>q<u@lk>U8vVNcXdroLTM@YV>We3
zcXwg%R+iJ{573(Chf0n6&R031Jz$p!#mg?vpm#2>e-qMr?xW|a6)}R%$&iau)z%)s
z6PRgUj*=KeR?QEBeMZF$+(q#hmv4(7j?0SqMrnFRMo)jgQL_)a=Y9(g2;TCEdjgj}
z#^&ZG>7z#@RW`cTbx3|^<g`qa_l@hxm|9O?Uq*d>eTHe%2YCgBmc6qwyJgwk>Sg@i
z-d;h+4Z_)4C&Qi?1|uV*cRD(QVBfx4TdxiOmijj~c2IP8Sp+T&|NG}ZYrQTfaY+Qv
zRxzD#^g;z%j@Ii(X6Exa4(&Ixve*Fu0r?tbfzRY3_Yw2AaUmp3I;9D$U}eY0j;DL`
zLkG)UIk~y?-~)wN#`o{v`<f-rIKq`Ck<Q{b#~*aOPCm&-QuRz%SRm|>PNMg^AKedl
z{VPks(^K@|;NVx*lR4(R+*~fD*o}>^rKP2=AdEH3ks=%-N^M-Na1eNAiMxxTp20zU
z-9j7;bZ|hr`}<c+C9a-qPZXK%&eT{n{8ht)^R}ts!#ockpnfNGJlJWXg1~u29k201
z?Y{ZjQvvUbKHZqOxXwP#CPgKsh(bZO;wj7LOiX2-7xs+G$?{TCm>y@l#`_B`)w9;9
zH#wa$tyv7D+`sT7^E{hrT(YkTj*9BqbTmjbag_k|V)14pfJ{fP|JTb!vtL;uX5QYI
zTNfufy1GLaMGdl=niMZzzMQn=*YUnER)nwU-mdCa-<%@c&KH9c1Z-$5#`9Z{&bqcI
z_Ix9cq4$=SmcZ+@cXkp#43Jk;Y)3jD9rkg4(9-Hpd*xTXp6a1q>&OBUoS+IOoT8ki
z0viO?$HM#WY93ysD*^u7b+=|6+(9vkKcd92>u}Cv?_e>Ac64TjilYzdOhQK11p=i}
zZA+(96#vQ(+i5>>L)gH;K)qNGowMPX_)Dc_%kn|7+mVU;@kYyHTaZTiS0e8w82IYJ
zS<_w0Y@JJNa&mu!z?{6D-OC}dGb#-Y4Yf}^oj<{8Yh6D=oljyCl8~@4!;RtJGp@@K
zi(Y9ZCEr0*1Mw(jqiDp=$BMi=d@*rsg;T&XG1#wu_N~&hvW5lVlBs_B)PYJ6P};h3
zKjq2G%j-=R>(}qM;_l4W30#Z$OirrLxo@+3AB5=^n+#BZ>NQzy_smz)M~|x;)HXr;
zmHXgzP{rH0FK*FMxMsI@o1D*_bV&S)O-M*+#u;HnT|pt@ub)cWb{N6S$4B(&QNYAD
z!;2RWlLi8Z)-ymMJU@4xu@>M@iU65lNRAy65<-^z=QF6G5Nw4vZyvfpPIqT1>t^xG
z9s7k+l)+ltj8nGjodl^qe*C!LD7-mUW*Wo>Ra-f__zTj3970S>OHi(5CMzem>?oX*
zpARWIy1aDX+TPv-Yr+qDc0IeYvVyM@I><nHf8oR(>|ERW5z&3D;E)h>P2qZ^vj}(v
zS5X=eBveVvR!%Xvy}qbFUkZE9z(7PzP2IR<TQ>`uTO!yU14Bd3$*4i4qnSrx@J%eu
z#N;7l6v(gk_V#xlK1hLN<A@Ify$7U^ys~oV-q{BY4FyX}Hc%rlHBZO<{jt*1(}hhK
zoCyI&bbOKbxxiq3FeRK)AhK?^HZQ|;JvZyWU}J5&RsG4y37d8;X;7Q*VX;+0mIA06
zH7Da=id;3XPLYOpw>N@R+6Fw~U#+W>u0VSX6}b!x4L#q;h_bEGLQCEj8MmIe#PF4@
zj$%-C3K2Uy-h_i4<N^?4-%o6mA3zJ%&-q_}qD|GgaOULXWM*Z(I7dY5lV2lkpv${;
zdyK-um!VlsE8RrN%1od-i#cN^da`@jzvRHVO*?HEHz3DmQcl*}n`?L&U6cV@u2K@e
z9m1+AK%~yz-oCr9&+z*E;24RB|H-NbzFRe8j|&br0d&HoYy_l|>W2>^Uw4aW0M1w(
zTxA8IN5<W~c9-4)2ZW=5coaX#59Gx6#Khh;!?9}<CmJ-)xt|>?7LGh}7Nt{`L=vU1
z+Gjc~%BBQu7trJ)sUAR0st`Th-Fkz`LZpu$mx87<W9@Q|kM~t5jgOB{n+q$rJVlSo
zwF4*p9kc($_Me&>jgxWZ!ty~yc>DGV=tK_!zG3;^)m#&bM13kgQomQkCgDF(#2P*R
z@5yOI@Pp_^uPbhY_F!C^?kD1&f`R(c$;pZ!3Kv1pbDYD&!^uW3kzxC%nib}c_UexY
z<7v2Kx<7%2FjM0&RqJGN&i^LF5>!i>u*WDIU@6<D--foszntq8tW3`=iQ$U|wjf!^
zAuutK0>HGX5<{H6ICjJFd=0@A<<2Nt6<JwX*G%Q4flkoYTvxu!fkR1Mk9Fw8z{GUq
zM|<t_#kyv-G$ZSc5jfb3pd_t<F2djP?!L(WQ62XtjBsFh#nszejq&|puu!tRNXReo
z==5}Y0RDQ0hM2^~Q+H?UV&miGz!6RourX=%e^7rm=TWW2VKR{D)Z#K`Hz$-cCk#8<
z-pT5pn@Ju^+RV7LYg@>~PiEa@CL(!iW@c7@Jo=W*^R$D=1bW^^Te;2_(rpY1_G}ho
ze%_~ctYEEK<&S)abEyU~B<TqWv$XM219T|?yUmU6ERva0Q&!j!jf4`H2pVV7kNN_T
z7u!j0TXEr$seD8=kd!rrpwcJUt0x1VpFgW4BYgGAJ>Ao5Gfdw~H)7ulcKoZy-0tJg
zY~8O&Pskf$H&t_1>rE~8h5nF_8QeNR(RK<?$HSI%3e#hMEyFN@&qPvgC@$C&?;&Bx
z{e$<)#D?18<ij_IP&Rs^SLpOabViB(iXSi!LmcUe9LUB;GUPPsU0=>$&7A=RfW82L
z7RbyWHeTCb6ce}uT;{#v0e+`(-=HHVCT^V<JHwXn$(5yrYPtB23KI}SDlL{BVuMvV
zd@vzq%1dG{Jg_x1A+A&YhsMQNfkS!7x`AL1HMOQzG)T_Vj=wU`atU@3cdl#r5I*u8
z`OE+#696jEER|D)Wk7|@GAsQ$LZ79S)Oont2Y~p;PoGeL4b*_4XHlXUk^sElyLWGq
z$>^?n_dfLc{60reL=-+OONfx3@ZrJYZ|=0!mTY4o2LHji`rES<rUwa(1HN&4cTrb?
zPR*mb+Vwh|#>U2=;Azs0Dgdnc=FJ=C7cbIvC1zKiveTRFMMOnqL+@@E<8hw^`p)l}
zwg(d(Z;bc>I6n%9qf&#9k6n!*%Pb}-sc){qqiydYXW_|0@e^sG9n<+wdP~mJFVTVq
znvq{?H0^{e7lX!TltrAAUfC67&w{mo1buxlMTCW!`JIm=u{bQM!Q%|vDX)cufz6;$
zl7fPwduXVOR|x>=0QhF6Tog^dR!z*-mN^QM@mo$x8XGfU;^1RxzFr&uV#C106B|J(
zfU!TB`SCP4e=)pk=w^zAN^G{dC5A^)UL`YgP)aEy>R?l>hh508&6+DyMbzV=h5X*5
zKEqg55nDVhv6z^cUt%8ol7i<9r807IJYJWM;Mq=7=v;Y}UiD>Hq-bT0d_Jiq=8t!C
z=@})rwzlMzlsYyYmkwNu5lGSN{kD9ahD41r6A+{j&;oOFLvuc4>+H`S1=+)XjD&SD
zWseL~Ed=Cfcphrpv$5e^xw~%rIydQ<>~7-&t$h{?<Cy89Zo9Of*l!d)?{xO|n&&E1
z+8v(@^ZN9`)lobiH!9lfaP$i2;%YD!ZAao$A8k`ogpwM&wrV?;JSb2oB-C_#QN#1e
z2OvkU0y80MY%ywfQ!)IyS7%?D@ayn`X46dbaq9+}SGnZFEiQz5%|5Kx$g}RBf7=n#
zc>4n$`2hS4>mqwTPW>{7!c0cH6n&caflG|;;rj5p1dbF1k@Cu$j%sQ}%J1R&<m8Ot
zuVQ#>MxW|HsC?c_A82Y~Pf0H}V;0ll#!1fKgWT<{7`3PLqjprzlwyWj5ZY*;wn8aI
zwXEq!-#?RaD`8N4O6&XJ6D*%p`~+L65zPYsXW5yNmE2HdIr@ojuzVg51hYt@+ab(e
zy)^q5d3?&9m_??n=B@$#3(1GLxH)-w%eC8um}#}OiC_fVaTbvZg!~4`0{{+q6BLRK
z84hG7&e`GjBl#I&*<JLs*DRxl%;@5l_?|o&$x9`585x=W6p?tR?eT+yE?P|X<-PM8
z@7v3XCi|uK2dI(t6p1+d(=wyQA=WnXB!#kaO_d@M)AO@V(R=r2{_xFfO(<9eC~`ep
zl+emu#m=l#OGHAfGucb@7#M8m@B57NU6-alfYE&7=r|b_e@TKUJ$Y$>MZ&0<1_mH%
z5oa`ZL;cm2m0~kEquAI?2pA<w0A#dxaNs0YxCld)(dKPL^)^A$c}xB5KM&3u((-a@
z*iaY9oQq?)O4GQy`tN9kC~_v~wKc!gW&2T5!n5;RyiFum_Gl_Nr2xl(p8q>N9V5Ok
zpbO4jyjbhBO%5hw&~~5jnCye0v&>?mun#}yY0A8ZN9ff1cL5Tl1|7G=m9!Ieqav|-
z-9~yt<f<fwH4=4D*L~GZVfrjt(1v>wcudtx4DNl%Qq(`$8bbq+!xpY2fX)TI^}Jn{
zxb;KWG*M^6L-fhRsf1!!K4f_otPzGl05UK=T<HPh>v7CpoV;>yr&$#dB%!^U>NwP)
zYFVrILW_fJxii9WXR_qingC@59r@+8(}U3p>HhZujEsTVFmDPt?@NduK-8G~00RZy
zt+n9LVx6m&bzS(eyji!aG3iQ}y6NE*X$!siQ}3$wI79Wg77>}n3Ki0A$Lbb}H`i_W
z=uzV+2C8LUAQOB6kISKacaA88mk939VE?jFk69EgD_HvR|3vN)J$m}3%G!9U)Oc}S
z_{EDC>HyrMW>i4pe8^TV(QCm74GR--RhE%{Zh-MzLIMUbQvgg>;%o}cg3_S%2I_N}
z4pN~;0{|O1UDFi&Se%t9#L|u~YW-Mrn!ma=ZZ|Y!yRH<8TnwGa9c=a-bI^hQP+S}h
z7VdC`ta3P9Q9uJzRrOxu)nu&`E27e>4e9K#p5lTrJUJO`mgsprBFCmvudb5eLcd<f
zR%M0ABHVlq&>#wJjOBjPwczSF5<>Eyl8*0RFWuw|t9Rb1I|nm(jD-$AlJ1{N`thTE
zyx>bnWF#X8M<kfl!NKYL!SX@SX^Rw0+8|Q^NpNuCDJ>(TPmV1Wgx_IZ1*n6ZNFA(;
z#HHosCj!<~puPCN7A8SLKv1WqrWQ%6Y+6kuWxRNFbfZB!yBEPmHB~4Fa>}!VYcNoW
zp`f4uxEY5DZI|y!NpCeYo;a)z{sB9U8aNbFgnt4WipQ`M0<bk2fZdGj?08A+29ISp
zp1pVx`tRRI_p{vsw9V(B5&$M8Gc(h3dGFGXF_zoi?ku<w>l@?pC~*zc@s9u3TQeIX
zhkHx5f}r9%pP${{eBm$7PIZ5AJ+(2C@kGr1IKI&%he(^jo<&rcmi0=L*(io^dFAJo
z+`HKg(?KZZ+bU*UFBW1u$7v3FD-NT?Nh`wbN><TY6xuJxg<D>z_!&A=|4*bR{hCm;
z<kMzA_WjMzU-=zGLr+KNOTwgFjN*<?cY1El=7&%hi;7G^Dj1oX1~uim{BcosKQB8f
zK_K0C-?Y-35Qqaz4B(`ZiOJ#Z)gESabaeC1R2f+O;{IYAm8dHrY)B3O_qKUfc6LAw
zN)r5Wq6hOTc#iGmu~~pT6A=WP1@!#CfB$Bf%FxTV?=*nhff0e*^V}AoQ{Ol=z#X>o
zDq$1Q^hZIj4NJ?)K;q-K=n=wsdj<xK!OgKSd?U5dE~ddGHdW)m1R%-7Xk{?LmbmQc
zfI`-EpcgvoB%??&4hC=3&|Ohc0R}F+jbS<v#`D=d?=HY;g$|fx{QiwzH%o($+s0K5
z#vc@S0}8=^pf3YMKv2Q^eHZ+Zkr73hc#2MgyY~Kzdo{H({hPWbCXBI6Dp}x|fMFf4
zX<mhy)L=AI;m6OPBa@TzZf-)L&R+u(wmxlMN~*?1`C-(c+#=n2*e_N>RHF=l(tybp
z4N!&nQ0F#0f$}r9kl=2$U~n0*yjrABN6kL+^E%Y0|IRR#k{%5G%gFXU%ud*HNBI3f
z)cW4<tdSuJWvXt@U7FB=aeIMx-To%f=&$jqIK*ZuwHV)nO@<F&rwe^iv8_-4&It8!
zLp&6EKX$+Q`=L9Hry=EsyW;OnR-8MN36E-enI)g6C-@AQ3^&Q~ViQ6VkHejaXlD;q
z)9<vPKb(U4dL*=m?lgaP<s@b~Clqy>%)VkwXJiX;ZHexQ6)?7~E>A5L=4`5UVOFr)
z-ETrbFcDKh2(8GsFcLmj%pd>U<hQ{zdpj@$yKW(Dy5fQn;Q!!u*TcAT=(KfjuWr;*
zcJPUrejD$QpYCftGKd@Zhi``^Umv+>FH5k~f|R-{-zt|GBzupps;(}ttxXNmRQtRP
zv|dYoN>uCp{o8kDMhitxSXx@<#d+;+PIs=Ml~{M(lYo{)+!(;FzOhT?RPXSMHa@k^
zrcN5E$eN&OR43q1dc|cl%0Atrl%U|R7_0E4p|$U5$4W|2OH;+zi)Uy~{8vUcWjCjV
z-V$F3);Bt>$hg(xbH5+s?ymN?I){B*H9|Zg^AcP!pr$adPn{T6VhyXE<nZFWwv^E6
zfQ{33a6N6xPu<3Tz8^(`O@u5Z;r@b=bsZj0+mBH4)({ggW;~eon`)O65>(Y0f)+{y
z!mSCpmBc4?WzWcQiUw&qHeQJZJ9|6br0nHwJzCh<#HMj`VER*`V=&`<d)Ph6Pe4>R
zkT$ryK0o0tYZno;Hy<6jOuT4$^_m)L=Ek8)@ENd_!VQD;poIVZ8(ibCPE1B72LLQ;
zSO!D>avOofbJB-)dktr$-Z$<bFHj|2MdHTsT>o|ab}LVFntq#YC9mBt4gqx2z9yU$
z*qwIsW~*aOn}h%&`PzNpW!&lyE^Y>NAHNPKMMagBj|D+76a-~m_^5@?22^N4mHO6=
zooEW1vdDyFD$v3e74)`C-ds9InMl6oMMSE-tcRW{i#A|8lYQhiR42aU<qI39>S$f8
zwmrfXOkQ+%mB{uFF%zp<ou0F;Oii@HPj;rH#2wntGY&#?HSV-+6k+gNlMH^N8x3nL
z!3~<#8|7>GYxy-xrHc8IN&09r#5@s`?%@kcyU884zPz=4O4w34f*dv;o1CusI;#iM
zP*pwl%V5qHV`?>BN>|q@<^m@yZhvCnp6fN;YYwtN{RP{^>#nuf{#otb$$zP&6YhfA
z^NQv0!NC~WWPXvo)z60uIoQ^{@(VCE&(1_C<7z5?%Rt{Y>vtMt3pio;qhiwvpY5wS
z$J)r&mu)}sNa))8<aVyvHe7zAyS>uZm`NS+%W(1z6=NKE`JUIken(6Lk&C|MD9q{l
zi%YM}Ff4SbLW$|(CXhtX61(P;j;cq8epo0*^Pa7K^_MNK75fj1es{NX==Pyo9JMK_
zx~PyVi!}7QdkY#9&L>4rn>&riLEz!xWTZu3sv@(p7&{}W`_g5?&U;k6TLI@r&TmQn
z`(1E~hzl4fTyHPd-By39xn6E&Z_YJ{0M2b?ZEfl73Wf2}Co7Ejoc8M5hW|b9HBS_L
zKi2Nw(je%Wi|=EB<7j&I{SAO?8wjz`2)P$M@}Y+Ck%zB;56w?mN_d=Fkn0WGukgMs
zd9v>1eQJ6&^~_?Jw*&zJ+z77(Ox83d^%<$2&8XhzOr0SQ3G$(KU%UF-^}DWE6>+ng
z!tJds6brLGUa(x!MPoRe_R8PqEY&;Y)tos$*}q|K{wxJ|RVwGCmS4xxhoReCKJb$P
z($5vgPLMa@zf}YTZqpR>KardapaWLqr;o31&w8I(vi9w;TT{~7l#>CuPl)fU`y`6N
zVfMB|!A6P;ewcrF-7zPvWZw?m)--eqibNcZBWk#nn)P{5N5`L>>`T<Ak%bdKbl<d)
zsY53!eoPV$O@d#9w5vwb-9}XSB3BRdrqn)$!L6U!R_gfC#r{YgnX$K`HK*|O^!)2@
zxJ#4&#Dms>hO!T2_&FSQ+in;oCab6z{zv%e!gB{v6$Ny=?<pzN-ha>9=5KV}p_Q=f
z7n~gxG*qW?YwfO8ej|SF!y<D7f!7Z^vrn+})b#OK#0)1rGl_$4@+#p&+bt&Ki>v-N
z#W&8an^RKgpYf?wTKFbP7PAEwidgiI;tE8Mho$1S^2Mb4AISp(5okjIN-x;g?7NCe
z%gKGv)>e8-_!MA+|EPL^2GAAu5Sy=JjB1?_h(yDF5ZGdy<jYBJFgoRXh4GsZKDg^c
z2VX%bi(G}7YitWmmg*<tMLP^IRqmso>?qqLpVjCSLIPUjSLxvG)@AysE@=+~g!hgi
z@Jw!D^rWfE3~GeW74(|=bnS78^AsHf;_3TTWiF=W-}dF=5l(_p)gnRQdHfl^z2GGT
zZGzEdS$B^%60W~|XD!%?Bg7zxtc?5OIS|K3ENmm!XeQ6;-XUWL`pGPlh3&KzSw8Ft
zGfnq3uZ*>T0N&<6c4{nWOAoN56ltfTvaKmwdy@@xt~LOZAyz-9NJ6}<)lU=q8dTIa
zV^ekQai2v@7Z|#9-Hx=7<%~eC1Yd+Wg%|qxN}R6k(qV_Yt%#-N---`olZro{eSb&s
zi!L_JEbt0yiCyo;E&aMYR`+g2d3R$xf)<aulTX3UO4G(=urXzAbg_NHb*lW7CzrxY
zbiV+7M#3HKRxenudj&lRPdCl@Dq*v<%+uzjg;*Ddx{dCP^3a&yDzb`1d5BAjw2PtQ
zqH#PCj9MsFrck>MH58$=ALMk!jOWq7H4w8}xFD&nOAjio5S@8NjlJz5ByC<#f&TOS
z<;#O_Eu;~1UbbhuZ5P_HgGb)Zx+`V+hWc8L6tq;#xB~jr%;Fv)+Hh%wg7_&z>}XwY
z9g_D-B*soUg+m)1jb-u08LGsS?&UF!vl-;W&jJg&a>?yZi7OqHi+thteqI}?+jRap
zX<k3l(N~f>+nr4ib>jvCFIoMUR%awvpsq-`-OJ5<!FQb-z7G^@p;*dQM1`zmTm>$V
zuhiDixsbbT20WUB)=dtxf|hm4i^Y`Zm~^+dW<)n6nsE+&C2wT$<QERE<f1qDyNeF(
zU3<`TDH7YZyC7b3Ki+As)49nLe~D7F>kZa#{zvC|s&TJjX=cXz?tZw=`*2$R!14{u
z9g$CcnHKJLxmh6}#%O3j@%QLDVZ!o+dgCLUVXeq=qIKu}_@J)nI8r%t0Vc*_$VV?%
ze}dC^9ml%7zcyJ@Cd-lv)TgQJrj<9u=^;_}+~N&PS1MzwBOBS(Ss~G2j*m8iLT_!C
zhgl1=k7~3yIwL6Mo)T84J?N8HVtT=s($dUnT^qtfBQfz@Rv%~3rbpsVlE>J;C+L8z
zu<ED^ezg5Nf6&nE3Yibz)){Li$j*$J8R$6Z?Iwr`Hft5`;{Bp}GjrF(Ob+qhyr<d~
zMQaK=dVzh8%YiaytO<03KII<_>b(o;TNRuL86O>-{h5UWbi(!|n?^SoCH3`wm;~$r
z31f6>s&|ij^)F6m8r^<1_H`jLN0Zbl@7Gxhv7FrmYe<vXMf-1bu{mv*hnEuGVdU|c
z2r~Y@B262#FOi=Q;Q?2F&CK1<4*!_7aR?Xbej>O{>$#>5xP~uLr6)TplHA_jJODgc
zJF=MQexRTRQoj?%?@-(LM3<{wLYchXg1|(V>xR{;Rd^kgdh~@(N}Kxh&F0lA`1qu^
z$o6rXQ<Q}j1K{Nbg~%v$PX>J}y{{H7a1tCO5Rm0>eytti5_hK~VK-K4zjQomwG!eV
zMmlg*LHPZ`3Tcz{0Vzm2=Gy8-%48ubXxSOxx_oU0wQE3E?A_?Vkw|#m(4}(8kQm#T
zOEZJ-_ax)8tgbjvjL9N_pFUx`NC7fWu~UqYR|#E(Wt$wVM|5q%#X_^K2dn!NLm=@8
z<u8W+6&p5s4oPYHS3lr`XW5wtNuSu2jOSr7vOnO{gDIrzjp?Rh?(MbbLqW2-hq*f*
zB-jVY5zN-AsL=8t4x*=7W?Ky#$md%-1qlVKVO4r^F?&kn?I{y~(4tA<2Y&ooySY26
zrbPuOYjSYjXSIX1#8q`pv|LuW+Vqk@HmP&jv#vR!BSpzn5Nx};*>W|qp_)jUY`9xO
zlO=VQ+0xO)w9_-(fY(4Pd#|fI{{oJS?N@)bTes*kBOoAPScO0!jTW{onIg-wk9^;4
zTgp*aqyTxNvez9bV)@#2abUy&Qs%Um8>B6`8=1-8E9!O0>vbu8Dvv=l5n?UFMV?t*
z{Y&zxk3Ko@oD^G3aF`6GDlz{%b>Rl0DZkY;Rx_C1r~fEMb+rL*W`IORbcBQQ<YN0I
zgef%*(5q^q13*uE!OHqli3F$*0Gg2Vm_PvCZ`G)%hHyUY=NTHVvYBU1)p2m(0X1Y%
z*kW(BX1X#K41|@Je<=%c-fH;*l^PqLiv0FwCWR3YPe^_O)4^o332sbZG%(CPV`neW
z?)4j*u(0sQkpwgdP}COZyTYHe73;T6BlEY-r<FCIiKhM0SiQC?Y>3AxOhvYHg$Lsk
zY|yJc{l1WAEwXEF-23x=inb8a;(Nxj2Q~slg9CF2phl4M1SQ6OkAafn2ln96qyMI1
z5cxpV7k$j3C2wub5vX5mIz%()b*7z|IBi1iP&Cok`;%3p#@^UfFszcf2#hc&mLHfn
zDz+Rs83hF58Q#Vb-p4}C;Rj0{V5BY{=mo>{YKqI8KHyvy8Nc7deGL2nb0)@*!L|d!
z9*AQ5S2y){w^zUr)G*d3<ay2&M#h;1e#WxB#DT{FJP|c1gLi@6dC$46r+nUys`~}l
zz+%?}1Hi&_LHwd5ZhwvCsa0w4eE#(5(c3sSeXxCiqaQEQozGgW2;?Ap3cw=pR17zG
zxS*um`Z-pp{be4IVGu|YK=1Gh2pj=k94HzDYlA5sE%4d`tL|tzRIvfO@wq@VK%bmV
zs~W84t%AbhN>41XpafM|OjIhk;XY<-HYqCsGCq)^HJzq)SL|!lo4mz9Bo~3*q7z8z
zKsE##^BCjmJq&ap253}SQv+|35ztegKYtFGEr7@c0WtI)*aHBGf${?ZJCfyOu`lR)
z<(f|b1;D0R9t^bXwSlAwb9e(eocG}B21Qx@8y_!q+*AcQ1)MU@Nk7-^Yc?i|XaNso
z3<OkQUkGk(eT~`$M&oIuGk7-O;BdiJ22Rc*r9KNhZ4ODGYXG4gaPI)UBaR^*To53o
zNvWz51KJ9t!fySMa@uNmExVDiF%cNUfKvrfFwcSC0?6g@qobqzEPFFnb--+tk&_c*
zQCO3u)XU1rX*ykQ2B8v+1~S3tlv6U`V~Pf*C7>^smd0$tvw@gDW3tT!h>GxVh?c3`
zfm7Wqim_T-TjQ^@R|I60gaoY*pwUoz4j^ZM?(wyP8&HQEV|ir20hLi$Sg0gYuFXLP
z$eHIrri+Oo2D1Dw;I;bs^QV<+sy;a&ftHUhs@D=NQ2I5f2Wx9dHfunmvQjm9{Tc&&
z?ccwDC|-wADe>Qb9}=J$zhjox(V@{PItvZ5u1*H3Bai}p<A~lK{fEAjldDp2!@;LQ
z>8bsRd||-M^Zw(<b|C0*ZiHJa0k&vm#mKth<TE%7z$j!Ldk8q?o}QlH`T6-{!kbN@
zU<-0{9~Mo#b#&wfa|%GO!)2yJ{47Pl8wPwLRWn{^vzgi1A9Qt9z<XqwRbD#qtphs_
zimP05t~0Z?W@KUt0;~47NwCHP1aUUt^MSGhJOGILs5G#>K=c96WC1QaAdOp9!q8Dd
z3ldSqV>QhS!qGi2unJ6G1-1A?YTzNx&isIkSOn)8P>EG$ZUw*l?gh_Uo38aI<g1r_
z1LmsifjtA1(+!wwQ2sJdg`@sMEQa%oa8ZeS{<5B}J?LSAap^UqDAq~fydx0ZmrVLG
zF*(@{&NJACwTpEalbBe8Qg0H_j8H=A`TnAncU9l*%~i>iCCUm2%HkoA6?m*?1TRmv
zCm2^#1nn@g;V6uVG7hDB?CN)fJ_1|59h31!Hw5_4P;Myj_9!z5fDjor^Nr!;FVn|*
zH-Nz*mQl$7KuXVbk$JO<O&vwxn*-kg&ON}=JrQwcqvW^z3b!Yq6<1VLgrH12;CwKW
zLsS8WYHj@03NW!Kr&e{fp!aDRZ1KQ#y1afot*#ijSYtR`P)0^%8~VnY3Gldf{$F3?
j|2ka%|FrNP-l6k9j?gQ{#zO$>HJXBqvUK5Vy>I^qaW-sH

diff --git a/_images/sphx_glr_plot_reweighted_l1_001.png b/_images/sphx_glr_plot_reweighted_l1_001.png
index 66c0310942e416a6376f03eab5f825a573c8202e..f617fd886684229a7ec952565e82feaf67d1fc0b 100644
GIT binary patch
literal 18582
zcmeIad0dWd-!FQWndnAHq!2}=NSX&iiUy_9Bnc_a8Z>tsLPW0SXi}O7n&*&Ik~FKU
z0hKhTQKeen!+o!{-t}4U-k<%h{pa4#=XoB|b)DCF9>eeVosR3Cs<Qm*m24|1idwC3
z;;0%$E#jmohI7l8;$K|LoJa77nC-E%w(6FbZ5<44E>TK`wpM1Awq_<5IPEXlTr;t>
z5ZWiQPneJMimk2HHE{s}^Z)UJeU>)H0zbpTv~iIYRwuNtQ53Tw`5!}^Osok-omixB
z^zfPMPY2r^TsCOTFZ`TR%HPAqz_ofg7t1k0##*KwN?ei4c~1Hkd3bqn?%;JfCj0pA
z3E!f;b;q<rShO_#_U<@#<nivie33jWIDP)w!9bgydo7?Pu!qa%gOZ~Z$D>)!sSm+o
z;ZIgv`+m*%`_}u0te&_P>eXW9hik}RWPcuJWF-G^kc*T2d`gay@wdC{%=qaOC+<d-
z2$rVk=hwe5lyh`+thn&<N=K0^n~23zzK!>z<Q7p+#ye|oui7+pR)mY|DBe@C5Mi;N
zot^)F>9fzD&(_iR{TNL&$UMu3x9EjgZu-8R;qc?jFVEVHw-h+0N6A%v8TnfB=XVaK
zA{MQxvC5&gKffJQdU8l<)uw%CO1(A+B}-U;JDRMMxmSh0OtpYr-1dZ&)3|D`P4~r2
z9gQqAZM)%i{X1*eZEUYJ?@cNz`Q3MNl<vMTYe0Lj<HH-Cvpnym=f>6Zr@~!E?=>dt
z#^PpM{AH%QeD~_-Wz^Qz;wn{4`%hK|3TWZOGfe7s3!Aqbtj==oDqk!0>&wlVFE^R%
zi{0IIr>mZb*o}Qx%FWNGzgP6*i7cKQwDfxT@MHQU1B(H#W|C^QMcbJ?Ta$_YCY8v@
zNH1^ihT{1Nfw}qFVV#2OFT>nsOmye0JKvf7{`FJbeZJtm-yYw=EOj-tuC6XFEv+V(
zds2y=hNh+o7Ogqk6B~3mQ+M=e8PL{aRBOIhJsI9Ry;<`5i_6tv$sd%}Q}y$AZ`!me
zHz%j-u%Ff9$*o=)+oYU67k0cfx`W3Ij4Tipa~OKc6@RgUX5YQ_14CDAXW64f<La=4
zGx4-6v*sN#GBWhENVz+GKg&3@s>7r*-51<663*;tqC3<F%76U$(Z=51{sg_W!JwE&
z?(19rv#%B$W(yNuzI>T>md|OpaO$A*q#53z5Ed3@+MM=WIb52*L~y)O*D~H^dUSMT
zM0Iv@(C+8AQYQLbUt;R`ug*u6m6a#k+uKd+Vyf6Jh3`ImsEHxxQm6_Mt{m5VDR%99
z+;MNVEM3;6tJ$J4Ass)34w=**rPar)b8GW$q>Eh0mRWS+<yqT_J`LC3Kg+Tv*{$7$
zJ)Rkkc2y`UD>uq4OsnK~$<Z*dQPX2R{e2CIPE&0TLP^eF-z>2kYSosOmd+b4nmuUy
zLsdWDzA8oea<zwEk&AYT=W2B;s~1U{i5Q$`3ckDd@bg!toYT<>x%?q0^2G~5%z0&X
zwaVkikG<~Qd+ay&b7az3wk6B_$tA`O+qR|T+4Z;hyUz5+CaA|gl$4YttH5rsMI%-v
z{Ntp<9tCcdu=@J~MRCivU6<XqZ5z1<@=W=OtHpXxT3{Mui(F?DFluQ!$tzc`RKvgt
z>6PX@k-jqay(%~>yuBmFK)<gp_7NtZzhYjq&o)?UQJTZ>xlGdrMLg8r)6Z^&UH`UX
z@bk-(6y=$s87w(rmtR)X)6*3gauu1mjE~_m;pPM7x-M-iT3;K7yT<f4rCu%f*`9a(
z=d*U#DSa|LSl~@qXcLdzf6EpB9=(70<J0)HJCFZ}dL*5eoNx8@$iz^a?%K6$Lk^nI
zXuiI}4numvru8SBr-rIVzZSgn-qQa(@LcWeaB*r)lzji|WgLE7^y1#^T^Uzujx1WX
zrXK%F-ZR*ed9uTO!8IV8Q*z<=*j}xar@NJ(#s!{pQkdv#P?wdJjj`LbPcLqMZZ^?v
zZpLXSw}(r?pI57hZm`qi!VqTp(UBuZ$a*r)E5LqQxnf04i15_|eEqGt>G%!Hh7Hfz
zCG3C9O?5OqKN&b)!5x0F`J9bS8t(g<f%}|Bwah}n-ZOC#=JYI+`nc2L&XY+oD&ZE1
z^i%_xWAgHe8i|^H9mT~K?FI3eDP;-!0oArVyK8om1GUfZJ5BUG$Kr`USLl>A>dUQS
zTK14VMA9)Wd0;TRqcLc|QGM<;<B8!8gP!W}rVqhF4LR0EgKhaDCH{TIEtyvXv&`u+
zD5zw%Nrh<5z1EOnqRuYuau)l@Vc_I}x|c@sq{_*8GLZuOF;pVfuD`Jt3nojt-(oi6
zsnh&)&jj7F_>xnl?d3J>;?ZRs?s`}SY+^Ridiwgp?(;4vVP`trW|P?tUOBJjj%nVj
zk>L04A$uZfvQoIT{>^2~jkH}ShxS->-C$tZd2{LN18Fnq!)^K5r(a$;MxR~Jt;|Ye
zLWSrY?XJukJsov?GwlJ!7?ry9g}+08W0lplYeIO8%dDP;dbu{GJlbi|`_I3k^6kHU
z`^IEo(81K5t(#>QyuIRchKaBKbndktCJ_;ls7sXr$5G*uT^f_LJ-IruvfkV0KRIN6
zfNOJQ0Dn3!hlIVEXNPxSU|>zY!?0QQGRnbbdi0VUtp+>9tUvQ=(?_3+T)TI(dH(FN
zj>kOMPn>i4c}$M>v-6Qm%lo~WNmm>SocfcqnFIs`I(x$1vvj92o6ZS64Gr~Hk5%66
zX~?=o#Nr^=W`V+SHP7p)|7J4Ac9ME8&i}o=Le%=(Ei}i+K0fEqCTn*3Yn^+UI5m_%
z%(n4<si@;<G=8^s%a*I65)xZ1i|4ZR3$dtK_wC!a`MqV})2B}(&9HCrP`)bCFa0ky
z1YZ1J`~0LQml3Lyc4LJShplsQNc%$Twv$fo`R89V>T281WpWt}e14fVh4~h2ZEbyk
z5>0>aHajR%LZ{QOwC21_=5sa3yi)VVqNC{dXr*U*db+;Bqu-?{#_YCz^$JBrMOjn5
z9ns@IYD$KJ@1bt~n3?FGnP@WDrO@A)T-RM0*pRF%qMu{+dJ^Sa6BCUN7nAR0JU26u
z=Qfw;xq8cP?pT?HRm`k{vRFgGZqCc8hqm4CHz?4uvmLS~Y6H1<?{4vDW;<9uH0QHj
zj;oDr&mLaPRUHOAMeVs_Gy6^5j4P4#^=HhR(~i-`r+dOHffDdoo3pG%nEhve{ffhe
z(#f-pZ(CG<GJvlJlb*10!`>5l6S8h7vEmLxW}6JopN~k;Om-@hep6CXGCAhGMPwX>
zBvH|S?_Mme$$@6y{+zn(-tXUIP^J1Yu)&RcuDGgVEfr0F-(vUU)9aze?1%_Xso8<_
zmOQ(wp2e{}Va^{dI;7`+bt)<;HB|U=E8Wk}>D<WoWlgQbecRrO6INDMG2Ok(293%d
zD$qhiEW>*3U0qy;_J)lw>RQAsd0uM>ABQ`zrk1FFiRRwed+&s1n&KJUr|DFYP4{~n
z8=Lp`w$Jf)lZE-Y&0!)6fzq61KRe3^Eb-r`FHZk#6Lm6hpH8N!f0kuO3I@CH*U!$D
zLg(x@a|6^3UqjZ-(z<yPZSTghhk5qyRa~`^?^9~=ylLP77Z+EnMyqC`hVO2rTHe%x
zXC5Au)Xa}4Q+if1sX?z^$to;KWhOuZwOHuw&bGV;K7M{B^c+oZHes{w?5Y0eivF}{
zKHq&^QtSTuvL=nKjVp0<T$yujdl^ANYg*k)@2XQ?s|*YbmMmF9o}sysUhH1{-k;Z@
zOyG)*&7~c<po;W#^znyVgL_9YAqx-bIl3xR6Mm9XQkk~BYQkopRq=%WzP`SLe~*m+
z0xB$;=}SB{>$`53;zjy9te6a}_0?Mro%;Iq>*&OUf9ZwOO)2_9k{83bpzYSFt4KY^
z^62ePEhfd2)Nagug8g8NL!thj{<(mV5GNISRkFlHN4fh`28L;_Ifi5I?(WzWJDU29
z?IM)J503n(iLwuTl4?+#g2L6Rp_}W?K-nasyc6)2Vb`yTHpNpiN{Tt)3#p$%e5mR{
zN95$phux>UH$8v;{E|eaefN8Re?LFlGNTLl+XQw||F_cn&F7X>RQ>#fK8g!Aq+O~!
zXxDd!UpwvSBboV}_W}G@n^Fu}SFT=Nn0WEZ^k{b<ifGKK;0wI@_JcZ}kK=~Ei=NB1
zxl}m%p4X|{U(0sj^Zteejg#szr)pxu-JDchRcHJ?=|&$05)_m>C&$KoV`F0<Zj%Zz
zY<|2`9>~z|>%%3Y+p0aNX|5eRj$n(1h+01bLVRpE9qRdeY;=_7_fxj**|Al*xw*Y)
zdD*jB-whUK0@tluS0AJF#F=?sV1KL0*<)1sJuL%Q4Zx2-{R4AJ=g&v;KX?7>5=SSe
zz6LFWsuvbBK1sc3`+F2jagzaq1QpP;Et{GQ+(QkUQB4Q{R>(f?dVom|wW?ZmIJN9^
z>o(?m+umpBH%c6Czarg!|CG1utu3{f7OIxqX>4rl%^@RQA~Ni0c<~o|+sEc+u@WaC
z<BuP#gtsRaaS2saRBZMXNzuzqD<~}Vr_pxww$DU(ZQy<1+}u3*TPY8y@M7GPx}4C^
zPz?Ub<a32BLjeW9eqHC+$vA<9p4YD%_Sij#k+Ew%2M0%0sDzf3^JM+bN0QIbp56nr
z0b<sob&Y>EdSrsGqn%=qsL3xTA#t@fT0!x|37-a24WO*<;o(%YIfBi-9y~aM$+I2*
zak?s4NFJD^vQA-Re|>xukp0!Z`uM%NSxHzrLQKZXDY4U;rhmX_UwdKxn>TL`e7NNX
zv^h81vGCFMIvwjV4qXbDte+gDPxQuy_ca@bYXI@4n+;>g8#WoZL;?~dsDw)s@R)M0
zFaceKEF_=<jo<^9>XWpECI*_-vG<A|vP;C2m!H6D?ad!9Qcu5Jb@2Mgd4*h=ZCMs=
zIzZNJ`!C!=t7`y+$nOSZYP_{#{Xwk#(s%C?<nF8)ot$jEx1Kx0rdv^Js`VOPeKde>
ziGs3mpB@jYfJW*$eKOtXA$A?EQ_T&PDFV4|fJ;f-Vb_m@N;qg@FuK2edxlHzm6kq-
zC-b-M_bO?dJd93+k?|#K2h<N0q`y+2I78~un3i2%{bR1e@!FGKcklY3f$@9Nm4XBh
zqdlod9rqspc3<FqY3Z@Co@y;cMa7u0_Yss;`Da~^o2^xX1=-B<K@Eagy(O~%jl6Pm
zckeDQX=-Y+a&Qm<Y&0A1tt-2G?_RD^0f)=T->2i%c9oFj_31W7=0Ji*;^myYJT?Xf
zhIg<2UgV83)n@%9ceCeI`Qx3(JKuS)W@cuVTPHWlm}5LUH7o)Ue|hneC7V-HQjVh=
z)u3;7;(eR@VmJcVa7df|9<+29!W#8P)7uOHbqI}xIX^F(x#TKqwx=A2u&Lka=ZYIj
zs;a7Z)MFmhfH8L#PPGfA7!+G1I$+-rAaERrodBVd>gwt%4GI1)ou@2xLz5of!DVK(
z+VUN;SI>s{W5ipp)W<#k+TDHWx})P}>_s8;_KRDkoDR}EIss}Q01mNTdhfR%MAY!;
z(W9*L54To~Px<H@jA{Wp3zAo#Nz|+w867>~;_B*+pII>gRjMM~uAc_y5)4`2fgyQg
zL9qdL9*<Xh{s3U9T@=g7)9|`wdw~hl<^yj>#>RwD(e)Ri>7HMOnMyvV8$Sa6H}vB<
z{bpo*{2;c~Wixsfh?M=Dt$p#PE~)#5BX3tz6i@4on>RPCT^q?P^ZQdxB$cx$jDdlo
zZtKiQvf{3-ZFwjg4`W$q50RwIUIj~>0iGCtyIw_9d@CjPivKJ{MSL=CObT{Gd0xF$
zR5eQek-e=b-nMu#Gzt%s>r4On^8~I!Q9<wg7IUG0Vtx8%XJ`MOn`?-aTgFbt%kEiJ
z)TNha2hI%*4PCyq3NPji4?3{2F}uSp@$k*11jy$(j>Q~*v@JP+U$+l>#OUbgdmAZK
zll-Hl7hZPdRApowoSURuV*7k-aH2etHhV_$YNf=PHK*Cw?A6&+(<?%84qs}t45E1?
z>E|~Nx!T%h;1)75xxyB$qLmXEM8-vnh7+ntk$nC7wL|7Uiqo02YHW0N_TypCIw+6d
zV@+-)+s8c6`A^Uk1NViR7e+^OX(XLJZDL~L)b&{YAcT;HxYIl3&96H<pUj&uBxr^I
zJmtZ)XaD}wxNMe@+36$243wilx?9QkVlH%V(qoMtNyXgR=EPI!#4I`DAwBbBr_}F}
z64bs=&clVV1eJV>lnZw251`SW`T4QXXYp_+^6a*3VqZaZB|p1<Uv0*i++e{JZg70@
z#ZM83#jlS<pyq^_7mkcr-M({&-Yx<pz%2bM;(K*?YIeI5HfvlQ4=MfCol7w0J`Sf`
zZ&;0dc|9{)DUb;as@EYg(ASp$O4J`{;8tNf9d|m)56Ig&RN7Ts;0RIPeD0jeWxT-Z
z>B+R2R2);3Y4q8ui9>4Y>ULjVF=Se_iPv=lB55X0q+~f`OY5aB8^fL_y`u&i8|7;q
zra979<ruPOMR~;Gn;@b)&_GC~^tylFmuq@v=6y+t9IZZ4O8_6BeERfhTC_-qi~1w!
znK&W^K!mu8T`LaoKQlEP&b=QqpOYi1U*saDl6B{|M%qP%B}<oXteE}CKzT@gtXmn_
z8WtN&^Z0-^M5_m{%O72`YGWLVJ1rXYquG0pdvEQo)xg@;*4BiRp=L$_R>T9*Ij&Qm
z;m9m5E(Wv`d&wWcRH6#b3i`*b08PNg>L?a&F`h@Xb<i=yB>Q|@1-SZPy9%VTx)`O(
z!0Q+lbN6}(9tK4je;4?=`450xj!sSW*z-}8mms;jd@&VXox0E8Rc8usVc)NLxLC};
zC^d(fh4OgjJ(lr?JG;lE^Z<iGS$Dyko32YB4Ib$*xZY)Nu%Z8-ZI^p0{O(`JRKQEI
zY+?*cJjKLWUMZJfy?V-0XfpIrO!mDuRRhK9^T+@G_M9^;_OVB6OoYQ?qF@0$p!V%0
z%oKmtqI8dO=g^PJ%U}Kc!f><aTROOLK|#T(d#DpE22W!zB@2e){jyD|27ynW(DhS?
zdwK%z)R66~e_B$dbF<G8(llqVFNKZXF+m@~f6?a@X=}Dg#%U(&#2>x4Zsqdjs*twq
z(2?nR5EUD~yuJy@D(JsQ{WuC4EjnI3b}t{_sdMMftz5IFk!Up7athg=3PbelR6r=R
zL7-9P+T@eYz0Gs!)zXRk^z%NKI(5EVO)fz+MJ{MXN?15pnf`a~+(F5T$D{UiTB6dt
zlEia7S{&^Yvfs!P#6}nlIaJD76Dphr^rn3VMM;>$M1Wn$N;TLD8#p*pG2`^)1&DH$
zRaHvRNeb~2ZUu0SKgtTacrPktBh(gQo30ZIYS`eyt~0jNzkYpg*@A}dnUm)*oU-t{
zM<#3GFWNY&E3cTCHj#k1_MkwLTgC-r$#uE|4KyEZlNyI`@x3XP-0wkhzq&70?g1D8
zo~)>=Q~&$#zq3ZQ(=W&4E(!EZ9sqw+!`-{hPaC(nFU--R(P{hI96B=T-wimFi5dgc
zG5%(iz?F|r#cZx!^G!`nC0)ch`#ul=sJ&*2UVNC8vm$LHzfLqRTS2Gm?%A`4T)i>X
zAQcrw0X^2A0QahJ83Uz|gSkB&=qesNt-}JO*7@xC4z@(V?U`ps?=e+A<dCtLElI)t
z^6ZAHrFG_v-@~n9&)&X0+E(DG4zi(+!F$NP$i~|GbkE1o2>`Ho%s`&gM1p4Ox#ZIO
zn@-Y5m)XV5-4CrOLw6YXCjr>TLS`T7P_G?Z79Q=a&br08ea{{RLVk8Dh4>k2tE&18
z_R!_!?gFUMv&@C9zQ`7$O%Pofr7;20RWB}-8{(Y-h0L@z`rf}%P6bL1l*FuGTqeMQ
zME+`X(o7UCGsoIyIrKEP?FggVU5x%7Vc~|9fyeUOYS5Zfe0=`O&FONP8fq`pML!5R
zaLMcGQwu4!4BOsXiLAh2?8m*LqGt)g2F`!394cPX(4av(Q|M%;ovaxuZs$c$re3Xm
zH!U7fwRnuac<Y0)jHI7`Ic?v#Zrw9LyS;}FX#zQ@f{_C(MBm>ekf57wK@=;DZwff6
zj_d4X)#Rfz5aK-RPER5ZF3b(P>*QEP4o1cnX9fECMT6)4cyoId#MqBE&ehVhWO)<y
zDjp+{Wn*b+>CL+Th(Pi0G_Z0bb{RK4=gEN@h-4F3?j5rNXcQ6D7fJrBT%5N*-(9=2
zYhWM=A}*0$f#TJg(=JVb;qVFzYw+u4ojH3ph!i2{SS<&>e*jKa#}a6WR`4DFB*)x~
z&8m)-pdU*5+wD!~R0H?vccFxQ{Tf$HusR@J5;%PC%d&0hx|2|?O+Eze_ww?py~!*c
z2auhtoqilD{?z0u6eJ!_QI*P&Qb8sx15FR;Pw$(Wv{tNGG5Yi8OI@sB;ZRZQBY3rq
zgM)+Wt@72An}<A`g}y81h6`uEs2KXJ^I5DfJy<$DeKtKNi~FAL3m@L`>DwyyvwZ1B
zyjrrDcq+*AUtseup!qNU7=0J2JP{hAN^r4P=3cXoA@`#e@mRf^4rx7iBqd?0w@#kJ
z{r5|*3ZQKu$i@<8nIwz0ym%B*8=E|Tb?it@?7H`3Rtu)BIo9M_pe%u@gSg&_Z<73*
zR*pKdY00jP#Nd!31||)rpGx@2YIgB6gd9Vf*`<Jj0U2A{0_}Lq<=dz#s%BPU_Qse}
zLbZUbK@MW&w>cgKqHDzLPJjmO5f|5?%{?!KJc%k!$Tc?o<NW-5>Dd7xh1xqDZgI;v
zTvUPJAijir{}61{)EF#eTn~&6oka)b8;do?eyBD6$B!!j9p_}`e?_6@;J?Y(RvzEH
z#xf*2Hzw_P<~a5}K|4K0GwEz3^krk69;kzLetXox_{`}XQ(G(xBL6LRE%^?qOPN?X
zAtVyrTR}mA>|t{A0!6=$;3->xh9Mx<KR>yNUGlmrbbtoz_gxU>b{lNOQs!AKs**0x
zsCElLTnP-L+GhHNVafOhq3X}=kJu$NApyMfzwPNsAZLh-dy&)BQ1Y?+8+(6GkCAN$
z1jw(Dp!VF$d=6SlIc5!?IX!+4)8V!BW8M5wT$$B#`~6Z{_b4$S+=@`p$X-R^F?r>|
zkTv>5#Iof;7q%^_-B~iGcxNm7zwi7#UZ+C9lua%Q@?a0?dbJqh2<-jO4cF}L1GpTo
zU+<Fh!+T1ZdraKFZ6Dup0aGe&^~FOR(n-_|coiYS^HaqO@u&&*spAV<0ry{E(Zr&9
zT+Uk!p7T5==6ka-F2du5!6@6{bA)Htu44c{kmNL6TnZXf^b>&-GtKFmg-#Rif7-u$
z#F0vz0Psn|!xa_1E?l^P61QtLhqNA4_`0|E))Q!d?AWnAJUqmE&_G!rCJYKB*@e(x
z)d8xz2M0q3eQfZ_$Jl#J9NyuRuYf}mcLm^O$IhL=pt}nf=4W*ZoiYd0^pwXT*%SF<
zBj4EwI6lHKo@#1pvOctsF6=VqO(+F)#oM=U_l7#Ozs#%w$3AxaI2P08(Xp|*V$?#H
z=>qC62v`7fq@~Tz&jVQBgqUNh)q@#@81Dl$S=yiH^bz<IXYlwpcPv%av&Qb&SR%+I
zzw9x`y7lY3`}-4b-C9f<IzMEH5ap+$k)J-P!h(|EInuRsmdRmfSEg`G_U2jTEp_yR
z>h@)gzgaeKj)&Zl@E^Dg7%Ev<S+g<A)@+RvzS_uxJ%&+bgNF#QdK`tcCvLU8>!I`o
zWuvxm;Y=xo+%P4!(k-rTUwzx8D+fNm2MP0Fw8W^{4@X5s5rqT<H4%;rp$^2_BoYkv
zd>|M9*%z_|{M!$Fj+EzctzT=nh&u3^y#EOI_OFd1eZ3_rq;V08_x?tHGdn<r7-+y(
zFk8VEJFz70-MeQmK*`SU1W#nyiw4Jb$Si1^^T2*7RC8K1KsP$#A2ZW#5hEkZrcHsp
z_RpR@>+0{n>bzvxvL1&Y6!qAHqE-fj)Aa)8!E*g6Wo`_A2Y36T5bk8C{8~mTtwe|u
zcV+b7naHKejs5}#?V8(IS>pg6`r&ujYaRa9*|~w0HK;c$N%QL8n`!GPFW(EUOQP?t
zp)|Q(rh0s<5@cGvP4X;wQa!4i(1+c6vC8bZd3mNlV_9h^Y{2#EE18)^B=jgpkD}s{
z^1-6L0a~)KXLcz-@w)BAHv5eWxU4pOc8%=M16MwAlCezGN(}&e9n4(={Du8jjxGbC
zz7F%K0ZS1};tISRT-SIpUxd$B$>fm=aDhE5PN30|inde!VL8ABx^O*dgRd4bU(9bO
zisj`GtV9BWs7Jgzooox&>YuPN5@8NQet*hEfZoRY-C^1H3yfJkp(jK56Z7ZOKA^S!
z7sVs~{bMMvf@w=ey#Dl!C9CRVRX7xC;UhqujD~`+7t=>vK_TOZ6^|Knw4TOp*yi9I
zc3F{e6v)yPY|(!GiA=NRS_{T_I@*L1wsE?Welf@xbo;}A0583rYy@hDS12!Z9)GQw
zcCmtQxO1uMcz4%m5qbb|yqnXFpZvAf<M&%fov$?d*g^GmJ`zft{=u7tF7q5Xpq5%R
zlfc!RoZViFN_g;UBOg}faF`KVA~rk?oC<=ebv0dX-rCNt`~5Q1wB<b}x69g|#)@@q
z_Mvn8y#Ii*MJ&_&G-ynJ(gsG>_V(YqHWEnkO)~k)0V#$GGsgH2Z%a$}aB{v%;mcp~
zTun{Qk4C%TT>T&mg+}(Z(Tre$qwwGLgF!WTAI0<Ql#*;Mw@(W!u(ri&!7G@U$fnoV
zwqLKfKZ+l(P#pnTY&F4UZ-GNA@%(XZDs7JJNo?3l^oLD&jmM2rKV#XUgMa&RLEShB
zWt!j`E|VXhB1uDn*8J>{dCL_@D1<~14_xRU_RI|Sd;osyE-tP(5N?b_T44a6_Jzc2
zef4|L!5b%wK;U#Tt~{Qfu69qjTos~I<dO^K<j;N010PITF{!6`bX{bO$7mY!Kmhd>
zx9js`UJF1(SuL)Ue@#*&sxD;PQIz2770bz=bv_>CkN;2p81agpcIkeYQ{NtceqGIt
z_m3rNCa3APW4{2^<ZE5L7;SW833c&4ke=!n{X!=Vl;!$?=JfQW^(f%!#?>w77SK<X
z9{ff5sP7}`91jT-sDuQInX@N0B*66X-N8t;CY!DIfwcSsDsUZ)E&r!a`C!&Ir0DCy
zWdR5k;XO`KLnJ)_{;vgfufD+X8c|(|IuI~{%*mg18y*yN<#-UKI&dk(!)3DBn9x+T
zTb956ia^yM(#R%(bM-K^x`&2RhT8J`ITn81As(y!2h`AS0QV%tLof#t1)5M*X0YMA
zIy<8PSh}<&KysYE2OG1nu_eGPxC)aJ1&BCd`;3iyPBq;F*oGZ8#<!G`{dn<&tZ7pU
zKOn;dfNHq=d@dBk!GPH4XwoA{ex|Q6Igvy}0P<s&Fi|{PmBS?C2^JN%Qv`c-8m-_)
z(@!Sp4HP6idS!Z6p8cTdtQZJB0VZHtNm$AK_qfB#Z9|FJ|7+xH))YF=<{3#!w!L!X
zuNjHqf+&!ozS2zwAI4?PGIfYKk2k46Uxm8se`*B_rJBP8ybD&-02tl%^{XMTJE9+%
z*h?<v`5>-)q5gtT2J-d8MJ8>$kfNpq<QUhjUHiDxe;5d$Azr=b;l$8Emv4iIEZd(N
ziXMYzo|BWawQMP6l|*bR@U1);!UON+kaU1I_Wdf|_~X+Wm;fme5fLut%P1e`<~1ex
z#?{-{MXg><x}h`>ut8Kzw6UI?$4{BUgM!hCT2{yZ7#<E!oM50H`UV8V)YsQDb>mu(
zVeh77!S0F%Kr}(9i)OK!mS))@O~w;o*9o{h)2M6>`UdH0rX59vBsr8Q&_Ge&%MUR)
zk`%&NwaiJ{9|Q>WJ32mIPnwNQSNT?v+5`<=vfv;**rwbV!zu*wAMEaWo3V>hJ)2?H
ztX3DRl3kd6mJh0b?dO+909XBb=DfRiABV}W3DB`iq2vy;7XC8vZi{7kbF(%ev=;GY
z7UrfB4T{}}T~ahVB%z($Sn&G<<ss=aC|Dq~@H-ugY?p!#!UsqzBoROjuS*Luz@T`*
zZ&i3d&!eM%MBdp}n`7NcOrcBKfnQk|BF@8W@!GLt2O-Sh{z`<s62IlZrT1M<v{z6Y
z-U~k93v{2GXqsZ%N4?@|-p%Q0$OG|*mf<$%L=Y^EW$V@?#MZiWRSzFNOblhRW}&FM
z{`z`{v`<1%z<jvs>{yl8QY!-hHu{N_0hf`auXgR)6^{oa?6gg;7p!F$eQ`7>T8N<G
z4V39fgbp;X^?Zco(oXmA@Nikx_xG}GpNLcK+yS?N@M$Q4bs*qE`gh1PeF4xu4TWbb
z`%-kVibD*+(yn<#9wLtG!u;^U85kA{wYZH2ykDL1lvc5ut9e_Va3xe#e`i;Q2ns<t
zl{Ec!2&BF{fXv8KlfBf&bGCQ4O4!rx4{6hvuy~kams7XPAtMjAcjQ>h0pbwhCpMNF
zO@T!nZ_>7QaH!Up1JB5>h^W1z!UvUvWN@$zd*Lo-y6PT+iUl^5fCw12ZHV;m5jpeL
zoYTPGRc(;%tvujgT;d_N?NR_*XdDd;=tL`T<SWYBkYz(WwL%ql%_xWAcEs`;0U2OD
z?V*nnScfb}4ffPt6%`eted`pv6}BCMA)x|mW#Xg6aGd<3ZQ?dvh&wV+q8?C-Lp@a<
zJbaktEQ8goS0%(q@kr|F>5U_HBPPy7dEA854;~f{D)?ro(2$j6h|oa&2Md&ml!^lH
zBOq@3<NQ?ng_q@!FHl07z%sY4XQ5twO;eJER=a_nJ(=hw5NWd%R@C8H_Z!??M6sO9
zp4tq??QlZYa}9d}_$-*pab~G;ACOsXYtnmpZr}<V1<oa*{v@CdT`pZj@w}~-7yEDz
zdggZ{E?|be^u=%CL>-zHqbLVN#oD(vbb5(HS;btW7L0n4U>m4Q4N|E?SFmaxu3?~5
zUtN0QVb+$Xfj~|z5s}!0P0Je^8Z4&a$Bqh%lD>ZIWY&rk6|`Lw;Ezu>tfC^|zul`r
z9)pavez9Ad#b$#G46;A{%o)>Q{HK3|rnMh*5K3Di3k)I4qUqxhZ|aGu2otxcmC@tE
zhl(0JmRAg{N$e7nnQHg>ooEDALHp%M*@b4O5qV^Z{GDZ#l@F7a>^8@)wdS8+97Z`-
zLwQJd{pA9UcbaW06)__2_B#ic*xB0{=cSc;4rzlL_>e?~x(a7^&BMQ_FUvQt^8wha
zG|g0`ktwp5q&#*!I}@*#XVdM)6^}F%F*ji4+9G%aO;AEqobrf5e?orhNpCepbrJKB
zcn{!$7;j}Zaoe~<7OioZiC**;E8FW7wXDL&gVpo(Qx3|*`4I3ZsVKB@_k}5U8#}uO
z7{2wOkA?Fy{cF(WzwqFeFxEf3McO|R(_BO(n@>v;D?LUzGy!AZ2a%$(iIV~AwYT=U
zxr~IESdA7%ab1Gb@Kj8F;a8aq@lpGrUY-HCCh-%{MUs*cclq^#h>?(cO+b0rM6Dur
zep^bhSb=V+673giM0r_RS<P2ar@Krm%Gaw%Mg|43q6laZDUl=v5pwH8)<G2olBgsE
zgn|bWRIzof5Ebzkj2lmE+UKycPn)-7kb+YWrVKhHE`E^mSaeBjts$IVFtMFNMyMg%
zg)V>i1o}YACSg4w{xtZ)^OM=dL?<MmHcD<WB}%OaUmNvCv;@VR2s@=^Wmo5aPtSb4
z!$I(j{2g>lAHGy<L?120&N8n5yg)*jEr_u6HKh*!@S&(RZ#Miz$rhbBb*diBWgNwf
zh@>c@8l;mTi$NljHD4L1t+#l=3tiR%`k;l!qt-wRRKsAhv#&$#<OD74f<5r-h>GOT
zTTh?nd|pHSs$+M)4Pes`ja>`bUt4e{g+KcSquuRnMR~j-dNS8EMP2kUuO0z5B3^@p
z!%#iu)X&e4OJN4Or?CCNS#?HYW~$m=eBx0s*>aT>e1v$pu?ch`SXDWC{Doi3reJ$+
z6}6HjJ=N0&AsFPkws+k4gTfEucYb`@Y7n-oRErP|g`c(WKm5)Xaa6KSz?``OHsENw
zqK}fi0Mv?0nh&rCAZs2+X3yG|m8@Sh3+VBLr+@i!o^&R>4!Yn=@bWkm?6xFU2I|-e
zY`g{tfbw@3FcoG!4A$4K)e#wKztOuNAmQlkWQ%GbBa!|MwQc{FLuQ}gsE~&A>afkh
zwT8D>Za5Qp<o54}ODI()KYrcDn(Y*|l?T;Jkcr=*NRN~vga`cs0xoB}AP)X#n_DTI
zI_3B1(WAkhf$L6A?@M>dcid#4ynM_a{U&}V1T=C}OBOFKzwk0>!zQdz;<bGYlj6mo
z5n4-kE4lD8j?iHt<2BW{DB1lq)GH)hlh3|9lB}D3h<6($yFJUg^Ei<gX(Tge)*k~b
zigrQv5d4J;$FD;j;lz!Cr2FsHIKaCRZ)pPWKIQ&V5e#x10&ABsfC=Mn`r?b=pKLO2
z1;l&>u@Bm?9+mS1VrkG5C+bvW6le)rsVOK+^su=RfTdHX9_+*(2;!qV%7URMVLeuK
zbQl;K8WJOT;2f4zH@a;e1gwU92g_u=Ts4RhXNKDg$yh^<6x5=({3l$*g4MJAdgp4y
zy@(5e-S_FGQQ7E79Yu{QW@*V136f|$=!TNF)==gC&`1-px4|rtu6WL+01;g+T-tTK
z^;)$FJP)z<#)d>Kjr7Y(bFc!MP`i8$r$I^*r#jpXh}w*pj4;H9V@HqvfQSKRQk`DC
zn96ZL7rNS*RD*?-jLQdaV8(EK&bAEKAKl)@wyPgoh3J7cnaRY+16m2*C&VNN#MlHZ
zOor(sO?tX(J!Y89NMB24f1CSqN;a%=a$^#fEB*Qq%Je%(T&cf){bK4y2T{iqjV6VH
zH1X`)r;G=G;ScvW0b)^0k*jl9<|3+m3vP<UvJem1fS9M6C^j(P{{1zR#1%LxP#6tg
zuODWG__gm30t5_F0QgOiA#!ZItf%*s>yIcP31uLQPY86x{%5<%x1(Sc<+vpz^LZfl
zU`qxd-31Y7Fu<VT`hIzDBnXb7yf%Vc*-Z{;lW9a0wA?-vc;PD83Hp-{c+kX5e`+Fh
zdNEUG>I-YOiu4s$PZFEt#h}o}>(y0=RkVG7%>92D6+0WYnmnEc16J*Od}w^+%OFt^
zArt(76Y_}^bs&#8_l0?)hax@__I5E<?(7iaA}F(X=~6`+(Z5(&F7jfJX<;d+=~<$B
zvq`(;47y&p@C<h}#;;7h;&<4(gje9=)X4eE#OHxXsT6zw5m1EgOcC!1Ib_a8;NN+%
zU>CtuK)bzyf~qJ0FLrGOW@yR177HL3)D7<o27W^{4+GVEfQu8q%34-EY%62tCKZ8{
zi+%o`6}$Ez`i7ldX%h-G4gCJO{hELPCEAV|nMie?t30&U5K*vkNYVO`ICEQ+)zto2
zdViA4<irXw<p|s)bN1%eiUgg^1anp*jkW)@H=sPG@xg!-wZM#!Im5r6Yq#|@MDggN
z&*cL{2TPv{3JQvYsj#z7yE!*-6R<m)Bvv(%+mH$i<WdS1*C3%_q*0&fDP7sJeti`1
zU~_wHhd9}wh?QUMDt}Bo$1vw1F$4&-al^Gxg@YF@rdCFv^Fl}@BA>#;+`uUdYdjtc
zA92d?l2}NqvitUB=RJX<pMY5-+9-~2&^pawr`JzX&X<7=+g-+#3i0VvTaiX_zjQ)&
z-f6+&%9S|GmXO|({Y-M?G(vU+awtL%*dZ<naBzg5N_9KYTDgi^qJ)Bjy{CXQGl`Ld
zxAea%r>KX1gOZ%^wU9D(@f-z=+k5cf!HyeHAe&&T4ZJ4@YDPvzT&}LCe6n3!T#!kQ
zXX?fW5gBt4lnR!`ScbjTPObZ}C$y#-cJJCnxMmBsdbAA3wo}7F1UW#qZtg1{d9!M!
zV_`lPxyL^2L82!H?9)#`Iz+6@oRQj^h$4u-9u5ub8CG^bA_PPiKrkZ}BuPvhIPT^}
zaVE=|bQVZ2M(-Xx#%43uMno9AD|qm8Uv_?7$xSrK%-I0Gg9p!$0z|+eG{iWh@n=R#
zm^a_R)IOeScUHj_P=IQ&4|gdb5k`uByJHXgGI@Ydzf<dd3Z17E^X!&%At45)lg@wb
z+BJ1B5aA1N?`{>biM_K7in)+Jx6~R+_G>3%#6+ls{a*HnBallYMX!12BtOy>z9x(`
z=@|`-2r>*0T%HUsqEyF`#sp6IKmsf)@0uAOF+ez7+Zq6X;#r#Jx}2T-ZFJrLueM<r
z{xh-B>lPde98~*U2=Ma##|w(ewzjs#i@W@Rv9CZ+&zjn|&-NOn`pN1dpzvm_`ISse
zwtzhzW<gG?8X?GSShsEmtX$zMpJZjgLj&x{Z0afEpG5ZSP_l;CfE<a$ho+)7HPlA9
zkNh3<HE)v@<?y2qLP++8!v|^mG!_L3!U1w#%6*7>3KVkHIaJcIHpvQj{uC1VEFgl=
zJWPjLbH#B2VVnDJYyDiCU3D|qGK!~8WjHfXqv5zPECmuEnHZ#Nq3JdiIF5Do_QoTs
z#8YX9{f~`pp6cO2rN3E1u^d3vR)c3gjz!Ij;z^u(JMbBP{k(&9Gl;7oik(@A(xCK^
zmGa_-oDP<J6^~+C^Q<&{wS-tagcH&k1NsT}yHM(de2gZPI4Jg!eEp+J8>ug;iUEA_
z&{hm^R_J6*r93t=d@DNHm%HVcQ>DJ3(a4$mg@lBBU3pwq)&m&*fPN3rEZ`AE;~Dya
zd^8X^Aea)7;EOupmT45_PDCl-YF-{5o<H#}gr5I+L%c7>JfM@qB{_$w4IsyPHt=c^
zdX4mGb%u1uMG((J7v3&+ogK_3hikyI#=%o7AszwAU$ebJc@zs?Vr_?}Lj<LtKY!YH
z<Nw6A?W;jnf*dXxRm*EI^=~jO>^Y0ssr%PsCP(;S@{zo11|%^Ys}cjVd1U|*AP24#
zvKB2b3~!=Yb1EpFJXwd*h+OWMt>!zCRDk=Yie@;rUk&Rf6BI@M4n>tS)a~A}75$U|
zE(l*FED{23hQxo-^Au?S+s~0sAc-B?`1dE4eKqnN8dz<FlPZNtYU4!N8RYe=RzfPE
zWVZ?NGk_T0$04s&k@o9yXjx={V6I+u7M#Pnv9XyMd+dk$NsM7c-_!x-&cL!FZVcu@
zg!dvYLpgrPSbUVWlzH3JZhkGa8mJ#Q3W?ZM)vY-4`=IBq9zwr{Bl2Va#7d&}kCuxX
z_|i`bW=pYX>yd;fb)Ib<;VM+j*NsS>c{sC-wy@RFAJ~Z$DsZY;Eg>|=!{q-g;_!d|
z-gl)C7?zy=4>7eBxYI)^9^_8{-#HaRk&9%{vGrWFabEE~YZ>)Q>f4`z`A;j;fAzDx
z5#S(qxblWiQ)5v3<)qp4?}tS9y>6X6U}(yC0LL%MQXrF!L<7m=Abw!1?G8VjP3ZE2
z3r0qS3m{0LKp4&bZ>`tLIff+6Pg$;~RwlGg3RBdzVEa|%9N7Owz0z?9#0M03g6qbj
z;Sm*qh|CDsUC}Iu1F$>Wbv5<MB}ShiSlliFVF?1*6lnwu!r36_25YaT76wGlKiDip
zENXH9fksY<nqWI3JRE>m8f$*`a3mp>uug&3jE?;OUXqIw!M`TpVsdmtH<*Y3P(TOi
z*~WPvf&c#(tL~plsi5$(P~}99osEA`(Dg?(J<2+YY4L*!YTtM`u5H<B_pWtv>#Oav
zH5Y<fBXSo+w*EaA#8K5|S}k~L{;8M`y(lJUa!bsDx8hdIQH9eo?V^mc`(;8ZKa)(a
zUXIn_`1p8)319FsFJJx%0Of!_H`+$s3;~CYkN$z(LVnAZE$g^cfBQ~F$CH!8L6H7l
zx$&`iRk{^GS#oh*)!3nGF&px&#ZvTQs&zrdrqb@~>h&7}=YBQ0MU_9eL7p_9cQ?Fm
zawwA|1>hPjJ04JsP{0N#%fLL9Wo2Ew+slOp%AQ4v4#<}Xl9O6!M2$6(t?A!BTe93K
zuK9MuRr?^zow*?GVnannwU|?zaoB-ZTDm<&Qd;^PG9<w%>)~xD&isSBQ<7i;J8;-;
z@9gYUzqnv^fBmFUjcWar(c%2j$NREfm8^$Eyk<JCSidR4v&^^^6m|$+V&Q`{0T)A_
zz<Gb1M=$|82P&+#4Mln{e}<L;M8?;@i1NzL@4X7cKxlAd($|tZi4Y=5QU@?DwYM&o
zoY2jBtoR>`9XXpzvKJs?r@{S+rc?U%?T_$<-?7@;7s1|_n83m<`@bo8XGi`mcsEYr
zrpP%VqO-9{Ib|R_gtDmF3-SK|Z`YC6rkSSXn3}~-*{^RfQJ)_@bcn?|=(NoEao%9#
z>QXxyod-8W19o5wdhDpk2>wJ0+<(uX|8F&HWe(T_FHXTDr3BOaZ#RhkeEr*Sl9Plz
z5IRARCdW+AMZEBRGPhd+XXw<CgzbX0hP=rfxJ#yP(S_2L#3?@{H-?=wx_{;yZ|c+&
zLsq9wZxI$27A~Avzylq@(XsTzF(h9T!DnE*@E!8~vfaJLl0|uZ%kA}Z@7lIA?0H3C
zGKTyr9+4!KNH7)FHq0$7Tj=w|%mJbe<uiUDKVLcMhOPbtohXK+hae-6eU63}ffz|k
zpB}7ALRTPee3)S^RdyT5cIa@`B1cllUdJ`x<9HM@07SS?{s=;ZwZV<lzF>#UPA};h
z?>G0{qO6{E<SbjdG6}_G3s6s5p^B1{FgZVU^w_aB>o!PYY$V5A_QS`&btoHI$t@5Q
zU-DkXQ2zWqR&5;T{O#=uIoeAsAxZDej>R3wUOj*zN{&@zbq5L<9K!kSpTq785+#!e
z@tprPGxPm5_hIYf#ivF885)O2yqrtwad#w`Pfk^8en)}<Ayi^9u6e-Jocm-hi!ja^
z8l>viZii0eLH|(<KBhcGiWLc-S4GE)B06}U{t#KoJ>%^x$EYR?6^@jL&6kH)BW?^W
z;O4V8p{-SJ_wU~~l94!zYmB=$XA7>HJO1_PpL=QD@^6pCG)3UBIyoc-ZIkW~A$Ba{
zJcGT+EWe<j&)mTI>6aa~3zs_H@co%r@_uW}6S6CL{`1}EaQ<TX$F_m&|FH$dC)4kZ
z8WsOTy+5TWgS&$YRs64y(Dfts%i~O>V7GvU<4pb`a@a)@$AgYT=J!VM2GOJ_u%>q?
z<(Ct~k#`tq%WwLOblh4<1P@sEzaf9O1|pVsy>($v_Us@qp$N>92jEfFzi{N~0W>By
zkf?WmFIr}1=e98CV1yO+4h?`w;T^m$Z}b*6e1h=Y*ai(eG9g+~s7mip9=zvfr?#N%
zi$F>81_3Y8?CGlTeb-*-><wRf%PvK~7jqmqjOpdbvF+H>|2f2n2{T$CD0mzK<DZkq
zSl1BnCCn-T8!wCxwb3V~UE>PrBwf5;X5_*?ky(U#jDVtuY-9;ow;A<V5C_JLvRz)^
zSQ0!$$5Vuin0P%ZXtzX-euvYF4*<7IOf%VtCTxx4V)^2Dl8?3>7A+ZjilujZiT=-(
zFM)*DgWts;iC9ZXE+T)tI?BaKej?wuL4N!XU%&DH=*L(3PJ@|E`2LECf%NL{{ctIs
z{q43WrxYh9zW(6!c*C=O`(e0|^*Dn`lx__ikCF>TSxJD!0fqU<!zPIFO5<Bz@dLUI
z(XhU;HM#L0ILe$Qqd3v)3Ktu$y#Ak8>XTP;alVqu!D$(sQLuxbz*Gr{y<1C*5E6m1
zewa9iW>%v@=5P}On?|D@7-T8>^=*~ZM7=t~IAZ6PQK}g=mJ}|dPscG>QazwbU2NHt
zM9%dca8ho>J^lG4=#Q<pAyYS`8jv|UjgnCxCC^M!6Ie>L=y10g@^t}qU>GQ&VNrJQ
zrZ&wj9pMJ5b1=lr=UIKdK{OGP*n_r0+&%MYa@>wP+@-ehUf0D?q#{X*#%a7pegavk
zABdTf@2t?k5qKQ@E7Ke%^aOly!^Vv<xJ!~;r2RQY2tAP&O$sHfrs+BKx30hN$(xpX
z=pmr+;OpR^xeG=PXU*}sB20-5rhGaX<(k}*lC5W7oIjk>h~)5o;%{PS8GU*6H}R9a
z5m)a-8sa#Rk@>_vy`1y2xlnrELH=2dm4krOd)(^H{g>XeMB(2Z7hW?^4!7MFy(ru8
z7bgOfy}Qs$6j3oBbKyHde&EX&GTNRv!;wJuj04OUJJ$oL8OJv9Jn6b+h$lkeDl3zx
zU1ar?{`F-MCQyZ>Q4lq^n9brtrt05wlLJN5jFjlcxuxf;35vd50X-%jkJX3Mt7*K<
zt5&7ktz1BU=nSl5oQc?7*1r`Brp4z{a*Pp?4_Y)@2=P5=gy$lmb_I=(d^-q&Od9ao
z6>1S&B;WahgHDfUZ!$!9M08LGb}%s5MB=vdaK_kk6i1NXPL2&m1!2-~*&i@~vI=D|
z)b9fYEOi6gALAoWJd&CW!gp_wZzX|iM>6BZ^D~J=9m3rv;E*=4^|m?vh;&|f1!978
zQWx>1CbHYE;-DQGUqGVWe9Pmf(ylr<NM&DHD?vjbnjC9r0E|r_T>#JE{`==0WT2wK
z3!#La0UxZ@wJh{^;8;%ir2Y_5Z53n^AX~cbzr=LDFED&N@vO}2-3csWic&bHd^F|A
H`5XTY%hBNh

literal 18550
zcmeIad039?-!FWXAuNP4Q-;u_q9kcD#hvDPj%c1IX%0)KkXzEAL8UpRL31HglIGGZ
z6`D1xRQq#Uzx{ijz2E)p<9Oe{-s9NYvDRv(`@Zh$I*0G~Go9U1ls~;{#g-KmMXjQp
zIjKxhi*`^HgZ{Fm_#2NBx1acO#PO7-ql%5Oqw{%tBTDYPqphWlqotYQj!Q=N4rVsi
zf_#Vh4)N|Vb#%0K5as8$`u7X?Z0t?=N5aCiaFOM<XS5tBiupYGmmy9v){LS~chOEB
zQ@#9Tu+7<hz1sZ3$dnvE2Ls1?#^tBh`mQm%dWMC@$@oEemj}aohGQo;{7UfNwf4c@
z`uazlA8Mkr(;`&Ya;>{9bnN2tQ#9(z-)olpeH>=?%yFgdymHa{L;v*UR;C=2PamgS
zWdjAnj|zU*Tv;rz7`H<0wM}&<Un!Z@>(-F3F*_L;@a5c=b!)yX^w-ATKYsl9z;++5
zOM}hNr0;GitF2YpaOc$Q<X~;KUDw4IngiRnP?|>D9`osD+m__++_|%_Hdf&^lT%1w
z;3h5J4e#E)3u?b~UPEy%;hkYM-ZJWuC&0XX+3iDj-q8~qzg2_`Za3e1gPFa7{H#DF
z&#ucy)OjfB&PIOo<_t6K)N@z3)e`)T+HSGzGw=EEXu4U#n?6|e^h95E)Qu01`S~?8
zG?pw`67M!OB<lFXui!+vl9D*v22}&)s`=TD1^PwS3mc?gAGe!qG6`?WUl^}d$h2+~
z-LIMYD&}|DzHrs1G^0mhVSKoX;E9h7X-4IBISwXP%^9?c;6vqEgKS6aPclpVijesI
z^Tfn(M~AlO*f+Vhd>0jKYwLzI<!3UzHeRzKhK7ct$sW|6#?hYg_3PI^!(|c-iab}W
zTBW9|`(%D1#jAI!!>cj!)SYqMfKLDYt%v2iO71sKk9}8ruA8$hD>OKmn~90($kC%4
z&kQteXRfWOaqN1(DknF0!-vGuCH0THrNbpWGF-;K<;i$1WoAoiN;lz2tgW+Xd|p;o
zMw6G9pI)<P;EMPS!{pQV*@TCk2R=V_>TfXO=BIrMJ-Uy^t(Ti2cZA7-TGIP#q^qwf
zJ(gSJdD4EJm!*x3jk%UTs-C+39x16$H&GbuC^C4&qw#5Ma&V&OK~z-Kw(Z*oYcA0y
z8u>pc3fcX&;pak|XNox<i+97FcDG^iiZF2<ydmJmjT_pT=Fc$s<pL`UCK_}S6rWyg
zZELH4dV)zW|MG#_ON~idQ6=nNx?y2qqobqBEibKxFDD+8mG#AZ^U}=Q^IZb6ZdU{e
zo<3yJkkS4|z_e<IUZMN+z3>-iwK4Q)JV?3W3(cand2^fB+<6z7sFBk5P(k9wOPl+%
zi43gF5BDjk%gEfGpRV+3=3u9*KG!|s@$)rPuIofXyT??UY`=cC?J>Pv#|K$iSwnfl
z605e0sy05?OOJn8FSz3|zd=Ll`C`^>+jy}Q)RMK;9`2P>c6H5_D;OVltOz`C;{5sZ
z_6`mYQ&LiP?cTk9>sB7F?_HgPmEIW^jeISO6V4a=B<N&4m%}pih!J(_D{X7jGb#^I
z%d&2BoEh)k=JlBFIND7s`1SRU>z6A`-9tmqXNU3>LXSGX%xU%uDeS&9GBVQFl9j@(
z8n=V2>7`2(PTW`>^xUvjj-5<AZ%gG58^Py%6TP*k(@iP^F1qnJ`6PugH03#`XlGgP
zx3RH##H}WyAU5*%rGd}qRrXyl@K~6e+9Khe?e5{x)7u*#8}6z7y)wM=z@={f!lA`O
zU&bmW8z=f34_z!>gBem9E}VUy>)0E)VZXNN*td$A63O|e2A<QW1Wi8G-CVnSe58b3
zr%?|Z(sgPm`R2`Ka=}70E-tR%qt4H_9JWp>aGNrowbjSCjn|yz@12|KXbe1XVf@QA
zW*PcKhi}U|?U$DER~9k#{iu#2?;!<cxM)G5MQ2Tz*SyOkegl1$ee!i@`3s_5C;G}d
zIwZ?;|JuN-xG+DbR^;g+KGo_#k47zM9ZXp(wR|dk;rD4&8?tWGb=#w9EZ6`3+jOF@
zJ}AJcx8@Xna{J!BhI`wOj?d2y_t!i-JA<8<ZpFW6&ndYO;rf)K`GoJ`Ui#~K)azba
zwFtS6n}&+JsblqKH4HYUgkwk7<vLkl%NrcFX|Hqc@QB@h^irI4Tds>J8}3}Kz;&YL
zxn7=XioUqTblJmwnHNe{%Y}(GX-`hJ*z`{{8fZM`)2l1;@(Pvo(#Hs@Vq7wuhudW}
z5*Ey{?3!O%h2%Q*$E;Y-&B@0n@9_PTTDozC<FB9RT`pg4eZgNQ6DuF;_@ip)s?7&a
z;%Tr)o-`!u#Hz%~Yt1I*4PLYFpZ*@+Xj3#7(6hf*>j-Keo7ZnU7dJPH?$QVLmo7Ch
zvAahYR|NOA*c7Q<Te`}!j~Q1o`ru#v`Nf6x8#l)G*2Y@={PJ2m|MH0VySrin-s>(_
zMI4j9<TZ(z6vjYb^zi!ncKHc<XMewiZlQaVwgg7$1YIdneSg>cyIUJZ<Kj4rQQxyK
z4F;dz<YU+KVSSWiZ_RsehDDamx9;CR?C$P<@KJ|#Z_P6wjtjLha>6~|KZOeF>gxKc
zCu?suGc#)#)k-(sO!vJnIWJN&Xx*Gq=68`-FSjnnec<zpAkV|76w<@(1rDujRBKsJ
z`QueI=^LxJ+b1;^dbpjSKk)azxaoa($+OA9mX<oVk#`$c(ca@yru!3B<LMP6eaTtv
z2gSt2?}mHLHyO7nMM-bIckkZ&c$HW*Ki|7s4{xFSKGt{TQW5Lz?QIB8Kg#wgT(ZOO
z;u8_OHv%h*&&bLO>FVj-)zEi&7t@|MTxg-myP>PQ?4i&3P12};oj-n<=Fd&G2v4lu
zCUQ>X$Ppv%B0U)ynV7V}maMIGUj@mzI_&^G1Nx)0kNNPBSR^Jo`}jSo2ItJB7aZJp
zH_r|Jo*qj!Ix^8n_kD4pM7slJyV};y?!Y~EN$dTygIOIb`T6;yXb<-`r1^T(rx~3M
z@t7Sg=iiFQ5cUvXu_5amgL4)7)%34lnjI$Y4{M@MGl!V$-6CWf(chTT-)P|Vh$C4i
zOCHrtT~*Z|E&80t5mC{EW7n3Vx7O;tOZS{})hqJUEB0CS@LF;B!gZ#Vn|9dP4D%L=
zFQ(jb?R#!>l$DqBl(=p#e@LfWeqdU)<>HuUsozez^<97e&}U~K_n|6i1RXT8y!RUI
zL{(cm_@!mjF51l3JNy6#LMTPDE!iPAJ`@xb)MG>K*GhZ)jiK+|osIp~r`grs-Cn<^
z&2@*QxVRoF^NU;jJ9bFbC#dCJVDB6nvYwm9GMhmmQOh(}d&H-w?KU}3=Fqyi#nH~L
z8pr_~{TXjR>q`mF{rlxQ+{f;UUK%iU=+`-SE+WihQp>6>_fTy|twW`_NxBU?HpjDP
z&w4R`G4J$h(8z3=Hiw?_@x90)A|g^g>U(VxQ@}ynubok9pV!{r?z~>%J;Ue~Cbm>A
zXFEG3K#wEdjGT@K*WK|w$#Jx44>#^J-NoRU%=#lDBHbdnHF$f5*TTGPWyr<SO@3YI
zvny7tPzJtm@ZHa!ducEMUD6!5$-FM^0Veo)Tb^@E>!j1b=L4I49`vRgW!GBCdi@@`
z5g8SA$ZOtxZl*s)xJ=&A@T=at!M?r(tb*Q08u|~4ii+4g#+5QH8s$-gl%=G+qu(?#
zkt%~)9Q5^pb(?uWSlDw^vPA6uhS8SUiGxNT&M3d6Z5(A^zRZ@X>8>?h&qaEl--ut=
z)g^x=F%1<NzU89!T)%(+KI5WAnRte6{R9SJ!xeRH6w1jt_y8B&7w)dS`}Dn!_S2XU
zuM(@)<k&D*88$XH?X-)xwj45#-g7pf^!4l4Gg6nO-GLA-8j{M=B)QZQRPR}>MCBR#
zUfGBaTbFwN1eR;5;fpw(EbAb)w8}eZooG&BH6I@FOrUW7DV~mw8N$}BaX)|l%xiz{
zKj(S&?Ag57!Eg1UGd8SZpS-;p=_B5e;o;%vDM8;S(vK>d$Zokj{zECGyrMD1pb=nk
zC?PrjJw>_HKi50z@vHMjbFNb&+SC)XX;RFyvqi)gevbtOpsLG0efqRvs_6IVotoNO
zhbKmPGLN`r<mC&y4!pqs&0rp!M!vq=mU6Yb=kgEqQIvl(th``Rr-Tjs`bh-R96NSQ
zcyNo))2gbfJQ<&W{W>x21#T7}{J9#^0MU%L-0HKw*M*^|=1IwqL_s-Ka&NoJU~}dn
zqYtbefTmM*Oh&q9+2Ze$o(FdeTN@Y{pbPv!qr?pYfv#S?I%r7Kdv0XqGDWFfyLNRC
z4kqIf`>;KN0y@k}B^4ADXhF<%*n$3c?%ZkjTJQ+5u^_mU>%ajetCp7qQ(qR=6^Ht$
zE%t7`JW;RF1NcGOU(v#RUPT(PcHi}l`hI|L`W_SYvi-&@%Yx5;`xr!zM$0j2(MK!G
z@R+^iF*nf|6aY-7g`Vi|Xi>MXSU|}0_b=Irc~{?x{*hEX>kO*aIJR~#hK0Z@s{qNd
zkAn7&j&;}&f*xbmo1|kpst1881(OdueETS?!IIX0HjR!yW|zCrbxO<2&&J2c(}$9o
z6HvzS!sN=1i-CeB@-ewNcU!8fm6EkHYZBBFEv?JWo;g$X=K5dtpbmM3J)4)@;MpO?
zFfuvW(3;~wz^wRa>F&OU<PZQLr@lI_Wc>o|`{I*<*cXomXRw0;fN%H81xLSqdlFDP
z;nbZCqhn*W81_S+zb~JXmQL`PofLH)kHb(z0h9FfbKJpos%Iwp4>|Rzf_88T2?^EO
z`cQsbQ4MObMIwjV@_2Z8d23~VY~<6eDq5Hu=Dm?6Zuj-?<}90(($X_l9ff+~9+N?&
zK;b<+Et42>HIM0@$3O+4K#EZSk3ahA6Gd&mc$0;Mm5Rb4UlDRxqcTi<f}j~R@Y}w=
zsz9fOvqMJ@*|h6kV`2{QpXxdvB5c*&)AMZYZs}4CRH&#^nwCB)lw6>|v8dDc)zP(3
zB014iLPhM3^Qb3Pp*Y6k)fe;T$f|yQb;;<+h|=9Hhsbm%8x-kJwK<2&6+D@tx6J(>
zt91Q(YhT^7vm5K<lum$iXumMLzHZ&Rx@4Whw5;bMtyb*aU0qSY=@uBUG)>+F<!84G
zXZm>h$3(o;V&%8ffC~$vfgI6;v@c}0{k@ze17z%wb*m<S;jhRiPo9jwW@6_&bm&mI
zi;?28&qf3LU&#*{Z)$J3vHPR-rWW5$a+i<dJx`weyR)n7LUvBhW^Aeh7_|#;-@k7^
z(9SIRyDCmMCk?$h(Co(>f-dS4H4Z-s4PE8s<#pl9t1Iu(vxc%O@>dqSb(gLsz11f#
zFK_3qoh}Tm>2tGFTa}cQq|rb+2L>z`FJ8O}-@XTEJy>5M!na+*{qwJuj(OwDE-ssa
zqz-_ISpt8u3p?4c`!otzuew1XQj~dpf)BQ#5*SDnP{(_>;Q|xgBF{GMVKlAx=s<Td
zzgzkE_<X=7s?ilMObxX)h=TXwGE*tJ&coK*Cl4Jpa38r2#(bf;wzl@?__$DlMvC>y
zEkdOt_>6l+B?MJqeI7#DH1X$F`*8E-O@Vvcj?jCCcg7ZWr;uh(ULB`jklWeYYxe8c
zuT87A2nhu6YSH7Zt!E-yX4L?LP?oo1og5Ss6Z1u%*i<RGkZ%WAC|mp!gsgVc!HYKu
z{sZ(0nWtr~Dz@0<^TMQ(ZTGHSZ#corJ3Bki3)!@{eg>SC3w<)v`^1yK)N4cWms?vY
zN-<*z6BEy_UB}ry#vc@%V6bHpTf|6Fw=;ZQ1jt<vD^lHzmEIKPy{o@JrKo}MW0Z{_
zxAujlI*U;DS9DS-O6oy<zKg>gV9yp2+lcqRn=cI8;%&>95wW87=gNP+`r#@RC2*H(
z`5Ckh4DUWH^6Ba6_zNZXu!BwS^}l`l7ER}Kers#%$7LJwVkw3LMyyEy(CR?P%IKwz
zp!Rn+3rdsv(ph{1lqqP?6-Yq%*6nwOWlkYLiKMz=gFT9!pr|Inl_xZw>nbTK`fb~G
zyYZCc%qptXOKW0l495->!b}%-7M9(Tl9Ge%1v;QgWtYT3TD`r!^UTMxIy_ZuZGY+|
z|3#HH0f(zSJ-)>4#)Fg}hpC7!zcV{h%f)uwx^;^l4I$y1CjXHmTA+4>jk39Pua$cu
zq6CKgG0ocCJV9ASC3#qI>6EC)Oh$e|!5Ei4#o2^fN{B2*!0iTo_@D&K+)}DLJZWZC
zEsG!6?5eH8vZM^Mx&C-+*)RU{)uNeSUH2z`7BlmpTMQRWg^_j+@@X`ktb@^ff1f=(
zzO|(#<lbUxRLd^^6jzsEwGAaTKWkBRenxD^BBs@K!G|m|AUKHW6&69jVRQd^?9{1K
zy#U}JGJa3;?AY9-ZfLk``*N!LVRSSH`af&=1q-`*OX?iI)^;7a0>*=vf1X3@(P#$9
z4IAEm>Cz?j%a^k~Q5Jb0=jG;psINaH`GyEqic+#Psg=c5emhTO7`|I!)t;|~_U+Ox
z4de+N9wBVi+-+eAZjo!(C0%IEGF{%(qy@Y@xq1;r1^SL<ycxFsREujYU9~0p`d=Hd
zSJj9b;qIQ#QC3oNdT3~f9<5*C8t*<m3I-X#Q84{2#1$QlU{KQfATu-o%GVcpEx6w*
z6Il7Kq@*9R&vcqNYZ7QC&yF1tn(fZ3Daz6=Og*D}Laf4n@x`Im9C|d`PLqFSxTJx{
z%=iNiY`PD|veVebX0N;%Zr{I8P_!zlB-*vyA@jP@$AJgu<K}+<?(Xb-eamX6(Iy{d
zWo5^q)+01!)Thwh*hx$7j%7r4T36P}<}K-$7q)4)9P<RQ4;Hr4zVh~AK!EXFR3TXX
zis)6{;l7C%!M+}bhqGs0V4xIz$=$V9xM+_AzZZUdaA?o?&!4Fd9Hq&=9`8IND2mnQ
zN$h%Y7jXri2g}ba<@>&V`v4C_!VNR?k-Jl76He~ifytmg8qLW5;Piw<C}jCr01NA5
z`%_BzxOB2`K}EK6b8!3K+rsG!^3z+F$OOE(945_ZAM&Ps?KmV1BDkOi)F){P0-@;u
zuJQ+dv1Vami2@un+Op0D*kKe!Zx6&*MSHujfPzRzyq}S<6;w{=g2I#(8wqJls`tuw
zBVu<=q;AJac+C6&+pYKAEC}v_e^@te+(#3={4)XqNixJbhzbVcZj*JG_oRMdiw0Tr
zO(Odct?1OrlV|8L@}UW-sRAypuG}Ibf0`2g;ln*`TlVJIbQH?S%4WPgY~3nC%i<d}
zx;RxSF`EQjZ_d&WF8N1pn(2+DM`MSbI(?c)NJtfH2{<JHU0mI}Xg`K)%r-1QaGl>S
zbCfaj&#6zBuG|y>EvIjwDSgH2)eQjL^^vFkx_;vZkFc;hYKYqL<5!``7^cqyTI|}m
zaU)T5f)Cq-ag@D&eGKAB;p*+8(FzjNC$}E9t}A`8dk+o0y9VRiGdO5vW%Kst+Mus}
z?FD)BW8n*$3n$LhfD$JA`?HGAjn~qnL0~;_%jAOOI*$!JwZvV=t5BVE6pQ$9zqRCk
ztzUCTg2{J({kn^pnR&_L#hT?N6~WJb{<K9Us@9%F|Kq`6NiO_OA1Ye#q$y{b*N;PB
zYqlIqVcthnnKyqgckR5sQOM-eww&zjQtYw~cizPmv|z|fAtwWeC<5B-q5ZXyPZb2I
z2MDIFPL<R-uviUn(>*jmpax*k3E+b`sDV?z^B_{E?)y}0FU_d@@i*!h*8M}k^+#q+
zBV8rQUY`Nf$}+&BCWsF7VE1W%fB)!^!!~`*nHDlMTG_-!DIl1>nrC+;7JfTS3^b`|
zr0Am-Bxaaa@1oW47yXVT<22p<P`ID0IW#(#xonKdsioh|79{Vvx2#)ny}SFO_(Yvb
zu#lN@jM0@VR~o`?*jMd@oC9cI`|8S~31~rymQ87ddjK@X;&Hy43D*E4CgLIfCF$&Q
zc6J8V)b4s14Xl!gx+B~VH2N)f71nm7&sc`e8L8U>n0(OJCVv8H3Atunw{F+I*w*28
z&`2K8XO$S)2iJ;|PhUfSw=bDYtBF(EjqYm!=AxZ$e3wY9CMG5{Wvq{Gut~^Bc6sZm
z&&49_M>B#<Jwv=hvB$*Oj`P34X3lW<E)QX+7uZ@xL7_3?*fqzIuPErJ%&;p(J%48x
z&d*p(47MCDk^F!mBRnMdh(pZ$RM7&#kPSD&=9~5H?bC!UKSy9vIrr?@(=sSNJCIJ6
z8d0$@0y^ad`T0VIZ<iJP9xb2geyHI1`<LBdQ~GHlH)6Tl4;Qgw0FHZ)Wehy<W<9@K
z+Sd<o9qf50=n<-?3`o}u;P9h>fEegWT)e!pSSl)DX33i!)KS-;LFqBCiKd~=-57^h
z#v>})93vkkeRurLU;J^{UZWEe72~#b22R86dT=1rr$>7xK3}-s#Ov0doHY*Khj(Zb
z^&g6Frt5_HbswMh7JhW02){34``tDGKDoDO@1*<tbFj0s@7cTeLrY7GaTADzBZM=&
z`a?xUh5XuG{GUy-w;ERl^7ME#v_|fYy<nkal2U0r^YN2)WpHyChxO~@VO8HZiXRzF
zPr8?^`)~gRtL;DAG5>GA^)vin_`(#T#h<=MWpVX>H<O=S9Z|LU`vrTqTP_Lhu?p-S
zqh<T(9Q*d|dxQ!F=$slE8Ci$FKO)llfeSk_5rDAUE!ztKT_jmIM;UvQr|J@Jg9B@Y
z>H4(TU7j7M6fa(k{)E?przC5oOM|C0!kr<ko7g0D)-`J&XE$1A*}>09f3t+5I1h*@
z<dG^^T3@fmz`($5P?%3}4S`;X>d7^L60AT@Ws^1f|F{f<WXH&bU@+yN<mGVmM%|aF
zxyCHPdHC>QwG@3lu*er?{U3t_Z{NO6^cJFZLs~UQySshsRyBC%MAv;uj0f<aW()cW
zXFRzB)V00Z8KzI+T9{$x`#Ore0-WG*J>yb|;S>{Vohg}nztzUHAxUcmGjk*O{GoHN
zD0GGdsJM^^)FEi96}r0+>b7FpvZ^p~H-E?^(1iQGyuPNLcPY6c#eknSI&xq-jT2~d
z?JJky{p<V;iMZViUHE1Bja6G?Ou}5=&v-)Z3_fUdJChepgIHlix+MZa=bP)IUi0}N
zJ;YLyp(m&(o9h*+#w**~+e7)0V({37E%NosScW9F$ZntWc*2W$`B14F!c+7M65uf<
z!iU--aLzmPH@HW`N|%m84?IrRyb?O+mm%A*$$hw|Zde1#rlzJZ%f-dTlO?et&eDnF
zb;z{pRZJLm9Y<MPTLBe`rO)XiqA<(C{B$D=X*pm<b=YOP^;1(*8UVti0>oqM38(lj
zY$GcQ(rhf`1>>Am{}iyE789%=&U+#X>F?J2?E=?RgYO}8_N)q8UBh#|MD(P8Nn0N#
zE&+L_K%u2cNJtRE3%qN1;Q~~@J*(IybivGmMeHI$XyfoE_$B6$iizZgDzJXtx=5G_
zrg&OnL*?h^+iz0_CX2}K5*z%K&8WOMSual=6??_<<w{t<K#kE*wG6QS()0t$@6ON7
z_LEfxeRlnZ4KY}K+<W)F57x@d&i)Z%Q}_VN6Z%aQphCger=tzR;Ox;L_k@ZU78cH6
z6!dn4Z=@jQVS$RfxVV&6Rw_Vhs;iDVZPllncS#PMqXw>2EKmssP+sQDnN*dxlnnQ$
z+M$nzn|#g>?^tnMS(%Q;>mL}%33DgJ2(`Bsst@ayE!<f7M^D|%KNsF<{kN~Sm1~II
zaK@nmIl&%B%l&5_Ja|Av50Ic(G(7W<fd>c#TFowDSzzM|3ank=nhAe{aP`;Mn96}9
zIv^N5Z%%DA5u6<s`ZTDe((=&%dsDPhTJ^uHQtQ<v6z4c*$YM-l?_Oz7PtWE8w=8tN
zNQftM{RRs)M2G=;p^-tGfa_xpjf(*1!L`EFj=%oA-+mp7jYm}2f#2O|08S_fnU+m8
z+OgOX37Rh=U)pqp^$3H}Jc6)4#z?V>!I74FK*z2k65)#yw%QZS;v6(!<;S4cGLQMJ
zR2qTkuz<9z_n}l{nKWqF+S(qwv6@-~f*5ggvyiFZYBsSrC_8Pnm!$8rCF5@U`vH)q
zS$9%9_7OhHd5{>YDTOmKIxj7iuCG|1)Pt&&>p7RpXHaPAu%2V*&SUrsE$j07`pl}8
zwB^(u+O0xHWmzVQC*ZS&RGIN6A94j)Nv7~cf{a<7tHj=e<dAv3c&SAbW}IkGhb%u=
z4cIbtFZ%BQTA#u0(xm{7XVKBDy=DVv(L`8jUPDKzJ>_V(8Md9rK|^Kq_4OORixTuT
zG{lD<1Gi40_y)5EsdCT&cb@Qx!_fG8FJV3KFSstahv60g4;YCP^o^XnyhJDvm?k>g
z<u6yZHR|L##_M~{Iib79qwtX*U?DqHj)=e(fO$hO2}fCR@rez0dcv~q{5JVF>wx|)
zs=)G#;*|LIDU^q1-jEa|H;t*TN5i1iz)k4>@+Cq(RFn%_dhOb^!Gb3D${sLgYv8`N
zJG+G$%P@|DJe$K}5@tCAi1@*nF}{Ux9qi+V_<pEmad7L*(RLb?6oyW{7k|CktCh1A
zvm83;(MsJeYSnFbc5tW#RzBo5Y00vSQPEFCfk%7Pm0a6Cc6oH#!3*zK0!j7NYZyG@
zAQm)We<)bjGuWqX4zduXvEzxt2fRMy;>IcR>C-3OX;g*9onpy9vO~s%^28R|&n{Nh
zYAW=|^I>MUR@tz2?NeY+V=YNM7G%kWshm*757pHV9_kQqa&s)>YuUI`YFe-qHxJg~
z=+sme*M;L;sBsadg5nHiDa)g+Oa4Zbgp%qs>D#Z_B)8+yBX4hN4+qC*{{v5faG;U|
z*)QpqXE|f^409>WB15Y|_X<=bITX$}vN?aAjUyiLlypjzlsz<Xb2YRdq8?{jwW#0O
zuwU8IGKo0f!D|oVZQhLMKCJi4arn*+LXRMW3JBvw=H(GXi%`^>UJPNw6d?-OvJ;hF
z^D3B*8Z?*;`<@T(L5m2@_Vx(0U&3*i<qQb+5fv@8(t2Eu@g`AQyu|}0R%{^uL%*~5
z&x_5@Na0&5(C7c)Z?Ct5=Z)8?C~$Fdp21S%(RiMrMNk$(Fd-V~F&?O(-8V5pia+!U
zTp=ga*I**jlUX)zjt2?UHdsLSl=EFf`RM?uVRVn0wdH1#C46B!2?Q)&9LTzs`eNM^
z=naSRc!$SSsC<}MJnF{u%nY}fm@eqIe@IA()&WM!_B1Mg0J;<`h$!OcgM#Hk2}ny0
z4-QTMx$2*V4kP$xvG+Ru^2OaG2msQ|ic|w+Qs~-f_j()o_3J>TdcJ*&LACB@U-)%{
zn8KGnLazG?Fwvavl1_Xq1Rm5;r~XJ|s7YvOJvv9xgIvD_nV?q@5MqHIWR8bHh37kc
zx`elplqle`vHeRasiBK9Qg4cjd%k{+Bn18rf1#qSZED3R?vjY*WO29UIvoOebQ)@H
z%J62OP8CCdS&w1EVkYGV1UMQkUdXJ5<5^4$36e}Vs3UlhBzfcd_1?}K?COxukfAUZ
zA9>57^;{5oAOJ+`j07Arm2f7gc0p{%y}pM!6W&7-8d_{O_4M?JHUT)RaQfahV*kl4
zXT@G!iPQ*$u6oGGsL9s*y+EK*=zSQi^00f5s(C_0E6e_Dh*n?&-_SKL2d(piIIElD
zjya1*iF8hKj_qWz$wFC@E#}fjNUEgdyQnZW9T0)D>tJRh$hM&h8$gVK*IcGli-f~B
zU(ba(XMmgyrbj97{BpZZ?GQ44fMYgH0kkNi*B&zdxCu;Eyzy?g@vpZ@+r(Y#jL)2s
zlq?cUW~NfZAxnTj8gWbFS{SU+jAXD^IMoV3^$#5_{QRLM_X5lT+BPJn6ea+zkYK`A
zVM|4L?y6X>Vjl5JDS^}XmX~6^qyUO*(Z^wMC}4IT!a54z(@Ox){+9ay49|jaehkeR
z>l3U(Mn=XYekP`v{S4lbk!p$R!4e+169w-KmB10--`k!@L_0H`7cX7_nrcAz<1qX6
zvy-6itv2~PPoF+b+(_JNBb>(+&$$^ghBHtYKxBi%*_Ts(%=X)Srr>9(gLm(tK~f{>
z9nwRIUP7>tT#51;_$$d+Q{v-QCx{Qo<~H!cvj5<dT@))H$g>hW1#ru!poP82@YKW2
zg`|@srm*famH`~v!BJ#JX7r1fP=OpJi{A>Yd=?vfmJW=c0(efNU%htC32>e9FKz(=
zMWkm`Fe;C**X~QqoIwSyg)1^qy8HgcoYlo#j8dJ-P<-d0T7ZB&m>lj%!Ca8u4l<wR
z!nvJNOzRHxhOeT*vR78IJYfF>X`Mtwv~F|jA09S1cZ|rkBuJDrycJ?C*<<MANXs1y
z=ocW*Btg^-MC1yJq2=v5vUdj*HVxPenf5(0a~%tF;ouV7G-97Z?7b-Cy}7S3B?Val
z<Gfac3f(*yBGybJG2@`yo~Xh9a2Er`x)&EAkiLRbbMsz>uqT5b)0P+kV6Ti_Mx81J
z=d)_>$hJFyJ!Nla_u$42CLl0>MZBrW&cUHvWe$%RQWkO0M}b%8WfHe~QXG&Gi6~@x
z>Ii$xID#%GpjSgs3znGqvD>OOTM3qaMcdzhQK{E3rAFLGFAW4PUA-+%!@#3{^;!mM
z#kD*csZnl7W!R-zg_W1pqd?6F<wH7-V6OgAl7fI~UX7{ahIl3Z`)9Fsk!L}LFmxOR
zEa~x2M~CB(m=m?{zFAPrKz+W#wY)UcN5S{jtrsqm_T8mo6@oyBqQshlve<NVCFLC&
zNcV;k8U~4|=3Bw}O`>j-NyJg?%9jThyGsiE4iZEdYX{?OY}7!apvs?1Wq^g1dNpJ(
zD^*id1AZHYYW)K?Wdm)wpk-_J3#n_1C|13!sm&C1#`%nt&+6?72vCyl&MZFehpe5J
z9f@Uf`rryE=)`T0Ly*b%-6DX5gVIMn-bFRsaE^oWe#4FB34A9ydxWB#(=a5u^<B3`
zDSuteMN0nYZR3xDPvMMJBZ@j?3K%iD8hUUk+KOdct{NOf!oXG2jpdL=Y~eKk#l5nW
z2_GGCx$FP~7lMRgo-=0YhFB4T*s;X@$98;%k#6DbI?GFO2H_6BRfgNBZ=<{ojjJP1
zfrvbYX+2(XxFZgme+E*_BR{`ak98=;t4BT`1vDBFvio`xPb0Sg2!q-exmnQo2jom5
z&tavQ4kLPdiQc;c8%Qh?0~aS7bP(mCa-?$%Rgl%x7D8D+`iXhV%j9UCC~oNYYG}cj
zp)|tq(ZyPh90zVer$I=R#Or9JVpiN^q7-E@H`4HN9&nHZ8WS^{s-|Wjkbe>y>(SFp
zRO(~d<)v_3(|>P9;We#^zC{EqFh6u7Eda54h+ia9N_xxWYG&%(zMyq$iUk747E!j9
z1bm{R(CG8S3-bzqn<Vx{$YP_uTg9hp25MlB3ZpW{x?tq}R?;x>@btmkxVKuAVzqeX
zJ%DKmQr8w!tl~%$X3h_M`66Y&K>7KfVm#>ZO?Iw}eW3~+GmfM)(3aG|!C`O#$nSw!
zsfmOI;+HpsV6Lp%CQ`*_w}@Ji1DKM4CI9Espc2DPqpqKfS=<fRO%W*arFB~ZiRTbZ
zfHpe;Lk^CrJ*mu)6zU4xoF8$iyk|SGh`Q~Kg<?MbquLeGtNNE#Iuk#tParz`AU!?3
z;KU;8v(ihJwMcPj0&WWnZlI(N5P(jm8H=Tn&_$FMSiyZHLJQOLIcRqC(UX+-c7e4!
zsfbs|<&gD)4Xp%mK;qYzMRV||%&Q}>|G7HtrG^wmmGT32lPW4PJ0M8<1j=ZZ^BIad
z+PMM@P>N&E9<%x1)2@)Z)^FRE1Q7ej)JG{HDak(z>&<%5%qu^kV8DT?%e2sde4hvn
zWped$N^u!+7FWE5HXgq(?DFx|t931zsNcQ!3f^E<O*C7iaB*|X5Mzads9)%m+&{md
zxHKY}N*iUMN;yixI!OW#i|W!)Ydj==9@u&GaVS#ekbPnn6GR>HtMmZF=0oQGdu0P7
zFpXu{zS`{|(`-8kzV<W0&ShG?t*TZ*GBmywQUcf#XKDm~M6oL1X8<D&sCj~qhTHNq
z@Dcp~>1XIewUYBA%)l6k&vHU42HP)?_TBUpyCi7+ZYpBat-E*Yu_^|(FQs0|WB2Zr
ze-ejHOH_TNixZThd<f5nG}fEhs4vJPjhU=Lx&mrPYdbRR0~{qTQb_fU18Makz|#m-
z;n9;PNpJ`-eHBT+DC*rupH*8)qS@^R1C{zvSMH?d3&T@tNt!PrU`F{<ka9^CI(Hfm
z<8$5=^p?o^=$WyAEg|HI_^ZBPSkFQdfz6pL`TZ?8s=R#84BJY0S>2br*@Z0w$r2$F
z%>4X3(g()xfW58~01R~f0fUB7`t+hB6}bfU<X|KeiUoI~0-@3JYNkrzBQ#L^KHf%2
z_?t8zY_~@oSj}7|x!`@;Irh&;??D18a4j1Zu?7={g-38<#qA}O;@d-3%|hkj=_+}b
ztjH=|vUwd;Vie5HfEq-`LM~Agevt9I#T4gWv;bmW!h}(vN2CA7p^h||NdEr)8xY*0
zat#Bux7?1+k54c62P{P}RuwECqLGHcxNag@PvW#p47KTy49D&DJau`_HW+I;l&_Dq
zFOk5DB;hQKD#n4_dWC(lz&O-QPJk2Kg?UrjGm!0o`{7F|KLh)G#u~hz_|&j}zU*|Q
zq{7A#e1*xZih_U<+)aZEMJM5u?_qB82yT~8ynGIQ5eh=@u-8Ifn9H{y*>a#;y@G}L
zSt8~W;E&#}2PHt%{Z}e#(&(Td12v=v^oXTJa$bp0g#Wa!y$}ga9=e~&k}h*a=n;xp
zftXbywgD)EQ$RV)Ug$MstHooO=+Veks$z!8V}U1Mnqp_9{D$rGmteVm77ar?aj^sH
z<3VOtWg6R^AXj&b?{9Ch5RVmDD%l7+5p;+vL~pVgq^MB>Y0(9Q?7Pok;{~y<rXo1c
z%E+V@f^1J9vtowzC<@s8>B*5OEDN*GspqjFlnXOAv$I3tC7d7nXBJ<%3=~SaeCJ@1
zwUnPEZYdFU7G{$Y=m%WfoiJcRdO)i_s_ll<RE;1=rrVS?$sH353`O5T*96d!M7D_^
z_vOpEg2`qJa-QOOGs<bn;6r9)%&;mnF(`f<#Cn2qnts-IGsz4hvXu9pqV_1^X)s`8
z&mJ5V7rUGX88i+u3o+n`31R_t6jFB)l78#gu6@9NfPotA2;Ixhy9^o^J^Jsz7lUMI
zm$R%}r<!+Z(6Vu>xN8PcLOjCZpMClAh4BsxWm~@e^y?4IN5sVfAA!Rl1f@#sCl2vn
zU#=0W3w0b$Ep=VfdN2D&8`$$iA4dU5T7DY)my|*1#>RPJiHV5_*vj2rNBQkkVT1y9
z(c|ZD`KjU!b&HUM2R{l_qysp>T28fWq{+#tVR$FRr$!mQ=ZF)(Fw<C+#IhCwy^7bu
zY)4O5*TdLB*ctsO)-MVrk+~x3Gjxpv!*{p0pgxf53UONNM(3&(D{{LA3*pEb$w>W2
ziQum-$lr{ieP_s(I_uok)s^lZD8R)<LoD-jKzf=3KR?I<K+o-)H~o+r0%Yc=kr-gR
z=d6{-SjAzN;40yN!_jOf;uH`#3MAt+(Wgcd?78mKX(TKUjr$<eGK%#$xFOnO0!%{s
zfCK+<+gE*kic?qLbJ{4=rbD0I^VicSPpq8|BXFmRAUsq=0js98J&i3yf16;KOhPc{
zPeK8oK77QQl6w8Y-~0E(d|OqOBJ4kt+!YX$<-;ZP&K-k*aB<5~38<OKv}ozi`DH)e
zNUaF@{8=-k`ve^bm#qK#WSz!)4}eR!P}7f}IN^leDNEL@D(rzPOX`u{wECR-ItEbD
z2@dx8E1AowQUhENphh)j6ic-nfBr$|+gSX3O~kDq<I=IS<_{k}#1lyn##s~2%xvr^
zOA=|=>q!#^mphltp^ns&fr4sD&VZZ^;H@s_FM851%s@@nqbU=51Sg>ZaT%wXaWkOm
zGl*+xAl6(^jLks{DE?_IcJ}-WO=8@Y7Z(@j@*)*{)OqNmo#jMJTW{xUw1H;#vl~#E
z06hA89vd`3^H2p5^MFdd^7UU}F-U6dLrMpou@?#lodY`-z)5j9f9zgp6h7r_!17WU
zGke-{8STHkVt|{i`56C+ga5eYfZ@sKTXSlQuN^pWK<oL#wM<mR4?6^viLinYBPiNg
zIuebBW6w5F5#4168Ok=zB_LK1eBgrgm5M0_igO+zEY=MhjE4mXB)cd06k(B$BoWf)
z{i!1oe?t_asOei^7*L1)ujWQZ<A0DG-@QQ)6tNFWeN1qcvnJ);BGQD0K#iUS1<yp*
zz29x$zT6`er=`e2;MY|!y09xggVjl`mMvIfF*(o#L+AlSG!mi5=G)nli`fj~dzP1z
z6SQeR6%ls?$nomctExHnM#MJ8A&oSkP4AY#0B>IXd`*b+jUWq|h{z#KaF(<`=ONvf
z{g>U``ibmHAa%uxjQ|Wnrd3k7HZckzs3)RK|3H$W5n^*%5)tX3&!O8#d<hEhmJJnE
zTeNhgBFMH{l=Ld1Ng;PlLR;y2CBSq$cD$NF7dp(N3!AYPc|t->2vsMnt7BnagM@A9
zd(NgMv2WkLA2vci>6P6#86qwo+{2Iqr{mIadomE<d{gSRYvh;?cBx4oaqyw|kv`<~
zqY9?V_75-lIxaF$^Ta90MW>&@4!xM;4hlfBF@>lzR`!2WiU)(}-nuw!g%noM^K-vu
zXL-<Z%bnmrFnAy?3Doe2pmhi+#GF89l?|Af-uoBzQv-05#0$`z(shUW`nZYjMFwGg
zvx^ieII@hcBz}#Pf$;(S27CwtkRx7N4c|jRJevrpC3!yVI>NIEPX)c;=H~v;{uFop
z#{c}~@<%-TASnHj-}iBFbh~Ds#2!cD#>9=Q4Sw|V=g+6M@j$#M{T_Ds`0Q{|IY>Ne
zG&xMvjVR=B6-9ZSmtp)0o{FkxN*e0t7f~EsfYwmUs^I=QWZuol5ZW$cD@6e0t22-+
zVU67x*v?4p*eDxv(-!hpUt3-lq$!dMu>i6n$JA!vuhR%y#<|86c$%~tc#ZL=**#Q&
z?dpMo54nt93>9-xM&UcmbCIn7>5mzUqrotVP27IDYZXv_A_j=8s)^U*&_~N&cUGQO
z24;w6ShSQ=LP8(8RT2!&uxb%TjJG(+kH`w$g0<_`(6)$OwtbnU3(3Ff+Z-GnAz)gT
zNaW0@a{<SX{v-K+Eve(#fxju;*X1s2&0pO*sc31p<bSM4B(_NnDV4_6$7F5(ZtVI{
zs_+ZQ82;4$E2p{YU-8ZVUw?b|VFCDZl%G<foQ>d6=*yU_<roR!=icN=1penoK#o<)
zbIYG}#kZGC)GP6?|6d+ifO(msrhpg5lat{eQ-!uON<F0A_(@CpH(||Xh2BCF8<NQB
zPY#ZgID<-XtS3%4Z0qOO%T-0$dyL}s=FJ<j?Hi-X!c)_q-9mZi$exuVISS-jEkHY)
z(@pM^gbcFSIDyDlJ_Agzf>ugGN;v(MfYdBm)38oW)k~rG?@jSoMZI#5X}d~-@2fTm
zNI@-r#6c<%j$t-HBS^wdHcF8E-#BnV>o0{h_0+T-wUV=8!MNs69s>R%r2_G)Jm=wt
zAJtLGMmXv~*vTI+SQH?z@n>hT)dv(ZVove)$Hv5PA39V&BXMA)Va|$1qfKq{XpfGK
z32Y9#IB&wcd?1Z|zfak!4GTqECodcma!wR47MOPUc3gipnBHs?IDeSiE}v`1ySd93
zm<LL(n({dmVJF*nzF7io9O=N0m{u6d5d6#RLlKzf?Jp@QG0~EQw~Y2(smiu(oBLy>
zjoT$^2e&hfUMstBjy4?FZg@*kd9n+IySqyzd){ZXxzWGQvzX8AZkOB@POg%Wug{|g
zN{S}8zB6V`g67$xdw_;e6*LwnFs3qOk`JU^h2>4^Y^a2XHo(;6Y9{AfbDCoQf6tk2
z@_`H;84<C=?S_4&jA4<|x_RldD<+Lb6xU5&P@nTb;yK5@r~Ai`IAHgXQ3Bur^m618
z4+-o0CpudnlZ1lp%<j0hYj9A{a>4d2-a4Y#I(6ZE-spp!DIRjPp~JUlJ4|VB3VGRX
z&wAz;bO;!+@*=WPAFpy6MJ+V|#hOqDg82gafyk?33dzwjzLpu-fP>aESYy*jp<4W}
zE1Od^os{nXo~b2AZ=nO#gLL8ar4GSzK-DwfB^S*8EVpm+WVy{_&60Cdh^Y79J>0H{
zh%V9a0iCtkH;6g)`ET1cpTDCPLAU+7*@`Izlfj5plJLD5SUCxOeO!@q#@5NH9t~Nr
zP1-y1I)*freQ;PF4_wk=I`SG%E7RwhvFW-U?<umW-_gwkpCTB8Bl`bExni{kic~*y
z<|Y_zUBt0Jid0jBI2bum_5A$S!0nR@muSnEI=`ELR*p_g5mfYYWS4M&cm^QXA|0d9
zgS&j><ns-?X#KhIJsO_FhyPau-B{&2gnxBJ)<7SX4EMf1@j9}DGx*2h+k0p3=AB?v
zzPbOu#2f(=D$2?PrjUe6)2B`MuRd;^!%<gZ#Q5S73Tno=VUnFeS0jO?22sd+YRE4f
zf{T_eA$?0_&Fo;Cb4NWIU<yc{88j~BaaCcHAmU`4>@YXuL|_KO-@HF7Z+^Y+8D;v<
zy$xNtay$_(&3oh+l7?}T-4}~@x2WjMy?*rs)v`c@lNqFM40JmcA<6U36q2o+7Sb@j
z@NXnk+JJ*L(S1>azFB-MU6{jBHfwSW1F7qmd_^Enoqc`gzj3htQ$_(!&u;~=nEpvq
zq*vRp%8hS5#WMH4t!u@ZoldwrqG#YhR7L>K=f#1xH^bB4zWW3YloY{NY{U9R{xmgQ
zP8AROk2;VQ56TyXV0fyH(vE)yIr5fgSG%7#t1=WfMQv^Em?pQmxw-Z1>^8A+w)>pZ
zJHB>xHC0|xUo*SE%V2W<$KypxLjT;$)PI=aE8<bR%5WfO`SRt$J@}6pr{>MJ3eUfm
zEjuUKqIGN_%h#(ZOW?0L>FSez-ru@)o!{{f@7|q)i&0I8vk}f7G^d`w1`K3cQtJ0G
zugL~7ct$)wt-m4pf@DbCj!7L;(>VBSmVHF=#AFTVur8aqi~A?@l6`v8a|K(I4s=p7
zfH(`W2{|+2Ft?#2<jAfxwiZtCRzuXgP#34vFayQwJ$TMt%)=IBCf?)tH<^<y&?wRy
zPjfyP&Fv^$n0G4%Ub~C~1HR#&Gi95OUP>aTX10mgzS91VG}YnBp*D*o{eos3>cy$b
z&scDFlLK0{k_L8a81qo)_5&otlcb&b;d7B!Q8j*1z-}R-I>1`VBqvzgW&>?`jhiGr
z^OX`*x%wA6;iv~rbt+j_{l{GIX5i{J{`AC5HA(Z&!C5eJ8??9L+$7{{$<bjPj__&D
zo2$<-QyE*t-#MDrfwRAe$QrhKEzJEnqyzE&0*;Lfm*Z%Bcuyg0&!pwyG2NxZf7V|9
z{7k<Oj$D#a6NAG4{!f5we5mbcxkny8<LXQx=ePe@DnsP=YOI(S{fpF)f1OQXdEiaH
z`+oq2|Ek3N&o6d1!<5*gjHOpf=tT76{0!5~sxKSvV8RNnQMXC{3ct=m>V(hAYGPVK
z?vix^l1xMdh`6vh`cO2*a9$(JcFE%Gv)C{gr0{LqwmBjSCR>jEIgY$cJSx$jSGto|
z0+vmE0RbNU`LhaQuHb&agAgQ|h%~300>`NxCspwKLBjH(Pr)$$l&10v2XvrEDWIQq
zyKaZ2BVELR%Q#!Xb0n!>0?3VGQ_btSL&|^TXyKmzeA3S~qP81x?2?@DhlHw%6H1Er
z*~A?om%x5HOZS}V-3L@d1aH5P5S`}@u$>IXx4sI`W@K!=9wcbe2Oq%|;s=T8A(2R2
zqlO;*)`DkY?xQq|ee@6~a_|l&NFNR*)-PffBed?=c2P&gXEKjSBHf|@IeuBF3mVW^
zahHkkh-oz!?u*9&8OhKIBEmecA5Venau%O#u!$P&M%dKn@#DuM7`e?j4vs?s0_3PT
z&J}3?X^w%CUY=nM*y#s|msu>Zk{msghAHg}T1SosOG7_No*-wm=5o9i=HDZgz6x<-
zNUoc(0)=8^gA5_TN-ju>P+y)`*-$U>yxrC0_#E7)MCf^ca=S!uYjAU@gia0_2aS=h
zu^(Q|DWCIksPMx5IH!gaD)^lMXX#K;Z?JpDA!-*7v4SMXlR5%OSPDK92?SOgHlV24
z$PUW#2q<#|=X76UN(PS55RF>62feF7zEg5x`aZFVQO$Z>Lm|dCt<1X&8H6Nv>5f0Y
z9J_w~I*pEKp(8{b$W*7u0SKG{B$;g*p71fBetsM>`gUFd6zjV2ty0DuJEYj~3pC_$
z!1ROad~a;HwpXDC>%-P02V9o8wGC4ypPu{)<&_^4rQ;(>4DcbeH{;4Ma*zShmp<ss
zYS6|qakr?KG-%*;Cz!<@huJq#whBKEdw<Mc@>i(|7#z;PtA;g9Tb#CsfCP1{g}402
z8P~~yc$^5<fKeYO|3u^|oV_^^@xIoabX1mj=xBG@#I<b-Ce-IWcEBf!?2w`r3*?hs
pDq9Ta2!H<5ugv-1d>g#Lq<JgJy_NU<3Gyo1DfyGnkDt5x-vB=;$guzb

diff --git a/_images/sphx_glr_plot_reweighted_l1_thumb.png b/_images/sphx_glr_plot_reweighted_l1_thumb.png
index a74ea19e6045ecfed9745239c837ebdde352da5a..067359595abbef5ad4947e5aecaa26b88f6fc09e 100644
GIT binary patch
literal 12506
zcmd^`WmJ`Ix9<r7gAhakNs&@oy2FKlq@pM-sNkZz%b)}VL;*<&K@>$`(Iwp=ETy{}
z7v1c6@Ao~=*=L+H#`&~A?GMiw@58%b-S>6P`2T-%J|Wtg%4BCRpT)z&BU4pD-NnPh
zkA#0{PZPsW91TBu;Nc0Rs-hI`xhJlSc(`AjNo~TdP3H{Is1Ohb<6pdH6{(@#|2seX
z^b6x0-g{r&y+U7K<bJT!iamRYa!a>TztFgFl%R;~o|0aU!g*9$u+mw@w9{t>RB8vh
zJ{&09YKA;gyT!uUvbG^DttI|JXwFA#HfgL~YOUgNTxLjcu!>Vdp))=TmEm&NKR7s;
z{p@*4$~+W4D=TXZmplqpPelYjeVB$vNlB?j&BDrR7D9kRp?L265BhLDc^t3NX-Z18
zk25h9A;EJz<J=vLpRE9GMotcCL=%=Zh2N$NS2Ux-rf6iun2BC_SXfn+Y%lJ8<MmDw
z5dprp?ARIJC7k=|Qv_KhDBiZw9-?Pj-m=VE(zsKyDG}OSuR6@H%E~6!*P{|OIfX1b
z4>gx^zki<{uX(w-_-Dz4Z)wZ3bgS1uRYm2a-^sDNG?RHpypn^1;NITeST;kCY>rWR
z46i{^WPE(<#RrB8=H?uB6*%G)@AWg)hco=?ZH%(saR+PFDJDmY59*tmPTIDJUn<BR
zZDQwLcfXn(y<W<#POhk^h`zhpC3_I|-KhL+V&d@f4xdR?;+HSikEi60r{cywWU9hf
z`W$RG*cNXr7j!xH>RAZd4wppVS&qvPio`aX7eu9`XjS_jJ<KyB8Yy?<l#@#dxxje?
z-NVKjbsAo@tW46<(h~3Uw{J6lC+akMb@cR9U0k*Vm&P#Y3)OoacH;Kq@|(dczfy1h
z{_{sfuTxl~$a+9gPmj)Nu5)I+jeqXT*Y9Z<%)9c{vQQbX?fEd&akfQ@PYh?W`-@~5
zuSqzji9q_#+1XbZ$fq{;K7arI#@6lHwQKE3QZ&UwWWv_inVA`l-nlJL3HQG&@I*%k
zJ7NRmc4Z43l5VT&K0ZFqd-5#)O;h-n$=k`9q*d6cBr+49%lqf&=Rc!Adx}7Mcl>3=
z&Zt+nXfGF4h_<foEf*J;P3xAWt-HFqmy8<McBZyG&ji(VciA28!8;DS^cgMp=e^wD
z_srBv>4cx)pAJxb_x?R4De?2b$jyTz23hajxb2bndD3ZNM{$O*%ObB;*lG*9WZQPC
zPke6NxUoCMe4?VEK^AM|N+)FAKJGPkz~r^FIx`u{p9x2?GF0reR=IQIN%Q$+-+lC`
z--!=I7B*AkaqGKl-CK48`2iXUqIvccini59uEw4V?4ov~>f#LM+sgy+WoMr~<I1f*
zmXvZ^oo=8pVdLkI$j`sw=;XAwiPh5k@d#gR)cq<U3AuHu-EA{Hh;IIhzXa6>BWv^B
z=_GjU?Cd-cBeo-Dc@9&zR;vzdZ{NOM0_P)fuv}2OR&}tOd$P|jYTf_SsY@mZ9S_kQ
zDz#ej(q^a#w`6h>zq{*x_S`vD4UKTyAhBV4LbX`Fr?c%Jq__JY!<EHWR7l70=+)|3
zBq^oH`aCSLzmlpQjKK)6maoa)y?fU`knj_YiFXvd-yI#Dx~)DFtz6^E{vsQLO_r6#
zuerpR73Y(^Rwp2wAn2X?jNI-R7_^PJjfI=iPESw&9e+t-Rxn^MndE)EH&^ku_8fA%
z%OmBnoGIRK(b!DffJI8)vu;-3!zt#1$4%tx>+9kthZ`TH-1wB=FkHEE<vG#0OFt~S
zm}a|Dy5O@Nt0Uzf#O#cPt@SS?J;(UVv&geig=|N7N;63_QB2s!EiPKFR%~c~uRiuf
zE@HK8i4Tbi$i7a9-iM`*?8T#=+<hi~G7)me?$HIIlvi&FSpMc>^4&FUZEL%B{rV()
z%}mjd?NqlibGUDEhxy~h-`R%tHAIJpb8=0>8ehJCr4uwe1$BYPxZ>f&m9a1R`Bdd2
znz4LFX}Mbq*dYF@-4`lsFWP3zYF-ev$BTrH9d8$(B6vPe?$z>^T|GrM3ArD)F`w9u
z1hIlAEmYqtxAl`fX6e3H?Jv%|mcUPl$u3imWT&|F>Q!uXN}*UOAxCQ?uiSi?|Kv<v
zAlX!>)LKS<{@aV(x~R4%w8T1DnlrK0hwO5@H5BfvrE~Mcr9yKlezMiO4g7JkKAz44
zP+}k^+x?E`{PN+rC0!Qlry}G!KgAliZgk0YP1FT>9IO;ONG8eG(J4*Q-4SQV%*+h*
zJ=%wI)Af=<wi#kE`^%SFVU6P*zmr&@PVtvxC&xR}Gc!0y60>*?-|7?+CU<}4lijAn
zy)Hjl8NZ5Xbqf~3p8}+$bu3gN_MH-Qzw#a{BQ*d5O+{0agOf8zOYZ0$R52Z6<IXnz
zYIQDZv%ljt{nkYsOCh15bpoW@t0QV%$ug?9Z>K>~-aYb_X41mo3-j%C_&GQ@X!KrN
z*RU;4fFR(auAOnjC$nJ3;c&0&O!s%zv=kKyZU&LhEM%oLE^V0=58>RUy>SgRImVTj
z6FM48=aT7V+^=zQ{hY8Lw-^7Mn`@WP*x1lO8G@n}vt>|=;YodvYy8B*B04Fl1J08C
zCzNazEv>mQq0ZnKAzIS)`Sb`#5JmW!`}gmkdZ{ouIr(E~g*K4Vs0d1Fdz|3Q^0lf2
zKBF@GHr`@wJ-rsFOb+EEn@fEWP~f9Jd^jVFDNJvpq1lOe>=0B(X+%5@QEEQ^^LdFO
z9|S;snadJ%vrUMoj0`sTm&X0AC9Q6grQdh`F_Gy9RvOr_?NZSyCi0+!Lx!^f%^e*x
z@NnbXLw1}J5+C;W_c>_TZA9sjJfaGTjE!xAo7VF2kr^tsYg%!Qj*R>@y+PkMwz~&M
zx$7-U@LV2Ktv{V0>AI|9Ys=@iTgMzH_?Y<p`}Y+`+e5=XJ0tgWb@Oe8L?O|bz5n7f
zO1ZK@LxqM6N8sV%aX7Cmmu5Fs^}DzCZCsqn-MdsuN=k<A69I5Gq;)^%HYRosAiyA)
z6;sd8C41QlQ-7~mzrE}#1;31Z&d_z>se9!p2Z>w|1@kP_kf#ehpODJ#wOY#Q`O$T{
zIf}#OEb&W2kD2#|4h<LEKL`bhSoKCmN7prm)3>y=m>z7es3%K*7#!5Eoe-d-qk|>@
zUpfHQN_Z36<JQ*J6La%hLz{FKi8ad1KJ~B8T-g0vN3k$m>g25VY_QjW-o>;KuTDTk
zRW*dE5Uyd^RtyfOJL-ySc@kU#9!Lki*ib06v3ZZ1*0Cwa@96C+qAZd<GF$#Oj#he2
z;oRe7yhQc%^r)$+?N<u>)_>ikz!jt%@px|IlPRl)FS(T(!rz^onp%M3USQE>K2&6r
zwX_te+Xx;1Cn3cTDDt2@8WJU(N+w@((~4TtiYAbik96|;;?5;|E?zr68at7>Opu+O
zJ?y>Fak$@CZQh%shg=Vw^K55Qn^2c@otE$V1!MOKqPNk}5=%J+y8~UmjlJItTAG`K
znPw>HZoGoa6EJV5#j~y4(lz$kx^G)DK@7jFrmCunNrb;m3ZEG$u-FKaJA9s#!^1#M
z5)cr;V^|UyYvL<WX8rUjyV#hw6GqK#bH4j<np19~u+R8$V;EgVMut5k*<)H_IH$x!
zhBDWcR`|c#AQjX+Jtg6XNWHb_Wa!N`(bCtyjQ7CE<)Vq-kq9WV=}`Wv@p!B6_$-2g
zl+Ok)8PL~ZB*hJQgfxUa6?pmh-hgaC$BQ$NR~Yk$goHqvANtCjIdcZy`$f&8P13Zo
zvUn&3o}Ql3+Wvueu8N3=JbC)mv|vEodH%d@@j={`Q0Q&&2?*HD_^_cB4qDQ?ckU?Y
z>kpnGD}A@k%|Nc4tCf+NNq{PT8qxG}#1_e;+^YT826mP2X^h?Ri-$BIe;|z5I60Nw
z(wHKGAk^;a=!8T>a8jRjnE#bJui3xg3>^;?2olOe1!HSj@K~>2y_zT5ycMPL>h<fh
zP&EWJgvvCSC7n+rc`qj%M!w-L-7VlBxP8ySKn+w*ps6sL8@gF*>+PPN9{3bA*T?Z3
zpsd(9I414I1xN|Au&P|tX=!OM$QdP}fW>1L(ud5f((x#xbl|-ln;6|96B61q#72Ji
zG(<(KgW1RmL&yLA{W~%yrgu865JT2M9%)qWii3JF<y!th+T#XfO!pWshz<Vg!?&Ot
z%sUh5WXPx~ql&Ty@Yh>7Q?wsEXsYr#=&$q=+ZlGE`1Pw#{Ba{GhekYVtjcF$pkPjL
z&1WiH4CDd5YdNu>s`a@5#m%)bM4=|fbQu&{US#$=%o_}>uh8bI_s1uU6EIcW2*PbI
z^-(;dWp&%s)~XYrm39|sXlSq=DSJO;TOJN?hi5<j6CbYn>w{eLo($FL*;)PMhZ@uO
ztSTZhKYyOb5H&++LnUY7=hw2iX{P&lsQ-uQ%R;>8f!Bl}%(^N(HsW(iaI6Dw+11?V
zlH5ubzkC(Yh)zpmy{)DO%5@^qsT1v35TU~JIy7{ujzUh_CS}gG97)j9($cIIA*Gz)
z(9p+Az2E!;=+B7R4mTqW(WZFh7R&jcYdZzDBa)!72C96dAiSZQDO>ailfDFTqvR*+
zdHqRBm$Hq`O3P;*bx5x6aQAP9YU<2P+&VF#E-R)fuA8od%XqjwZ&G?u9?2^MP;&ns
z9MJ|UpwZV3cGfJ`r(42k1t;Tn0u@u|yTTuT*3r>%5NEKN#Eb}#zNle=^6b3$TUF;p
zFy`utLrU*59-WM*aJq7Y?fPJbYBXNSdMme<yG+faA1x!jHUVujlg&|AA!CSwRit{D
zcF??%)9DU|C1}MYITa_uApw?wnel|WHVth3+y<mh+DZIX>a)-kLBdhf&_u_?G=8v1
z9td+pR~+DWZ&D@?{h+a98O~YbqJ5+7?#fYDW^uR_Fhs{amW__Dm$EtBeIae=+F|Z|
zqp;XVUp<U2XM^NJNk*`x7M{Uyh{y~3dN!GwmM;nCR2-z@4?CxDKOWT(K-Fq#Z=VJ;
z!pX}U4jE%DO24qMaEbsQUKo<@-8*U!Vd*+IpqIyk&K0!k<pD+W3cFB0g+pas7I&OJ
z4Fv)k5xA9>&Q95dF;#<UvOqnIC(V21!+x#5Yozw|%#)*S`&{yG>RR0{zm9WOamLx*
zEW0MJ8MU3S6&ivksPP#`{8_T6f`3S0xA#?Uc6Mq;#w&<hkjepLYZf5G;DPw64;@|C
zMmt*HbC-g;UBG5hK-XiCc8_E}{u{JBR2%_HJp!MQ_%BFjY@yWy+1+tMTkt}lHTA&4
zMRRJ+&cvED3Xqn$t#wlP{v}NEI6OxjIgXh)H>kC{Hr#BAeY5JG<DE4Ttoi~QSk$YN
zRk@5c%e>!6M_g1gdlY&Q(<-7F9|JNClmcYII0&9Mnr20gGq)EP7m@xUcZ`dGQ{WI0
zd0!VqvC&9x7YU!<#Z7$>cVK0TAPK5dfPSAKYLoW)b1?F#t=zewS5(c+SaU&{ZsKBC
zzrDoQY^pX^=dq$Vz{H6Y;s?iZOf-kM(FR`Si1O>H>E9)N-e_@tXvH;<f?3nmRTQqP
z;&^|_d3Bi1zlq5!3<M^&+|BOGmoHGvy0dl^Zk|j8&=@!0q>y=KN^9uy=YGY(YB^FX
zYHJnw%9n3+7;=9Lq9nP(pW}0y;mWhlQWbGLnZ}<Vh_Z6rZIrzFvl>wTw8R+=Qi2-s
zE@iQo)}&|6bn5oQ=5^}9rr&CM!&o{SZ}kvL4hRE75PnF>PGRURu+T!T05ptInKP34
z6W{@5;5{$%#hwEd-}l`Jgd(I=2=sdp<d)!WB*+=W<s+D6(z*gKg?I$wPVjFqhe^i&
zp`<)r=T6OXqUbe%JPunqLrz)g&-LH{yW1)%k47sz*x1+vz;1)YL8H;oQcqQnX~B~v
zNN*gJEq%AT5mXKNC|12kR`VzlI_G?F5S$gy{`t=!B@oByxck=>>VLd+&t0}{x^eYW
zR~0ur^CSE__Ht>g<_6njc{vE`1;ovN*SHy^P%;{Uz;2pfQto0n?3k&RSp0~?mry>K
zy=DX*CxL#0mvzhto^;`d-al+<g+m^c5;RI)BT~w!Xx2;HXh8?ESCJI5?o=<BG^Meu
z<0}5CVlOHe+KZtbo$RHYq<;DG8jj}H8wOQ~8sHx-4xAo`qUCODvrt*!22bjjSXp~C
z;V)6KoXBPT(mB(3rW0otcqNrEQEV!Xe`sza&-cjFyenD5#)fxeuFC>!GN?j`jfa(<
zB48x$q~FT(IdBD^i%4-9&&@}WY~FusNJuXW2b0l@{y<|n5RC-#O;e#kDMQ1O^8nL-
zI<oOfz!j5?Dr*bYrt&kb!Wz(-7yf*u0f}&UwB|SW=Ya`3RmiVWCyQk7bydXhje4!g
zxULK$&ewfDH3*TrVBT0ki?GSaB!PTc-`J2?13PK!qLdMC#e9Hqbs}Tq<0FUW%SL4i
zZHAOVJTghj4P->kb3&5&k#aF``A@*akTOcXbLx}``u+P}ne#$g3zZVvY<;aP@2yUG
z_0tUK-M&OjQ91qS0a>7_J=`)l#A?Jj13~H%=`*Z@dXkRCm#^Cxaa$Q|P@*wH3DDLH
zkRk%mb7xfpq!hf>@<2flkc5EX#2?yRD9~i&L02U;<^Bat08qJlg$YYZ1PnIG@`@Hc
z@sbWJZ5t%wJjmN9EE%!A6@Iw^QDvah-9|k)AsYbbAZ7;&1Ad&O%XRJt-=AzP{@Gj^
ziVfu}3;vp06%Q>lkd&@AnDQzYFYnW~7+yoi7Ix0us-DcFtz5sSD}#l=&G@Qz%|MSg
z)drkH2oQpX62#Wx=u3p773i`5iVNG4@tEi$N$_DH;v+nKK;`2h7C`=HdgXF{TRe}B
zwzj-Njoj>BDL8md{n88{T3K%p8)1mXuh^Xo1?PDO4t=sA6k#C!`A^lHr`zEQpm~B#
zS#-d&QI&$MyyYK^xdImPkw7z8Q)Ld#L=`|?Ns=zL0;G<Pj$HiwB$xTZE?vBMM_<3S
zGfC=CA^SJYP}@A4iPViItnLy!XX6y^FIYXtQI9!6v-;CO^mL#cfDm0Ct4;x=KtOgh
z2q-A@cvxJc7-MWxxZ-C}8PCX4?z%F$QaskU;wt2^KFtyB@&4VrzgcoeAxtu!_6wh4
zkFi@IWx)!8oD^qxBZ_MfTRo+is_m@*6zvzoXVeO!GXUU2!|RLciQ-Wp!hzb{ynXxZ
zOA-p%S($Jeg-=_S5eFD^lHjhn$3Ok?k!FLGT5wAwc-3QFa5QvMt~^$K-vJ>7L|s)D
z^;qksISY=EQPTNiQBhH;Qm_a7&u7~=*FLQvH$K%A!Fm4ld1mH_5nGGZ;nK?ie;Ny$
zeDn*7r-20EBXK&s>Nm^_j7md*HtY>C7l?A==3*l*=kgJa-YY$qK%Ij3>z^(xDJhx9
zPG+oF+8qe11t}U7S(_g1Z$TjN?++G-vvdb2f~w3O@Y)Ec7h4BN<TzwgOnLeg+#bSN
z;bAA8B|SDe#Po`7882{Zo^nprMrC}n+#XQ_qG(zdm<K+z$ads4Iv${aAs>y?Y`f`P
zXA<?LOH58jg#Y<zUi3B{f-OJ~IY;*)jT#Rz(Y(dOjyGD_Sy=@&H#q;hV0<@WtCD8E
zy|_9+MG%EhFM((SCU~WIr3Z^uRaduOV11oPefE4*bPfPUrO->DwGg=TRri}2+&EHJ
z1Rpp4%F;@CJbHm9s}m3H6K`Dz7g@14n^3VmXayN_vrB*`@1sGnErMRL!uS9S`mIGl
z2%&aXM-q6vKs9f#3@Jlh2sFi@9F%wp2;s00(1c)DXwqP60fNy&7KYIYHMg{&7aU*L
zzXD7B!>r*fVkIFV5Y)vN`?ci;xX{?Pi)63hZXAK-{P5T?N5CP7YrKEN^gyW-r)#<7
zsSKCAn{^8tP@kLI+Cq<6sn6aqGV*D^Wds?U|Fo^Sv(p@OZFgT=YpXk?0$`KrxjCJO
z58dOxusC``{;u~`AHNR^BL&QwAZ7;uPZpHKUfk0N*@WiS)->Q<8J|AofmBq-LRBQK
zdo+kach9`68ApE&ZEkr)VlU2s6o`BGP@0<L%s!ixP|T)*>7~1^jcSaZ4?(HizHRpB
z%Y6-kCxDj_uaRI|76L^0V3#@bDi68={;q8?xL7!GpvSbuLq6?};J%~}|JH#F-Z<Ro
z!UD`fFfa5lcpMqq2k*;5KZ#i8E1v*S06gt|h-*3=sy?EJ1eD>B2fB?kGe`vVx+k2y
zgoFf-K~b2(OJawa_Ycfyi8Imp)&n;XM{@o8YbJ*fj3{{Y>DqJr(|dC%b5MOJN#u6%
zL02{L#3m<qLKst#CBQU;n4Ccclp1(;UTw!Hope=yILnRcXe~!L*|*8bzQc{c#jcDo
zi1ef@Q(WRRQ~|;a=3NNMc6N56Isq<#U~4-%ckZNK<wX}>E*`n2lldr~1FEJrmvj4v
z!nz6ZOapFN*<;LmooQ!u{!9de#9Od|Z)0M>0X>4!(^OG2w{geN5FE}K6K@<5h)S$5
zzy@GhgoqXf`^!OdiJo33h@3I)st!sY{xS-{%I2eIYonEit8UdgeK{C|ChI-0pm<Fy
zLqx#iEI=tGLTf^92dM7_8X9Fl>Qjy3P&Q{f65QIGsh^2EQKG*Aq>(#X&`Lk4J~=+f
z{`wU=5CDJcRPfkZ(-aPmENf(p^z9o)sT&rZiBsRLL2P7w3}(%PC&mG98HZWIDFVPk
zi}8UND-8lTIN6EptmInT;tQuw#R;RS;2W*7^F!*x9to&$QBQ(!3pFNeZ*7%a{ZORh
z)y*dV`t@tXQgZR|guS^>6XiE79~T*Up3#VwnYk;;Z8V#Y2VES0N(CA+A}*kRU8M>^
zWFg?ZzPnE^gwUnm3O56G(7)okxU_U8K(Sk@Q)%!k2mXymKk*<C7o4fcgMggz-sg*o
zn{+J)h*$xxF6Uw13phqc#~am0D|Xc`KSNXuEJzUV1EY%uVGT<wtGuBVuzKFICTg)-
zZ0112aK+U};ZVy39j8^mGzT?J-SJl`e;O3nG6r=CT&VN>FMLqJ;tai~FGQuo6Cu17
zp4_f#j|1u#uMu0+rHd=ImKDY64!+zeT}^<|P=<yKV3c`v<)a=(fkB*ZiEe3ag#!e?
z(>L#<!@3g1_!C~@Y`{azGfLxJs8!%#!L2CMbSoXFjIGV(H8G__5`{6#B|~?_I7gxM
z??o>RSDzfM>D<4s#71=mIv4~|3}1>$yNZg+_#3HJXd1HPfapJg9?eSjyuE@N{Kk>~
z1GF~KH{Sb~qT()vp!y3eC{p~6B}5a#jVnE&#_E?nWaXWW%T`y%hII;jp+BQ^T(h*b
z4h#fa?dRtU%o7>Q^(axH??5adJa(cs02!%(fzdHB>4IDwL+<Pv_{(9}Asf)mQE}bV
zsZ5Akf=mGY15W^WR6*^%8VY4*7>x(6KNVO#3?@c>c8s7r|Gqmz6@msi4GBJvkM8*{
zawZZ|E;1@g0aBAZ=w?L9#l1_J==I}+HE#I~BP&NlbH0J$AuuFlcnG;ybZ-D<b_;3`
zbQh3-AmRf}1$+<P;nGFkhhPlg;lWm#LnA=++am$@cBd|gDJ|K(S1?J!LH&Z+9Z(H0
z_lS&ocTERjF={k-rZHAjERcoJNpC3|qeYFQh>3`Z@P>U4R;VvtR5Fo*_=2LS5knSK
z2QX>eu@p)v!R5|J(BR;Gq$`PNBmyAXA=BN*pOYYC$_e|2T5Qx3ultr}z)Xb%{#^X?
z^@E(BEHd)SHw4ZRdWpBzydV$e9<HmyEijiL|H&R}m?p|@b1NDT;VNz;E;L|W?XuT{
z>vJFtL9-T*`AXsKjQPodd1-1N{q@a24Cz>DN}+fUtGp#3nmz%)ABHIl^tRc!0<ZF;
zI4@5nGvnxQq5u^Tp#(E6O7fs<JUmZ8$ZL#h|I?h_bPf;Ahl`Cq0!5)7FB}5ZCuP4U
zG?4==8Pif<t_A2l1is>VZ~b}DR7qcnQw<0gl0ev6LIQ$05ukqL3$XaV7WJOb%4nDP
zQ_Vy)hCcy7;|p(2i5`1nfS<$bK;O5dVp*#`k+tg0nShasIc~B+dh^%KrZCO!J24pD
z@bn8n<s4zQ3*EN8z1^--(|uD|9(?7yckd#Bx5F4rmbg!V^z7NQhz7oQ@0rREjHK(c
zlApy^eSAQ_>BVeM1<MW=fnd-WjeP&Up~PVd(aq6)p`oEjn}6K&=IKo9duYD^VA9ih
z&=<hDD8g84x6S0Z7A{SY^yQl0$pJDik9*qI*5<K05p<1@Zx#k-PXWvw!emISEtY>S
zzm3m|N)l`l@#XJw@tGcji&q_Q^rudEbd8j`q{CbZ1`*6)p#*5zczIV!>|Uc-Ii>3#
zDC_m;GDtcH|M~jBX=lWZ(0<HkNATG%mK#Ds3IFswnW@h9_9w@OdneA+x-dE}44{!9
zN~3tU*q?Vbw_^P)yu>Hv2&U!X(rA~ZUPh-^GaGv#Q!|0xK=mbvXu8QpH337l$1q#y
zw*3lt5jBRq>p4K&kdlD!fU`!aoC>OQ7@oOYymt%yOV-yo)+~+9wEB1vD@3m#GXv;u
zFwzk8+24FTs%>Jz>~pvR1$AMxQYsUD!+Bo&Iyd(~-eVHbr4p+nrIM|0dwC%2ta2-3
zkX}U_SOZh*oo^NktSbt_@_-LL7PGUO+uPZ$T)_t>TzO{k8D!Iw&P0hOqXE!S&dT30
zXc##{Z!dNhM42fH%R}2EBqn`UlLq@GR%tEY|9kgH4yuh&sH(5PyMZeG7_f%P*kjsT
zYBZ|TQab*@iDp1BN}cB3Li}fGCY1ns0bs~$#1pmFe%Jp$-8b;BEKpxzj8dZe`|C4p
zdw}<Vn_GaZDKY#Wt*af?3zi+K2vjsEa7<#-%dU)Vs!P~qD;DG2SY;-USJ1d{K(Bs0
zZM%dA0qu9Zkpj1P4%(-rt^4OF6|X17L&)Tq#?T%gnV;|OPDMNqe56<AjFgzk%AL`=
z344sas!}}gY)6pCU>s<SoGGEefv6*VL`w|bDzl^{4(5HyfZpZ9QJ^lXb}H8dt@{OF
zU=OhnQe~{I{S!7oYG8t>#zk#8@%FWFmzym6vkKa0w8UZ~&KK}tcsK!#e*!isUW-xz
z^HEX?xEz9!0_y^=aVob@4c1zI{1DhL2Z2opqjM5<taUc)>xVtdcrc3+0tAAHaML>C
zO!W`7-~*Ku6!4L$-Hc<N*N!bfiG&K0-zphCuOlMn0_kl*y@7EL?=Y7#|DeLwL%@Dv
z&o+J};3DV`kcqWT%zo`)>ggn%xnxr)K<|Nf2Q!9D%3v77uwg7hh0c|9o=2*Z!|>np
z=x;nOrb#~A1H+z6-?~q<;)T!nb3<-Tzq^f^e!_L_#fVi+Z<r}|0&kO+(Be{JnPVa4
z<xA|Wf*r!Coy-}fm17*lI6p6vKB=AO1#}j^&jrGi@f`_Zv55-1c&Na^cP(x3!p}=?
zlWu>#90Y@o4!z!9uos)U5Fxx4_K6XqNkv`gf5BHg-1(p`O5za-`&QVLVw-IM>M5SC
zhi_IVuk$8JKR*1yKP>D2TB-`KeNB0+{c28hxq<$)*b>^M#{N}pfe>A<%oKHuCTjZO
z*+|-n|4oGx1n=R3<?1VAw_xw88J&d}LiukFn;_E5_|sHJD<e43cW~oho)Jx?-l4%p
z_40QBdtTS<3mL1LeysvgJ&iVA$J^KB*|rA^e^`)t<ETEOOO?z52vGXZPu6Plt@Lf9
z9;R9SmlZFXm|&H22bvKp?EY<kvG!tmLbAUoDYj>w&oH<BKQhM?6t>?!HU(J?iG>9R
zx9k1gIa(W(nST?l17|*^VDjtPwl<n7V7~=&>>4-sV{k^0tHsVyY=OV`UaEUzf)~wY
zUAjXP3|7=iyQ`?}lXgjnr3Lmixk^3_JeAQeAgl*L)O<BUz+?T1?sr4gTepz3fVG<f
zcbr+{8^*T5LMzZTrZA4x7_~x{I})l3LhBoF32s#%^JV=zdc)AN1g5@E18HL}q<Snc
z;6J#YwV@Zs7fUGAsldkfqC1R?Ytr6yte9zZh{Rkds3cj4+_BTnRqtEAp)y0ijl-i*
z;X@nEdxo?O9khim)LGb&f#JFuzeyM*PJu^DH>viOh9_)Kkz-C+VSZXLC<=U>(mF1X
zv3SIKbmz?2f>hxb5gRdGi6rtbJ{<3jqi5B598N`l#1f`D#*6%{Un4r~;N|#*u@RP%
zM_>+6*tAf-H0Vr|lV-#eOzJK!B2dt9Q%!{0hKB9HH3c=Ifp;O0$ZM@qSxJcqBogPa
zI2ob$a{-5jPi!Q09iFI9=cKHd_rm^SKcTG2#k-Fb*iIV-$;YOW$)%WuyKLI<HS`Wk
zUM3$Mte{Le(qVPz=*JHrwBHW;HI=5Ttt@3nB-=1|+tQqIMY{lBbTHw*a-G2EFJD+C
zB=+j%mQ0Xp_Cdgu04`$|s!J-YzX4nXQna5mbjzt)hn_e&ID&-r%H65x_&ysm+`rF2
zq*w6ACu(J(rMO4^dW`+g$KDUNOO}0fy=tv*oo$0mJ2@vVt2~bFDR$wd3sitg0~3~i
zIlxSGvi*a1Xi@JPxxJ`w16^W81J=@*T|r{wb#7SHiFZTMoly{PUU2`s?)H7^YJ||=
z9?r{)v=TK}dF%YOKVDJM*I<lIjFA+zO`)U9vVGCoXp_f27-&lCo#)1^X0q10B{rrg
z&S2S@P&ktK<XiTeF>87lN<vYb+BwiGFbC9Y{Rmd9Dqt@U^yvSIgsj(-W8NkuDMbu3
zIF=U~(rFeh=cs!+ub^x-b=7EQ-l11qfY8bL?X{WoSNq99>FzE_h_R%vid5Cql`OSz
zwvOtOGtxhBbz;@ee-nsKO)ti=dIAZi6|tm7eye1&M{NLp69&MGs@1{AM#$RI^0GA*
z6_tJ&S60{}?ZIXKG-j~3P#p6c_|v?knOa+0fp}PTJ0@Cw^C_30<keGFL)JWdW^LJW
za&zaP)@vAtqFC?Uzklb!gFeMik`iwoh&>aJ=i%qK%Qpm66Ln#QA{?JnUkodZ(9`Gg
zc>W-GA>fdG$oGYXcZalx#0Lg{hra@s_Y_1TUKo>1d`mPZq5a=+1+e$QAhSa_lUew_
zZ_UgW|FJnaK0<UP0QOF~6K`Y}3Ldw%=aZTq;3i}ZL+#SRt3;KNKr;G9@MIpF-O2;y
zZv3!EbGRN`{RB)n&%+=7FqN^b+MA8ZxuvdyJi$f%vEFqpWYcFBGRF+`7y_oi{ov&n
z6d*lVGb`F&e7N=#W_D<?$nwv1gc}19MtELfVIi_Bf~@*{oBe!<TGYLFuO3Jh>>E%}
z$Q=p*<H#|oPCh*J{$9Q+`JzTgZa81$U*9@-&;Q<o2UrmMAW^{*?;eqfwF2}Swsk95
zHH#1#(=CQgCjh(l<s-0W0gD2fSp_W`mU*kF#aY*K2!boz+@U2U;_wk<E5Nm!5ynwp
zPKvT#4h9Z3X8+10K*)bj&s|VCr-@(IH#8_65YkuI_hN7iEAB4mh+jT=^5lo#i7${T
zXwxtx|ClO=DRfqb6NDdu4Sp9J3kniOx12&k7k0P3BBz;ui~bhX3=kDLyEK1X>oC=b
zj8nxKAfj*h9Usp7a3hM)L0of8n4Pl@w>ax8xzLx(eC`qttW61kltiXb2oT0<$!Wk8
zY8MW@6NZp~YCoOKX<}^0mgg#ANet0b$TH!@ix-F8r@|33&uf^_pHY-r<ZylPR>Ily
z&uH_2cEDs68ST;%1N~cHcLcYZiGKUx!@%_j{c1fiyyBtorYWSh<m3d>hamI107cle
z06O4MUOg0+kX0Ab+aauh1Ue|-1!tJBN_g*r`y6+cRMoPZx?tf;Vo7gF&uNfw4z`>n
zWb~DlKL!WaGyHqSIm9&FC$0(GAIsgtlx<T_Or%1}&C5GuA$LGWLcw&3<^1U%FdPCb
z4n~&;0UHJ-uUL6++FTXok~ez@u&$+}<0<US$k#<yf&F8K%@_<Ci!3C5jmIiaKe;m$
zOa^`36_NqgV(N~M4)77VP6t~Bc#z50VNPW$qH`lSQQeY0FI7V$vp*&-vjsMKnto^N
zq`Sqf!c67N1&$vue*y_lC;aR@!mNPNLT7+^63hmXArur|m<KQ-(h#5mTqfMQ(39p!
zU9YCR)1PhXsgj=8e+I~QA-1gnS$}z8;?n>#J-{BY$gu%XPXI>NozffXvL}a5$UZ4d
z==#fCxFPht!Bl~J`{6XF3Eekj#C8ra04Tr0_-hQPs_?;E%fIMU!tw{K+Y=?CVG<78
z7xf5)fwfF=xQGiJ>K}nPg6SoO82~bq8Ytidj|_#Lij){x@>>`z<o@&LkN4h8EC_F1
zuh_ii?%$$jL8~Y`cEFFYDatM)QrNNtZU|Y=g9SEiW8*j21vX$X;OW*u$UDt`uz)3(
zG%U*|%%B1+SsX1t+hG2E#hP;eEVuXL8%a=6Mww&jKs5c&-f#c^Z~O1bm7~Rrc4&M~
S;J;bGQ&rMLeOG++;y(a>ibx^=

literal 12504
zcmeHuWmuH!*Ds)gK^t^9Ac&HJbcsVrsVLG2NGaWoiV{)+3P_iLk}^mmF{B7cHzN(w
z-Eh{k_y0ZDIiJpTKD?jKxvssxz%cX7eXsSawG6zkDo=8r_B<XQ9*Lp?`XL@3{(JbJ
z`Ya**#9lYq4G)huNfCYLk!$?YsE6yL-jw6Dm3Cc+YH3QDm(=&e?MVnwY$HAP*{A^b
zg5qS2waq7wax}m6-Bi1&iSAL~GVgWoaX&whdjImB0>V2Kfv7870f8viDLKj+NsN!r
z)U$iU#m_E0i>OSR9d~ma;K%e8^n34Udn|n?=@y`%ps>g-E#`h15KyFco}7Z>8yX*l
zLPcGdMx*N~&jbVnXr|)9i<B-gqfkbH1ZXsx^8vg#;M)I7FCM_h<X*UnLM?^4GxIT%
zpZ&AtuC~8J=zHg9YwMb;xYhbm*BZT*K`0v4)8B6#+t>b>j+XYUKej2&6{p2aUNgJF
z&|S<J<}b}``&y}|c+pDlWqbYX#@bw3*u=)(*|#zSmVy(LlZwj9_4AImJhm5CHurAd
zzD+M|`_lVp|7h5%N}oq))bZlEbLWT|#i_%Zu<;@e*Eu*+l9T_4eY`8`FNE`q%+)E0
zh>iW(KxLq8YRZ=XWB&$|*Cs{ves6y2FoURLfX8x~zro3Ve|=Na$#4e=p|9uRT2trI
zYJc_d?S5SclD)mXrq3fup6ll%JhoXyM9O+s^vdk7_4f82R-f#gXdfn6#D4hjf!Fqr
z^rEwvn#9(n%tZGa7ui+C85=?s|GtVx3&w@Keoa6~LR<g!arWG3h0vKZXOutPdizB^
zlQIy!y5|X}qpqWKL0V6b!Jm}j`L@X8$B%n@dWv=rxpazGhjGVIM+ZCd9Rs9{;tF0~
zUYI52?q>^qxgVmV`F)N(xy$EJ+!c$re*XMfdbB+}h%pOuXI<x!*y>Ri-=9m{TpFgz
z?{EFyudiff#WikAozx+(qN3tht)!@Eczn25I%=J4mz0;4WxnmEpC5gSi1g3K-av_s
z0p@6Re<7cpo!#Hh&u?z~l{8*h(<y<(7zQ|g1=g<A_*c>|*Vot0>CT@f7;(XQRl<jN
z3ih#52Hw}#SA~x#Dw*1H=iuQ9v!2{JSaYv>gB_pS5jxsQx$FPWgO)k4laP@7iWiA|
zkS2#ue*W!s9v*gTVuFyX^E(G3j&t{ewnr1Qi0mmv=t&zx8I)Kl$G>DIP8{G)Bxz`9
z4(HW;Qj3eDF7Q;o*TR^tRjqi1Re9{#AMeeoYiPjh#Yz|QDyaF4POFKponEQjAeBrC
zQ{parGvA#lFXaR0XE|8-wjqehp_Qv-^>CwOW6WcPg4F+$m<zXYQ|K95x}P`mE9>&~
zDqNS#W}7=YGAk?ND=Wn%k9Uo)aB0^&_3JhLPPscg7agIgVPMdK>5yz)@j3C>-<&@s
zZDGN2^GRO%lYD*4vC2<k%cZ_uT@RPb7r0f$mIRi699SPOR38Va#0jo#cB>i3vK{Sj
zP4?;7m>g^`9nJ_<UzL<hOie{SR98<s7>Ao{U7v3KB;g@QCUwZWG26i|C6#2-8fEwQ
zOXjVoHF#He^p&q)LUD10RIL1n`|#mX@)~-+Y&OxjHHxdJug|{icK`2u1F2#g9HWF{
z#o>DEt48wjjdn2+*Hz}VwY3rNgQcF{Ub`j+m#5T(Wo2c^WkTo0tFk9~?{joW9&qlA
zx-fa{EKh%HVw{Xn;vSDwlQe9JxTfni{c&-mEb{Qsvt9DQa;77xV;t+RsG#t)GgWqR
z5aa*$0+s!cMdmk;9qTKcn&c#L+h~dRg5+mU2cS`%YY)Pje48e((vzyIt7EN7MjN5u
zup1px9ZMy+_Ftd);oYAMV9dy}pXASE#92?ozQ`L7muIx-tDu%}7l1pOenrHXoSOQ2
zV!|lV{V#|0k3CaH_qmk0ok2@M(_b;QBD3+>4xeKO<?i&148P7adG-7EYq~xwU1nU*
z{rdH5!SK@NTqn!^&`bPNyHEguXjG{#H*&WCBfFz3-tx@k0gMIFTsn>LG-X?TI&KTO
zC9W$~JDSCoG*9%LE~`oIu|^wsyE#hf!cVY%BteWlfB=vBa&_ccDGi~<f~oml!&ULw
zPbpYeED6G=YJPwIxJcR0kk`18*n2D6U=7Q3{5~_uJ2pO^9yvwtwIBmthnZh<#Q-st
zJ7XS)>(SNE;IthV`a<Lx-QHZ_E`4d1=;n9VpLDH}&gPG!=&MFXS@p2%aF9AWt;u(M
zuV26Z=f}aa@kpuNqbE<GyiN|PPn33+NB7{fq`kc*J0y1KGb|={4$9#D%hrG1+-o&B
zwzofC3+3OEHhO7DP3?CFx(#YG^F@E&)YKF&lvzY%WJ_Dy^kkTnDZJIAM~_zPF7QCt
zQcHRY3&!OR&ek_H7*Z2LoqV>z{b0Rzjn8%U>7ln2V;K7oUTpEfOj2oSsc90&x{7s4
zoHVoa^-Fcr_V^)H5xcvtb>qCgL}YANufD>xiw3bLx)S8~2aQ@cO|Gr(%_d=-m$WSf
z>CEEo<Tj^17Zi~DV_#v%lT%Vs2ED1NsF0(cs1M}SDGml&2v0~*ads9&N*6E5Ynx_%
ze%^6qOhUW;DR1*#G6^)aCh!HE%SB?sz@Q+kb%`;xoZ~Q@zG+)*1B1)RoiIj;QaIT!
zU%wjlWIX_UM4+3>!1E69$M1o@P)-Bg5vh&_22y!Px|jLY$FTrt6ZU1)f|k^Df|kI8
z|D0At#9VJS6`hc^_?;7N_9%k%5PA{)7qO;}zrN3J*$R&=R^Lq_cXM-thfC10E_*vQ
z^^BI5HdgOL(V!*0q-Tr*s|<2e;v2*Ol8Xc1PL&L+$jHcGR&4eUGgiXj$!9kx&Yn)4
z>_y>vzQ1?^P`o?Ibkdxog~|H*bs|J)>`$oJa`M-&@0Ny3Cf8hDp^4{*zTbq7N24ex
zDJcQcW}Bpr`2g|L^YTLL>(LPr5!bo7XGR=4u2B;kQ4>lgkyYcuy5|AX;hu#jgZWca
zZ0A2m8<h;Vz#qfwrlADt7q$e&_5iQRZK}4ev_|uo&9*0uxXmVXUi^;rClkt6=cS(N
zQsSAO2;wmvt8~jTs7@Ri(G|2AQ?|C|7IR*FI#gr<00h|E(%hU}SxH}w>)2izPD`N*
zM4OwNGxPKF4`o!ith5s`dDcv|M7o_E?X<MC1X^Ec8+P|dyP6{ixW&Q2(TTYd-THQT
ziH)0^Ta>@vx+L<$htn^+=j*5K?HXy!cn#}Ld+&EA*^fH)uT4d&3C4{LRQvddGm<&X
zwqLt>GpMdkCa?VOjpgbSZ|`kPlB@^QZ`Y`v%{}hu3(>8zo}S`f`<sT#BW0GjA4kuQ
z=)2knmYkQXcJ$$!0Es!o2-z?~mGH61ht`{73`>Uh8hE^Gp<UX2j%R)3?lRxL{gKYP
z^c4;e9`M8;VFjqk!@Ui+^%nNM7B%lW;1DHy`&)>BkUALb7$~-)1E%Z598$2z6*$bQ
zHij{g)u;%K)+_N;)g!zBf6M{6AH@&v4X@#KS=K$?ZIX)k`0*_?NLE$=@Y@F55xN8b
zZ^DnxI)HROlujr#>2Qj@5}4pu?Cl^RjCQFLPhxD%w0(G7oKi=kMA_f3*{*<f3TkRi
z)AoGso8~iZaZQaObc#w!r~VYSa!)SgSDS7x4tfIH>e@A2#$#n;BWN2o@-U9kU^fdg
z$*_>)$!ch5fCA-Jxd`przi^F@FWgo*7`MU>UH+6;ftNaXoB8?k=Vwn7q-SOdT{vSN
zfBT`1j*^DP6%b5-1{2N|I?7taZ)Hvs;6u5l_CFi37OH*v{^!qoU0q$F7C%2b>W|wB
zyL+K--@eRT&f_7@2m%ZW!*i?m$%tdOiul1I2D%rz3aKIy5fOLtEmv`<wD8#2dvKj8
z_XFT#<jCq&GcU%q$pIcgB@ZVZc}s!FWMySV^CSxaCp^;7fW~-j5!>mt|EJbhm_BrF
z_U_~&Z&=esl2hiEYEDj0&rD1VspT-ml%)QE8VD2H3M<1WWLP|W@ZgS-5sK9R6qeUk
zG~tcR)f+c&&UbyjK<b~S6(?E9RV)OKLB#NNEpy43`%S<DU7w>ZD?NIVcCZ%qbuk%=
zpxP`fEZVrWm70G4#oJrNbhu822-#MQD#r<i1P2Gt;c|7$B6z9~5g;|~NMu+dd>5e5
z)D5o%A)S?<KO4mJ<0(Kk(v#AYnv;MKF``a4kiG|?<{CQynTN3&7F={zIgS{(0*Ef|
z%B!BCEW&{+F>RxjJY2o|og~l=j_c}OKVG->sXnzy1_5*O-@>xX&0i9M$~<>%@7=qH
zlq)E1e!X(XS^*V}9L+mgS~S3clhM_O*CgCG3n~j@n*;kCnN;lT))X(iw;C>fpsA_g
z?fqlOuRbTE_?tyH+Y5Yx5G7N~k<#!p7p@2#{K*J<1}esQwA@M8dC&xC$<XkscXiU&
zuQvei0ATyqzJxTNqoks0I#{Xp>H5*J<lLR9#st4q9ddc6Nddnowi?E}AVg^mLhN9<
zA}r44$0up#m$(%(;Ok45FMl&`BGu)l0(`M9PJN-ON{R<9_eLgVZUOXSd3pJYc2Z1a
zL_{jx+E7`srlhz=6erb!D3))j`KYKDAAAb6i2XCiJB4eLO(1jmh<ytSZvlEW6<KtH
zOmPg_VRS1o>8AhN)ANX)RNjBCn9lDGFLgTLvh`@Wg(7*lYC1n-w>AUwWMc@z*lqf8
za5;F<`tI|5KB~Obmd+(k^5gbp!k)XlwNLNkqq+he8;!N&RQJqdsjOq~uaI>O)d&(R
zIE(DgY}$7VtnGOYSQgJE^?`o^pa&Iq?b@~U&!6pna%2{z>c*~k3mo5nlPR>+KzFv}
zuh4lvmP-jyEzM&zC*G5)jX!Ok$Fi-d(mhFx3~SBn4Gzt0cUBGx6!G0UW>$~7tUh_}
zX^>d49C>I=I=irz!`-VK*7BI@4Oeka?V<pnuv`56Wpk-P3Q9_8g@u&RRqpPh;1izm
z27uDLcI|C!Z0uYFF^CF8>LMrzhcNCa3U0CUGa)IHq%t%MG?Z(Vejx@Ec3{&*;pS@?
zm2(vQ73;m&w7WuVJHRxNcY)}-bZFgA8`^g<j*5<JOf2pk(yH&<?cquB_*AmC2(u(d
z%9)0Fcl+Mt=jMJcD*AB$dOdiDcW~?787crMh-q?O8EgC7n=P_4;y}au6c6AEah6O+
zbMl^R4OEDoA*Q+Y1|j%F36J&ViWU2#t$y_f4{E{2#Ryu_@>H%8_!<7LaIx{;#WH=J
zU9+K<e%<ZIV=gu0Gs{LAN9rhVJS{{buS^*)SX7cIoG=}8g+KRJJ8wb5_hq$_l1F(K
zokidbP21xYp{&8CpdBeNz1dIH!A>-XvmnB(rM;aEupQBQf3x!+f__a`jEuQua-K1a
z$RB%0U!PH@*fOQGGzMC86$tj><Hxdagn$P2n`=?1C0~5irfQrr=K^t4QOw*cJhXD8
z?bHAfF_t2-xtjW^rMsVWH-EsAvsjv32ue&WZgbO7A9VNJ=7c|!cMSNlPN2xI+1cm(
zv7MPacP4ko!C15&?cl-;sgaJX1XT>)I|keecjc-~uI|kxe+r_jJe6)z!ZN>Oe<%p!
zc{O4=-m>SoGRT?<XEdN+QWK^(i1Ddt_l_zBxfGWU^z6FHZE_gh)W}JOx<J?j0Ll+)
z31KA=ooQ;M3&SNmD?bjnAuC{%aJSu`&)o2rI(!Avn+6zv)H>6wBhgiyvD|4v?%_j9
z__2tZ)RDl!&Pt|Ql4PgZ-Jp7L>~1xl)>%1;T-t!MZwx8VzxPSQ(03c5q$Bi5`$q1Y
zyEWDcU1Mb}cv?#UWJ4|Lc-^@op{PheE|~fiZUvFSZHaqcz$`*KSlqPz8sf@<#U4I<
z=-7zGl1b5foqcq2d{hRWQy8QWg4m1}9FzydwrsB{Qk+-@{cC3qcjBDvkt(=|Z$w=b
z<Fq;|*%8prJva3I#pV{v6Vl_EZCt$5@6DI-EM6F;=no%y#%DZ(nqB&}LM=qCuKZWr
zq?mc`J@AW_m6Z<FUG!9Gz*}Q&W!+O(r+i(Qe3Si+JBrQ<Q~t?Ws2TH|9<~+!b#mBo
za$iqR1&+qJG59JN@-WH$xvU^@@3_$l7l++dLqak}&qXH+ih0(iBI<MRPvRcVeIb~n
z?sw(GzW;ruJ-&CMX?xETLFF=s*@sV_G}XPmfcWHag*SiyKJUxZ3xdW1jSBxEMwd-B
zULi@+YoN&DB64ehv&PVMh}{-(m_Z2S;(a@c`S+qNd|ZMA=@$14oBN?#KQ(t+zrLjs
z2|&5OwIDAeLvR98#ufA?xJB?Fs{_ViYc;3HR=0<3HU`a-_`s7OS`Cry4jmFrU^o@k
z)YyfDqM@hS#W#LU>gZ9-{}FBC%k9=Zg-HE*N61?Mu34d-O#1WmfY}RuPQ3SbRyx4u
zBR<9~!RZQkG0`tyzJN8n8+7fR!mmI$-r!AX<+JqGhuyJ7#T9h80}?+&+qvIYkB@h%
z)!{hARw}WVpXk;CnxwVjOHXc$N*yfnw~MWa1F!oTuI^5TbweL@b$27?7z7qXH#gB!
zF>htAf+MsQrtfsyhqGK?w+H<EHf)?e*mAB%B-G;DH}>I4!^5RMZs9?lynawbG-A%2
zh|E@0R2*?0wt|QBg;D^&6UnAR@U$_QCP%+2_WO(eTM$PfpU(|I@GdBb24avj`LK+f
zoM6bq)_s`{|LpcY5iP1dTC%C!n~8(wLAnhtu6d#%NW}IJ9_VRK&4T0x5?$?<WZ!=a
zgIp-yVbbnEIIgV`4qK;E5)?IFyTp!e^0!WX*YUS)#T=ZR=>@Av3x~SUDowe%T_VA>
zx8DIZVt{26J(fk*H#WZH<U9ifG~%^0YB^edThzf6`eygY8%|<V_+@B%B(x=qG+LA#
z3=cR5HJ|;9AVyLkeF3fjg9!-*m`}VAD!HJbAdZ_ijo~I~ha#LaRdxnKSAXxN&1Y>z
z7$c5nF2!GX;}<^-qy<bqhkSqsO#&*w&GNV(UauPMmLLsc-N&i4Iw~FXJ{UeO>CNuN
zUx&*2$5uO6d=j+M#xzvkQ&QU7NIFr5iZ1>0`oi(_tS~*PObR*JGSGsgYgi$M(=sXd
z9kop6J)?LG+CdtJwwW6DetnFDd^tIspq=IK-8&Cn$Yw1uBPhAc+X)Cf`xHc-e9xN6
zf9y_NVD#QKT&~<u#RI>r6wPf?_vRdOE$Qj$ybxr}7&xP;MolrDpOvWo{{5@1^CqVM
zvr@E%)TX{@C=I_UuSFLt!7Ds)cHl@Lg~`{_WY9Ge&-m`H#t1qM4?@MdrQ_Mr2+*~s
zeQW0|Uq(bwLaq2(#O7#zdkumGN+eCs*_{5S<T^iE^fh=v%8M7{s;*E`QF&>{JL)DZ
z#hA`NGDkwQf2A4P;Gmd>)z9$0re*-ZV`1S#JtvBA($wC_4k55ZfDdJGR-o<SV)ebY
z9s`csi6&HT552G)D0l^Ag~mljMq)WK<7J|`^+fASlOa++OP<ux1jIM^!%J+q*a|D4
z0>{A}9ptt#^X)}H_q%uRv~^EjHdfflqSKAdx5)cS?Ix*(taUN{2%Y1(Z~bim+v_U+
z^BM}pr@FxY-<8LI+f;U>*mEAYjTsu!2JHyV2$BL}$+t4#Wg=r@2Kr<#iEx90Z_)w{
ztQ5srYl``1Ie4otM+*z;0il=4rZBUbm-t*`7pXF9@hp5|m%e}g{24hR#F!zm+WqT2
z8q<^(CCK*zqsOOm@g)e?Hq$hn66?+7(Tc%hsAxpogULUS<sC{bMMVwb!;vAe6cH&s
zvzQoN#b`Sac>vTSeDu`JOtF)QtWh1{tJe4Dmmx|tgXRfll%S1n?X{$fYx8TJen(gr
z!)GjfwAt<UpQP);6;8Yw0Trl`|8(2H5-s`r?FQ12(b<+r_Q<3prlX+y*U=F91|^gi
z7DimYsqsD`p$!ypYJPNd^af;W#wEiX*RO*konw7`XKD&bC~SNV5O=%lDG9-=1R*iC
zw5qCVkX6{C3fQU6Tpb~Bd74G$6d4u}MF8HmT#H7RkP-T_=W?T(d>LxSZQmy)soB}_
z0U&s;))FDehlu;*qXQw2yR4K~IXD7uzvvUV@E(Hu0`NC^)yM9TzZRwWd}=R9XW&=4
zSTP!ENc|xy${w(M#M&ij>Bfy4u76$;!3<&ilUZU4us)K&fo0?a=1e1fY*DH?lx8Zw
zxr6gb0%nDyZT2~A@>!jUvnf9N9(o-?P)PscsuCfFf_MW|&C`JbLtXdz&(wae{<nTJ
z(?u<*!0{vc({1LH*)9mY-ElG>1$TG%pe5jnGQgoUL#U=^$eCdX!g~AmcanSD_D1lP
zfZ`ZE7WueK{}8|{!-j;DzfrBXX|W&?#o9yzOcWfjBti%<fI{IV_JwQ%3ghF)kD}cX
z&e;~Y1W7MZNC2HHpk%q+`@oGh6li#Q9xC3<Y;NhY>|o+IB?g1Y3RYH{x#s=VyV{kt
zvl|Z{K9qs{EXn6s1oUUO0OU1K0lX2}{N~LYQ>rU*wCd2#sr>5bp*W}rXpGHg<||`W
zhd`A5zjGY5xwa33gNa|Jg)R{Y2nZmV1~9cUSTs3!6>u~QJt@X*!<3$s2&&GUkN*s<
z2RCZz<M7f<rqq$e=E+LgEaTsWh2Dzys@bwH0bC%y`8}|7Z2I1DY3L&W1;ZA0rai%f
z;N5jfI!{R^xIG@2YCzm*YMPN}p^(|fDGlO>h0*8G4$=a!CNPD$Wj9e@d?QZb*I7be
zaOinI_W#b^uF1*CN!u=x6%8`Qf2gbb%OqM)?!kixj(Oz)hZ}@-@O|eN>%eg%u`Y?B
zVrDO=G?2F&%nEivhl6AxW)S^w$CvQ?<3Y4h<G?p#k%b%#M9%QFAxukki4StrXd`&m
z{Qhc8Mj&Kt^rVxHqQhTxQG5+c&Xdj+Zs0+XCI#CKTG)H10_U~x=3;ngJE<)ww(aH7
zAd6TiOvste5&G_W{wU^7QuqZElvx-;i1`wd82yUj1FXdlYf8flS8XK?399~_nK`+!
zHwWnhXj@z192z#1JUkI-tM{KiwY<o$k^}3dL#3FmBI{oeyS!wsjr$4XxLP76a!6VA
z_4VT}-)GeTFUe;$Bv4pbhy>`1gZ%Ee4+>__i2jW?qByl)_b)UEsK6DTA|(DT>ll6l
zzSwPd+<(3=mkwk~Drj;mJ=So+UQ3Z3c-R9Ziw#68C@NmPdNn1tm7S6w6nH5Tw?ky<
z1|*)9mp8N4#PmcgTs3=;le1z}RYOBV=c6oZhAfH#QYb7)ws(Mw2G!nReT%uH6yhwt
zeF%ia=CYg<OG>kpTb2Hp>m&)t++Shdn=C*^!{rJntFKxwJ;Bm(>Q}}<aLo6-=Q?CX
z!jKvbf3-yMA%PX(Nm*{tRZ9L?i5}yvzrA1xPs7w=)OvQ*1!t9Q9~htz{Nf5Th^0r5
zA3uOGhK@~@Y>D;g5OR-PP|Hx@63i;1-%asvdF?yix^)Zkn5J>txit(5S0|t{<~GZ4
zCcyD|Tiz>3&%*&pt6m<ruqVMZvF4wO%*xKjPTN0#ftoOVBQG(VN-SDWFCpp{oP^fP
zN=V=r$ExC9og$(nCe#t%(M`v|#4ij9o{``f0*um9VUQ}&_aI*q0IW-#C}dU3P4Oeb
z!|x!~Jo8aw^0LJvP0h14TqPqic6N4~NNB6BtvzH3-T_Po7Z7lII&1PbqKjuwgH>5*
z;VNxD*TX4Y?l@mpwKLWVd<HXJcT`F<MWsHE%b1&)z-h1qt)KyDL><jBOE8$!EPQqj
zQazZ2AVkjSHbsHwz3R)9rK0HH$tmr;I8ckse6k*-4Sf|y@!g~*j`;AQ9v+CGMq+2=
zE;yZ_Ub_F#pccFp9t_u>fq>au=nq3?ULY__=0=B?xR&QH2fDRw4}oz*BvrBHpo}nm
z2IQdOEb@XLcTG&L0U!e}ij~{jL_RU_`lwgo%;)s?k^9cF*Mt&}eFWnA5s3qXS6cv3
zyw%<(1`J!)n;Gnv1#_qFLp}+X8%7FnTjM(iAZcPUTHam0dHv?iw9-;d4CCifi!02t
z@eZWmtcYlv?Bb3O6hPU-0RsPab4nu@-j{FC0aJ|*m_(JG93SZXQti}=Dte-VxyS=D
z=J|^kd@7N@e*LnkQgxlW1*1d@V5o=~7zj`XN)q<-5}&_tVeQ}u%**q81w~@c=9M)&
zAX+H`(F2uk{L-79#Z~TG&$s>lS2ogr_^@uSGwl<vApuO&G9Y|{1U|t;(f@avJem9E
zTrEt%pCCTYklOR)0GA1!#8sS|l?7bLn4F&e1`_Fa0F>*vlOtRXq>V6iEzp_+;J^#%
zXHw}Fx7eZ(VO;#ZTRE#=<^BMsQ1Cz@0BUfNQBlr4Lb9lE<HU(<4CnlH7-Bbr;VqjC
z6~ilqK#m5+CZ9fiYAaS!Q=5f(R`7Ny1pZp3w&T_DZ!ZU;RV~M5RbS4X`dse3^b^ER
z^P6)Q=eje=Dx8;$z*%y&7XZvd#07`64yOvbAFwwb2$j>|hq$O^F9_v1567r$3@|cc
zhK__;q;fYL<sVz&&~m+QNgPb7rk`C_pGptZ|DajzEzu!$BEC}n&x{$g34}9)6Cm|T
zI-79T!^I(ERUR-<Lwq6(=@FIo59cf+hv^X&71j9oiC#m}$)OoB_7Cv+ai4y<f*Co<
zd&ox+$^nJ|2ME#yZSz|6A&~zkP|FXbEZ3)6%2~jV0j-}ue}1T(8pU^&mKo+`pp=UY
zAE43c5}ig*{(DzVj&L?86z(xBA#81J!Tc8jI6mUE+#uHV1P*DmYQev=ik8+^M1LYN
zjc`IcU~RLDDTnm&TA0r&%*%g+%VkrJA@}l{{`t{!Z#o*bD)>FNEp@s3?QUi%c$mbp
z!5%?$YeUVeQ(?4>&K2#`=|>h@x5;5c!Iw<38HSgxkg*JwJKe;Cwy8dx=0PkjWVDR8
z#d{Re`3r}aekDmIIkb!T<4%ruJb_lEjSJwa<F6@2A<4j`t?*E3Iw}|9@4;#x#67^I
zI}GFx1k`vSgkV6e4$Z*dA%6Gq<IBl%C=i#pgRvbX7J&2}sz}a&hm&&(ID96Uzn%Hg
z`R{g$vf8s^F6Uyo32FUmxB3mpY6?wSHA}1+qV=4$io!vKK&aoVklqd1tN898Uq~;(
z_sh!4c2>8xwBRR$ql?KH)R4XOExi{X$j2Ab5}5c~4iyE$GQtML@9LVG&9JdxTc(TP
z`RmuOO;_-5Q_jc3_DdN!0%-SVW@g5iL2XN5%Aw?ED4a^Re*yAa(1?4hSh7z77L-U)
z!pIKG3ut5vX)-b&fT4GZYd&qDqXwk%DKMLrCm{>p{@F#|XOWrE2&1G+M<>wO1jtyA
zG3jAo06p2~hJpi?g6~E{j{lY2<>@lK$&d|HQ07<3+99%7nkxysbBOC7*LD#S*6G~1
z@K2s*oQy{i9;2~BcFRPI?{*?yaj9{&FD@v^ZE;d6GX7=7Wt?lNlSGuNGAvT@wF#v3
z(IxNOE)dk;rI(bacATv_mxla3l^ONGM6kx00FqN&BF;t@xAp(8cduMho6h=Nuj`K6
zo*;RXMm8^crPlUMYWFRgL9xlxQL5<lYiAUKR|&D4S_d?b{f`E8zUg8OeF7@D>3`Rr
zs_&U_M)4`&Q?d{<*IHnyib|eWG2!{=`i7y0bPHxxc(*&h>S=U1k&C2UVx~N><Q~8<
zO8PMVc3?moy(K^oy*i<*D_kfeMSV;64fAp&ESF?PS>2U*zpdI&gdVC|*&NU(C?%E(
zdGb(a?QN)QK}m)@F5)xZ9t>h@VckRq)HHy!%;pxvXvc@MQphj^ME>iQN4z3(<&wX-
zA=1~#(^YhH6N%QfKQ}ly2vREWDw57K-pI)KsHV3$F~fFXpNHWg(rr3^i%%9@L2};4
z(n%|<=Pw>C;UM*i6b8i1k>TOBtK(QGE>6Aj59cnjAvOj^U5NjSt?D<$j5`*4mo}Vc
z^r|4XVea|s!P>smA7Us$sdJRT9@jcu6YM}981%Hri0IwZre#~-QBu>J?Q9dYH4HW>
z>3v%HX{Rc<NG&ONS>QSB3~h^0kHTPCfltRnbe9M5xSlZfB&HPzfU)j-*c36s9APXq
zhW$R`1YVPL7fcq%<sL|kiydWXe9t^;yph>`C!{0&=Zmgb#gG1c=R_sQza~HKCpyxv
z!3`O~>`^ZJ)H66e*x*2NFJMs+Wrn+}<FH6YPE3gO2Oh9QFbzLpw}93+qcbKZCJ5Oh
z)4lL;yI&#Qyft)ZH*h`&hbPvBV%vp&sDi6JpEmYg*t=qvM-@$}|6F*pR`tY>YoJ`T
zoLDTMNV0ME9*VO6-KqNnq~GJ7Vi-`heN^QihtwHJu*v1+vA~YTCwV@T5Dq^Bd?#j-
zjCWe->-_y&`uX!4FeX1ywZg)<1zU4*&CZ4u{oM26f}ctej^0JZiP(}0+xb(JjoP`7
z(0Rh;b8%rPch)emPL%d@I*2A~>`Do{dRTUg6JHrpU%GN-V$C(?A3tXlJHxlbs*G79
z`Z3c~tRaOPFvZxNM0_S@5R5-OJe+iwa}&SayT-1uu=|IFxaz5vF#}7JYlaCm`FR2T
z3nJ&pLao)DEQh4{Z>f&7eNs7?STm=~JyIdT{@JK9C^-C>sOx|Yd4o;qtJFGZv+GOz
zyW3XW`4<mnuC*&W81Ex@L93zq&G~LOP}LkaZq&^?t`EJ)|7M6K%uxQ+wJ2eAO)OIk
zEGlHs`E5UKQ%%fFx-!j#*6lGv<=5GjJ@O}Ez`D0!aPsqq*8v!Uz-{ZHuU5M<Y-L*W
zQ%);GY&f67qMf4|%_0*9lI}VSw{F=@{8NOy&WoDI`~B7SwP&abY;ck=I{O?`9`M6u
z1(_lN;hHa`#`AtPUVKp=;QgltpSx<aORzd%dCe8sIFpj9R%x|M{r;c#<wCYQ75pp0
z3sO>E!F*WIg9|p=AU=o9JGzg5cZIb_lh-T@Q&Uq1ok0SiL!HD02<cypDtAu-sX-2e
z2iOmS+qG&eQloGCp4DBeN!Nw@LX?o$M1vS1;I$qR&QOUXgjW-3R=V<mJ(Xwj`cPWB
z9UX_$DDgVjwu1!+z<I=pd=hbB<N2|RLe?q<3r*r-k-<u9(b%5Ti4F*h->iKq#h=j#
zyfYB(Mz+Dxy96<pLi0`(WM!q}esn{Hw0tVa6QIp~JX)()VG{cm;X`OQ<Yva#T=AOQ
z+f88&P&h%gZrqkGx+~TudpP#?3uJLFh>FJ)Bo?^LHP|eP0qQJ?!>k*B(Sgtf+1r31
zTYR<V6hvMayGc0^E~Wc_v(FK#9XhsdEC2JUbNBIKiS2H_ewBi%YU8})#E#I|b>K0W
zUgwsd<LmR(8HLyk)`}3%YhSkNUIn7UROm}Xx$&2{;iU|i+<}apn3!mW%_FD{$Twk#
zT5((=LDM7HBdGd@^fEK3@edzAJ}@v)<E3UrRzDaAHd%@@6J{0UvvuG{EG)iorsErX
zs|QE*O+sUO*0H7)G@-&#<$IozFeuK-%7SzYiG^XmXLV_Zxkc)WRhCu!S%&9V`obsQ
z!Yb##a`F0&8%E=GZ#xe(kTsGCL}(P{Iuw<QB&Z@ALaDMri1Kvm)6(_cn?@uF>@mHT
z^x9vA^f+E!P2ziq(n&tUi_sq<?H<EcC2P~s2FO)`b|{0@z};Fh@86Bz167{WJ*Q(&
zwQ|hDd(}vD0jlHIFJ+Lpeuhs=hQGkG!opH48z|<b^_;CZr&Dm0klurFZT|VZGCT`_
z2Z9%2dNv6OhLVGd+)}q&{}g9B<mZsQz&IG*5#}Edq^o8|Zf<TC=g29m58oda9uw7t
z4d&*SPDw%nSv2d+e|DV}h+oiR#bYzR>4vu}Sg{CRmp0ObOcU8p5r#Pd9z-Z1bV8Kp
z-@3v&BC<8t-+1kk1$>8Zckn5(``b&1V;U~8A)?LPtb*EuE95SlqJ(Y4vAn)b3K9Jt
zK_g!RRzGB|bZem>1O1CP=5ylZN5({2gY3CL-wd}!5TlwRhvY0_DU`#RHN#>ZcGLJF
zlbd76fW2Ttz*l7U1Zx3)By^3nzC@5zrh#vPSrP<9P{S-RL<RIkwnWJw=7i-K*c*h6
znWAB_P+4g!mZ&Tu$~1kCrP~naBAa~%UR#$zEj5hWk~1;=v)c(9KEl%<)OX>7VUNNE
z(8&-oYN-8<<AW7-Ev*(9A!$~*--c)fR*H!rP_>S-CEpW-g8=@gHo*+Wzy>Ffl5&FR
zM>5G>WQExdaz{3`L}h3s8YypaSZshr-)`8w1^QA{QGqS2b#O$#F=lG8D6_lX%41ii
zyR^-19M1B^s-AQ-a7+~{AK7MsUt{{Q<Gje212!$-%MoODE5gtLT+###BxXLSt3m`V
z8+~UEu7a*tdaC~Zu({a3%j&+~|I=7}J5_GaibCaPF#i7)_y6O<|K3TyA1ck>ch4RE
Q#R8t9tSUNB=IQJI0W!*$Z~y=R

diff --git a/_images/sphx_glr_plot_survival_analysis_002.png b/_images/sphx_glr_plot_survival_analysis_002.png
index b9a5908767caaa974b7984dd568ffc200a0bc514..be49c0e43fb3e25b1012003cfd1a14a4d1a40e4f 100644
GIT binary patch
literal 25194
zcmbTebyQYu_caP4jg)kQg3<_5Qi6n(h;)Y_-Q5T`q9P(7AW{O-Al;3Hbc1wvcXQVL
zJnwmb-x+6|G0yk>!9b9`@4c^mt+nQwbFTg7=@YpdI8-<&C@44NA4@+&LAj=lf`ZC{
zg#o{z7we3LZ^EuJ+OBF2FI+uLoXt^`O<Wyq9b9d#Oc~tGon5RP?D@C^x%fF4EL~k4
zT^?|A+x_<sxE!1<xObA1b>Jk}j*oR*P*4a=kYA|T5?NL#C?>=5(vs?)i5nAM`X2``
zn|8W+O4^y25?H@122p1Q*=*`%#t$<)2&%T%$mXcVY6|XMKGW-b)|bHfc|xzkD(nS|
zeazx}YU><C2FEuBhO~~8i=9@i5}_tOJiMu+wSzk2WMP*sFAArIM3>~Pdv`NPq_7wf
zgi?a|6`O&C$bZ*IqA33JD+^*{Y~+6}D_!G2e#c6Z4`fDuPvyfdLVjm-WkHSo=a{bB
zO~lA!d*ZKa|8u?v(f@lVakm_=azY>wBQ?}9rb}H(t}A`$9v&We7-%~?J9WpiVH9{6
zdLGL9l~(;M@XG`NJGMQWI4*;?&BoIa5|=&oMpD|*@Y(7&F@l@Z(naY_hhjf`XbR(Q
z$UY05AI#ID6n4J7)Sa@_DL7?zvb%5{6O&xey(>{TU$<Og_|U<8p)CfF>OPi{kx^n|
zA{Gvg=|Hxs?MPATvol6Swr{!Z1S%uKJeJ$YaCt|L-tPx|ruMvV7#XK_i5WIabPjK=
zRxSc@{YB?zneJpUH(DJL|8pUEd3g?<5=1DScdPH^$x@2N>HbPcTUR_^DA6rh9@`O~
ziBax1&)#C+WGQl8c{W;ZBd4T<hl+}N`_7$G=LH4$>+jr+=TdN+`IF>~2)oCD%+mZj
z5dr7vJ#$6@7cuk!=f$ljpKo(=-av71*pgbZ8O)U_(5*Pz+6Wh%G+yb;P)d3b<F&U)
z`uUshX7!SGgKxF+*&RfYcsvtH3;skM#v0s5lQGZs>004UYHG!F*>Fb}m-K+E%X7Hg
zyBWugS7#FTwHpHZ1_sqj$v#2>m%fqSjaQdGH6B~<YimX5=;-k1L_a+W#g7?{%TSEt
zb)J8W;(f5Hv97)-vQ_t}J(knS-JRv`-Pg;@mRyE)p6xd4z3G@Jx8&acFy`qgwNMFU
zvMGWGs-vUBp;yr{ep%%-r;slqC>Td>==t(+V{C3=Ap|~XqSnJ=?pJumA@=m{LVHJ&
zNZev)!X8{Uc`e!cOH4WjCZ?A2;~#HszM~N)rl+SLA0J;H$hn!PRVW}4aJ+2-H$W-k
zZky`ue|ci@PWD!}=$=xJAC;gzO}WkR;{2cJalfO9fo}#0rw41W61Yt-(J+@MY6Ukp
zH({&nHC|m5KBlc)ecsS`c^YQ&{ngCo=KGWsdKJ-~>ssIRSu2djDjh7Qn_hZ)dCe>?
z%0|$M9WTTgDb$lO=H6gNWXGnZr)$VJwy3JAvIz-|oD1E%r;5rZDykn65|RTCC}$}p
zz))CNIOj>ygB*(#KTWrlKAHN9!!hgK`9E#OJlQ*QzsYm=(j&y5VFXLwpb>h?sa^E2
zez%RC?e=YHFCU-0A3uH&pLv}hJMnw%nxSVS&%tTbppX2@`1<v3#>daXE}7TVryBeQ
zgV1pYo=1pfzLQN~9Vt$4>+Iw#o-EL-%taoAiu^=th!E?E8n=+uj!NGX=Y{q4e0bpY
z_5o`{1<XT!iyf4t6cpK}rltdFGGPOGg$=o%`E4?3XlVy0>wOe`lb$KS>7JmjPF-Dk
z<SgtWFCZWwU}0t^HQ8rVnBQSibg;rsZ((}+OT+0ZyG!-3p^q}^>RMr=rUGkBe0&Z(
zm;9EN7WF4jp5)j`%gC^jlarH@k!4<<bO+>)jA##LD3YUR6G)stMMp=^vFgvvLB}P_
zj^)&2OYhBok|?CQDE`92A}>=pfens4*zZ@#JK0;xgYA?jU^jaAq1|{@uf6~m*AwKD
zcO{AB2~K)Gsy|sw$W=}dX#4X=-fpU4LVxo+frR?QoU31m?6;8L(mu1P1Zor-in;B~
z%Iuv`^1ZnGu-tA;)1cma(AU4?(>*xw&v~=oN+0vW#zq0llSJ+8q>mp-78VvX=FdCD
zug>?F=!@V6V`DQR$EaUjUT7E^CZC+Qgy52X5kH;|ezrDI<&=JH0J#~K&pT0LpD9$=
z*5+krXFpqO_t+TacY`b?wZx@g_2vBHVi19SQ^9XT&!Ij8r5w^l*y!g6Zn@V6R3t9U
z#693DNU5lDrKF@fB;xAoL<bA>1Q)ip3cHfU@{pUcK(~AV?%d(8th)NgYXgbGE;t4e
zd3kxf6nLxsDiZuD63+9luMKp>^R<uvhQ(ul=y&+nV$m4_HCH~GC1f%n`p;MU#tV0T
z<0cey-uMpl($Z25OUvxl;lh-Y#OvB=HBqcmjMg6_Z^_LrESUS#P*P%BuGQjv!+!hj
z-E9bPwX7$o+j9{T&aj-n-cf%4TWZl6+G}~6m$xr1r*t;os4;nbTtD+k(%azg;aBfr
zJzr;KSrn^1jcK;LKRZ8vJ@Aj&?D@%_eyw{rKJ9}di|*uSMhTITsDYuOp)z&_I;F2(
zzs9O@TYDj{RqJj?=YMQ+c|1ozOiYa8=;(;dj%F`j@LKjTOnNQRzUcTQ8H!KK(eN!u
z@S3NOPk+7+-}uA?CBOB@6W`Joo!f_pi;!K+H|JodW8vYM{di5Z=xlFqFM7WHJ>-;#
zp1yml@u~~C;bB!RE{{IO#K@$G`;C;CE5_8wKYG3UK_gabfZ+zU0LQ+N#E5mz7s=u!
zZv85hCO8iN;nrk*Six|Af2sb3yn@2DGJ=C3*dvw$*^wt+dfFAXln!+}7;^7vW>yFD
z7WYz9Q!^U4t}LA3DTj^10lb(s=;fBZE$6;@ngu9OOYd<v95eN2DyM7YX`ZhcU6o|t
zK!H$zqqUY3<KoCGC*1WS-AoM)4Hu6ZIQ*0p6w=_S%B<F&Ty$O?k&u#Vm<IIs_vff(
zy+_w#ksiDag`WR)uO&ynR|ZL-*L@RY4ZVJygw?{q<A>k*&hK~Bg6|<O-m@FMe|~<>
zLlut9(R05e;!Z~)Q=^TxIN5wfXjj_PqGYM0ERI*Z{A{5A$-3p^aO31E+0bW$zvcV)
zFitgiP<iFt(&j~lSEM0?n!Qy{&r#rf%aC5<jf|dY9-M#uA{RlIp_(qcxK~|WeVc_P
zHm{)MFjdINcLzP3ihmZe5!-%EJfFqnZaJi|bjYZ_yMOM8o~@TMWAK~=1qTaz?^{Oh
z!D8N%NO9Pi{ds;o7txU{mQd}owA*?sV*6m&NY7hm;e{1_F#NISHiJTrltg8MQNq>M
zpZ<Qe;Un$Wxza;`jqI(O0#QA^y?HIV5~cdhPXJD+t*%W>OrXB~Y6zFi4f(4Ml6Z)$
z(qzDuUtfPerKqRlPZ6Up1eFRGAGjs%vW;YIj8&G}jaBG140U#fTO@nE`W{NK*TL6K
zB0{{s^P9f(`EPub&6+hX_w^Cg?9Kdqj!C~G0|;v@TwFAUd-ozC`|VUVUW7yPjvY1%
zNOEX6c@-gk+L~sQi=}wxIc(tJbh>ZVa6F6aQoX)z3r9~#9h5svCI1SiMi$WM_;{VK
z^?DaR#Ay#c5OFa6;PmqLF0t%oqVt?a-K?6!->lsdhkf(=PxKv~dT)M%T6goAmJsF!
zaZyp4^xjtX!a6B$?+0C7UHh98)DjXB5K7JYI;9jsjtqbc7M6R{>-M@tT;cMEZ|B0@
zT3nnSbR-ByeEGs4_2^NUf!1}$qY00_K`kSvt@<Ng&uwGkn>TgpJnxy9nEb@JE(lnU
z3PN)h326WVyrZ}m6c~u|?AbHza%)l%_jO|^N+?Px;z=svCkV-4jC)W3eSLkALG$L#
z8{d<KI3XdSPki5NYif|Ug+k{wPt<U}tC*=2F9qcg8waOv!@`YE@f2&c6C0PD%WQq5
zczv=yp{eO1;Jx<=33{ogG&D4xaB&p(EeSdGD$LH0cLoZLQXqpBJI%?8iiw>ZjyV*#
zi%F^DP)|1BMPvu=ezo%4C?Cd&VQ1}>ye!xe%t9G~+)l+~9Jsr?`;3c=E7xi6mo<Q1
z>)~$(mFrtk%!=v&zdk~UMU|Mq0_OrOXzT4&JvlqegV4-@D;qEAD*0+91c|u@Rx}2(
z^)&_Mzs9`r$n>r5=vd_J?yk8pTCO$)TO=}Nt-yEZCwVTQxwhZG<py&#6wxs-a(mO|
z@}SC*vUwoEg6R(|WU<oM*H@R8Mx4U4?&x01@;}~f=f<ty1YFbxJAT4#*dXKZ78@I~
z9FYr8rOOcvmYBCQEkS$?PB#a)5B&B8tZTj8lL}x$`Q-F8msKq@8;})QLgsTwZPv%z
zGr9H?HEn<Y>Y!s{=0Wu4R8<MNP|Ei+<gy1p)71ZS7uis#qWz)zKWH_i(-&Y7g#1qL
z10FRW^?!tHUj~Dpy;Y3kA|fF%dy=U~Mr<7OeWgFks^`ngg4%z&8TPZiZi(uRim|ry
z2iQez3<3hN04=^5)M8jlxT>hAP#-BE&*R<R+WG~ym8YlY$F+l~KlL=cyu91HyG&wY
z2LJR$E%+MyrG{kHagw&S9I$6$r}qvS5J4LxP0Dghdx7EBKP{18-m*n~elzjV^S0ox
z1s~#Z$}bqJaD^xi0>Y7Z{_}C52F6L)_}=cW?L<s2_YAf68J<~UgHsnD7DmH0lSc<W
z@OSDrS)$)SRn+r%q`$Ph3<<ejV%q#|5slf~Z8GA9m)*av310QrB=gx0$jP8(?$*g-
zG3Fu<fRMTuuT-)5_HNVISVCZ6ps9;)8;J>X6cGpRr`<BuY^ImQf?a-lLi!3A-w%8^
zK4J@WaV_v;Xp7K(CWu{;Vw5gV|HvdPOqt&MI2hwbTTI!rWkpsQcXt6RD=U=i7#ORQ
zuihEQ1j858&K+&{O1IVbr}$E}dRg|2h@vRK;m+$L_kx0gVC72OJfu;f$GgSJ84eJQ
zES~El5r=BIS*EV$H-F6Snq*4B3vQuu#m1Tt!Ze&eDF=?u&LvJNKRN-_qi|vk`H00r
zr)N}t!O1>+^piSWHTJ!?j&3Eaj%104-Qe@iBWaeP_J_pbx(roe?o0H{7wP<g7Jdc#
zGKRNcm#Le;UXXNky^oEJy*^&uKV*<oL-S5o7yXN6lk3G!Qyeb0Cinj5gqUqAc>Dr<
z_TxgYZcxA2dyvbEU6lJlG}?4*$HkX6;H};7?nsN=%p-cljln)P&q-IUR~AqItdQ**
z&CT1N*dPegFp3kpUsFj98EX*P5U}vAja`rGcj;KVR<Hj_`1ZS(1^bPD1FJ2stgT4f
zqGH5s@2;Pby#MxU`z&>X;?F!6f^cNULiz^mA9LdZer@C#R_4S){GCvq^|^f7-=Q0#
zQyMC-SC}&%AS<b)OZ6+i{Z!#f?QGOEd!Z2apU8{$?$>|=f`$t8=2lk~pc<|~Q33=f
z)3z~Q9SwcjP8+*mtWVv3pW^Y!$;`}5^CJSWANwo)*aQUJ#RbV?KD$sG`P^33R|j%}
z8yh7QV>uYv*~6e4nVXrBjAl_jzC7D1aLTEPM#;m=FUDZ+{`%>_PRhzkI3B)rQsOjt
zK08~T=&87XJ<hJG5SirDnKt*VOXG#VOZGtbYw||<Y4|B@m0fYQeSY2yi30)F3i%%M
zK-xEjv>~%PQe=V#CFA+w`d?8fj40<vjaTOILmhNK$UF@X2l9HFUPZO-UtV87zrV%L
zF`=y$f~zx~Z;b>dAw<mQu%|H~09nyWN=rE#@bw~DrKrauJsVn$n9*4ogvCN>-z=?+
zSJ^MPb&xstF)pPY)_b5OxXWNM<}d6HinbUYOUBiZe29vCv_?Lxqb@!(h34o|vYsv>
zEF5n${Oz9au_GW0R7+UnR9<$iZ^$B2cQonqj!taU*S|RkT@>*BJ9-JaKxAL*wl)+A
zW!E|Z3jYFh+L*?0<8|k|?Z>C5yiT*yE@Hqem_mc#x;pSDZLNruvbm(VOpzH=Yk^g(
zz@ScCY<U%}i>_iLj8JO85l)h!k|JJaKk*sBc8S>^V(1Wz{r%4lH=rAiT4;|W$N7fx
zo=!|bT3R~QMNoq42`p~M6m_`lDr}9r*?ULByNrttPHHiKuwuhcysDOUzv)J=NHR*_
zm|nHK4fo2yR(oqlbWXhPpyXr5oCq<|>T}hxSYE1dBzCf$sEL6)DmHGuhHURD9H#Ie
zLZQyltK_2h-9`<@B>2_a+dEsM=<Gs^5uDmneg5Oiph>?i@}^!trH5!xP%M`J5*H@Z
zC^Tqjg*sHOKSG9UlWApCxS(l21tEI}BN$aMOP7px!lBTj*FySeY)wbqNQW<u8V{q!
zeWMNF#z?IP5^m+|R}-TEi6dY;(p3EXHwC{niB_Q@f8#fu(ibvA?89X=Q8m}K`7F9_
zv$D!Z<mTSD8JDqi<>7I8tX<r_a7EQYL;R{Y6~ngpLjdMXC+bj5K(lSY@r=&M&bPw@
zW{KmO>b9+P$RTM#DMZX)?JGtten3|VXcNEIePb5trSHL@7F6R`WSqL6+99ZlO}=Bi
zrx9-XbnkhU`vw&Xa%TeK7YAl!?r17tDI?Ah`Jnl&YZ(l#<rg+yeg<6DpQS>73ESLh
zp^a25x4=0+=8vCo-b1pBIecFm@=2`wP3<8vS5@hY^Fr00Vvo5gj0?`lZT+YRI=?sL
zRB$3xerX#vvZ{<7Z8wpRYN;89)N&vQhhncNiKaC?b6%HA+ycz<x5OORW5SIbzyNX&
zE+O$1bR{8YM1)NcQ$mi@C~$!iUS3|~0aq8$-v(zTd%qRgtP1||LmDnJG)(b^_NH@^
zWXwWx=d<;R+Bj&)KOjLE1cVbnkb5RA*K5Zsxo_P<<L2fD^t-daKTt4v5ny1p++wut
znDeuiz?Fe%AnV0D{e`pcqz`ojzkG2zF|7InOE1E(zuEK7KC<Mm+vmleY8tg*nA;C`
zNtPTX6N_Rt#4?eA)*Zk<F%^|=D$1?yuOpG3$z_Vd$6i(HwcW4%FHVUF37dc_vYQ|+
z?VC(^>p!$sfsqH`6$!`_#r;0Ry`><@94PpGnab3Y-fNt2i3iqg_QF)<mEWslBfG-a
zH1muSarhRRtLH|KbS%n$i^PpTsJeD+Z@rSP3Rh+E92Zjyh4Rnn3Bo~PGy0t;He)VU
zdhkm)(pkmxg6)bjGAXIE&TCJ|b(s(v4wJ=>cw}Fjoqbo&qo8lupFH*1Td0ju2?=4(
z-ac?!EpnWZx;!WhC^q{O#Z>A1mfBuPdHBQEv=gZ9TPd%`*G^O$yzm4%Q5$E6Gf5+(
z@r}f<qZTxk<5zjw7#Y+z|5eMY{c+z9%C^DSP?qtKr`~(d_}#XOpC4e)-oprf$!{~n
z<8x^1?Cfkgluv2of5Pl{c3AE<%oc8sI<z@a+o^^CB>#?91c`+{-?zGEDgT>AC0U>O
zTE4{9c{C=SXJA>`ee1#8pN*FX+_LX!tk&#jJ_MXuDZee;m5Z~-59ED9PV2<hnOazu
z-a%rP7?iYS(;r2L>bQSKeCq1=qy2M&O1pX{VGDj0;Bo&8Z#|FinXcGHx|Q~wkiC|t
z8YPh7@#qy^vn9tl`4Pm;CtmZnbe$U7Kv{f}2oQ%d_4Vu5o(4Z3WLpGHy4`5uIVwj|
z>`x89Oi|8HJ<oq|*-{7%zG^2gv)zv4ZdwTY;D({3<*;KqeP%U^YV3G7Rb{xl;C(OI
z>{7%BVV}Sq@8sCG@5Ey$Y&E5f??ArMf*b*0+_VLO4LHB?2X8||g9ghzZDV7GmoG;R
z$bqAUOznJm?lp#~J$x&lirCQV<zpS$9EDjOdRgD6z6{aFH#ut0hTN*otR&-CRttvx
z%#&Vv_>hQewQJy(IF~*o)y`X3#<@`|WlYUU=(1Bs^pRAQW@6?nDyZ{WCT|ElF8LY`
z=WZh#x6<}F?!B#mtJ;nOyT2tNjE`Osl8_*pjg0}(%e4LY^$Qq{ps+C9$jC@2I_aVY
z7Dt;C^9u{TUtVq;HPl9?R=h%J$LGYx;No#F$fM!;R?mOARTU=iQ&ab|c=ze{JdvmP
z4UF#6e3h=`>1Fr1o$_`^=5$9#bjJf;8Y2RU?%*n&qEq=}YbIZ3dJfwB3E>lW5zlm1
zeKyXKI`NqJvcZx|&<ToIA%8)ZC9Qq=9AF`ECTuQ#fHlhO#$Ydvn5GyNo1)*lcdsK+
zI2u`xfdIt7qj+lRkhZ<;s9SEWF%JO#SLg_EO~BN0Y&D$O$Ddi=IC-gIpZw+@sI~P{
z(=MRB?I1v2@#nN}zg<c4frFUcqVd2GH!5@HN;^Rh4NX&Jj=5jWR=~DC@A(>yl8FDB
z&AnyU@82>ca{z%P^qIeX%Bo@OJ_NgdW=`#p(-7kZmHO>S9m&1L&gZi~-Ygo_kaHV$
zK)-u39Zc{^-0uMj2oM<+$V?1L@fTntAdcEO+sTxYKSxHI%*1HhZ#0`-mah_`2x}?7
z)7fQzX~+dt<VE8f**&>9-Y8ao>y9sCDUI)0*OPY(9B`;}(M^AH(FIgH25QENqf*#N
zj?U;|idTMizfTeN-k<b$-xqc0JCrr<Hsz{i;X~yMZhi<%4lcS5q++0~cDn;EOIwmS
z?gEDfv^d}MUwC`Mn3$NSha2q`c4Hj5k63>Ki;iTk2}HM>INQ6EL@M=fImPPw@Wjj0
zZ?Wz=B}XNH{~B>L;vjHz-kZ_=D=nOb^c;Kqe$7WC@iL#2QMkwgqZ*As%^`8gzp^oj
zq6|mRdiN?jj@;Cc@sXyG*HiZssJD1Q2Ra2-5Iue-I5b#^lflY$OI+#C-Z%0}67z`#
z+VT7X_lkLIYfD8z;jN4tARO*|?cyLHrHJV0o_Kl+VdB%=Ek1uo$@f@AWeJFvEZ9Md
z12?@}`B-20B|IG!v*|b$X!g&-g|!rQz+n&z-fy(wl6Yt%CUvIUx2&aerXH)BF&o7G
zWnYh40c#yW*z(r<aQ)9fjv6W!Hns@}8H-E66)X=I(gT+d%$}b56sSUvJw2;Vc3_z(
z1Z+RqUldrg-4BXMiR8b3|09<{&GhnewqjO3kl6O-HR-*7|5935SS0)I$|JvAPJPAJ
zAB(Cudp~*8&Hg{w@f(+&rf}?gE{#EP9o~XvO>+A!%ro(7w=sJ+#<TJ8-#?YjaTu-X
zs96{%t6eNoMK-2<_mrcequ)eCa33wr0FP@vZ1tv%*4W(4?9<Vr)$ljkIBugws9`^2
z2VEM)0o8zTk&_dWW&TV}4I2xqryyw-8U7&L7$*DcWqvha6Ru)YskY8z>JJz)wMyZL
z*Q`5JZf!5;XyT<v?X=vsK|L6I<fuhiO`Mil>O?#X9RR7fFYo}Pl@8q-l~abB4-TlZ
zvvb?TWVExM(C!)dOioV1Qf9SA1{kUUyeb3U2pac2;pjio`ua3UA|6BK1qm{8a?v2R
z0rBaU*WdlUBwT`%^N#J0C@UGoEaK8W<0~BPE2eyZ9Dzn0D>>J#62i34ceng=AFv6@
zP!JQh#2NYj0Zk2=(tr;^(XaE21j+0*T-?mciagK<e=8k~-ndPQ%8Y<u1l6bm2<q^N
z2m(mMNNF&{&?Qd*P?%hyVg2^@c2j?UH2hR*Gt5=*k7xdDlPDip*H2r7aPo9}+H0*H
z9rGi_&li6KJM`oC@430TN8exFfFBsy*xmw>Oh`o3414=eTO03`h9&0^Me|#%1IC{A
zkdOO`Q7my-+kNiJ<Qn4znWUWY3j0_H|5An-o8t>PAs-Tf#(dfX-G}Wyse_K&oL5HS
zQdrb#INx6AX%2m~X!g(Tewdyd8}2GnLppw3q3^yq{s%4uP|<ur<z&m?AZZ|yEMHx1
z18fojB<&T*9n1Y$1dSJm2|p>(LEn%Cfe@)QjaAqo30EX`DKwG@8?Cy!JVS06Xd7BA
z?}Gs5P`aSFfM4XjzpT1`<ULYuLtgKFz$@|o%^TD*ek0&T%=edjkvJ9Ej(|u3?5hrN
zuaO~J-eK?Gvpjg9JIt>wi;2b;ins?&r%ZDD6)E(r`3hT~xp3~<>GnqD{E!?DJsW!$
zl!#|%^y3lx%!s#;3xxpfkU-0!13G7@Wc$q1%3`-dd92jJyuUnm9%r?E6eu|;3Hei!
zB6i~m=~($!!+LKa5fNxjtqni#0U3kjI{W(i9Npb9rXd$tK!HOtSV)SUY;XcV+KaA4
z>?YW!T_Bmf2@6X%YV;>7ciNqo2O2>ZGI*!kh(a}5{>9mmo_+=3TJzjmPz5UWj}J$l
zN9@2?OwMGk^?#7G;qfV<Rh<3&7*5fVs5ZZ<j<a*Mr_@;QsM84(jX^rP-fzK&tCAZd
z+r-z)dSxqgCvnA7WBlc%u!<d~#CiGnpugGf)JBHr@y;CCu!~C_Aq9n{;d?WAYAAD}
z`#q01j2aRf{LYY~D3xGECw(eB5m@MZ6nK?*2t177Z)(;RX5p+YzG!;reKie&21b7a
zgHbQfp59@S@-v%nql;PkCa>jybye$bmNFbv5Ny=^M)^aDk;v=|$)_TmJvh`DZb9rT
zK?p+g+2zde+d52E?B^ao3fd>GLd?B(aqM?8(mOhouA!m*EwfUKT;fokLm`Rp;-kP@
zti9y4n)^v^G!MW+1Q6)r*;b<kpi|$|m8@}}t$MN}Np84Wpp6dJ<;soWB7>|c%Z)*b
z9C1*<NX=<Ah_)B+*`yzhbi;q|iy7Jw^n=sUyU_joILIB-w0&k&J#E;o#@z@@6%`|R
zq2D#!N3Wt*dTvsne(d<Y$R-0ba|qOYB4T3Geu2*db_!77yk_6rq2POgM-M{H0Zvg+
zu+5~=?;)DzAZnX$>%!cxd*XczEa@|7bu4DK1xH&o`?wvgE8X*b@a@BT$bN@TxcKea
z08RBYd4k#Yi^Eid>*jboC=xrXd4+v&`%hBDJE5#sLsP!elNvlYX8E^-89w$oC^?8%
zuXceuFLqgaa(P(UxZ)UMp4D!MI00lO<h6SnbcT>F0$OW&Y~lj3pqQ&#rkR+#<Q_)L
zJdz?@8P?51sWGD@-;G4do^QBXj+~s{pB0-nnHYDH!lM0XzL6OY-i4ujo$oH)V!vUl
z2)isIMgDQGrKFmPK=(WW$m+m`JqPgiH8XQ|ZLNKK=6gZSph3=s#sj>#L?Ndjcp!0T
znD`{RaWRd8slXz(#nBsQWM!>=mOkjnV0TTdrwf!Yq|4W<PkYt=xPizviKEIWOH$Ev
z)cF&c?cial20=)X&3`+lw_%Ft&Qzm3(Y5t#8}&KM7JeazOdW#eZ6rX*%PT5=2PFci
z`(pEUvInlqPuETFtJdLUvJj<tCch*+!BW!3Rk__1SUviY5hu$36~9U|b*J%P!jXfQ
z0)xPPdvi>vXsBb?C=uC@RfI$8onBiD3l2uZ$H#YF8wv-Ng%=q5!_A32IX(d`9h;(u
zdZFeYUMntl=83!dtBY_Y4-kK*)6FRuR2WaA%}eRK+_M*^(Z%Z}8;p~_Oe2$eZMs5i
z45}(L0;rZg=*iH`0R$nUq3Hl_`@k|yq3Dn*KL|DE8E+bMen%~%tCzkA*Ewpard6}1
zb6xD^%nz31O>E9bu6Z-M*4V?>^o;&`I@|}FM;o#aDYyO!F^hHM1~V^jl$Mbn>76@w
zkdg}(zx7wSoBUF#L>vhRQUj0Oq`r))ZwLiQ;g;bSAU>iwiarr(NWuyCvBI6d*OV(y
zauWBsQBe2tG7Q+=M{vbrAOM@0o1<<+$R*TnH6%d0rB-Qgpqa0oiac>M5&MQLDrF0?
zz;r)WjYJNuMTBIFf>0H4AAa-JT3NdEiBG!nuxy=K&C`ERc!Pqc>D&)!E>Hxpz}u1q
zQkQ%A`_@yNA{-?m-3)Zbw$@7QababVE}Fqiq9d-*eF;Uc$e?#>qoet|ENY}J`10?+
z=iZS}tm9w)jG4!v=+X7<2^%{BnWaq;BKz4Ye{gcQWrAYB$#g-xMW$|}ij+Fq`}gl*
zJA=~o2nx!&^>k{@Glj^6gm$1nKviu5?fggpmJu3E?IL4TYC(H&rxZMvuCnN+6Z>53
z6OxH#j;E-uQO>DEWoaeMxSrT+IbShmKL^amEN~-5NNyNp)%&(1gqWBHC7PsZih&0J
z#Y^lb9>~6<kOY-NdTBvjNIAH(hEaMYs8+i|@G~=YH;(Efq<v$|p)m>4*=P;>>;s^x
zp=Hb`D_tfw2Gz?0x|3H>F;`WS2-Dcu-5ko+a+@YCrPTRl(}~hm{kg4i41WCp`P!i?
zn7M6hQ54)nI<V$m7-W0nVo_<5SgLdI80&X_Q6qWC+sw=>ss=HnfWEY~DQHF9%M>m`
z_Kf8b%(+!XsXPY^(tel$KJ?y7!(H@KSaIB=Wj0b+CN>Zj?#n$d6BB7*Lm?6ps9`}v
zk|hEz>*zX15)b&Qr;ZI}i=@jxGzwA${`qQG@3P2adlljSI~AzDr_h%^|MhMLAQ=er
znBb|<nKyhAg{i1+gOA_yoXYnes(*OMljx?5b&(!iMub=obJJ*z8#};-bYKa+)wQ0J
zcFuE2^WPM}^EG-^+SvTwW2P(1cH$QFZ`cMy#$e4zgm!aHHKyuG3%e?(Fa`W~`Bo(-
zf<4Z5j2`}q2)aA~Z1Hzee50VkI&F>%gC2rItI_#0j2sdx8boE8_264(#n_P=#Xt}|
z+|vdfv&XJ;83iV7{jCv%MYxHP@m*itA7db4*pM}YhXPM?thTTYbg?Su1sr(z5GB32
z8tm>HqcXs!a(eq+G9B&jyNZe(=sl0}>6GqYYQ!|2!oApWcB%Wxj>t|s^uuANN3E>k
zF_#b;u5gQh7Db8eRHv|=DnFcBa8*?~;~rIbRdqG0BeZcJpJ|wOMvqf7tjp#R6`ha@
z1zdCYLov8BRwD<Ingg2`P*bWL{|mhe+nMrV!{^XW|HqL99iV2m639G-^@miftgOi1
z7Mv%57)_k0p>O;O()2kn;&q3kHb4-pdf!-J^eRG3@C*fQsx%Krv~WATG@_W-A2_^&
zx5=b48A9k@IF34Wd2wbnm`e%?a($|?aiGL*w2W14;y=AJxWjmSkGGk+3!!oP^oatR
zEG09u+j~ph7N9=Uh<THM=86KaqS}2!H*)x!UgdjGR_p4KsF^+|CkKySJQ3_>(_0Q>
zomR}zwMFc9vds~EjAwY~M-GBrJ{VSd8;>hI7!kDWM^qfpM?$x>y|?$hxfvSIr|b9i
zWQL6?@m?|A(guP7w5`YxT=Humw^{alK}C|8z*Ta)P=n%px8CRI1r*bt;sLPA-}I|Z
z{SzZ2K{>nOY7}hF%St6yixDVxwHg{x8*T4~tGa^JtLW}7zF7-+{O<H|5C@}ljCm=)
zHX!kVJgx3?Kl4AGo#wFkKs|o<?P_j*33<81ZcHF-^|7+@&(6*p1Ox;?*&%sMNP=v|
z=in-c=eN-eIrC?Sm>9XCMwidY8eMx0ugLwueR_`m5Gkx$_09V?nwy&!Sqr{?y$!*x
z0$D3H9r(hUEfvqhbsgA+Z<Urnbd8RRGF7DoS<`iEG70+XzIq=Ix`1;hAgT;Zk?x-l
z$;tNOm*;!kI0ld8ag_f_xNP!$nyo?x3aIFmS!K?!)k9r%<|p&CxS`MMTua%7n%xiS
zi7hUZk%=kj_}DY7T?7#D`Z{t`6&DxBIE(QHGUt|esJFLWoW2)u9b+JV7ZodwWg-Fl
za(B;Q0u(Eks@mGMUPOFhA-8OL<iJ47)yu5ROkzU9FE)A=w>^m7S{GD=2)+XhOZ$Y5
z(EHB)m2_G+sr*qf2;T}EbnO$5V@f;>B!df<os(+l(hs+$RKR7(HuOC~NLaWRsI}SI
z*$&((WtN*xPEMkLo@*-HO2NVh&Y@1+sr+W+DsgY-otW&dpjt~0+!wdi&Lv);Re3*(
zZgBLyWtI|>o>z`mrnIemDm5og*momwoUM2dg3E~%r)W1AFM|YhTA=?vr?@yfr-6#D
zlqgi(aq!~B3#3T6HwXOk?A#piE~$Xuz}QAeN-6^mfY+~I8$&-2+2Nl5`SDx*E>4YK
zB&nzybB|})RSSeE%GL2w!WVuuuKEtk)rQs;OvqUBO1Ak6RnWcv*_<*qGk$Mlq=DjH
z7vZ=LZq?`g5bsWA_Cmt6wtn-xM4%mZ2Gmqm1Zb<kVTyDx&iksXlR%?{ZVY1lqI5ds
zH6(Q`Ei1d)cUhoqlSdQytNR;c_t#@LS^8^r`mP0fuNC69!`%f<n$fqC4@{=pnX@cg
z<wJjP>K3_i;P<JJ`A^<j{-ti3o;<Dl+j@drYCxa#B7c~<sVy!x_7`9mRN!QFB4xG?
zTRx*|Mq3O!c77cZNEdu*_^s0aC80(Vt4ggN4`2Cu{CG@IY2)`DAv?rcM9PeASNPHm
zd`CSwJGNqZBP^V_G6E?q1NArcSlk9RKfp8CmX+dfoh;5r6%KOLXCQX-3bb-<V|Zme
z{T3f>GLg7{Z^lZUXg|^oMqT9}eE%eRdexj$TZAkN-N53#rP|v7`%q<<vflh(^@W8~
zt2gKlUFO0F)4Y4luiHQuiDdpE3HAV~fdX;z>m#cg76HK$^ymeSa(Toy;l>N`B%R4u
z$u=xMgYIoq<+(A)V(Qp1Nx$j4su7Q=qI+WdihU&{<;dZN@;6NV8xJHo|I$ignZ&c*
zlUd%}>;hHcVWs^9PSVv%RtnNp0*R5-E={3WHn9BAn76b+nL+sC*S_4BI9z16TBPxX
z%MQzAc=ua#j=?Z_dtqq*4vDV**0P}xDMICz-hn!n>J9ZbtS#USBNMr?ys#jX^x&`*
zWl&cQ@gIp*W|KxqzT2elw~zLIgvjky;#tji<zGzhY%E`hnk5ZBdq~sRpe@dR!~e`s
z<RNa+?y!A`E85a%#YQ7x%eQ6TpaZZ`?1H3H3f@VVl5S|P)6&u?#C?S#M0X#Di0H;F
z3kStmJahFif0P=T7lijPFb-EjqF*l)H{>pNH1;!s@hZNZZlh+`jbD4RcQ|?4RnTY1
zS#lT<*)zF6NK8O!5^`Us0N*JxK2Q=xy@-LH0x#a7l=;#6U$x#v+b7*9HlR`i9l6HY
z?#EkvGl_b*;HjR}>osTI>4+42r8|ao!lcsE2fYQ8J=x?{CnK3DK#%4St07J(2RD-V
zGqL))NF|&#CLxL8R(VS8HIJr}B#L}P*Zn=muyd#CBi;_ZO=vS}&}P-W8jUM_d2~iZ
z6|=#f@gNCr5Ho_d_A}a@N&|u_frq+MGfM__sZND!xAgtWIA0<N-Ln+l883|0`v`^-
z&|`ctZ;vgr8Gh2ey75dpn*i9;uV0rwmOZCNHRO^;w2_1m#<cV<c%!)a3BK7^6C`U(
z8Nm`JpXpK9K(VXWj9v>{pmu+y+r_MDrx`h)C-vG|GL9%8Gzz3`3Cx~|%1S{LWN8K&
z5ZT;9X(>!oR%i9jvTAJ%BYZnitPBy5mfKp3k>x(s$`N2q5TYhoCHeSht^duin&fz$
zcEiX<a8Ox2wc@DwV=UE`+#gIw5Ueh<Kd{A5mv8BoS)x9alDY=g-uqU42++U2!uhe$
zhm9dyI6xhG{h4bmQ6M)aMMs>ke0$GL%KmE**2cGt@MBvJnM#*NIz;$C)i3uo8<8<k
zyu~1jJaAvvUOs<MD}vNR;Og@&Z)&{d=c}&#Tpio4rQAV180*urrf7}9n2Vdia4>x!
za&vBazR<YFd(P_-r=w$F-8eBDkL4{LHG)>xR6>o1G1^dq;~iaHIh-eZdx^li2`xuZ
zZ}UUWbTobb-inu%*R`b7?_i)^FB^JquhJCL(UQXeO;g5XU_Qg##|nYGIv+tC##}s1
zF}4+8men$qX7*R8ZE&FH_<2)BR1&<7`<7|HS_yLIC!a9p3NMZJ*b?ITqh+oKqq=bT
zOmz0YgJ&ZjhhhWi)XdC`keFBsL}DwjDuSC4eo+MSvDoMAdw5Y|x>#o<J-<qv%P$JX
z2pFX^dcffVIAZ;8Y3RY)u*-Za5elrjB$&4jvKwASF>)}jD>#zdg=@(2{=D)xG72Ug
z&nWR>K@?q^*ahr@0iJHTM*oZb+!EO~`X;0;b^Tqis^ZV>zuIL#QGVVwr#vp<a3pVe
zvdk-kMMzHmd%|PNE}5zBY_nFU(ZAksPzxJsGa@8}v|8_(3a{i{1T8LgZieZ09qYkc
ziw|Ed30u0L$EXKW79NcdA;1#wmbE}0+y-jzTkR&3fPiKMOD+NmC@=~m)h=0sV^OGZ
z)#@WpKY0^AQ0}7>T>hHz;={B8R&824$qW#4o_n2wUhBobJC=a(R059}Qey@g-TS`U
zt$A+eIOI)y3RcC<_cw5zEB-g{AdSfw_%tw>0E)7_ipp^M8>;V#nUcSX5Je7~4Wq;B
zr4<{Yge`1<C0^mvwzb8uhb3Nt2e2CM^uV%(4a^bP`1p%CS;-}CS{=6)g9-T&V+Sbg
ziePkFSFO&%n({xh+dBf26VfjU@V^(C^!+amwERyN$blUH^Y*Au;YD|{GW4Rh6XB+R
zNKy&ixdN`UFEPy<pg-@O-~nTn7EpA8Q@*7=bD(&B`0xRuC+H**;uPr?0R6!C#K)4k
z@#<0pshGBYV5AZC>=~M1eF&CI)mv3VsKID=Z<?iG63FwgA;egZNK8yk(`0_Xxp}8e
z{|Goz3SKiTNl8iQQZ);7HNGN|6(b{KRaF%sCFLIoskVV%YARws0Y|lu3)=84HAnGl
zON|I(-l{saoN9jnclzGmt7so-;Dx?PQE8{h^d}B@-9HJtu=pIVr$QrzN5O-XL0#hi
z7wPIgOl0uGl4O7(GsaDxESgyx{Jxtd_=;FQ8H%4@0gnH8dX7o&`!w?uM5iC%DO2$E
zpRDr&@DmA!8nhzM9aGIh!%p~<CXks>UxO~Na{v-Gdy4-lYf8YSaD05ce!Vv-c+(6$
zXF^QdjjEkPz-#h}-oni6<BE6NqfD%jY}E1jC$KkSxKM`xspE(>cptD~641@h{tN?y
zrd|X~hqH^zUr4@uFFJTqTpS&dT$$}yMQ7TB#jYfqWIFvCS5`QH<6c!8w}j15{#&<k
zDQEIju>blPY}gD5j~zb3N)>@NaQ4~#Pf9E-%AchnkYW3*Lx6(rIJ0HsDVg2a*yxpR
zRo@A=0ccwG=H9)7@fu6xm<l{z`Mc7ck&PJaPJGAZJp$gbU(=;V&?f|Z8sdQ^0_t>s
z{ckiNmQmeKb?{8i+3WoJSAa2n{`^gNI5RhQ_v*zzL9E{*=s1nLadvCUD%9db1!I|Q
z!Y%1?i$R&q-x2gnFi%r6v;y1{c*^<t%X>}+2Gp-zyXI=Q_9CX3Jat^=uJnx$+c$g#
z$eg)J#Zw~llT?m>30x3MEz3d$3F@~%QD6nwj`X}jZuh=#vN%644W1exo}Ys!DNulx
zlAK)2l3Us)`yCO-o~AId5w=N+l2_3o<=d$~1JwWV&gI^N)q?Jb%zJ%$`XTTpo^#=X
z_FkbTPCz3IyRUybaWXY!0=#>CeB4uH@(4ED>$h*yfPw?V4iC)Sl$4Zoc#!HUf@uv>
z03^$4aMZ^L;FYu)V4cO5k@7x?d-hQB7i)pO<&nY~jAR)BQ2^Y_Ksr?^SRQ(z0oj{7
zCn^EoK`fU+8(4CYBqVp;_BCi?5uu@_&}uW`P913G*c9bY1`)Qf@Txf&WO6JZ2sNT4
zbKa`R8Z`J)W{^N{jsm<#F-#l4cm~v_fDjR#!TAUz7df}Gl4CPrU^Lm36Beg)Vo5&A
zpPz0sA@!OK@PfAcQ9-+m-yN_*jy#o-SaZ{v_wN=v+U}L3M|Fg(#l*yfT$3u8G14xx
zjHH7C`ftYVEq*SSSMKyYvX-!g3*Do4lZ=CtIV?=@{<^-77_pT0OW5XMQ(K-iI8$5$
zU5|l*ArKlnWMQ6!Tm=P%m~D?siV>0I%wwHH5l9@IpBOeef%v<`o~H#9Z9wRlTUdaZ
z4YdjIAQGb?Y0A}sH*gX-*l8Y)UDqCYN^wBsi9sg5r9<CBkCJ|q#^6R+tyS1W2;ty+
zuCVch`&KQv8hKT~EuYMU(t84vjq+GgvEMH|ic{msXAGA#jJ$Fn7_WiXFNY$k!Wwtf
z=4`5Uo=%uZoStAQ;(K3n&m3iy%jd6lglSZOtJgxv4W}UuGu^(8Xl;G`z+-3#KOZ`K
zcup(5(lJkS-pDdH5Vu@H(-X2@AZKZ`hnC$`vE52$dFJtyW#m42?2V9pwTg{G!WL85
z3b$n6O@lXSeY89bOu3)Ck5>9aAp@&Luvp{i09-?9sumcp#d^h!nl<`ipO0e`M{a1O
z#sZ+<a)ri|hYwM}RQ$bpMmC0si0Hn<q<$fb?Kj>{rj9#kU~NSieb*;jdhPqnF8Z-6
zUcMLAw?*W!ub*kVaZCBA39BJSDPom~a2PvQ1zuqWAi9Pt(Z+V@Jqx)LFD5D=-abCQ
zmd_>&xr^{;?!tAn7vC?80u#8u(gj|PWBW+NHnt+NgKjtFrnx>Y;zkt7RUQS)H$Jo^
zsQP@vNbbAVO~yOlhs175kuacYQmVI6aUh+ORiqIpp2p1kUZ{AoLfgXUSt5y}c5BJl
zjNke>n{*p|g&``K@7`@q)OUrB()Z$E*qh(PC8h*p?No~rMj1c0HDgVrAh2zk^BDDy
z(I&SD8tZPbL>Ktnkq*yN41`e&CJ551wQsdmo;zv6L|IV1fRXv9;4=6uU@^9(NnStS
zyMslIk!iC%*ErpHi-E5bE;CKGpSejFm{D+2&cI0Bcgwm+t15_e8X?Dr@Hl9gqY+YA
ziq+oVTQeMwe6ts>{hl*c)e>rG{zZE;ZLyEJNkK`8L*RLBX(=D{LEiUOpMUyYc)I~E
zIiwkd_Sg6C3119xTNQ#AzMEWxeefVv^7?Rmljd0J%(`SU)(Cd|Arl^O+v4KM)6>&Z
zO8_hsK{P}{H=b4|4@PO3<MO~z!4`t<mTv^pOM;9u)UBNZhoT)zl-JRWdJm;Zq?Vt=
zp|G>CyaF@>EbmhVu-V~@pUmHc>d_8DMHT?dSCYIY*o<s#y`-D|H>nN~F;@cu;SD;n
zypU{V{ccOf5^8<{TDRvB!3?OKp!}(TV<h0}YzoZyv!I2}Hk@ttr@vui{MVDc$K-Dl
zRSWxrbm-jP@vvvtoHP{$Ek6mr6LV_BQWq-~H@HtY-5kyV6%ssmD>`;DQIU}^K|)<w
zS%CqKRo$be6Fs3Q=Qj%kvOO1v-Zev<h-?$$fFN^^sT%zO55m!j(+i?BD$qjeA)bxA
zSC~QiLWN<-(zWT_l2)u(!n78gZ`Ev_#Rt17?+G8kQ#=Lg$Exouwn2^SOp?ds3{*dF
zb-i2ibl_mOw?5O;+>ATfesX7Ik4V3>*g6i8%?hv!gz4Wv_pSUby#tK*1HdWB`N^w`
ztpGC%a={}9AAl-lR{gI4SO5uOzP3ri7@&y5$i@Lq=1Su?g54BNIRZjMulVGq6Mmcc
zvyYs!nI+|QB;%NPxQ?-9K`#ZwaOllDPl{kL0j5MgJ#b?KvB{f1wHDkAAPjc{?}PNj
zQR^){ehrg}44CP$Z&B}3nHdC#+7gMzqg-4Ph^{>vq8t&Tp2Q1w$ZGjJ#a{efYJ?!q
z_HZ4h9`o|p!3Y32WROaEgfpHIHUyh(cI?|`zM^ff?l07=RQh#a*33SY5L_OLUw`$z
zoMz;Lkz*a@s+QpLSZBxzTNR@Q9ZB*bAtb=HzSKvUAO^JqJO5Z-p4G(wXjdfayN-@-
z!b1Tw9vYFn$Xb9vOwZ0r0rLqXA9vNL!OKP`=G_D3;qq+iDz-Xeh~+9Ao6+^A07)j@
zz$V%G*xBfC4(8j%M&1WfYec#N$4+^vD-0Xw4pWeAx9wFpZ~7*m&v5)GSdPNX0}cv^
zwP@%UdHl~k!HoL|9LMbjMw*&r0Qbzn2L?21UF#mZW<CRKeW1E;0+<As040wx3Xs11
zQ_U1~bX|~K-`2SUXAKv>1KV1y(ccds1k!a4s}CZzz`|y!VaF|PVhE4@dE+us(%@zk
z#)-1I^{paKvO~XXe|EMPg16_CUi)4jm9t7Wl6uYL`X~5c+@;&;>f-1o%s|Se;DCS(
zKG!H{{^Pr4WM%tc5y-gov0xr94ca78m~w-0nEDPor38U+SYclK@ha1lM&Q<gGPozv
z_3u1`=AU4Gv*vCZp_nerxd%P9)gc}zVB;z`Pa3S4{JZn5;t6ikAGA#h+@wqF$s2ZK
z5{#~z%)7>&$zo${%^FVs0g>RN;{ShwNWdypt}DbS5Q;EA{SrPuR@;Abyc#BB<0Fu6
zt92#?M$nf4S=4gxd+H`w)6}4>@`sKqZrgCwk>aOBgsRorO(-(vzH?R~`t{?-cEHt9
z08pMcZ{Ae1{tHtvF_rO5Fvh58Xh?T?86xKavyj3t6C>jW@~JaS>0J8|7xx>6^;W=K
z@n*rI<CE$#)t+2YzWh)Wbg;MO$n6H$&gmUR@hi9je{?!7vu{|AP(F~r=*0MQD3T`2
z?h#n-v1PbN8{~7eIau&0CKbk2*2gNlXLrFG35Er&HbaD9-+k=@Sgvd#Ow6kK6!+U(
zp+-}^WKAjKnY7~JU;5Jt{UwB||JGDWsn<ATYNNZzX}q9g{jY_mG7u>2a__E!|8S~%
zC7Mot@1M@HgMdvmg0X3U=V#b6WVvyQ(v=d(uVN)!{y-Fqh#)vLpPiIHl*^`wMd_2V
zVZZuQ*q}FAlzP&!`9P)w7mtd*e&R{uA2TK{uJ_=_SsNYd1lWv?i_7lzpg7G88}iub
zE?!Urcgsy|o@I^fo7cb~sZwY|IjbdxlY?EcnL!vM1x-4T>iRG(-g<&G9|B5?scrx#
z0kE>*R>CZ!(F3rG<ehqk0yts99ny{y3|S*v0k|NT`YY^l8&~+B5sTc7Kr?4@DfQ=F
z<>mMBBer^@;wQ;ay{ZYXjI;+RW2bO`&zf-}ifl;Rf@2_WKf}o&0Xn+n=g*(#`tX#Y
z3AwkF!4?KZn{<?(C*Y;mztrsiw}&O~Bfs`=XwiHlEiCAgF`~$)Y_6NO$c~ZblI*=h
z4}N28kFc(yy>9#8p?Xp&E%;m*Mh;B<Sa%Kb3X%(kfQzYo!<Lg)eNTb!JjrOpWtZv7
z>x&qguI7Xv4a}wJcim|@{8?lntl!z!HFy~M`pu{=r(g=Txw6*>0OV@ryv+vlUDB#n
zphFp@5+QY{w2x^XTa(CvwBzIP<DW3q!UTh9U%q^KdtWs}+vhqAPgH$OpNRLmeM0P~
zcl_M-gN6=Fam?Pt2{wG0+;4L9Y~+$2m^W&}pp-?)<Eipm0yTo>f1T08-WXO1T2pmD
zBM3D<aK(><dSMAoN=~}Y;SZ`?|ICj3zot*FDSBmq&-D4QbNRWB9=X)&OFfc?OUmd*
z`Eo){b!Ogb2Z&#oA4zLHfv!9e9`Z^r!a(moE<zH0R3Pejz#Nn>_Q^IKuoWU^&^f5$
zs97Q)sCepOw0{OkgE(E0dyRnKU1`!x#Xx?KO?{Zp2unQL+j|ZQGJqFm?OV!^lI8EM
zjUi0Q%a>9T4(zUZvFIFfA&Q=3pxs|ZHOLxI`|^)s&B2(<18r3F9TlYKd)^nu-Gs}7
z#EN7;zk=C~uy^{TUjyg&Y20ro4;LsyV3z#fEeuk!<N%EfL=L0*bI_2o^(w#7DUTHV
zXP^QMjf1*&xcwX2x=*mEdzqW4z=Q%H^8&hBXlElKnSkyGCS)AYkW-MY$<QaJ{XNHO
znDqE~6G$!2e`>SUnXXXvqu?hoL(GhAkK$^vTwGkdre9(n15C)#P;5U3&5Qz!tzqQc
zzyAEmX`wAp*slxY>HfNW*oFZEJ#P&{%~3!{I6jG=j4A=KneVnQT1t5R<rhRmwm|w!
zvPYA4s_i4GkJ)J{2Y~5M{LeN70ZxQML5AYOA|#|;n43pZ-+zC8N(L+Biv15k7II)f
zeh_$L9pVO776MH_%X{i$Pjwe0OPhIS(HuV#`KLCrks2>7E=GO&^ce29^Onxx<6RhS
zm1e%n8hzHgHB)-Gwf0)7M~mMTngxO0k>*G|TO=mCh)4p^6oi6;g1Qy9Rt+xOFsZr(
zFFY7Z$F3DD(qI5KJ3&)&G=X9w?;eub4uZtixnPXw*&H<tNu6-8jqbN4JfFjL%)RkO
zZ#v!N1E+^>Sgs5t9;m&LYa^gkl8$ay12qADQ7ItLuG?@pbGI%VH;pOl6%PF<tjMn>
zmU<n9n8a3{3bdgWD#wk*f2GJ;u^B`70#*mcpiO~j1j9)|kGn9;Y|ex8=HCrRFx0sl
z;Cre(jssm`?iAgx0D^Kf22NCQ3M_K)9~2@`TyctEdj_YZ7)9;@$mV|$AXWd*t53+l
zL3wj1$b(1!@jy_1*)8O!onF_bec)<*usXQ-`$TdO5U?D$dBBG9tM%%yyf3^kg=`l)
z1Y6=^J((jbr8v#?tizESt+ZGD3!074h4a7wI*@|M*${{N12ULNLypsQCJ3(eN{2^E
zOlHk>CV`H4C5mQpPIY|j;OI@jL5&k|MIu%6x5gsvRiF9wJg}*Pk<kTS5P)>iLmvzJ
z?Q`I+z^!SBZ1MJ}IL6#zC^DB`klt>_=dnj?{a19o#tY*$!#7k4m1Y`{oL>BiV)oY1
zV*dFbL#bi%;3!X_b~F~e3#P5>FWTnSHF0{(Mn8g-OWMtgh}c!h*9z)jt$g?HozzsE
zI}G6~$90YBnvcsY8yf~Q69EbFF)SVTogsdK+{+az-xI?1)25+%Vt%D+xgwY*d*tlQ
z>+kOmhTJ&s$}jL17a+@*07j<*PrqPSBHMa(?=!NkdA-Z`!1|UJS8~yw7`S#oPveRX
z>&E9sbb(XWZlnkmUiN_^U_Y*H{THaPBI`jms6sy?Z#kdYu6&Cy{^zlg&P}6u)Z=p*
zZue6cUCNB?KA3+od3O%k(Y!256<EtGQu(h}k{L$dEPlVIU9fxN0U%ZB6VFTFjliOX
zB;vq9T(PP7Qf3p0-=sq*@p_ifA4{O0NpRBWqnh3RZt>h?K#tf9+hC=lRg9ZEB8FYd
z8R{YmkR?cJqa#853?_n*#?;Sw0MLK!^eJ-pg09?^A`t;vY(edH5~Y6QetYnLU+^q6
z)9WdCKKmaYDU}-XAGC`1z>@_cJ#v5p1#A;Iu=VD^_mTmwYVZ5$3Gf#)%gc{JI7jVJ
zOVW7x)e4;NNz8U3@A@-Ho_=t`Y50Uvd4zezlRZpBW58Yz)^Z6Dcr7)QE)7!rIT$aI
zL_1jYx8Xe$$k7;hbHWVoa3#3GUg1$nJ$*_DH4;9Hlok2Pn8%N?V3-+H(O9d~9q_GY
z0Y!mF$%{bF3BmUM2*BL*p};M8%aDON40v*xkosQStsA?!c)d9Ye{_E)Q)_y=(s5Jh
zlhV+7PSLHJcg8<||3-RuVUIi)*8&$Zl1_#P(yssiAB}Il0W@xC>-`$W(wrbH{7+Zz
zAQfc@3wTEh(!>CQe@u0ucQg#VBYJd@V+G|D0EMin<t$QyT^eoVT80JNgleqP^)5}O
zhj)Jsj~?XrxvBaAU&L=W`VqDSBtD&GUtL`)<<Iw7UHri2U<%Vy%;FO8b`yA4MJwbR
zRBf1IN17{x3B--zZ9aiy`!QePlU>081V|Xg5r*^O%@IbXf$(CFe;pJ0JeZ#*@xTLs
zAUws{Oq<~R6sfcm&_GqTB*)c1WXy^lZ}q|ZDQ?BHn+zL-iNZKqrR4>lPo~O-Q){P`
zC5I2JWbahARY|F$$A~nsQJ5iL!T@vod$VEOeK~4({-NqGq1oJPV=q*%i$kAZTrBZ9
zvWJm<6Y(YBCP7m9TVa<V<Az)#V7OtBFc7+Eq!T7mj*j(zu~^t+@4e_oh5<qxhG*Gc
z!4ScOg8KH=@LIqizVFGeYaYR<qUYu5@+s6Z%3h9_*T2%$(0LkMd5aU3^i4}sC61_9
z*{8_L$kqBvncwrf@pKbqZE-$LfmJq~O5^rJAn{*+(`!%s?-T*S$?4gdUgC2|?7GMn
z7v3&(GlK-2zDpqVas4ww@cPZ0f4Uo<2>%t%t2=COWMPa(&g2pJ0&G%P>!v689^-!F
zknf!~zCJZpp1x2-z2s8Lu<=@;Ek<HA<}=D)i<xt$4WaXm@ZEC#=0dNTUaqUyr(-IW
z?W7f2aCw@4Wb$hClpm4Judlx=Ep=H~m$95=lY?OW_7AVO-!;EEk6&3n{66Ua>h0X4
zp<eqoPAZkoZBi;E$~aUy*b<>sDBC$h4!cA-pN5Q6JM|P>rJ5m!5*p_yp=KPjE20b;
z#xXjOaw?K+;koXf_OqU6J@32T^}c_+e^{%PVSaP}e)oNU@9TSAA1V#=BK>R@6~DMx
zjx-H?u|K`-vG&cc$z@xVrTh8nZMBL;B?ME_tyHm61<2D(APuq#SFc`u5LNSv6)SLB
zqOvYX1l(Eo2w<%aHF4?u!0vqWwqMj&E^l62%Me~ZRoCP%x@urYsE))ZVw7~8?WVsu
zl%X)4^@~ibWA<xbgROKZs{}mZlUz@Zfr#UE2rc>NIbjuC??UAcDofDtA~+xyYt^iT
zC5l<*N(fOjt%`(u+wV+=&zm6k?FnioC-W7yX<N6^rq1}S8Xw#{Zl!FLOWoWMIW??p
zqKoDi6vE#Hg=TjlXS`?6%$ptWRS`2oy0mh&f4gz&pZ`mtyZfR&l|@JV+WhiW30~Q5
z^?B7hO?}<?u`;b4kxwV`Q=|?_D%DYbZZ*$U`<b6p?6TyPp<omGkPqvUc2n~l!UJ3q
zkV5<fg=sU=%Z0A;Cw@Qunp~r8iTnJFo{gI>MMNAyl7|=~kOuu&A+j;*$`u_XQt)F+
zc^_>&m_@S<)?#%?|K!XMYt=|H=1ROfAon23(l@{Fo7aP&HB(>p)aPX`Gr?_5+bm9t
zuUe|QdBbY436g8W`7ut_!?rgFjbT&=QtIkOr0F~^WBHdL<^L?=WNB+)#EP)Ek^PsR
z5J75j#YLNv><jKJxt|PWj7{VMQ?~G2=1tws5r&VUDR5@hRvFQ0_(gn2(N=_<tNT5Z
zw+!t@B6Aa^8sin6qA;_d;o|AMTG8u{2h`^ux+m3r3Xh*vTJP}~I}y*qMD9p6#J6pL
zpsWe{sBIfMN#ppwpHU9dSOt_PIIts0jr&~-m0Tc4J%go4XVD-{`m^Ccm6OWD$G4L8
zOu4JWYL-jAcN#!-_-lQA4(Fgl+})X+Z)W7z0q1qJ^5V9kX%GskZs}b;i4za0hoW=`
z0Y%8JeW-0P-L-6<(!6!8rEx+|m&Dc#=xfZ3mG}H~%Wsv9PfBt9+@Sc5t*xS|vYxkJ
zHAJGEBOsO0ChH$e8KF90VcOgKVb=ga0^^pW#?i939@wy9A_y0g|3_5(+zeqnB+G{+
z1Y^`rIE#4J`*(Por`D_;i+d9Hdwb&ekzERYlTS@2<A*JW+Gw|}F^TIgJe~P|y@ypB
zQ@|-y*t|mP%%MM$Q*27_ht=O*HSq3zBD3ri<!+Tt@A}Z01{Jnsu;aqQqNkD!=WH$O
zdou2FsOwpMMT$+vT+jFUVo|KbGo4Lh>6J@@;+=ns39XS3>DTCs*?Ifw$Afbx^h=Ml
zlwM(K&B#_se<rUJo=dm{1_cSycP{w-#0a*Hw<uQoSY6^9ocLt7XV-{C_w^DiD5A|N
zpB_09+kmG0mMaA#DVmu}wIQuXhiyxn;}k?Y^^0G^+wjtZu`do0HC?B9Gb{h>h^ny7
zyAa`qg4cFOH%0H96|ODuG>(3=qT#HH?r|sgNPE3@1q@;Biup$CE!E=|aJ;5kMctGR
zC6)Irc1WPeKA2%-6?^M8^L~c^T&VCm!v`YrBUcGb?Mys)L+@ALsJ3}{C*CGp`%R|b
zPx*oJLF(O2xf|sJgJr$%(1g!8A|A&n$@^{eT!X+@Z>~J~cY*L#xArX<!Q_aKcRfYe
ztxX;A%}WWE;hzW%V?Ld~=71wSIQtE`^1hm~8t0H*yDUY}gaS&UVNQ!xUwC&uyV(&e
z-p*19G*=`Dj~zSW0O3QfCd2xAfZXG3k7W@yUd&MUru%c>#yT8CO?rN1?Zb4|5rK%=
zRKDb*ASyq|vRhU#`oh5eopugyGo@y&OM7=m=2K8Njle(BGruqFJB36N$L43k!l)(2
zZadt{x)PQ(jHBNi>3-X0qvM**nX&B(`ZE+IMn*jA3la#<L*{33w;K^jTQSB*mn3P<
zq_Y`?hc9VuT#jT1<(@TRB&G}ClFL{6FIl;IJ_3WY84nI$irrnYP7;?n$M`~(g<7}$
zh1=eMJUHU|&x#cb)7i+Wiz6%&J*<zx)9Jm9qoB%ZbKUu_qE%e(XGePlCAC%!x3HXD
z(XbNWE!@c+i(EJx8GFHWyGrxaS-4r}opG(?=l=!h_%zIX;rQ4rv=)19{SE!(QWaX>
zIvlfNg^<^^TTMB61`N$&6wRr51=*UYjD>@L!!MA|?uRr$4BQGMee238!dIU@0%heB
zQdstN{?hcldr{|X?ZRt#O0)LRr1(mFn`KqB8XEtnFP;PX^2pllK|#52;W0a7S<^ER
zVY-ZRju$E-o}M+Zt2!d8FNgY;<08f$j7#I#6~)9HV%sVF-i{CP@9@dYBMH~9$D%<G
zg|2W=pLL}#x-Vz$TTVB}B_b|pVB7)yLo(*I4m1CIw&gLQ|InEWzM%#-M6w7r$Rre#
z5UEqI&@*%M4v)QH89;|UyKwpP82x^?$Nv${y&v=(g71Sj7uAtxN(I43(T2nOFq!|i
zJlhF;BciwyWi*4qsD6KEAzCK}C~?WKak-bE>w3XgXI+}ZnGT70#~FAvCShNntc%t8
zNkU(S&%4?nXu+(~R8Co^(U-o$n6aX4+-HX$;xvpRd2rje7P{4(5)?ukQlao0Hextd
z=P%OyDg{yV7?h#;65As|X0Y()Dx_xf7g(L<0KYMEnZYm%*T{JQ%oKj!z|}Qdw;Q@E
zuvMGTwqA>_5yF7O*i}D)jc3VD`%|H>;GH_QlBBVAE|;sX!RrL=EBxZc9>i(7D<wnU
zB69P2s53D;JA0|!(M4sQ-E8;tsj^N10LUb;<){?-4>_U<1#3cl;2`@P4lknCFW0>E
z*f2p03oh<Fmb!-(asIp!+TJ~tFh@}~Y}o%CdDiA(<g-oi0hE1}Ra8tg0PeCFNPAgL
z??8Mg2`1jmTQ8}SH!X4wQ=MaBCa@}SZON-+6ZjFT6wA9#O-@e>JlY=^3w}VBjQJw1
zIto}mfda_OOKX~IhWxdo58p4D3|Vf56UgYxz#-9z1dY<pas49qz>qQn85v9FkNj-|
z!ioQyzl{if4j1G+rm*-Xd#&*9tmNUts#=bYaxfilJzm!p9lE+jDDY&i0_cwkKup1#
zREcpo(l=<f9~dnGO-)VN=)09;?N|C{PfSeK*3`U#lTsGY709{Et)Hyv^f1Vg=pULa
zJTudh+54B}V#uzyxU4KR+~punhSg@J6bhH7>e3X6JAR5lCE(N+M`jbM>Q`VEDJeDE
zjb8*_zkK;2=pb&qtm-A=;#F{nj3b~+zJN&*#0k&=EYre4D>iH%;PC0=KEbDM9^T#m
zYOF}gAh+#CNZF!QKWm|`0_C$Oar)bOoWa*ADrPt5efb;`)6E%pKdCIz(oZ$hLsV4x
zGv5do09MN^l5G4A7nS&Oz;ZXT;yBF4j(g#|KZBbJgqa+4=lj2MtJh;m$6X#qGo_8A
zz^s1JSD4m0MB+Y@lD}ZQ2=xu_4A+;%>__&0*fM^+<=i`z%YWIUc{=Y*zvpE)eA$Lp
z*fENoDDxf1qX%c?e>f?xZR1$VGCtb|(&-3aI;YsS7{Yf^FyClRVE<-xBGn6BTPob{
zcUU-61HVqz9*XWOrP8J*{oq~3v=$(lSbVzSCTovR7=pC@)IfS!jdzD?R_%~KwF;0W
zv`LfED0%ewaV$6XfE>Kl`Uyp7iTET93+$TudNjkQ!~9@C&V(Z#aP)ys;jwU`HT3kq
zDrqtW$*h20#xx-&Y|`{Wo@(*RmGl~Z^B;k>H~LCrH6}yi56Ho8Z_@Ac<zq<9t$gyI
zC7rCiQOZPI4;~(KEQ-#kbt%%)+8W4GR=2fT6C@{#L5DucgV*SJcPRW@IMXYQ!VQ)M
z0$mnX5z4-O`)FO|g4Q-VY{n?Z>3JLrN`Np8dEo*a8fid$5HKgLw@w-_iA5h&Z=j>-
z5cv{xhOx;R(On8KEAlMz9*shacG@e4!ZiuGF%>Id>Ox7)Y&0jB^~cRnbQd(rGOVe&
zlK}LZ;gDwn16a`x&#qH(pg(wfdK&fdbMNCNEhs1%F)=Z%`6sUG8yaqxY-wrfQTg(9
z!sbR_;kix^R(@z@ysC!<v>Ru*;GA*nCj3;a-SVvdkx7fpe!kmA+!wu5$Am=06pb*R
z1T#umyH;lspS|ZD34g$@JC8Zte~)X1fFSeg)vL}ec`U0EI!$)WQt}vk5Q|`34bG!k
zMS~oJftGufvrClAzVU=qCCB+IL`7L7cKdc49OHx~y3LulKltmbbIg|>8jR6ayMS)d
zM%1s}uU{wQ=9*!OH9}u@Z5a2a4AOto8A{Ta`myR;%d^nuxe2fkxrWR`xi>yEq+L?0
zGOE_Z2dxh2_%a+4!_<~%Q9OzdV0T+wG4+4qr~Ox`<~^+D{1bpF{|3F<g5LHI$;X^t
zRQLs2!qPQs-22>7<uvA@!dv_$4JljLh_=A0G{VN79+i*~iM^UNwjOvSnEcG>MLOr>
zh>LI!i?ltYq{`*o+A?h*3S-xYGz0<iA~eqt%po*=kq81&Qc^c{4i<z6SS#?Pt$rT2
zjpPy$Wr+e)1^79{`;*{4J#b@ZEFug(lW1<ff@LNd0t@LDqCbm=On<=ip?n1<nNZln
za-{imkr5||V{~=ldAt3HufRHqi1CT$JJOgXXxoAw_*P}~Jt9s?_)y3U8Dj$gsG=NZ
zx@WnGI|GPn2{U0Gc&aEgPlUKlaUp<RL~y(d;9s(Wk?nB1ZOHgBG$yg7cRduAM1Cl(
zZUodN7@cN-YmIo$`_=*lq)S*+U>0;cym_9AFi;Pc;7siA&LB}x)YPc5>(^&tYGhz|
zC|WiTJaMae=OjbYLE1Hkp_K>c6A$~XIXcefc-R}$jZ=6@yU+3NcZOsEAIc-X|D0(k
zdnTBi#!zvLQDvASyc9=BfiR3)hX>E#!3N4S(U+Bzqi7-jb(louKENX8j4Tge#g^dP
z7Buri0k9@4Kwwq(%F4=S0OXkE-x<t*tx4v)n!5UTNFfp}w2xid1@Q9HKnZ3sOwdTN
zfXl%dx9<ba<`h5#8#7GzVZk(%XFc0h1b0&b5*K6<X@n9=$i6<3Boc{&j)>`LfF~sr
ztQ4UKqbA8>n0!D@QC(AGpfSob1;F?yS=W?oIA*vk=D-3m9m_qWyd5MH89rQ--osvk
zVmD#@bW|Pz-V@ulD(K3xBWx4`>HF3M?2%`gnV4Jw%nFjD>Ddp1m}7C9kM0ehlvggN
zqo4eYE4jmWzu*VNavzsY0h45e?u0o)!MZo@-TDQ2JYzH_l?n&ZAngV?g7^{QxLD*e
zk;58A`k2CU|JL7>Yv-YSk3Q-=D;-;#0*mpXoqdn~xqhEW<LVVFPAt5G?LAEZLy>v}
z#PPZqVwiG$f(IUcjd;n4iHT(}Ok<-Jxo%)!AOsath#+2&YX(2kO>?`)%V+SA*6{qk
zlf8sz&z^CujongpS%w`nt^zFT5j;_3`@qvgZBN5(!n$bc+ChAh7){@S*2bXP`ufWN
z=3-TL^-5eMsfm=CL((vA)b`M!H*fwRtPGeP0OolGhYnDHY(`?#;^eiYw61W(6Dgvx
zU3~uI8(U%z2!K*@Wa~;|GTobU*NA;un@VlNHwAk9+?#G{3jsw|jk;$Xb36be5HL`{
zZ!QkFryJ}|qbZqoe(v@0(){*uHz7U1#7PA=1_y|pn8Y&IchFw2LUqndkUBm|9kWVD
zq+#?$j*U7>K9ssC*pSaIbm^4ug|8?|diTkRf&0k5FGDlr4Tel?Gy}H|`>?3B2fAHe
z-rlb<(Kja+!6U2gKZwfZ+3P8oyH4ZjUcO@tm6Dh|iXL@>G(`zbLV7SMO)I+Zp7X2P
ziifa@D#M_M7muLSH3B~Y3%YLx81<<HnMEi{0858K5%<Dmz_x$9A$0+&nI_itcs#HS
zEL&@3WW0QQ-T;IAq_L6N)nyMNTVRbQvCZNQM|x8(5!}l%h<504RN;Pp8XZ0M;+7t)
z9ShL;Sh?M`3cy=^Sf5bo-FzIJfHoyQSqkcC+*IZ?C}OL2_@0NTDj0R`S`}i)HJF9O
zW`!9}qKdf4Cd#fY(#~LDo>|NZ{E(3b@9A1a#YjA+SFc@5w<Z;}^v06UOA&2Y2VZvu
zpC5yD*9*%;U(=K4Sjqw*@L<1{M#qC(Gyn}I2+95aCl8M&{-g39SJMEZU`|dB`UEPD
z&&0J1a?uSviRyz`MIexMVO85;>dhm$3X~T&LjF8~qUmc}=hPGI_t-oo5%y>U9fMF^
z2|9AqrYKk^2+H+(s1-bjVJ7-59C9MZZuFhaOOD6?4}Z7d`=RoG{^JMz(}Iy^iAq@s
Ti@ou89|ZJt4fo{e*q{D8^o`k3

literal 25012
zcmbTe1yq&Y*Dnf4gQSRnfPf%KN=hRlNK2QLfOJSVNC+x|G)RMlbVxT!w=_tIba%&@
z+y8gIbH*L#-f{1J$8dnY?1#PAv(}uyn$P<e3Q{+*$+3};kZwvpmrzDRLeWG*LT154
zhkv0K?u~&j{Ep8w9aU^!JGvO!n;^*>I@(#;I$D{(y60?S?_h3g!^OtO#=~;Y%+b-#
zL4cjz`oDjG&DP$OeK$E-3oe3X_gu>X2?^g2@dr6aG}{~r$&f}`;)$wj;^wrAXYbf`
z)9zSrw7hr-E}i<PJMW_jhCk@$TBx4W<$l+f|Fo7a>oj0k{?#*YHfHUYvOvz4v}c=|
zk3Mj#zl(qF_u0eS$8&**j8lZ~n88hyI;`LSym~p=Ve5g$))fEpz6g3|lsG0$0D<^Q
zTi+xl0pfc`^ZQLih;IhsQ8oYlPk~4z*?&J|N@Rfb?}z13SP)+^lcas=5nt1|u!<31
zX&f1lWBy&G@BiCHB(S(WPwXs4^UXJHZoN+ubRoKXHzY5Q8QuSFN5}IS_jOiWbpP^N
z2Ahw;!A%An1cZd)+hc4ms|fs?n;NgL@;n*?itodFU}V5Wf0`WX=lNbsTUGUzZT(@)
zmVN}E?TqnU)7zIdj_<X~UYEG;zVh_)Dq|INT2oce*G#iKKQ8k;wRb<-kj|8ewzIQq
zYHtsE;*VZp_~RYzW)MO8N_2@cMgW0Znd|PnTv+4T8uMtOZo;DHk9Rk<iVct%BEK=#
z#^&YZFlNcci?@eTI#Fr~zL`gT^ym?bX0dNLoz$29#;db+xx~khR3^TuWOEucA)Otr
zXX~8vE6F^1^%L_p+so%=uY2P-3{mv-^nSLrmBRa}aa?VQVb#IJ!>e%B4<Hz&fu(Kz
zd3`y1ZQ_1$JnK{DygA9TEE^dRfQr-=%bv388P93LsM6-%IP<zEuDK-;hfLI4h|9FE
zhg<h_(y|Egx$}+`F%8`~pD{{cIc}5Z`&?g=!U`U&4q@TrXT$ZXs;UMGbodlfL_R5h
zaG7xnvKYzzv$4_h17EZ!Q#Q7A`22jkaj3z|<E5@HlSbj)dXM9!fwZU9M<#z-0&(v?
zCh%2A5-<^rK3&aD`IVzYcbAfKcXyYDg(ZYpqwx2!u=C0QvhUj)7_T_)tc;h?iN)|}
z!Yz)9x^<gX+t|kEWakh6Xi`i}TY_za^vBz*b}laMBO}pAn^R?P{<gh5N449W<jImx
zkn~5t87Jc9USZk%4v$jCz##Q>w=?nq*<-tVR8&-m4^iV}sb2Wf-1}<h&yUA~f?QVP
z#d3!ihhzG8K1u4iEk?gQM{KyR7}jpMAFeBDL5w;TENXe`^g=>+eQ#jk4OBar!e`t)
zJY1ZzZR}NUzdVck(-Y5ius(*bp`nqh`mJTZU$p-d1z+jr7`(sL!3>P8tu4f_PJ4<`
zVKou4?m?J7{z5$X6%z}K+~R@aV{UHM)3dWL)KXGX%wlM$sEnebqC5^ua))Q@Ma&o&
z7$fkW57+XFa_8m@4m%&D<dhh-KlE=-P4?XDlymYv9MR}9;K-p8cF*B=Sjxc{IT0Jl
z*HVdQ*VlU{CAB=>Py9gP-%G^m&(`uRcIR6We;%^2DKKRYFLXwdq>77+4=3_F3_J7-
zXNO2;z^#pzk(KQ(nJLhz%+=8MQbV@b7%%zv%HA(L@oPQBuwxciR&pjQZEbCQ)<+BY
zN4>5t+!i-C3mD|$QkQxYvPCX-o~ku^dzskU79DL(C;VNHQnr9sW52L%ymT&H<owy%
z%A}~Ms8;xr{}49ew|Fkohdw?&CZ?v@-e;>19IC@a&YmN$EvNXX%P>VJCg!gVf6KVI
zxEMZMA9Jv#gGJ-#=VyYYI!w8~kO~Y6%7f)W%VBT4P(D35$%)l-m+p<{lH#!#eoppD
z(rIs@)ARTd50BbNo`xd4@>{IFw^nbmuui$zpg529q$;w--?mRnrR!rw8hChkx$+6T
z7HuJ9P<ga+&4;qOnwy_k&o)fI-0Eg;R23^sc|>Cs4rTC`3K~}NLyf}x!k5*rL`2Ph
z-A7}DPvNw`B)@-3^Yk4f+`oO&WB8?Nr+bSdIZ7Fv<;0|<*}tc!%Mu&M!bGp)7ID;>
zGKq+YqWt~+*G3BxWE2(oQc?t63Q>kh^*zHa>nZqbRDI8O+o`#d8u^-(Wtd=}k&u&r
zvGgLdnXbvS`CS#h`g^fEHaa^yoBYu$iNd1GlfPjksWI&O$=MZcVbn!Php>!a`gQBw
z)kbpFWMO*@dFuWe8d9ru-i(4*)!6;<!^Xi;G3km@5_H+Bv!zH&OT&JllC98t;GLM1
z6y0p_V%F!n{`3NB)~!?q#pJ|okv>>F<h3AbZ;n`5X=x>=_0isgeZF|PwYL~lFQ#i<
zG#DZ~U-BYW9Db$K=H!T{0t$=L%g$SA1eqO8Z*RPDap4oi!^TGWeyHnGgoTZbP0Pyq
zsm!z=rPz5>8`iVsGtJW~cxR^^<eyN{&=$?tSOj6O(ZNsc{%tohby*%8gax&8aOif%
z&SiuHeE05MB#wAFL_OE3v>7Zlkv-a&h+MafjEF#D)F^zJJgHr7LGt$P+wYcReV&C@
z6Qx3z2SXQ7x+I=GLpCrqZHMA0t)P%)+Mn!swwf*Xh5J2J+ND#kX)ii4xPjmF8$}8W
z3vtQ0J3Je_FVCW2i6b8<F4czRsTZ_NO(h`TK#%LLuBzHO7*aG{lQK5Ge|@o!54ZU-
zDZ3uGPqMPI^2)DFoCH4Gif;+r0gRYhCQK4c${CWM<ifc83f4yQOg*ml7Q2xypo$<l
zZvEz;a-aVF8+$k<F7A%lxpHHIzyFPL%Q0S2+nZ1aU%?i=%Woe5*Za~xmaE1*HfHkj
zW1dFQ(zhgdfkUcC<-3~{Us7y;S3N><S{wdTP&qT78LO8LRkLoljl#|1MpI|!`>tqa
zQ-S5JEroK6k<xr(2e>oW8~a1;a0$eY%xv)RX&_6<9vB$7d!7&%_dF1rWN)W0pjaOI
zzS4&~pV~Z=G-QzDq=;{`>xIO|YQ)&nZ-!PCHC}!B^CM`eNMBUXeU%<cqu-#J@#J^&
z#K&)F$HsWg29O4Bhw9MM|NO#ZiS*cI>n^NO&(U&tlHb(^E6v~igFz{3LIQ%%*72n#
z-P-wHG#|t}&k7g=ih~85){tIg%8)?i+PgeUEX_5&k0$JXz=&9x6h&ECIst(iG3&)j
z{0XfZM<)2!rR)^%u7eur1+&NHo0H`k;(^$n$I~u#P!jpkImn>z^Sd9I#q3Lk-hCvR
zVt93Nny*_Mb3Ee_2B&|whpT@FZpiYXHHoI)`O5lwce6i+l8i}>&2a<o-~EN9rL6US
zO-vnHOMpAaBlR9GzkmOZ6L7M~S}nz0?oa7IJcMn7q@~Or5D+joH`nwa#WTYE=Yr4m
zIraPZ?+>=7Yj=_w^YD*lFF(`JG1k;7PF32PPE}YRdlh{Ah*MOzhaH4Zy%epX*XewJ
zWxc5JHJ}NkgSC;5l!DI2xq|fcf&C)qA+V0~%gY#Ij*i^y^#?>yK^co0&K?eBD=tI1
z`;%AH&^6@oFr@CNM59PAQP3q6TKAXk%&e@v3DXpl+@fb6Zx4S{30fV>j!aBU%&b56
zN)BMeGe(uge4}q;W7Cts7nb%k=vSSa1C__dBPe@4&?Ggh?dh$jDonP1SMxWB+`EUA
zG1!4Y?J4%^6>Vr}C~V>QhibVTP#F;*D2z&YWvY^0DvWAzNKrHs8fy)MlJ#IZ8XRn+
zrJneclam*4l&r=`kcAEgo?gH)658vag9S=hSutDcPYFv&(XT*mu1D-BJW7F|A!J<R
zRd48SpyM`mcYmy^dJMp=Ih0axyvB(c2|C|G?RG;39On>9!Q1%w_y8v)V0F4<Sby~O
z^(B(cOiaYP?Jptkz#T5On&52vL?QM3Ii`q+NR4iu$6{t!>U1K9%hpRcMQw1Cc8`u;
zL-T}gZk*fS*B9C^dYzao?14Zu(2b>Y)eE{N%1kHV4Z(6<K>xn&Au4_gi+QG*F@Q0^
z@2mOAN=6uGbT?D)leU80<18e2?-^?|^|kgcE*c673b}xiECB3Sj1}r0Z*Kdc;i$r9
zjUF`Xi6}LM1<nQZ*)=nx=jQI7JD4Gr*B#3~TiJK<)%>yjLa_TzGe&e-Ss9AG{Dc8#
z{4m_g=%S(`dq+oAXuK+omq$7=>+Yhmv21G7HBKr3h;rrPIlHE&bZ$}!X+D#XFn;JV
z{}YP`K&V=Sr`zJf!dH|VH%6dezxEcPJNyBMkh6XZuG-bxD}T5)!aU|3@b2A14UI9D
zWkJW4RB|4RsKq+?jcZW-8_wcr7#JiEH>Wh99OlSGKl}zY=NmLa(}XOe?V0*qpX;mK
zSFInq1_zY@{vbZtH)!n%-Os<EE4?I(DxH&&L~w~<<>#aSKiI{6a#p0s&Y#r|%ZS3?
zP%(>O9}>H%QS(q^VVA{4t}X?kYNFXeiOtn0ve~pjFqxJXK)CWy1@|rmu!{F+goF}z
zcI=?bqM4CS_xAR7)&FBaGZ7IHIePU1i%>WR*6AZXknhve|AG4X>Q(wbZX|)FI==0$
zwUCVe=g*&$)6+pH&6Y0D5Wj*~PO=zVOcFthf+GrwXAeu!Wmo#6@E+;V@KbfU``L1g
zlkH#C!P`~pk@k(oC;Rw&XlSTlnq-RP5GUUoEuwFle|hPKun1CP!?BMZCXEbZWT%9-
z_LuTn8k@4p%67BwzkRk$J`^Jvf8fASYHY&);p-u)T<H*q7WTHl!QR_@K!eB_i$><_
z_wU^485xh_<wmzYRLMej+uhsi;?2&KBfL#R*4!1yq)QRsMI`iQ`DI13TH8LlrZ6G-
zLCmT+jl}!(C|X|L+Zlt;VH@B?eOF$QWqM{|@em6O3(3{PV}0gbhyjUjh=B>+nWE;w
zi$CjFmp3iS**%5DF%9W#X6ibje*^q)o}QkboF;UscHO0gmk#G?uy;GAeKjYjpy8jX
zSR)k2qzHes{s(Z~VsfwtY$GIXf_y$;b0S^tlj?|z85wVDPZ@QlyhqhMz4u<y;SWXr
z^v&mN)aw4aZ2c5X1-a^B|GkighKA7POB|LJEL~cU#>vW6al}UNb)`~TCwKojopP^Z
z$N3#dkbdWmi09+Wv%@lX$_k*f6cIMcR)TNBQ|S6jH-A3+CEvPnb)&ICI~T3eW2?<>
zf3VNR<H7lI_hab~=rjRUVXv+!mAYUHQ<R(;u&8){qUGJKIZdprC5}8l8O$iTstnT(
z(bs1k5m$(x-1ffuO{ICx5GVH{JJRW#6*foyPfw{RT28-c{1|Qja=1iB<4oi#HJ;eZ
zc_&Gq`k)aEBPZ+r`Hc|m+(3pK-$l2Rj=Eh%Tv=R)(N)Or>YFAtVABMADqC{-`72Y_
z{MT13sB)_Ln&F>669Dumjhd=<co5C3vE+|0!ja+u4e*<CCN^wiqlJ!eprPTuzP@*-
z)_{Q-wS^d7ob0ORyg&w&ew#&8OhyJ<QRMVK{BQ;wh>NT9?HJd*nn)yOYL*go=IO6-
zdzPv;tNiisrSmQ}@14}<&EHy5>e`j;u$3jEl)q+8cMZKcJd<a*wr3pM4zBcmKfiXE
z;FilS`i<Sl$cT*77?VyiWDba#goMOEqt7*q6R@+lg}SwSGalQ!UCc#DJH33fJBYtD
zV^9}+9NSn7=QP*3?Qcyjr9b<aqxy|V(0Rk4C+<-QmGBo}lOPuKKzUvnEueTBgh$WA
z!_&Y0L)-YnK$o+Ge%D@OL@m=j-dzE~boRk4h1T?SO(pYjj0$g*_R*#SZbmw{R{;d{
zQ{@S(C-tX6Y)(w~*j5GPXS5@;_mpg1+FNOC^)E|GN`BWkeZ0Cn|GU19BNoXZ&l#=N
z;K>gPz+V6iraN=ZOZyH>J%7GsC(i?SL!E<9U<~E>{(}e0lUa?8qJWYFkl=Omm&dc&
z%9%1D<UCz3`{6^4H~NSI0qYw&FTPFITv{5C$B&_b#-y>KYW=NXs3siq+NQ$U7Rt!)
zrR{CSST<cVxZS`|jhp?@UN(46TwDPCL@?ivr0jjSULxNWc8_XmYLZ*t@=8nsuKeZN
zhMWAGX349p)V?8=?&!FZ)=#XFge=T?%6-yf_N}2wl*8L-wT%Hk8lOxPSROyeq+#ir
zn0Y=tHNCIuEOkKt&d2TAgb63b>X`v+m{56=hGa%KK&?`zb#-7(KY&rXZkCUYH+Tsm
zpl2G?de7ZfQfTEA9v<gHV@Z)7>l}A*<NHkRqqE{Ayza(~iAEdUYnqy7sys+=trc8c
zrB*00mI=b<DQ~3KMy}i|)_u#QJ9WO(q`avZY1P0Qz$k)?euE9qWY6)tPpW-#1Kp(P
z3x37JMLvh>W`k4!CD7jjL5i3M#08)l8L>ksc&+A6cK-r|XB=|of{xQ)s#pDe@$)c!
z(-$E(PS9rDRZWIi-jJpZ_BAbtcu$DL?)DqlQ;UW;)7yOv%+WV+HqzwN^6<Rm?6F8q
zYMiyaEsWZo{Oc-lT2qkyZ7e~0K=rvGZSAc-gX-<z&8bQ*P?JDGX$7cs0_w+OpDUr8
z6ujKAh?9qSGacb{xMZ9GKv<ADUiaL&I&Qo!F&m`4I$fro@jBK$+Md<Vt-$$3-*k4m
zoB|iY)2(s*1KPtPU=?>&O9z(d&Uz^$+oel=hle$<_2xDgw|>@LTk3H-BuZSOh~oNx
zb%^iJemT%*N@~1#@a3!Ki6Nt>Dxbp~&n5EXs+L!_PcQ*kx48twdMstq#$1I0p)%TE
z>P4VS_rtYaKrviKZMdL<#aC17=^>6J?6RGo#EPGjllzDjg_^%gN4mnDHJ_>D9~>OS
zr53sSAe*S>H$4S<bczSBLy60_Uc<$nY{U7MAXJW4qcGvNU_v^Xg$6HaY3Zuk+Q_)L
z;Dm$(o5L+(r+6umX&E<D5BNAZD3)$fJ>kqMHN0z%wyDOi2TB{mMC~2~je6-vew-`O
zzU@X6Q0xbcotTnRzCgPio5bL2!Yw$>;dX}tYF+#!wcEi^J?0k|TjyIpLJKYX?ado_
zSvY1L?g6Tr*=Rl)*Xy3SF@2v!VD(>Nzi^uk_+}@&e*oI@cWbLq!+~CzOwd|vaH3h0
zQ-#wwc)~qJO!~<$trJ@*yKkw&3iU6G=dizX6R7EkWq7;v#s6~X<7J03-eG3S$UA%U
z+qZ8$DWWL=9F4fiPq%8fH7ji>EyoJWoyR_eY`G9V%&c<TXD~N6XUxobk%r>1+_(6N
zZ^i%)u$Il@t;`qpo-<vo+haw&(d=gYl^Y?Gs(;rsI5DhjL?&O49B&0NzKtUgUr1;+
zFaVBS2V~`CqjwSrh-H9e#l>a%ljdvCG;3Y9c@+r-1Ozy(#yMJk{P?C*$(Bk_Ng?Q*
zFORC3t(fw9WgzYT{rfMn<j7`R|K3E$yKA;-6JBzpm?Eu1p+3sxv#2AAgI8F5n&DR&
z?~WXLiJYY~DNyLKPy3+NVtISUQ9u;&J}|jVO-)A&n5T78SBVUG%zp*YNrfUrL&T0A
z%vPihY3sS}riLmHLd#G_wK%g%7B0|zhI{vX0k>;^H+xc~S8wO!1gj&1F>Ap3Liz)}
zY-(g6a=Q1sl5m1kc`E%#|Lq{N!^NuA0AzK=VcrapQqEX0D@I38l82+=N~{2{Cg7lu
zbDI%BCrJe%i$CU@>bHpA`{?^eM{QF&wd*aw5&(?8S^Y(yfK{ye<VNs(OJB|>luHg4
z{6O2Hc7b}>9}BncFFvy=s+}G!RkC~Kg=b=$@-fbRIV!{gWbUdUHgj^$QSIsV$#ToO
z%cH5&LR-XfSslWlrly91#H^m*q09i#5kbcm79?K0Ab_^S_<U@l-h*#-byXEUrdUyB
z0Fmf3Ol=O%6=u7>v+7QD`-`Nps%Zz=QWdH-yF6^K<%NQRr`>aB(RI;*f<qsx7q<a>
zD&?*a8OX<T(ofyItD>R;!mdoiI>;~0v$H8XJ3Ce>`lZHL#0+u<pm%Ldlwp7tRpYXq
zY;K$G>x)FrXOp-HJiE{R3y^?EupLCsH_AMZ2Z`h=>GIb3h9;cnX;-s|sLW3f1Cep|
z)p?DkWAyd!_z;Nu_)A@9U!Lw^v64Iu#U1$TmGgqx*4hnYp&)T$Gi7MIsHsyvVnr1I
zu82y2%r~pDvaP8~eEhHr+udJqk6GULZ1IAmDfx-qWG2%C!LDCcN3b`X_`N57fI>h<
z7H<hE*Y8yQMsEZP0fu}YQ!nUd0honqR(hsD<WBcGGoHnveq-xrG@rklTR)*P({l})
zz0=;@E~8p^R$!T}^_Fy&+;nnTKMIZcV7Q1g>34){qgFoGRaA6)XCbqo(DsMMY0{R8
z-QmlnvCiQ%0`VyAK=S5XICwyX?%<*WF$jYHPC;DGhC>P+;(1X6%4blW;Q(_@ubG>g
zBDf*2@@^GM8K!sm)C$XeNn`?!D_-4-3A{I5TwM0kjveRM{7XWv-0?{xhB%YLWb67?
zijIrV4Dt&%C!#_QHN`?Ho*cM}+J5YyIP@nCGohL$_nA}hjU216s2K9gxa)G=MHeq{
z3wTf73*=F_hw=Ol3^GhWo?G&@%6i%_PWMki0I*M7V$Q)KWrG?Vnr*tIb15t={Ql!d
zcxO5lR)(Ds^!b`4esIrGP*5J*&54CBK9K(+0hAoHfNT)1mWGMlpYt(gU64ml*;~LD
zwbE=H2W_3jrAU5jaXsS@8y>RMOE+7CnAi&Yn?Lvd9`#!tO2o6>62B#gkQ@LSA`Rzj
zMFQJRNcs5jV+Jf(RZWd@gXgKx$nmI-E$B7|-a;Uy|MN3^tmV_PFrPksikzdCce}Z{
zd3<VW{TGugyh^uW3!E^=pM!(pq<XGz;ZJj-h(114al!;w;r3+`f&|wj%E8zV^60ZD
zyX$$GQj8zICcBw4(-GA2;j1@HH*h_AZ8>T#PZQ9A&&(h)0|I@P;HGKsmuFU1EaBnd
zXjdW^pob2cSBfO4&LJ@~F?H7QDMJ0Pa^56|-``!4pL~9(maB>de8$sr$!ZR$iL0C2
z>ZpP<luIx`9ti6;4hH{E{FQG~S<&^ydYU3wR|uJK$VN22UmsiA6lWOT0aAYQNtOL#
zS3Fh9<2QaFHtaOrz`s@17t!FQ_cj}oSvAL?!LrEnbPsSfO?UVkJw?!(zncyEA((G}
zyd1-5kv?oCYe(yKz5m5mrq}Y7*=#9%mdX<cCZ<(FDfK(O-5;@!%D;WuEh~+OM-xz(
z?mal_FeCUyFJN70JDC=C3Ik{ye?jqSG2q~U+kk`+Q~_#yds)p1+5|$D@j5?R)LgDt
z2|(OX5cumZ_PSkNUFV=(>e$vtuksvQ8qGF%Ij#-AhmZ7UX-N_k$MJH@SiM*peQY7W
za>F0!`;>lgaEA+?B6W6l-WB!!eeqa8;0suDLPA2I&7wmoy2s5O1tKcZojV<{o1x=B
znpHDnin-H_oxk`Nh|!z(V-7V%H`!gR#~clBY*CKbe?t`w>l6M_*Q`+cW2gPk+|b<`
zg}2)Hw@r!y2`VI*s^Y&?mS#DgT$<R=zMUt@nSNwD6nMnua%94s3JxKn&4K!M0{8rl
zvopE@7<^##?1RV)ExivEH+_VG1|G^@qLRU2`ZKvWj<*2Q%B-h&5l0%1_bL?j<Vn~0
z_YV*d0c!gXGj1X>GV}69Sv0(2IPae~XAg@l$JjuIH1-xUY(W*%*)O;IBaV(w)pNSP
z?6%*_2WOEE8XH3I&^jrsEH36~X=!oP^*8y>#Zetbji7al)geA8Kno4OGcQnDi;EsB
zurI|KULEysgkQI0nvI`fB~IP=cqTSYfa!))_vMyuBYgZnuOG@C4u0dz{A5+1(q22I
zFPy!(sf`Jou5ECT5IXSwuSD<^5xU^#&!2rWGU$-t$eXr)AYsw1i2@Z=K+%cMdNK->
z&>M)&i4eCCn+^`b{%{+@^IFY}H5*8MTa9PQWanEBDveOAHHfwiYg^+bf4xW>gpYqa
z2*IDz+%fe-vyEQBntf25U(1Zlsgds)M5FSv2D|K6NBZNWE@E2$Sku*S+I6z?6y~Nx
zP15tnZWvvy1zXtwb{23Tm6BIW2_*p1+Mp5az3i{w^{cF`bmLdnn?#tNfFO)PZv)Yk
z1Bw_pKq{XrRn6cgKv@IsaC|cW?B7Bai&5kEOf9&=<Q{)@2BVXyuixiog!c!P(-y}B
zg}fvob*ttou`*dXr1^e@Z{>=Q#};@g`@LMNKvMjiw7DF0mg!-@z<8}0owOTa8A$_~
z-L<_UOK;+5{>lp0d;I)lwzjskw6vu{v18nLVlk^i{>=sr>!R$|b%&!mLRTk$5xorT
zktA5tePD6>9duy})%fnPd?b{V1*W1O&}8KZXISat>{*OhCYZ=%a_eT$R+T8wFYx`^
zn`$cO79>y3RmRVGzEk29q_7?kG!V4BBg1vO=F`h|m~l?c8LIA3<~7a+p%Pp(CBSJ2
zLwtl68kSVO`=KjP1q6B4_dYe*#V+>rlAVouM=w>pr5>^G7hHRS+Eh$p_sajaS?iYn
z&Fz*@w1GFY+!WYt^Wi!Qn1(cD@53B)mgC2!*;p9w-w)umo*V$h9^v1?_1i%}LkPTK
zLVK93Iu+ktDL#?&ARL5OO}@y(xoRK5MlJ?Ph05zdiE_s9YPFQq@EsW5pkqGM)?Q-O
zKRaS2^C;1BN0nu`A^S}rDNSAK@d2@wD+b#&PNOjXR7+ZS!a%WY+V&Zt_$?zh-DUG<
zVElqEybGq_?n0!ZJs691b#<;FZaUQ>Z+csU&%>ry*HLWvBRDJ!!I%%WXH%F|vQ~b1
zH=O4=-eKc++q;i=zZiITJHE3q8{JrmC`D&}D(M>?7nfs6G~zQ~^vh}Vbyw{;@#$;6
z={1tpjsouSdIDZAJJX@HT-NV5b1W;$EXD!|&>Sw$oIoS{22D#rA+W>TtW|D{Ud=`{
zp~ZWQ)PlVQ7ocp0dVw~;=eha$77)kB>)g0tp?aJjed$QXRmSXkEH*I2!A2w?hw7p6
zxR&Df(K}mo@!wA@k6&OKvK~)as*T^$Mac*5v3oRStMW~meq&>!xU4Kgx7JxMf%g;C
zskOlXc}#r#_Wu4b_*Gi9%;{epSTK?p%-^c#DH&*>UuoUgpK9K_Zi-gX(l?R`S`qz&
zV8{~&+L$!?lCen4!oo>Vz>z2Q-Y)l<JN#^DXaFrhHH^W^kG`U-Q;wn~f@iBG--={K
zn6kxLEaS@8&ug|h@%B!;<!5(kX3|5P9vj?`RDLFk=01HkFYoJ^0nF==iABl-MR%#*
z*U!)IWakI|M5zf5XvMrB0gPsVzF7Ie$zU#)!y==iSHYd!DglT3{6}fH`fRow2(hng
zP&GmMsNp9Z2P!KMl01}PQ(!URX+Lo1hEAF@dH3BV%Jp$%MQ`7y`;T|%__tNi^OG5e
z0s#g8bd(+%0GDNd7JP$lFeR%TSBXd-s&dRqYZ7S+;H$S~MhI!rYM|d3-b5)EY;HTA
z>tC(DrT*Dby>_3GUHU~J!EL(6DQm^HxzF?~Uoh5f+@|?!AD1{%vC!<b_TlAn_vlP4
zxON{s!wjeM@bCas4=zqNtRlU|Tw0Z~OgNxIsPrXeWk0|fKIV1ADlXJ->;ugM;YNU%
zsYTl&>~XX*DEB^gS5e_AjBYRAN{{%YliiUfztxOexj>r-e)>n_^_Au3R&Xsi3NJDw
zZ*FWjYCVd6&xSmzqA1)7{NWCV0TTF?L`gVGs-FJaWr`El-I6UOvuV8!ledM%gK5<&
zthA_H=a4~FHQy`@o(hqq0~fbMvnO5F_Yju0#A4*(WQBE}l#_UHUCp+L33l$u6Y;2u
zJ6BC%le)ppMrk&|#+4-Lm|&8JNJ`^rYp-xJsFz!)IZ&}E*)1_@FJV6OHr`NKT1q;t
zOTfp~#_hN9H|2dA`NV)j{NVM;Yb4J}^ds+^l{T8$vv6nppaT5aZoKXThpPp!Dg-4)
zGW2;@w`wuNC<U|5^#TS-l~=fh{p>?c9&==#U+UlcIlWh^I1xQm>?N|Hhia}-?qDKJ
zvqEOZkCMoI%L|%b3p$B2yEvSZ{pEfxvw?fOwlg8%iX9yt#V0p9t|}WQ6f?z4Zk*2I
zFgYetJ<)3Cyc>litNH6tg0<t!gX(2{QtI#w$rt><(Tp^4`8;PuI2Q=zy(Yhz9O@>x
zLnhE#T;1KfXkBQmk|G764nMgDaAmlDH&`_Flg;J{F8R^OsJWV@;bt;6jYP>;S>JMy
zxhkxu-ZeHhQjJ7w6UBa@)o4S$z<4_JgrHs;=X+?8e-6o{65SoYcCmUY?HL3)b%UES
zlKSKZ^t&u6HI&XH=KLlD-$}}Hx+?-JoS;k5rf&t4O1sEr#OPOHZ|lp-J(xjEP>j8f
zXt=&QuiI(C0cF%~cm5gp51f=0OJF~IheXMLs9g`ZMfub8G*&^TbSuY9A)-i|$rw^D
zRVVltyi~7w-UZ{|TTCzaV1g^k$3EL4BYUx}>ya4+_)yboyTK1)2g{%+!AATX5g`tB
zZKB+gl!RHmXK0<ZyyJ4XRh8<(jE)UYGJR$%O{ZS`R@bLIZGj-`6!Cec&6i}Rn657=
zMshXVYc7e!F>#9DKifh8!;CMqAbq@f^9FeiF3y-~tgR57_&8>0i^Ylu8}AxdN967J
z^YKH`@k2E3l>*k_B0^)t3J3HK&^N6Or1?{g#97bO#e<U>45kt>1w}XLc&lSY`dOUd
ztayS$3?1>$_I5v<JWO2N9tR3=d!P*CA^da1|5nXal^#2OK~UPEJI-xHW_H2Iy@vlt
z=R4YNyZVpt+3y+JsTKDco^0ozR_;Y)K%7nHGYW?0`SEr(6u0q-i4p)nFT0^Ba_c)~
zfBiZKzY4UgmDxt0nB`a&t(E{R;y6(skta`{fYO`*MA6}3MG35K6frR|q)MCVuDwYf
z%h9>Nf1iTEGEr$u4depRuplTfw;Dh8$$?cCeJDVl>rU!x!Oh1$w<>xAy=H&JQ||nD
zk@G?}K>*kF?%kI(YowZ&*!PoaRe}=2Dq6$z_1Ek1LHkNYDQYe)E2Cp$3k8oA@(V%D
z1|7~{3W_a@A?_AKn4g+@4-yZpzpEXPAr}F{L((IoHjX#*KRqjVH_G~d0arrDz4Li<
zAILY{Ai=5kU}P!PyE{#f86-7C$(-Z!&0E<n><EHMb&IYOtJny6<0aA<o8@vMv)MV7
z<fZuNWr_{{7~YwTA0+?eA8wH6qs%_L19}|z1ie0@(27A4)Cz@=#O?3ztuiw+OF*^<
z+rDQR`Wz+}mLcG6w_G|Ak>os$a%*uSV&Vztpa@yE<l?*e5O8rmY5>vfzZ0tP=@SS)
z%x6KLx*!+w!E~1&lLX}WE+?ha+o5-A83ds$d@yhV0Cz2}uUj@)axgHw%T>#pUtdpH
z&CxD@2-<Fq)9pFPczn&yHjVGwKRCG0#>S@T^z-M>0m!ahTwa>bHVA|DktQ9<z#o&s
zZB7L4{L)|0b^_h;#iKOw3JD%#w{5%&Et<(L_WN2kISHSd(a=$AVdk64gI7jIz3VBQ
z_Zb*|!QT>OhY0T6d2VcMoZKxaATYnO5@%j!D*O4_R{GCkt|eiwAE$IL^6E{_=>iCz
z$;C=!kewE8Il<wEc-f&@OFkeQfI9Pk|8gaTg@v_ucgv1d<UhMz{gEimiprK;>yY-M
zD4}B4abBzYNE#&E=V_9~oLf6wcW@C3EQHDS9YAc_m~Bh}Ii=-rhw6W$nxqX0|F~v9
zBJyRaa}41hIu@3mx*d7UaGY6#O`GPgnx*<dXm=>@3Exz(9Sa>HPID(XIlR#r)Ld_&
zbfgK1PL&)aX1b0SGD?u(N=(|`P+FRs6N7e4#$%z{ur9Blumm#uQajv!2}#NMKYyM8
zaztQyh!!D=R#TJM>;7c<|F}xVhBJL$THYUCK)1nqWjd_<=z;@xv#`0=KYuFaQ0f|E
zMBBoJT(ZDd#b*|);wtWba|}WH_%D;ZL;!v}^B>vQcq#7?5<baRNGeUlzj+gpIz;M8
z5*$l3;|3N6zEwbKYI^5=a8B&|^r|JNj``+Zw1paN9}O4i(d0jz64F#lYd_ag;lAJ6
z@<)Gez@BxyMvyVmKo(Oe0Q+3GNY*!+hKA;ChEy2Ji$?EDmX9;C*s^Gv%Y`rLG$GYQ
zL~=jU!=_oiT+Hks7*$6c^JB^TdQ8yyy2|=Rx*?Nkn=Hab{t1^#C=OIQqB6bCUn&8w
zp(aIfJGWiTGOaA#e9~_g#6PZ$oLynA^|yaxAd4^$O+3T9(!anWW~Ck>^Rm%t`eWd~
z2Jam5hCxu0kcxq`A`%NwYJh)ur3qTdvheshE!jWZ6d4a5G?-fte}W2@irItp(twVI
z83X3-4s}tiiU5MZ*TgCT5J|a!SQ&(&*oIC;uP-7YLkPuRd7Q!O13gMUc#UrR%TMpW
ztS_Sv`)>!clFb@ZY5sXJeLyimo^`ZC<p&bx*Mc`!!2O0;(1BSCLqJf_-{oZ~h;cX;
zxEwZXf^Ydi#SV=i)-`ERf7=lLMR$rg{bxs<<lS9CxyQJ|q5<U1zJHEx&F69C5qeUg
zOtsMLhLn4-1TfM&k~N<}p*n%Lsa5A%!arg9^5tFFL5LD+n(US}Kv_=j9%_)8a0@=M
znAQBDRSF-H-j>3}p^Pc0RkTGBum0>5Rg!kh3YLzn^-Tvavep*m5Tu`!0%3gJ`{m8W
z$r}jJA@(NHW2d!SkXR7{S{&cCHBpvb+^dz8j67OFOiETdw|yRSZlh{BIs8pCry_tc
z`u;eP_$^=g-_PR0mFmBi{nQP>(kP>@erU!|^HcbD9nFgQW(Q_Py1JP;UEjuVC{K|z
zEZIPGp<A&G4p}o%I%*ZU%L9MU9(B%7R%+I65Bg(Z-$cpqK6(~sc)T@zSH$y<FUX~X
z`C44{xfM(?n*JoQuEJ9X8c>RZY3*hW@@6*;@+7woZ{1JJC1_!oknoNE{J9Ct6-fyR
zREPs&%mF+=h&O=7lHps@3ao^5moM)^YB9GjpBM+I>|YgCf3+Y;&$-1t-sLQNkJ?(a
z&7%^PNPJu<0kasPWdNZ)h2W{#crgu<l9CdnJ(MeLbiv1|ZQ(T}-qQ1T+8BS-=}#{q
z?ZUIwhjRMcufQR^lCQk)PouPdL&QDV;8$%7C@dh1juz;|(u)p5Fb84ALXT#$BIa#X
zj{j{$Tr~K464E=B8QQW80KcSdnq!iL<7}38#Qk|Ac&>cqNU@bd#wF5~B0B@>>%C$?
z$v=vdV-WtrNYc|Bg$y@@u{oNc;%g!)YN+1>jb<FF)G^1M`g)INq0J<|P^2sUIcX%!
z(*SxwyMyBboYrNODpA-Y22zp9Y`QfogA(5i#84j6d_Z$3E+=yD?U=HR_{%!=06N+r
z*n)yCY#>`9<W3NLUSum!g0b*SS(ylY-NB6a%Fl>QF_l783sWlXO@vvepfSU%2F)NE
z0qQ6|RK&layCHE~je{UVAe_t2gUjhlFR}ZVBEO(S<^%mY?eSEXoB3h@LFPL)<lDaB
zT!Jj>xH+j-^boM`(F7zFa7A&S)>?8T+*@CA+En;q!TO=fK^)U^(BuPw0b~^rQK==S
zmLMwkV2EL^s@4A&Uu5r|j&Ki0W}*4O1d)Y@=#?JIxlxCxmAQ78*q#%_N7rW2TUaz7
zlEKgrxjaxotc@@74YQ%Fn@ErV=L3_#1f1M}z(A6kjEbMi=HzS)v#G7M+Aqk33rNx9
zxo)MqcCtl!Xk%X-aNj-GWggmcTe)BuUp>0kp$R}6m!Qgrm3s7MULx6Jt6FaP$B!TJ
z0!|MBZ?pcp`^cl2QQTh|%Wi*0k_{$mjl!dtdE5BfJZ$|iK#J7Yh=2V^7U)+(hriQe
z7m549O8BBDAgbqba>KCdQd{?LrK6T`iFfD>_#oe*EFaA8uLar6@>mW;4WO4GZ<+w+
z(&~c*W<O$u+_?=2!I3*TJKKYfp<T=Dn-Aykr=<X9@&#p=d-Lg!;#4k)Yx7M#EE<;7
zCKp0);G#>7=clK;5IG=((E{_Eqo8EE{s|<-o|VOM+pa%0o31G@uMKA<Ya&Vy$Q?yV
zp23_I#aljVnfD%8r!(VknG_}#U&(1ghIK6C6fOj_0S*%6P~!U`pq-3xK)`;ipA-K6
z+2`ss2;rE4Q-((+WVX456oDZdto)-_ilXz39~Y0!e>OSjgmLB3%@D0hN@DR_qG-vq
zA)%ptLvIk_=COLW{nuv)tNvkO_@JNwayMoorlL}Stic<5ds;@uV2}hC@;)Il#Sm#+
zthIgiEXV^tfS{shhK%Zx+R^)|N1X$*=QOo4yH=&XIHvCSf>#R@6Vu80Ib0`sb6+-w
z#SrkWT(VF+P+LI4icqJdpFih_#lpl)1-Psg3J&Hj0B@BdJz)@!G*oSw6ph)%X=KR!
zk*Pch(S2?l6%y2I>BKr+<&=9c1Tfy${ly(}C;WW!Ev!U(lM@9lI;55rpu_eogREdS
zl1l=YL~AnUM1WEVyZ-q3o_Jel864*m@TDJ-e*z8B0m9aLu7A*GJokI8x@W82_yhBZ
zqkiyTNOp+(e%8#->rEcW)aog2eBvd$#-?GR*}CV+hf1XQHY0(zMy)8J+=E`?n+Y*&
zas7tG)#uNj8$g5u5kGucZT~Ecq374HU*jM+WO?Cj{%^R|2$5zL@%jR%B^FYXT#!e(
z>z^002Z{fs3^30Fy{RW6;QYz37qv~!ACouR(P1@EqUwl*C$d6zyk_HBobCs7$f>Oi
zW`On@N+lnsTV?n5;J{Hxm}^^Zdv(=(#!TB5;vuhK<(JyY=wR|Djd@IlW5ZPQ5M|>O
zd$a)=OKE4os@SDVM5QiiaI*fkxbf4<-sEja(zgRUL5x#CD84lF3)6S7wP!0sKof>G
z2VEMmJ>1=ETF>dP%}LUu9{xr05_$XZ(_ENz@wQ-<>)kK-{=w#-9k&#;_^7Qj%o`g<
z*NX^fb#Ch}I6zJikv+Feihiip0_r$5HH?{fdJ4gmN==1R5y;)?>FMLu4%yGvV!3aM
zV~U-d44jGX_;fdQQub{G=(ic;6(T`z4l*P-{OEMX6wnb4@Ww?k<!#KXd%V0cpl1`2
zljlkizy3d<))W(ViF%S%%$Tp_B;v!XxOAf}7X}<A$0=%PcMJ|C%m#F_Z~}`FHNfL|
ztM_0Wf|}6(AjA-IK4Tb{w6I_tN-b*R4Kpb65YPtmZ#YNEA6!!ca&j4PO|WorjUhMw
z_3PIi&>wG~)h{r>oC<uT{UkK^NVuC4x0$f`Q8mbJ-A@`BeLZ5KZ#+#TzMv3`WC^s{
zUDc@Ke_8{HkHl}2lXoE|ZiHMpKtWoD^bb&uP+;upK*mh|)Xo^+1Ss1<-#*(Y=iLRx
zIDpNq81#X6%Gji&q)4w`z0!m@k+-+E<GT9ZUezsLK1qcL+u!NC3!+Gm_Z09GH%zMZ
zgO%A4LA!y=?o(0ZB>fx5xiQ<31ew0_3(f8)e}WJ_90_bGY*3OQ9OVz0{W#E)SO2fN
zqImGH7Cy~dogMFwm`7eW$|!#D-T1R(``NqfdCh}>j%1LfAWda9tCxWAP?!)CNk4z?
zjA-cZLb7QEu2y`a8*uxZ8m3bIzosQNI!lDUiAjF5VN)!AE`_If0V{3<6KpHl(^-6V
zOeI&iLr{>Tm^JWxfh0rZ%LhmylvEx>3Wr;1t%DrsRXjM*2P9Q)rXt%*8A;oxB0|+f
z$x}C*jhZtfDj3#_;&-IEG>1u{KI-t57U^TA3F_Xi9s|4qC2$9%_?>oYpB-ocH=zBd
zu9MSSmL1T#0BC_N2SGfNr#A7wnMVR`w#AnD=Ek>+JqQ!Wkp>gK|CqRO5hV8UMb67v
z_w8rF%3`3IKto&vB>DgZRe%k}uRm27?gB|gXucridHloQ<sHI2o437fEk6FvnAr;b
z=-*1Ti9;jvvHMZ+vE^o!yGHu2wSsc7?jq{PRhk7J;^VoXC9e}1AcK%WNI~%tB4N(X
zlRbX{DnSq@17fkzx42kH40L5>v;QRIAtNU01Y{f0+Q-a2?o#R8TswMRmm$banPc7h
zchh!S1l#Z_g0W2TRUwl21s0IuJ<Oe9cn;F0M2QK@jYAJJDP<N3?+i2@jaLTC%aea7
zQ$JboP$4PMo4N%NjHXUus++(1MwkzQ0P~%9ynaq*GA=JEO&|qs_=F*{g4eR2mbKh}
z)!72WgVgLd^%wjg3s#f2IgHU(uTf=JWYhI%0$L{~o+K$uhT9__d1ZgH(0%M|I;~D3
zYDR@EkMZ_n5MX<Wk~G;bkeDQBO|+!Pq`F}!IM+EcIsMKg@g5DfdLs%Kc~fb)_G`EJ
z6N34M`%PUz={*Z<T$SROc^r3|DsKvLQtLg&uFvu}C4SzBjNv_8u^B&va`okAC)Vgz
z0RgXUC6Xl=N8hcdZbK-;!F~{+*gkZo^^t%PEuqL-76$_HaiZltpz~Mz4?-p}S>O3T
z)v9tt8I^w*f77t@x><C+$@nzECxB4_v|X5J`d3Rw@<;)@KAJVrfPgg6*8(XJyFk!2
z@XhT_>u;aTQJ>|1QBEIc)BH#OPlD8+zevF+NJRbGwUy%Is32g%fL5_VGm4(aCQnG4
zpjIGw^_`HWIxtHg4#yG);Aqd1x7b5oP9<p@*klT|l1TB%&*3Qw1$;j`=JMY_5R_|v
z|C3k1zb;L5adZT|2HKMhn!nXC@FfKyiYG<*Y3sMF?^_<z2A|?wojw)pa(~kFS6ggJ
ziFw~rezn^6oKE6VwyZCBEUdJ2bRf|7I(h?G1%Ubju0*Nzmo0#icO$8s8&4jiIq<Kx
zUO|T~4Zp&a-C0rmi#-3BpWmCG+HK%(_Wk@?Te0<K|1A}@nFkxT5(*-k0%j^yO3=bX
zl|KOB*qL^jeFWl}$n}K-0<jy1*8*{`oqaciXF|iuzrOyj2XL$~W0gw0PnGrMv0r!&
z14nL`hhx(6@`%IM-rgRB7?WzJ@{j||LR`rvn(iYUfmUrYj2#b6K~CPpnX)?(Mh>oT
zQE-dhX}MQehXT{bhrIF<hlf=oA|oMZ_aS*?@|G-+Rpx1l=`n8%gV}^>JC~BN0viy~
zYKRGoPyU?8;vh20m98!U9rA{a>rT}l5$~N|K!6;Hlr)xz@+Aog{ZmcIY%?5+&vb3-
zqwp1KSFnp~#z9qt_xBtk^;md#OEA^2*b*poMMo5W$CwaG`+es1_)2>Z(jNzBJNpiH
zuVa57(tV(Zy#V+W#6@ob=Zy}~mp~RwbJ?EpJm0MRnw15X&lY%m(nDGD2y6v|Sfg7C
zxA)E^fO?;xTe{#Z#YMCsg>KC$k=kULy&YZFHPF);ztXRAu4eRLmPlJ^K*uqFR))A7
zBK&~Jg?;^czdnKSJD3XgJ9AH9!ZBfcOxxmSOy}2}%e99{M-79~*oAmCBZwF>2#%XJ
zPwwqc3sw3TKGA>jL&VM^!5QP#Fm{!bu18@=WQAFSTh8tL&YZ^LdkLEUPV6v~2KVia
zo!uOql^C=1pZ^-=`-L$Lob=yyTqy`gIKA0m043+@I*>F}tUs)&tmL*CtRs#`^KbRy
zMdIh-^h_<htG5kcYH=R`H_Ud?X~X%LsrP`G6E<kbr-hdPFM?Jb8WezkVK!s{LiY`)
zO9_a8(}JiGBx;(A*h55frddx`uD^@6Q+J45M?0Wlx|3h+lZO0GH3_VA3H*57G|6Wm
zdE<mewY%JeljPEakQSq6q1ghu5{$-u1?Vhvx)6y-Y=X3SyQ%?ZR6#c+d-MwR^8C1e
z?;?U;7NK&1Rq+5YH<X<rfQ_M4!WIqwk3fpjsVv>{FMr)j4ru8Fay$rK5<<k_a1VTy
zCjP=}tvx5Q-iv~$bcTW(A*Zsf1{`ALw<ia^9^QSKB~Oz)l_5_DLfBrzfF~pBh5jfZ
zC(VaAHp^l}O-Xj0$AS^bgB2A9SOU-h5VKeN)&EZ_@Gn_*#6m;z0TBLwRA7pwcA&c;
z7V94KXd>}EC`2_lUZ+uIS<kVY28U<<C*=T20eBUkBsLS9F%N4jEz8&5`6|Vhpel8Z
z|0K_c9AZChvzqgNATuB6W=jsv&d%@=L9eP18VQexXc@`V4vhvN#%nb$2E%XmETT{(
zU?{Rz3MvuKF~go%TTuW3Eq>h>vF+8Jc_r6{#us#VR0ZzpqRrlXHz9BE!^3Vds@&tD
zOgu`^A*4*^l3Zq*VI&w}+=8|<OcJ44GkpAC`6$na$;rtuDR&>l-u8BcLyv?=uR=rx
zIELE4bX3%;@aqu7q)9R1QQzEkoO7aRmi6O$K*H{Lygrx-qJ9EN<u|j}k5Mu<qG(w1
zfYZU&>~+Be(O;tyk0cOMBW1R;4MMR{mLbkm^8A=KiXQN~l34WmljGym?gTsI%-dE;
zA-=@x<v+?fQH2WSY1H$LHfKH%<-*l~ewSU=X8v~)$un#JKS?C5@0<V0kM!)9#u#us
zw}aXG8pgkvpUy4YGtF4vtj^fp<HxJ;%%|j;vV5!lL&RX)c$4ZQkwh$yCgdPFip*<t
zKxc(-2Qsj@FfcG6YuOCBtlnig24aZMX2Y;60=2`a3+x{IB!yeB`TT_}%;~iIsr}j;
z0x@d*oAItLiM`3OxG0*oiD~{6BbFXhcZ-&r{1c3&zn)k)1(XY^DU$>?8ovKzf$sW(
z!1f4If1sKG+rWTPTmuNEfr_*@dwmuA9|5=*U-Y604BXQE-cnRq6MKvt9&>t`=FqX4
z7nz9{wlWYpXV?wlO)qzObU(Slp#kg~^}Ip4xjBpaE3CYS8!ldHTe_eki!%i(e}KWq
z`31)Z_ns>%BD`cRzrs4mD`{k1B6Ho23wrmi^#3lSqa#%H@B@sTFW^BFW+d7gN&U;b
zGxyQ%72Spwrj$<*m*97!!~15OVhQ30SjEn^&mO7d5V8mg3gT9OrRKYXi^!Re>3I~S
z%8%cd4K)}^{!SF>M_J^Y_8|eWr6j^@$nnn`0)_RPGb3I4JkH@RMZTSL?BDiM5%_5p
zxZujT{>^&d$dqN$8)us2sI~#3e-P1sAu|Y)6-W;65V=HtxN|V42z!Fb?2vbPi!k4)
zJ#>`{A%5c5RbWT@xhZmI>(H@>IG6f_5{Td8kHmU)^(!Kd>I&`G(D(f#wt<{R-ayR)
zlDk(VfUu;Ox9UcFG2A^rUyh2Vj~82j#0`cCOmQs;d{K+ge1&QWQY}OSpIOa2?ep8Z
zJYYE9&5$yh*n_}$KEol=#<`yCvE5an=SLegnpG=a_Nq(WxDvN=#uzKw$a9#O)2wkc
zkJ%T!I-!H$*sqNDMArXDHJJGj5&@`eR0B&|u1B_aO3+b~kup^9nn~g`eibI?$Wwn%
zmIiJGR2$4ZrtVY3I3Z-}g})SI#XKMSY@;l9%JXn4aFc8q@n&nJV;XR10@TQGU(jKm
zd13o+@wujcN7^u!P&$wszXy#Mx~UR~(=5wgF!bq^yRdP6>-KF)cq#<c_u+y*$QqW0
zyquOtE_pvTLFgr2$r}iZsjO{3)8)m5dHXO>)U;`0nloMP!%AtZEAU^$pa#&F9!GD;
zG{LZq5}>I3|4Iu!2p~wujOsPgzUPFK&S&0!R@vOC+fAikOE<jr)p|rlW|Tda7iiGa
zcesWS!2>E33bT_016v$-`E2fn2_MB#$&aJTE;M&paw<s+JcQo-6G729m|`}{NXfr1
zr)XwPe{$tSNxu$Qe{Z+Iw~1t(=D8XABJalV7SbY1?P3C>06@k#$Sc4D2u=Y|F{wy?
zoPdK5`i2n%aUVZ^d<)b)iT+y5A^8u^)K>S{cHI@OC3fvy=8H50WV-;Q7t{ZnBeM~$
zI_Go6b0N%{^&9uHv(Mm7cUs#PT&3nzwaG!#E*JE(oyzC0ln5_*&wY~&2uxdp*SVdu
za~8O1pt`N8mCFwb%0F$>4fwn%s)QI=kRG`Q0u<v~tks@@S|HW-tmw0++7A=2C=_gd
zNmji&4qw%2G6=|JwNZTFJg{dI{pNd90>MPdHzZ~l0*8T6u<ebgIv>8jkB@k|2v<F4
zb++Cw-V_nCG&J2<1!M1-nP+*mqnm_m{cQ%>-Srl<6ug86Fz?C<gBA{bd2oJ|YMre?
zS9%KNBL4B4SO19;JbW<y)okZ%`!_d}2pN&sHiHN!D<kN$Bhf0_sU%(=0nCe@OUt?s
z#S1WO!|a>_uz5Z<gwe;z-zQH3%kcbYlMeD?L+|>b^|5G`O8iUO>={h=nwd744Wr~P
z@jU!j!!xx`$)Sr2o$=S5;D^khPiHsZm~KvE`KlN2wENH4o1GHT9vke&kU;<Ldy9VM
z9{OvmoPJ<!b_2#huw=yC7Hqj;c;?98_OKxs%Y|IP!Af3l@xtr=n=>a#>wNMTr}htW
zJ*rl;D`DP{i1sooMuXYtax|XOpeDLHL?d!)=zS9KsmE2?nR%^h2;i&5=me1~u&^T*
zJ!NEMzD=aOU`JnYM<n0zGlYlD3MeV{-%@niSxjvn>JSq65X|G5wO3RdnBtGz*p_>6
z#*JarsmpAQ<5*n-K`3K*lQ5%;-gH;h3vz0}mL<a&5RZfjCvGp@5!2X2dM@H)x;fe@
z|JfD$RfSo5+i_Fc-C--22lNt)=n*@rrf@po#AbqO46^B)^-)%ls}m!Hx&#xx5S4|x
zqD=R9zhJgL38<J%QfYKrC(`B%Y&NOcc!iJ|)G@XnC9~x>zSmN7Z7JkWBmd6RDm$k4
zV?$PlhaUuoge)NDu3&?LKKl-!D<TH^h1)#+AF+e^e1Bvi3rnf)2hGNQCpTQOh>a@D
znHcacQYn#0)@M37^z(!tYZT8@FeW?p`m6n#u2<`W#1>0wNBv+zp0u>A>5wetZUG|>
z5!4XP@d64v_x7a0g0k8t@moT+#W!rP1=lNcE78v*8od%m_6+Fh>EkKtW_A6*mZpEW
zn?t`%YbA7w4}uTvd~4`%{(Zz#bb@J}g$b{w?M(yLTLydjDJaWZ56On!&mg18=3D&D
zcP2FiBQE&UCwOMfLwG=e$XG4fP4pZ1LOVz)-e*inpun3&!cqkkb3bq~8;}(gn;JOV
zLfhBqHxJ9;Yd)J=v=nxQ)o&H3W{(?V-mm|}3^`E9`a+}^F|TY?XabKDXa?Kwj<7p7
z%*!*GMP`b(NhvQ`!-pJ4<+I)KkrRya!@2&ARh51{ELn$lolBXz*&6pKNi*+8)P9|c
zMui98nOGS9R-=p2BNb!IVtD+;3UF(<BM(0;Z81ZHn2=7AfSNjN&};$htF{pyTNV45
zL|QA(Q~bi2s@67AL1o!2OvLXiJ<yqC7<s^FfjqXl`8heJT5j>3cCpDeu1_!RIzh{3
zTHq7?3DwUg#P21M;gJ`J;XD|nM{j~O9q3@=bw|0dOuWP|k-;@jbeZ*vg{e6Qm=3v1
zX>A;2@9+hk><7$U82$}0N!7;LblrLOTsCmX2eG%NF_!Tyr9y8R_<}qBWSfXfH%Hdj
zX=Q-s>}XRKY)S|*V9ddwWh%tR5VPRiHorB)UOPI#Pi$;%B9aEaZn-LOBM1WL=1`mj
zEHT8^=ueIngD<ue`U<R=o(tp^^yZE<Jg>1ZudzB30%{D+`6oOmWgL7ZkE4x0z#F+C
z%Zo_$&K@FiG9q4Q5H*2(ssRLkwYmUxsg%8zhgp71Z0uL?$N;y(>lB}>G<fg@yfg^5
zI}E1{13ywmsE|+}eze}Wpc*AgUV3I<XL7C*0QT#LkPu@C2dn1TJs{MAjpd4PJ)m0t
z_t**Csei+o{p?z$pf@vc_*{8{b{{9~!3&X{Wl*-4gb8n;pj1I%DrT$@jv0))?Sl0N
zk!opZpYQ`1Fuauxrq}6K-VEnfTND|fFaml>t^V`aPQ~h1Vh}adt9J+BJ_LYWZkd>f
zC<m_5u)aP~_5iGip50Se6i$m_M#K<45<DfKAN1Z+FnQwONfW7PUi{yHWkE6e3X=T4
zrzUiCblkm>4>p0_-?p3kf6o^$5p&%?bv@x(QfQH{I^Tr0!R&Bh0OQUrLuKWN(>Pvh
za%dJNtG_ZKX!H&ddWGB>JX)n6x^=3lkF>0;5tt!BapL+XCMF1ph*}^Uwx^0P!NE(I
zgU}LKmXKSMgRblFzsQcH!D}oR<)99Nf2$9JNSUvYwpiX(2Wboul{HafKk{?3wO7i#
z1&=kVvHT#5LjVTPQ+IcFnp(H^KpaxQx$Q$-`yA;IXGwOjie^%I2QI=MY`XNQbXXyT
z>-C=)3os;(|Iyo-M^nB2dmQB?ozkGvd^DI!Br1-rjwm)VmU$?IPzYsKDixKSy$K<*
z9WrKY+tJ`C<(x`}omtq%jN1^Ye($I5xvk&2_pWu<x@+A(8rH)8?)`nf&*%9J@Au2z
zLsV*>V*~+n(e?xeV5&IP{lp{N&4Jq|y<vZod;P%O>w;@-T4b1MS*$kst`;=N-90`3
zDp;AsxV4g5#Y5g?<g=WdsH^W=wTFU&1f8uy3s0SQwih_t+(?lwfN*M^lINTJ4L(Pp
zV}%lNbTnpI9x7dkSE`a-LoZ%jvSP*YLn_BLms;aEk+}IG+|+Q#+#Bb*=0fZ~JG=IQ
z(FNa@teU;?y^*QBWPDA!)hG_+@aUtKP=~$VsxD4(&z?8Q>I<M+Kc(Uz7i01$kOF~v
zb!jQ<;y2}i@tt2caixkW>ku9;FWqV<_)&<LzG&}-r|E<BL;6YRQh})uhN48)xqch^
zxPO9zaSJ9JY=xt*6DZT!GKz{ZurP4rgTIScJo;-Tc>p$g427LZf#VXykK^OL)ff=m
z8kQ=c{TOysG)730N0$*9zivE!j*w@lIlH~-w}=SypUyU=YQ<h6183w!dFTOh>-S)o
zc?`bm@C|zt{*K$-lTilEZ#9%d+|F?RZipPVSv-gUXXG@FW6ZX6IyyQ81O#kQAe*iz
z8_1IQjzwTZRj>bP`-PmDr!ny(6L&{ySG(mLA57rHZI7d%K*@*s5`MUdK7S;**aDjs
z6bf^7-U(}+oNOLzDYVcyWIOS3QhwxZF?wN(lT9>A=UR^kFF2TFKoJ=YjaptNfO5nt
zT2giQ?%mw+mKMTCV0v(f&O^9jb>`1NhT+F*Jw}6*_By>dD6=*M@@BKEg?@Ay9Nf^W
zbi?_43}qh~lfj8V1oAtvoR@xHg)??n5<UgzR<I7DpoC6y#*;T3vxLliR>8&LzTH_Z
z8Ok;G3+oZ8c0km$X4hX2zb2NJQXs<V#;dT&5cnHMf#ypeDDT|+IJwNtKP&2KQk$~t
zZx654FhVgq@`u8OxS{>0+Q>31IIqC$3mx@kSH*NfD(<bow!fO#f6nZuvmr|+gM^&h
za=mYisj(Nosi?RprV~q=G8PsV<C%?O!cMV2X(}KCS9R?73JEPQd~_;qfw>B=p_1Cy
z`SYSq1jV`xEDEy3sr1F}svbVj1Fgr4H57BM*<=OG`Er|=ozv{o`uD*my~doJx{|ZI
zn-c#l5dC)hR891o?(2P>RC#)4o-vQ+{gHCn%QK8h7JVHU6iJRl8&QvosugMXEoRAt
zil5T`oSERbtmfiUn+A~(&=T;)yp{ROBVTb0_Q9K)b(^jkY`_1l&eb)=1!;a?`?jXU
z*@<HB3LK4En{!HWK_pMrflL(<Etb$OH^swR$&>w8Lkd3G%*haXX2x{Dt6jJz5%-)D
z$~Aj+9-a4jTGAfY0kToh?uhOh^-bl%u(Rs(#QOPpB#q<wrWRWpT_aL3=;a?0mS@}$
z`3(tBN%@WhC^F5Fd!;k2K*dZB-~a2~sH?n{tL<6>n>z!(zg2s9z3f*pOXB#BEz8N@
zG-FWhg|h!b?qS5JzlKh>#=K@tJNCFD$KSlV@E;pI3M*CG*}#CL-%dfT)^*nPZ~CcR
zop1~L2%f5!_h_fz7T(6y^rhadN;=|)Uid*dbw~Sig~}t|XI{E8*S8KGf-p+r<w?Y$
z#mS4meavQc$M5Exh)xd*YTa^2@%b7ghIBl=Lvz`B+xp+!%G<rN;Yy@g-<6X-JLP@m
z7#!lh@{_MPbaDgVm?M+6Ul?_*xB6XNcA(YV(&Fn2B%L1C<m|XHuRh}DR&rI5NJ-T-
z#|n0~JLa_TR;S!!%{3<LykiSV^6(-*P>GpOq=Lgg@gckS%=uJ5A*{VS`*(kZD=C|=
z1d)5{T>0_3OmAth$I7F;<aom9-JG?CyhLs;wwxhOSN}(c_%ivR0VUKQ<ReFAds&Mq
zl9E^Oy{Bi!omvd`(dK$s=83bn`2G0DxG^W}dshWXMZDz~2bvwYBAC$$3jf_58W(4t
zKD)2&hDG#;7hW$Ce*8o49_SbvvTA3)+_(WB^xtuiE@b`;IVkK()s2?-C4XG)Kd;Sj
zBV&SZw}7yR$V>cpgh_jTNPwZrStc$nj)ddgecOZCoS@m!gNS#*zyH2WCszI_5*78@
znq+$70O|ce6q{Q9;zhcEpSe6aj1yG?Mt~u)k<A61;8U5{#*OEZ6Lt3Xl5W1~1Gl7#
zPo}kFewM`j<8K8M0GiP1HSagdqFD)DtvUi>8eIHP;B+F3nF_Gf{QB7-V!*&G@DX7p
z5y%49qR^@464ZGVHk&s7IxJ!HW2s~RlaivcXcj^#;^on$ObBCm&43{mWWgJVqbuBB
zOK*L|<}c_vgMM-MCNZ&80Y3|QIAlkrro@q(>vHbiz1!Eq+Kmi<qnunM7%L>u!k;9x
zV5-28GgV@@2Z*cG*V;(Bw$+v=7dzF6!Ibt9&Nx$+w=`gwDaaXcs&##geLM1r@e7-u
z`g}MV5}r}OKJua41;#)fmRBFllodUD^@=!J<vr@gRN-{=@|NRh54@d_kd4N6Ey}z+
z>AjHZ*TH?RO{5&PJ6|i?RtK&qoVAK3j6NodwNiiBWzIKBopTJK@u5E`QNzT6mw}F|
zd9Fh}J$fL|z4HKn8ZbvIH4|NaQ&m0BSBtyqvCR;-xLI0;&5-{R_(Cs9m45<o1cy%+
zPO}e%34A6Aa2xO8RvitE%hQ-ktV1R4(o`hf#3p|m`UK+&Jtk#)q$<65H4Ytm#%4!j
zBlOHX%6^+7?{6gt>mCnWUnSmyn3I@iwyX=IYHQDZTwcEExg2AZ8yP*o#~7Z{^7U0B
zD+`efKCoC5`I3RiQe6F?#`7Nzsa!Ok6|?z?H`jQ&BkdP91NK`2!kEQO+8=W+B!=fh
zLNFQB*7s~U&KdX7q*6Q3OFq(1pYZYX+lh1<@8KEfO9VNDnVKDvl=`7qb^IVQP+bV9
zothuGU*}NEhof1Xi4t0Bn+&{9caY`7j5!88VkkxdkYq-?HoTqmk=)ks!4%W5#)22)
zk1_$*S-s1plGq|!eR|eyoz#ApjMlIr(2d+qeDTojJx@>?n?QWI-0%1*m{Qj&pN)XS
zu!?M6pVIA&K0wQ6Ca#2sZ{D%vCeB#sHeQl7`Z-Un^@YHZCFmq91sMd)CGDQh&Q1au
znpA*xU|Pz0?eeaY=Bn1#g#S=tYI)v&SJ}ty=<dD*juqY4J`Tfct*#yp*tDI+ZGpDX
zR~TVz$<AZEpVGQCYZ2{7?`o65F<*;$KkT+-V`3UJjgeQrw0W$GjBf-_hQNPK7z3`K
zZY4LsXg$dV6{BRX3wl@%(J>LgG<IZivdCz6*u}RaKGj-~Z(o^tD8V>8#KZ&RuLxqt
z9k{Tj0(6`{J`~2=&yfEd2!;|?U{g~MmF$OlD|_ok|F55?2--^8;(?yJx_UZlG&6$a
ztN`|sL0hflnHIapJ()R+r8UBMs}@O49ppK60CLhxOG}m9+Jg&Q&#QM$m7uq*{uzC{
zen_8ApvdUI|9rh6SYfb7?63D-u)=);jWuL><PU!u3ZZ)-p4<bvk3lm40L=(y2!&_}
z@l!n9-SvTmi&XOem@6nG1YC=LmQ`t^*{LTR5#z7~3<A9++!3$#$Gi$BoA(eHCjyc~
zR`DGcaAR8`<XpaFNfObyA3Sm59#h+C_MdnM+rB>Ts3wMY85o3Rz|Kh+rI5YdrAfsn
zuB0F*;fG*SAG=%+3nnEitM?gMTz3!rw0BhZ@83!P`aX1n5&Vn<gb5+lO*5v4{gSB%
zj$J(^NReAD>T7BYk^E{6Hsn7SEG^0mALpA|zlyyPC{R*$S2@EJN`Cg9C97y=tV$;A
zVOvyG+rSuDL21bLC2&k|-oH=6z%>nHC!6?J<{IAYIhEsFnQM=%vHsMm;b;`I64ut%
zoc8uRgz5>W@UcbN-#@@An!&x^XI11p6Q!z{LO_E5{htBHm^MIJ2>Jm0*#=#%5Wry(
z^t%Sln&81!-?uLqSTD77qh|!dOjE}0-;5z5riyW|`RC1pA0~{&^1C@D4#||@aw%aS
z7BoCsfH14wB)DRfT~e;*8J41IeOnAKK+?AmJq`?eM)5e=%8I~2OW2pDsRAp(8HQGJ
zda$@H4Twv0pMDmR4aWSO@baJ_Oowg}_3su6rN!ba*tP`sLTYL#$U){L(n#mTp@fZu
zK8#eI$mSqCo(};NHe*APtNHaDW9lfV0Q&H_p{m3PgLJ#6`)ly{6M-5o2uJ)mMCH@d
zWA@wuv?{c4O1XJ>B%>z!(9EovZ47rV4*cwki7-Ox0w4t7VU#yqgYJ+DA}nc~6L1Tr
zSq{D`ILC1!5frg?nKEXMTg1e~*v{e?B!1a?Y8S9UM{%>0F;BqcaWw-^^I;d4%o4AG
zs7BEE_-YX>=?2t*X-+gj?Ri0nizn_NdAJW_X&|uGSqL-)DqW4w=;{zMYydWU_Y7KL
z>4@ZYNFnNQ_|_fV5RifoTZe|A$A@L)pv$HiNGU0q@e<fwKjN7oIEaUVF4Ti1vmZa!
zQ1%64B_sIrap)eF0zVJTS<8UQ1_z&^8VnohIN?`Qd<)Q%F`6f|-iYicsLG6~k_z4*
zOgNpL#EZ(n=sN`)>M)u$wH0A2=>{+T2XO)$si|Ec-t`l0Wu=0P7VVlMln4@{NME)e
z66O9vwv{2;wT555e4$D*;3!PTM7AzWrYsG5xoostYcb<HtgWs6CGfH8W#yZoe<!hU
zwKQly?UkY}gZb0YhmNaK=Ki(wWkrQLc(Zl>*+^shV;mCnsglc=ExQZ8u#(@TOKJ(r
zvo9G17R@#&Z$U1Rhj6)}i`E$gBRaVi<s|N~y#y>@0^tykSy;`;+f^>ELrcYF0Fmzq
z6yW_JLGMV^($X>ntFa##Y9ne3(;H&zozbRD4+dW?Y^XE?Dw3}Q6hKa5`aznNn`Z*T
zs8R6M%a<#QHmC{&5Ekvz6R7XfpuL>F|HmKVnB%12+>Cj(W%FhpH-tg74AYHs*4tM_
zDQrNuxdTCX`UDFqfbK`|h0eLv#AJhI%LGaS7d#Fj;Wp?;kw$Ttx?oRuqLJ8CqKh*W
z1-?^LQP%Wr6tIxttp#N6b%2WhM_D+>VqPK@dp3Ipd`dzfK#Yf$RA@uc><(mlObby_
zwWF1ACSIrmHN(NzMOGw_g1!e2Nn9M|^N^Q?AFxgJI5W?F+O^3>mIv?27;zFjQwiJ;
zTO4r81oHwjoJ(VYq;7zRxA!aKw(;K6-h|5Z?E1PY<UXe0#t;-4x^N6U;w9s{0`8$?
zcInyE-X7s}w8RR+kpoD02&fNGpPHk6zw*b&$CG!O=sz|eF#xI6HISjY(sdDH?2!{R
z)SE*r4=C@uJo~WeX54}X;OC%a^ck{7AFKhEtJKqw&Yp?#K#Z|{o1*E`QbNZVoQpWr
z&0Hu-5(v~nXRs=En5vLqMUyrs98b&sEUb#I^5|SL#=-zc-nl*;nu~md2#6@CFv)m1
z^;jVf+YGTD1Q1wcq8J7df+TxrUbhzOLROoM&Of<3vpsviBBq=}Y?A|N03${s$@KvX
z0e+o32owIuc_GXht2NZ+FmE6au?41+_?Sqa52{W#oIs#<E+p*L#DK|FQ1?_oxrAQD
z7$BjtU{Zux>=dQ`O4NHLeh8oQ7cQ(o4;uG-2_OOHEpYE|87}*@UFP8kA7&j4*uM%+
zw!}iyuTi{fhyXcTVevs1Y9AH>rOkWj#u+GEeEJP>W}4(5Ur11|c(ee!SSBp|C*a9E
z(xvBL*c}7j$)ePkj!-*>=+d=SvA_LEh$1-j$iwAerR)J_*B<>W;spvWY8^x<*k)F6
z8Q8dv!9g7p6Oq@--0)Y%S^~T_dQ?g0Go&zsuIZ-2|J!c?_}`wipMtDC*XQ#nB;Uqw
PP~g#2KctqeYIFWyN}6>x

diff --git a/_images/sphx_glr_plot_ucurve_001.png b/_images/sphx_glr_plot_ucurve_001.png
index 8bb0d05b7832c20e301e7a7e198d52dfb30aa7b9..4ab98259a623d8d9b6eb18d60789ed3b8f0b2828 100644
GIT binary patch
literal 29790
zcmdSBWmJ_>+Xi^(25FE|knV0!1OZV%P`VM2?(Pl+B&1V7ltwxajYv00w{#qu1I#|Y
zGw*yqzFBMj%#UF$mMovMpS}0J@4W77pGZ|@c|7cg*boQ=Pf<Zu4FW+9fIyIzG10;Q
z2#(ILgMUO_<g{GW?af`>jh)OOuZ>+CZ0ucZEKMG{nK?OI+S~DS33CZ@Jo@0`;@~XC
z&29VN4{+H#S#Td`Wod(#U^ysgJ3}A@#)w~}@6v^q5QuZXqU^J`9vS-!?j9r?i#^AO
zc3uj(OiwYf^y3x@A3Vr+M#WJ}{5HeR9i_D#L47pS7KD+{&W34Hz%o{#S#ZIn_+eX8
z{E!YM5Hkd#$TZs`Jzb;ec5Ux2G%K#2)oV@>78n73>T#PU(GY(q+^IoG2nh)td9efN
z!4KC9Bvn~i*{*IDNEG686lhPuQ$q4Q2n%?+YC-~$1CRASA?F}Qe|+=*|G{Vzr`zL|
zvo_SCPF*`{vR!>(=ORBouC$w19rx1t>+vq|;L1FI4nE<sJ=PtBEhpQ>24tl?G&h%C
z4!ygMI%d}>S?-D?_VD!7R{atYp{0nV+STW}GZ883y!kmRE6WT%ZRr2{_WF!fzp+~G
zi)i`jsf(JrdMOu!JP{F5nJ6bO?{U(Ue##;Sp(c>?>blq2PWNn;g?6)t<7WND#M{lG
zOr^ucmh#3=ohwH3^^Wt7y)+))-o(_@1H>v4?N_j^s)38OSn06_XN%uSoF9{u$-l6@
zTivf)Vq#!m==UkrsrmXM`r+Q6nvYYZx@)_;N(u^?hpjhuF$SHh-5)I%n?>b5l64Oc
z$68Gm9sjW|H%r|D8*<s5il&!~dO$+brNrIR1vVQsXl{CSvXKf44G$lO`rotHjuyRR
z*Qv4kvLX3U%17jEcbb`p=Q9{I@>LEeiBmVptbsv3#;9u6<No%7QrM0ftgEBH|D%4B
z8&ZYEFfFaeGEx*dcT8QKFy2Ff<-%;gMw1(H_k;C=ro*1z-f*Xl-@H~6EY$e&F^kz5
zc3Z1vNC;_f7iKfDutb2TX03M*1+6MPVde*oyIelLzWhsT-@mia`kuZC3JyL#KmX>o
z|7ReDcP$J^msM-Nl<TSNiH3)V$m?{Wjdc6gT+`kd2!~$%=adu*uk*b=aN<7Qd@2&h
zsN;ny7Ht7Y68CTzwUkeaiu7$_sczj*uQMwN2?+?l>gClHug8%Q91hoNbYblbCtwy5
zO1fC|=I$38&X!gQ$ET?WBK|v?hvB?6WcGI?ccMg_zpAWww!xWMztJUEqg3Y;wI~rc
zH#a`DNZZb2G5?n0%a>olx~Rq7K5b?DXZ<wmKiNn&{0a0f-EYan)HGGXldr|=%w#0z
zSyy*Aui0;sKXvxQVR%&E)eDuAt_okLzi@Kmi<Q0;F)%P7;Wo&~6n7s3v&zWGc$`gY
zva01{F8SSXE;PEXH66C*d0!krfH99Wx>}EokMnQcz+uD3$H#qdN<|lj<iLgn3SNFO
zxjb6Yj)7!JdMCxuh>ulSsL=Xd8UxuP7W=MPgIdHPz#`i>37eFS$6|>3`r=T<#YKR-
z`KZHn2}nEum(cO$__x>m5Yx4u=pJyuR+l$7RFRR9Va60a!30tts_hp3@Zff0;^6R$
zxgTUS%on{Y#%;UYDK7t6Vn6y-PV*DAzO`TSxQjTfyu3UvBO^n!%&b2N-N@&-n>=+*
zO-=37Sb-AN#ge~&ZMo^*OeMNehLrDh8tfC5@DSvaJ@i({Sx-hL=uLx@843?VCq<<=
zI0pVsmoJg-ewh7PFbd>6umWt7fsv81Ut()GTNrFFIuHfJJVxB-vI2<$2?^=N+qZ9%
zuI8#OO%cb~@4ez6@=5%#dHlOt!OP8hazmd{$S0uv7BL^AqWF+eQ1IGBd@hd!4R4}t
zwpJ~^dcrneczIoumecVty-N9P2JgOb+Z;#*%U4&A{`NZkp`V}M3mcoge=_rM!Pe?R
zJN-KvoS*;1?`ng5b#?U;C_<U%wfj(pC@MO-)3D@)0~kLb54_4T4A$?2eR{(1*3r>1
zL)1Bpj8g}w^Dl5@KTD>8QHRs{i8#-Uk0$^h;0P4a#+qBFn#Cx~u62ra6b*+=_1h~-
zV2FVOj;*b&5_gwtvt{Dqc6N3iUS7zXtj9oo<Fc}pgVFJ3fTT!aOaAw_iEQuKoG<4a
zoH_Ix6KN&A-nR-{3}vkB%~n<0FJ;}{TsWVs_YGX93s}nJy@=*Z=6+X1;(LACV3#TU
zGxH-l-oude^f}Vqlamt@U}M@noLUtD;1tJ;HOmJaX=!N@R@c?rTV>K6WrI+olz$=t
zvqsh0&Yhf18z!WrD7<*_BB{cq0nA2BO#EC`b>v;4yquhw)Do~T`{rYGrDQHfAt9lD
zA6{Nw?XdnY?BRpv{d{Fwl_qfThS{mRk>>=wo+s~Vz4v$xe9zwfskZF&y;$f)A&>}I
zu)8DhKK=WKO|#6W`?~mDangQcxkb+>N}r>446x1c>gwu&V^X%a^(Kq{G2#x(kTBz;
z&2;+{@N3|D`{vD?kdP1u@PdcJc0S#w858~tbA4rg0iQpA9vJZ+9nF*ffzTXp-%;SD
zvqW#5rgiEfNH{=X;ri-+xQP2u0EAaeF}J-G&-eS;)c6?Bo;~vcHmsTHcl=iwSQw{n
z?UP530^%6t)qbckHn+6Q{xrjR9**C#`e9)*MkH{S*4WL*$49HdiLpP4)ASC0#i3Ij
z0iufY#ezF*&K|1o+%kPK{Mq9TLX5718KMbyw>RUZy21v2S7t#$K@QH&roeGhOL!a=
zxf&d-?$YtS$K0Q*gYF2i1ED;xF&rGU+0vON+dNI{+*|JvlYMf?3W=(ZI9O<cbrGvn
z*e!^e4`=xfUa`F^>N<qpD_5Bh{=DTgZ1oYf7|t?VZ1$XYoz_oXBcPMYD0W4e*>?j6
z@H}0}_VTEcfKt*+00bN{u=YILx!RD>&`Q%@eB91oL`UZb3*Oz0QKalITwG2`*J5cU
zUv@{4@`rDMC<*>Z6}C4>a*d(jrI!~p)JqNz#{lC8-4-h<c+L-({JSrUC;i#y8Yqp8
zjS*@fLLI{)kHnp|N`QmYe`Xt{usALV4xt0s&He46Km2Ct{+so5*>py^d@Swz_wSF+
z=j;u6IVB|}`6cuvblk28n?Zz@y4kDJE4={*q4}xF1H<m+OKhWB*Gu4(_}vfY2Rw<0
ziLZcnE!V0Ar(U94$E9@R0)?O8^v2Q+^DRvpEM1@NYB#yr#_Zr@V<X(^s|=Ai;MDbb
zZ$WmFIxOzG)79v@J5yo&JZXDc4BNX3>>yLj6_K*_OXF>CZ~L6|F@unVaAvwX&<L&P
z@(?*a1*qR`nYaoNqUUNPGX=jn?p&>H9tgf39v&6`_kO^Ibjs8Db^(oV%XD8<tTgEc
zevxii`d&gPJ(?S=u(8m^e&SaggEr735KWjkI6i^!3bxDt`wP1~Y;zC^_ALw%8(V#E
zeu60cJCU_B+0ZYty!?=K*L$zBH$~i?i$=^P2S^X7(&tou3!twB95C1Iu~6U@M_atz
zf!iNg=?qWhF-D#&)?}8Fl9I29rIXIu986cQGJg%cg)BJN$WNc%@BgXE5OEBWzK0VB
zXc~Bif)&Q&(}*2);BdbeLUBI+t7~VwWsY!R2>Ahht^!Ny=<G}divX_09OO<Q`YfF`
z%5}+*Ee3MXd;UE!H~I-1HaFtB2ImtEj)4ly9|na&-+-ytzGxco-re0C4tQ#m8`uh1
zj&&U!AM=77%ewjGH+v7zP%pW0c3_JF2lZ<)ARQ_<lxE=I;F!Bk;?Opk_daZS0qhb_
zMMWjgX7=Zca)XvnzBlJ)sfQph=z-pyad%EnPqS$M%s>9O1$}2vP7oLAe2xMMiHJ-<
zMoBrCCde;j`^QC&p9ep@lS0y~nvjgl9Pz|;cS;aMhh%oE)){yon96<fnt)c~!|DFq
zGz-WyS5{VxffJ$t$p$aTAk)oSLAKhDh<Gde`?iQDAU4*5EFzIw5G7%J|NTX3{gf0R
zZfC@&PXg$;6h?_0+G9tjrvg?}C6=$!g*f{qroHK8^NlJgUvS+wanY+i<uGi`02yT!
z*c(TS4=5MJ^ilrTI$;#p6!_tK0kraes~>i-*2zSu@qwuA;kUQtv+d2m{gH5l89yK=
zmj@E8I(6KCNdmEN2f^p{hY#N;&A{_tuSNXeA{)QIR5ju%u|O7sKKdHbO8aF51VBC(
zgQ6uaBO*GQQ9xj#;O3_Wq#2D|<pv!U^<QB_!dK>Vm1LcDNrtF>jQ>BWecFQlP8jkO
z(CleL@WUr3!Ga?a-at4arhJflr#njwj-cNv_Njg;R5h|87D0p^fcB>8%?O<F2s>ch
zdpVqWB~Ps4!2ffakt@Q^)qc(5cJh>PeXWQBy8>A@$XJn)k)8Ze`gkiWizwl%41<nR
z-8shzBYz!_IJ0i%KwanP<au|ikBJbD&kdx2u`_$AZ6|gaZtR>9&ONb<@j5%~>junj
z4!mC(RDMWDN$WQlzv7UZ{6eIAWV#GDMbc@C+(KRi=a{fp%$~I31F6cq`u9Ptb0j`o
zk?H{ndXEoR1;>oVX@z0d7pUC#@M3vc+587@Aa5bfjA*MUE?(R=8Mt$;VG|zC8HLgl
z+h}+g&wdn5O)!E@Jtd7ewExHo)`xtC^fD$?+M+#dcxCTSX^p*B$o}DvKL&;2Kx0^p
zmX^rgfHOopL3vNHX}N-zV6-H|!zJ*9Mh`^?H*9C2?ZDhD6l^1j?eo&o)hnbUGLlPz
zk*mB4M<Savx9wAh$lvBv!8NqV!UWd4#Ds*EPXq7ohA^-L>T*VW0*HoF3&W;w3F7kx
z9RqVm_RzlH@OWHm>12Tl*^f<NX~|Oo@y131)t4Ew6{l$`k01Z0)_j6_;6wHhcOtVB
zl*UBZ3m5LRs$h+w=O~FusJa$ugTK0{g?q1(0=yT)#9CadfCh-v$lTm!ew-k$U_UwH
zrHb*h!<iSSjnWM2sybR5=V|#kTm1>~F;29nnwtlf1}K*UULS7A>TlXcrI-B&jr!&`
z<75!(B8DAcvWPbim*dCO*oVXjXq`#<Orhe#?fxlCd_M_fBi_V##F|V(fEeE2amBy`
z^Go4}tkRd1DFX|1ibT$NZ&ua5ufH;jH?mGyF>Aafaao-^TVLI2?1~Vi(?{@%oo0A$
z=LZiT(Zo^=*2)Sa%vXoF76dx_nP83c$Y@Xd3{w9b%^Z$w`yQPJn_3&mIBVTE#GnJ{
zfJ6;Dz{%H5xN0_~aO$ZCj)GE1D0e6MJzqI7(91X(IV4r5GA)YJ+dhehe^+qiHQFuQ
zJ9=1i&bz>IW8<McZQymg3C#;Eeq2JC&)5cUKQa@JkTU<z?ohzHouf2E{Gd<6@6+%U
z=ydF8@#ZC-%cCF$ixgr7fD|PkgzZF5U9$AS9s+5&!-827x1lm9pN>vWzJki}RkoDm
z@U9D+o49}{Q5GCUhU&Q$H5@MCbNir6Ui;q;^@7Jy+d@K6t)|NcFOS!p1@1wZ+MaLV
z2Zf+Hh@{Pzs~;uKXH4POyXBsGKZx11DbYDXycf}hxo7U#;fKmI7wMB<B7vuRH-p=G
zH8eeIBh-3-=M5sc+n{F{KFx3V6>PTBRDQD_B<n;pG=pE|J~~xr&dg|m5(3mnj~_i+
z*64S`j}u>MdiJ1ZoZ3zV^G^#Nt<=;MFdnDCR%qs5%;jjECO1w*lEYU9f=h{h)700f
z4A~<Zo)4@7o`0wpMX;D^L)87Xb(6udux+W(7*F13F^a85OcsbRCgTMFbwb6!u%INT
zrha8(!=b3CsHxAy$QU{^Gs8j9fK6ckmj|6T3&XG7!JPY5BbA1KNl%`f>=Ru}=2{0M
z8E!b!RG!>=GF<RbvRKLKB^MRMz9)s(6cZPBz$YXuOh;d>U84a{qK2%<6<z)zhaYik
zGr5Vfj(#dw&6611nOhv*?3!*<Cu=r&j<#UjJE!tz4yCozh6Z*i{pf9=Dj=q$Of}Pc
z|NaT6!!9rNZoO)ZJHrTxi4(4X|M&1q)Tt+BMUsRDmC>a}%hMTHE<XBsLte}djJ&v@
ztMfWC-}fzn`Nd-%D+bFjDwwT%<@Z^e1CM=zBEdsD*DuO_3p>aqW7U@GWE@(b&E^q+
z$e-$M!Q|tU6984Pmn%6-2~kZP5*QDzG|8__2&F7w9C%Ucw3rJqGtkqQF>n!0`2|&a
zqSq#BNd$bz!!#8C6R&6SDAvv9MNWD6<U27zppT%g29@$7XHZRd+RW8PgIUyDeLO+!
z-qV-Be6{+K8{z>nxU17~70-<?niSx(pnTZ?dAb=$-8EWJw={q9Ni<UU(IM?h1%Ekw
z*5y`z=zN4<-6%##*hj!DZ0fT>tSdtzw49TNJI*I|<gO)b&{8M$F8H>zY=lwdH9t>(
zGIt@!88kl~f)o#+8-@P&cL4MX+2~KU(B3RaHZYkk(@$jAj1gY)LDT|U9_fwIa5A_k
zj-MATpIVMA>EA2C#3C!#cl7k}$!@n1yoXpe+lqURAt!d%&-VyE{NB~K8o7%xvJuj4
zP0T&CXT}cD7{T~UmN}1HMHe=cu)Ixhp%P?8Q(E>CIG}XnwjIZ#8z_=i+MDouegh+;
z-89_BV){i|yDOcTR;d0Z(*stB%uf3n`=gT7+30y;N#pJ&E4CN1U8(tn9V)}-VNlEM
zea|V6)YzMv<p!a5Cz9|;fW4J$XkuUAVMfymhbP}+wECi|t(KH3$;#U4k9nqe(%EE2
zqKJ2nUmxS6Ts9jL>%3((b9Lo<4aRM4B4sWH`}<<sWmRmEr_7TL(N;I|5fV0NTM(rA
z_s5pd&u(_MO*|f`b$$O58YWP#Hny{)mxY`|LVZRXh=kJ)#m!hmNtjES#)^tiSaEWR
zz7mG-Aq_Qt`Ko(@IrV3k6;ApHJ3y@>(HauV9sc=x^G7iEL8$3&$v=T~@EG+cZ!R6=
z!>CV6UWA9#iR_ULP26GoqXAz{B5_wRxQpDG_WUJ>&nyi-XB{GI=u@7cO&OhRUSVMP
z^17AtFUbv?BW-u;AL!#jHg-7a6EGQ%=hgFUQPpib;DF1$+FA|&N&k+O5#P@ng-IZo
zp_pghH$~Uw5kZzWvV>2{n954|BQ;23Q2#i&*I<xU<{)~W#aW=fw2nT&HUm5x-Fw}S
zET3z^utdfbOXnlo3z+zxeC0HtJ$Uu*lD&ciQefnK5k$+M({sD55G;DTGCZrH$w<HY
zNBJ<6`wq<+LuQ_pMI%t(UFBPMrTo=hh4><^+qZM;UzuQ|N#mzz9LfWOC<p1N2T~Qb
zOg{G^o;vmI5QH^5a<U#*tU}HG4`N&#E|~IWjii(he)ZzSno_X7WqTK*;x(D)r9Hi8
zoL8X=N$hApK)=84lHhD!alYJ;Q&0-}rlO1RIo6ln>7K00dsxo(I?Q)=d_flES!Vr}
zYY>TvM~Q^@MKR@tt8&i@C7*<bFW}QGp%Kt&-4epZaL0(2%aTjWUdeVf?B7vtO4?=l
zvOYwoSu*o`X={*qOzzmXed*o6f8CnqQmQ|-^O7+f$iks=HfLXV`3-7PY*<0w07IZV
z)KZz4FmB$IVpDn^dQZiZqec~fklFETDHb+}UX^>-xbgACakw#r?4qXnZ395%)h=gj
zMy$SjK5;lsuh>E1N3EqReZWE=uREr0e5X|5`b~7+;X%`GbB}Yv{p!=V5<*lG3q?fS
zwWU$v=?AQwHeeM;R_D!>$mF=Z+zQG{K|^$Dg;8>{>zj;O!S**Daqo<(o`<{7JUD$6
z3%8eK-Np}sw@}=u(g+@_MX*9%u1tOSOo{C8Lnfrc|9rV+t^GBfEn!@`)A=Kxvqu@?
zm|1s6ZEw%f1vV<NtQ8d9N1B}!rFOK#iTn&SG}-E&F2$=|PWn>Jw+!MW-lw5$_mQVX
zA0x~vu4ChfE#;QyiTraTmeZ_3#qy}r8ymdJzU>JD_MZG!yrW~6sz$X2IWbM9q1;WB
zku{tO@3u0!0d;cAK7>lCdDvuHp0{LC<2}Y(u4I(G(ac?%z?1D_QfK@X%X&+4ml2Ab
zBN98;l%n6;;N968dh}XF;dM2zo4BAs9-~=`%VWx_-U9K1;5+E&^>^`s<HAml*s{iu
z76&M1BEGhYRjqQ6K)M^=TH^R_rVVyc>!9-=hOcWC7ZU+MikfmW+J$|QH~i3x@*rEL
zapB`u>ug7o=_sm46RbNgnnib__m2*vCV)g>KSq$R?;r+94mNl`vv~1Cy1<ka+T*jZ
zmn6}tm{ElOo=VN%qK6Ik>@FQHUo5=wXqZbKgFgP}yEU9r!7d6%UW^<vAF-k{3wNHD
zJ@EgtQv?#7utbl`MvT1-3j^`nHj=7$XhwUhz5|Oc)J=F&2k(-@R=PR4G){uR`0}5w
zx%bE{M5u9d#$L=3w{wW`C0`uK%v}Vz6OLJsTP&VO-ymH&ExeI3@PY)qS+E+riy70<
zU;~M-YKE?ZD@DZvWSQjwsgx<gA-Wob^Rbqt3%8_MKIB|hpO;WqYruox)McAo#)?Bc
zDs4%<gt_d^3gUZ#1h2l7*o3v2tT^07P$CagS0GLR?&`(;%GUmlpZWnZ3i-1Jbcl_Z
zQw+JeeJ4+j``StK$hWBO?w1DY5Rz$C4qSAOp!ms0=lPV#4DJX&q;Mxv*@x?2$(;NC
zoyVy#%R9UPbzVgYQD>YV+xQ1RXb0vlrrN&wO=be;Tpts_$76(qIvo<ilCnstY*&ex
zzuEP@;J{__>_Zs^Us(1^n-k<oq1jPaa=IvSc86`j?bUY+n)y9!|FWk@s(MOs@gw&Z
zRbOuMVMpC94mO#S?;<1Cqw`_r>fDxCs+EGLw^&h$kTxFNu)(<@LuM0}&E0l{2A{np
zs4SBKZ-`alt;%exF#k+s?dI$9q2b>PT0_fa=9g5PiB91Uo8f%Sudmj4mh2@X(-(R!
zrOq%p53G%I<ClN{1wF6QCBxn}QjHO<6w_9?$Nd|gwIJPTSk-18en8v?Q)|<0I26bu
z-$P#>9IA7+D2Ug1#|kPYlIe|Y>vI~tQvUb(ZF^D`yYdM2ZE(f1Sl%KZ+!{3Cq3VS$
z|D8q6x-n<RN1-h-jPY|1(^dPbr399p-b8+yVk{K2qYy5NvRp2TFgqt0E{dNgA<^(t
zRo4%K9L{P!FU5jeXJlWgJ#>}%$XC>KaOwIE6x4RK4ogqIFHn8LO%cYd9r|Z?=~cb}
z#GXE|%er>GKyF*%p^)&Vx}jggI=L`eWjH%_a<8h_s*FO;!3<cyJD?#ryFo#RU*g(i
zm>`l4VWx8uj!S8nXMe8hH@D6F*yMw)I{i40#@W<M1>Yc)W==LWY*3WMecW^V%ZL9^
z+7f@54%b?`s`?3G95p}g<+i9o4DXGM{hiuozMMEc#l?*u<b`YbvxSrHk|H&1Ofa;6
z)@Edd=EM0{QTQ&8)>&==BRJ__Nt6yoy!D>X>aaF9INMr(iQFB|*+NdC{28=(GMpT0
z*w|pG|4-oSp}8Byn@90=z`ekr*1*&Jg%e($xxzftg&+SFYO~q%{Oj%?_c`s^cod^Y
z3v7RCk_8Ry_tzwF@?I3xl*ahw9vFW2buu{-Bog=-@FU#(p66HMxqaiQ34ZhE!n(0?
zYrES$^;xb*4&&3UDW{4L`jkJmkNADM@;MW^JQECk3ff#kuR`+t1LO(+3N1GOz*Bhj
zdMh*38{+ZJ)okJPu{J<J^i;+Rb77&okZE6ZXiF}LCup6ix0|zA0dDrskE2o(m`X%s
zWo`I2qOOiUz_6z#n9-7ua}<;Xd?uQoQyA?RZn_3y2hy7`cJg@ND5!)*mFBj$4Jb0_
zC8X4^K7y3HHqzJ=)LbxY1NXQ7HW&#Mh_EX<Q{&kW*1Lp+({Aeou!-_?X80~K?9T?y
zqIP|Fq6<be8%d#Q^cWc#E7}(zlIRZtzigP$jW(Jp;Lxcl2KqAV?og=o*|p55((5LK
zyU4zWQ>EjDx&J6DTU;*!t!cZr357WeEU>GS(TWq_e^WWT;8AO%xy35&Cz-fioKFom
zK|H=cV-b`==44aF76a#&?zrV_t{VtZS_+$_)7FWJs81n1Oq1R{?{hjmY~o~|@#906
z=61g;*BAU4mCykk!n(PjKb8*ESjtd>dMc)_mX^>FuUEJ&*UZR0d^4C|zPB(Rc2Mq(
z<Z*uRJMrY=XORpmDq?kVAO%08IN7zX@cUYG77+n!@;EJbQ(+-X61}c6!w{iKmeIm8
zL>IuR`zFLHfEe+vSTnc#K6RM96`6`zS2s37q&y*g@XpWHF8S^N(mTw^cNN<iFuV6L
zxhc&zZVyO%o-G~kz>hes#Rz}+UV27q`cIl{muRWSFkE{#mm^v#R=YlZ!Ue-!gY8+<
zPCFSXb129z73?Yyc?tlIoG)*fSC-YRt%(S?oOV=Z>K@pfbI9LM(TMd2?Gw(}@-!EN
zi86p=>Ejfla=GN5AElwT<hdafY3K$qrAu3{`MkUbMfE2igc3hEjy%nGd-A(P-530d
zA~DwYU`tL33-}AnGNDPI9}*Z!Vj{qVFA*eegyKfS+yQt85939>hQ6)siIIMU)`<rn
z{;<;bRr?Dxc3sGw`|lNL;t#w?i{*a12js@UC8(v8HaqMu(JbU9&bAs_H3mCE4>6WQ
zh2F8O6q*}8*#h&O1SLzKUJW%Xh+k6-G}FbKB9d+FyzdLZ6~eCcxh-Puu7;tJ5wE``
zLdjzV@S1u$=ry<AB00b>k3+)4*Y@{cgU01-g>k3g)@vmt)Ae6*Z;Rey+1uMAPytY{
zpH%8NdifPBuXw6H!K8O@6K;9TISWLQU!i>^Wz@ERAGEv!p$Kcwe-{p+7~#+xZ2g7j
zoCXo`KDR+9pdE-HWB_e7oEq!tZ=iUWm6OwVy%S#a2#zFX2K8Weh4nO7md}Y|qx<0y
zKw4o9PL5XH>}{P(ftnENTa;FB5^73>R}Sso_N`bGhh3S;-)roN4`xM(M0qrh7FNdG
z^+ki}E?2_o_&$t00=Vwevz<u<)()C|uQJ8XieUZ#L~cI+W9<Sc3!trXGNG<A`a_*r
z>T)^Qr%W%yAXR4sX6*nI;9kv=roin)><zd^B%`M_Lz_$SI80?cv3F-SZ?m4#5g-g?
zfEC%u%8D&r*!~*;B~sV++{D4ZW&peY2P>f2+aGeQf!5R|U@Vk1WF{t(fYx@EgsfA(
z3{Jg2pHcAOxbYQ5F*{LQ(81obnY+@f=8b|p|2dVeit`ce`8^?Yp>m@)j%qWaau5mU
zO<YP!N>mIB+gk|{k(7vt;w9hNR$p%^_|bFF?uiA2iR0Dofpp<mL^s7*+_dM@V;&wJ
z7~#W*X~pHwGxW~9fgfZTV86rBh5iJV%sQjtiy4MeR2}})m0w;(&r*nUvPOeaChA)2
zT_%Z>G9f@WK+8lH5E<(CYv%wyzCLS_Z3eoSGyBtiN}!;B@#@v9%{71*U1?Q{goee+
ze8jQ!xP>EAw_Foi5jr{oFlX3)V2jhRkJ=ZmM~_pRj=@F?EsiJs5}Du-1=0HXUU7bY
z9)R!zd{TUT@c;lZbWH_qqvK8jY0y4w1HD7g#v=e7<F_SR$3-v{&<p^~pwAnRojP@(
zc8h~s3c<6!%kVe<MA*e+boYRVZvaZV@$kKJTJz2Wd=yagDxUqk#*a_mB0(AfjbYIC
zsCwTQKj4TT9?YneG*WFjZj*d=e$MpdNeF0+x859xAGBO7Hgjp}Wg^V?>9)>i53<&6
zELKN^@}YGbE_GB2F0bIwaOrdVNIm{yFPsTk#-h1R&|N1%dSG<4-2N&}Ad-^byu$m!
zF3W3&6%b!2c#I$rL^lL<wNhc_hR`g~4+8z^$AB3j2j~+HBGiZ`%k-ZRg@a=b<QDTl
zLmt`vphNT@o)m<E>sngyZ&&cL#}Bl)oX3)PW8xE&o}KX_2^8yfd`0fouGiInYUlYk
zH(bQ`S^&Y$pr;QKbKR-<F#2_X>N%h?ByC)`uCKpe^jO0KEu@g_Y&y_STkYqD=7W!r
zwIDl+20k{2@r;YLw}w`%bJknrBeJe9Nl=vIIMuIw0=`1dE>55%n4(5nj!Ys2{2($?
zmZ1)F3aih1=G{F#L=+SXuU>sB!$?fOn5Is-NV#gByU$U!M?@uS*@F-)&H?Lt0IPkO
z1IN;Ez+w-t2>$$4io?Hwl-5q!8F`rK`p)gEdK-PHMkz+sBBhT_CWsjdcl@|oM<)eT
z(p)(Hz{B>DH~M;4d!mI(_mN(cM3SJE8ivm`K7GM9FqYX)&t-8#Ez<a>`8e<PFGm8|
zVbt=kSy^OX36Jl&f3m;K=R#Q29?a}JEN!VaGCj!(wD;N`G9uKBqsD=@7)#k7`>UR+
z{o=<Bj$|Lq^Cu8yRUzaqY;ZB;_y)$Q>NWG%D?E_h5Alo*29WVP`uJ(C#mhI%MDtgA
zNs8$)Sr)sMA0!s`PMi4W)2_Q%L2M#<61XuS9=y<ObclSp?~2gRW|9V+DHWb<cQ5k!
zmY=V(>O1z%_wTMBqh5L6J6p!3Bc>V&5el)(P8|lhinUZFu+cuM<fThTUGneqLjCI~
zZye#R;@IhUwAPOM21T-VVIZD8dkaIV-?D{W+1t0BHz1OS2KyU5k!y_loA=wtYhM-F
z2~bAP&yCo=ob(^=boC(?_tsqsBRZQe+yDLvwT`i^Bz?R$7xZ|o>^}JGM%61RKd!jL
zx-8V?$R?Ll6{l>(BxY5qgU{X`qQ?k;I#$*O0m<~?0Z(#FsM<_Fe+4H^Sh~06{y~YK
z_KxdW@X;z63CoZ&l4|~=s=3@7Sxjqu)=C~25(3qJNAJXbw}`!aXCZZXxkxJPXCqD6
z89R4}Zlu*a2Wt{@Bm@~XT4uQLDIELdf*mbyil_9GEJWw1fwHfX$0ERUP5#VIm2o$k
zF3gz6tKAOLILWFM{rM?Q6p3aIkRin-NoY7UIN*Uu)vyhV7-t{w;_ELc&<F0NMf@p{
zpXct@vzu$T9tv0!WOk1a4&B!jC*2E1^{YZiSUHfN=M3u|;fJj4E;seI#r->Mb$>sU
zIVf5>)x0#NPn20|#tP5nVTU9H9QUV@)mDwStX!b*$>+f;!<U3239HP&a5#LrC5j%=
zqs^t>^7uzwHJ|aYEUiarZ`z8sgK&l1yx{rGb;E~CRWBL{*O+*5U{9EW3bc-jr79Zf
z0oLV(JwOAo#CouZoJP=UlC)lpD|Km02N_<}PyO1Ob){c`U@OLT{EayPDLm|Itj@SU
z8FBQ}wI21eAqn<a_4Y(F3uP&0J3E~KCLmUu2ycDe34^|H6a~@^eV4$Td!@7q%Qg&X
z!6_}3mo2NdXfJ5>Yq}*s{;kK`-w^Wlu5Mrt=B{v4<l(JH`9RJo9kI|UbJU4#Qz2H~
zLTPGUdsjJnr#AzNm&cYc&jgDctDEAX^X)N|u7Z5^mRFFADg3@(<nAvBeKQC7(=BFE
zb@1tXP{qscBkEGe_ifG3KD*HVo(QaEC=;8jtcvt$IKYH-iY<gK*!4pnb{8Oqy4lel
zExWKxKZwsW-t>Lo+yPEgZ|?WS@?R^^hr#$}Fgn*OcX(TyT#Xw{_U`4_*OxzhxYYCj
zk`Tdv1Y9XJH2&S=^L#HH!pgBY>48>1q{dI7L|?#!*}ZvJARnrnU7n1o+b|8J2a*};
ztP_!tAb|I`ZA<QO%64H#ZY?)GGZM*hv;RK1R#Zh9)}9h-o@~Fc8|EOd?Rv#@1!~8x
zR1!SFrcbcO5#3+&fuG~-fR2~7^DOsTJUXyvc|$4I;y$RId?-2C_<ZZgMKjsn1oV6e
zEr@+EPP5SL3@Y4_;3r?f0{aN4q=fyq<|PJq@0#IdV!B~#W1~={!453eorLUnQ4que
zS@)D-IKRxpOAXWTM%hrK)-%=p+wJXBh7UcdeowS+mhey0m>Is82$Z6gSWwe%jT8i#
zVaDXXehw^r6DCO489OZVBCSR+QY3F81SeTHUuH69<TBEF=pkkLfH(92ZF%E-67%%z
zS2>&9Phb*`=6L8&r(@eU(dW!Nge4y{E`kHd=nv9d-t&W@mw5N4?>ifG&O!I+r5H2C
za^<!czWYd#vBrb4CZei%iNMPjg|^jU*&W~!t9c+H?Gm-gw!U)?9o*ZhQvlO8nk-rR
zUG0*%QGMFccpfeFu>h=B@7d(}$RT+Z|F!y@k&~{7a~qfxyUscMUnTF}f=jBS=R5_h
zBJ92%CE(6oBj-g#nlGU>P?-8Cr?ZvOZq*WGS-Mcxi?-uv5VqAs3JN>s<sS>;{}v8S
zdR~=zTeX29>LnBh5HMUjO~LFak_I7gCYowr<0+B*r$Y0cdGgrzw2i<?h7OO_tpMal
zpX4AL*Y03C<7|1`gNZ{Lvn`@*@)*{5-b{Z5I(%J|mtXi)Ff=VOwiS7r2yD_2&4bY1
zyQM+#-S*~fhKhMX3qjmqVgH$DR$;*!lP^beUu^rZsXg%qr&{)_pTNrx#g&e?spNWt
zQA_NlEgC3X^`5|w@~@g#OwFB@W-lG*mP$lD$hh6b8SdYq=nA=Geq3ipEP7)T|Na_P
zr_~C-+=pQEo=$`s^vu*$MX8T&*FHNu3_1vK%lnxr$;2D|2Ih`*<de>p$i0$I_bzHZ
z0vr;1)ojig&tE>WiuP};o8%b}Kan+z6%j0tI8`YYU=#+*pX=ART+Go9brq`pcuR<0
z^F2jcFlzlX82U6ga5Hqz&3#4+!(1BIQV}6(hrMg8Bz=+9>0^DCcbs#xb+qEcn!B+`
zJVm;%B^J(H+0i~q#nGD-T?KgR=QIiL>7w$}1^h!AD;JmMmTzztngqR&cO>KY_qHkV
z4etFg!jLW|adIsStcCG8O-R9p8}hBhJ=VPecOcI{Ww7f>1_Gno?UDoa$Rkcdwwl26
zyz-`B<kqo2ci(aVJ7}KTKKEf!Tkiqtp`c;xE(6WxK6}Vliqe$<pqe;%a?<-SXgxId
zy&$X+YV+p?B_cv`GkED7nWwtbHt`hpJNBiVT?*K*-VYlZiVDpCfP;UK5Iqwx+?3*|
zbKKl?7!VRJIF4HVE^Hb6XBB0v&|Uk<3oJTEd|2MDZ-TP!*W{NXYybcrzkJ|4fE!3S
z2ig-){7Wvm_cAct2s}P8kwEYCoBPjSR2q4s+PP8PZ)*@Y>D%5Zwif_K`m{zh0JfF>
z1MO+wVtj<n9n)~onMxg=JPr|t9?<t$QsUk_Vn9$1a0|U=1DZ=z*OL$%K8FVll&#66
zKXt+dh-^$rEm(82qf?dV4tdUYm1?<t4B3s#Hd0Ks($`lx!rY7Mf&@4ueKbHToqVO_
zR^mKWSGW|S!&5Ih+s(iM$2XZMcYwlAk`mMn-Y+&kRHD2&ex+w=z##s0m*&#b%s($e
z;Ted-^rCCps9yd!Uk(FLpIlx)o@x38jMz!u3fX6zc$MMN_v_HG(S@?66l03!<4(6R
zKHP&=&WLk&)UD^0^mQP2kaZgB#PIS%t8)$fL{ioJ`Cnh=S81O@_#yhUF2iMABCkQ@
z!wmh-i_Y4F5^)!NdjfI$+g*Gcp@qxGp~2w1<|*8{JsiX)>ghn;ZUKRxm4du>o;n=G
z&*$O!5QZ=}j4s2(SX4&<VH!|?CpI({KhT*M=4L1DL*RN&Yx5Gzt%|CJ1Qe7!hmJER
zp@{JmodIaim&<3(rDr6~pI$ME8o;5@gB2SYwlBeOiiy?h!7kA4MGSY|vpXtR$!*UI
zK$4ckpkmV2|DrHNfex2K7J7|jGvZ7Cb~k0MyN+6XSePmX=_uPy@W^F)`UeTubyw7{
zL6AAYFaTe{*`xh|)@>zl+3Q_J7shAuo3JvvXM-3{%^$qNf~TQNfyQmOXTL}5O(G?L
zKRHCGS}AD8-1_HqKXCRqP6wUGLF-~!?60oGM%yef&zn(I=n@{Ml<EE(x9lj!k>22G
zv4tcIw5WASgmt+JQB`{qANY+^-<fjw+>7KFYUw8blD5sn`WTirtcw@?ZFqiUwa_H~
z@~Ex_g1hTM#RzY9YLbcb0ZBG}!&cd2heL}^w6nRbw&aUvW1awo<=uxbHfxX$7)a^h
zx2(MT1otGz`awCtZkPx_Mdya6aqzYO1%`ND3yY-^jsZmdm9@3!A3m^fx1197_4Qr7
zEsaPhn!~srKvMAsiS8>)g3(?OTJ#oLmj-I@aKmR?berC6#_dv7*r<@;xXUh^*Bxgc
z2v*KbZf27hC=){fDRXeSJ(!Z$1XZ`zCKvGL5nSHy-@g|R-2<W|3^20!Ou8`L4i^U#
zS=Cb{y@kqK&UK_TxmH8i7VuEW1O2r{s>qq@LZNlwOzVpn{M+hcQ}#8c>ggIkHLV<x
zCnWY%c+MeF%y4AP^{<!YeuxYDZHc@L=*howo&_VyK|uer=m^2OyuRiG_a};npyq=Q
z(*&(wfXfe91O$bExOloVx!Wj@C-)@$<#a%=UwDrABPpV9u6(f_5$lDLu$5wr!8_k#
zH-U8TzGvk2_W~zd`iPED?t4P?C3Q{$Y|=hMP{OOLtIsxj@*_AC)E0GfmUFe7fGpCo
z!{@%}$qaytAzU9&fSR-ip#e-}J8ce7A|4}{#Ls1a5>kMEg-o}`XSdw)$=Blu2?JI8
zbIemrkk2TyQEN^hWJfF9T(so&;%Fg>c*geVF1Y6CxcL~usx~6>KM{-l5b=MAMMN>z
zc$2&Rix_G}z^6PuI?4gXeznyk+njpa{|#3Z;RMK)^@~678V>x^GHCD|K=4VQ1fV%`
zCo(S@XPGw2%aPD0=pZfT!QmHp<vV&ZY~{Z8S=Tf!nsPhgq{AB~awRJyNJr)f${s)v
zU;wnJH0vh=9LrunLh-!WuPbjD04P^PWMrQYwX~lE9XlCo*SuuzoW;n=J7<qc44xZf
zg{azU4o8{3%FVdV=mey6&no%k%H>VJ&SE|g#r<Z6QZL^Yq2C;<%@s8mmPg3h!G*s<
z*tCB>0c>*F7ca~y=_n}`x5o;~8wbGMi#~=}i4N0PN%W_H<fy2iung|oYynN{wS)3e
zd3$=UY!nD(%i2W^qL+q08wD3G<UkQ2Ya^`EA39OhDCUUH%e-7n`=Ew+-go(~&%rx6
z-C$wZKhvcB9CLnONXnxEf00(~jD4t+H3s74<)sGo^98KTc+sI1hl=-oL}@~{UXx~k
znOlzOX=k^|U6-q_PVZTG(0TN?T(w#PktyvP-O6SXCF#O0G*TvdaQ*$-d5qdQL!sYi
z-v3LoEDp<F4LSujy!)XLxcvVT=y2cv^98{y9!L}X2ylrik7JWXAK2?gxBWCqRJkXy
zArs-u8tu0F1K&i+T1|w;bJaVX_|FG~5BXwa19EnbX5w|lifky*J2!2a#dxD1K1Pxl
zZ*<}!V^`x5>1H(w2&eUm0PGHapGyb8;sQ5A@IR7q{%+=7pQ$v}j=5|(?xC8o5kkNo
z%2^VMfQ$S0PYwJ1)v$kw7Z!b4cOi>}z_`Tq#T-?|jfKcrvND#^kbd*{MV&ANg{_8H
zT6(c=>cvvVAhXs_W^5H<aXJYt3<BXoQ4RX|KJ&jmTdzO<4|d`N%lBWe#TzI$Pldg4
zE4`oI+&|6hwXd%<aD04RIB2(U&M6SSOAw%ZxHK_Hn=I}q6^Fm3L6{vsR7_*P)|d3f
zDOZd2!|Cn5Y+-<0{)2H=;|4-+@(zpF14mn%STu_b+v2|5TMiVsw+Gy->SI4NTUmm2
z-VPK5$zjt1r_D;il|^)LWC&wfKFBtfaTclW*!i8GqYxRMoh6*^sVHH3&d2|OQoxf<
zGltB+=8mP9AftP(bLfU(s^_)uZ4@!+KE?%gng=tfq`6~}<AkTI9d$#E+Sxy+eQ<R(
zDOjT#@W-qjfY>-}IrDt@F}q>JCE~;sIEr%)IkF0$`>yqyTXTwCO%;)Fv=gh*W$EGJ
zp{I3HF?Dp2Ir!w&xGLWDD+Cwh8CiKFYVhCBpxJ}yS@vHoVv7|OhnKr&=M-M{=U8?V
zo!UOkV&5fY{ROIke4~>S{_<U~OYtLa6bpj32{&_Lr|lman#@-#1AZw^1SLKMxEK>&
zm{~e_q&n+=m~cCdx?;s=-6YZ=_U?<mIb?2`*QDf#Tm(&bqxR@j=lg6YE>f{nfMV%m
zlkmweiUx`j7=W3;RLw)>an6NJ@GAI5O1PYKP|mPax~C+%7u1Gp2Ci6~2kFJT<gDKj
zYg&e}q62CVtzm#YR>V9~b{$`*(Tm2(NkuzFG5f3o8cpIm3Y=|4)ygkd1W%)HoP3?r
zV=~rCv)z%8oNlbDi^vjp=N_<r&-nKE1EC|*<YDN^>%rELiy>hZjKgcJ1J8b3Y&8)^
zxz5J>_Qv5yj8oK_jZD>Ectx%Q3FnU(OyWJAHpxJ)BsR^_scv<sqW9@xP%E01f?5}S
z7!trKaKa6+L3JdPzM@(dSW}}o#MqlA*<Nys8gDw_*e;keS!|yE*>mE72f*_Pd|P9r
zW96N7E&VqoeVJRc#AsJvp5B&1Jd{q6h$^*w5<MS!jKBR=?_jD>Gvsiw`@ko>rlzlk
zvA3<KuVRxJR?YqN$6U@NPyMZ-(Qs370&b_x-qqY|X3W#$<t>YJ_12KG8)w1mJHSY0
zyoN-QJQ(dYy7FHcE+3?ls{k2HE`UYK7OAmOI|GQ*jbJV_$C;kKbWbDMHVL8Pjlm0?
z!~K^|`-F>dM`kywS9Bv0O+`t+VYV;Z^A7FN+H}7Xd{#hm_M+psA=tZar><>!z965o
z!TvyXYHUnTLrnnWdzA36ZUzqxcicYkQdrGrbGH*1b*?y{W>v?ygy;JC>1VV;Ng?);
zQcvm8CYI)mw&y^l>O6CRROcGx<(q4=>w`nh)z34KB`%~Ap1EWw_Gjw=&wU=_?B@B0
zJc*{BLo4O(C!=Qx(jvs!FOzl5F7q|@ZLaElS+)KvM%Lw+5~1Mb%9`XJw8Y>>-yCH+
z7BX{v+aVSy?AqbbQbU=EGNrZPqUX0rs=Lps6tjJjGLqV#6HlLPMN)tkW<V?S8Q?1a
zMR}s8ka69Q%joZm(>F@~;!lo=5{%^7R^O>qmp3^f3HOGqBfHAz-UI;P#|Kj8bG~4a
z4PjSa)rU^tzRvG?XiC);s+{j&y41u>)!!T4|MR84@)<j`42}B99Xz@Vju+(y?Ru=F
zQO$O>$uzL>z`0Nv&rMf2FBUNAwe9sbxN*anGV~5=*34h&IezBFxiG40Cr6<gZ+@v*
ztQ3#@bZH5&jAiK_#mVf!31_ZJZEmnlRKNjz{YhzHC56f8lJJylE4Izhc|eF_n5J1k
z>e0GPzyVf4k+#O7^lc-55VM6{508;Y@(!+zLHka8e!$2@9r7~gD2758mI&YSGG7%|
zsw58ALIq@Vv?NV(Kt0uuYvy$dL#Iun*^<mm6Q=TEtH5CAv`tmHGB>9En`OI_Sx3Gw
z4CWF>R?gt?^u(=wZ|lF`2>SB>FBOVNP>c+ggbiAw5pXt*e0yZ%%-@t&JPFz&C`+qJ
zNvVZutINx3{73uuq(s98Y39zsN9LQXi*xsA27XhMsnHQB?+<A1G2$v&tG}zoTi#Nw
z9wfO(D=Dq~S4ks?KgI^pz>IF{9GqZg#5St6F_f6>wLv3Hp*QI9m*bl>O>yD_=FTy2
z#~+%ckn*SB$H)K^{kJW)&*f_c#lqgMnmxF965BIBpNrE>R1a@0{Uj2e|C&{OL=}ed
z4RzJ)rg86|%FY*IC$G8DI>u^Fs@bd3lZY@1FJ#u0BmUS5{id#=B2qRKwLgBTTFUU&
z*1lD!u7fa#8b6rX+|K;COyMQT@p`yL+IEa#1DmfuJ7kf_H$YZTb&nN>zxup?&>CZn
z;ivKiD}I4>nk_CFPWad4TMjk8!vmfn1%7+;NGUtFXmaNji&I_^e-e`zebZ-DB;`i6
zljMJ9aQzsI*I!3_+6)9OywdZ;RB;zy!nJyJTM^`2tyTggrK*$ySP%3$pFHcV_p{8M
zSqd-GjE_c~V5gR2qWq6&TYvwgZ6!xGrnpSzRVlnCX*#NhUjf=0pcp;xpj41Br7cww
zNpKaptEM~$8rNuktOI}fd}JTEtSkQSjh@qG7I6|Dc_-5)c)cn0Cl4TplB!B8D(hk{
zs7GXfr7;Qn=otq0!VUna^<DRLR7u+O9^LMAxOBfK$q^RxHTMApwp_oVhryzc!A~|4
zR>$BuZt5v&k>^W4s1)2O0EY*gdqI5?p|Bu-hyfSM*J*_bFl|3b-LXodnJA~&Zf8oH
z(6%T^SkSKYpNL>6j;6F9%3OkA(#n{tz{=l6%Ia!-*-`&7uw3Cskms#2t(b>c7O)1U
z1>>XgKW3hlwks08oCuDLoa{+(Q}S$ui}J%{H<3q-=ws@G<6WWaOrPLxpdTe)X#Ye=
zG}8-^>kd$ACZ@6#@Ih3LAQ5nD@|;j9W995@S2wiX1l<wP=;Jw#L-+6|De3txhs`%F
zrum4TJ~Ydi;?t!i^Dcg7UPK|4`i*dYA~bo@ERqCq7o1-0yHDjgi($ER*dRLHS&353
z*qui5lYf2d2!-5At9#2AcRz6ga%t@HY|o57vJ<gpAeNl3D&iuRWzqUaCaiP0)$6f@
zqo!Q35)o_lzlUuf6<@7YQl|4AUeYYId^@-#Z9w}sSx$k4h{xtGC$(5Yj<}y3g31r5
z)F1LT4>w~`ERK#HOx2dq8`m_Skb|l>LLu;M@;dL8p)+bgCBbAaf+1O}^rQ~4eTBg4
z&1K3L@Ot8a<T=f}=YV^m`D9h%KeLO1r4>T|$TW5?FN&FYHon7e93hUQ200O79P&oj
zl`L&mbJpK30jqGb4d_L2tYO?dXx7lvqrG@*!$+WEFucUu_m<;b2<xp&Z@Gd2Kuy88
z6(1Qk0u}AyxT?rR_RUVsSix`k*<fdVIEbB@G|a3ETq`dChK&cH^ITq6G9_Z2RyJ}F
z79!9d`lOeD$uhc4HS5P0wkjuawNu#xKs#$!ZF*(E-<6f+NQlpBCpNS*UyFlXdIK6>
z+l#mOCv}MNh5yDY^gA5wP3vmc+~c-p*&%}G!^eRd$&9N+h&?}9G84fK8@Nj7ZlXVS
zspD_3M**c%6&Rk3DB|g_e)HfiUK)!_+ucyPH$`I+Os@_7W-R4W>MHmr=vH3b;5xs>
zZh%{u*0lqQPW?@-QE@+!qv(MKIZ&dqc0hQ#8^<><69;;7-vTAmxB60pczr(N^`1j3
zJRbWu`vZSGsMGNl!D9j;%zuynZQhfu^IycewCl&(6S~Ziw{4u}v;ccYaF=CePhJ7s
z<FWJIdbq9snt2`Xd9E0|xK{ZlO6xz5Wp%=Wlg;hdlA+2gc6G;!<&7>S+C4izWl=$<
zOxTA4h;s%hHD5x->5m+$cw5~i|17-u10Js<7h0j4|1V?++sPtQq4fPId2sI%R%t>=
zxP_4ZxS_4_N(?6G`@Qn0Z|%rdKw1+(ayZ|c#if-<A!K+_0cuV_oOpe6z8@JC)wlx(
zCGXnA5B1bo8gaMf%}vCW@RZ>%A0tF!0g{~scjAX03VB~%HrxCDnJ;s@lf`}r*z|OM
zmyf}HWW?R!x8(+zSj0@rySx41HRWRdpxqC?XTfFD6#M~z1U8`O^ZEG`xahu>k!Id1
zYx~9)R?skzbG`?%b0YnJL<0s##lF2tl^M>G1o!3@48E|^uU3O1V_Vo3rT9F%y|*(O
zgtgJ;s77~Ukh!ihedf0q`WzSc%)w!59Jjr_9b6LLYV5b4sR#r+;{mtqY_^mDi~#PK
zJBtVjoo`~35jxWIMp~Yx*9O}5<{e)+SzNb;0q(0ESXVq=p$M`^F=7FrHnyDj{-zcd
zP01hD*48#4|KYn%B{?8b2}y+-BurKAdkDz+N?v-Y5{F91$+Q9XMPEFlqq8#uA0J=;
z<*#1~FO!GvhN{3l7W2Wh?z~tjuZ4YFDxps)DJcWJ$7{XzKkNV*mxE0xH7mQ^boz(Z
z%WF@*-D9Q+4)PC(rJw*2f$Oe6453m=NgRFDTMi<Cqksjzs9?}-7V#Aq@$nsin&7jr
z|6d{k{c4o*3+7SNRh@JBZgsJ?^jQ^wjoPv6bQU!&k#4~NNH7M$H$$kOz=x3la4DVF
zZeHYn5CRwh>wwp^=e@QaxJ`=q5(*;5ms)RXfhTo7wLxxSF5P<XLWE$Z4dyT*c+$Xv
z8jt{hMd$VRv2wQ5Ah==y?%d=kXNrL<&n&Q>&)_>AjO#2UmmAy<4Zz({0>7JmVfL3~
zgsd+3vfZQJV_siZ)xK-}#5BaSdX@T=4w!5l)q8&2&Wgr}4mR+u4sC5uO-xK&#I$5H
z7R#Wfp6`Eq4Kd}ban59?0tfb=Krqka>n}b^h_EGbZ-&*slAbi9k(LvfSj+u^(X1L>
z;|N<~`sa0P0j#_p$GRJ-{3=iR!y9v48`+%La1cy>aXLgb1pr56FWb#DZt>G|(6GkH
z{~N)F??=q52nLz#deK*6bud4AOr}!4k0ZECT}>$G>-tBtixT02evuNXjZyOWGI%qr
zVFFI&W?nQF_E9YmP#+`0!N$wI^g>q|nje?~JmqVTqm8_1RzsP6c%%sNKQEA&wX)o;
zJ*V|w|M=Oe;1)PISp*3=A_8JWfnM$l%|ti}mnzs51KxMt))cUC1#s^T-0A|Q=4sG^
zPt!K48<o#|G5%n2Vl`rJ2E^PQ7X1W#mXpSi#Re!Dja6c0J?QBwA)avAtwsemU+fw{
z7XSK)OFj(I(Tj^%7d%4+x9U22xWzwH-WhyQ5AD@?Ta6e)@Ndk{8sp8C!?6^^z)JTE
z`5%pcujBsrsD}r)=dLw7HhB*N(YHg;9zPZn#laC?c2x$qbNP1ZoX>Ce86?cLIht8V
z<7YkS+uQ#e+5tSgc6E3pJ3i7T9$n+6N|zww$m$`TJQRZd!s>TC>0Vdv(>tg38ttoq
z?)ESGD-VHD*Pi6#(|`)HZzt%Hub5IJ7a4(Q+R0|5bQ<%(3uy-Q+;IbVFDCOf^G?VB
zXGQ}NNFCKS<Di*{N9lNX-iDlB0&f|H>Axb@S-l9T1N>98Xv-}aR<NhHP9QYXqlGpG
zfr}6_*__g!XD9mS7OYs{kqTspr7LfKtdaL{kDG0o1K`Ox9v!_C8sbcu`P^Yx0IVVo
z*g7a1NU&PhmPIg<WxEtP=KZ2HH}A080?b!%TMLW2Nh1}WAwtb6YSMo6>Xo*xBgIJG
zGttF8J(~1EbKnZpp8iXkoIn<|(Z?;EkYqg4ph4L)uZh#7zoH^rhKMxz|3_<Q9*uR|
z_5HI+B||hwky6naQIvV82u&1a9#SD?o<(UAWmb{7NGhVtV`RvjA|&&iWX?Q2-@|?1
z&wAGV-p{+<Ki<~5mg~B5a-PTUIQFskcklh#>u1Lt4*696kUg1sFUx%0B@-X9bn4v0
z5?iuCH!72}A#S^3gxYuLo%-FXnePQPIxcm`yJr3d9>a^p?1cQ}U&V2AQN55i4f>w2
z4E;d3p*daeV+Shi+^{LIVwXxiz8u5RFr>SBKO#`fj@d7iEGJd=l7ICtVcT8uRb0%P
zO>X?#O=U?77B=Wk0x(imK#})>tDm``@JGLlV9{eRIc;$}ZRzjH=1^Yq(m=#vqJO`M
zh!um2x&24<0Sx;ekHZ0=rD8Gtmg9=vA~wl?tCig~Fe{8Yzg0QM<o43*iSMI)8>w)Y
z&3_DA>K>;zm^#<RY)4Hi@js|tWG~RZuL|I}QpVx7%wgz`W^#t)kB&#O|2VlD&%%5B
zOj6Ax&TCRj2*u-YPa?;PdV8~pbWy*w$!erxcm&@&u7U%}WJGp}a$cOILja4BDXQj`
zI!DP<tVO{x+KlnCBlwivBm6>s!R^mcR_mFaMk5qb2{W{WIC9^(p8E@TwfS$F4&`cN
z?H3}KIo{q2A&-&Nwzw-Svw++DQAPO5#L%o`EBxVf4V?gY5wZ+S9($q<4_JJbV+d1x
zZt`4MkVc5}tNi>d*a_qpe>gki%GZ**-Gd8B6!HlAj*JiSa5|zWKo)|G{_xDCxO!so
zpAOS&i!}fA&n%abJ-nGU%Yh?Fn7!ESo|032)U9Wk4jcUDi0^gj4u3s+cW3l!%9h%9
zT$AV3Sj;*7UC5e_Hbk<=Y+2}Jpy--txwd$fYr%M=ZzWI5Ie8|JH9YrgxqO0Lx8sX5
zo8VhBv5*JtuO!4FS$tU#Ycu#<XZ&Kz$4Q3q(Vd(DWW7sQiz|Nq>he?0zgpnxx#-;y
zf8#x}LKaqfxF7yoFE6{>XTZlZYUn$C;hvb91KC?5q^l|X_T<%d<kOeR6o26w9ey$%
zdP*VMP+^D4jBsJG3YIml`#)!M<mKhZDg>(tge5yW`?Y4uzIB^<OJc4aFKh^VbRu%H
z%U1UDtFwoH`0!(y^R#rbAG(hNW3(f!B3v@$)4kP#`uQsX8L3FO?7Z6gQ4&y}J$v{5
zdNhq3M@DbB(;8aZOM^`r|D=&T+Ppt_vB<-{hO{{pJg&FyaAt1b;~X@7_41=lXv^Vc
zVxpLUK%mOvw+%n^bio69(vr9gwRmXd9Bs0$8bKOJ_u%6FwHq9UawdPE5OMo<;Ykih
zG>+c6bEmMdaQ*uA7N^EN+{fHcoQ$gP?f+rBFCg(wi}UOim9kA^OP4hi5k)Ink;8|d
zR)z@Ifap?TX_pm-uZc)H#qy2SC`o@qKW4aN4=YGKZ_Y*0M9KRAW!3pVGcjFOYpH#I
zBU10^#9X5G^%hkWZ{>k8ywiy+M@5hzP#x>pLQ{C6+uANd2anIJcp>9PU=%19<pphq
zG&wR#P2D~hCHH@pg`%ti-Y4-G-xWu${X258=du^%+$tO%*?0*3!-jwrfRQo<W(4tz
z7$yI%C$y3cT&L4}1h8nkxVX5+Ea5df$EnNGCVI4+SAI&`IAU}8bDmt2>d(VU`yVQN
z>fI=O2aE+~VPRp;UAvMzeLOv1g7u`P1Gi5AgpRfzyEmc_&;p8X>ooucjr`3Xy<`et
z6!wW#ngbK{uYZkzLq~aK+)o!8Jtg5+^jBrWzD5#jX>~@?a^?S0*@<)$5JrGb-7aV^
zVM{Y_m4cgPT|3OpX?2*|$BHf*L7s`Q&v0<x%x|lHeasU65L0P@%E}|R$#yfY-fdX%
z>=P(2j=$f;C@3gggcnW$bRrO4xdvYrQ9W!q*d(g!(7tM7pphh(2+O5PV$Aab|4~`m
zpeKLn!h<thyZ`BmFFrW*Mm>LNVautC%T`GI{VD@~Q!qNM(ZG8YHJR$kSI_l$?JV|Y
zFEGn}VlN{jGmRvXV_$UO>kF})ckbK?*3>#WcCqho19xz`DxeA+5p{mG<Wd{e&xD<4
zcyRqFw6`4^plBW)So1Lb73$xsoO4`Bm}iwH7_BE0tVT5^uCB~KOD;L5cR~@cZhh8X
zRnz>X>lXdS1=*KhmyY*~WV+Q=E=f=I?1S6Q5obf+DkJZHsTM+k;Uz+I%Q~UWgYJ2#
zqAQgDU6c4WKf-;ZqN4Q%ump4;o>J^=NJvY()Y?}uT%W1wi_62k=@2njCUfjzg?Wf>
zH!!*RxNYm4{T@ATO6$U@oK9|C?E>$|2kAmT`n^uycyik5V6@(fYfpKu;azj40mvZR
zwJ+e0cUOjas{ga#qqP^MUrm@RWL)})%`lzUZs8k@(i>K*bqfu(vD)a7mF?nql^|77
zqG#8B;Q8^V{+>fBJ`{7t*m8eod}|pNKgnjpK35g+w-8JnBKL+e?QZfe*9oxEm@g)J
z^j-I+d`;5dYi=KV$1P;*X^T}4^eujWr=oOu4LyBjwbPT0+aKryQlW7E{6-4GpAYGg
zcgb&FN!n4S%~MBa2Nfs}?9#oswN$@<TfyQF#?+-toIraO)vLwKt`p?o;GjfCM#6wq
z_qD|m%N)qU^tj@(p2?NsQgw;u0ndSRZeLPtzUS5Auv6Ol`iw32_;2HTFqNEqXjR@n
zUXglnma02@OB}xVM*facG-bW!P_6i#Sa?kAcRyP)M-mYR_^3jnUM`Etj~%rXms`j;
zP^9`*eBLu3=S8>Y?x8+fH6}~Ze#Wrq&h!fvjaJNA^x<X~9u7sAy&C3TjvfI)gfFq>
zz7oU*;M=uF^4`I9b~V=3VhDkl;DwOb=P5&4dtbb?o3cpCwc1C(g&nM~OS9~h+wL+q
zqNOPxtNQMJgGu(-7jd=I{oo4#zT>hu+Zxo#z#(-S$#1GwDsM|m3u(><=Os|mY2wP4
z%SjFKGtUlRuiIby`z`QTvfBCK?&`XsDPEFHmXyC&&w97(4CPgwa-k`jW_@r}PL5`9
zuzoQ+5acekRMxFq%K#Q~1W`(`I84IPA64I1jDz>=-|yu#nBj)%f`jHQ5(MUhJ_Pl2
zi=E%TebcBDHT`)=%;uM`fj0Jqb+oQpl#Ao>!Z)I9G>_HTJM<#r43%WriaZ<yM*L>7
zXOEb=N>l#=2v>!sf~j9oJ_0`O?EmOD$9pF~6u`7%zyPHertTpiw8DMD%*N)4N^JUV
zLS0}dRV99L316PtU~)QKA_xYDkPA_FKLz|>yp~oVpN<I1|H1Y7!X|wAGu_XW>+)y6
zMlU3TxKs*DHT@^4>T31NwE#UGx?blAL;vmNiO0PC*+0_9x*5@h5aK+kjS5b8k$3Fe
zmwz3!+z$nf>u|xR=CoBGeidy_&W%d0jXtpxLC?cCeE|}HPo4APxcQd}k*@&|e4%2t
zkI+;=ItfJQYt{6+K~U149Blsn_JYUMlmm**3$d&}GHrPX{|NO3;QH}44rGtlX%V&?
zU_2fOQNP6~aeN6A4;&BxDH4yw&A8^3fp>`1O|ccL87c(ykf43vI8YHLE}TT&Ug|V6
zsU{nWH2cPuMO%L2Djmp!p%%_;QXG=&u03?WOI=x$rq*roT4}1imx)h8?|5$HiWMtH
z{{AhjuMb4#eXiILEiAamC4&lzJXm6I;|3KFNFK;KlG5{kH&g-cbT|!kM;BlySLFdz
zG>uLT(g3q5=JKL(Ti%`ItKaVWNX@w8@+ROlJcmT>MHXh}e83YccoMSnwajylL!P<X
zoOu|~w-<o8Y1Gx@Ir04b+n92l1;Bx-RE7+Hvh-fC#d>j)`SQ6#+<bf3Z@bj;l_P$M
zGC9ZL5xb$(1yslq0X7bfs@%nS^PZ}Rsy8Yz3Q9`cRBB(QJmq@r+o#jn3Vxg`SFXHb
zml6<$J)jbOauaYz<p5zhbot50-(c?$tdfvv{diLY<gxy`cbCD?^V}{O9C2%gdg?aB
z8D3Hck_d*8L4uasl64HtPs}-&JzvfT50~;>PJyy3v*l<33b@s{T%W==XUV0enKWj+
zQ1FX8vyGlr;67Gz445btFb2`t0+!xtg)?U=?IeM+(|BMiw2i@Sm;7NmHW6RwElFtU
z5i<E<Q8^9NQ6G};C~Gl*!b7TUut^Qp-i|&M_zHv)(Vy{GlwdJwQ)FT~9cKe|cpIIz
z-uAy0p563uE8};@qRRN19IG+l7Q7V;G%JB{EUy3ZPB$l*fZV{<$^)7Q&srjm;l|`(
zrSDFe*E(53D6y&fCIAD|XqM##>(OOw8Xwo@l6YJ;zGLXuZmxG?g%!5)C&~|4q&1>l
zDYW<J`JH)<oijY^Etx@-ibI<Qh@Ne|z49X+^-{le$ZA)H$O9dL4v6gOo)A$ffBPhy
z_|JG!@n}N&2ET3#YKnX$M%Un%7o*)|6OZZv+u!E_KnXTz`St56{If*QvB}BY4uu$P
z5`Apkm!(uBVBEd)*36iBHG5B(zkTcG!u~Mx7-g#T(!P*+K=%Ry5>fvENrahnu5^?}
z&kqd)TL~eVK>}kHv)TLM!v{y*%XX&YPhmQThb>xe^zZj(6TJ%%Alof?5Q!Dya9?-=
zI0Ow(2c**MF*1=;96B7xQ~0m`4LUGKR0DT)BtzR6zmtSCjqioosw>{+${$3O(%x3+
zT(FL5*R(W61s0)%f`a+lRFvDddExs~#zSFQKtTJ&PZKK6A4k`VFADi6|1kaf!H_jP
zJ(e@2Y*R`qZq#EZ$ukZW9HGv=AKzW0%Fzya>tf5M>9YlAX@p{rOf!GntcF;RxB&`x
ze4hk=>S*?J&h1{wGa0e7Gk^T%o_UGB`a_{^A@|S4xyKY;c;qg!8dvp)(+>(RHor*1
zmF0h2T+z?>#v7`~&+g@puk{@VV_ELn^QAoG;^gGQx0b3s(0@OBo=w_Byws;rk(o8b
zU-X^!L5|FN`WITw>zlvvaQXvhaC~>wo*0Ao6Kxs+gAXq?wCq(3RVoYJ{Hyb5ucpB-
zUE%vQ`bFQ>65JJ}a9jIHSPF=h(m6Xts#_1!l^O?b4H>sdtI5=)LbsOhZ#>!F9j}rN
zRh>8mR>zN>6w~q!18E|Ow{b&JOG!F!P5H_jH8ZI@G}YEcYIHNZMtbaLVls-(&EvFE
zLEY{9z7kO(HmR!5yLu5~2#tNwIndZ2tM+lrJM(oNgN=97-zDcN4ncBQkI>Re2fZ={
zM$A7@2I1va^<kt$lp*Xy`lXLnMb1|@>V^i?p0reZs(dXKNtxS;S~|A@AA>LaM})Ni
zxiN<^H}<B<hHKY6r=F&A#9u54l@7}3b1<BW-TbGj;7r-rxhk<Zb~|Agx||S3j<*Fl
zw#VQwj-J4^?f^)swQU#}l`&drW}?AWpbB?AR?<B<5M&e%#PGLVm&FRp7^R4wH!<On
z&eNmNP|s!B7!$DT5)&46Z^ed5IwKCfVfI#?(WR@zmhbI|gcHVH#cerkBe7bkIEkCn
zJItjxQCh&;rueDOeJ7_M8Br!X=<_|k?adaU&wsSxSHJ$zZcb%i#Bc)F>v%tVvwQgX
zFkYM}n1v?|4$U<0SPQ70FJHY9z(u-GHBJGv9f-XI=&8}OP1I^P%z2r>m)2HkbG;?%
zv!&I+6UJ;(kMmxvI|}^MN*Kzv;f9sUHSB2eL)<T3MnfE_Yxh@L;mf`!e+w$(uTRO+
z`V%PMr6O9@IQ36xqg$B=R}&x62;P16+-P~6X38T%502Ln$D2HbQ@o7CZ00%Rn^!Zi
z<>3kgJ82Y+gQG1@L(qunZddf%1@709!&1j<%7E@9;xN1eIfW>t2q_9_`y1^pS9x*F
zwPHFlDoP{e+9pE5fsaW6QS#gCb5T5XeM3V^C`q&KJ>wq>7B;mD8j=Vqkv8-`7IBn%
zh%W+eN})<fAI=At-^gryb2f6#rc;|<oP6(M_-h^Ychoi^S-KU~%<mdf-T1rO?1!PS
zU&3tzi+&Gu&gEy{o~Pn>XUq;}If93{1@{fg2@$UlZt8oRK4mxo#{mX(0GG3`B}W_(
zSF8dp1V`{Tv+ResAV9+>SavqsVZ@}LfPsM1EvLH;*MoE&qMlE}db_l=G(g0p$&+|I
zp#KU__b8a*6jE<j3$TSMo$j{Lrj6-T7TMs^H{Cb2QL%WBwHTl4>mM6Cl;hcl`DjNc
zBQL$-JkQ%LAFX_gr0q`bnO{TgiHf?xS??dPCaPRBX`p_Ari!yZZ2y4+>j5+bYdMMS
z;4IXEH{e+!K!+CwqYhMdLz+1Uct<Z|Vys(lP6AiBRNyDrJ3cZ((jUe0?F1oA&_^gl
zT&A9L<1omeWz3Z3=H?*yJtVP%BGB)#YCfR#65=r2(WeGC0XOaT5^gCGFl*0ug%83h
zKVC8po7g4bR<Hwg4ELYdFe~%Voj$>R@bJAU{D(>nqAgM*_v_{PFWj$~tVbM=ZN;`N
z?A`HQCtj9G33|i}?SrlBr1>Qo@mYl?Wwk3JToy{<It-6xC7vpRS6Yp&NVudV4Fh=v
zU96pVR&Rd}h^%l*4hl8Dfw}!kDZ*QSiw&aQvGZfygSm^h2+|T9WA!vM_H`_LUCs8b
z1_0t}m~D1MuWb_D3dk`{=S{6_XH<QU6&;|%ab{(40b@pxu9}^uj{()pbZ+;O>8x)c
zYiHwjR5SBl@;d1nb_Qj9gCN^WgJ)GW)$A_~MHJ5D4yO-G?3vG(=wpByppY>vDB67V
zq^4Y*r7;5pEUWI6pQdNC-C3gN_$B7}oNhH(G(@)0I&eCEDyWxWHmXfEKkDo(0T-kj
z9@fyN8u<M8LPEBAu<#>Wba)Q5Q$iNaUYO~})h|_U>m+;$E}#T>rxZ^P>4o4iM<*vZ
zlam(sX9<A&<Fe1eONH*dcM{N1dOluz`qgjW&~IsS%aX4aQxO+aDo&lH@a1yXq#qC{
zzZo1f#;OsRRW{{O9#oU)Elk&$Uv;V=@YS|0G$M4n`sV^A?XNiAvRrUs2$2@#^RyO0
zk7~h=0?VCUI%^!fUGC?^Hic?lY%po-?ek6u+;%*>iQ3Vp$iy#<?&r?l0mXA1jE`n+
zD1Wc-%r6kM);_4mIF}=7b3?W8<bQijk3Bu(79Hxa-N5T6PC`?2!omgfq=u_?aRk44
z8O$zNi;?-sRzZj1w%GWw8wi4qpv4(p6eNybhYVQiF{HqnzNU87Q7M~`IB|ANS%w^6
zwB3AZ&~DqHnNZV!9U7M<noQCgD)mw4f&aVR%ub?dz(Lz#aL}+|tV79OSfaVYF{i0A
zV!N4;M03-yP41zV#sYd2@V5+_IXG%>vCBR1h5WbhTz1pQ#QT0ZTBcI#fA`6X9D9O{
zo*gn~hAr#>zVwB|NC#nbYa|&G)-h0`Br3UDAI}b=F!AP4D{77EU!(XjQ0fg3Y(Ck0
z?k+N35)z5#6Qz}vUd>tdMpBD0Nk&zpZTA=jEZUCXJi6xQN(0r~ft}qAQ(P%vuC2`m
zketTn>6sa3NPs-%PqpQiy{BZxUZc+C@Xtri3m(%Br<!PdZcIF#Hh+0pqrp8PlR(ME
z;;D6=CPIy_5rMCQs!V_IZxdov*4bjOA3UG`XdsZ@D87cqY~obsfRNzp20l7HsY+)v
z8gbze`mE-;K0Ey%bPg<mYCo8n>IJWqtVbWSJi6O~CNDFgF~q+SGm8kv&)@%;y80Hx
zVR>e08`iJS1H6{d=;0ocC;MqN$5&AJC?L%i;=VBc`g|GeR|AN$8Gn`sBOdkICTs|@
zI04u~TY!mUJ#=+FseTV|bYtZBf`ns|rxOUf&kpG4d*vRrx**iqAu5%jkDm`(H%g47
z7J)pY%|O2~iRfKvHw{lrq-L^Z9B&v<y6F<jBZ8%43jDtfS;+7l^s-_2>C-?G9Zc=Y
zjka&zyxFYAnX&WWL8i9M0hY{tDTKYcJUxrV$>n`DF(@U^Gpj%7pI>r-4(|c^_DHlg
ztC>F04CSSG1F#EPy$_^C0y5x^a3-Ygc(EbGpsWWj9tp6-EwY=NyI1}V>Ra+qGECF}
z+}5Hu{=Tm5F4?~ixm5%zshOV3%TdJ-Pk~j!;Q$oS=a!wH!Cv>q(MUqV6*6gj=IJ9x
zC{U&=*;JL7A;uzBc?>G_uR|6)d)wPz>~g7ZKeP<gXXz`5tApgn!AQ@8a2O<JyD#D6
z%<l_<h%IJ@j>yU;A{PF=RyH(TMZ1j%RR}#$yl>hglGUt;uNql)1BV0Q+i-(Z{PpWu
zG_C8}4K@XFbdzBg>SfgBZ2%CLl$7ic5fLe$`j<3(8lucI{4NL)yO6~%Y`3r=-d0E9
z!|P_;{{kW60r>$?xxB6N>d#~<`qrayaVMKJ_@it69PW57#ZZSz0}#SxE?&IYv+K~I
zL#Uv~i%QuQE$RO)C11j(*NG2)v^mpO)S@l#a4)Nn$rfBTU)a1KK70W$P>dOmQrTU5
zuEnnVb7UCj|EOO-mS)yWfgaDRk$4~AeIfaNe4dM_r&wxL7=!=V-rfCF)OlpZ%UyR6
zS-C3VWI~!(4b)#7v*v$Ey$cM{4mPIoLrnE#6{I0l@#>c+HUNHp8mMvTGTVV~?*34`
zy?jInFzIqcvBaNg04IRHz6nrjA+x4u*T)Qnt13^=_D+Fci8CS!I4`0n(YN9WO(jl&
zf&FEDiD{XOi6Mi(XNOel0wZ$oHJquJ;$@pX>mom9cvJO{Ou+$n4-Z007S>KkvH-OL
zgm%lBzs3(3IRGtPiJlOBC{IDc4~Ud<Ds~X8ek1G{@J!FfsUQ4ZpHPKvk=MF8H@wq~
z(;BI2@oa;GgTmt;e|mZXI=ig|%<aaSDw3~zF|JwjevYx;E8E^NBR-VRDR!1Wt1tfJ
zqM5<X>1pw2%{BuC`5zd^?GKsP8P!VJ)b4U&5j{g$=`hh?l;Buqj}aIe58$zuqom&2
zAPd^|>2UtQ$Bs!My@s!v|MZtJW@t3Ni{ptYZ~47YZk8E4l&htocWZ&cP*`X7y~Q7y
zdG^YX@l8)|{<>!JfJbi`Re1M*R~hgwR$VX)ozQEKxv4&7%yIH)dtzl?iPsoN`GxLG
z8fN$X@?~ZdhHIpF=aLztpMND42V_3WF67VtU&goWQ6h0FP$DWhH*7(6K7r8)wR=kG
zY0+N;@B(UR-0+uLI^Iq93kzlQ^Yh_`0u)+8|H#|>)zN!2<Qp(?;{ehCNT+dko|n9%
z@Ci|j3+gy9L}b4RW;p!J-uJRDPUBugL`1DY$(fldmj#)}Txa6=WX$g3-T<4Gh*1Qr
z`u+7M5clODGVh7wI}>`}n{qO~5%hk@CUEeU-v%YKb1;7Ku<U3J;s!#!fLxQDxiMoc
z8cI;yvmCW2clEctZqO|~LwlR=Dp=4qXs>|(b}6ZWc*F6}jQ+iJNyfV#G1o1)%cSuE
zcf@6=aT)44M_PttC?S<{*x$xK-tW}IKmfiAe7@4N<|V%5RB#3E3vem(xw(a%{S3T7
zvblgC7l3eD5&iqP4egITs9QM;hC7iU@S*IB{^AY8W4H)wb_5I=a^ZlVZ-gs$<*HT7
z&c-ORW9&wP`b1jmtzD9mVW7T$Fl`iqy~K=$6+{9JcN#1mtMA!2dV>R=Z~)OseSO!H
zcmcNVjYbj|>{c1Rfp+x9qG?mQA3RRwlxr8T^pATJiZ^u^4_(A@nI5Z>20XnB6c6IM
z64O_u7m~h4Mtkc;XN+ah_<Z=<_Yw!V4FYb00s@{`y{+gwD*Yx_AE$9S8ETdF6t4Nh
zE%;E{Z`9(Su*?oeJfLH}1Sa^v=C@C-mwL=kA}hERt1pNR<B4(|GBV>cWWc8(2IU5$
zo%X|sZUsP_x8^Z;bmfIgp8s*UPyc}sBdRHn>=|yV4w!%B*i_@SOUZx)wLRt1U7w{+
zD=6&X<72oDc{PAfr;Odc$&Y*dq-w^8McCZ}SS3-bt~D4nAVp#bxNVLX66a*h{uIDM
zR<wQsFdIMb1Qu=drLN@}u~obpp%|Cp>FLR%C<p>BJe+X0cLmMOVTent>vin)x6t58
zM*+m@crIT1_|gwj0Evg;2-I}^ib!O<GIV3LslpQx>O8^|?RcHz+(@yY`GUh>M(=C*
zT)Izq_@M8M%EDaOPU`Pe{c!2{!rYh=@&(#oacW{u`G$ObeQp;-GDBO|jzS3eiQvKi
zd&AWJw<pc+Pw2P5npY!6K_BfK8W_*<A6lBO=^OZ~{B&~-Su!DeUGFF^%X*IWLcY&e
z6eCIW@tfpGjU3{}|E&t^w(EuR?qtSw!UMVaO0#EaB|V0YZC=N#g<h9<hrj*x{MO-0
zQqzTd^*&~|BkOVq+Xs<OnC+_aP5<Ln;vxF4%~^?cQj5HfO#T1Iu~B}@-{G_n{PHS8
zNmcWLpr${w<Jf{7-QCA=ysUHgGO@Iq*)HRZe+|g3EHdoyE!1@$l|K_Uhi%!Ok@^v~
z?i^0wXRmF6MvdvLYO&?lGIVO7a__~#Ml8bymGdP;*o;J7ynk3%E19L4P3;e~<I><$
zV~#*?U+U=PZ5c1#4-~l2($RH&ljkThOG~;rYZqJ2M12CijP@B!;<s!^YvkQe!pQxq
zj$XGK&?kW^Oo|`}7_$b5yNze!@9S&PY)qw6$><Z<2Qd22K*wlEx4fJ>j&8t5WXKK@
zM}F|$s{uHoU?vDoDWnbBsi`SqcL?J@LmQVp(cGblqe=4thGytmmhnqS%sq-ahU>Gv
zb#Y<xOs;hO+sRQ=4pFO>A(eJTQCLgSpK^-}^BNy6twM(5Tu-9BH(N%>xX#eL8Ol7#
zJcxwsQCgBOK!UpA<Ag_iiceB(zu|<w;<fCSTOU==u!ypGZ@V>5!6*tC=0ud-u~qVO
z68_u918di=CFP&Un|wgu9e7X(DT>=IC?sD<W?kjoS<#T=Bn)Y$7e=b!fTu~!GJ^hb
zh84vKl6#o1a2Cos(Zn*WdlTOo<FsS<lUYo#t&}{_rm+?H;e9VBv!j4A?ttw##Lx%|
z`Mpq`NlC#6blTym7r~T{&dq12<^-;zr|({>Jq6EOIZ}2lMo7i8nRY_JiG<>U87zCP
z*<ZYPp=QF(Jv50cU@rzJ_17o#l5ifvI<Ic_RB(cH*3D2I9pNNKYvs=IiLvpP*ljoG
zSxB252H5Zb$i^!pJF0_09G*zs3CoX#OWpiQ@x(>24RsbUfkKo{%-2w>2x(5nF)hF!
zGW$rNIQrGAtq`h~%<grwT>PsX+{Gl%Q;Ld0&|Zjke}#1sX|=70^aOy=qGx1eB)N^C
z8puubdU|@Dt}MDGY%CZT;mIy8R4xWT-_#RxQOYOy^?D)}rWdr|wfw1j{N%}%LuO4s
zus1#Y@F&Vt`})m5pYqb#=0t;2QSfdx>b}9nyNp|@-Z`gXp!3h~%x^axrW5Pm9*mit
z)*PCz%)W((hR4f_9F`pFfxQqA8;K{lYlorgf8&AK$)Omv1YO{&{4kpV^Ej4?*!|f;
zA<Q$p2m_TKujX*}J@J^!60GZfMR7=Y#>U1-+VNUz;k3fOxRW4MO9j)8lIS=P&B8T-
zf!&TvT!2H*d1Mt1RoB-SGB2Z}iKtg$0DSrzJnQgWX|-5kab(hV^YMA3LSt)Ib9kqz
zwvuA&5T=$HXvCw^f-u*tmp)h1FRARNlxr1U^UhHTt?d@W7-KVxEMj$<xgKC*N}8IQ
z-f7&NT+OR^mGg+BftfN~XM$ysIkFUCgW$JoEdM5|IX3g_qPz3@6mG65h`kQ;<MkT3
zw<O85pq^x{b1xPAcnc6<QYs>(BS7@xtwBLYywk~PPP{wE8#9spH-q$s<});N>IE?T
z?AH9a8uCTH-928RCcDz+#^UuuDg(N0CZu)D7xl`$1`D*_jL0?p%9Nh&+JNU@fbArT
z^~Dk*D09Ob={YIM8Gj6;H)wrv$0Z^tBkiAL+!5a4t=cGY2pLd&M(VW8KiKxGN=Qmh
zBbk*Hri$ca#q~VCli=RLfkCom{=x4!-5fE5hGa+?qG%&WXv)gC)G!n+v1xeyh7AQ|
z5=>mo`8Uc)knSV3&H-q1Z5^F`ya8$;JIrbqN?M{Joe5`?!6>Lwj?Y+Va?V|(X3e|g
zWb$RUNL4%x9qbLQvytRh3h$~EZoL90ZX;FsK4QkU?(XP*kP+`<XQ}Vz;(9>Xa;$1_
z|4%;I`w3&y6oUi|$o4|ucrC!OE<vyAKkFUh?_eV(#=ol&-I)waL*ze(v2*?yokfbI
zWm0BmXNjnwd!ug;j4iUA-1v1cCJ+#g;5=hej$ThH+%=Njvdcn^WI#qZ=BkAt^sY&{
zt_GW524w)1*jYVx)}squg`kf8LX<^>OmFeQ9SGJB+&ih5#<YF=7bq9RcLCH5Z$+Z~
zFV=%UuvzNM2_2U1D;&B$xw$Uh6Jxd4G95DK<;WFb^XG>j^2RIL1_!_3DaKRwT5Z6M
z!<*<lZS}kEU5G&wK15lZMv}wAbT28Ps(pK*6T<lylkk;0FgOYcF#*&qz;3Kvx2_Xa
z_s?On|Fq=H2G>wp9DCXEiAlx|YBVX*)^nMEiZpJs#XLzBld*7QG?bW)I{=8v3A&QH
zP6LN2gt@oGDzxc4hIpUc7?^KV+}J2?%?_`w6N`b{!xh?63o5Wpd?-H$Ci*{0i+(V!
z;etNFdidH#NDyUgt*|?o)es(KN{7_yt@)XA2vz+3{o^(5%|5oW7^E#)Yt|pHZ*&v&
zdpz%)63nNo7NkO^DZx?8>9|pg9gnL2rj;QhRqHI!8FA#;vE_(t$>gpyo0X?&g$9iL
z@{i&;>{c>q&U}h$3gUsG!g~u69TbQn2}Ms&m{s4}yws($yDEEeVfIgR77I2j2{GYi
z!$W&nYjA?E2H)dMYA?vd%-5`2{t=3y41|ozT;QntVzjPt{bdi#`lOsVu86|hi+BGE
Df=)bT

literal 28775
zcmdSBbyU>f*ET#flA_Y32q@Cspn?)AC?MSk7<A{*B_$nFDvC;r5)PdrEmBeggM`!|
z-OnC>zjfd1eb=+z|6kS@{jQl$oU_k9dtcYJ_c7|eh6*Ja6Bz=5pj1;;)<Pig0}%*3
zG$}EBB0exN2Y*Ss-O_i{cC>Nxv~aOT+_P|VvU7B^dt!Of!`j95iKBz4fTVyp-^Isn
zZceT;f`az{bAf=Pi>=^BPL2UQ<cyQ5fhz(*ZGrp6d#_mV1c7k;siu5G$17`f(%V>P
z`=n!IghG}hFZkl~+0-W-Yj+HuYCU~gVqjSF)?n<9_~6SY26_Rjh6a`ChPs(HmHL?m
zPd_{ncwACb8i>XZ{?wB-&Hje-QY0Bmy0;>G{HFPQ_iol^hV)|XevWH>hb2RJP$c|R
zQxuL7;r>T;Lkn?%hK9yjm@JS3egtmg-B(stZf(1Yc#gY{f#^DXOGEz#aTUIuwY-41
z1z#J-;OF7)j(+g}|C_mfc4vr|>+`WlIW?|mDYtgQ%%wh%&oo(<V-MFG7JMd!D4JVZ
z;0m{;A8nyzx0GA?z*ZW)<Kr*;6i?+wj%ayxo#wvL)YjJCeXp*p+?|iOIy<*kjiH+F
zd8wkSs~fkU8*n0hFH`K2nAp$!?^$}fx^#?;j5cWxE`Ip%!MomLXJzb}q34qW7G>jD
zFl^HDc$MH4jV$w}A8*GtMMOl>6$2EG54KTrA|9yh&ZH}@EtHbz)zy2ds-!L9l$f6%
zw55OfZ}v#KtxaT<l?lN>-WybK3)b&(NWA1%K~>t1Yc(MKD;=jbP#&w;?d`WNzrWZ1
z{mx6|LwzSr;H~@eFd;GV?81Udwv7AUdcf(+H$4fLRTih}WUap^zp`w5F1P(to6q1A
zVRT#b-``yo<KP2APUuGC@yFwp{PyFOJsXU%%#v~3_cG6(J$pSv#JYNa{^kDmh?y&&
zC%O@jkB+>wZigx9isn0;o=h)q?{fb`4}{0sgxdO49gkVP+|<;RnEkJ0$A)8xk)100
z@%G}Xsps+W)DJ)E%)CuAZFZ=lU)<zb^bp9RE3<4PWAmP)xu#pBnsV*or|AH&%vhNP
zLDhr{d3}BTSnkFKro$T@L_&JzOcSip`}bGjrS7+KLn4#;)@<l6a@^wO;0V?&Hr!~V
zH|zhPeFcF$+E7VMOsrnW@%36Ru5#fQwHp<(>CNiQyAk4bw9zrv5TLkTyGiv@!rpS|
z<3lemFGSq+nT3UpNSe#o%}q?28PZ~++L`mpmCZkerStj=AG8Y=N6RnE`|jjx<|=gO
zD(2eec!z)c_HE^>fc1E#J$O<WF(s3bMKd8R)@X$-b)jzYcL&$C3F%)|j(x9`B5<!g
zLr#tv)KvTpkFWXD$1g3N5wCC%S?_-+;<k9dH%q!@YiV$){}sXVaPb?9mT>7`ed)p>
zpFfMS`C)IsW3Pf?;a!+_FCAXPg2%AQXNp*V$HL=&)cJbWAN=9<*;>TAZfwl5`s>SO
zUS3|*Bni7;V0^r0&{KY}wbxvN3Mmzp4a`j12celIb)!Hd8)u{H_4-3F?#x7u>u3A%
zRG1}p@lCu|osT;>d`DvtfwcdDn5NuDtF<Bt4NnjSjkCI_O}7vN1KUMx%6ozJjosL%
z_OS|E$`+v*)Y{js$0H@#77XDJiVelzxGwaLbsg`FI})4w?Tpz<4_A3@%o=YB7Oh)?
zuZ#Nm`^SF%{CUOO{?}(=1SUySp))5fC&wW)Je;Y{XZ!oEe=WEku{kFv=iHdh-?>hB
z%>M7Ru}v*)?O4MvPZ<L@mP;FK`bu?+Glk6?1C5s-559{Xcz5sR%7TG`!EGNOpEwDa
zr^wBlH)EPYNNoGY8v+{e7*0-3+QF|}KRZs>*U>OCW?(AH##h^8C8r+$EHRPwoq~}(
zLA3w+@~mQkRS`KsN=hn(usS_C_O9RM9qSF7jv6jDVum-=Oevb*eJR1;aYUZGx-wQ#
zv0(k^5ks-z7m=f*qnNq5xh^CXht}W4vt^Owi-QH!3%%JGTo{5N<#&3#>&<qOE$PUH
zhle-eF>cSv&26Izfu*UhPq5=sf3)v=a0?|%OAg6tKT+Gh9Ba5+dUzQ;)qi&~?Ra8v
zFbS5Z0_NcNjGBUikR{^frVQsn3$^0?4_aJ4yKC9<eIp}j;08j|?h73~J(QZ7nvKq?
zM{}J?kF2c+K0ncxaF|Fh^e;871J~$9BLB{*A@_gNA8aip9go3qTxMI($ha*wu1%t(
zeYPHZ9P=Ah=6i3>k9Codkhreb?V48hf<1_oBB<JWdiY1af}=8sJ-#33t$2hbc=hVl
z*6O&VZ%#x6Nxb5T6b}zik$ySHbl(kSb#7LCI%mA5rY2<-mCgNyT-Tiy13L+$xH7l$
zNt9<&a<X#Ub6Uau16ZXBcE|C`*mA3mzg>b2FS4?--k4}<X~iu-IDx1>m#$Z0EL~|g
zmgu<ZIqfgW?!QH&6i#8WwKxE)o0B8&`vc|aSftxAT}DF9mJflVX4ni|!f(W1y27Ss
z7Q&h>*!7W(4Y-kMoli;SX-<`G9~(srx@I8<1HY)bgLrmDjo;r>q`j~g`3*mR+xEHf
zry+55PEYYVd|IKjPK0~3$Lg=K+-)L=zdgufOYdSMWpz5pxAX(0W{Tldtm?2eAtHZY
zGdPq)O=`gO_E#$U1ojWYrm_bMGzZ_^V?H`Q&JeP|7k@U~3ZV{S@ZWCHzPUfeD4qE8
zQtg+!;~|Lh>Xpomm<3o5yuQU}TAN<uy5Isq?k_H~)J)YL$#R;SSB$w7J%cPn!6Nkn
zR&H}^P*c0y^47$}M7r3MbH2N4b&s;?QP>leI2N@CDprA-9@t7I4_4D$#K$|ACz^wE
zT+?%#^&=uL$p0Jv8x`tsO6GR<fTJb|5}S5<6*l*fC;PSzUzgtbHFT#5n)ODVzoK7e
zjt5cu&d$;C;V%8tNrY};%FgQeV7<TZloy7QTO;dRMV~@fX~T)So0}K}UMq-n4B-iv
z(GakBN4|bf;q6!&EO6~n1S^X_{rL^45OCu6^ymAgU?OstBG;q64Wf_x&0po@<RatZ
zdaLZeJRKT^&{wjb_VLuP(pCK9!@~Adek0h((mr^CEVB?43|~2FD?iWa#k=54gG3^0
zcc*;iPLDR-R)2~2dH5e}HQ7{b%ti~&`|v}+l$A4C<VRJL2wuJ%0-jWLe_6r%@1qiL
z`jW<bnwlgmR;+BY9x?ly3$|6#zB_#E5SL4=pA8pvf?*P@{JyR?oaVGdP%k|8{IFZ)
zv1&M6WT2h*l@~G_#BFcb?NPC@J=;DIigqxvYt<o&Cwn2_EWf?vOHD@$CwqsgF>$hy
z(a~KSDoTJUaOo4q{Q1e?eCDtq$NYWjyhw6JA@fG-ari78QiSl6ADot!memknH?Uaq
z-L=W?my+>3+VA?o;~i=@ZsSZEh)00SJ^*<94weei<nw*sa^FpFgs+w+YD8ffFnb%A
z+QT)J&Ihe9h-w6J(M-6gez3dV2dn;G=fk;N<N*)l;+kQ}N8irQ_vLDK#IeFkQG+L0
zgPZ*<G7wck37C8>GjGC&By<C|6E4SsN56p$m7kw~;r$Ne&>vsjpK)Bi%qZh_6_)p+
zf<g{BIu2?{I8ISCzJKs$4;*y~5-lVf<sBg*dY@^sXR>;jCKC3!<K1anbYqa+Xc+)r
zQW!%YtaiGbw}^FDYTY0A=*XzMwfF|g&->e28s`?KTN;hAa2o>zQ(-?pznD+fT}%lH
z3GMKb7^{vr>nejUPg4dyS>I{t=;$CNCl|dP%VZ6Zh;yW$7nh10s%OHD#~Fmo10joj
z@mn^kw96PMHj=!|$JdSB+7fGSZjOP>%@lBY>>1|)n_q0Lc2lEeWMqVtj7%iQYc^`-
z)iZ&THX#ZO#23kHdL?#%lNco@-IB_Y#|OqOjiaj`R<o^9x50@i1O)}(Ja}`veRnz_
zw{V`kCAaqAPhnb70Klo+kaJ@D`|r2F&wJyVPXll7&|94x?Mwf%AN$mXOLjA>tM)o2
z#$#T9e_~-(=pQqr2RmG8n2xI5Nw8qJ?dMl3{J8((<*Qe};YJP%5X&fPGxylF1{a3u
zqAf1`&0Xf;;A#aZ_++Gnwc+I8Sscm_$bw8Z!T*&d3`qZ%Tgd<4ImYCfIVPB19INc@
zu}{{UwN16aEt2`dEu7}N8NqTc6PRBqR}rQmuo5(EYz`ibz(rhWsn`|n0tD<9Xt_WZ
zID&4L7ouiYh-qkO=p2IkRc`<uJI}_p_G9;xE~1v0p6c-6@Q@crYm!Au{0>r|vF*{}
zn2*K(i}@_F0Od)D9<`GY5f$wbP$~&}`I32UZB4*GZu<8>T9lQRnAi*mS%{)#MQe^}
z|KLCi(x;rfe2>6eWtU7uwaagHQ~&LU5_Y54?d<r>0*=KYMfFu2LtOHL92glDbxB@6
zo7tf{XdyQshl*9^g4qBEhn6M0Gmb?%xU^Id^4uT!Yi}u9E<$Vx4JF)JpVow^XW~(|
za3pmParfRmu4~tx13($9u<h%rG^=sps+#m*K7U1nQ%L9qtXKb+XL`dWCU@T4j(z^h
zl7@yy4Q3-u5rRgWgMW7msFVajZeU|$W0CWE5fejETUXa3@S)W?6EO;qSjPR24tfUw
zd+&xHIGrDm3%}h-Pn@L$?#c)B-l#+FL*&hwbie_VWI?kjpic1fJQ*MDjl=qoxdEMC
z>62JmT7q?0NQW=8)ZmYjH;-ohE+B#c3ue)qdIbqK93?)d<((gM_>r$hNJ#idb7p3y
zP5&({paU?A-<?--`95zSL&BHaLx+rfjsrGhS+2j=aC+EaT<X3y!3d;evrjQs-Vr>1
zYk4Fd$AIeUK4`rqoBZl-g$phAd~{u01=6nb#@fXt-+vRJv!q=aL>`fvA@{ik1_t&k
zSw(Fq)YR0*24`9#qCcp^@)z1N5Cd8xddkeg5>r~bF77BQCzq8WYCE6qxwZHJvc~PV
z>M5TVtZi&u#_g*Tl9O9uaYw&AqXz)XB_fgt33@>v*&f5#+TK2&{+D?}>F@jxJZ!XJ
zy?KA$4L{g>z}EIZt=84nVt04@glVJd8XD>SEmjX-l#z?Er9?&=Hv^P5D7Pfies{NX
zjYif9jp#}f>;PJdiB|+;od@W&2Vrgu<tYd=z0Ao;sHmuzDgLY-XIrc6Qw2ghFcd)Z
z#ep{@5RL(;zJwgH^XZE_@Q>D3Wk`1^@{W*oCXk1#(m2aj#@6}n{#`?@|A0`3OON*?
zCtL3b)+u+T^1Bv^f@}5n_p>CGVC!Tw&Lgz7v~aG>%}wm&cBP}W)oHQ+QsaQ{NJGHs
z$_|e6LSA2ha8__Dny$62jfS3HMN2D@@%O^ku4dUX$)59tyY^I$Kf{PxF3yvyD%g0q
zgL^dzQS{n8$`GM-yZ8r}nLa+`1uT>~dD&xQN$rhF{QT+3`srx76$vGi*hMa`FkE7U
zm|p&5kXy!MMRlaaq@sa$^KUb`V1I!oJHWhVR?ms1y4|TRV9uK+?SO%O*Xz<?i#+Ok
zDecM~LPFg;ARr~hP&oAcXn(VO%#O%s+GRD%h9P}>d1SD}1huko`F0Ef$O2w>mUPnS
z=qO7A8ge-v_y+~MLKgUpZC^wGluqoS(%9M2Di*4eAkp$N>o_ZX(vJk<zhvJ1kqpr@
zI3G<*<E*hDV*iu5sx{f(F}EMjtgkEJ76~CCdTA*N_B^298q-{WbFh<*b~Yiro~NZv
zIX(iKItQd>e%!I)_kap}kE<*J2~|FT_&}@C&s9@C(SQjC0PXrMGs}4i;h;Zw3}kys
zPA)D8|7k1>rw>bvGjLZz(%8UR27o+Jl@74VomeJum!eX5T6Im$@AN8Q<{yC@IkG0c
z93*T6mzY1KA8b&XgwaG-<~^g)>J$EgBopatn)wnZRP89u5*^fP68B{C5Iz5!G<d`_
zT*fodtgPlsbfjt&@P$E^KJXH+ll=vQ&yQ)!%gcdAIIm6AbnPG{Ajup$gIECmzW|UT
zZ4#-YbK%Rg5gSNXM+eKL6Ob&cQ44WwfRPFUPEYvE>a%bV2<U{c^Q>|`pcodsogT^Q
zZufemzW%87@{5TqZiEa!b#8#qA)+@Yl%PSn89il&{)snQ*DCCP${AYa7TWaHB+oe1
z;9+!RvX26rgI%t9@9GVIxe0;k_lWPwJV}`SZwZ$TzPj<?;uvJx=8letsXE^zfCKA3
zfF)hzs`U9rS9kD-guJlk?=-T%186m<axlqR{hrEy2OMw-;HhBU);U%gH#Eo>-)Rs^
zN=hFd7U~97^6qOR`0V|<l<FMfI)%`CRHHSd9!Qe%moxoOJGKTZ&T6rjtj|7X*6LgN
zeoI-MHo8qKucgdpb8CwaTjL5UfhB+%U@8PoTcm7KPI;t-Apj4AUFK9@zkdB*uhe|{
z7F~20cuUSup&tDc`;wx$Q=2<ee5nCc^i|$W8tfj%a-T4=Qq@eeRWk2rPf1dkZnpSd
z`c!pU84{y5<V|x(GoXbb6cro%)*nAadDzqO>0gIg0U(syop2>u>qHp>iMcExA)zNY
zOC&5T9K)`d0~|HiXZz30(o!dc^SwPxy!U*Dt;=V5&;p#Dou%hf_>Cm=Gz*M#@bg5^
zP(;wRundkRVRtQ67$)j2SX!|RCZ68R*63eUct6HAvBi6lLn||uMLH3WfcSo?seEs?
zY{$v*;ilbTJF_Fs?l`m>931QgFM0mEL^miX$h%(mtvW3r1FxAddMSvObkSUg``ViF
zDQ{Lm$Z_FU1oV45p3P*GOIAMmLI1NJuC%$iNBH>os_N?3z6#@!EB!gM_l&L8kJ#Dh
zB*W|-hw|j=B4d&MNbP~LvU(sWE4VDY{bxwLGVA;|$AHw#^>vOpF$_Xo;k!;?Ds_`m
z1N^7dHG^g*{Lo>sJ&tv_vxn6k6`{7vu+an2iR0E?y#Ark+c!I+u({bzeG75X%0+8V
zZUlS`_}Z_{hb~BY{#sE7aV<acD&p%kN$L(K;axxC%N{pT<#9^QohlYfQeJP^j`k?B
zx(|)d0eIobzJO>G_Huh!Fw@zPH?H$upISuQVpYrwCEQ)dPE{<9f@DU+!$}tSmzvtK
zW2lL`H2jcGs^px+N*TszZ_i`zDwD_Q3PS)+^i}64y58Ai8DzLRb3%m0GCg)vtoCkB
zP@JhS6{@K4A%uac3&SO5s`86FRB~S^4dwq-?_!%C#<4Z^9bBI<v(dxBKRs2v`(*8<
z9BzB;a`iM!)QISpn3I-4!<HiwbgO9vFUZXa_*#f$-t+f~*8IrBOG6IoST-wrBRn&k
zJT<TX6?R=d50`YoC57j8$bAART96^XszNM6C`D(AZ)3YoFK%Xy(Sp0GvRy#@{-yI2
zvD+JRBl^{!*cU0pyk<e3M$#n=bTkPZLzVVMIevdTZ<;JXDqVt29Q4b8kB-e}@s)8Y
zCA$I(KnM7uVu6d36N_!3*8cMJAtY5&O3JsOPZSw`xrVx@tV{@gG_n<Nx*HI-F+3;d
zSwVjG%7bEBii0NMN8RynW_}_&Xj`_wDs`x#*=;<+0LKx6UJOuy%CUZ*TR}k~5Tj6w
z%Z3mDdIl=>N=+X<dL;C0m=Boh4L?6QNUfOGsPodl?xhO_f%G?@t_NA`wf|*-+wy^0
zUVL=#M=04$7OJPq+9)dp-LlxL=S|A2GfZB=er1$$x(IyH9OBlPyfq0G3&I>YvfH1Z
zX8;NSZ>&B&IaIrQx7*YMVyg1(+p`#qQWzO6&|z+I@e~~RG^nuIfDrnq)OG*wZ|~v$
zYG*4Ly+N7zjh$K|a=QCq5+?Dd$@1#}=Z-*@t3CSL_4(OI+|!|t1GV0+0B~BeJ*P6P
zvLOd^iHV7I=)Aj|(Y)VJcps7c8vStCxnahFH&0#kn`#i36}qL;=~3O-P*miF7>nz-
zKT2!kkxb-@Sd#%~vRv$c1z71EGxHrUFR22}-1%W6$G4EMip&~v+M?;tv$KPi>W_$w
zjeQH*ih@lpomJMO%x~YtXJ;kKf*}Tig@XS<KP18xchC=F_lJccPS|v(CmbIEVzL3)
zCH-q}e;+5@Hiwa`X=*k!XI~1t7ZLb&G4ysbI@f&gN|{@&Rd-|j>sOuEBiX1Ges`DL
zSLSrgu(f!tR#1Ah%iv{N-@$wFA_8PFP$JJzQugGW_R6mJzkhHU<cf6pb$Ags0L6yW
zlb6ZK7e0OZv^Z9A1voRoMbPn4#Q?bh5K9kd<N|+=7q)75!lG{=0Ee{!Bz4(h`HDh`
zy3XGP!1En)=i`ZLVIY??IlgPJZiEm^5ATAyBjm9%%Cd=*tR3q2=kbRwrxmbW=P}eR
zyo{Bv#Cv^!M{Jjlw?0LBsGDw*4KDJl)>{-u(L73D=H{}45G0QP8jfQRkhab<GIpAH
zPPRp`BfkORF7fdN;(;q8fcuJAwNuD=uHC(V|9*)FXsfdTy8I4TtK3#b6MlE44T7hD
zMxOCz5|pHSckkYf`|xD&ot{bV+&dT#n{{wrhi{xOWx%df)+j-)TgXec2oI~K$-xz<
zpt$_Sx^J~+d|7?=A~EqM<Q8wxXkF`cfklDlqz}2bP_J~pAQw5iS~WdCW}9mbNlB;F
z^!Q^XC~D?x{#y?aka#m~WgQ(I9Z;(hxL>f6bT|cLGe8MfV}6zk7nH%PN<LEYFl-~#
zH=b9Y2`_frBC~A#$Rl(tp*msZVa>rIZNL`?>Gdw%ouIC*t*us@ir?uYFfqVAw_)c)
z_Wv$yQk~vxS?)A=PZ9^6$^d5q)7uApvT3~V8)pn)Pr$7}iB}3rgK@J#QAh7He~kxO
zpw^LG*Yyz|r(L@a@4#|l>`70Vg>I0hAMnlyP$ht$bp;WSsNTKX>QE)>x96mhBd0c9
z@4pXx28ZhJr1DvS*8(fVT?sM5paCfl`kN&PTp)1YuX31Nad#N|6lQ|*NJvjtwXvDE
zUUqVI1)9uiH(k#vDOqJ)nk(fTY~)y%0(yIebr&6^u)bRFO|Rc6dW#_5#8uk%W?d2x
z5a>wG$}*N>Pe+giTE6>gGBY!D|Cp|$b9h>w%4{d1;!01+UFH0r4AL%~4h{}K3-#iG
zNCI(H*4JkMRMKoy;S8>Ub0-kO)y&N=>K4)l1O(`o7<b~}xBG}m2e^;-@%G5l&kyGj
z5MJMEWXs@+K4>(Wg+P3vV|toBAmH>){XEp6qUrd(fO6LSPBnUO)}YWQzsmx5{Dwf&
zIp6{H2%V2QJRM>XKc5}uagsz^yzlz`lndBC6rVuo&1t}DAv|P@kmN{jniYfl5)?l5
z>!jfv9BrKZ{IS4Ew4qe-kMF9KJU||5Pf!2BOHOQj8aAxi{37?&O@1j?3Y*C@UEG1k
z-+OOB{!Hl7MkFU*E=s`QC$t~q5g>h8=ch=Pcx?2zp(d(x0rB?H1|NZEJ*B)NzNFyj
zHi6#+O=4WSNd~~FL5PEiADT>ru)bH%Q)~n@MN~B>P}=Bm!JQZ2&N(tiDy4v+=i3*#
z*lecojOdM#((ua8H2l0|yP$_NLAiMzB;V9S_Cw6uc$oz6D?g7R?r@^|6%Rl%IkB&h
zZ|fM}kjhi;bUUET{OY538AfT5QYqvi^tO*4r5i_G2nT%tA3ZZ;;ljn0v_{1DHe`rW
zMk6%u7q&$ez)qG7DIg3aZBL1Ya#kP;B?~+eTnlap8v3Y^_~RR}u^jm>5Y^96P}I(u
zs3{ZNFgs7l`6SW5i`A1t?M`suM<F!xRkN0rkIJoQAhVSPoFZ2|P-AbuiIY)KaEggt
z=;B%{ELwLYYJB$E?BaQ@weY^f#!mmUXQ@JvPFaFiIb34ldMbi67f>VicaD}0jUXg{
zCi_Skf(&K!Z0xKPs))vTU$?H{J)fY}`}E5nU8mVCx8RUuQ)?xQM$h{K_!{h?1}2xJ
zhhYqWQRIlO^GxDjYYALujR??O3)5T+O~(AQxYldwdkACDN?^lq@$peW{)bGkY0nNx
zrvZ6fz4mpv%&=H)e=ey_tKdRVqmzneUVydg5i_{Mhp7KN!V(_wU-<<hCf$2~A0HVA
znlY@hsqf#v=P1#5Zd^mELb^pB7RriMla%uP@t;v~7SxdDYZquha{kb`dYOYG9r7Lm
zlwjuvDO_zIZr<xS4zZoU+c26LT<bqI!k!UBiGb^!YkT5@xP<`dMo>>MeKB&9qQdv4
z;is{OQJ`a8B@5KM2uc!^jaZVBuhW)BoJ<F5N=tpkQ&^_N`6qapmL(`>L260dG-uDg
zkPnN9cxjS2$MsoZqTy^U0UB$#_Xn?Ryt@Pn*ZIo#5w-YM?J<`?z44Xt+bbL6QbRT~
zHP@L1p(Ky$7*ge)dE5eo#}$l^rp5mK+dMt3*ulwV>?OLmQIFa9wQD{ZF+U7b)56DT
zRBCD<4TPtsGazud`R-i`2wnIrb)K*T`^vk+cHk}GYG==%Z-Gn{jdJ68D(M!ngJF#i
zU?B=Pee_fsUS(6k5N?rfqHLjG?<WmOpi4a9zdRfQrwi1UQw7a(jx*aVJBR9_CIQ4H
z@A~-9p8PvBErE{gVfQMU2C0b!owKe0G8Isnp4Pa4=*ct7)8giB5$5X-nr>_$qcEAT
z&MgD^@`8VW?7$`Sgf5reP?TaA*_TWEq<%=rNBNm1AjhbdYa=d<%xY7u9iQ{#y>3Ur
z=XG|)qoNRCIFBw_;H=O0kw9Ms5=qU`efcA~sg>v8iqBK<^Qzx{<bQSO8`n`d9ZV^D
zk6#50EyaU)zWjWJa<!M5>iC#+v})p~3-0o0xV+LRC!BaU&;oY+rTz3VT3F``(nQyt
z?dizfvhiP0qF2tSkbbuQ&zKURdt(C|u29j~1!zCW%v21;0Zzm57I76njyD8#&t1+`
zoXF$&v4cF0egHC`3-<B6*`>k~-+(05U=^!k!(ILwan+@Kkv`8wTUC~FllZ3Y*Z;g%
zhUj`ibd7Jbi74!K=}H=;Pw&5Zk!9oOeQO{JSBl?djUTC_9<P`9_n@2a;6Xei?t#Oa
z(rf{n3|XVk{=GO1Ud$=D+J1A0`4Wc0PZ}Z8mtm3sHhpy;aiKH%rR>%TgI~zS6T!U=
z#tL4XMM2T!*xx86tW&_}D|4o^%dz_Bc_VjW1};x4Ln_~}#?v2@Xs`->VSyPug9&S$
zt&*dBXmsN?86OK31R$K{jKyIF!DU}>?0Po6p<p>;9j>?i_aMD{c=z-3V9BTirP)r8
z@HtmR{>`8enm1gyGAV)$2h8t8ltu*c5B{5RNjl6prA!~)Th%0sh_er#HkJHmYF^xw
z-yXG~<2IWde;3AG1V__)4l=QE?6;}=rLv<H=;eruqCNky8*cbd_4)euDX(hNobzZF
z6wLf*hahkW<6;q7mIVU*q6{(;q~yQR08P43L+37Y^UWb10i6QHl5Ka1s<QIjCuqYd
zc^DCO_CBo!doWjDrb%pM<Tq>FLwFqnI!taNBzl7z($bCMaUa8AE`<M)RQk8-{uEl8
z|2|h$fQOLXIdnd&L!vdydZ}ONKhI9YeH<M@y>&oZ#8&n!0pkRFL-h>UMJxT9c;HT-
zG@BqBL6mQIF?jOaIo#Dog&lbZ{4c!e=qT`v-u*NDTpnCMTBN(EKv_$`x_ib{q4S>w
zTwuX@$na8LEkEkOOiu-d6#m;0bC$5)w|<AS^<zyN13?eu|Ig(JSiaXoOV(HE@zf&B
z2UIHxAvfY>T}6}_*Cc{8caDxOyv@>Lra3Il#0ChAn8O4E6i0xZw1F-iU8~_NY-;^#
zA;R1_q@g~PfAilITfOnkpFX|z;K75t*)m_QukJ#00+|?B%K$M07yT(1MMAmmxG@Xz
zJeuj5Ni6_hi*=e1)rZC?Wx5up`f*x)KDv4U)n7f=b8KJYnma(^EC(HK#Tz6NYG5xm
z8?$XhzcgH2M4(m=Lfp*ydO)#tv}lEOQ%s!GoaTF<%M?d!MGm+aJ)5|7ys9i8sQ0ud
zQyj|M3+W@Tlt@6AwbHsZ`*UJ~H6mF47~TJL@Aw{Wckm3vtxk8o(_6bJtf6x6d2knT
zV|BF@usxS<@-t^HL7xa{&?YV!TU&04NlC7EIf!)YW!!w)HugrMunQ_*&B`LbH82vt
zq}Skfn&%gjK|LqPm3Q#CrO|0kfF8nCd|}9A+FoAO5dWVHma&V~Qd2PV9Z>W65sC;N
zlK^xrl=hrGi!yD<exJUvXZT}m{z2~dI<a)*1E(<r0~3=vlo;Z?H#Ro*4i7OHjBBSg
z@I)nmW;<?s*qPa@53{1$@F%B&7U>%;!+LUCw#GrCgY-@#9B<*|wNLSB%0VqV**%|`
zjW6CSQKqjbPfOXPqR>?x`ghDzfi7)8%feiGmD`dQ1pSl+5mQrBD3{T?Kbc+OqkDYJ
z87dyUA98U`(5r!RE0gNW8$nccWTdFDR-CK(u#%3ZK5?VO9e$B@zqHC&qjcj8SJzCo
zM&N5u8wEidkENxy6{lvuz|0K*N9+HmXH&~>oju%Hwfgz~!OBEU64ZOsaJ7qq9G~wo
zf7@t>9uu_a-u%t|!h;ENFRv*cYI5Z}g!Ef~-w|bbzhN9YpPFD%z2RQd6dJo18%`E*
ztM^HGkiq{)R<p{0XFp#ZuhIrt3rB8<saW*U_J2fNcrF+$VKJg*7U7tEch*-Ohv@T<
zpBg}XpP%5FT~N63Eyxiu`b5XNFShi@{tlHsUn<9RZGA)bRYc~EhU;4IC3?cwRRZ~0
zzdwvX*(yEu$0acyo!hexi3)!-%zx3?g@@Hl)jjzUaL)Z23B3{J!HLIfrACgorf&|>
zL0q_?vl<cq+G70XzZ=Ce4v`BhzE$VaJwjV{%oU1`+2l_)k>%lv&bZC4*38sYMtgr#
z_nQpKCb1N8tgcF3WU|YGL~oGhSMB!HyD@d#%#OJoO1s-_9on9^L_+ZMRMjv>YzI9d
z^DQiW0Z6ubUlzNJcd1+)E)OX}L=_ui`l?#67n9!J7EN3`$#f`Y&M?Wi1KF7;YGPt0
zPA?DB6}N#eiJYxg-<x=Xb1NPxBW^WU_SY-x=ke0*m(&*t%zn&=Kjo(9hOonB&=rBs
z>4W^RVnoI68K_Ot188XO?~g>agZHv{vJ2*-e0Mi|UP#3IUw<_87eGG*)MP(*nI<&*
zD?2XL(KGB{A}%lEZsK%QQ&B}$$>WtbxDI1yf4g9oPWMhp1=pWJlPCzEs!#6okcIPw
z^sB1g3(Uy2IF!U~|3sq|8?A!MCdGZnIOd>QMuiNT7WM*TM$HP|q-fl`=<rf3`s44T
z2p+v{N;{s}?fiOEUs;_3Gs^u#;g51gjIzgAA=Y57BZh5)-jUj5$>;~}E^bq6)&)Wv
zvrBSpbjKts60XbVXw*h;{py(9sC@fr7%5*k?$@4F=g`(;xaGO?g0}!7_Ag_<Fd;g9
zMb>&B#3zw%;C#kIA40OSr-*Wk2JOoAY^Uu98GqJWlur!bD~ZERXz1^y#baT8dGqlR
zY}@3E97~Uf9J5b*Tb(_8==}!cHmmAH8pH!~>m`>ZQn>&&s6NqcxH570)+#+63SL(x
z(JN{=FVb_~;aqq(9p(7_64hA9{--q~xgRw(V-MBva8vORcMQXu4#pu{HXFC#IV+sm
zdIgRto{0LQU47^`ZeP@HF|&2cHF6)VN(QfUCVrUNoY!_=qFY!Oy+!q+x&uc~46NTJ
zC>%P}#&<Y<$6^9#xi^ji1*%azy2*6Ev^jcT&u(KI&%M7-o=7FV#^l1+j=N`URHQRY
z7E#(#C}p*WofhfaY$mAN*!X#0@NE?IYKopI`m=bZ=mL@Lc?%o*(bJbWU=FQe<^IP)
zY`%0!HeNzM{3-gV>hwb7(r$vc^|lK#5c^TnNGi26rcaW4oS1{bw0<!X@37yVt#oJ0
z2xq>`cRJH~DdfGDMwKUTEF{>hget`eSttzPSGuK3B6f;}P}gX9+G2WK%Sv+NCd-7r
zOJtoKv*$Z-1ypn;;vmdykzDOeiNXC+=1~!I4MnNSPMAx+n9O;V`Pa4xufgG$<7>#!
zRPk%+#>vLy(6eJZKTjpE82YE``falAHOebHIDDDR9e!ijQ*$OGqt$GOw5T|DHtw4`
zn0^G{T<chkZ_{M%v8%<mg313-ul&!W2efMer`N-0+Mg+CVx10QQ>c;!fth%#oixn%
znVb$@V47dJ7U1@Z`qyMayXf2aLgIz8yY9;|RT1I7&ctA$lIBv5iBs{tu*MC0Y@7TG
z?_S*42}O|%QAWRbJVpyY)eWD0y{s#S;^YBgo1-M(m~f&Bm<zg}DP57l&H(2eTmw6f
zWp-w&sTS5T#}@5VjBm$lrD<TIdgl$<50H2crnmy++8H4V@V<n^xfQjT^;3dIF(W;P
zC+F{dA+x?KpiY*A8#xF^Ynga-ysh2gxw$yZ$D@YyBkmd_JgL;1=WX!y&bNoGoQQ+I
zvmZ;=DNG8a*h(`gp1sR7jEhV@i$aDr)VJE`d&#&m^i7T3r$<U%!VP1%TJbacx>xO-
zffN>1vmF^()0^GAx2A?r`0W1fW;SzF3B9GmJuOaHeq)p6g1j03Ks<7`P@E8>s@(l)
z&VY}{dF6|q<u!kX=`T5Bv}!cai|y<VB*BOA`$*{A8|)<J>s94e+Er6B;arOj=S5Ga
zmtVhPTcQ9Zl6vjmSYI;V4rHBfY6p99a<uXAj|p=igf16eKJ)vvmp%*=pfim#=`+IB
z4nW9KOy@i@08U=WWzPwpd&44j(@VUN8iSh1BUjAI0y`bO8G3X74eO)qkbUaw|M-LI
zgK-h+%I>DP6dFcq=fUk6csUX5=5LAs3xz3xLmio6H_35tzqRxePhSi_kS9Zaud7%9
zgPJmjx%0u?ISW3$>R&bKz$m_JUd!zP%=tW=pS+%lY^k%=Nq&9w>!#{>rcN;$0<0DV
z?)oYj)>r!>aUJCB!O0jL84Uwe?(~TD$*_oA=7|tZ(+LbC3^$ColRF$bMI;~2lwmE(
z#&KpP>uM4&`%8<+(O_<+xty*h&y4<ELat~W2J<t{jF1E<u{Kq<V5={SgY)@9S+bPm
z;lUQR<%h+7whamOJ%vANeY_!RTf!FP@FxF2Qq?f$b+qXfIcW(z$i<$5pgw>sk;?TK
zitJRaCJp_`G-l1MZY)Jj*!3eq3VJSpq+xe<-T{~_GO7~jtIWvAxC>PQDC*Bbzc-Xe
zqv-_&+MnK1kFrt9>t!|X3M^&oFP>FaujnyrtR!O4^<-0JynRif)@gx~aL)@Qs0*S6
zX958aTDL~h;OcbHBY2LM_7+faDdf?QE2Z(ETj1J=s-}@I0c&jbD^U+X`<{gTFP-j`
z=wv54@`-(_<YL0806=Lob4c8F;dQqZ<YPM1Xj{KqeyNaqz@7KU2ju(CuqZ~p@EI1#
zW4nRZM1W*5px5CwI#fXuPOnwGd^@xuxPGCAre>U=3pj9W6AFf5Ac=!SCD1RCl$3Oa
zl=LDG&l%7Go4qe{a1Cv+3ZW!9*@Hxy`W*#y9B{cs>ulxS4tB>5iF0}-Jd*7SQSk*N
zcT_%K;_L5N`SUy*P)<IH5cJGYGE1CgWrcp$r(e#ovchPd3Dn#LsRmljrRVqe_Wt$u
z*JS?>MizKuX~OVVH1xJi2}RBT<A3Aw*#?bJE9P=vBA4mTrqS}hZV@s9j3=#4Jox68
zme5@v`{G4bRmmAZ-2wt@iCu1s{XvpbULnx!V^uYdt09xq@m+*MY2WJIOtGlIz`zwA
zIl0>V(-Eps3ub4Ck4eSHWgX}xcTM$oUqR+jMeY;_PEQk}TN<y7C9$ctV6U$^(por}
zSY9h24b-!OG7qka1IoK>P=Ub7;Lt@LI+VGqB4X3c04?mF3ZPa_Z`H=Z6{=^@F*DtM
z+*H;=eR-x}K}-Wsm75EDH!~>%%LgQV5|wELYN$CsC!yMa-$w5Wx65+`!-Wf>(CY-a
zU3wnc+|D!kEF~F*goGGWJ6!@%8`mk6E$fLpC;&|Ze<$M7(&(WtTAG_CZEhaVWIEWW
zLnfyb3A>Xb<H`Im7ssjUo#4LpAlED951%SGT$V?8VN;oY7U!=E>MdNr>xWJZI5AKL
z&EPBEaIgUiNCVJv;+M7!I{+Hip;KUENY4zq=&D`|=(S4g^Eq`5_GA3Zd>SiDK_p}1
zzMaomX5CbK<3%<3_wbF|VH+$m1T?gM-nIVr6@^LX8hP+9HbRU39+dVn#Gld0daR&l
zBG_@2FQ}&A0jv)IT9P!01YL<wzbw%IaAyO0ETEM_WyM1l8mB~-`6H7x9_%*yn+2sQ
zvKYRxuvAtL&TVg=vPQ49>{PO2*TO3hcUNelsv;bAxzP?DcwxVxd4~Jar66ZaJFeda
zx-blCy_|Q#F2Rob?u;}uG-Sf{6N1VQr6}hPR!{6wLGC#QhQ~i-&+@cA{T$SE2VH?e
zw>m&&_MkTiYXf+5)@L|?;@LF8#ns+juSaTh;dpgSdNb{YH@NAqaD@%V5IiQ1z0CPn
zu(KTN%^o>ETOHvP|HMg&8uqF)>!EJrN{)G7uu(2115+jHd2Dfw$g5&6%7_#@#Hx6G
zdK#@95qK+~uv+MDw>#@4-X7=ene`;rU3S`p*D2K`^(*V70SW&4#wF~oBkb*#L6Qez
zI(tbh5Nq?Dy!V^rbCotN;uaL+Bi3`ACZet@KX0$m#vwrurjKHmOI8bGPJDO~Ee;>l
z3NvZDi#63*Ag%ED=}J~(XczZH^oz=zQ!)KchKXY-QBv$i+#n8Vtl*tq8$I-E52&B&
zT&l?US-R2}IbI@)y2%rSd+p+9zi$J~i!<cD0YUMy%}*%v1hn4M#fKmBHol?QdfBp-
zuA{pu9aI&O!m}&ZJZNIJo%j;TtDOIMXo)ZciD)=0RnMyMl)6JA12B%O{xQ|jK8<;h
zDSLU3X^LLz3Zc1i8uNR`6<Y^?ElWt2ssrTD2$fQAv@zo(`UyM~QVex1;YXK)=S1*E
zfh}v_o7@HCSSmT;E8%xvHra3ATpTyvEljJ$-MT?a5z!R=%EtGb9J%XSSZpr9D%GVV
zZ>o7pr-@*iyw!l2cURK4FjPA%nwP4J=mmG52DZKSJyxH>O?1!eP^m%i)r`OSD0*DQ
z;`cM0%japGGY6!(!C&otGoCtcRg`bsJ4<tzg%JG(_4=w=`{3NoxA{PZ5@vtTQvt^w
zz6>17n+ECE?yP!8`EP1jmowGsz&4tdbXPF+JxWa~srUW95rV-_+x?d@z1`CcEg8s;
zsv8yh7k%X1jubmK-Pc?6LN4)6PJC5X|9H%6abWLsa4|VL@RFiZ>ph0>ruX|Vr4urn
zwqIE@4wc7^rKt0~g7i24WM2pprY`m9&KHbn^OH`1Wm=zF{qS;k++yA^PBlB<czE9)
zVN;^eSc*D_b@QaypIx1?aVgg_g!f0*aQL>=Z8i?}38#*A_r7qM01()1%Ya?x)ZaTh
zD$2S$w06p$d<GwV$lu|$akz?zS@V*2zuA>(P@F6RsL--LbJZNZ3#}N=M&<{rPA<6+
zU(>H_vVd%>o27W~^)~6^wy>8ML7AP3oMvxSlS89?Pcx&NcBuCk^$)oG&84J&*$aJz
zw1~K1NA@PVoB_5I4^mCV{PvZ-<vV57H%6*h7P<{t&Bh9MIsam9nL9r*PcYK9pZX|2
z{YGDdQ?#^Pbl#t#<_iP9C$Bisc9Md#fge>#IQ1`?wN#8Y+uA8pCdD_TcT1Bt#r`9C
zMgtFrLi*-q|JP0x*N!?<aLbtZ){d0o2#>=dJVsOqMZdC#gTv2LEu<WlBnk^L@wqQl
zy_}?fK2Dl^v^>B`+^gR)M!_sOLN%L>jKw+)j#L}79?wUrc4AhkcY7ngnlz>2K|znQ
z0wn+Z$B*8&c*eK9J`I#7|Fi3oMfIS>&?9Mb!*=m>L#8+1dVudW%R1yEK~IwnUOi1X
z8mP`q?O6ScPHc@jd~$<me|-7tvq-75IsJ~5b!J7pXj+GtXa=UvK}l3+)A<h*hXaNh
z@1XFmju<47Xc-xtY<=CvxG1TrtS+#;e;(~0g5H%hA0t$i$E#{c_P<%H`oZ0cwY!d5
zYZgRSX#Du{+Eq;2<ACK@!MJNfbm`|Ap01)BJsZ0_LLbbXxewTNl+`QAJt>!j1~!i3
zKj3hjb*53;w=bi6o%_Oqm`l|H6vyIvdw0Oc{n?9uFD|+2P+w~)P|DQS*xgIH`I-$v
z5vw63$uwCRmS$9hq&GVlWiv0O!>u)I{jMqgjX7rD)o|G(Y;8SAulGilW`4fw>DHH(
zJ9po@<WW~de9p5pa6C^&XZsfSAo<s8LG@;VDC!kc{9F%D!W#)c&|4Pi0~>yO71vSa
zP9s-pn$0ZLWpI<}trjIr+n%ks#Q&aU(MqFd_A_q3yVaBsY?&=e+5W*AraO~xvWi2X
zIa)oYooe`fw1@nSuHWyUUou|!#GR#9uuc+C#(P3R+vOv?(3Gl7m?He)v9k@%=?--4
z&x9Ze8}37#dpAwo;h7FW$H4=I`LjJHKQp2B{sIwo$v5|U_Qs7lv9*13N>qHo*+1(&
z6Bfm|uy=D{mh9hX$ZGJy`NHy+FYaIUo4}(wasc=RDc!~RH_9&9>gt=@yG`gU%%REI
z+^lwa_M@h$P#s7uqA(gAy}sNFIGV##hhNx?<~lp|W5fFJ3H9jLMsVJXNE9a30Yw+|
z@(NTmCf~`lR~@BQbE@p2oSHf&NQ{=sN^unr8lFkNF4^8QwN4{1_CqXsP_TJQR!aA=
z4}7I+sE-O8(QW;Y&|ELsWc}k%iux@ltvVl*H!XTY_&*rl(5t*NfZIJk`dT`iM_$+{
z;Sr?91abAr^fsL15wO*44z|D!5rAa<^mB*>?<<$2I}(aM2Q0gOSP~M=EUwzx%ST~p
zdV5K;ubdIL0DDZL<Mb}&!5J;NB5#u3R+)Bv*If~mjI&KR-gtdDaH5K+Kq7||>d%Lz
znBBkUsB-1*OjDntzkP7kDMWYCD;b@8_sTS%0dhj*W|@B8wl-9y>%9ppkwiQlGizfY
z;CZ;u#w1!<3<R(&9N-a}QG3=C%=G*<o<m3J9Qg0`u_`FA{|Z*eBMZ=~btBbR&eHrj
z>9j6~<s&CShA?L1^wYv~ww4TW<2Co}XGjBdJs*qya{eZni>+GlBJ-azYZ(|^cu%7f
zE=^sjvv*z_7h_#Ndh=2o;2qs$u_CIUCf+%+^LRK*p|KVL6;ns{kS$KqXDg)0)%H2x
zuT8zxk8rNC(Ya(yKiiRI!GC6BWOcHY1sra+7U}}88e}w@HDxI)d|SScu+&L%=)rkL
zV~dli*s_4PnAFg!20|sIN6ej!_=Y%6=RT5v7wL&N>eNrVy~|S8R(apyjv+75-5-O+
zY9~MtAiW2Kr5vHonWFY2v8+pYbT3gjj@R$9LwJN<Q*8Wi?^hi}%v3qAry~NK4(zxo
zTL|_oAK&?&w!!zV`IZ0_ca*boQS=l0sa8U!|4MRxUWJ%DCY*|ue0!^u=^mmi`{!$q
zUUfZUz_RMR_a18-S&?r#9vpGumXDqv>Js%;CCB%Ze_rp9&;@c5&B(x`ixh?wCn+<n
z4l1M*8`+>tsnIsDzEf%WMBkW7^~E7XW%&BwjX2(3L94duGZb0naTSab4kfIuxPZ0!
z1I?@@f;0KE4?hUmjPZsHm2N-BIAQzxWP8`o;gwl8Ki8W%HbRebW2u~Cu6VD@k-pc;
zd&oD**4J;atk(Bw@Z{jR-)IAmLe4(>H98&*oY{Sjf?u6V_Wya%u{k{*I#zd)t7+nE
ztDkd#v`=_||CyBxCzC5)kNCw_^zuv<18*5wdVB2<i`aR{3{n5CJIq9{zQSJNf|6ro
zHMgYaB35xxoK?=xV!9M-$~#)~ZRL9_g#jaWS$QvyzttWu*QI~+^5P-Up5C?oYZNW!
z?H`ZjcG_-#S|Off4&U&YzU}pOxlmV~t>B<IOUC^glxe(teFtCPB8MiEG{-~ey@uL1
zE7S!@si_w}6y$L5^B1Roq)j+?>60Xd0Wqy(;7`KzL25dxk%!GCbNVrt_lmt3p?RaK
z@XbhRlL^t<_MNWJ!Y)&Erh5*u9s~yx)C=1Rl+K>6A4?6KIW#t3N_(yu-;QBug7%-W
zYUj&P-NCUjOh>%gr(<Yz>jn17sECvU5?~Ie6}z)#(@sxz8`_L(BGU!(+2r<kC$<R9
zgVM*aQ}@bFC4t1hPENsVsGXCV^Yvri(+!*w&Fd?;FOlC^7<EskV{*N8rZ85;slt$9
zyhh->ukx%7zrNQ6;gB|=EGy-obo_?>aL5UFED8Z!0**Au?TlK%5nLI)Llt)aWL)Rk
zKi+e?-xr{d^5BPtG2fudjI6A?*RLbsC;}8t=zE!IXk1y1Da{lK3#essps{?lBPfD+
z)%{wBgtU)OyrZjJ=?JS#1C@1f$z$>5o!q&EJ_+2mk~X8BS{g}z`P5`gCVy|~a9qs?
zAu1v=3a6yuxWm|HkuER%U1U%(pV&14opNwU>e8Xm2AogY1Ac@%RQ7j51GsW$XD5(5
zXOSTXHwx^yS0HAeU3(@(BChe`TQgl5K3^IIyMX!;LDIF%GyXOv5B5YULNrHM@}?u=
zP`*El=lZJRD|Pphv|4zGIn|-Z$)<wlTw+C+0w-rE&`4944Cq|OwJkx%A{<O9D&oU6
zX8KX%X#!P-qam;EKci1L_pPwV3ca{2+aei%vsnX3(JwiIVihTYiwQp#M!EI(A~93_
z2+6HLSyR-kX~PyDGltH=YVDu};TB7VNin4q>yA#orZGU^B@Ueez0gDD4d+;J4g$>F
z9F9cR?#)Jf3&8p0{|9~6S}9pZ#2?{pf0Y)X+S?8gTIJq5yfUd-E!H`(Ty^YMI?boj
zH%%Mfu!qZ_XzKYA#h4*{D;bkMqg@<5((2$jiJ>GTYwqcZ-P_w^*@P+>bPLwaiQsyf
zKyOim6Q_946PWn(rw;DCc8>p!LDBHU*l1T5xu9Taa{vK=$CM9fQu2VF2=-i$En*~+
z{^w(I2&1nOpwD(+x|+B?;79p1!eQ}@XRS|L6Seu^D{Z;GCE|0v`RCD?rg48}=PviA
zFr!jcVdv*;*FiAI6nm1038q3#_QLVD$3q{7xYFQUqFu?0h57kC@?N!)Xqp6Qb(0Dw
z^p+!b@auL~oPjQ%-H2lwDb9{elUGOMV@Yu>yP`5O&AE<exAqCwGMyjJtQXijPi!`^
zb7rY69Q1$kJ6ya6N$R8YE$V;l8_vWb?!A<F3tjtnbL28$zMxqUUWdavAB<~6Kmowd
zdp1&1yV1@J4PBA2amwH<Ew0@4)nhf26`GD+uxH%14#b)96&n!abdB%E%G`sjNHdt(
z3Hn5A^p<{65J&{zDaM9XTw&Q{q7GbKXsSATSo}9hoP^<YRf2-Rp{}08ao1j%adJ>k
zUAkbOvL1Rb)l^l@p^*$ur3{o=DA5U;X8BuugXUY8t-V-QSvqJ8f_61H!F(0h54rj)
z<@9(xz^0CfhKK+AEh(QwLk0XhC*&%_O@GzOEVw6R=w2TLKMYW%Lxq}Pk4dGxkXK8e
zZ|8Nry#M9*!{S~CNiXkCjVZHo>Ojy4|5{cti&a-h5fw0!^iwdYs45FljMkq`jXSkn
zY5Xgm*Lm@=S1E?&CXZppg8A=fWlx`DPU_Ya`aQ%xGf_FP>3#XW#n+K7Uq9|g5m+Gm
zb0(ouy!-;T_oCJ(=$yjCW-ijJDu1$m+wzs#;y3S$)VcJ9`6UMP>Lqg(!Q^J884-%n
zP0=H)Sk|xgbUjoEh9ll(sKn=snysiec5u6BGT@wNVVIteWS%YNTA0q~WWs|JI6lXH
zKy)s__=mRnvR)Zdrq$uZ7U^W7ayI5Gb1gx4VIN~YS4(OQcEL?`bS7O-f<mTlr8<6H
zoN)C`L7y-d<G!Mh6P%kSJGl1j=JGG5iLtT8{~79Ecel$$r=!B5cZa1r<fzTwPm^a-
zgnqyGlh=!Dwo3ZubZX6ugxU0knPt*pdlgw?Y(4RK9GCFp_FNXlB#p=hdQoDD0@L&p
zN6*{;J5qdUPo_%0rA6Ai)r)J&zdtHHa*V7c4_;2!HH}J5!IP~IQ&bFAqM!UFx+v2z
zH2Prof_nTPN7Q+*#X-t^o6%~4*7%H&%=2`Mu06{siE7i+aKKQx;W~%QEeZmv<C#U;
ziGBG;ImZiMUZ|=t6it)kxAYX2bC!)?@;JY;$0p%V_Tfz2KAq(I@iv+x4hMnHQ-jvB
zvygsoL;5uwzS?)VE6LTk@JE0x;m3{tefSXA7tXJ_*JExH`wlCXKD7T$8nXS>Y$&ho
zt=4zv(Zgs=#pq#|qSYQ=a`Bwhqp51ey$ISrHfeUzn1MdY#3!PQPD-SX+xDSpsBo;a
z$N$>nCY0CgPTQSJ&`l2;!%kEAp$-XJ(;s`5rvAZ^Zgui_{_}B1OQe3j$X@Rzioi5Q
zorX4NucL!rJ^8FVNaP-GYDdGUa_VHBYl!;YdekL0CQKjK|1Kh#$zS1)>{}<9B;~@=
zX0Dr|Sh_eN{Wp}M;>1;2LOBhS#8-B%uh2z~uk3l-ozK6rV5bKIxF!4UTF-ONEfehI
zyp^1^B=O71JfjvN(*|C30y)%K{{UWu>G@;)DAfe@xG)eNPRT~A@Dt<b{3g4m2Y4A=
zz<Cq%mRXY0a;3L}t?ti<W-qi%(RK9g6-)Pd);1Q%{d~Gk>L>0<)bPAbYj)&i95z=_
z(n>N<4|A>d2MW(Y*CcjTee1Xw;dO6|9BLr#Lg+b#(cQw!6l{Bxz1d&gj83n`UH5Xn
z%6Fb(>LiZWX0opI82|at^Dp)!IoE<)C*wNKpwL^FPNYNqX{gOeKddkQF8tZJA)1dY
zpB8bPI%tVGcuF3Q=P5~PtJSAkL>7Q+LHn;OSu&w|23kD4vfp&@&-gK*oZ9S=MQo>`
zH%d1Z>yBAB1w*S=O2eabnO`KLNqrYI=IfB_jRC=?kCQRZVVUoy*}Z$eD#h#{2vQKN
z**{33M-A!ixjr~J3D<J^3w@_^mhK^Cx8s&9Z8dY&V^LQx<cwk#`S0E#rsz0!>C5@%
z{dRSTl=QS9LiqiTqI{fn@SO?7NbZUAiv>CdB6D;sP0YHX=I%Ca8q~EoX72(MTaz{Z
zV!bx9;?cn>_9KbaLWqCqbw@!{nc2@hKdA#x)xT7s=;ged{8&D1WyjY}J;q;h7aGQ?
z8ea?#c~`ocgef|E18Rlx-c$L2#SwPnIgi!~UG{HO0n}eUbc(VrUp<O~rXdavjX#5m
zR3azi_6WHjmE4-1_xH}%Tn}M`Zmz+7%gI+Ot|#GNJ^*z6D~NmCtWN#WrjQlg5s_5y
zKJ(;;Pncptz`uW%s9<9#{IXnyIXSxemma+zRB!<5MoNW+y^IraIk5q=IU3wks*9Xc
z!u^v+>UEBi>{hL5!VJV)&%m5(wKrxoIoWZq(E7;#>s*(%t5q4ZlV*jJ=?_)~Wq~{Y
zZYpjwvhw!p<+k$(nY#B;s5}CwjMmwIH(kTs^d?uf*{tt&ig%430SpZP$Mt@A-XFi3
z6QMT^zk(+N*C<=Owo9%~TsToCgXS<8M~Xoq^`q$-OpeXWQK?FPbM7}<95+hPviei%
z>*d}lfM9!Ul3mWUpgH#u$0$^z;HYY;z*rlqYZo8ub@)q;*2HK=bM6|BYpO=sR59P(
zL&}QkvHzdW&IFpv_3Qh;6rEB@kw}w5B2k&ja1;%WIOv#%C}SaHj5L``Nui9Dp;Bc0
znKF<4Oc7;D=FC$>rssQk&j0zJcfHT^ylcI!)oNw<-Pe8J*R`*`_qD&DP5GARs9STJ
z`soODUx2F8^2G!~+rB+XOc!!5Viwrql%(S}a!IRdC|OQ{a2%NV0uCh)Ns5Sh%Da~s
zrR868l>eAw!<s_dWV3ep7CWPjo1MB-?#(Apitn)6@^z%e+kcAxiCz$?dd849ad#*l
zJt3o|T-j0Yl&Ryqw8)eCR_!A`Kg^fBi$`1C%OVYH@5)FZ^j;yog9JEL@xfx*0?cvC
zz$xxRVv{dUpRA5HWmYK&zQ0=Z#mRS(O~yXr++={lZGD%R##_a7lA6{(J@v@6dEqR{
zfRYOYkqd;BT8Yi%l^mNr-y=<NIr6%4Y09^l)>hU1K`tzG`3E|6J@4`MSs7KNJ5eHh
zbN$O&KG(HL$@!8)Yj2k9t_pLKm(%gz+tF{*(~!+k`_|ByJOK9P2k^$BBpl<n=lnzA
zMOb2&LVGN8R-Hz)xK@>5l+fAmqbM8w4Ta|BAD`31$)%-0U~35(U1lugY~+-NE1{=m
zC2!`XW($Ar7?nL4>lnS-qsMrS=>C^8yBY45gE*{!40XsH9}3l`&1XGFR=Xa2T1RmB
zc<axP+7IiK&9vWgGVaI?uweZVf*0}!Wa%?RItO$R{4Y7^IYLA&IZ*BG3f4pLWrJK-
z=i?_%q{Kll+BrBFkA<kZYRhHG!dhj~2jIvw4)Ra*uRQSEj|QQGTqZWYl#qac0?-}^
zSR=wTP20!cACx5#dhIgkgn&S~woi>`H1A}6ef=3FrD|ax@6(IYzbtp|$!)z`WmTJ^
zV*E8PZ;f;WN3VRbq7pM@^XAPE-rl)$r#oclrcK9nbz6;t4X9Szb*NUx<qvkxricE!
zDs4uj;Q&`>C2%W`OO834f84OVAU==y78X_U<+{izh%E3If1G1uXFoI9TQzAZ_D|nl
z&4Y_8#;?k1tUYn5v#saa3a6|S+e;l1{2E>hFUl#<#cgHa*|+aEF)=YnrR)elg%_fx
zc4IX)HKOR+%%hkw<N%Ers-LrBwdLi)^t59!fr{Tb>2RtuV!`fp3)K+81kUtU`#}Ls
zhzPZz77voDv7)A2ZaPHj)kc#susx{Th9jEzlE%Vp<6>HOi2X>H@)m2qqWY>2**`_q
zXEyWQ412zM<9Qt1uG=bf1BjwJ2(EaPBNS1*c$;B8T@*C2069ECZFi5R6@uEUqhG_<
zvhEL2Nr$Zz4-BestN)xdJ7)gDDQnztDYjLUzZ^@WdlN*mUjYO{@wJdut9WN;rx^l3
zoGe%83Ux!pqWM<hn55%ya|!q))ygsp7f)8eD(?}I@{>KhcyY3${?X*rx4?n}-fuRY
zchKT`2m!`c&nEdt>WCMi;D7Mw(H%e-ntHJQ$nugs{CT`AH*n@!_T+4B(5J*V$T|e$
zoVJ^A7KBgX`K^}GtJyiH)Y;cl&C8Sud(TIY1f*Q|>x&ngoPSK-n@>N+Y@jS=vccjc
zPDHW%0_c|Q3xAMt%4Bia+4-C<as3f>X^z@QvCKVHf`7=ZMTH7j1Kw3uLY=g#`HTkP
zPqftwp|%&!>cK!iUACJyn}!)`T+=*Y_lw#rI@+#ntMAP^Pd=TWr2o0-jkYpe$ZPsA
zzn+BoPTAIbo)3EHUYBL&vsoTpGO{3rKU-#9hj+tvzEEZf=UTRXy4-Q>ziFfN&(m;g
z&RClRPB<v}wD-%V!_IG0z$H{$b8$H*x2!L9%YNMC&PjJ3&cUxE>%}tnGiRcd9Bhj-
z4%r1Mj7yz<_FhSvoKo&<-MOk|$t~H!(w%fq^KjngfBv5qIpFy6r%HSoVo<JWsl0Xe
zRFd|$s_bn8VKU6T%eJ&B`Le@F*UTs6cB?Ep-d^MQn0Yf{n$<>>T#@eg*sAJ4J0r~>
z8UiEgjH5p;$(8*gS6zOP<>kqfGi;%nIR`c{i7~3piS$lyJ5HXnuQ#!iezfxB!}F8(
zkgmEP7y8)zr&w>}puHCpv+1()P<Y7FVaHsj>x><BqF~{ynNHkuSn$)#$j>s{sA!2E
zQ?9yaSRsml`V?LfVpfx2ZmyEwf6{GXwviRhE~uZRjx7XfO126pX6GdpMtU=n{oO_w
z&nSdmXU3)e<D|wMxls*rqn*bIh{-yAqQ|9dyL4=8z4wN!W{u(=^3F4UD-v?*$R2i0
zpWe8IG3R64GY01^HY(%+2<8i|<5|cEd=k$@8M94fxN05UlI@eRpKLR{WQScy^<z>W
zS$W>*{=eRpFsTpn4!m}$X-Dj07TXQ@3^%VIc>Pp2z&qq@7cM8xS0ofNKWP;!(6T=N
zOSL`wl?!(=<$X>s`#GI=h8%9Qv00jyzFtYWJD};cTkNwCbssaIFqvIpia}M?-|5|j
zu`}n}e9@3{=dy+N7GrrBVcb<AAv~WUvO$alY9dh@f&0l5<6ed`28xFUl)O_G8*3Sn
zlWcpF)24h|(yu&*I0NE0qwSuoi<sp_;<WH{{XM1;YlXpP+2!ATgj+ey7aa0XX{mqk
z%Z|{ydDLTv8ARgFq}eWiALZ}(zQd0n5+h4y2j(`Zt(lnYtY7_9jruT;yv#MI9^_J0
z0F$s|@qVMN0N=e2H>E3g)ZH48HLk=&QCNY&2dJ72=WbpX`u=9U>ETz*x}Hr;sk=zw
zMzUNqsoG=aGu~o#^aG9~%ilGbZ{1OTYUR^S-F?O|GRr0P-j6nU8jdw4OXbW?wo|UY
z-uWTaQLMP7i0tRlsbo^{(buJ1zB@Z`aLeVRYxCcUia5m<O=-T}5(w!~UAYC$=glF%
zfz*QF6QMHGgls=|7`oD`v5ulwAQz#KSl6-5ejX(QCFkCHpDqZK4wc0t(R|mx&q|#U
ziM~UzYVc*5GcA<rSBn3gb0k_#?C}I}2%XwcbDAGPnpi|)kF=z=odmh-2@sBt`89gK
z72F1Eaf|ynEP%KZ*JIhkOBW4}@4v1;qWGRFY?;6)ioJuc2ke`zCTuEOwNuSx!5nky
z3wNZvYZ<=d<u&=`!hK;dlXQpU&W4u8C+2>DF9DcS#%oKSe<|QLEiQ66!U%r$yv@u0
ztAYV%a2vOLt17$w)W+@QVj_$TDtaHD$OnXy@R7cM|32gl_SX0BW~ezwu{OvGI@qU5
z@JvwJ@(gJ@#V~Q^&?<FZV?Xci{?L#YKNhYW;8i>9{f0BYutSV*-ski)8{ZT8K0ZDW
zD=t@}Gp$);h@$#nT@#c#Kr?)>71Ry#Si!vWI<>?3PGzNH>HVGV6~=Y&5tNl{u17$C
z=wy9dZq;d8>UiDr;mXCXjg*h}C!FG~Pq<2jcjjGU=G(11?auNy3+>x&CTJaBVE5yc
za(Sd-;3iT-5r@QIANqfzgmZGbB{+hVoM<|kzTIx#W)^i~^L^7tGy08~nE5U$+59H{
zf9>->9@3yJ^chUZ-XKPGFRgiv8Z592AZpXKeS$)8>}-#uwoAKajD5Z1osqh8$v1|k
zlb15Ad8o%#Dn$4t`-`LciDSH6@`ZeKQE_p+fJqqwXw)=`6fjiYFCYL9*6jt4(WoXu
z3IvHEXAmyLfG{eoub+HKLj^lj0+AOsF=uY+*Ak6ba@fY?o*r<xU+c(Qb^o0a8BZMS
zZSU)YjG!6DMWMU9RrnT@%=`l=xS@I8u;coSAs~nBN6a=-ps+Hn3=wV`-ia?lR9dlF
zS+bzreW~iXykCg84Ht@p#8BJf?I+<nA>epcr>L@FHKpuZwSnZ9hYBmQ;u5`C4#Y{U
z9{XQvSqd!{ct<_GyizV!Y>IFh;Nv)AM$mT7oj$yT$x!LS2ur-h`d(K*ssjRp<f9T-
zC$_W7&t>fnHuzI6dZT!i_zja<@#9>!F?EM5b8?(`536sdM5#3VPEyw&ZVpUs<`X1^
z=QvsocxVN$W8U<u3Cbi^oIgD{aCg8-Dq=**P821Un~3?A7uZ9qK9)CRGBEKTdtia_
zd#6l|y;S|g6S-VQ8;I@UN<A$2gxO9kMXFIa-Qe>}A)Z#}l|3JHT~b#t2XvCiQ*{->
zxStv$JykL8f#!$z^$iyIwGGzC$c8k<_<+dZ>&=qxwuX<4Lv;Jm^}k<j^Hm&TKU0$Y
zwRZ~-7pd>|KD`_Jv$U4ojbOIXy$9-*v<^C!+P^(D%JbJd&8w#{@g#_GUt7@l;*8|n
zl)+``hpT=c)RxU@&A#dBh=Az~_V{j&ursvodok{ZFHNl&Vt6!VYkIEiXlu)J4)wM&
z4LOB<<Zj~6Z)K2~XDz!kN@b1=+aR7W=9XZmxBu6mt_(9@6S+u2EvEn7Ig@Sti%OIz
zZF&R!A@xWd7?{jM%mKHSjb#+l<vZVeDLa-ovRjJgkFDFQrLnssU1WUv%HeEDU>e94
zow=?^NHh6B1WIFwsiKHAyw3M)8&KQx1*AbZS)}^&G@%~`hnG2D=x@J#tT!X@)PqRg
z8v9pe;VOn4%>bDIEx{b%jx_({?9IzRB<;#UfltTK#!`TgbZ>rHQ7at=9e@ujclI=V
z=Q6tHtL85;Y*lpl-n~UJ^$TAPi0~!O{q*$?Bzwh*#u2+!d1-UXbl-ZcdKQBkwl>D{
z;JqtC$r^oy?02*esZFysmpDn5VPt&rYD!(A{9ilFb=10MeZ1>QP#tVHugGEWMyk^E
z5|i!x-3uE{;oi7C^_=|vHti)#R14L|+m1Yf;H3UqK9jf7`Oou=DpJ`g<m~?ImcRNx
zKi|(vdtIh1rwh@7w|E+!r(b&UA-p%yoo7Bov}aiCT4{ARN^L>GDNE<<8_fF|GO+rJ
z^5hf?Yp!-{|DL#E8@z5h&c2`aKCg@Ox9u_ele@N+OI}WqIH5|rX(pysxtm?{B!Ws>
zI!_-BezKtbGEeMh+M5qjRq>m;<M>_V<?y<rKP5M<Zyb>+_^D&0l_TOG_=>$Lk$%?M
z=G4V~dK`B?rfN&PQEy8-A=gcZH^KLpf7nslX!hn=OJy!puh<R@$w;~N(46X$uJ6<1
z;{W*kBz-Fjs3(GCLMo(orz$Rfxpu4Ffp=mqLT;Kjw%182zwEZZ*qypeS#GO0T1au}
z)$Zo9mB3n51!P)Q*Oo))K2J(dxw-nNlq}-^FGD9Cw!vX}xqn4!!uHClwA*G|cX)dp
zkMu5QC6iGNg+xiwM;4DhEKa|0X5RjY&)ZXhrxh#88`53=WY_+@_PPD=F71XlH}a@n
zYgqlf#hJ)`s6G)rc1kI{E8pkbhP|=2W5rt2JkjrlHEsR<L<QlsztsD9e_J7)8Hb90
z=%v*vF=2^o9^-cp;&(H79I|DNERNcoF)!wcvH9^yVnf7f<0YRIPRW1|0cw4=oQ3a0
zvKx*MKS-8SpN_lULmAN65M}jJMmoZ(WL^*uY;ttk%||og-Ce>w(iymY#pv-`zq3bA
z4b9H7S;YrIiows!`eCEWvvm!(Twm!wTlZOGf}D^zUE4PV0)`j{!FFLQgK1rg_>+eP
z6)Egt3RU0aEEv|9a>^bQY+@ehd{}o{@eJ?%>TFI>Kr|cu6jn%|TRRr3?Qux%+!!}k
zIrFdM^Y7nj9@z`p(is$8seZWFG!#ZlwGk{TMHP@~T{5D;GZV9s{(WpxN;Tc*EO_Fo
zkKX$F>eIU9wGRiX?Wo%1V;@kgk1xD`#q#ZhW|Y{o6<^P-y&}mZI2glkgO!V-`2g21
z!wgyuE3-Rho*1XVG?pi~R3x)sY#zGURI?;0BeRFLtH^|R&*1#U6k{2s8e4%;r^|a5
z=4%)eV@#VX-i5m#sIP8X*CKCt&@Xi|k&OWhA4m0-5&IvFKUcls=upcz^^<3BzF*WK
z{VYFE_|z&Cft%Poddl9~1Y4T2L~ZU<J$v^L0kiJ$;*9SV9wo`{`U7dYY3#VXb5;5E
zo`;j`%d*yY8Qia-R*Y|bw{5ACx~-Hu;@FOsYMJzwr6<-Y$C`0a6uMwgKii|Q-;dr^
zs&#DAA?^)zwei7U{iDM-^QhORbAC5q!nmTt5_wtYIc4lez|R;|#ZVkd!#hGlE|>)G
zp1nfVzT%kj%`mNV#H66D%Gax5yuQB8w`?|2zt7zzOJ_W1CMlcspd-Kjh?SGxd~4D0
zylYu^SHJ&#uWh*zR!)&i{aqcN_v2oKSI2LL9e+mAq-(R)6EbELueFrTT|D@hSum|N
zuShPJ7VbXY$-~yHQ(d#rH!G89;Iy7rGP&LEK#+U$#e>&AUTXB)VV8T4zRWd#Dl%Q&
zujJ$O^Qla&0~z7gT@H)(@eLmvt1Jp!t}l$J^Q-eqT+h-H$jFokvcu@D()KwaCG#p4
z7Z*v;_AmX5_CycQ9TAIOIxmXQ{|$OqqtA!T9g^IeRH_(Iw&&Zj?48rYCG29m!jk-W
ztgT0l-l)T&F6IOf4h|m;1-Dyq&c02|YA+ZgKtx;thgC9o{O3<z6hdD+3Vw&Mx&W?W
zv;kQz>zr{0xXi}DTo9-P2qGPWgS4`p?qPk00I~vYZU`dr=&k$&)hO(xwkjXA4Z>Bt
zdiCnGWPAf5!xEyZT_)svs2)BEb__;%0g1K}l?2?8_bd2HJUl!`VBAi9^SE<oBb#IY
zF$*x$y+9Za0iwqfWDHbmXW|4>0vxS~7I>q)5)u++dof7WW^s8+4hKjg&yj68a{0>x
zRK^o*>8;zh388gT*W~};7AE0}r#r6vH?`Cp+EMl(hZ9r`5U7Esq5p6wXPyo$-`{Pu
zRezt;7UG09xvy#1fsh6P*jSsoA+)L#BuRN-2PZCf_4Ih;<+=M4qGEKa%~PF%d+-Lq
z$|7X9r()LC05+dFdzL&_%8$|ZNWfbU!!@P=<oY{2QLxCi5W)`1t9uZUK<-*xN!?9u
zxBLH@8(O;R6mitv7iXoLP~V(s+hbN2y$0BmSdd%E2`SXrMux)h8M3|W_Ct$$;|2wV
z=~u-eo{fG~`;)qPit@DMxI^O=AzUw}%wyD2V`QZo0G$Altf>9Zuht^hMhM#3>(st)
z91}t0!wHZs9tp^g3Fdl4TL8wPKwZz;x@lX>377X#lSDm$$`XVuv5wgW!>0OQ2KlSq
zoJ11McWxQ6EqTv$P$V&W=!8bA`|SG!_uC~@rq6G6%E)yX8EH0dH~3YUzKH<-0H|%u
zpN&1ey$*jZJxIKmc`Nn3ZfK5vUfnFW;nwi@g-XV2ZY7yaiM=AjMTL3^89xoL4_AHl
zO2}B6xTd$n;Ag#4KjYbLe!HqKI&?rx<&+2WpKhV0RSZt?hZ}5Saj=Xr+^Ju6soBdT
zTi0gv#Bg4rp3>6&-tISZwXA@Cv+(OV|Mkm!XVCR#ub9y3NEQabYmz1=CXBF(5Fk9!
z^VQ`ZE~-<Wdv~>dF|aTS?lJzbz4+I^a2o!(e6dC0&p-M!s2UBo=S6}%tpchV*sho*
zBH4R0kK;p*nAP4uO1BO?^AuV@{8a$92?7!{yedKYgj%3t&6`vCAgqW+MCgiA5tHD(
z+;^<#dAS=w(Q@vO2`1Qo3FnFE(wzM<>WTXIA|sKjFBh3<Us^ncX-8nH#}N=|fxv;x
zb98Jh=}Q27(-|<SDh1Cf5^_)g*gJ?TRuhgP1)QO>Ha&O5BUETAQtbMaKqkHDz-P+}
zP7pd7(1U)ly-28bERSyO4co`S&71*E2;lr8qoj9NItq0yfe8ZIV%68LUsKIK-hx7#
z)oCHyakR9i=05b}bzuDf+@$H<z{s*7z3QE&077|BWU|MhQDIf>v6ZzMDgx>Cka=D7
z@z<J*Z^frO@Em-6eV;2l-Y>cEqka`VJ*(=J>eE+$$R{ZLr>dNwsX!6$b#VjekVI*)
zYN#|}B&l1d00<-$EhAP%0nrGJLyz_}Ku28s{C@4%=k9@)K{^`%;A7tShMj;Q1M<W)
zR6flK2?*Rp1Q(8#ycx6-m_6_a@y4Rc+e72`4Y<vq#>uwL{<97KCP7!j<t~%(Qhaec
z0|P^{^P~x~+5I3<f$qrzbBf=2bj7n=gzhfj+DFTB0EOR%n6T!$RZ7!c@KZ0qrqOHn
z4K&2xgCIg`Zk9hW$B1-{@K8<cR%e+TYd`7FX^sf9i{tg9d6p7sMBu3^kSV`F+TbK!
z<~p!{x;)tQ_f;nMaFvXYj-jqd6jyOPVC)-MyZb!f)&Dm4pcZ8hYkn6+TXqAV_z*RN
zT^?=KmL>mJ^86o$Lq#Ya0pNb+?+OZQ^X4ZQ#ageN2>8UG$yzEMhVub0rk%Z*h&j{E
zm)31R`UnaP*1swy4tgXpq539eX*&-$cM{0FNC2trol@wBq*?X*rsac&%niagdw2(#
z6y4jRwzhD<$!%!|sEDy3i#nzU9gnYV&!)8(c9xcv)k7Ucxh%R3yH*TBZx8)=U2Vgm
z6T%u8A6TLjNK2R)%#R~YgEj$LEIkT*$Vzgn10a93$B<9`#mV)cS6_BuS3M0qx0%Bl
zTUOX>B}K)%O~#-skTC&kh{VPokmJhcXJ(vmEOYj?Lc*PC!~QroP1ptn6`=8s_x1N9
z^>J1+{j%1y<bRd)Job5fM)q&>5jp*gg?l?)P(18!Atq#U<S3RtX4r?|gUVo%u?rim
zTFJ033e)q5_17?z=mkD~`T?j!#W&i5vCEetX2^&mF!6Yh>HyGTA~$sFfd@}+$pL%w
zBtC-l2Y7+X9)fx|)Gv7G_g{nb)e`W0cCC<DpNvia0~eb3{r_LBXC9gT3k@thD12>O
zb+?5~B2;VG73G2PS(~?ynXLn29Iga2&@#iM-JW_dvOU%Fr4iyhAVVJMxiB(yDEz^a
z)V<h}!4dqLcG*L(K1BG!OK^klOHFUZNlz3&10Kox%o-#Ik;%2(vgL}r;3plNm4WoB
z6Wy}k-;%|%E}d!Vz^+|4R_}14fFX-L+SoyYE9N1z&TCSf#*MV8R%ZdkV%>i(3f)$Q
z+Oow?g@OqWR+bO6Z}{@Aqeq{9`t*sG8x#>y4S*aW{f|Qf1!FZ2?(uA6JvfF$>;;zL
z8B7j<@>WtHn8NE<0m%wAmEOL-MgaUZ(cqVP61puKef$ZS+E;*^b-m`XBiK%@6@l{t
zBeD2eCNgK>y;BQlf%)i$@U$799%V_;|FB9z==d^Z%!L08F6fO+TA{xSX~Y0A)N|L_
z33S=$o%?#Q{L;=ojn553MYldVWiEtioMl5Yi~sr(1w`DLU%!$uaJn?MfH1_R3L#jt
zq%8B-)Sy=SWp<<)cSyOMr?#y4pE!QAxeP)UpZ1adB(TFJrT4ZGZ-t;#;nU1E!afLq
zzXlvTD{%2E&%;uI@W_jn8uu_K#y_bk+qAp=OhT6!3b`ycuo+Mbi4+gy<8AOt1tw~E
z6I~8&*|9_E^5x6Zu=hOs&%Mk)Ky6ixvQ&I>$OCU(7^R~__Q<yOH7m(Vk7Ajt#455A
z2pbGx;hi;`fv7W&7?sI$3p3EuNMT}B?dqAH?2(S=cUxj3Bq%te2%F-KK@HjlCOpxw
z(WRkZ51OFB;vRx`1RwT1rIbL7(yB(&@bd)NK7;-aU`nGQh@qijDX#UfZyN0W3#g3S
zSC#_+DFMc7;+^}?GD(iz_(-)j*drm)G&VLS1mhaK(2Tuly@7#2mk)15d4vY~sdN!M
zZjwKm;9P(MA(8GrerRcWw5CsWs+Sr8d&dr!UMTBoKhWdpbe4M(r}QKfcSSm))6HEJ
z?)GV(LiYJC_$E9xKSSH4s}mFW+q~EghHBiSZ>w>3o7R|IL=%_BL<4Il7Z(mxb&6uo
z2)5Q?&O9;&-M$cZ+E~CE!OU2R9Rc%@B7XJzMhc2^Nvn<W^3x!02};k8od68ZzUlMy
zNZTJKCOn|tt4*f^wQKe@M*;=A<<wXR1-qScQ5i^RWXP#Qi)}Fd190$nj-&d(N1ej2
zRfUSx=<{GD!bf14?X!F3GI<!$2I-W8dGFs~abQeTFfJ<65r?O?0vcE!{t(3WxlFc{
z&1TlQIVI&-2;5z8jTvnRjSn4w$dV&X4bTyzqOww8Y+9~ZR%3<VBO%bkVzAK>7wC?M
zhat7)l|kP1X`eZe@dWg;9)Y|&2`1t==G`QDq3G=JS|c%{@6zYzm|qeRanaF~cV-{C
z0QnE=5<#^mZHH=ZzD1hJ&WPM^wg$bybFPAaS?lY&^m7LZXccDJ58CSHs64P_334|O
z0n*!+vwD$A8(zxM?U&98;z%rhoB;M-JCCq%z?#iG1i}p*VN<3ZFFXVf)%)J@XasS3
z4e87Rx32_sjppsSQdpz!FOK?`7<C{Y-HEOGJeCt);skizeRJAyu~TYRWbZnBI8bJ3
zJ`fIto=P~-w?#CXA9g+!ge06G#E7FWLw-@wo-%qS6fMt#MlR8~nUhnItgK1Ebpifz
zuZmQbb)9O^AVOwBOo0?ve_Rx67$^SD33);?qwX~8C4+<nMnNhQ<MXJw%qS7PYM689
zk<7jj)RL1XR2VtD-+oCfD&q16f~dp1d2`?eeFPvc|9nPAc-%;&cjEY(R9T-21xGp=
z$L&^m3l4Co%0atE$!8RNiP>CrtYwldkGTHJ{O_&M88}37wCKR9%4coR$DRrYrvmI?
zk2b3~JSto+$)aID+O43U1_hgD`B!yVX;M%Q(4k++!}k${v50V@4%~gZl`GlBtT-_p
zoUN$4aQm0rvK|w9G4P#N(7S6R8{2B}TMDvN?0y9pnFL_PXxssezI553$E>q~K%SfK
zgDk;)<<dD}h<dh|$r!wjW%O!EEGF9nT#4`K#6dw%f=_RCG=P-akK}a7MZ!N?HWb*2
zZh&?_1_-kn9Beo|MPzhRl51|{<;lROx`@U|pz^E1K<K+K*kipEqxr&w>-uj5u8wev
zptnIJ^+eG@G7|i7Ivq=<@=z+hYV~RoNRkzYJq=>=-n^xSgA@|R6F?4JrGobMNKjjC
zX2uLr`jYZl808m){}wE$hmh_YVCI$7)wLFB`qPtq-XN@#K0ais<8RE+EO4iWxDWbf
z5_XVhjXBqa9x}mivN2H@Ct8;1-wxiQ4Hw<oSAISB+*!8jc!8Isj|yS=f1Z7fDUTGS
zMuE}%+$v5a{U1Disz8n<0G>>IwOUa@bl4Onm*_*Gx)*_bG0x7+VPi0NF*{e>!1Q>c
z`-Fvu{1C}}`c_^_d|!$<0gW)CvE32?E*ySTII$ceKXVGoSw!HvcRg?7RsOq)UhmS1
Z{gGaeS&CC)(DaXTMqWuS;rRJm{{c{gL5BbU

diff --git a/_images/sphx_glr_plot_ucurve_thumb.png b/_images/sphx_glr_plot_ucurve_thumb.png
index ad0d00c71797aac0e034fedfbd15a776832cb7d3..fc6798028b42d15a7df8a676770652dd2137bf9f 100644
GIT binary patch
literal 20374
zcmb?@Wms12)-5P0DbgTF2nZsGw1m<nB~sGe-5{tm2#AP)lz<3GgLId4H%NDP*EgTN
zzw_^0=hwNecfb3+c;H!U-D}=+jydKScd(+oB+ec3J4i@KIMPyL%1B7apW%O`7-;Yl
z2dy9ONJt9W(qb=F+!D5?T-^wlCR+D)Mh;~NL==4erk^v3^M^>?Q$$gkT`f$dYz_&L
z6hp>i3aP&J(Fxx?^}ztj&lp8fX(=hG9phr-X<=1W+6SajR@@cBqeP88n?2Ld#}j9I
z1fKP#DHHk|Fpc>7;fs8g{XH)rr9JQlgT^nDlKut!h%gi-2wx<~AQI`W2!_&Qf(IfP
zG$G=*#4u>Y3Z>q|?=cZ$;?wxyx6mTf`r)Ul|9|--<0m{kF$oC?w1fNm`!;c^kwk=q
zyW^(GkqHS_JIa1c+?e=EzX-+W;!KlLUT4VvOixEvR9BbE*Qp^DaAPM9ZRsKUh7;P-
zSn0eWrJzve`bdN@eq<Uos6$+dP>$i=+1c6VOkJ|~)w%uQsuBkWhn=HilR@DR)of{L
z>AB|Md*u}s{l8vO3|VlYqM|17IR<?FO1M6hCF<tJixk7AE0I0?<;xcl9UUs`i9av%
zM`!l`MMg$8x3<1cmm&JjY5q|@gYMx&|Ap2FxVig;;whRX7EB=`T3VDi<eW&ejs94t
zr>C<!I|@!toH0+d8^_1vzOlZY-`tdguT;y^YG`Tk&sNJHC^VG3bLS4P*D1S4P0~p2
z+sXCdh=`T0xMwJ6X!G0K;hOI@TzP#P>Qh7m*~!EHSPb96!^4Bek&=@WOY67Wn&4U=
z$(6OSc_NV(nUnK)qQtT_nnf)rB!s!hCN(wH&c$VUeK;Eh1x2yUW@_MkFhoY`me^Om
zTvgd*p?71o@19?sA1$;}`}~xQpwg;#yBj9a)DlV($7xRR=C{h+`OyZa<;bV3*BK6{
zyZXdWw1RZ<5jS_aB7Bo#J5%em*Fti6dC=>${Vy(xQDGqWEsu`Rm9q+S^cLS}jz&|A
zuD7(97z!UBU-ixP#i|p(L^x$|ax!&KPtWF5Rm^a<T2xF-t0eWiX1Ega>8vl_>R<+5
zQvlAwpn`CV9y>k>iDV+bbF82z_h_DWnr@x<AKRHEe&=<3eEj-9=7W73M@GiRN>wh_
zLf5B;d0G{Xu}=*#?%ZiKD70G}c=%@`mAWAzfl}w)(VL>8BBay0n+xIP?)cWV!3;7k
z3&O<buD|3H`L$}j;>-ts^{2~{z?&5@F=5!LIib?m*M|j(48F(6VK@6i=ydjWPXeE4
znnV~Wr`czeASNY(2dQ0TVG@WX{~Jb8VK;Y+<cXHv?rcMXun$$VObdyw*UQdttnS+t
zD^k+Z`lq{d*kl}jaJ~LK?dNx}urw;29(BgBWxzVBKPD#UG^^hnFAfd~nP1CDUiK%^
zsc&j(I!+M2xqLQN<!ZXW*j{Qrn4!SP%Nu)kuyXJI{Z~8FHA>mTnZxU}o)A4$VG_6N
z;Z1%Tw_<quG_sY-^Ai@yliBt4kQRfvFrMij`b`0H@!U8*K0ff!GA=IMx6yGI;gVh#
zTV;qBh$nRV{vM<J?rswv2`1nBSMl7D-@kXjD-4!NH-NjhnyF2Mhh3PTZ=AHPdu3<G
zE*tZtWpp$q?wRTQ+FH9Wo>1GfUV)~dyL4ma_H>TRT`BOwkp5VWf9Vyzsph=-nfm1m
z(p}>ogF<+HFBblW#R<4`+`W4j1s&axBTkXf-*KfUGC5hn$cUae6jfJO_m3tkY`W@;
z?JC$>@LJ^M<&n%d$;o+au-^v;qVp$)H#RoX_VT%Gs!zW=(Wdg+ed%$u-g3IPP~ot2
zr#qep{f9&tIj7YaepEY(qES<xa+YG23Lzw#nVajFG|7nh<z)#uIc#@#ci!jE?~#+s
z<Z6~6mAmcHuJk6GHu_=4cAHdJSHqpbR^5knxs6XqC^l8;Y~AbHlr9^?>$*+ZD{z1f
zcN^OBQ769@B9@p~qWe1+@<vMi?yjAHpx`4WrU!z8iLm)^^+CRY)MRdM4p$re{U-a%
z<pHKX9Ww1E#OrRKDBd2t`ThA&HsWzTPY(}7*hxfCg^gQ5GD=m=CKL4FJelzkHf#+y
zhAr|~yONNpvQUi>Rb0u&y;e0>vrxYYjh2qi{rYSL6BqYyZ*Q*)@z(~u2Tz~gEzt~<
z&q(|6qyBiTxmoPcLl7O8a(1@C_pwG1im9oo$KmRSaVxYO8kYrkK>-28CVc%%uEB~s
zt>3KAjC1wKWAb2WWhE*h!Aq7R2$nwgZJEjS<r%CGBv8%VZ@S)Rs3n@(izVw${PiAE
zijY0Of4`&WiL0ijHt9tp<6B)U7Wa%4dvRsub)EM+@9Q&DT3T9Ox7`Pjs54cV|9&EC
zg8huNfyPBc^P$kNRXJCaE1X)mCjf^+{f}9sHJz@ydQiU^N)EEJqN1H_5aPyt8{R`M
z2}y+*?|^4<U(d?O)T!aOx3`BKO(w2{fs4!V>{%3_<MI!tJAruA%k7NG%@D8E)dHo~
zlYf^wV-TB`IePuu3Z;-2Zxq8T+6NDigoTBl=v23r+b@v94{fXW(DL%~68N1%AO@x#
zFg8*GsSF?ly1sLT78J1eBn$WUrHEQH-u<V?5oc9aqRAQ|u#W*n!FVWBDMzQ~dr#u?
zg|#(l9i1fWDGa3n`WIqx%2_IhJ2T0%vwDb3DVI@S|6<a-TF~S0MgC~9`5*(5W|>V`
zb~a0X_5PiM<?e;W_9(en_J-qERj@;#a3RkwFaP~aBWh=F|7gFu#By{wklK3|wsQ4u
zBbL+J04ii3Sk18_V=RcpGzb`2_l1p(PWYjEg~LNu%|Cij>xS~Q;~8FkhreMfupE5?
z7wz`ksR^vB6W-q5hFsO`B9;5@;<zh;FAA>Vus1JG&Sl{T_a*3gO#l3OJp5l@rL(1o
zsHnB`4D5{N%d-PSQmb;^VR-x)8y2$5<k@XB?1HUAPP4xHkterzH;a4c=jIwd+(uuW
zs^T&H`w97nX>XEWK)|=dWF-tFm9ojHI`kHg&-hE@@O*AiUe?A6(`sbzrsiaAd&j(O
z_xBUg6Qn1^$A4d3Tns;N)Ni8-!mle<e_Z-3&o9#`<mgw-e_%`wQkF{C>fYX-R=M3=
zX@>h98Zq<NSt^_sLl|&lo-xX_%x~W3XcPxP2_J-){r1T$fk@E6&%w2^Iv-)UYrW>6
zyT?y5#8iGGL6#Lb9y4^>80nr)$Dm<Ou^9gSCX*_IWlBRG&9`CtA^a>$*tKwBm09~g
zc!}))g`xIg%g*Dzl>bu0a#tkCLGspOKH+j>PsD8Jh2AK3+ouv&ae!eqQtvN`7+u{~
zU$36tO3=c^@@O0xeQ*SSbWa4MDhjEBqIbP-TbJFN$0-iGzphbjO7E&}Zr$5cjnNO^
zg$~&v|8To+AJff&Md0+2(ck)_xq<?=1T9olvg15O44RIwC~Y(8uKoddO_?~mW8}Np
zSi&`!2UQJhgAWA%P7}d$#bv4_KKjNKeaE7-==z^4j(omvn(XYZh?xM&gaA%0&2gSG
z28~SE=7+B+`Q&|be57UfsN7m7<T>hCtUSmF6eg?q0Xh;9Ja{O*bEliym$B@{O`P2o
z{o{4?w+t`kQ){yCi3AZmL{=t;J2mU0#@UYX`GAU7A55ed`60&-pU4+Q%3Rqpr2&bw
zs(a!|(HW8YQ-Uu&6IC2Eey6SUFFHPk@;LcmN!X$oQa#?psS^IsDB!8a063KaS^3*<
zp1{@-k&`UDU0HPw{Ji_cu5bPeZjE211*)dawNp2A<JG$3A9tZdVVAJ@**zv=enBW+
z=!$vL!$TUt>YudJblt1h-ee~@F#QITrdiN27HPcHnnE-H8&OWnji+H#xIwOUpATk)
zU6p?SHw#1b9nJ94?J!>V#`$YIJ*wES<2*SGn(Yq#TR)-vej<Ao5F7ip>dqNpCi8)x
znABdo$l0p7^ZQGkW!eEU>C9Ukp{;D#RN83Tzv_GN^oJJ)9){pAY3ZkU^~2?}AiJfc
zq?nb6d+I6?`j3sNS30eUz!uNauHptzS=h}J-~BTqBPufTrHhNp8P44jw25$EM(!8#
zcf6*sdaVAyA_NH+5-xvnwR<X!74Vr{y8`d|a}`~6lj5_|K_c-&RA0^9zXpY4%f==q
z9FEI!0GoJSHc27hKZ()|&DDHu8MK*zyU{9o_+<4w;7CP}#_!7LZ{wP?I_?vCL9Y9a
z7pcxia2b~+J7fns%1;t>?d|O<`8xLjc0g^)Qq3htAUTun@1?D$=jZ#^7dy8{M790>
z{av!IZ2PN3db99{?v_^HvX`8lbH~Tm-|6tn^x{~Ut-oPU{8#VKNjLg25Px6?(_%O`
zVtTq{Jg$kd5l{)qPVL1LM)`!<l2M(7!?i)L-3IhnIff+fbMuir?M{FM=A(J{=jP_h
zwEuQ{GT^4@C|J5O*}T!)^Tj`ks$n5Wb*OM5tzGM_P3GAZ4sbNyjrilaFDmj-07;SP
zU}%&>SeKiPEn;)i3P1_eds|yufN=KC&f@9Hk@4}efa<TWuU)nN{;P1r#KB2dNER|y
zcgJKW)4NdNz2tIAG}t;Tt1ex_P*O^mL7}+r9$55fF7-){#QXEBv{m@xA1f5JT-DCq
zxhC1nHiFkVB+Kl91zOtLa0yn9R~duj*E`G00rBx~HDcay|B>&zxQ&sCC;Y1LIDOc)
z@@PnxMyy?#koOSPU$~}oGiN)ts=B9n`l$bT*WULcy6oo2KW>x~d4=IE*ISBhA!U$a
zR387jHFiwS7e8a7d~A28vbKqZfXbCVyQP1~!ctJS)2JXo>hBv*HF~iYJ3#`~hzEP4
zwq&Mmck?fjmX37vSB+TQC5tR5BVDaYYbg`{<<NX_xxixd$3J<%M&#{I{Y|jBeHA;2
zy+&Gng{FF(5hcuu=H^QdmE_W0j3%}Bl3b@QL7$$|HqRJ8ZCZ0K$i(|~FI#<)KfDzu
zYIQ7&#r6bKH2d4VH`@}onj<St>-$<})keu_F&VeCBk{v&=!jK?k>hn|zt6n8>^w6x
z`Q@VLP$yA6p7DenGs2S=d44rPa4s_chXj^7ceBTICH{QRd_>bgAk7MS-JkQ2l&5%6
zX2(Y3tG}>qwGV6}h<PHBvM>?^^k%&<*6wr$h#ux2=Sc<auWd4@b7L63xF>Xc6J&bA
z-=AH})|qS}o<^>7oYyla^<IUL-_&i@2fv?0HQZ0OZmf!yrsFQE_r<MRPxACLCLz0|
zw)uteE;rtUvU8sL?sokr!7OS<xo_yn1rIM!IOPlWpOPN>;7?CNiqY3)xNYUrIuRv$
z6A>HlF3b`kwzt$~;*{sQlEl(H)lMZPo;D(qd?S)wSm5$FVO(z~#IJx_l-rAkigvW#
zuv6V+y~R{G+4P(a=b?bi`^^t3V$3*)v3ER3)P<3|{QKMrYO)DZt(NE;Dl>9@i^HzN
zS=+hXZJUsP9R|-f{2JcU*7kpZtgMu!(znF5F_*KGjBHq2!a_*H?0ZP6&eg#$aJafN
z9<$c@ljwrfRTm>|0)4!y@{Z`OpQ#&{9A-xa#<a0Gedmu5A+shE{E>*RewaieZ&s@o
zkL~ez=X;Z3_u9AGo@#0-ux@BuH#q!ir}iNw#pTC0r4Wx_!!0Et;o<jTp!G5iZ}uF>
zo(xo%;@Jw1DL%n}tWWSDT0EOEcdF4g3FGgra$DLK%3Cjj_8%?$KH6-W*qTk@BjvP_
z)AS}yl<5A{5Y}(Gl#ucT3PBtv#hGV(Tk!~Yq>XJ`y-KP<LF{bnXhRLN04K>|Uf1~?
zWvjuU#<kP)YAvJ7`T<yhaZW01%@yIE9*^1kik6$&(fBuA!9;dvB&0`ztJ)f_oZ;ei
zI&(fv@%N}!^>@~pA7g&*$n^H<^Tx4EF6GWSNkAi?nxbYDbN}iw&cEoG5tAYz8!nb{
z&=W9NJ(NZzNP#v1>-Gxwj;gA&XXED<JA?EiY}TlTX(G(Mlen;gBjFLMtVv-}F9~F}
zTz>i9fDfOl*oLNu-?+gdXqb_>9lo_hcZ}C5;?BdGlL~(lkd6Lh$LW)Wsm{$deXWhh
z^s1Wf_mhtIC{o5GL+$*PbLsaB=4SM|ieZ)i#s73HSa`R}LH3*5WMR*eGDI2z=GCrC
zw8TNt17zBqg&aIx+^2Po=cYf{M-g{4Du=d_lD*FQV~bbWG3;_@1xuj67I$v=XMc4L
zutVAF-E;j!cFw|9UN3P?gZEh-5}oxjg$7^d8n$1X$r1Z!@cWNd70CdfbeZDu+Rcef
zY^Ss|SgNU#DMAiyOc-7B)2+Hnv5dtcH(kzNXRM~ySGd~I4}twr3H!#r`L_J9_~^HD
zaX;LO($Jdo=XmVg)?2%4y-I$kRZj~luqEG)(7$NTwI~fpk)6FH9@`H&`Y1}%;oB%M
zcEiU>jw1AOv+8R{L`MFzLQ}!D-_5v2VWE@RQ^Wz@w@quyT<)f*Z$d_~yGFN6R<mw3
zWPj8(;gYm{blZ$Xxn(-~jSRtq`JRvb8f@GG6}MdE%Uroqi=GkHUsv=QF`c3m&oy)P
zFfQ-T-#$|86sLBfL8iUdY`fa+GcDNh*_|Xg&XZ6vS@C)Qho$eDOxYo<YwB*xD(*|a
z^QuW$lFpR#Mh@~YXrdAl6K&(vpu-ur`rFddf*bwD4u|UFCFj^abNT6uD1AsqD%>V?
zoAI~9&_5`!&+OEE?9XBUW%#*<C3}dH;RP<W5TJBN_Bc6g3hqyVR32^Iwyoj!AHOZd
z7FSvw&L$BJ#4RyQCsOI`{~Uuxp?k}eI&*Yznkcwqu`%MaH!9V2f>)Y<I=TUA|1WMK
zhn5O$zGj-o;!3B0*GZx6r1n0g2_<1;c=QMp2{<>v`G_Vnj?0pGV<eZ3pZ~nHv)|%A
z16yZ?Jex*f^OcQ&)>ZfPa8O61mf6%r&??^bT`T^(3P}yk^~7A8+8XIA_{d+NYRb#w
zLdTs7y(?e=0%~f!_cG~??d`#<t7h{H3w-YT^q82KrLDix)Ay!5cb=LL(5vQY1?#ZQ
zsE52t|A;>QaaX9o1pBANconS@nV7Zj&u4wt8h7M&w``vt+E7a2*z##v+v47RP?mOn
zm4gLc(MEn<Pk!BXM+gy1rbh8EYpVAd8H_*#ELW@~EdUm2HT#OSaDBIg_}mM{`$(_*
z_2*>$oe(i`ldiz2XFnQ{b|ds|(v4ETIJo6#!{XoLm3j~T*xdGZPb8gWP-rL}9o=ne
zYHDa8gHoA>ax|iVX?o_cDCOnlmH9d&e5(+8Wnc=GvQ=4c-MXbn*b<)q^P`<ZV4!fB
z&QQr7Uqzqd*Ce{o2hVV~eTa(1_2;Hp1QhF-oo_F8W$)M1kfW+F6ZHyR<mBFc*A@Ge
zDWq{F*{6pw*{gATs!F5S1cx}(-!#cR0Q!?D7e3-pEo?Odf$XNw{=40xyPfDGDfa$M
z(rZr(T#22MFs7Y(siJ!J#;}RQ&8>Gjns+E^Um(HFwR}XyvdY+F4rv6u3Vg}oYQHZu
zv1y|l{cA>MBOLd}FsQ!YwoLxWtlxL}6tX?@7j5bfrd(!cC`h3vl7oM~k+5!@j#aSR
z`-)OFR=x!6l7|dc1(+KoPI4R@8yf<uWBh)+%4)Q?3=Pp!h)qJMc3eMx)OTgwS3wZ{
zwL2-3P{I|X$klkB0hj1OYQp-itr<V|NSN>90sJ;^Wuw;EHMieM;O8euIlQE@thOZl
zqX%WS)~Ohn_<OxsLwy^!Gbivz!u=D(;QlDb`)wZxl2_KEF>{R;-%)oxBu{&emNMGL
zkS!2d7r;cuPqGoG!Az7xf#=T$Ta7pMlls_m3jP`lp_fQG0;Hi4@!{y}Y9*$|^W4zB
zx7+96@vRLQk}cR1C;KFG`L61hujCuED^8?6vNJDF_JG#O;f6u+3EoZ$$YMp8>y3Ra
zc6+9$Yp*_U*fiJh7dP-v=tZ35c>byXC>XQn(*HEane{mdqm3jmv(@CRdb^_|hU0ZT
zMXp?pcYlgEeGBe0sdmOde!%=F(782lZ^y>je)xG85qfk8(2)02v{P;#Z&X~i)|EhV
z5_m?YcOW}UVNqkY>zaJy5XtW2dd`je$IyO0Un(bbV9<$lk)df>t-fV+0y7QY=Xk?P
ziWk|iRI{ob8CNo+a7&I>S7@3bSF?8Io7#TN-Sq6Hx|A(9SRXaKu_|(<X{`y&I`2C*
z8NNP(NS{pnZ?3!|!}Z^okuV!3#+U*ndg$zVn_K`RTe8zXU>=XJeMD3(EKnqhrnJ3^
zjoy)$A<V}$J5=H?KieoQ<=Sh@lo`3qX-d^kf08!Xn0Z3U0Q!5ly3JZoE_$WKDU?Y*
z{h{GZKDt|E_+eg0>U~gAiu1=0x*y}IbEr`Hq~T?a4L0QGWOxXlJ}STL=bQh#u-9z7
zxF<Kt{+yU`I~o?~c~8>Fli9Gd5q*0s4nU6uf#r0&otzs<l3~Q9?PTr}%hwK(@XAuY
zMP%Q{B0Q0Sx4;@<!M;e*7W5%H;%)B$L3AJu7IzJX{#44*7;SOw*gMP;&6b6&jepmL
zssCVVV%%Ft^-1e?fZXY_zEou4FiV)V?b{3Kjb~F6m@7N-&Bg^rKG^w(If2PlQE&;L
zK$M^YycvAUpz@2iqV(HA(0-(EHOuhU{`2^dbwPcfj+nIj(zR2AKR`9bHElA2fw_Yb
z6JKB0=9=B<(U3Xb3~oT^Bs>ZeYx~Mv>*$XI+0v&%HmBQTTs`zgdlvb2e=PW{M<by+
zUpPQ+S$SD}w_qJ}gGWWFF!qk#pe+6v3V*Glk*b%3Q1Ufp%jtzk*4yDGdDM{kFL`o)
zSLRNAmfw%;Mt+gx;^bPlPF681DSMwG4LoUz68<<+(UvnlnU*>KYo$HB!G&4=@*ceP
z%hgj_UiYSvoA)&6+wpE?l0p01r)9g7UY0vuLgKx76BN;nzoPcM_S|H(4#}SrF<oQA
z-l+XUZY`Kovgf`r&ObGcRAQw?IS@TnaQ0@EAC>l;I4)YvY`%%QXQ0fUi+?oQzUMyT
zefp5&mEtlTPN4uqoxVX0o1T6XTt|2IvE+95ucn;)7T4{m@A~OYC}M?wQw>;cm;b!0
z{Z--a)oj0QF5Qr|8}27mh?;-d)tsj`lq%LpH9|}dHR$y%KPi`YYP)3dhSDm_LK8y>
zAHIY*&}}$r2pYC?XUw@FAzX$kW6xtk2js>b$GH@~(iC%B#rI^wiHz01(M2_f|0sl|
zHbaQ+=bEsP*4{F}Yn;f)3H&2=c68v0UxP(;c1Iw4hQrg+=+?!PU%Q$!=c}AWXZB9}
zB!o-4N6YLET=a<n!-hG5QetNZ8#l9kYV3McDz+}2n}-5)gFd{yX}OkzIp194Vi;wT
zP=f5R-%!*WTG9xjK_}#%g?Oe-NE}oy7u|ll(`7ZCE6>>sln9g72sHn;1P{RIj#4t3
zowSg^;kAAXMq91TP6$eUd-U!C;p!O`Cx!P#6@=+#8L7?|QFugrf|8mNgiovTLqdt1
z0`3v-19tkvLRazVNx-Lj;^jUZf5==Xt75x|ud9|Q6B{pAXT<nrbEE)^3=9m!3ccgE
ztvMowIocBrHa@<%2;uA4Q60~)WS?u`Wj9sQneI)O<XY1TIqS94**VwMu>4MVa*e&g
z_~ODxG=pTvt!y}Of!9+oRJw*vv7K~<D5MFc4Kdr$$<KgEja(vsbal0`{8|Da5SRkY
zH3dF1?V)_4UFkGwyy{e0#H{v}dz6{o=<d)K@(iy3Gm1@Zr6mVrk{ove%}*a~UT?dV
zWt+Qu6>Z&G*ukP1nyJCYA~bHGx%KnMkN26GOfWQP`SH>hDJVR=v>Wfv9a<KaP?*%b
zE;e}ujK1{&Of1JLomnV&Y!-q0HRBv$mGDn-EDUGuce#~AfsS5V04voLyi{O7&q-!t
zeNx_3vS<?8I24wFdF*KX$?2Oq4WF1A0}LL4`c6p5p^L64FIQCDZH;aR5`j{{4L_|v
zsNIN}fB@=wSC<@22-`Y4qqJ8;By#<2^55q3ovt}=C{|3(R4(1(R=P!qwRuf?+~ap~
zW4Lv>&nBeM=^Z*Z8_lU2ZW&Q5Ut9i03@{KC6%|}fSy|c6)s<0D@IJs4Lh+aB%7rG~
z<kQ|4)-Wb)1)7A<d7Ti}5xIV9DiJ6v<dFv)adIwR4^xN!kl)tf(r+rp_qj!NPI$RK
zQT02dDHcyyE@047zupumJl*Ll17+N2W#W4xWtE(n+h;K+MB$w@xccUt<Q^y6-9SIP
zoh`*6n6Hr3-3iP56s9|X{|t^d#{tdw!lVx7IW|?4Y5nSdqeD5~l%ao`zD17ncpWQa
zbh0WngoY7QRR~Q<Si=*wT-J@Wn|ON4EdXs;&@EL|aeRnrvytZb1rm%*Ucq>zy|*_5
zP%mki1eNFJQ-Dou)w{O>aqlDEB}7sJaPFQlESu>vq@UDLt4!4|dMjNBE9KD=vuju!
z<<wl7mSbLS+)5kq^*Pn1-tWFPqg{BW@ib46KUdUClBi{o9XVAw%WUvh*j6b_R+Q2t
zkTQoYCQ7X>^9RJ-(;N#+eTU32;~7N$y<9zcs$y-l(?J>dl=`^m<CSB|+zU57=JK<t
zBYo3*MX@!4A<{KXS!!>3Du1soZhXirr#)|oef7i9*Z*pER&69eZ*P=<ZZ00#&W+|U
z+1mP`JaCtZ{k!|d#o(Xc3!Me3C9-w$`mT_L4YCQ)=)Q@@+|;w}H0{rQ_PE&aV-_@|
zx;%dP*aKh7OK&rx&uCRpPzBlK(k&@=NNd`=0NI1j>CK}yT@4j4a=|~IXh~z(s2aP@
zY`mp}j~$dz^#rw?Ex5d<&|O{yUNCB)AME2)tcd6EQFiW&s-2nH-TQd?Po_{^DT!B(
zA_)IRU60OEbhnnTz?maK^LC<cDX!KVGL0@fo3s*&<009Rny`i&(<#dfity&22R(nE
zRUf_(tn@?ws$7^V0JFj9&=JgSPt^6}KRSeE1}G>~c!|12xH@mh7BiY!wxFT<>_JA3
z{r<y2!_E#W`U|LEY`FpwI}#rRhNbhk^Q{0<EUDyTVqc&c95*E?=&jKymXZANA-+Z7
zTlrC&FIWA+9%q6&@=BR{U;@q#7T!BF<hw1JMz#H2izSSE&W`iB*Em080ueKE6&vRX
z(L}D1B;%7Vzm}^M4_Pw-4BB(w?ED&*OVj<qxl2_`3dw&o8*^Eu_kLXamfd$a-ia+d
zqBPAnEw<3|tk(LOHqZfIgz>&$8-D=Rz^B-SxZIL|(`}|Q#hMp?-3ck$H^I?>TjWs3
z60c>2=j{gWmPJiSwW{TYH0n)Xhp(E3J^MHJh%)*yUGi=g7LZ`rC|c&cRLZz*#8}^Q
z-|q75M!;D23_1N=O)d%jx3VTK|BBCE(-Tu|YdblfRNiyD_0RDggxyA%3Nt3|MK=z0
zo5qf@BwTEq#f&-lF92?tcM<sg@*r{~`E;SqHg;5#BI#;f<3w)#2qj|uN0}(kqmUMI
zg1y$+O@?w8(xUlXXCAUUM;_4~(I+eUrUwHZ3)towp-=Do%6OeMB9tQbJj3xCKKu%P
z+32&IXuOv`1q*_}QgUg&%q<-eB<WQqEqYV<LWlgP*voxBsj!cSW|sL)f3+-QSi{g>
znz?Rdb)-9Gvhq#VJ9b9&|M^9de-!*ed-}LhEG`OeRws)Kf2jZb^=Hg1lH9PJe$QM_
zrhv!0zm4ygO;vHbjTWMXWK;8cew^+;Pjf`?-k}yWtUa<1u-Z^~$*{=qQi9Ux?@N;$
zuE!3qb0VCNi*z39JjAAQqgZ!BiTFKYuOy@Np}*@V7h|9K8!yiC;7>a=NU-D~t>c~N
z-E|VO8&LY}KYJnPB||qqqVrxMAKj70%<0E~`s6E{%aPTWN;zsxEG7l&IiBPp<L8sZ
zz5L5nKD<L));JG)U)bea;C$vVQX-t5@%mPbIsVxFH_n@Xi5-V$;%5($X~m>>f16)n
zZm2qW*l;Y;i+C^Zc#>(1BwxyHP=#gV@m{<S<>B>wYg_dObxT<I_3xiuGw<R`Bn{8%
zlyb8uX7zU*#Q31Xi<vtdlwI>?Qq$%?^PEJTrWlLG)py;AR8x6^>tIlYpPG|c`K|ry
zj{M%?I~sU*gVRev4u}x758a4EtuB}+y)T-*>*Y%JR_7l;4E)p}0pH-@Y=zlhqx;iM
zVSVOC&uDBECO|X9O^hmZI!8$o<)9ACnU_|lf?+L_4?-Bkjpb9*c)l<id*j9;@-u?P
zo}|c@*$+u?u-#2)!Xt_2g6oH<8IflU9p!tt!pAMsd+x?Z(n(vy9R(vLo7&-a_%J(D
zae!t45nux2PlHw}udkFa#g`E4^w5Q!KWs&f(om{XvwDI>OM-~d2ccey9%?VgW_Zje
zvJ>z@ZcR<MQS{Op@jW7tU`bE9_4$X74A(_*`hPzZXr5t2e32u+P-sRSpWMNOkets?
zO!)z>DkcVa;mnb3mT>|lr}(cdJoeiAewD926cF2!evhFfoe<PH+b$kxrqLQmCi;Eu
zcg#jmi|~Tt&JSO7yf2_zxH#VG-?&*dm;(}@{^7$E)ogo5$28;4Z{?K}`Uw^^9a(fw
z*W}ba{BTp=obE<46s^$6_yMp1vC!GksP5k3q0vfDV)fZ_eCF`T{mh>~>GO1IWE>ng
zX5O9Le_Lh~4V%C<Fm>>Bbb!tY6HjlV{x7fjQoSz|8I*UE;y4)6aJ)Nr`uOo9DYq5r
z)2B~CSaG|r%vKi(SQBZ``(R*Tly5mW@oUEoKNfShn{{Bm_l+vW)c3_S^%zFYVGfJT
zXVL2TqT*sDLR4YljUY22At4wikyJ#q8yO7#dWB8O{-f;>$~1gm`+Q62ILuu%b487Q
z;opqqVYPp{SjM=bP&54Q7s!N)Tgc_?;v%6$h>dm|#>p^%o^c3b=HMWRY8NXu=?;-o
z%N}-fcb@}&3QUwkQorgiiHx<DJ`1j{Ugs5kJ33zi(+i75emtdp0F@we>iqIhHWTJt
z<Y8zeBqX4x_*<ag1mc7o^}>(9evp`Jq1!n-r@QV<BY{lIkUM^BkHC-V*<1OCyB+KI
zKCR(*W!E%qX@_UlQ<@~Z%r^&<VxxuMfBqM=EMcLcSfEit0vQerv)4xRXLO%5nQ?v-
z&k`ByJ~te7Ghed*J4M1gM<bIcy7>V~B}>`=1~gFOzrrYZdX8^0lD#W)WC>EWG<-J6
zqec_wThQ0{^Pdc4=P|v*z4t&Q==r1zL*%0w_+Z;PuJ7Rx>^qr8eO!G^8XMc0bV401
z_fL>UA<*KW2sVDWU-iAGc!tVQG4#9KrbaK~w<TbgqC0cbg!-@w%8UjfO>fR1m3&&E
z>4=5s$7N#Ria~rZsHr1IF@}+-MbEVPV%k}RHx^(n1T*_^jdrY1=A$PZNr46TAj&D*
zqQ7b-X*ppCyD~0g+lH5AAMP-0O3E(~HNMo6>oB?eLZ@=#qQ4zP>{cP=hH>#l5dNk<
zeUUOr(<}D7P#ytQ>c7uOcT*Fd`v}JOW8o2OuKc~BvVJm0(}Bb1_fU+_h308IVtcv0
z!85_J3<xZSm=K2iz?6~qYiTdvHiBS%rzazVyR}<dC@<A3QdC{!t7xzWs+_t53fz+k
zt~IQ72hhMY9dSdHw{J8(P;w=@6C~s37Jk)T`1%#xMtl_6<we)aDlB1>5_c9+_Mm=n
zeEK1#y(H}XulL;jO&Id{<$*-k6`T%v!A9a4L*K0X<#x8n*W3OKScuo;<i#EO_s=gE
zA~nQk+U=kua4{)mlACh+UZd3RwWp5Ncw2B&`wRYw7%czu__INGZpsML={91{+$HNB
z2fuPt2Mk$^=ilotICOVHmwh6*GpwZ$$*p;Y*7JeLSbK<QVb?#G4<~4xsgl0GkDo{e
z)xAP|)tCeGhb#((gXHNsY#T%#{;%%xYvJvf*Jr&GQJ*#_8)L&85Px6*wJ#MrF|2of
zrZXJDhUmEc$gAnRJ5LQ2JoQua&qWUJwYw827rUxA0hw=@5sxB_C<bIwJ>0vSMQKC3
zE^}!OuV5mQ#z>i59qP|yT!D5kKS$`p_^o822};*B4K)=+*8NJGHr}`aQR?7;>nO?1
zSU=bf%c*3rnEe?)z1d31&9uqfMO#Y0KX&9l3^qND%vXrM^hM~=_pXB7h(B38)czWD
zr(OS6vlVx-VJohKqvKkC8ahyZ;>*d;UBiKCRfZ;t-(@pMBCYUsx=8(|HHo@6dq4%9
zW}kc|(_NykO@BThRE1V!1s_1SvbSBeV*t9EgX68qRv4`!w{>(hnsH9MkXe*8Rcp=F
zFK<;g$5gRBGn9G_IC89t5)=udR!XlJ?p3*L8GvXG(JO*-hv5Ew>-ajaGlzfw{so1F
zJp`eTsp$%BL>T-xH}^NlwJHwVHB926-ZhOm)+DF1krIdID&2?|^BZ_k2L}hF-TnQd
ztDcUIqHbZBh+C9=2NBVJza+Ii2xBIzL3~Nfs@V)I{1{|K{=_`1P?G4sApnwh@K)Sb
z8UnDbo?N{R4(X+tcmZQw7~fxBdi|T4YVYo59Y7-hJ>z@i+h#DM-WbhiWMBvYZ4r{!
zyLavf-8{$}`UQ=RB30FBN`<-nl$}Firf#_kD0CAerLKG;UoitQ@N0AwKu>oD+ObTf
zG&IP!pfEcC1tPnCBML$mr<$Yw(!zonw6CDGnRemhDT~KYQdYemzSO-#7dOQ6v44Os
zHB*do9-p|rldVqf#S5ghwKZ5fB3D;eAg@6J$C%{y(PeAGet%J#K_N*_QSoz4CsB0s
z1f6|#h0qr@55H%nQGS7$T9gKOD2V29Z=sbWj@xRC9SNGR2%+;#!487xv2#QF>O!F}
z!GcZh=e78-Q^evNV<EY8MD2uY8+N!xR`0&h{wwVkB-5dP4C#WccET}^OgL#}34Q(P
z-1THx8Kx2$0%}HlGQk2?q_wP0d#MIFjC7>(Ta7Gp87PiIQpc*WtCX{ciLU33Axmt(
zG^vdo9Z6h-j#$Vi2UUCN&cSn;SFaxO@Q{Q_yv`C8QzC>nvUhZpTR_4t3*SxhG^+y8
z5U&O*1F`UUK)`U#rnYj0F()}_84yBWvyy1f0|V}OP+r%!w4j2<%`@gTGxLeAdQjDv
z>!WDtYz4)pR{wOvN#&`@vu(WJ&2U@%iMTa7wmZt3^0}&@p>$}iJ!AeD^6tHSkU<Am
zETVHzU=<Qu@wmR6Ssfvu!hH-1D^OqQLm|_a@kgYS%l52tx#<afY3O^XaV;7}Uu92L
z(c4piL@W*9fnQ+Y%uy|Q*m4ps9So87{yFyUfwZd+WUb1tUz73JP(p~LS&rs;9Sy4`
z2)!c*t#4Z-o%!%@A_ajx6xgKD!GgB3%MbG&bGID<WQ5Beqte~(Yf=;Ix)*_c%w1&D
z3Ezc?4#rB-V0V#+60;frV1bbnHmEy5Z49F1#koN0Wsp}>K6m*9+P(cT!w4ls#d;Wn
zm{0t9r0a9V1smsJrT3Xx-vd}l<F|0xCm@iXuiB|i13l^G`9?m7@nS*mT;#YSkB|U@
z-dOt8tCA^USP+<fV1oeqz05r-7*?VnNg52Cowgn%NVm;LC(V7bpcX?FwExw0>2}Vq
zB+X|w-10V(SiGTkDmW$AMo^po!Grf`IOKgbp01THTOFz5p{H=W2(}7n7uXoe`8tBN
z=j$w&S6Aj^1vC`gRtOQ;W6eK7AXsy_xjK1f_!pa)O~;Dj7AQ9RQpK?z6JvqyW)9Sc
zfQ^uf$K~|&9zc*FbY&ps9?Vpt12iQqD{I)FDsEPy3G(8`cPCRP5hL+*kiDZPN<MiB
zP1;9>F)pm(uebi*+rq><EO6BzYD`(>^yW+c6jkc>Dp#V|pWWwV5-mtRqiJ$79%O^C
zBaz$|BPGqItNAJ_DmJHU;vvT&6ndbsLC`Cplm#$jF_hT=Sp@?VQ-7|}|L4!2ip9T5
zF5ZDwwW{j*)<lWn`Qe(MpWp1tiln!<(5vs9bD-g!1;y`N-OUx-eSRlBpk;XZ_<-HO
zff1#3i3%~W49XC$r^jX4UL`9W*HrRe=Vp$T;%z1OAFhRY<G7{iuCdO5fQY;W)I?k2
zh+6|?I0w*kp!pH^_O9I$Qgkv>j+SxmEMW%O&*A!TC@L25Zu30_yT!J<Nc#r|m{?e8
zFMZK~EegOBzWjWbP6CAVy`Tq_+428oh`59HE}o)2F@c%Q2nM>|Hopr|(*9E1CzTzJ
z|0+|>#)}?;O!6TM3s$1MX_Y#gfItFdPoc|QQG|dyLq74n@C1Vm?q}gLwSVdX5fONA
zORW%c{2A{HoEn{8kiVb7*mL*zco{_L*!TGf0d69s<q*~V8%HRps83D0Vr3l3i+{(e
zh43#fr6aD!$@x|2-T7D*q0(Q2!F_eWft`5Ts44<uKM}L0*WI%6-K^3;_<@f>sm5b1
zxx$1yJ~%x5@9;1Y#52f1t$$Pu4XJ>#KT|n9WD7foqU`nS*L(Z>BL9)nhLn|+u}XoO
zYJ7YgMx_%;jL!j6zG^+%7;W$A`8zlmWigy1O5FPQFN{^O-UFU3u^2A$SApiiqh^}>
z`#ON<K8q4;YN6_((cT%AoB-zW+jognq2)bmwi=-fU&weCT)xKjX^T*#j$s(%@awFd
z6Qq)G-_|z(cA4M&{>QPkJmYgMtUjq~gNs-1;zDFMsh9I6d;d6?0hyMWFf<iE>U-<A
z*O-@H+`$LI<3o6dddm5B=bU)~**dsyFqM>nobk5<SSYGCaJnW(3tKm9ZlUq&=+gaU
zE<jA>NGv<j(ZQd|oZvwtP5Lrvd2B)_%mG~zWa+tD--qhRZakn~j<Aph?Ts7%#-{c!
ze4_6#4mbgcL6-nse-<BQ`PaX5v782j2U6`%=-Yb3rg3hY-`kXJTGw&o?R={lR79K1
zt+c&J^u6M1LQ{p`<5U92^3PwJ#{J|R?sWORan5+jrQpGZ{Zm_t2w1I6?Yx4ht;sSF
zhZ_D?&Gk??5q4b0hV}nrJz3gzc>lQzJ`qvNSfOFRMF}OpQxK@>cMk^@UaPCOfwUV^
zAw504AB|{$-TKfYZkvB*&zDXjC8<(?Q+;@SPWb^wK=F%Kw*m6fn~(*`_7(^HJ+gYj
zxCBD)U6<kch2M+<eNp<kmILJhJfP_?#~^{wc?$VQv(!rZ=K5-_FXc9PHkM!(|HrKV
zBZ$p8o;@3{aAX7{2=0CU0)v)2U^Qr2?N3_+r7c1^k8oVLA1wEmSgIko-fH&)UWY}D
z2*G1guoYy$$XH;v{&w@<zaX%#uYGsX*41SMqu*G$UmSeK71hv~Jj|;%OQ|>eRd2>H
zW5$qC-(1q=W$^l-{JY7ObQg2IUXflY6KVIIJP$4!@9jUB$4l+uFNgM>>H0c@nQi2c
zz4=nhmDyf;lV^tTS@;<sC$gy298K0pq&19>ljW+0LL2zpb=wH~Ot86xfT>^~Axw5!
zQw9lsre+CqTwEOFqPfLIr%rKi&}1KOjw9aB!s4RQ#-NN6A?zKnexyQGDHVjm?yH*J
zV#JZ6oMrV=g-rUrib2S&J0V9WJ2z!aBTuMpE>)3=HE7^DM)hdR`oxkN<iZ=oOg=<@
zsutnc|75gC207-7HM_=+DSM<Ld*o|9+WfvxfyLj}r=O@CI50HNgvsl_$(&n|!J#Ix
zaI2bK(Qy~{BN_qutOM(+R>+B)BMl07fPIxs+|tsLxVkzi)bex~(Swlt19Z?(ErWuC
z=iv3gGessP$wOWW3JQ7vy6GAnpHBZ~c#ZJ>?7_tXt*!)63*%+B)C^I<qq;tc{#YdY
z$CI`TAUlAv8LFDSG?ew)&C~NXWO6lqNy=&p%+Gc3G7>7ky_BfM<PvAAQ!97>sJwU_
z)*hnll+;K~VQh0=XApmqVIK0)pMt^(wZ)*Ory=#<sr6Oec&c8hI`CLHJq)2sC*~nX
zrQ`b-wy0?>23OXldewWAMk#$cdYti3V;_^s5#M=5jQva4h9%^qe40eQQrhTm`+?u~
zBj$Gf@Q<Y({KueW*3v>qUEn?MZI4UAgxwQ6+WICk1yff3F!DrI?P^&a{{?qD?!&n2
z-ft<j2Mir%oD=7TFm(XEbr{Ta*u75eU;`p(g1W0?3P{teuvm?)t>~dG4xJ8!;tivt
zB*o)hK%;0>xv)YELHF?*6c*DS8>~o4?#pD}XG<g?+w?`lG1wd{yxA;ygGBJar$LY7
zna>H4ymse{4Aq&^8{*+GADcUP!>q2ShSy<}%#-ilh*LRNJ!!rXkUq*_RLshGcbfU;
zhiX<|=9|?aW=?V<{u1(B3F9d@%b``Do4Qb)K%KO|(ynxboU0`k2|jperitVviCm;G
z@zgpZpXB&c$95v|d9`F@O@wf65i`Bq5aQt#h+q8S(O_0#e73&?Rto@Y0GIGUzXKe^
zL?sWnuW&(TV+dX%#11VL)j;z_21}S>grEc<Cq!G~T+i!s?S&AeIy%y^vW9^vuE^)c
z+hu1uEGC8oq6T6}8DRkPOj(LWY14(5<95-2Mk@F1t;kE~100*mT;}~wV)ozhx90!S
z<xpHt&ZuUo_*|-fIG2#xl~jkp(a^f^H&!uY4)04h@AEZAL}xc%%=qEM2Qdi=Ww3Mr
z*7yKk6Aq3qRDvD|%<05`{~m+fgJ;i*rSPho0`XeAyK$hI0NV!`*{aXivk@!`5I2tV
z&6rTwz%lVqK!6f{<G-dEr1Es0nGP}j*MuTpV~aNB`tPb7_7E<W`~DxN;@8X$FR7JY
z=JQIsOzzJxC6i+qWYuP;z($)CK)6oUoJE3)(g)W-6lG*;nl|`myS5Fk1AddzCN-i+
z;N9TjXZxLO$g7iOY;=<0f4bvgoCwQK2No;vMX(z*V|YeC5=;NpZ?=1M1c0J*WLVNz
zHC-n9&byQpQ52B)JXSAw4-%Xt=+?jkCV8?wg^7)glH<t$E<io#=s2w>4#PXo4tAqO
zg1)TG%_$k-Gbq{fiBU*_^a*YSyh(lB17lFetq&k@KNS~0!y<k>TXQnqWW_zXOhioV
zaDHeGpboIYH~hOIW3R*~Ssa4|KT|AV&6I}>859y?euQK_04yI^i(r{o5OzP{{XT=S
z4j?Fi(K!J52;1D@;UO)JV=+am-Oaz#bk%H-<cL#cC}#nO21%Yaf`>e;tma*sN%8n6
z)ogL4ug^#JT+K@~yG#xQxhAGM-`!mD*PnZoXl}6%-gwOj=f}h+?<~7plt3Th*ts9F
zv3%sowtZMuJ14)f=Hhcg(OGp?a(O1`cyJlN0s}x5=3&-OmM4(tTMR0;-i($xR@EIP
zuf!~yiYO62CJxmWzUS&;JF^LYy*z3MGIWVt)yWZG5QKqu^6=0JFmK47MF%5vyF5HR
z)ipKa#in>*(latKnO|CJfp!xCd<zQ;5pe}W+{rg-JdhKg+0Tn1d7sXqgXygUI(D0j
zU(gyR02whmJ4=i69@-HoAV5+g08>o^78X`;a4-!-BB11Pt7Tw0#l*zatLT*o!Rn&`
zW=8U`0O+Ao`ppntBWQn`0DM8gpFTjpk_UMcIL9(zOS=kA|L!*fA0{n3JKS&CbMJD{
zP)WMExkcK_L2hSJ%fr8WS3g@LUD<y7A0Ny;ttcA7m;|!7EVX<pB(M_x052Pe%Mj{v
zFE0V`$<Z?~AUvty*u2@3SU@cI#^K)?etLS8+~bjm4-#o}aJOLfoZa5Gd8@!29Rh0g
zFHundr;CTJ3)R5~a|;5k4At1fgCFQls2Pw!4c;oi1RbC!>}YU}G4S(~hqj0)Dk=u&
z1j5W4TtlH587ta6Nj!CuHx`0Z8>WqV>~=@%ZzfF}jX0n_r9by(n4h2TTXRy$Rt+mQ
z?VaD+diCN(z3{ah^pl81gHB^73uoSCXG<7j{2%LvG&on`8?>~vb`K5^u3Y6TKpXU6
zll%M`A6hDCY2aB2NJ-nk&UxK$$)aX+sdTff5EuIClL|u8kMLG?<Y`xZjg2KW29hUa
z+)8!1D=tmoXy|pv>E1##;DVba2f*ojF$Z8U%7qLt!*K;|S_f>?P`S>wHpT0cx|?wu
zNEb^{3PKeN2Uh^zI;Lwpi6Hbrub_T*wvv2P^sUZt`(K5Orsjm;lxxFBR2eI)P2K4g
zmyJh^jFNHSfMOjKK3lxk0-4%|k)ECjjBx2&$IwD$0sLbMNo4{DQmH$X@7Ew;tnvJP
zreif(`U*m~S8&E>Bu7I=M`!x}iqF;2sEnLk7wceJT3R%ecj7pcLUqJ#l(+o>PbL#Y
zT*=h+Vw3`*Z^IAxT1sK>f84e$Aw-BKcDrJ`3Vc^t4$$Oz*51^3F#ug<6L$^F_M><n
z+b0~Xf`ZzBr^2^fUcUSQu8ywfLNoB)nTTc^Zm|OlxIZr;6P{?DOhLCTDXpMTmvJ$@
zR`ag-d2Yr|-Hi?(Xc!~SAt)8UJ<+zP0`Wk8SF-R;l0{*Zf$;~(rBkup%bt}pVL0Q7
zT6NyqPvhju<a*oNE0(&=IO4v7=Wx<Gv?Ucjx(e@&iHkGG%77dCX7(Q5Gx2!PO*HJm
z$$(9Ho3bQC8qZfy%Gv0GF4BgCm>9uYl}^IHWc2mtSP@KM2KZRFk}PUTc=BgF?O{LE
zdCmMY7X$}66Xd%x?dfh%XoUEkcW*r6;Th$&UFwMLfUU6u>jYn;1TE&y+4aeOT=|qo
zM|(TM9UdZg!I&J3SZ}f0^iQAel_oDuK8^1#d-tD&^hi+f>f8J7-d;73Pgs@RT*fg4
zhD4)A|7MBWr)Orq!^Xx2YR*mVkg1=)etiLxWNQ;+w78WOWHo;<D??XURSBC9rYWZv
z7n2}I&Q#C60D1+3<|o@@FxDGO-ug+|(c4=VCaI_q*KGo?kLmH_Kdz5l)RZ$y_?dkh
zUdYRb*L32S&EaS4sWJfWHwMxNTG74Jg^0;2$|`s8@Y=h$`~np5AD7zFST~qr_`EN9
z!OEOeBU=~`(wjISRF%M231-XJb5H&{8Wo`{-ysyI5fVy*PXL74#3v;rbOH4O%TWx|
zq=5=YQ(zApikM1NG#(Qh7#Q$EHjT9X3>HfSNdsK!b0Ka<#y~^`s&HIE{N6)GM#sxj
zdq`APgV2%iIjt&2QW0H021{~Vn-rw&`?YGL@xu4W$jD09;B7$bl-c%2cmqVrH6BOv
z0OD>=&p|e6g!hw>keIm0W{Mv%>j9vQV6b4Qg_w^aQaqFdlb!@=eSI3Bjl|RX^|<2^
zhy?aqf}rO&K#5_P_ZWDc6q2B2Ls*;<oD9O{ouNP-P9@j@+Ey5cy-*^AFKLIKf4tm2
zZR=w#kPV2587?Jn{q@xanXvbzHQTWx3cN^IkNs|*nZDY0o(QusLgobaWuUJhs1xCf
z?*fGj%2`E1RG4r9E`y~)0TeI`-Y%Fl0e!M{<q(ovQ1BUSV?|cu9Dl>_cfM8-6o&%>
zY-`ULyuH1RjEva9as}b3QEC;p_3SAafDsf9ls0RUb!k_zr9eY(=1x8!DImvzDX}yU
zZzNw=0YbrjEj`9<CDC=PK>s#CnZ><UYRPnEkk<ee3?~V27`5R8J1r(76C#Qp1d?Zj
z*A*Cqk@O}|LL&}T0Bk%DM=j`sIm>BtED{)>aokr>J_n0U6@ad914|VsYFx*9&cN^s
zY%UW7GaRr${T@8@P_vrgbBzK$WERyNJ%m$T=O>YP!|-r4be4{*eKb(|T47oT?OD+R
z%b6+ok8jRL>p)vv+S?3R)oiMg3oPOS2+bIXs*J#FE-rL0DJcmFk<hCTdY;45w3*i~
z1))eig1-WZ7j3IWzIGJ?{~~3AHw{QQ&>1U14X2IR8qKFhP%@7ncg^KvmG{^6bqDG~
z9n6B?Lwfo0CA7G34ujU>WJ!yGKJao{6%PIglkWDm&T>i2PvJTeC}$WrIH(VR)B#uu
zb6ndB1~ce|7Z(=`-d(r^;^Lhy)Iia)+MtEKgHGj!RN&2EU}$K2CI_q=RR0!8@a7{q
z<E)0rm=|A=9*OhVazX^OcXr;+S&*2?@=1bI2nL}8=g7*>0dV~`FqS{D$0b1H<;w<Z
zyiTZ_pkmVg_V8O}{OE_!(9ql~u$KQA-Pk$h1%ONiaHfc<Q0^b)10Ej+2p8Lnxbi}D
z#tuevF}8=00pSS=;N%1unL9vLEP>i-5gf$O`S?QaoUCeAHZ4?#Gec&9-V&3PjEaks
z1`nconazC!p8&rJ?RUqt@wZCYVX0|pKyp6-Pd2c3_Pet$<&y;3!Gx|5&uzV_Z2%e6
z)Bt*u1OfLi5C{Wsuz_a`(=`L;XJxQl!z5z`vXEAlOPKS<NPm^9Ex*%hJv_A=nDhSQ
z>+O8(49y9IKU!3KGbrd02U5U{=ogR?2xJX&)$f3;J!3T1T{g!WK<4-8@nfU*FAq6w
zrxO}{(Vy8)3m_=f`VY6av<<Re6Z*q-B0qnA0jU=;Ge@-C+Ep%H@jMaWi3kE(WUNYV
zk4IXK0g-xH?)^9Q{$I~IVFpLPgTj?`oZK&i6C^D~#aYF7TvOJGVAjA!6D3onrwjw~
z=^LaT`?Gyx&lr0Jp!oA^&zT_-D>mMN4`3Wz?1F61p&^3aq<%vCm57K);YPKFii%1w
z>qY=Y#_O2^yE&0>Y`Wg97p<+W2aw8(;e8S0imFgLJgj8;28EX@QNGo&R}fu2KQlZe
z1dx{^v|2E%K(_2d(7Dj${#MJEha(KYCjl-?#Q6+S@$sDyelX5>NKgN>%@M&kDk{EE
zRwkgKp((GdBp@RT1{o!si~@lNkq(^*V!dHt1Uo<;evQj=gkD%U8Qwm0-3a*syrPdE
zZvkTgjoVPZE;aC>1Ox<7p%5l;xZ-P|han32AmQoN81VD*l0w0Pd#bIi1yUUd$<F6H
z)uOOf5uWdpDOYOv;JtI~gr6TuV^b4woU}AF$dEMQ+DGfdRq3|*FdFMg5)9SMWq?);
z9tPoE2T{cyphYuID6n!+^PxP$uni52m;jz(J_X+f?HwGh0C;_NeqIE=4VnSOF)E9T
zkofKZj531MgkY8b{nG+k9#I{E_XjAMCiMz3$2`pPj3Jm|qo=7%OEwltzYj9#Uvq+E
zOD4Af3Lh*)2!Dx=Zh?dz2O|S;xtl;J{_ofTbOV3}xdEjkP%t7kjXMZ|bAV<-;A}Cf
zv^Ek61_xcR?boJjo+FqD$lB1le2IyXgt3hIWGO4+r~yR50EQ>s=jQrwV>AMGTC#X3
zu{}_TaQI3a3_yWwggF9&j#d!95`!}_fcIaXa-A{0Jlz9u;Rr(tL__^Vx3;UeSEv*6
z8bC2PD&i|Je{hyY1DslcINt(sMgg1xfH;`~)&-xOyaSQ}!qfAVJWMWGC=r^-NsqX9
zAl3v`O7qmz{j~lzsF+sTuHZSj*r|(1WrBctrdN+NFgPeJFTV`82Imj7j<Cf-b(!%x
zz=7jz0QGqsZz6;zq|aR*!_t5l6L1p3Py{0Z2B1d~XU61cJUc($l9iA^^E}y3*QxP<
z=i|1WmQz(F2AXFc7QHt8rUV)~XeHpX9Jqny{*;Xb4{4d0f?&NAA^4zLzko6dXadd&
z`QH<QVBrBi3;@|l<#TBdX)CxvPY(dMz|G}Bj%LYkDGF%senEz)KAy0ItPL(<SqK7P
zQw+e|564URK`#K6XQvN9(A$VQp2`c=>|ycBYwmai@Pem#8Ai@ER^^(X`?}M9;qOz6
zAtpG^2+BVkTml2qPjEB^q_Sl&6=gs*g{Z)ReFHb(3m?GIB&Lv$p=hEP93TAwu9oWE
zp^^|Z60qeu;LZ905rC-U(3;S|F+-=PZf~_=8$!E-Q1`$B!s#psHW8ajAXbvf9R*xo
zGj+o7Di<L-;Y9?#e~$!JR`%Ot#33G|8wl)#FmZt*+XNmDXo;_g*aC?28gc`m0W+A+
z<<}gOLymy+Q4lAUAlS3@(fnQ*eT7`U>V*mpd<lRK`1tFDrzGza5b#^^*e2A$z^v4E
zMqWw^10=?vdwS}&s|V-&v_oN!g>!@;0WSfvn1eE8H1Jaz06*Y`+wb%9*+2pym6oTP
zoepE=_a(xQptQj@0w{%pFmuBZPb1H>>+0%q(`qR?+2gw*dM0^H85tRCPv@~fGAs&L
z1o#C4B}fmRlR7%1`MSN(RUj^^p`jrc#{ox@ypoa%%Dwj@s$GTt5IRc)`$J{??%lip
zDdAj0nvA0`K7;5>S~!zspn`>Y!K5Tb2r6}@rldiM*4lEbEEjHMmV#6UC0;6wE|`*v
z+Q{;f%$0p8#V%eLODRoKaY}}aB3Woq^gH@(>;muJ|9hVEoadbLNY_OCbuw#eYRaNU
z=Fi8%14~nzovik2$E`-Naoo&j#_RXc;{cu@$UyC`qV0q=NIx+#K`9Rf5C<Z?9^G|~
zc#$xhRQ2K7SgzIc_LCKXK*1Fz_zg0TK-s11zY`)s3d$7~AHUHbw<_MJ%yD{#CW4e-
zBGA(sbvm6T^;^lBo-C}P1h!{{cb@%RX|u)94NYay97m*%{2YI!vX`J0*2N5UsCSaO
z!C|edKtra>Nj)!oqrbl%X5W(4RY_uLHw^e9mfGz%nN=Xd20IJ`?}vt1VAn;e+A_ka
zT`1kLo3Xl9!&t}FZgrp-W66VZ?(0hvtlnFxgu?_yM&9CTvMw{f)qcwE5Az`}$2X=s
z2o=Gprj+VMSJ&j+h((FcVEtxG8}Ov}B_#t9yI!TDJF*tSD{rvshi+*sIuiXUjV7dD
z#oL`9vw5-XsU!Yd(x*g}qihxyp=-~`Gn;!>#!CS4vWJHJg;?MpU_&t2&o@aZ(?nL|
zZ;f|l@}C~>J~O;U6rl0ZU?$r|4!c-bC|LpZhQE_dGEqqT79$*q&o|mfZ0}2m08bqz
zr#h>|HFT^xASO>yle!NT)?+@#FWP^SKP!9&LE%NAI}~sgRss^3re1rXyrxjjV9-*B
z{2;?Dm>8{t<j{wcvZHy}VL22`9g-v$>~_ymcJm;Mxfru`Ih*V=B%@i5S7Mplc_PYh
zW29C=-6hDxO&mQVhc4hx14uUXkhbJnzjWrr5#UR}&i4^Zm9h05x9ZZuNdPdBU&vq}
zClRiBb)TM|oCFV3uJf^+GoaLvZkeNo0@$Vj`DOqalov-D8pXYUMQM5Y=7<O{a&FUs
z2R$IeB928WD!8Y~OH^qc1wS;Xx1z%_jtG6JMZTXRck@$|*^E@-YgqJB8d@?ukJ9DS
z8T9(L3$9F*r?KvWZ8UzK8}(t`B>%&$niK?58Q}pktbxAEi+8h(T;TPjBFCef{=7SU
zFl-0gsOaOi;=i#36~gW4#Fo-uGGc{yoln^Ae*wuGLi%E^1nt3ZPm!3A8vkss_UeC-
C=PL~W

literal 20286
zcmb?@XE>Mr8#hu$Wf$2hduEf6Y$CFfmA&`gLPR#n4hh*IGkau{m29&2-aO}Z|KB_>
zo|jL@;XaO=>-zn!>-#<D=lp!W?_VfL-Nq)xMnOTjEh8<ijDmt12|py5H{dt+TA$re
zP!zArh(A?vP2QMv@q9Ua)xM){Jg8mGa2?y~o{<eY>eG)Z4UKrVC?Ru{BJ>&FcC+G7
z?Wo!CUfj!IAr5A#r}l^yOD8gwJJK1esu#I`qI^$3{fn7MU%NjBH(%AsL++y#hrX(%
z93qKy>~H(jGy!-#(chSGm323n0`SC!ZWB<8VbXjPM-#`S5s#Dh#iR+yA;7|e%T({9
z-VeabkbNi$*QvZm55^M<W{N_=7sFJd{r_=OFmqvTZSwc;-(@RQRaBClURxP+CLC`!
z<977)IJq#2sU)BWS7s1N%=%$aes<ZKA|fI}(JRrY+^k+B;V~z|32N@SqjL`@$bX?L
z{-uV-@}_b?4rf8lO+2U242D0JydNYVysoe|TJBF@pZb}Uk#XPC%gc_RJZIQ^cX#(s
zg$-3*U7h3B6h*maScl>LM~|XvYlRZI&BVGsvZ6}M$UMy(j*gF)l90GQS>qyAG@5oW
zfrW)-XlRH>Kp=K@xG}Ke_-|pM#AQndZhD)+z-`)(BB3O$sF>}vHiAMfU?1Rqu!?{8
zZrgi8dVc41k{kYhY}(a&93(k8IWxyQ^R6BqF^P$?MLIt@66Dt>stAATC{|ccY8}8=
zM1J}tEiR7k?d_d2Y~mBdmDIETr{r4Ru<U4|_Ri5!c1}+4`nn}I>FsQVlnl4M#n||G
zgK|wqM#glh2ugAh@8nFW2sL+`aLOkkje68PJaOCGwuyW;I@(v|8931>`0B}Od9=d9
zsqmza9G72spDYmHxJ@c<Z_h~@_U_+5!&L86Gchr-U!9#92JPXs`@iMj0b|(f&k5im
z_4V)f3GLk))%AJeu-H9YG3}kHm?peFU7vP&ake+%(3h^9tJPO!ImRh`@{e$CZjOPO
zd45=3S0a1RZrrK@pOh3HN)J|<MY}pSBm{fPeT~)ocoqXi$76$MZhrnAm+`HwpGTC&
zr3w4{j<C2*`L9a(omPpS1>t0w^rdo}{SJWl`Kp)}u2Yo#<x4<L4uf&`r^cQ6_Nu+!
z$7X-a7{MsuxzmT$E-^pG#qA8eR8&${Mr&9~^XYA;^675$$G9irE#!5&pDq(a_vOnM
z4zE+k(}OiB*j7e1w#7Jgo%C0w#t-S~_gV-QZTEWk_WoKFMZu<w|FF6zVE+XSGykDz
z%N>()&6R;nj6Ai1xnEHakUdw*l#Ggv4Uo;!?@bnPT=^XsMCJVn467}StbKVPb7iuI
z?}__<gMLY|?W|bc$s$={{iQoO4o1Q23geOdSL9FJV?8#jf~2Uth<to}Mjb8>$8Bzr
zaQHI5{90-40Z;NiJluG5qAHo+?k&x;cc0brl=6n-G&$UVQie%hYxr=Ft9X3F(8R>y
zY@@>A-!FVmPfuHZa-pm9y)v`GcPoS0gkX_!>_5fL%ozEmJ-LE#DD($1B=%PZgGO~c
z!o1E7?FOIosbUNG4(F-Iq^0%Ir+L1Gx0$H4%YR)F4ohA8^JfCD73sjh0E<>-<jjoT
zl-H3q8X6jjuqQvR&^AiNlshvli2mXF*ewb{B4052?Ext&)g_<YoJ?fVmTdVkadES?
zYn<g1I0Gvy`7<&zOU5i2LYiiqgC3EFJ$(3Z1}?$DkpvHnjBMy)um6z2KtxPDzp!xK
zS28ml?nF60JIfqgDb`?#OLh4DMW*(Ax9b*_aPnGVO}NL_&-nTFh{4>K^eCjfmeNVQ
zR>fnMC~-LJOmWz8vVk3ja|ffkCP$l-$ET<EyZ>aEn3y<7Znd<v{jRh#K#7itk&%_1
zt2^6lwyq!~Aqmc4`119uJ**09f0@}!aHCE4y1{UY$8GSqZZm!uu)h4{IQ#BTz+jDg
z5(jdW9yGWy6&rPZL`O&0<4j;+Vfi8#&xV4;nxr1dr%zEK6clRL==5F*(`Z3-eGRu;
zSI*v;^4NMp>9JlSi?QzOi(*}OVoaxy96V_)l{cKBmY2Y9cOSOL2;Medt13G1CNWwZ
z?qNvAXNJyKeWT?T>S~2rGGbzA5kk9|usH$m-c?IJP3STHURnw!vg{}*PZzURGB!50
zeBl1WYP{`mWBgUI9`Y<t{cpUf7)Lb;PVv?OqXQRA5OiX4uAvB%^q0(hS~Oad+Golg
zos{IEMipFNFEUx}WB?}kXTy2OoJX_Z<FwBe1X1mk@>NH{1~=Ikux)8+xHIlJR8&->
z&NScr&yC+R_geBAaFd!1f4ftEaY$xoXJ<KGN0B)YESojg`*_RnbbrNn&o?n4q2r58
zjA{R8RF9)gqlM16?j*iGr^pHo7L5Y6hI`a>F)R6a$YaA=_xHabl4xvi4+5XFJDPMM
zymwE=-CYo(AVY|lqa$}~YioDv6EX><KUEI&E|X5gjg5^eDk{a!8=A_>%KY~KFqz_3
z^Owl^tjQHcF1TlBXJO5vlassub;R6*U-pSy2?+}elk?m94h;>JD!iGrZNAN-TbG=`
zW%4`!6<2h0bmhXw*INFV3Z-L~g=yuQtPdnV%w)yu=${|&Ai+)8>l6<UPgFu8aP24l
zoP7)g-dOOZKgDmvrKB<;N<K)rfB(K#vn$4J(g2+zlY`Y^6vwrZ+pun_p1U%_m&XQ&
zTho1BN0UpZUnjE_sUU-)ZXX;N?Jo~N`cQYD0)O;AU7^RJ5Nyee5SFd?zIc(VBA+Il
zLgjNF1vdISPn{j|AiO#9%hpih!72w69rqR5jEoG^>AEKo!lw*meAZteQP%EuG6w_(
z&Mqw>fl51j$o+7=qh_nF%VpY23NnM$_>WdtY>1sNH8ndIdy-a03dl-IN(!qLEg<AN
zEcF`AHU(P#Eqf16fCTwT>!mg_8^>}D%?j&{)p%6b?M4hpUM3JjZ{6bzc)V3h=nJ1l
zNJi#Sbww74BI;DQcqx(d+dm+uLo$N0#~|W~BxJ`4?GQZm#MQJdd)}(Z7nsS}0Zq-#
zk4SE9vPob@(6HjsT+Z3@Ydt6o$T9e~FUS7BngvxkTX2AjlM{CNmq3skHjTK`%UqQ*
ztMQLuH7@b?uZD6_Uj4AV2UeB^;lxC{fi?hdSSEY0#A!{<x^_p>`saa~OS}Ve9eB@#
z_u0n&*7Wpb7ADQ=XGoPp-|l6vO)#*Oj;(JJ1mHOhO=thvadi52{eD1>H+W9znAXx6
zsVdVQ9UYxGw&Q#&Av#eUN}*q?!+B-{8P_55I%HWrc3#(;Zwmw0l>9(4n>eI^X=fxr
zj>E~trIf2ue%j#nXx;g-b4?0-U*#fe9`YHoa&nK7dcd~!^HF5v<T6$B)SxWdZo0|7
zJsU)o$|C^<+i8lh1o#5u=pf3gSFgb0@kmIZIQkaYzV8Qy<8L?LR<%ebI2erofJD%d
zfp5~0aA9GAB&^M0^O-U2os829uaTL=21CxM<YdoQ#k}D|wNFwMO6)_ZM&+7r3vucs
zoQBd_k4COvzwUWF8x;HT<9CC0oK&C76K17vco~UaR~JWM#XaScoq?ZvjKSA}!^7)l
zo<5<c4}#oo3pKz%`ZEl86}O9n(R+|KAS3FvhTK6S!cU#P9HmTC2ovnJ+h}Tecj9Cv
zAe)egczYJ$h7SrciS}yU+mu(!i-SB9O7}dG!|eAL*>T7}pCCTYXi@I#KDR3PGdpCx
zjq?sKVL&A~@jf44{L!>e8U#IiNT7uH_)irSaB_2V_czAlV`GU%mFpm`N@NbqY;1&S
z7A(TG?X$L#$;tAbo<d4WN`)79@7~>=beT?9%UkJBzlLlZgugP2QJk8|UbwoCt54{e
zx3_Re#tR*t-rkhQ2|N~ep>}h1b4zhZIs4mbfc9^DwwXn<Ecl7XVX^CuzM#`;ONK<~
zC&;5kwY%7BqeV-wea6GzR3X=Y{;e;V{dl;!x%o<)aK!wr$dvz30p1@H{j+C-Q+R*T
z(5_BiN8mZVYRc9BE+?Netok@5ivinPN3&WgD7m}`d=zXPDP&bti1O5n7SFcompbUv
z91AG#5fV1Py?(Rbyqw#7D0r;YWU$7?=Iiq$Nr)Rr%pLst90}}&3f4i*Pa`OwFtD-F
z!KU>j@j1+g@uBr9souGB2kfoC+Q|aSLk331r%FT+W!}Zd6T?Cc73tQm+9U?;d_sJt
z%zQY+dE;;T=o;3oTmP!p3a{xk_@aUf-pw0c)x3M4b$M`INRX=1`^LYAVt><Y*IoXc
zTqk);&3{AlmY?x8njC!7{j18yG^^3<R~~P9&bS^ee>)7DFn7Tt#&2*r()AkQ)z#G{
z<+GNDAYnR^e=j71L%#*{9WGTkM4kwdOKKV#zp^rJB-i~ZF@T5xX%@n5<;CGR;)Ag0
z^k5mkeo3Q@5_CdzTWH1=5zihR`1+g-Lgr8RgAAxG!NGPe&$b-8SSvW7Qif82-}4x$
zwo(+Nw6rE@f7+zB7@&QT2)$coI*{?(Dm$jmX1Z>8?+eqNV=4Ztmct<LJ~0`(<nwFc
zHg82l{HAr@{_4f^VC5!5zx`KLkyXs~)2LLL!Rd1=b~7qs^V&ZfhqAhN6(bz~7QeY>
zF_QlV5+m3Vtgh0EonT7gZ~KPqcv|5e{SsSn9xls298=zBR!E_r!2s_Ex%+)Y1QKFv
z4@O=ig(GB$RQJ_fIro%Qu){BHn5HNjk^)zsLZ~0(6R0;dE^JSa;r%5S{NQjy;V0c@
zp&DkzfW^>@Ft3hBj~+qx#fA2$dDgbah8RM6hoL^$sO4mJoX$`8bcPTMyzo|iyOc1b
z4ioz|Ypcgi`u@WQL&)3QmVY9_t7bPgo<mI%dY3f=Ds8Cj`YUX7Yinzz6%>pYyAvRZ
zfjc(#r$4jUaH8^2&;IIgHRF#P92N%M$>r|Wgfi+FRxh<cS=$)ubPls)^|g#bv?AZG
zt7Neu4!W54Z{MPgGN!MpC@D3L6uic|dGi4`H&MCfbLgE=-oJkjb?_imW9mK~&>NVI
zeWwk=rJ@rQB=v7>NPiX-{QkY2HYK>M0#psGnq{`PB@`4DrLqU{iHKUDmFei~qvz(1
zh58j`8@h%Unwqf2=olD-1+V!=_0lFLv_=_S8iwh)tS2hXeS9uyU*Y~sQ?c+(?$AP$
z8htu<WNUd1UE*o`>+}oO88Wq>eXIdw+8ro6=;cpD{}HJ<ZDafXc#!Oat=O<xlzc(!
zQ;SQ{F3|k2tW{)F5|hUKdo`EdnD#19AnKK$hqu3XpXjdOu|B4!50xR6ql*)C;YI}1
z>LE5Hjz4d1>R|A7zsrc>uwlseCdXEu&`!Y<JGX$pCyt=lJ$o}pN^(e<{P%jA&dvI|
z8S?uq0eF;Rr=k-QS?_rH#6lY*w<#=BZlA@ArnRT<x%Ru&)QHiTmjwi%gL`=na80HP
zI$>uHNS7M-Fim=I&pfraw_hGDqGE~*A*5G`jExn)y1am#WjRq91%2sgV5pgy8Cc0M
zk7Xq^?%Y;?_wDP2BB)vMTn62qWslDm>dvkP=?#u^`0ytmH&4c4DH+7Y$M4OBvVYfW
zM1u+()kGAsK?*cZ4Q^rW5TE~)TM$D*=LrG4cBhSOeXO+IX&XbM+L00IH{k0~5Q7&w
z{9_Os7x&0vQGTRQ8(@I-@rsD=-#NX{cchRX=!W{&4xG$4^Xg;nN<V*t8IVWyu5#_o
zg~j%P>5^%IMJif-H5arw65X_VH7&GXhBdbfT}FDcm%jW3z<~Ajk1rV+Z=o<uhOkk0
zy2|`jKA{O3U^z9lyO)=jNC$F$?y3Cp<yeUU23+mPx;3-4rP!DH<mBRL8a%Vzl>2vy
zffO`+XGdEmkd0f~+tsW2iy{0J-VI4ho6h<viA#s7Jc!=jl+!o6O{#H3@bCijny#Po
zgGucKysO@C?aA3DdG3NE)_%dx9B%WklA4;DmWOf)q2>Py-M`S~@w^7hUHIjnGBZM?
z)?QgzLBc)$14%rH|4|R5ptUmI{nrt9XjG!HJY6pWJz$f5Nk8;UIOP1k&9gICms?k#
zxJ+=MN6&;O6Waat0E!f~0!;}?$s6Y7iAYboyuADko|PU|`DN>Er{43k_Os$ueEvH8
z8+Dk%jhlt?!J!Tt4*m1qCO*2V3^+>xUzgrMRrE3%$sdK<kCDE9K0-tRLRHnu*CZ_m
z{oD6A(c@&Dda+dRNYLEBkCr#wX_DsE))aVi?577GZ&Da!RxX!KEigXU)0OO#jf!c=
zWE0NgRufR7s(NUb2ct_rH^vM=Wfzs#)!5_rbbE3uW%kxx8uRK%jnAhe1CWRe_ZEA=
z9%G^KEVUZv1&Bt%*qBZ-GYX77k=IHV`UYqqpnM5S^SMm8hJn-nO_h0+5emHz!{$cN
z1tT>iQjpQ<Y1O(u0tg9d@_3VS)r%+rjG2Qxzy7!UX3eBR4Bhk9A)-x_Xz=)f(S{bP
zHyeA!rrcV45BJ^Chnx(WcxieUQav``K-CQBj{j?$pK%Qhq7cWVTwVE5QG2VFyss`?
zR!58aK0PwnK0I7(c~5Bhr`Ye$9}RWiTjczapFdMWJ_`;GhFa~B^#niKwQEoeVZHug
z*=Ss9*Z1Ve_F!1ufP(~JoH+<!=+wf(b-g!&y8{<wh%naA+T41io7sAA5?nl((2m1n
z<0MuyD)(Z`uYP>lO_)2>zr(vE5xn5WVZ(2@>Exm+7Az=>)<PIq@pio4fBR#x9y<9+
z>*O6QnpJ8nJjDgVTbXqMI2BF5-xpuJbWwlU;wB|Htm<4dSzah@^)O{QRX)>yO($^X
zYw@AT3s>?R^4ALDYnTb0VimHVDG}K_IIMMl^IQTHt3-n(ARwT!Cf57`sxp~l&3KjR
z@NS15)4SPH!+?*rHzB-VoE|W;vOcjW*r^7L;P>aJ{+bkD2&l^?_8M*12#(!9a0$Ju
zyS8es6{#7#2M%9;`nuwnDq`iDjH2QaltU)%;S?rTU0q!UNH&|OEU`}Wv6`T0V+&}Z
znpt9;$r+zU3B()jGSH-!k&yvtfju@hb{jwg^}inz6FZCQFJ)zAv7v8h9UUEgt~!G9
z+`Vzxt=Lc~hue12(d<3ose$cfqBm}lG=PbR$B&Ltw4^+CF<u?Yg%(rCma6<j%aHcb
zh2Fk;u?tOp(&DCh0A6`4+|nTMZB}f&*Gbh&Mh0uz`-~L|E@aPxDBW`e<?;pe&jf-_
z*itavb(fFwYnTbdhBztjPz$DP)`nIjPaPhqy#q)L9t?VOsJmB3enl?&&wP!TJ?ogE
z!qTRzthrAk&N}xfyO@)c^B6*sjDo^q{nZ6K1W?cY0V%jGBqT(m!kPkBXdV#0to;1$
z1AhNTQ3zoOb_Fn1p-0MkfsHRJdlUQ1nPp8E^#=ht5>S@|kc$M9$1a;K#-;cK1aASU
z`TW@z3hWy=<iAp%crdWDi!2HEYPEe=Sh1sWOvRj@#vOaVStyCAB<I@Am=BQ1?{bS#
z*@_gM>nwD1@5~0jlpepm@e^Glx7ZO)UW2yu7@LMI0amGn_V@B^8*W1v)yL_1ts+fK
z8a6+4Y3(^}TuE=7>+Q4jRsAn8?aUsEt|rcylAPRoSm;Z^>0gAGu?oM+$f4voR@Jy-
z`f@osIgCjo6brWw8{IFf!%`DM&w7PC(5xR^2x!rbdNfU&;~kme*Av=NcumdHzC%V+
z2e6<s$8>)VOKS4)#av0SBC89ZVvy|Ku=FHU=B=(Fr3t_Qn7n^S%~a&Bn(!^!><R5K
zyryS>t!JZN*`@wIbz@?1vaI9yAKp&CMZLsX*|DbAcOO~HZ4?6PIUiD|Kw-ru8+I}8
z@5q-7*%DA^d{4y{HJ#A4_%=E)78;00EP*Dj^7Y{D+VtD~l;XXc|7}Dg0BmEc6@K0?
zQ(mSR(pSR6h~K6~z294~CT{9@C*O^LWl86=9Hvr*jBl{R;4;mn(=WNVB1`|f@F85t
z{^j1SRGcIeu~pvxKJg8FVxSWVhTUj$Vtb>;Kg3jdf&l>Ki0iC!4$eaJ_Gv|y(Hsv8
zZ>R6P>CN)3r8}!ZYREc>-=2Mhs;s5e4GCOYPVVv2A>eYQ7NczPbc1&a^TSPf@d~SP
zy*+PWEB@~RK0CRf5?esp@!V-8K0ZD!)@yt#JI-tMw<TG?0TT-gL3`qLz3;KJv-7EH
zh8}p`P_Xn%OACB-^J#&di-Z^^-|nZoznn(3+X3PXQ1ftJ^n!w60uVt!f|5RbxCKCo
zW3Sh;WA!Rqsq>&)b8&+0F7BV5Dp=$ut94POPD;y<)n0AQ25904=$58k*gX3=KmUVd
z7%A$DdT%etJzC?;OiafoOKGpl&F?@jD4H2%(-Z$NgJS;*H$Umd8F3A~c_F{tD^!Ut
z%61_ibP4*Ssv7c&-A^wyP|V9o<78D=cmamgENC-!P&hX!(O^ocv*#xVfYX380fqzo
zYdaa()l*ns?`xbe7}APfEu}b^Vkw0(inh)k@O!)8kaENN_GSx@XmV`Qs4A`<T^g|5
zj*gCh|Nhwmnt6VHehg8ka`qjW>EPGfUGeND(09R5p}*G6VRRn;OacHq3SCUc9qu`T
z93I=Am|lTWHDUCz_qNR*cz=E&<rwSwqJ2J@vEH~)(mCA@dbmeiWH8+(Ti1rrBGCx=
zvPOyi^`}ptdLE9MAYZ$`tYpbc2CcA2NX9lm{X4%NMBv@K=ho|`Q82=+<-|J5NlBOQ
z{6li_wv%)5O{K0sT)28y`V1#N-49BqAB)Q;)6eb@N0|9(IrQ)#KXKL;00E#tO(mZq
zI80&mep}79-Trt%;Mko->xjuQVR`dgKqbq%Oue4AGkg0j$OIQsi4~@S`x%!J?dZWd
zjko-4=P~f6+N$tW-!Sd_D7-ttj>tj@*WP<n;2H8DObY+e>T^=`lbV`*QYkcXE#W6y
zL7ydy=V%xdoN!JKqH$BUSwiL-N-l59j@Ca@PHH#x9xO#a`<1?{CG-_n+%;IOdG<$E
z<d180_iobO+bon04mDwHFPVF@JJRz{^oY;Yqic_Vl9(FxK3jd;<eNG3;)K~eGHQn+
z2-<UL30<}<N#dvBs^k>=*C*di`xmLbyU+!%%AtH3mcklUtX8b|kbL^zW=H@(d#kzF
zs~zgx_dd~X^ZM@lRN+s<!>^>L?kjoKt~LHrOh352y<uhRT8@0Or8Dz49>vIN7mDn{
zvt;W@580({2^jbc%USrYnKY;#HSKm3D}EKh=Lv4csh6j{=rek#y@#*uaTNci`syuu
zL5{b1m_s!oyN|{}@m^DN-hY|D;oXX%bz*yO-jvDqXR@7Wwg?{xK3W1lDV^T^QheE<
zg65TG?;?@GcDCj5ufQXM)k@c3yWj3udT?GB0X@}5Ym$j63q4z^v^IGZt84Bn5mQ`7
z{Z7+asQ6^E%%{)k=k7xMF99M=+^F-Hy{$XAZ)q^Ar#D9I#VKAU2JY^aT4Iq)RzyW8
zO-uE@R*gdswvG7lpiu+8ULJLe+LW(z^3MBe{XH`BUz{0adn}I*zO2|W0<eDOn1tSj
zrg0ElO8eXMGptYfWJx)Z%|&8LfsTp%+1<^;w)qYYomm~PR7bWK2NzEwt0yOiga77d
zpm=+KOyeH?Z8xLgnKgJGE#r%Mq@1faBiFBw$&!RpB?CJ?q?R<Tx78sQ;vpol^sa8R
z@T*!R3jSvHNYjDMGY#@{TM>=UN$sc{D`oHOxq992-!J`dZB2Ge4iAw+G&k-)YrI{W
z&P_H_zq)BIZl%u>pqss0P(@#G7V+S>Vz0I~8(g)BX^GxE6{x*88`+&i%NEmP^Ie2C
z=~=;L#4H7AWlf?c?86WXYoWijOS7WlQ@d&iWT|Xsw~ekYq6LJHIGy!yYySniyy_Td
zG*7Kld@B0vc!$yU4<w{*6$!KLuEB2B`j79PNd3-Auc%@E=JZfhyk~v^x7oO{Lqu=9
z9de=T{hB-8UWq-~`<7;od`AcGon0i#qouA(+u1!jIXPLHP*YJ6*U_1}Qz)&H6#Ro@
z+vv~sjV~PsG;_Ltcd!Dn$XuVfh{RyH(1{z%=HI1spgje^>G)raB1+}u=~~ssk8!PW
zm|S$=iku^}9hmP;K>Xfhda0$5(%#&`x_$A}_RJphqfOM@{p7!4-&d@3D#B@hIN021
zzg|<5q1p97lNy4~m`NHEbhZJjDfK+DhhhW;24O53#lFx(d<2*~Sl9a`5Qc&&4kG5|
zQngxkiJBX=!e4Z_J$bY=?}JN*%r0HcXYY=jD{V=pQj+*<F3RTfQrd~LGBY<r*9NE%
zKwgl8?5+oE7SJ&;1Vd+3;&EhsxH&1Opb)-R0(dAgXG6Lrq)lBu;B-rr+9|{^sC%2E
z1C>9y&irPG`12o^w47WHuG$F=&#r0y&Pi>YyfbW0iBFyV8V5$&2Nja@LfN^>xd=9h
zbkS3_$WTeu+zkE3ja`^VZbLQ71xQrB+FG0s>fG<;i<z&uDGtV7d|1w&v&a{e&2#-Z
zx?PmFnJ#0=Ea?(r3mcXyweJVhE&?K=f7roubNbLKjMccXLpE&a>x;-vbq{H25qI94
zkdc@FuvP*?U}%g`p>>VOVAyoOH>aFa&ao|l862RS=Kq~#`Rs#3l@?8SVc@jxzo5vo
zKc5B@Y%ruU;~dn7dmo^o&y=)N&i_1`YPRIvgK_?~{}4~d<pyNp1U-}c&r}#HMi&<s
z1K`}0d3Ra(p66O~XVB+kBCV^t%SQUOf~=46)kNgY=Lh6P1|n*Tti|(gpl%j^Q+T1e
z^I3B~yagv{p6YRj$nnfIeFFn%tLw#7-<!vr(zDbMiX}Bacc2#inv)aLX|LD#{L?j3
zuCZeh$)LpRF3zoT>RhkC-n`2qnm5b|9g|D^u8XR9p_@3>uhvL>H7EHYRSxe!2i4}l
z2Zp4_)xXT`*+TPGxkQe!4vi9A7urU6c!kU#2&`iF-T6yyeTDNjL+{d7dTeGhq$RFw
zQzX*E_eqKghlh#4%fiqsBQGBpb=7Pr`}!8iurrsQPS|RAK3><IgH3XB_%ky~tanvG
zgqzqiWJOwNb+=`iEo@TzRm%LMlx4Bx_+GMiKd<@SezWl4MHKFnO&1r{48c#bdItu!
zmQOyM$GXI1FtW<cpX{h-Mk%SQCqwFZaw>!)OtZbm^&%BxR3Yw7UAc_R4NsQ}hr}za
zu9I^9MB4Gof0^TggqvFno2#o2vG9~?91~E48+&_)Lp+0h2+QnkcU)hcy%~GRf!_Fq
zhrD8VNu-O9Yj*Xk78&s-+`<C4oRyJoTE1sVL1{y)+_|<Ly@n<<5~VvEbe6s{Y};(f
zSx_fbsmEOM##@N<Gn7?%gI@@>hZLR{z8kOe@AiUB_2T0iEP$%z8Og^9)D0UyrL>nb
z+w*?c#|OUgbk}BG6LuGnSW$r3cWB-}C2+g81h0<rb|5`O(n^g*WsM7lb76{iwOG$|
zvRg7{8v`Fd(p>lS^e}0<W|$%F0@<egwRAq0xTA2f?L(vhzQ3cm8_}4d6j@8z=5<q>
zUfKG`-ZEd!FE*Z?v4BGEThR+<7HG(Oodh=jDM}%xWFx0t3mGzO=M!MZRyEUWVlw?p
zmO0!TBX?-eo&BM&-6d1$36+aF<0a5zR&Q)kLx0htbu-l3?_W721(l43w6q4k`Oha#
zw&Th<qeL;oy<A;~rOf4#ikr)_j-TX{U&ToK23J4Zy2)3In_}&s<s#PP0(YW(3ejtQ
zwmsg(u}w>I$hPeDXq;TA+)`vbd}(jxxUlSfY?LEU*sb|HPN9+*qt9J_NPD}YbUMs_
zJDbo}os3Iat~GAgK&fE$r+b3Rp}8ScZb<c5;qLB2%O(LOPb}C+QlT4m?X%f<#utg#
zuBAGpx)7`P&JGvvDh02j8xV_%K0sk&aZc5$j-R9tJ@$NCCS=YD>*C^6LwGKh68bW3
ziJv(8w+x0=dET#}=UjTDLY1pJ%Y4_%dTFZe+<hFuKGs!Nc3(KDKkmovYm=|@DghQy
zswgMLse4nr&ZfbMZ+-eppt&fAQp(#;Hw~kvM<9pBI>+K=!dTmU=#f1~W`LWEbGA`{
ze2>nc41G18(vz9+t@LelmdbnfI8`+nf3vx7LN_HxFE%J_d-PX!R>`t#cYCx(rkmx>
zRLbN-A)x@ED51PzeRGaGqdV<J(KO4mS(tMut%aI~`>u?_whL{EA&*PbP2!FkrAArO
zyxW53qY|<nBPE6i8r1tOW~fmd#x#!AlLA)9Z50jH0C}Y59TTynt&?wzEE-p?nV1R}
z|JJsb>yg(fj+s#A+S^=LxQT@uao>wtELq0F*N<K7g12}PYiWr7*|gnZOmL`+-o1P;
zR8O%K;_JLsr};I8cati8X4zPrG4{4rCH+McQd<UxzTFK5eBA{r<A*69X<LzRZ6=4!
zCwU+Hl91msH8OyyE!#3roXiU}&l`FRxb%(tWi*(~*T}lMGDKiSLe!>HOzjT9BGiEH
zB+Cj=#id3JDno$xkl7rJ-}qD+&EtM4^qzbob@UB<px2zFmuu^y$`g~(P-m0kMRC&6
z8;XWei=tza7Jd3aAi^%~Qs#3}xMI5tjWV6&z6K>BHJ3^6zjtKTGmv8!V9I=7L}a>r
zqqI+WKGCA^4bVLDwQD{CKmyS`dr&*qUczdY4Ew>N*!k<?JuVFnG|r=%ns13&48v8L
zGZBhnOON%MRvmr_5?#N0;TGMYyOH6e#h*4cp!(JD$d+;dLm2>hB_;e(Er;8vsHo`I
zuScb(Do9IXqWELrF2X$80H6qNo2g`A)YMSER#9mIR*%A~SNC9qy?~6lM!WD*^XChC
zXX`{1{Z5)cPpd3@^W{95d!XWV$o@`O*A_>#(o2AQ)4R7uRb##Oa-QbnhOCLKCplrE
zwY$+8(QwlP?=!_CE8vNZPKRVv{-6Rr6=0tO00h=uU3vgiwX?T(tpQq5gn+|<4G;}X
zZJ4jn#5eHBuNe?G9`?PNPwJh`9aRWi7Tzpe?_?IjTGI*0tQU*GNI%N#?O^JoDgE<{
zQ0~w{sOSTPg(e<WLs2oSS^Vr`XwML+Z>G`z-_jDW+N5o5IT8{QfE?p<b*76TDS$94
z5e33vYt-Pt{fX7zvUDg!0Mb+u{M(wPwD!5(tCuDh_4ds8>lF18lhLS9Tr0^x!o<Q<
z(1^Y{KkD~cb!(m%b$>L!c=^87cxk-j39XR%y$Dmk92F**r?vn-hT8xPfRgcz?oW^W
zkET2#5NHxe4|8j4avB;W6px*Smb||l(6A*Hf8<!b$zOMmxAulEj$P{2)(Q17pRJ42
z@B^#I0BhAM*N+<Xy5&1AZJ|s@#&*p^40W=^EDAG~@_l+rZ9z7ikp1>Dkm<>?@*U`Y
zOUIBF-?{?Y;Ozq^?>$B)rY1{XLYO;AI&8jD{lC(Ev{C5a-Zw!3M#U=>OJ)D8Bt>hK
zmNsy=`4`-_+dTjb30}dtmi^Q;^rrB?n#q;g?Pq8PH4=|#g>in4JWy~}skwgh?&o21
zV0k3h$elb^Ri*rg&l)+1cI}FfD_b(tE)YZX#fwNXCIYk<7N0LjmSo+7)EZN-_lja!
zVQJF~;#PYfOlmLG$9s?*vN}7*uGPLdKI=dHdQ-uph=If9C3fiu#cYXRPlD%jmTycH
zi2?N6IVkuUl6yGzscoyCL3@sU;CozstDY$o8ynVszi?(${YNV};+x*H_U4WZx(J@R
zxr;1Xs$Mj+w5rwkkNOC9^NqiBJ6QbjVSQpbB$j0Yde-38{>qv?#d^kGwT{{0?H&#2
z6?!lPvZHMN43D|n@fpx0pR*?)!{E_(&e4<?#rl*a@t{5@y%Rq1Dxd#X;>Er0ssb*r
z*YDJQCYv^(U&>MNp6$0Ha7>s<X1nZXl0L;c^uG3v+NA4?8GEql*7~P<0$I-VZGJ<l
z0E7Ccr#MSuntoyF2haPR9|pAO>~FI0VzB2+>uKxrrYS6Ust(W$tJmZx!3S3o3~6GT
z-S~NhTj<>)AT!@l5c!>Qqz)5?r0CBKO-%cz#(d_eW`$UdG0%^PS(uzT?)v9qcG@Sa
z_nEBQ<M926eLIOp$_|<MUC-N7D;oY&-Fg<ttPayudh=0QixUHZ*3bQ~4uk`ipxK?1
zeQA-rXTh^PQ~R|=1uffI;EeqHVN+Tfzv6?HTQjex1qh4d(@Nv?wlRuCInd9`Z{_cf
ze)?5^LsFCZ^9+Bip{N^MtdV~f`6`d!HgEswjU~))#dmu}YXX>8!tOqP`EhG0;-8Q(
z#dMs4(g9m@(gPuc9J^_Z>HEhb#fRv+CEe1~dg|-;G1ON%4it8B2@YYm#~&3Zik4iI
z`wuoJ6TL8N<ZBV2xLEm|Lws$@uh6NT9V4@~cd+ZQ<U*WASAJdcO!8^G)pU~s6pu<V
zIV!z0rGhp~IoS@{a&yIgJ##|O7d#`mo)IM%B;C}Sa1kIWa8A|ykv5qt;mMoMyg_>7
za?fA90N}iUY$c1QTd5I+1NuuVTZg54hNx*_Te7dQA_Ska;ocF`A}a0a9ooVzW~EJ2
z+*-arO9Q)wch#-$usv~vWqFd<T${4*hWXzw8{n7cxLB6bokqPI+RK0Ve@)W9Clv$Y
zu$|6geUpKe{#}R7W%+nIP8Id@2*-?@G`$_B<y=lDsJs*%Mw=AcD}Tk=2#DgKm!|1e
zHvX|Q^xJFwxdYB{Tqt^L6~a7bVq_@YUY>iB++8%&^?owg;)TBn?p<rZF9U2lSb6pu
zGA4;WzW(H+?vnAW@25%u9>6OB$8wTUQD?-PZHu<s3n`PP<m>aB(c-u1Gm8W8)}J%$
zuTNOHt_*oTQfhjpMGnN~)k<AlW3HXh4M$7z78Am1Ws#ZP=hY16d;)>A;uO4B_4Kau
zOU)B*PA^kQCJ4eb?{W|mD}uKP{(|>EWDk0^Z^G1vYt?F{&SU2y({#;E-WR|%gnpw8
zV`zlEHvXf)VjKWVe%Ebk6d<ap>$n91yR;@x1%|4erD3^UxWy-?Y*Z6}-(TA8zQR|X
ze}uXVJWrrskO3DmLD)+$&HI!Q7$hhY_Fb$n&4VP{43yA-`uZz@qx8X*Lc<QRA^yOH
zyO%aQI^m14CAPm+%cAZ%wbE_}D>%ZW7AD#cq{0mW_#?S>Lm`=e_AezlxxBbIIznDK
z-kv=^KDR-afvg!y%x-*negZ=f{DKh!&X>VmOr?KXz7#n}D_QD8Sw$<<Lc14?0z(-{
z59&9?ibi2X!#tT?TpS6sXy6|d%U0wKn?pZ|g@e=9-i{%mRH!j2Hbf-l(D`Y*HSKdl
zyo+<Do0o`7{6llph-1M`XeuQ$XNPiMwzRaofCL9b3y!27;5eWbziGaK%-~=-v)WIO
zk2wr~;Wdm!=vR#Q{jJ9<AM@MS-+Wyex_EH$N#N^)1JC_UHkly;$^kTlp#?2BLUDsZ
zNpN0K9cfT=V=UuK;C%uqLaj*0OKoj_>vGB*KjX*aNSzk#J?DM_reIu~4VGCNHfBhE
zzNMM9z*m5R>2JA3(6IReP*#Bwgo043`itK@<5Oo8?}@Vg_Nk`+Z|iW6n5%^fMr|48
zj+c@vUIZ4NFVI*=>wnhPn*J@zU0dS1VinhR;V+HtxamAlZuVq~d`ze5naa(R0)mW>
zG9An&yKhs<-rt(pP}+F3%#FN;GBE0g?$hseRe4O|=!<!|Eqx3>UVgIr(0`i)`4a*C
z9&_bcUJ5rR=k`>k0y@b<A@PIgX$~55znGJSiM?SqG0y`}TMHAkpGG6sl`g&j6D>%t
zr)XyL<T2LH;LdD;wqZ_Tr9OD_wqCnE?9GC6)3w^;<q@`kV;R?AW+{a?ShMNjhVh&y
z@GV$7Zw@Pa(SMmS+Xc$p6vxDkO@7(-<>c{uoIRt%0s_jWC&&kJwA*)8I!xhc2Au6o
z{rT^q0z7^(vd{=g%UVONI{no06MJc!bCd47RcyZKyTN#r_j#t>0^Ks*g=<9=nx0jr
zLExmM&im8y?|Kd9!Qft9@(%(Yq>ks&GA?;vv$WOFC)b5h;runV=+^KZ9{)1Gz??>f
zJET=U18k0@JXF(DfiFbZ!L|ZI%OHQHD=ezpj-FGP4l^M+fx`z2SE&?<q1R}Q;yGqR
zHv&%G9vy`j4&icC);y%?g%(cyb}4;%B$n>_XE0LXx}RL>zjyJFehH7~*Wvti`NaB|
z)BzgKzp_KF$UflJQ*^%heNJ)9jIi-_=J<R;d>pz&4;=)X9Mrqkk306Ma39r~5T;hM
z;S^{tY(@thOTmEd-o1O^#bDd+BR_zSppYUM2t3-`q}-xe(YT*;PAqW(@70qp<`v}1
zhegbXr#_$5j>aSI0p1wQZ4F^SY+9~aSa)i^S+gZP`U5x@92O%iK*Tm1%E3pNty5D|
z*7<6|K5FLQcPEZ8{y1_+9JM<c>gAn+Z3X+ZK0s@kwJkOsxS#5_crTI1q751bgs**@
zj2C?d_|*@A=Yp_Bb6;lfy$e!X!)I3?LQQCuq&#zGzb7GCS6lixg%qO-mh6%zVp*IW
zATjfD0y47Dq9S%08X6b@RQ|9T`yL3xJ_t6A;M^U(z0Ou#3GxV62#63MXb9Haz-NCo
zr0+xCWzbDOUMxUVSC<_W5VWm9X~#WmUd|BmcG4Q7!MHSi@WRG5Qz6>`n9f4`{m)>)
zZVQ}~QQpwjCckSK2%8tkd#2o+-z6%8Cex963rwwJxP->j&4iB><E*lVh2~rLgWeUG
zAmIVs9rzmZmX<6)a)gg6cH1)sPE|8ZDSCh!W(q4<RaFJV*D6XfJiG?r{-Qzw-ez82
zy=D|F$Wk%X?W6Xvm{uUdtXY%mhevFU9Kl?)oA^!i8E7^P*G3AuQy#|yfQDVJ+4Jr;
z8A7%{-)<(lb7u!61CPvy7=Zoi)$1kgis^E^=Zb2uXuV+g^m-_rHtj5p`R%w2230q$
zILd$4RG!<Yz@`O0NKFlq62_-|=S6GT*_bd!t!cKmOYeT%|0VTRzo_dji-xaS9zsOe
zWLwz!<~ZpumE1(oG)Ugfu-JMBLm*YRQI7A2u}XX!vh!UEkhr=aadG&pVHp=Ge)Sl1
zXQ|1efBwvKQu`y`FeLBbu79@1j79k}4<BU)&|_50jBF~I@=y-;FtpUuL0e4A#5I1b
z-(@E}?wPC*=C5I|tC$#cR)s8a3OGr;o`Eu3(UoIx6E<_-ZL3xQv0ikE2WNx>6rO>t
ze3Qx5W>U#p;8W{uiu-P^q7J0!o!|U2tYJPnr~8YEcV~EnkYy@C!7$RVYQDzS$ug&=
zi=c@Wf=eZ)Ko(uyC0Cq;&sY5Af`#{utRw@uG&fS@!yc@%K^|wLF6jRi@+4cMc;lh4
z`)yopvaVpZu<fJhh;Y1JM!Y$jb)3L!*&!#!9!z90_K%i%CZ_p5l0BACRxEr-+yuCJ
zaD8loRlZ1ab2Nn+6bc#I&7!IABMfYlPLEd?GVixccVvqVnMsDGSheq+T)bebr|4}n
zEFUZ~;zzpo_O^rmUy)igy>fuS3yjekM?1)#Ea=@Oi?-(&Cqg3l_FrSIL7`a5?~+0P
z4Qq$ov*?&t0wsB=dgklfcli@^pS}qURKrJ|7^MG5SwdsKP>H7s6^;3wnrAdW`=yoD
zu!6qJ3M6WGMy+%H#%4ApSbvtCL;bEi;)|3#N7$;2+_PXgKrV#BJIu3-o5XYy0s)l(
zcUAZo@4#9A_Q6<IEQuBR=`!Z-2tLJXmzBE2zyrE~cL%jQF|!q7TD-ev1UzriJ{35B
zP<sroi5;4vpBb31&vma=k~1XPhcooKipKgM3T)n>w>^b={kiY44`~F4;xFxH%P^0!
zy+8W}i(b=QK_|}m3xJl-b5tUUr%Fs$<92P+IM>&4FeLL*uuD^vW5Nhx;^4oWrEvr?
zVHw!W*B6XK!gxDbd6~L-JG-4_ud35p7eA*)gmd_?uVRp3-rUY`GNhip)Co8hunE1z
ziMP76uvq~Gm<-xKv}EGgI{^G>54lGyy<Sq!-N#+ny2^K?TKMod?S#MefMnsr8Iuo1
z!&QYyyY7)4!_~OXNyBBm&p3t8<l$(AThOz{RolwCv2hA)wVFMXyh*>-#RhX1y>fx#
zKX+JsCXGwRO8(yIcE&IcrO52@>o9LOcliAI^UJ(}wIPx+(ppE-TbevNf;?bx!tSiz
zH&}0@UF0m~(RV8<UZB|AN4p(xvbHl~*8gpInVU(4$?Q4Q@6x_nzGKEFn_uV2`BeiH
zF8Xd>V@x)A^JZ9J^Rw2>qouzc_gqxc=NubydUrnWQJARuoCU4!PpN!X&V7+5TA-Y+
zn%8fZjA2!v8K!eKN)<-Rp7IXGf3ryR#eCNAOqRo?RsT1z@8e}|*Q+}?Q-VC2VWcT-
zRk$gVVLI~@_3>(~H`%aeff~Tp@H+d=-|dl4R^g$#c6GlSPi**7*9yNmOt0<p4hy_}
zx-j9pS~RYJfy(qEtu9SA@U$pt*e5VR)9G+}wZ7>YCe0;l1U6pYFmU_Z^>S)ALH0%x
zCJ8KcV4&o0sAehW=1@^v2w~E+RJqFgBAPKE8AD%>Oq}^9Z2c|9LCW=~+KD+TD#~$l
zA^~WbH6_ZRI3yFgFA7ROVBi|gwY=xH82Qjn>G6T`i8~%8C61`GGcRyj78a9izJe@k
zX)@yU{^=tSC2EmJ@MnCCkAJDACJD+ovx&+_pmc!pvGRutFt|YSH|4qa9MLDmUfnio
zGA@n!_)$_*lT<)Jz`jlwSR;f``6YKHJms+eCoLIHF5~7VKu9N#2<89@L3yF9+*Iqf
z$H2;J1augXNVRLz-lP#Pn8QZ}C?qs3jjG0ZBLkiYqP<d)&J*A*0_l{7oxN|EsYJ72
z`QYmHK0QDGCt@8}R9^vyMG!Ae9B=|dUu`EC1?ZAO$20zg3j2r<LPF_yIWxi)L|Z4P
zYeg`H0S=AbsJ6=haGM-9+l@;*psETOANnyLvD$t4)+b#Sw_3#nXI_=)dQh{UWF`>4
z-f$)W)hSMs%7+h%1JFnybVKV3-R&wzK?=3J16`X@Q*M;F!y?d~IHkJP4;IaCoei3o
zLxBu*M8Omwvu{>UXPst`I(>B*_9{7hyvLLL74&7fDuqXq1!|dUdD>@JXM1Mlny}%Z
zI~rYsmsHs~I>=ZPj;|j*aUph3>-D^yeQ^Q&!Fn!}Iz^x9ucK@1ubEG^%C)7PeCT@5
zQ{dA~%l~Q~B!J?oJ53}F7;n}<Qe+x^{rp?3CDif)&$9=i!|}bkkb1ip-_^I3HtGO;
z-~~YbfYoqWITO@oT@eH1d|+P-Y-KCwB7AIMs26d6m&+WmT^(WsS>(amsEZd~XJ_ZH
z-rldEk1X_W0u|==(GhG136v2f2JOM2q0b7`YJLQQ<|i^z)c?k9Ol<7xfBs7=)k!=S
zi|$)&EvDS|@Nt?yEZuq=ocPH>o422TB$GK1867PKupFS1P_w`+kewYDS}tJQgX|Dy
zYuGn$f*CZ<+9K3qAcn%jVNf$NzK7}n%%*GVrw7o_TaBvPq?Cw2Lj(^3+{te$OmH8t
z+Oo_Bzb<)F$<q-6Z`0(bBG4_re0d0Ro$cLS`=5L;5`<O>WTiLwrSJuT(Ei{7pn*U-
z?d<9TY2YH1UZ8$5G<*mO352}}q$)W|rJMm%MDGM%2PM(@{Fa)NCmm2(fHgmMU4*be
zvs^4o*>0`{3uF`k!Ob@KT@#m-)Y<6+9V2}6TOiJ8mYLf1P6zJ&RD|Imuy{ax=Q{VE
zo-KEgv{e9qs{Zn2-m3z&g%p>msu?d(W2OTnT57FZ<6Q6^)?c<FD=UjJ&Fk=0*7!#@
z?Qb9!)m0*rc$6UD3H$omY#i!3A_=9~rBn|Eh!~2hI&A_!LCYC3W?7eWW^7`@40@&=
zw~BE~UZb`T_m(DpfO>KPB$sj4<5dpX-}PIuNBgA0$(8Q1YW_Ptfko(?sB+L2ft%W8
zot=MVFof8HDy8PqOQ06$|HNL?ur>@J%6qbP2t00t`7gBe1HPF#FfcIA`t?uu@?C(H
ztG%bn%woP<PYRCqc^s}QA@Hw+5|I&zb~<h9FTOH_WPtY|ioB2><C^8a?1muvlvPt3
zk(eCSFVPw@X_~c7b~*#T;5W;$?^(aC=oM2{gkeT3TfzG3ySR+BG$`1VK1dp5^|j2k
zhQ<NH%>*3tKH;<c@1P<G{dT5bi>yMceif{Ou`v_qZ3VWR^M>80`dUE95sS!tbK96d
z%f6yUwTJQXJe7zBXnvN?i-AqV3~L@|4ccK@YisMQacpdCP}_z$2u~-B>$uJac9Bve
zm)@##faYFSL7~^JLU-8<%;+WZK;z&%Oce5Al9{08^O;)e>gcfZDy9ORZm94d?GN4E
z-Jx|hAZ$~9@xpd$1Hwz~)=OvS!&eyR5Ox1!4LX>a<pTF>{j{pKcH%%dE;hFE$CxGW
z0;n8+F8fc42<%;4&=a2}@W+Q<$OG72<Gr=v?2&bmQngXFlv<OIYG2p4Tcx#X(qzb~
z`oEW)QSzd88UX69*`Hz>pdW3rwH6@bcL)VWNnhIt8AF<Yo3spSkCZwFUXXtPZEN_X
zuf<3$OBm>Bh^8Govigh%==CkMQ7Ail0NsC*r*ny)RRel4a0Wst-@Qr_AP?{2?H(gc
zq3ob1zNsrfKuu{5lr+R}@^yaxoy@8n2Ek7=29;;S=48#JcF1dZw{4Iuz@7=a#Lw{p
z5uiJPD~NB()gQFfF7YqvrF|=-Kt`ICg~-K0_@rH7VrW<lYy==Xc9@rc2T39*H>yC{
zUZ_$vj0CvoXr$L+D$fb066pme28d+ID<7n*Uz4meVAP?WF=P((ZD~-df_fs3O+`*t
zR)4zAi@5gq00`;L)6-W}?2!lP@P_0<u909U)ogbiA&7yR5p>ikM|SDy=@WMC6o3!?
z9vm!Esd7q`(w5d32Pyd;Fsr|VN+}im70HSq_%5nDWdeaDEEbRvTA@wUsPh!iEZD2J
zm}&4cG&cSY@5O5~)eV%=2>6$z`+Rs|X^G2bN(d$=tuXEv1G()ONLtc7cd?*jJU%%=
z`XhutfsmXbWCJT;+@?MSMk<Iv6A|q|k1FK;EjWf=krP<o$ftid>Ix1C(eH?+g^d6e
zhbZVukSK%@Rzb-pQJ_W$C|5k2Hox2MLy)h)P^$9GNb5WJ)$U4mnmy3efNJs{T5m99
z8km)}o0oIi&LA1c`6KIVM5&L0G=#7*v`_;yHHiVPHF<V_HH0br>Q200^DNLV@^*jA
zu`^D4?uBJA^sWgn_N5g;96*!@ot<&acVare7#bNdg3cBRJV3(DkWb=$4ytlB5Udv>
zwaK257H7?5AqZL*KxFKoZOiMnYiI<VW^nyZI2a%fsKQLptQH>4c3Y5ud^ZkKg7*IP
zXr8zp<Cm7aNeFX+mX=lm4m0QkYIrQ<hgM5o=1j>jWrb9sTyXuGsUVO}7=b->LfI0R
zWK;JlLBQc_jcyxlsb+z!jLdsZsCcp}D>WEdSX4lLm<}ZjBXrfBtQ8X(uxt!q!w|?m
zM#sfv0b#$gutWn1<q&oG>ainUX0b%x4JJ!T-GUT0%C`iX4@Aa}qOY&7HK>Egz<wKm
zq_qXq=y3Xizr`C?MJM;dT(0S$=x739Y@jYxbU{w3c3e@;%>pY%v<481Zcz#yzuo)n
z_IB{4S+X}3gd**i$)HvS(WDf}>p%hp{d*-P@&C?W7^}3KccUjbc&J%0a-aj#I%F`n
zvqQtdfd{%}wLAzy>-nRJpy!3qIsifeF!oJ`pS@8ixu^<?sTjQ-yCCzLNe*S==H5?^
zSQ*YE1}L&xpGOCaDX!DWFB#@Q2VP+Dqgu))zt;{Bk%4+y8}sb`+04R1U<O0sWn?|O
z>b48OhCh{Q)oKRQus|!-s0R>6mE<Q7a*nP{aPdGX3t4SbHeNGZ=D#QaqNJFpD8EL%
z&w!yq04as!1#buy=40UeB-V80J3}1-%L54-j-g0@p2S;bGo4B+_mRW2pBmtmr*KNa
zq_%N6?Btj1?3kpau7OM`EKt;guoICdgW{{(xYQ!O>~A+zFmMI{9Cv|ae?S*tG2xJb
zXF%Nosc}qRzXfI$!DJXvH%m?WjA5>7%1sK|h8CFqod82iT3-Gk91;OC_nMk+b9K-*
zf`>i_DKKcp;VTD0kZjzWECc!pgSHR3gv>IKsP{%*+eF311p+Vxj!VbH6oL#8p>)}-
zJG1`#!)k4lGf!JtnE+<WU&_j2BO~!ZYjTMTLP|ioKIjxJA%}gyi)^x${)ozz|L^1x
zc_TQ7WUll7ee_7M;H<vMxAHMd)zqG0L(Zy2>s|D#{ZuGFZuUN-11K>k8r(T5I(lxC
zlMT@%=VqtQftOvmPJ`rQq)69i;0(sgpDQa#d?70ML1tqbb^8hdQ|E_QigD=}D39XT
zmf-A*Xd5`Mq4ucn*RPrI$D4W^f6L>`ag;oM1qNius^x3=DP+&X5a33GC9eYnnr>q*
zsH^Os3)rv!c?3IP3e5@BU&%sl(cRthP`14f3CV!-PaqiML(qbMSxXEI42bMDCgz#E
z{4J24gI*4*Y(z7USmdlW#0ZcvLhM8?0e9yh4I{dHR~g|YJ2=1*3Sg{=5^&f&3uKGP
z`63Ak?_2fLVV%Jy;3$OP0aM54Vb&ZZAjoZl#{{<UTYvxeVPRSDAjG<!QQjA)j{tNp
ztXd?6KcE6<`vDhM41`7k5)x_HXULMt$>dNA`qkI_WR040lL|O4<3WkKyle`IK?15`
zsErU29-LKTZOsNN&;-d04$gS-0<R~L2b3P5Zico2VhSEHaZp7CAM|377xD4&{Xj(O
z=H@1}(?W=x3jv<-66$B5HbZ2mWnuXM<+IwiTriCf6pue5;v6`(K(%WtYTFXHLbA}~
z2&VKNIDjsr{(Sc)ghzyfS9`LUgecNs+>1v>CI{&PS_va?ftJ=*Xsxb6eb5TUB18w|
z$PDPW9zix36<+E~>jRXOghLMnrYMfu@3oYao<grzx%w^7_3~uNY_ggMN=uC9cef27
zz}23v=7CZ&0&2?bg?QavAYj0789o#_iAC4vA`Y4YsCqjAXzd1C95O?Pn1Mv+FJHd=
zS2)1n1Dx<CG21`rae%rOjF>D$E)Z=kPgHS1QHluJK=^~4lS1kC4?kax8LW{G3<^OS
z0T$i`&Nm?*o-Z_|Q29dzFwXM?^l!+cL6g#`Hv>}qFA#8`+Qp#~?gsh3fcH5+oPa_@
zP3;TW9UBgIfi@d5B^|gtXf^R+dxp(vg@lqJD{>lj5<zhPQ|}`J)Bk_aq9RAo0Iu7g
zA#t-zlNE%j=^&Pc1>k^CP+I6H;=a!S(wVqLM#4F67_6~{0_*cP$cDiRe?fBRx0^$U
zod;S2Le>YxH#BN(plRJGAGL3%a4Gh=@&*S&h$m4}ly4y6$DQIq*9r3Jf51X%oweNt
zJvOp~kf2|EH^c%;pWd%Xn%d*JP^3@iF&RoA+)24STDjl>u@?+yknaUM=7*OSObPFZ
zq37~GcO5-{iJXi8)&`1cIBP>1w#01w2P**c4ddfUaDoS%V;~E%u20-%1dv;N9Vqjl
z?*9nw<1QRpfd~S;FAmgi-MCgU=}fj&yN3e?jHLM0?_-vTr^8Eg+RcdzIxdSse7^^$
z7QqP-qDn+y%WZ^;LQlb3llW}tgoMaJ_YMU;P*x#CBa9w0n_A?ravxbhsfvSa<1I4Y
zTQMEt5d0AlI`jxEP~m{bSG#QW`dsc&X_lL}!pbsTq)`fceuAt4a`m<I&iJ_rAoa$j
zxY=Ir4=TQbN(pX*(^Fz%Vwz#yD}jKE)N;rP7;pd#D{_Jh;y}p%Lk%;Or;h6jXT0RA
zS!5|_7pTcWmx}`C2r&pB914m$x08i<C`3TAg#uXnJJ~E~C94kMJQ48nLYrPV{{X3l
zp$8@;B=m=@05^k%w&u{`D-vyizl5AM1ahiRaO4OCl24yVKx2)v)Ogk1@4!zE=+fwV
z2W%3YGBCHbwdG!L!#`bF>L<inFaS7FXy<&duM`SK6{a|g`Wya@PVS5%B5AOlP-ff{
za^-lLt<Va`xDXM+>{J94Jz(ZaN>LW~d*KqOI(d_B1MH#cD4RI|dIc=sq9U6toV~%p
zb&wsz6&G`Wl#~7RNNk7#i80`G&;`zd%)tcufrqrT2<QW-6+{pij>7Q{K;-%i5=_X-
zrBDyUdn|)w3fg!$S_V01260~Dvvsa&cwWX>s3g}!$^K6ZXBX0B9EI^|gcpY3@NESd
zk)Y+EMo6qkl5G)2Ec$XWx3UX$tu7{rf(aXn8eS|h7$_ABGGp2z7b=tdT6B?O(U2UH
zsTYxo&hTS}6#Wi*KiK=e|8t&m&htOdqtS@kk40+waK6G%x)4!pP+UaC!6@%=90AOr
z%^Dda+bN{ia+R{segyioM9e@)wk%p|&OlnAbJyJZ=*_=diSqc&I{@t}sKu2p4-i!C
zNFRaP+iz~*5ki(UNlnA$XN^hs3q}A;e6~*&VP6#P0TtJC%%MyslSoPoMq`XymmKqB
zjEnlMmd@{4IJdB^-QB&%wfStDL1NMncL}H$IS=3QTi(dr@Yonq$O$gLVK!dneOw$*
z5dC|P*~Hz7ZtJj4*8=1eO~+eTZ6`Fom#xdweQVK6XR(m_yimEm^z|XcUplx(aS*#E
z$OytF$wPXiqvB>DU@+p&B#NQLCRQnd>s_|Y-g2{L5OF6p!5*$mI+UOPZR0myEFy;t
zIVW>DR7P(<p>gr)nKoJDz7x%sqDDF5Y08kViv|S?F>YH5I35rK)+WTTB6?AZ9Q+Q2
zSFKjhTZ;{sGd_4TMFC7IzaUv@?8QNZsY;A$P|Q_LT0sgbzy%r~h^F~KS5S2H7y>=v
z(?|nYlUP}HZg7^=rU=?|p)x_z5rm!Yf3+;0qw@5F$}GMe@Q^h8tR5$s)!7lzqbk}W
zUCFQt9K!EG@S8ssQj0ABB1A%&m;AN#dWda;K3J)=0%oMCp3+n(EPBg>98t5P6-76h
z#*LkVzUp-KO6_iU<h_#p=xB4e1GtY5Zh48n-g~i8qQ{;R3r{%mdg$Q^oT>hbEFj~p
z_CbR_MfN^<6_yr^m>sd9cIxgvr~QXZuXoLgB?(Rmbu<jJ5pZ&QD!<a9`r~f$s{Q9Z
R$@lTaQx%>wn}UXty8pew9b5na

diff --git a/_images/sphx_glr_plot_zero_weights_lasso_001.png b/_images/sphx_glr_plot_zero_weights_lasso_001.png
index 29d61c94d79e9fa3d484a4e1ec8a7641b2484940..e757e0bb1987b49a26f0e2f6fe8f7eb3aeb00d82 100644
GIT binary patch
literal 22668
zcmce;1yogS*EYIsMFD*fB?Jr*0Rd@g6a@k4ZV~B{*mQ^rNQrcpbhk7pDBURyBHi7a
zGZ*;0@B4lK^PlsN?~HTy7!EhE_r2D-W5zYFYr0E_31VXrVWChcY@ugQq)@2SL@3k=
zwex4;H#D^NZ{a_7i>C?}(k9v#RvKnnC{YcI*M=q*hI%h=T56e@>zNoc(Xr97+`p-7
zVe#6WgPz{#Umu_|G1H-6iHcH$K`y+0rf7~r;c6iNP9*at>7h{aJ3>z$$yi4$4B03I
zqA`any(i9eq1Y}n%cw}z4+h_Bz9ym^VJItDOVL<waW^k<wfAa=z!x!=1lo~yVH(Ox
zy}sx1ib{#@dI5Gbql{-?#=f9VUfwwv3&s;{aW)_2xj4l|*82l(v$)}4vSc1zxHV-y
z;Kd*ScLT3$=hv)`-`%v9&^>+^6x(+E=01PyiKBN4u8&V0y~)1ha`v>B&DvzWcAnnW
z;*c!P+Lur!od)f^9YY6}9LED&)z{z7=yxUi8nTAx=H(syB^Yyq>2&7m+{{$8)27uu
zF;krAYxnk?_vcE{Hy=dW40TU(<ljDYeaRD$&W%%WWt!br#U&N)WH!W~QhYWs+*s<-
zBj+f)MfUQI9%cQ`MDNOKtn(BU6eUX)dtJF&A}qRXd<T15y{i+A2PD_t+#K`VC!S}I
z;vxy@Nmq=6dz)PztH$7*K|jf!N-ZImSfhp~$miW3!{u<?f9Rg}R<+bpuQiIhWGc#z
zS+9fKn~+IBG>m!0O5udY{`QLPW}gm=?Yb`a;r<#HE^e=^ylO&sn*1Io^XYqLUIz2w
z4feiH@3?FIleovz=3Y04u?UNcC$){9+{w;;(JG4;ikXzf!alD%Ha6zr;ej<>TVqel
z%d76!Rb;Apc=i{+jg5`c7emFs&R1GmGecG*vH}9GQMR*Ga0^1w4x6F<`Fe47OXa<S
zzMnI1)fUxMq%>XMDa@R-Ud^|9O>*|^M}93di+nc)nmbEM_dMTtt&jY`7b6;M-n>Y;
zovBDpD}8tf{lTJ~mztU^8daxBI4zrI$_5HedYjos(nCTBSPXieTMgSJd3bt~Bo0+N
zRT>vhTr1FRf9&h)%ds_RzTxAvUp=!rsJIoyZD&zSVwEhNB&OdM;~d6f0IO3{pJ8c?
zKk$8FxLlbvym=xzY2c<%ntYbrNQDDFK7PXH!oXahyjlutB2zDu>8~dRhJ8{nvapy~
z&+mGF^snz1_LxT;_cAieR)v3j|300b*V3UCteTiriRquoP4%>a+b3ey{X{+k7xQb@
z?)-6G{k{-~+3ydDqTz{Oj0UI~8KqWMR!a6)8w*Uw1@`ulXB#fJOB>4UzrmCi$z>zO
zHf$Z2C?1{J(IFMwNvo9q3>G7dS<lZwK}IIvu2@8JfkCf6+;7GC>vT7G?2GzU##&fL
zTZ)Ty+i&+7vSuh1q)Oe$DqHn{J<Vz|e80ZjvM|~SQw%>|qgPYW8yqmHqz!LQgkee|
zqW6incX}g>9DJS9l-xKO91jn+a{CN-m#c9Xs#y(x-h>aTmD{qX&o$6w$sA%1(!6R_
zf9oxo2FA%{s-*BZ>^@ZC&Fo*34|dXNiAbs3pP=`P|NZ;b9(v)v_%<V%aL?qEwAI-7
zm{#}U7B|1Wt@w%BEjKd9wh#u*tHi`<MMV#}w@2SG*66|$m2CDK;N#)Ny?=k2d##BH
zXRXw3aERSU-wtDUn_yL+yE20?sCZ{GJTM|6-OKBe(O{7zOnk|&@p0G|_metde&q-W
z33qqrvNozN(oeU=^4qR8vr!10!@1G6GZjq_50fIDL`_F0M&`7igUsdU&$mUxS)0ev
zS?ZN)20iKRb-s7daEJ`RT7+@ieW9z^erYsRB8yGTYGJ@R{_EFzkFp8-B~F{w_{EYr
zx~ET`{B$cW#SUjR>cE#*iG{~$j$nr^7b&Hr6sDA~J0{1)d$8`myD^tk?S^a65ic}5
zSe&=HG?Mky^V-bM98Hf8AKHRwm6>0U@*x{)*ltPoV0U8!W_$XVCw)(<Y}!1V&0@`d
zEiXnAmbPQuhc}b=aHn;4pfHPj7?x7i^BQeOSBi9?pWkP7$DOCJn|oj)=Gj72ODx)B
z9=*#qn`odeY);kGcHvwqlZIQBl9LOns;*WQ+gd0bnT;@Sk$nE#t9se-U{}ZwtYBb3
z!0&E_F(q_y)sVs0Nb|MXw5%RQoyY_y%v$s&aje<J>vUbkQ&DM|s--0)d$7gR<KxL`
z6|(tyRjneci_ToUmh7;%DJ&|AHX!@tByPfgb3)x}eDf(LL!V^v`_BRcc^2bA1$Fxs
z7ssvA5h1^d{Z$bc7nhmRr3z_DN#B<*U&_eICBm9M91Q6>Ytvs?aj>b3O~N7L?(Pou
zN?uPdh179d<HwI5?2#2EDy9b;y>rN(g2#&kyC~Ki&K69b)VL-KR=U+)Zg}9dx!C$s
zo`AthN?Nv;-_N6GI-ZJit-Qx<WI4^W3zErpCW;=-c!vmmce-L;{)=xXhs;~KkvXO-
znZ7L)MYfgNawo0D$?Eh|HJ0-a$7{TqjQaCan7L?an`O9OO+=f38FHJ+EM4U2Ojl$|
zTEi?_Rhknxf&n5BnuO<Ksv$t^iGN6lt|Mkg$8AR7;|)D9a?LrKwOndeJ6!k~s-->8
z7)>g65*>aH7MpvM`}_M#Ph>NV<Kp77A-2^aq;%kc{q}1-e0;6V`995P?YW+eyIfp@
zMdo^5B%D3(U0gyp=6dCc<mKeF)dinCdEZxH5bQ80&U^6cb6i~W)9b3gO`B>6*5D)%
z3r&`94V#Csva<F(4`obf2%!jl^zN)|snIZ(c1wVXKc^GeW%Cj*n1x;y+nIzwSChWn
zAI})gC!0(;Nfl~`RK^xrMi-0O^}Fujk&&5*MRI771%-zjR8>_qRpas+l1(nMb7|IG
z!XqTq!y)4_Bnyg;Hks{C)9L8w80<gyd*o|VWz_ct&;2>r6E)Tw-SXkE|N3gxoenl8
z;Q+IN)tiLP&RRo|2Zr)az5VjHbKN!O+A3guIMzzcElF^QS(}&*f40q!RF3e>*OECg
zZ@>js9DYfQb5E?>Zprg7*X%}28}nFv8Co11I4y!~0_K{72n9!bO3|=aWR|mCROWEl
ziNl+<qpW{j*jwgz!wEf$O%!4}Ueo7R4AzI+N8cP1P4#n)T9oqbS2!+<%QmB5op&dK
z)n(wN*H0y#_|P8H8I2(q+TE%=q}BAoY&jjuSG!{SGYIe=?&lqt)P1@|egA&?`cHq(
zIb(1|2P2f?QC;+v`=4=0IFjK~fk9L5Y$Nv4@T=S?`&A)Tvw8}@-<zxfK|!<r;ympT
z5_FfK4fYsIEoMI7;&+uPUChm8q*cs)0+*4F=5e6*F6=X;rlsx9)oNshEdr*!g@=xT
zL9+7jV5z`%Lmzxww{9h8C%Ewp<w7yUBsJ=JF&N8!EJrt~%ygyrtJy8ExK(m47V-O8
zG}0D6yd?hoxkZT=cxW0%#@vqO<m7vM*4yH;#Z~?WUx&=izkj&W(AsMCaO66zQas#^
zHvvP{_;&3_k~BZRzrbtN@feO(xp1Io18Teo$9ov-DxHI!xZO*s<tEkT*Gh=TzdD~n
zJRk|TtxU&EZBACnEcM28J3pAPg!oQz_Me8V*aQ#0E)djpI>+Yc%S1cwrnusgWrD5j
zce9#qdnB)BlhA3yRV;4dR=5@`lQ)-Jvami`E}H%OiTXtK>cO~>SQ*4%3x_zkxbiSZ
zue{p7yW^4c45dg@9!&8b=HRZa7s*?ZDw<+~FK(6zce>w{V4xDp-P*PV$06*~ZZZF}
zeaw7sa{>R_wZx%PtGUr{?>OwYv(+3nWXpGF)5LmZhgUo)=Xf0Vix?RhnXML87Dg)7
zYY6t%!I&)?C%&7XHym=@TOeZ54*-X5Ot!eldV|YaqSl8jE1x&M^SZ;%^+=j?xTG?Y
zl9GWw>SV~@K9V5bgW7wpBf+zwp&{R7ga_=LP1#7K&1yX**v*F!O{Q2a4VTQOrZHR2
zE4a~LwXzuxk)>c@Xdk=ydwXTvb~03_tI$Nfe6@ilRX%H1raR`*DKP3XL~6s6dB3h>
z+h?V!NUEsJZyudsi)o1;H8uLMS|#(jjpUIPSEk$J+*;VDZupEZfwS7(Tu1|x{?kPN
z+I@wXwN_qr5)O;k283{g8!N%w8hdXZgr+xDW`3z1A^=Z|L|l3c{RP6p!cvZo6=&f$
zw8~!|UL}%}lIqIWlP=V}DSpm%f8U;j*Kts9*fh}d??wFG$F80UtJqU6`e5P-&7x3w
zAZ{yHMNUo*1f#iZ8z~R>gVh~#oHqZ<!o}V~70<T!$O_)f>|c;6`!uL1uc#PGuGbX8
zK*VMo(K%C!o|{}LIb2WT)mt1a4r4V6laiLkCnlzWaKQJDV1Mt7JDmF&S7PHf_Nj=t
zUUjD-jfm0#3I7dS5z)<3CODdT$W#3Ayl&8!t7SWzlEe)0M)@ITkM1zI8b8fP5sgF2
zHQi=90$wP2VW4mhoJzMJg~*LW62?~_!1&V8-htD&x3!eT%*Dng?{@<;b*@5QHgowm
z(d(^2FE6hyIO7G@D=*6-FyweW`le!cmWCU%J<c-h3Lav%g>$LP>2OaSR&jbRtFmY0
z5VJ)@@3T)#kY4vqZgQl*Mq&8T=yi4b*q#h!8Xg|CsR;9r$UVb+DZ+-?U2UYNDWK29
zu_;+hZFcxfGu-6nXJz)8ZPDm0E#+A=9t%tR@)hps7RC&Y!*|EQImGm)yQJgXcAH(c
zeY}|MI6r^?g=nkA!4x>|1rTbPnwoli{3r+B_$)4|${adSIbYYe#+!Jm?V{n%;S1Z;
z^z`R(f<8TC)ovRzNzrn%Uw;f4-J_+IeRmdn9Zb6%QK4q-NB!{+o}SVWfRs<!%6B66
zd}j5F$w+2)_5)<6OU4NVEo^=N{$$9smZW6-BWFN#bQVOUGdZ=SYY%gjZJGviv>LAs
zS#*j9x*mc<&tNtL*9B3QdBPaC?Rpy|a>h&Z>4PQcx%u%cAzby`Q4!&H1iadaP0C{N
zY2<X_(~LD;$$d|XVKeCOuFu#`MOs<xZ<}^!sgJ}>_&SMSjNhLV8tl%YKI9k|^15>6
z%4}~|G!nEF7!N6(I)@Wru`A?;m^}!A-~eP>1x(VdTD<t+<0B6tyXE4bqFzLv#^UOF
zfj2Vm=w7X-aw7w$HO2C0qq@R(+N6hzk}?){)ak&@I53}%fh;5~J6mGJ>9E{E!Pq#n
zQ$EFdHX&HeY<nMKLSc@Bu(ajn*F&$Xv3Exf57d+DXTdbFDCt;RTPLLMc+yuWOr$BP
z2#JgLQbyZ<hIFZR>|*M#lEFS#_!>4L;}emf`=-_fEf6NI*HcEz5DO|OgdoN>K_qxK
z+@vCfd6*y$KB{W^1Alk)9qtXaSo6`ZC&~{t2Z$cLcmwg-{l2AygoHrSLl1by#vob}
z(j>Dj*~fCEOjat^V=f%l%Q1{hOtM*O+y@i%PVL{`or}b<+vQ968JDdH1O^3haLvj~
zNqsIhpGtK)#H_y}{mNB*mW!PYjA1%tN{BI1Dta9YVe=HTNl_YP@Fif(Se!8S$Q;9I
z%~-F~mi^;X2Jq9?vY4W39zqO_csC^3DQ{~lvoVK2@5k^dr=8bUzDcL}t*wh-qoz}c
z1U<}m9w|0Y;(-H3L|+J747{xmy~!kh%4Tl@)A8D@L-eR0jK@9ZZP(u)bvW^sJ8Sg`
zKZ=OmEPk<ciAU9H2)(-&eV7hgSP#N64mff8LyDqeVjq#9Ru6oLQmJLeWNUO5a#&}F
z%S*{;+$p1w{g{b4*qi$@U@R>x?16=al@D&vVRt5p`rf@|>&RXebMIl>Ia;vE)D#p?
z4SF-XlcF6{wVT6i@*^Nunei1QZHHJc9xP~cH1CLpz0JA%i^lx6wk0s-EHDSiNkJU1
zn*GYtNV~;ioTsGnq)z*$V1=K<(SsbszbirH`o!eqEChj_5Hoc{T)@3M9j8=bUwrlw
zK@1WGKrGcslU4Q^_BHp;Z~slLZp%tCcC#ls(GFjbRYWfhtLj9egLn?sr8CqkhhyK)
zEvCIs=g?kERj;hLefze>fKlNL*qvTHajw|x2d%R)2=F1wHWW|Dq<s%i1#lxrhY<0t
zPM9gDvcK5U$04e$FIKE{bWnvr&~|Ic3cWrp$$hZiS@6+({%20fN)0iEkPp}D2f;Nz
zr&COawtp2e$xW%tbGe&BTxL6;Q;WDvx3^Wu6hVye)M(&Kcb`u5Li=ugr?ci{Qz!r!
zp>$=-k42-n6k!wZ?u{H~LMB7a$yqAUI6qWcNIv6yj<gLDL0PyBNc_?pY|27$Ddx;S
z2X3)`i@<a6#ZBm+dSYcW7xjg8%BeV}HBGLv=46io3_R0ckmmm3!-Gf;G~?P-tJRKe
z<GXXX98Gic_*@g6#zQ6DgVy8TcR4uv0loSDJd&gP26#aY3|Ns62qi-xk&=e1`hWWL
zg06I)@hJ7u9dg4VVZ^)p`~f63Y{;Awf>-IQa46LSfT?;;C)KW837z$3#e&$2LAyIu
zHW9*uySHyw<wF?Q6v~v;;h5O~a7YuRo1!Wz5$;}I%;2Ayq!XVrfQjAIV>KSsb-hIR
zvV~*rerL7@zj-}Ha0sMdGD<<E44ih65<)`mJB{j?mnBw9#^B9F;{?2)LxSAgNMAYs
zAZyS*@RMw9lXO=3W*EE0^rzajKu(B>AVai!{`m1*<Z_S>vV?|)vZ#^jxB4I96^q4+
z#yhJM0)9DthT4$(2iDlx;sgf-gxnPk3;Fr;r*Lszi31-7dvJfwc(9_bGE<Fk%Gj%h
zKw3uT738j>78d!{G-^|=(I(`f2t3wZGlpf;UpnZJ@yi3B<~cxQqL2aUjZ~D|Kzi5=
zSe4P|&!6Wzt^H`c7$)HiF+euWsG*^;xq@kOt5FEY<KDp}FFrhNC$?@0S;AVca+HmI
ztv9g_gl{r>yO14)BFPY@Z)kjRW<$z<Z6xqTa~i9remmZ!Sua5DsJXdS5&P91$BjZk
zNG+}3mz&^<O+2}Ic=eOhiR~)H#e=I#GCC-Cw*Md@=fB3=)g}l?rLj`(D|d(>lEKyQ
za$ZH<nujP3C)#O$MeqQ2XAfl9^LYeyZs)Zk4!h%g0Od{1F&^S@0IbILx36GVo{l&+
z5p&(htrO=pU|~Dq+ufO}oC|$<f_XZvaxf2z<y&f5S>bTinw=#cENm1?MiY!y7_U=>
z?fP^q*w-D-8>m~xV6_nM7Z4nrR9E*bAu(}wrIyU9kEuNH`09&CR|jnM1>x@h9)3*}
zvpTB?89c=7LhbGCfG1&a&Y|oZ2;%MteoWSC48po-UTOS%@$}`{x@eRH)8b?+hpO&2
z#0UWZJuE#yea!+Fn3UJbGq5mIABlQRw1bDzPb}G2OuC2-1Nw=Xyn=usI_?B&gqoTf
z;z~d6>m{Cp7jAa>@}7v%9&C<Am$XS>>Xe>8zf@1kSGT&pqH0EoiV(n@eiN2L=hQ(+
z{x5v^=*<g4WKQ7qmGLS->F|1^HEW^s3EuFj$|`9G1WL#3-}&L;^HHkucOn+<TKYR-
z)cGgJpCxM?j7VEgP!3d9fHaqYI3Ztyim0EZ-N*L?-$J{^Nf=}~wyC)|jLRvG-}jN<
zDwZdU4MsrAxMf5773u}{(WT$`1OyQ#&<|0Mzf#%BhS6k|CJnJr$;9`Vk=x<%3zU|a
z5pwnLiXW_csWowRo9-Seo4^G@{J#pW`V`SIcD>^G_bs{me9X1@kDq6P2W8xIB+WMd
z5(yWpboi`pCavyz1{L)A2s__?W*s7`S1m=JB9#-nZ{DAM)y1ut`YDQUH17ECZEczE
zVcdzqzcg=?xKe%>(Ax}ALmpAw*CnWfF|MH0vHe`%UV&vVec8;ZYzc0sbI47<4~tdB
zLu<HAa+pjq3U!oSaA3*SydloG{cLhl1)g<w`Ru3fKkp@Ai^R5_Qg}moH1oX43Nq8d
zg8{eS`Mff2>HfV#OjEASoRzKMKV>2V-{XCH_B8JCYVFqqixLPodUXUF2j;N^;EJf1
z1ZrR9qdlJUH|pCZ7x?AG9wkr`QZxFQUGxluRqWbqzOT-EncDg@QY(1Vx)XPSA9JPm
zscYEtb6?qy$F*-z7bW-=7k57rC;0<g=zfjg<7y<f)~I$(f!z!H84j#a!CJaXu~fP3
zhSaN9@f;S@F>s0$uOLO4H>fEgyt6bLBF3rkc*V1por#^IY`YX}1s3J!9Ye#-Szh$$
z5N-7LmaGjKdfP4-V25ld)>+c&3&oB>?uEn(YPK^Jg^8{8_3hpA>KP(*#ZUVB`jFJc
zxO_v-0Wv|k_ehE3=9`_g3;h3m@rF?+eU|U~H{74O1W{8~7a-a*o9h7GNSc}DM%hdf
z3P}aispbPNzzvW&5jYe4TgYsjm%tcSJ;*ostnD+CkdIq9%_r&ZVEy(wXY7;Z)J(!p
zqDkQ=RHOVZOP#r2S%!WIriO^iI=ea>APHSS{$*Q`n9k2{#P1QHcx+Xt-o!j!gadcW
z$yVG3&$X=TWfE_jA;&@Wh_WbwtEFg{3rW~qS!g4<(AwSg?ynzBZ*A~j?im%m+p?3u
zR99BU1-O2dh63Q5fV9K|5>XmKf9DQAq=TE4;3rMouuy2yVz3OZ`!n3mZSP}$<?d0i
zCDgqk8iO&qQwqf(1>}cZe6te#&kiR!>b&)DF;p_)YLZnTo5lwpcMF23%ftzNnzrrl
zPoPhlFq*c|i&Fme*IyP8?Ye0|C@%qK2Jrk*vkEYWPXz=@Zs4Mly#_C!SoksPJYf%F
zGS{q_V)!P4DTyGbEO_|?%T1#(m|k9AKb5x%nbJ{8W(_{oYv-Haob;t|aRqcs78bIh
zr6tH>cOwwq_Lo{^#drEddIN&Ta{vHS=HdPnZ^_oMU4g@%C6W&A0$7n!x>(Za@frpM
zTXG@ngsiqur#oPU0>LU~?gwk@Go2Jd07`b2*{lTs2uIKHObV9z2;T%uvICO#EU>1^
zfR=M)eL0Pc(#6hv%Ip<vILIi17+RF#BcN8mm9%ERw@@gTPKta4Rr#5)&1AMqneyl-
zz8A;F0EPPJLfg@Y{>hLYTQC%=@&B-5_y5^I8~Y?M_M2`n8M$Uy0Z#D>EA5iAU>V0W
zSe@L?p!lA}Am;9^@UeMH1xve^eQcoG!2mjXT|&KS<LkVDLS1HNWSi1o-S@_MdRC}`
z=6(;Dsla0PUUA<OsLLG41HXxp0Jl+Lk>=%ZI#CT4N>(p4j|zS-d;Hf=|6tEik1rzz
z{qog+*>iA?h>ziu_;75`|MfFi$bYiue04{~(6^8OyGb~<=TIF!T7yTLy1h@DNjpVW
z7<_T6@m|I;+dZepUI?2`ssk%P@@Jj`|LKf!)x$IREI$G3bDjg=6YIb)T<R70Ah-YF
zDvMmirY~Rj>;Gbu1`y-M(ulgEq9R>VB%4VT6s)QNWeo@%JK7N_vb^-R?AK$DMWcCC
z_4M@mi%jDnAk75p1KGsL1Q!DXLql_O!6C#wwW-Hv<qesFoUS!Myl6;9^s3%p5Ec?@
zk6<_Nf++W3ob2#5yHB&WGz6&!NIgc*<+yCT{}l|rJ04}eM)g^2Qm*a7Gj1+=%cI|r
zgsySN@nA!~K(A9kNJt1NSV2e^MBOV5mJe!4ZoDYI?z%`=rr+XVXhg8E5eQgVM1+Rh
zx9#)OixU$Q?2$<TknKV}Cl;4yn-7u%1g@eh8a;Uc_%XUT3MyNW`}H6C`T3=vCqum@
zIyP}&bH)n`#o`ka^&t<M1&9uS>iDe@Cw0JKY$pQcRa%$;ZwIil{}7_5sp=0WVvKKF
zo#bo&gv3O%iaDPqo5N|ixKzNdy(Vx36GuT#UR2-(Np_k}Yg9iZ#Ep8l@7%#BBvedO
zIB~P<6DNG{C_oG1YQKEtN-WeV4SsmwZ)#|1X({ErbODn$q$B|N535msJU}!wkC9jh
zg>rmRkB9BY?67O$P0S|du)8i<>3FbSSa~2S=tJ85{sJDk5M<4=X>u6>zP`7FY`4eU
z(3qX6G|1}_fE6hyBe84Yux*;`kz+y~t^ETi-H-rAip#zJX?1nAD^3tsOIy32^TQz6
z(5fm&yEng};#acpWdMm^A&*FgJV+@|+ZzJJbSyZv@Y)X1aycs$#(_>v6pNg-7;Ffn
zA>rW#FKD}1Je95jkvqJWnJUZC$B%NF?0*dm|A#yM|7M`BzgZ0m3HeR;Xp;DL1U>M{
zND>_DMe#|s*Q^w7xg0YZZr+2KNOrd)qk;DR7hDl9Zz+h?kb_%W_xPgdv7JV%f%y`R
zs%(#~yD4Pj@&gGw!CLK7<+lAHNJu0F{zI&(;C+0J7_r6%s^oQY{-**`#u_<bDebSz
zod09{lRN*mO<g#Vk6drsYG{$&eEpgZW~)oh-Xb8OGuHAR^Lv8-_`O0L1e+8!0#sSq
zKlOOAS%og-Bpphh8y4(zzx<d!eg4#GoR+6B>vQkj`T|(*Kkb<L{P^iabf;^=^6z9H
z3ElHo)~gzL_(n%nz;R^{|3f6o#jxuHHoa$Y_bJ({rQUwkO6!Rz|3ti7RsL!U(gSR3
zsxNrH^YFzSjj;UjAA(X939kAVPe;@$^h-oyZ#`+dNVA7>erJ9K44WM_W*aBy<Clzk
zmjl>_xYu@ERufC*cs3rv?LYdD+b?SLs3^A|vIu_6|KZ+}OSdd-YhovU&WlBtK!-Pc
z--^$}BU&lQ-dv8K$;5{#I{zP2^p!*o|NQZH&&OVu+dOgd+|+IOe5$6b^VeQctC>rT
z-%p#<r!T%FyD4!r*)i1cjfU{WN&IV^My?v2Gl$E$sWWuq%<p4f%)4|;!>nQ*&1%t0
zS$HgCcq}9wNv{5{^Q6#YP}V&w9I;H+;@gmZ|MGV34BgQ-dc#3&Zk`El3q5_Gj)7Ge
zl7o|}mC>oc%BCY8@(Q@}Ny@iqIo+eb9Pd!{^qhzAd^Mhawv&Cg1w=w9-q@EHQ7HX8
zFv8u|5rq`NU5Nc)M@mBIdvG;0ZuXZ-#m1ZH+ZTDz8<^b8h(-)E1f4vQJarhG;rA6v
z4kHOH)K%M`k>^1<^8L65n@d11R1)^R7$=&E<KP(@-j;Ate9JmGEW*>NsVv-km^vyW
zzeGRRP<-S3tkj4aS37E(0ACq}l2BG;KNGk}@vjtMv@|B-HDycL?Jc#DbG@^x2U@D^
z!DEmBzQL81&j}QhH=JwGG=`Pug{{6dJ~%W}_G4sj>$*CtM)KTNFnU78RY#ZO#P%mj
z`~k=&?5`8Ryh<OX3oXj~<&I9vXgt>vM<+#?A!}Ed4zkb4bfmY@7pMHC8An~dm>D6O
z<TSV`Vm#ZJO>Iiq@;+pA=fq6f-Ank<A=~;>t;WrG$$u^cwua{kOhCG%yq%n&?rO~r
z5vjH`9D1SD^-~!~samDgwJEU(rp(#-&XBSn@}dMtX!13^m3g)x3q4iZx4SpRITS`O
zaWpxXTH7p6F2}BV7^!LGbkg5~WDqylgBDdJcQk(LtbO@(M3EA$-JbwKkIcB;`SV|U
z%kNR8q^X{HU-oWEY=6GMzk`Pt6+k^(b)Y5MIsqAAdc?^IGAcQH3z@H#?zaLIlyt3U
z5`x#eq{XS|>BRxz>jhXk87jQR5f}t?e22@E48KhSKLE^}piHR#(eR%<DQxq@5~8YH
zAsqF;WDKv=rS9F%(&t)Z%nDqi8ws0QzgxPW^BdxI*}d_GaE4w1qac!JwRIzFXaADt
z#cFKrt@NIkoI_t)2ur@m$=qwh6fER7DLS3zOfsH2NS3)7&9-GAYf8##+Te#)$We`2
zeCN@5Xgw-l_{*Z_{kO=U;cOmVO}LS~3}eDJzj4)ue$e-NQP#$%nt$)!a!z{6CEj`^
zEfe2l>b{doXLruOqY)&v0&dI(;^O#SFSUOp;k-&lrmnXOAewGhvSdSDT^mcUN?hTH
zqcYGydZ9?t4J?SXs*@+hRCh+s#i*bU?Nqw0g_;`6f||oG<E8wP)-o<P7{WJ+YAuiq
zPhvhfL$G5v%(sh^vB+Q>9`EI3vSl6G;>awsKPI92t)|;2Zv-vg<u)>5i>@s3=)}rw
z@l5?P*F+W<6{XygXjL03O2FklBz@tqFhUSh&^*=q+{j00hy9l!yEk!9@0cbfO-5eo
z?Wm*`YmPx#=js9WMi285V#68&5h4~~tEfW))eShdm5>w!D-fALG3`JIf)A&J89^_C
zSSa(%#`*B%;&KA<tQ#D2b0|}72KpL4J{2pg9Hd|#?(R7+nI2c=;3WmDwo&@UXW-&W
zR)1(Sc!YO;Z}i!UkPl|-#J3dIsh6VWF_de?yCX%bOz4iM^|l+HH<HFx9f}{&uk~1{
zS(Td)?94`8Ri5czo!psk@xPiWb6sAxF;joI#M0h?GwV0TbonVprF8B(l2{i8s}5Q5
zp+z4=L@q629-9B~SS!6PKVnl*PQRAV{%FVYz~~k3(2nsJ@|hi9Zsvs{!@|~#er62?
zFhQGk6=C;lRgv5@$4^SsW?*e!YTz)3*JAyB=SIn|`4B?~>kk#%V=ufn%{@czP2l7A
z>a#~JGh@Qkly%f-qKt#n9+sOP>eXkZ;Zcg|La9ECLDSVi0ctHs83Jnb9FaPZLx=G=
zlpN>K^?np`?|=I-2uyC9W@X-K3ygf2P$!*qNLfDrlOZq?tNbzl&G}QxE(E*zAN(@5
ziyG`n&+KRXD%O=8>Se?^B-Zh|Eg~)xHrd=-Nb}iMgNc64R#Qy~iZP!W!n&>M)+oi_
z(~KXO;G3p?pRi|9%@UQW*&CV7{lY+c-AknLd7!Cm!TjuGhL|c>Mav}Z1)onjeK(?X
zFTAfRuO45aGt+%4JLGeJBuwY<$_+)0?O1hTW}{&xZW~QDx3vTRB_~UcgO$&JxwS4P
zH}5#ymETe?D5g)FLBH@{Z#p;9*SoiVuFqKAjYUysCo0^==K3DDx>45H?7Ejv*OJ<P
z!}xx#y+IV|;g;ndE_!A?{1IkEHZqcjVbXB-K{(t{E7x60AJU8!47<$oK(Si*gmqQt
zwt<<EY?+*XiK<S`OvawLNYoLE*p(_vlcAU=kdl(JIR=?@mRyEX0(!1THjMK2?Z+l%
zE3N|GL|SB<NKq64Z6O{miGn&uV8H&DJDwgM2ns35QUa5{Fx@>cBzg)JEwpPJ*`jZF
zaG2iBY$dlpQdB2Rp|sk5%DDC{s)BJptX4*9sW7^AY0%QKfMmgj(@FTZ1PcAGaFKVe
zYmSzQkrBvl28SZT!mHMVa%k35D7iH`yr{6r7Cwg&R+P+-hC$hIyVg}U(VfFSEY#~}
z<*n$T`vrg~0gE$hoxIBV)VY99qh<Vwz>V`s3`R;Q519+^Wl`Hz^~5;4s;`BR>s70t
z+e<LQz!B@(YZpHtdRLi)yH9`lqcg9jLIsPs55^#T)kkLfQ=UidNr!LnHTfXr&OybT
zyi~vC^LC4;PfG6i6i*3mzP82AYNTr0E>=|Y));>8$dM)M)mS!pf2x?cm&R#2gSVm}
zR$%RHDSXSVRmU=4p17cRD}6gL1yAR9+iXkk9SX|sr5t>}c;$2Ca~s|4-I+@%#HBM~
z#<UylTf8&$o)uO!HDSrtYNflRg|mA@krVP|tokp9-*=kSt^HOqvx<Z>Cw<7IMy01U
zkI!D}w!$oEP2>}}A}>=hq8~Qiu&29iv3t|St<OG+#(=0YsaL|NFpW#hWXMU$OrGrD
zi3$<0BjY<GA7v`!JcgVL%^?oBj9j`xEYwZ<+-TJ+24iP_gdU|lGkT4Tg1`vHSgR^j
zyJD?X`=n*tChe2{YYr=4)&a^}@A)Ni(JrQH{EM9r;ZTKcxVDepWeM)!UO7>7ot?v)
z3H)n~GIf#oE`8*hDpsR%_>)@0VNC}9R|m2yKF_*v=AWZmx@z*l)1Xj2_^f{a{mu0q
zL(`@SlCq7t4q=KUVbw_%9rmn}qK4q1QfZOEu7UaXvhEAt$Cblv7HY@5$Seya2adK$
z<YK@|p$>5>hdLA66!k(i?WXxeLY9-+cBUBxU8Z{#5vz=f3SQDqNn*>qLe?)Eqbjlf
z-i>DGmv*&1at@WHN@738$~g;UI<+~uD{!dQ3jI>cacCCb&Gsmo_G#$kmlslRUSsgJ
zotz}y-Rb=+IFhBrU!B^x_nW^4X(VTqn0h4Bma?ds$_{CWse^boZ<;Up)@mg_mv?2V
zL1Uyra^iq&_l_P#W0=@;$kvg{(6IG5Hq_!ofhXP59w&&fdxbinB1OXCS+>@~85kOx
zq7!8!fbe0VX0-u)-=h1^QE%{5?S?`uN|<swa=*@ow39U)vWZD02s#-ZZdna{eHA<&
zc@S<B+UsExU4-FsKp!U9PJLP1*smi+{x8_A5J|g#*>=r7C>}t0mf3qSx2Z~BIM+&u
zuGiHTvaa>4KeX<IltRfN+HT5Vhxj~o1*U|9HIhyx;JQM(VXnquyQQp7QVx!zW#GZu
zcJUkn-8KwH7<>yGlJ7>3HiS-ja`sC7aF7U}>q-5ogIQEj{TW-FxzH}0vJl?=P*7$t
zvq-nq<19=WzRq5A<Tf!|`qBfgTdb)a;WzYiqq$`}w<A@R7ju6NZryFf!bX@0o#MP1
zZlt!)fW9DMY54`py4`>s2z5aI9T>i+{Hu8M?Zx%Y0h2kvhGn4+2RP|%gbfDV@B(Ub
zR5CeDQcS~^B(YGj%a4Wzs%2G)6Y2X-<A1#*A$v-9cstru)}wL7yk@=Q#-x^-=M-(F
z^CjFosse{RYLnd7-_2{qtm2Pk|MsfZ_8AY^PWp0A8OXZ4mQUX`MJgP;Y60m_uJ3Jh
zw5)7uKzzNswbgR*_Mc=k$OJu{Q(9P5f|ELs=`kgO?j4(;q8(I9N}BqNpVRwlr{znU
zRrp=mYbk~tLr<!QW)eEZZ^{0iv6KxSXWid0EA+6ATGSDD>m_&8nF>rQbHH@DQkX<8
z2ZxoND*m!GQ}Foa_feDiQ->GdYmWL5@YD^PIGMbbjUwyg@7>g$*s#{GzP{v4FF&-G
zpGMSL)->mBV^AH=ARpM9+Grz7TMDcGV|;GGygGQNQx$(iYAkd0E<tA0$A?3&;>_^k
z8m0e=Smf2&L+|u(B{&rQh)uq%?XSYUt>ZsFK3cHjXoVgoqvA<f)28z$Zw_zk9e!2a
zZ8KKC^Gjbzymw}XRq=|QtP1T^lQqN!F_8b<tIu(l+95h*&2>BAX^5Or-*pgY(YvZR
zx3IeDn5|!wG>1m>Sljx{FRUfqwpzPcxrlmy)r#{%UM^G43kgi6DGTAhAHt3)Gpt4V
zx=&VR3$~NlmJy*ey<XNw+RteR?Few+`1UP%(%w?jfzoimcvIM8*^G<czuTOIjbsdK
zLZJEqf$+SGkJ<QCrq#WoqC8buQ&uCJO2Vl0PU*~-`LilkR_q@SJoYVb;9*Sc=;zx8
zdF}%m5u8LW!(N_t$i8J!T2@Y$PP8~<oAC>7_b#L-%fvctmxqn%C)>Wo`meAmj)X!!
z)NNY$)j`9pNa;Y=QeHI7OrgHdXRii(Zf-fpKQNKjBzMePuu2tQxvW^s$y_#Gse&a-
zs`AFfnc(g%)!9JGHVG@m9;l33SE)xnQ|xxeLv442x+rp_WAk*Qe_3lM`3%dcMhgu8
zPL+PH^Ym=JT_n4k*|`p#O&W+H)L0J8e%lP1;JhUUG?iy+_Uz~V#)g3(XVW-dpG2vj
z4;VfeQ+1X*Z&gTD8`?7epiP>^!LN!qlD*S;T`yMU^`mA<UDT+w^QzWF{ag*|tNOlb
zh>OV?sas`t9q<)b8Y=pEgd{sp3K<1yO>msLa^`y1X$P1FD)EZ+vLPGM2bOow-n5pH
z4=I?j(AZoJI87>za~tk@nx=b0k6hMq?|mB1DejqfDPMT|Gv=E@apJ|}M&E`3^87Vv
zJNG1N>$H@{wF><6F3YQlH=yPfAljm?y1gYCsIICY>*l=3<ty1qwlF;1#D10IlVy^o
zF&16&g)lL=fr@n0SMICYV<z1!6Gf@YB0e0KERs^h(UC7X#xr>CpUfD!lvX@&C=!E~
z@hEWyuqXOgBNl3qU*nqd*C8J7^x1dykz`wt4_FZ_HpAyrMlR9VbR3FgUso9=VUC13
zS>*Dl55`BmuhASxCLYvl(9n{OUM1U_5h>)jpI3%^UVTgB%Dn?^g@n5Yl`>Y(XZ63G
z83xShP3wx*8B}-N1J9E0vxz+22Sr<GvTM$5Ca$vjN>{PIyuDcF`S{d+RpdvVQB@4j
zS$oA3d>i+gK4&fqNJqceH6?qdFMah_$XC^?M5E?_Ap~xYUE9+Ux4R>-aHfG_6doc~
zP~#f<gs8(T&aRU)eH$UUeS<#?Msnj6D%q~#V%|jgp7blenAqyCCGS2Kk6na;nB+DD
zyu!v|v!YBc^sq6XCiZx|?DENgV{FNCCbp?z_=6-9d>u6^%Dj<6Jbid(_?nc9E1zC4
zX=?hSH$l|#H{);<OPoJnE~9#dg*~y*(Q1FQtW1@|$J~aDq;;FD76wO6+{5~&{P?ra
z+5UucXE~x)ct0XY49eu9`{Q?%3ms$J7F^>_$dyp2;cI`nARoiAy|elE$FSUah;)NV
z6{f9g4j&@rUQ`(c`O(GJuskC-J6{0y)107MvedTXb!d7C`G7V5Gv>$`RSx7ms^SWa
z>25i5^uzXJ_{%-PQj4RX%4p)Tu>V@<IJ%$Z#7}x0NAFN&KL4SgI{M`QtgQ0}C8?x6
zf7sGM^?F-H{!jJiCJHI*j_Szw@%luUs-4nyNQBzBqgR9ZpUV4k-k(}I&egz&ms*dD
zD|0eA$c{x`jxxuEl?WEUdt^Zs76y@e0+0SEzj*tzr#?ZYSM?*?pMo!-J^J7L9`l1N
zO90XbL_!4G14ziapctk!!3A~Q9Hh`N0;M<Orw@;=&_#I8CqRbRH-7(4rsoys_$d5~
zgxt3IX6EJv_B&>XMhD0^Z9t<!oUlCmh*9BCA}~LHVv3<cd|a4C@kL&_PBPPD0Tm{-
zkKc(3=nz_>C~#nuDjmy#WOzr&cIC%)bhoVh48n~B4xMVt08}LalkZ|Zs?eck(}#`1
z?g97)epPHwFI!0XgEvAysKCh(&>ylu>aU1YsvKZm;26n4UJ2w+`H8ZN*X~QKO*SJ4
zLo*aW5l*(zp9iCYN*9#deJx6oB_wyIASS}X<uo^Zk-CA04cabX^I56rmfPxswgC`!
zm6k1_pQwS-K@#q7<6vuKu3~So6VQ!y%-&LIL{fbG6AFsT0#zz%>Qtj`)aCWROCJ>o
zpodhEfyh9@;rbkI{BNLonrl3_?KMzWKy9Na+foKGAyJk!kkU4tqJgR?{QP+rqNsvc
zCIvtTv8npHItg1_TkqZrsMrt3mDkgw2K=RWXTi1t5#rwAvz%XX;%vdmG~$c_=8C#S
zdyAzkj{DoxI>x=3kx<PA#^}<6eIVo@1iR!z7hvCs_yH?(Cb#VYR7{|pG=wQmPJRXj
z9pkK1sK;0S6kNWNRaF}b8;vxLXmoirKIRAN)URK^QdRjNivUSOI#9!q3WHyKE6?6@
zD2Z(Ta^?j2426)pdnVaApgqaFinl8_J&6*#^RF_D{t)3SsJ$}->3Ya=zz7)L9rEf9
z37|^S|6S+1$tDJ326fQE$$><NT(37%6(RT_btpu!M&`I3TLY`<Nmuf0s05v6^z3N6
zeiheO)%_$^$^F@Uss-V2TfnZ;`}*dDEvPCM#>Jz!Qb0%dbA%e0tFfR4k&#z1b>XmD
zOb1<xGTY$mcR_Tj`hT`QjCIg+K?vCZ!;X7K!Q+TLPt(zkSQkh(kVb3$NIkL<e5jn?
z?nl-Lg>wJfS{zj`;g<eb3myrjO=Txp2oG_hPoU1<Jv!fI_W>0?%e$v6_!FX#-ZzTT
zRPYP3xRa=_oJWNfYsusCg(RTLH(aOiuNeR!c$FB1dgO*&=S}b*o{Ww+d4n5}NQ9ZN
zpgww3_8g^vYe8kil709Y{3j<T=DhS425vkaSm<w_2XK1$5QFhm<e21vdQ<Y|c<l9K
z%0{1xSD}zFR9BGGb_u}StRtSxD8--S6EFP3b+r{R@vEehsaH>+_=b)uD_=d4is?K<
zdBEklpIo<Zu?pwau6H1nV-K*)OUYA9$a=jUg5Lmm{Rp{PUa9o$>CR>fYrhky$MQeq
ziFGEf>b|HgirG;+szr~ULN58mWD;0Ws1r94OrB@(<sp~jqAgNVF;Wng;F=LYO49JM
zCqG8qlTSURz^MlJ@Nb$?&(Q?E@MR)<m$6NVdIOLx9L<UOfmnM5z|zr}y>~}5a0UU7
zm-Eu*qib7LvrgpkT=X~$J&97d17RR!6LhuzkyNN1O=3*Y<1q#W6yL%GnW^fl<a!Hd
z4`d`h+V>5b*{C=Do<1K_9dF|%UX(dnfBjoBr0>o^u?VIXX4o$+SfQVsq9)4mjG3wc
zvM4dv9#0&N2l+ix4)UYMCvM<P!WNOV4oQ=GGIV^=;}5-I`S*lxmw!6ua_9kc&HGe4
z(iVdQa}q;G<8+DjPw6SL!X|yr=i5nCIuqC@I&mQh*dnVwS7~^5*SDUiI{C%kKnn10
zl<qlQ`h!FNaD8EW=^E-q#`1%3DJUgb5vV;vru&T&mP_r66R6kMp+<p1Dct-?29Je=
zDDcz$znYM%jM70|cJ%uHm-^rZOp{ttyXV(cUON-98YTnWT*~6ZAb=E3|H8vV<RDPi
zRzNuY_+!U6$eR=2_8R1^2#-FD%O-bVjSEP>wqq{WnRQxjAgpR{B9@L%xA*~8Fns#Q
z?%>|1VvbKrnnNfiHdB#PJ?k|D2MZ=no$80X@~U<V1+zUF;luW8qCnpjM>vkhG$TGf
zDq5yZC)nu^?a#h-SKrmGtFM<-Q;Pyv)R62Ns6xR)8-RDCo1&qj0&R?MH5n>K|CqW%
znI2HgF8?lCRqN;w0COr7DQ+@Qf92el(yeQ3^d~=}4Os(iXns5b@g9bT=?C;wRP|}5
zC;1-SJT}D|f&UO>P_f>eF1pV==yZ4h94}dn(MZKmaE?9*m_Wake+a^k)}*;}Z%Y2S
zq4)TIbI%0+I-+<a5T1D}%q@Rsv0~5In~b-jA#?zIa9rSBa`I<>I5!@~K#hEF8K}O=
zPzR@*@<cqut5@DV;MqKmJ==SHgqWf{T~cj#hpI4J0ZatqfUKNPsQu1uyL|n6Qc=;v
ztaN5}cEt-9E+94d@@xqpM*+BLvp$J%&rs)I?FbIqrA3u?o<KWOI4}dFS^f{t2zUsA
zk&zi7OF96B-KHMm13=Uw;s*f`qE8kibC5xN3{uVqDrzb)8xRQXc-7vF{16mW<RFo_
zUS$9I8y>Qk5J+cl3Yf*hemsZ!=BoB9Km6js)g)(UxA8PXnI{eE0{Y<Ek*YqzD{pFg
z{`9G-GvYzc;{bDn`k#Uu>}4XraYKLr;>8e%Y4E({_sYw3dhMHb-prUAnH=mB5Pu_r
z5=6P0bNj~8L!bZ0z~;V$)y~955b}x3)c;{B)-4|!1|Z=Hl!8^8E_8_oL3)F{idg@<
zU9d2}5cP`h!7)z~#dDcddmKOL{DnWCNBuuEA2=HHe-=EW{!15w|8=1KnEa-nhmnEO
zqsxTIU*D8?J$~l{q|$$gt&cVa%0NQ*Up&<RjIF6yk4B&6!ouc}`meDyP4m5@zoCg-
z&XUzPP)PtGpmF8jynw4mf4vshHhO1AP~|$IJ|n`gLZO^Lz`cAw+Ash9`G4nJ;S(Z~
zy34*`VLMzCo=GLV<o^?zn#+0|ef{P#{5=*Hiv@T}6yG^meV%>r?f^%NgqhK|kMe@A
zL_k~OOD61Rkp#TrUnB~LlH$|ij7OuqJ>hbe_xusk#pL9MY(mtV38>^A{qg^d(}v#7
zz^fq;k^-9)eFxabG_4L3#cir!nqoihBMcp5OnRq_E6|*uL6OXF|HcK>b}Pt=0R5k9
zU;hgN%OgHb_k?)(N`3$_AH;uP8hNQWIFtZ<Uf2Ph8x$YnA@HTTd$-2(%2~S26MK9w
z$)2I@$Otqoe%D<=IudATEEc}MV`WRW*WfTuvz3!IC_hxOipE$_CceNE9v~JTpw21(
z){+W*>1f^a7cX{#{y}woNAu;&$G={sVC}texS4|+Us8(8wG&-r_FHU3L%rC?VS;Bf
za4)rA9#a?H{o|_FV?8^5J!Oh7Y(KkM|A8<a`sqIN69EAbor*$(1Psm(lb`42Pq`x_
zbA78Y`}eI~|FB1WS#vc}qJLPTU(S62+fvTRBy#c6Uq{n^I_Q)A5fc}()^mvOh1Y+a
zms^TQb>yxuM#@j0K1~6w5X{)B)iWM{U*FgYhdpYTIB2JUSSic-pJPke`?H|2`2>v@
z^7T6RNj$_6MW$wrmt>xHa~HV1NXRgDC&NwlNrGAtb?vuvWHLaM1d1mjx|#!7z?Gq%
zXNWQ#i2I44HHB6odQg_b8TV$eosM-22nk6B+LbPl!FGWee}I=BUF<LI>FLR$)o>Mx
zz0VPKhH=r@S(1mZzlvYkYn|`Q>&{SS+p{Z{mX_XtuME`YMMs!#lKiOHZ3diME3fYc
z2JEmQ+0vW?zV6EQWAql@xAn+&&}IqO_QKU00;v4n&TW^NM_+wjE9dv|2K#3yU!;H%
z2x)79@Wv4J4NnD~Nf!Q=<(V^Qh#*x3i5AkbLp0){WF=-dtr6|25~~5CDx}s30&GNa
zgy?smRxJq<%}V|?cVJ^7mqL0c{GH??X~IWR?%Acb{g=>_0O6Q*!LvBI{gMVjN{7pJ
zwPfIq&4H4y2NKVXCZ^V8A^*D=oMs&mD<$;h>C7!T9hSm_Br4`<&w$XZ3pfDNFxujO
zkerjoqe1apzFc)Nj7jGsjFSYlIXy%&O-HwB8jqP06Dx-RBi~|13LJ`^Qh|P;<L(S4
zq7HKO>gR3Y*)!x``Ht70t?{m6XOdVIwCaCYdDn#IDRopF1aq=(AS$RMYG8(1c|T7J
z*&cP<ua&!2i<EC?I5pwXaB3|<i{cj#4MK#ZUX!R>E&9Km`6=9yQ8=7B(HJ}j|960c
zwD^H9k1TyZ^c{&WDKD2B9PlRPR#N_A*be<%HbC?_15bzm11ecv??Zs?5DJ=}pS;0A
zL1D0$!hqSwD=@*kv$G@Mrw9*H$-hzzrXmCq>M*!WIHECu9tNEww|1HwwuW@TWsXAT
zIt%?5SPKWuJgc)o`#J1I%H1!1<;-0n?Ui3YHd`U_ynE}`H?EB?=}iqcTryTTj+3y@
zSNPvMPc%i?ZZB&filQ~>EYX+L_r)mO_f=1(>Ri@?Qv&{XMVny67DuzHZmX+%4U*8a
zVg|_Pou-(d^3K<M3T3T$ugs4S!l6k*lr&T{G#5+8m4_UpvPlMjxe*E)!8Lel_Df6M
zX~8z5&gb6h+ZmXd6r$(k?$oCv+p-yC`Gyc)^ud&MZoKcp!=@CAXoJJWY`y$qFB?Rj
zNF)j(w++o#7-<ECdq=5f^qrM*n7L_qaLb2ekAEPtT4*&gjr7;Cw=c8ZoKNXhvuhsF
z*y%cRT_pxooYU};PvkIZR!Zfz1xR}qUcqZW&CJYz3OMz8=zC2KA~d`eE=a?{lw|xf
z`w>X+Eq)$l+R#OT-_{94iL)RQ>W0!(U|3kgpD9VrC3D)B1mT=aspSGQG}c*wXv*(!
zQ#R(uxM=r~_~|vP8X$@?gD`Q(tdSO?#z$86X{F*=Uo`EYxd03b`jiF@YmnUm-F`M^
zVAx}@?+cXvQ$hAW+nMC!5C+;vn)~;k!(Et{4B8H3b7t@FueE|*m4yiQhHD}8wt<@1
zmb45;8YDGv^k<;8i3}hE2P7ZG%K_^^c(DAXX%ef^hx>@$rw7>J8;6*Y!%jGz8$cdc
zguEY7LRVI=x$iAis)Hp(nvj5sN>)iJ6-g8zOxv{5JwXdp9lwal_Epo^21T*v#ztw7
z4#79)9~dPsW`V__VquX5H9E&!dR{(=l@sFQ?{RX5AILnC9g&QX{<sVJW++($m>~wI
zO%@c1hy)o#YkOMiXOy^x<%SNPJaE#v52Y&Ds68N4qBt{Bpm@BWrl)lgy%CsTNdQQ?
zpx{*wWkG~;3)+gy76Zz@oC86?>)hV9c3S~)qBJz^P&IaXhEFuA3x=z|%qEY2;G0A?
zsQ-o_838GyO+u<x@;w?Fq(j2`_PCFTAF^@ab_`k~*l$VeZX?8cP_|D4^w)#*Gl9Su
z9#RyR`%<U_%D_cmY~u-t@dF@cuBv*BwBkWH_Rx}LpG1ytjI)6WKl(AOTR8IE5HnB1
z0wjSexxBh!t$c4pe82T@-v`->;IlhMVy&Ov8>U3sLX}JeDI*jG7dJO()}n^A2Y^mH
zqf2eG@t^cL+CjGpf03U(*nk!Z!q1*9!_qULXJs7L+XZ)FYas1u_Bd}uBPMTkZ4KHx
z(ZfG+M6U_Awhp@^8T8Sxfi|o@s0D+=3JAFBrUA`ha(IUcu`sK_mD7OU^&*XKhHa*1
zaGkc_A+2n*UcK7p%zq;nMN=R5QDveFhz5}lO&`NTS%&b)Y{|n`KtwK3gQNiw2Eyn9
zRY9v`YD~#e@-<qe?5xx_@Hmm?E$qM~*5Qjl3`losY4rJ?B`?sBPeYi~4R%zY?3!V1
zVaZ<n3$8DaQ1v3W3^oX9XoP@a_5PF~T`P?l95!1tRDS&UaY8U<1P59e{2E;$FTmA^
z`WiHvP{r3E1CT5c&i*|XmTn*4osSFxu<^Qo{qWp?c!!Fey$|f@Ac7?#JC~Dl0MnZ!
zPKFq4NQZz{u?5p(1qrU4?U@l=-u)O@!ELJBBb=7=Ayrj3q+fts8188Qs^rio=r;t@
z;l`zOoNBo|C(TGMi2&(_38E3#a9a48!*92>h+ThxP6Tm0sPwa=sgoBuQ^lis=F_P6
z!KwB`FBZtXPyH&x%C}BOPC~Y$PE#bOJkkvcx=Aclzg^gtvx>f&4o)7n1j45Vf$THz
zj_#S|oBc7vYF1xp;rt;BVqG#0at>*rV!_s1vi~+bUC;Olu?2S-89SQ7*;)-jX!<;i
zxt`g0@aOM|i5*Vn=n%;2-Mzh8Xyh}Qi;IglE5U%n?<=ti3dLS&X@!D$3I$uxv=1jm
z``g<yiGXEmWA=9te+~eG4tQQIp55s?(3VDX*y+##v?k`}OXXX;Fa{?yxY_|Mycq~?
z2IQe25$-{BB-)_DtRc`uTB{(PN7f+P3R$W=9998n;4Z8WB7JT3-HYO6-+(W(0m{IN
zG`fKfPbSX+V-*q=-O%j9m;zQD4&g~?M8u6Xz@WpyhK)dbkUYaaM$JomDbzZS;Nu<5
z!S;s-(#VG)AA`bpC>YjoV7!NePFmAzX$vvgMvH^urw6PM%DUcANkK6*Wux}Km^ePy
za^%t=JzawkEgVtcXhWxmL|D^Aa2u1*IHV@z*x^GPqOSG%zRq;z!j^Q1CXg;6UPLSn
z6CC`RZ7;kS=qt8ABV!{XV2CW3K=D&h$-hb{Bz8F25;+Ia4|Nw7HZ-+?fTk1B0i-_;
z68|AxMV#DzOM`Kqgl-Sa$YcQYf#BeiAV)@eA;3y3A$fL^&C9_e)0&XZtVg!c6h{>N
z#B-3K6$A3y1fj$}aej1>zqAh-ukm8hI8JB05Z;lfcL?1ZIyk6ucxHhT2__KjFmjWi
z9+=E+q|3Wb=%v9cJJGasZ?oG(z>7cwcI%OZwmUVDkm2R~KF|j^lw}>w<3+^v4OrDc
zdx4ka>_e_EI4TwgDbEb#sDfx8Q0qu!U;rqz!7xvP>~(VmZDW^FRJaFibDEMR<C>tG
zlorIOO`rv6g7iowvs2rPL5|DpI2XGt1qIck*V3Q7`~Q{XC<ceAMi)>B1k{F51R(kY
z5<e|?7j$1yDoqQ{i#X^MzL10HX*B8>HSAK+)=q@YlLQ+o6_!AMxU86>5gL}10r0Of
zN48tw8y=np5`%4{1Hc%y5rA334InEKXi-%H^bE*7q(NK`0Rh&<wf$QSU_R)dKv}2E
zI}s1q%#xsH>y5^&QzG#MYq%&BK?2DG0|TkRwAR<xBauBKZ2?V6-xLk>^??nzFwTjd
zjfd_5!#s|lVz(XlW?P2?Ri~YYG&h`R2vjOIOMr($nk_}Ma&kg{JJC|BrJ!|XX%{eW
z5>!-F9?)k5?6)jvRUjZ11Oc=ZD7m{Tot!dMigOzl(tcXJ1Ud=av@M)DG^oIn!3ZF7
zL?mK-&<g6ZF6cwVV)#=6>Gb8eGZ8G@cmNF|y1)k^`5DX<^ylh@0DxQwu9*!|`<VwL
ze=8UvQtc*zwC`q=F3eerEl57PD;(^QUOGro4B`<AA!y%*h?3y}<6tY#BH=XBDGq=y
z2#!;rp$O7`076-AB-ujxwL!3rBxTUGLJ?FPK1j9!I-F)OtbRzd8K42yWE8<rM>)tw
zuev;Bg22tqLMkrcyO9n9&^1YfR;^3{<R{(WE4p$t1q%yzkRAXC5eK;q<UMTLeDf(W
z#MQ$0I$=cuAqR!tCNxY;(jZ5|S#yH<AmX%4SF;{FgMH(n1cZ!8uLW`;_^T}h#{ocb
zQZ^}<^}J5sBSK#u505YnC785!X#jBElLLw|n+}@!7D!!dFvUehKaplA(2lG|qXc5i
zFlbXztkM!dCEbM0Hu&O@3t9o0FBJ50Jpa?s3$!nr8c1_O-&YX#dNI_1cj<k`2+nR3
zX;7gF^V<x-DoK!4pH^@v5D1IBg!~lCtaBJ@LcnI;sV~Fq&TXzhHpAf81oH?ZAcjo^
zYLqyQ1YiNcd!QAjAl+Mh&<xL^1qK58!7TJ^K}&#UBwvBp=LPf#8f6rK(?E#e;+LO5
zoh=1C4blk0iog*N<_$2Om>uY5(!q;a!A15z^h4@}2SoHdA}>FD_#jbeJTxE81|TG~
zKtuuxoKEnl$&f^n(w6=*!^6X)5P}>EiD8lDhU~CKW4EhI1`z}CI>Ca%hRRz1sp-?_
z1_5|#QIP`DRSEdFh$tRuC9w<2ujDIt1g1f7RW8VI5_NeNIlz~nAkr?Oj>`Y5lCyhh
zGLGW-$V+1e(M@6n4UxRiWfzLhLJE_>c2Qa)GYN(Eq*Bx(@VSX15T|7?%ms-kGmEx}
zOq(xYVf2EEnF))zi(LBngvMwV3w`<={R_GsjIrNyp7Z;j^F8OY9?$A?&U~n@B&%99
zsS4PW?ULy<CQ3oxd;u`FF;DvHN6eQ5@aY*tZDn3RA<DaBwj$pYes1-9T2zn<DKAR}
ziLn-m>OZ?zuFDRnrfES=JJZdy>2u>RVQxomxK`$#B4K(Rw)%Dx&O*cs+4Xi)TDF<(
ze(gY&8owv?zVJO(tBJ0Q9w-k*q#5%Yj7dx0h-*I(i;zcTE=Lz#SWsYMJUri09OSm?
zJryJV^Bb3`)o7M<0v{WwoOpdBIgT7uNFYHj-uSz{tr<33IMbSB+Sa^zDm>oHleXWB
zwE1QFTN=5@VHI<csqU$P3*Y388RpB0`8&t&qLo<OI;PJC%IU1AdBWci42|@Kd4x@F
z_er`3(&ZA`DHNY+5nmjxFXR#Eln68eW+rE{po`YBplzhj4Zg?sBrp<ct22bUlAJYU
z$I5(xG^&twX!CfwMeDq|>ywPx(?!tVrZ$|w5y!92ubUe>-CCG<w2qnWjLpB7g#-B$
z&}WvsihoLgpJ>@gTP`l8>r%&7)1g~ev0Z<W+4hl;g`>#J1zTwCqvPlw2_6)ygGoKR
z4E+Pa$Snx(Dd`Wp9)Mlxrclj*Fey(9PEdg2xtId3oS+tJlwels#{?F}#&6k`h}tjT
zA&n@g1t-Fx=Br9#)ca-~)3Ua^93B+tR6?&Zma@k&V5kCipW_pO6HXj78fArBK(o!(
zom@|pd_APLP)`SP598u&!+ln>9@pmVNUhl1X#2cn!?5Cd{XbKx-AHLkFI*hmKAe!N
Ox#URcVSmw?YySY#kz$+x

literal 23226
zcmd43by$>b*Ec#Ub|Ruk8XyV^(j5v2NOv<T(lNx)X(C7{DIndAgfxf>NXGyp4I(v^
zbi=nU@P6Lsd*A2V$MOBM_r8w9`(|dYIIr_u=gQx2dCAL)pFTl;0);}ImXr`xM4^bt
zQK$pD$Bw`^jCWj~!~b}&V(M5Wi~-jEv9&%*<}udN41+Z@esaZ5-`d6)W5LG6#l&&@
ziV+rTX~WCRZ2qqkm@w9c%=1xE8gP^2mJ%8^DAd`<$iD+ALdnJ`lqS8T=sjhJk2C#_
z4$7n3ySsSWcSLuOYu!J1<k(wb(e4+-=Pq67P;5&Rj=jyK{8H6hb<~v9@@-L<=HXDe
zdxt~o#pTU>ubjB(<?$#;Czg2a(fxOHkEJ|*?`&6LYEFeo8U(OyE?VgnXElc9m9?<0
z_KgdSS70JoYL3BxP^d2l6`q~h`wR8*?D^CC-?7rd`@`X(FTuX|2dd$n(f%JHSKB;d
zSTF<p-DaHIYa{-hvWo?m*$jRd<gJ@K7kskl=hbnUuQusOd2hxUv9Pdk@Vt1=M>Znn
z&GE_cX!kSNn@m=xqls_VK3l-M_X=C`ZrZ4Yw#1T-VrWA=0}94tnVy$RCuH2Wg$G|9
z>qu29nM{hn{u#O_u-a$iwAjSenIsb->`8i|zP=ti=(?J!-w<Nb8uM&pWhNc|Jj{C=
z_4q5zD5fZMJ_uK3yiDi}XVH6ohDMNbLjK&fujLsh=&Cfc0*r*^H~Z(@l1b=g<t<fm
zZN_D1dJ3fA5&k!%Kg6`fi>FHl-gtI~@um+2mu!pRX2t_w@<bTFZd+Sh$ym5)zDZ}A
zu;=S5VwXp#Ym*q<gDQu0Ym{W2Ythl)I@8rm#~Q<E$@^mXGZ^|B(&W!)j52r~Au5=d
zFsiPuKD?}_x1*}8Oe<;9S7i0*(0u&YulMihl;rm~c4g_L+YdO3cXxL?5mJ-cAAcos
zUFkM7!6o_5mapmxiFwP5{OgIan()NUBhlZ6zBr~ZC?n4yz;j+-VYlexE->%?T)&jp
zYut^V`}s*^q{deX{sn*dkTO3~JGC}H!W7#7**wCcV%>m_mNrEr&!A)~t5jK2lfXVV
zQ0{EoaHlXu?CnKwU*FF4@#t*#tq~d~{CM4G`<b3ZzPU1`NM2jT7e`KaWN5JUV{tOn
z1vcq@H)d;T+=H47Y&OSd1SQ(jRMI<=<!IoIQ(ZS#oYqG}O?@V(XZoC`GEKU36QAhm
z1xH0?d6Cc)e%A&1IaaKXQRuiXcRrMmXyqCI7|)nhbRX@y9Nb{rd10~0YJ}Z#NLaF6
zhO3`8qA5v=xm-B_`}+s6Oc<*eysC<sSw=u$pzUa|n#s4<r~O6?&3lvjOYQMRXj>jb
zRX$rnkVV=21DQx(mC?~ri*c<Y%h<U1cn+B0+Wev-?tU0leN$6slz{UVF;gG6%bKOH
z{ww<E19DkSuTEa7H^?&{tq)fD<p6J?pVysljJqKjRZzOUwb+7uG-``um@8lI(#q{N
zgWK4a@{eNvgjPm<pYdNPEHQUDKVm=VN)O{cE40<hV!pkx3?~x0^OA9&vnn=BmS=hk
zdmQ-1cxL!!>r+S>Ra5Ne24tgY>}S?Z8QcvUKg0^{uD1jvEG=Pc$9q(GhFccob3aSx
zr|}0}SkaHmZV;(k<W1e)a@*ZrB^PkYKTkpu92Asrl*UEAhHR!gPaLmfA9?=#`JjM+
zIKwE%8e>PfmQs4Wf=;$O=A}rnCi5e$*ZPPkBqk>_^6_a=aGJH6W|t@Z{{6HmoTFj1
z44Gj&yU!IH1jg+dj5{t?#B-^J0wWQyGSi#<mZ?O7GeRbiUQSg}F+gx<O%It(hFiDV
zf^I354Xl*dPUSl-oA{Ek+#io}&T`wD7jKFZ&<fRcO8sos9ryY)g;RdCq@*OWyCai8
zupYW^;G!>(kc@m7n-nzbD^gI?u?-Fg2w>N%ITb1Bs)<XMQ+s>y4!`$!a<rSGfk6^%
z?~LOlH&SfIn|n&_voA7fr;OD4Gv2uIAc70i<#FIp(7Sgr&U58z%TsuK19N2tB0h|2
zSy<%-e&6-g%H49!0+U1#UTs9qJ3(m0%k2BCJdqpK1=2HKzb<n03}q__biVDBs`%kN
zsZb_zZp#?e40VN>3Ju%l3`>|{W7uGtrFNMhSN+YaO<aQ$WsA*9_wPT)Ej2buODgK>
zMzy!MyO1;XrKe{6Sz;TiD{!1Cz!N^&G`GR}%885pnC-V8+}m$iI;BpOn~9;iY5Do=
zchqx)@o4vQob$<xx8*liW`cu*N84CpwgmFdRw?f|krxWp4-!5POZw=P+Nr=Oma8wJ
zC-9X!1+<fJWB<6gxTTrG!L)<~(c{ODr=_H{5<1f(8Io-8@ywtHT^4;`pSm2^(V^V*
zkw*!3htt}yCy38)>z(j=ZwiZ|dxP$~<pXY;=7<V587M2JE!K{%tQV=<U~Q`}k>L*Y
z+PQg_u4O90?wjXXwNNIVrSBwi6X@NIP-@pD1rGBbW#-DQT&Bt`YY-+Wc!gHDQM{n*
z#zcS=)BE4sj@h%kxrB{XQT3qoBlirae|`cL&#96m7cJPYEU3SNFYL2^i*rF{z_=?T
ztU=#>dkW1bQ@Pm8=S#Z?FGjBCQw)1g#>6CT9h6n%7zN9dmy*)b?6v+ht4_Tu`V0+C
zO&kG@6qvH%51jV1+SluTkB_VI%}lSw$HzArM!Sw7l6H!Mv+;X%br?({&scY^{#!3U
zKh7u@_8OnddN75JAjfHCY6=R&KOaa-`*auEB#)FwyKU5cudR)cVk-VV^}&~AY_Ua%
zE1(f%ZJZtR<Vjo@yRl*cixlgjqh~U#23;JRwLi@T(z|OFTG$r*!MTIlZtJ5eNtl7M
zZV6VKvBr6BYV}%N?pIukM>2lw$NTFt*4DW^t{eG!n)mKK^YZp)1r3s;7oLp&13E|=
zv>uPkT8^Hkg2MY$wd|-cLDv%TT8{-T?gljG*NXW1lm$gpjRQzoIEY~FCRd;3i{jze
z6w+91{yG?MUzQrPpE*?$tTZlv?)lDtq`EYkkBQtawpr$(<i(~K|ICT)h7G0H15bSf
zLZde2^h-%jV02-TI%T1iY^p(2u0cd?b4IfNhIYNEdyr?1gYJ!<mbsV5Nhoo$(Sqj9
z0;}43dEJd9a{R*|+$LJD;$rR{VzZl8uVn`nlng2;Yd7zebMZr**Imn6Dtqp7>bs9T
z&QNm4z<y^~lnrOEn_Q;Z`EyjVy>fROb6Ibxa(A<m9LDCuvH@#N<^JO6Syd}5tKixw
z{%$+4NR+$Lpg~}}%6p!n>HxdJFtrG(sw@`tjij95dID4FG^^x-15JHH!*U&cw9(I;
z$0q$H1$vP>CD^v1su!kxc4A_$E;8$8fR;-GMXWMV*lotZ%#3eOl;SY^dAGn6-#o7i
z!sfQwI|zP&3aGl&PAmIFtC1SbJOf|1jY;~#MKV49ddf}0$6{Qn5`FmwAzjuTEPFET
zWQdl9Rd`f*IKi5VUiPEBnVDJeB)ov%DlvCwFzJBH5&m*c&$BdHadFp@r(B1>zCt$a
zpW$!ie(sGQSQr=?abI7Zq{0ojuIeDlFkak$#fU1XvPx~oiNxME<F<SMohf6VLq|@v
zf@Z}||74Lgq^_+1pQWSR*?xOmX!kvK?AB4RMLXM@&Xk8nv6eUC-9VG3J^uC@ai{zy
zsd%xiooeVA>&qB|SJWR{$D5tAHX1m$ja&;7CjB_r`<wIgNz!FApL;2H#l^+bVU}2H
zCc3mRwmo*qvXM)JyE}N6rrt$eobA)<Wz417Q3e_Mufn7EB?IDX@gkmQR2+Sh^lJTN
zvwI6M*(oVE5l^zG(!CPV<Z8N(V%cR2-VNs6d~Wjwy~r$q@fmB5OfPp{ttT7#P+3&;
z6W5Oo2n|i<n=KN#F7DR>o+1@apx<+Qb9L&w7m13WCw5E(vm0k)xVy8BEYQmQ6!IH*
zJ_I;7c(nvO8RzYN0<n%@dvs~+)uLP$Bz>v)RE>>|LDWODipR)(FSFH5o~%}?;5O`(
zl##LBILKr5>t&SV*yZI(oH(M^;LPF_!~q3=G)=m&a6My7?9gs{R>sXccb3=Kh9(Pm
zp8hyTcjE>DQSGE^+E*u~#bTUSXL;w!7R&r9!DJC|$^I8_s}H4g+y+H|NeBC$7dtW3
z#(UHKMfR4!?$)BI?R1NQKI~<8eqXCE9tU_qA6VPi1c!tqfsmXgyDJQ4&h8=m)2B~U
zjB!cSwZc49l-}On*k9j|hmQJ{=KhFqilcYYp?)}^vuMXLlblW999F)esX)K1;^?rl
zE+Z|itgZbqK2>gX5;l%y*~&^89!sTRwcU$s&A6+_il^GTbi}XB_RJ{nTyf0Jgo%@E
zpJ<EcfSnw=h;|r$rHk1->%R5^S)^M^NjDi8rM$enz{w#8Rjd~=z>S@F=e8KJXA*JX
zT8u44C2z_{%xpX~_bX-O8g$7!a*8Yl98TlO_QXBc*R<<)`R>l9h5M*wEx|@L5EhTm
zwr_RVlg6a?v$@k!n|Oi4{F7wIqi3lz*4EY-rX*x!I>B>Imu9|s^Brw6+HJ`FU;9Z<
zPY>C-cQo@Qn9G;;ycSqIMSr0khs@Q-u&og1cgcNs)g9SiYG~I)&^B`gx)4ZM=ER-)
zoV%52>a%j@VCCAclLZ$2nss$`$ojEA`hqH7YL`MTa^G2uW|(<vdY^}?grF#gtqmQG
z{T;#DcF8uvl9A*^F+bMU&2?U2qK-GSWanGCHeK7fm*MQjzpN=oz~FQ`_6u$OdC`u;
zQDxgt|BQtR1`l2pB-7j9U(M06_{eE;<#F@wJK9C^uy1S$hr4!8CKWO;FtD3++@un8
zQ3p$89DGKbzu$9V!J5OURT%Ck61@5uGpMUa3HJfF+Hy&I&LRr5LW%a^XY<|*q@;=9
zU2vFm{Ky>vg_MT)^@xrF(do!DloMF_si%8sS~ZGI&TODegR38_6g=s^-hfV%?k%xR
z=LJKm@3cDGzYKbqkC!h`3v}OzD-k`Wo*I+YH2lqTY8j8lFN`)6fYWqY&!p4rtHt*>
z#NoWQX^1xw$!&EX{IOLN*Pu#6BBk{ynLdtBX<N3{5oCH#o&<%3rS5Kx3n8i*MB~b2
zW8fKmeSOWGN6*0L9G|$^qy+KZRURG<(49S>^u&W#wPhBA3cqn~<njk)Y-+0-*!JI`
zY!FQ+Bvjl!TR!@h*Bsp4QcGsF$@r_s3Se3wFxi+boukB^IB`PzOyaN1OK6w2$>o#G
zW$?7SY7psdG-2H0cee+3Z8g+q1}iI7dwu-;vI02VhhJSvqe`oe3}Z7S1j-4v^<pY#
zmJX{pYDG<pfN4DfGFkoh65r{}Ja~22g*tkU8e*7jP_V{z0dyN1Q`zZd7#{^VHLHZc
z;QFWPsI17ubZ?;?rhGYMV|6b3<x3(dKW9fr$KcvZh-*4?^`(+)Q~9YL{;0PZ^K#qX
zFb1_&UstD4?zFrFI$dLPb<Tc$oX~b(vShx7LPb}1b}0p|p%I2`$O_lZ_CUvTi(_m<
zO$9tgtyd7AY_4K63$Y>k1_mX!x%fNF9Bq*3D_qFE0+dnARgu@qj!*M={~DG;KGo35
zEC-4aaHjJOn#f&u*2n!sMHa`JIuoTXiJ#qCnT(42c;*aD8acbsJFuS99Y@@1{irrT
zv743;K=>>fP|JmerARDsXVUppVnO1!XGu~m^y9c(6p}Ni_U(imJ5en0^#>Cbo3l8O
zpKPI>d4FEFts;Y#s7&zqvYR72!FiiPyj#R=?d#KSiWWjwLRiw7sl_GP4xUDAi_oq%
zgq)q=8)o_VoTB1)kVirq#KwWU;8fm(-X^Fa;RX^9!wI8fV+5R>kV>9GQ<-K<jW5OW
zkL&)3cd7t>)mTgP+)_pPp^I{>k0GpwNpFKmz>Y`S6P6}%P9O?cm<&ET$3Dlc`Pu^P
zq~d8xZh4X$((>R6E6dBrtPVKKrK@HtB98e)yg$Ug`UDWB*|J4h@TlYvD+0!2>35(U
zZ<h*w#0{d_NaxvNWF3)LinM9w*;vBKf!^UwK9!X$avTB<ew9i$*Xr-zCsukf_+T~N
zRQf1~`w9vJqxT>@NRO~6k;Em*D8cB#l*EANtRu9u7QZv@zEewY^^G^m?MPWM=M+9W
z8*%SJk@@*i@g>3Zlr&rUU1od0HQ*3WJNaImdfr5%esJivd%0IHX?g?NzRU};KTgx1
zYV!?SsFPoQr3=-L3q`wWK!Dg;JRUXGxpFJ{MXk?ewtSce*xhoFU{MD3wX?HYnQn^D
z0;emzlh=8fKqu)m^z116lkdkmv(aeq2r}(1qag#dtREK)8k31+TY>taEi;M0XgX}%
zEcenJX0=S}nuA^HCD_!-o`Q5RFDwu%vM!Fd*z=W*f-@5tV#`6~?>cZt>+=V2_}UP^
z{Dw^U2Z-PzU>MwYDlvnCY{u=hfgvFV1bj9|QBkoGEb?Pr$I)O-^Uc;+5reI@`6S21
zu`!6rSU^-b!PKys_vBlVyCo$JcbbqNtm$>VTk|29Q7v4z+$rqC2Qpli4j-Fg^GAl-
zV%}1_8CJa-9~MxzH#j-F^1)G*1`jqUI2ZzooX0ON^E>1~YQshvQWQ=4p*jYt*}6LV
zI#D&#Cfu>Px$rJRyWQBSt|nyYuw}w5`b)yVZ>=-daO4hpr!wBCm0hus#skI_-_Q*T
z^lP6@i@*&!y6W#WHQ^9ZefaqCqh|J5+i@9<;X1=FI8&SjNZW&ePu*s)e*{zpR#(WG
z7#Z1Q!r5710#)jp!~G6Gsv@K^ss>UO`2KfRQ8h+BEOc~qH+Xq_U>JO-=Xu?CT&~w`
z&DvL{vz)QNne5ttdrTV2jJ9xEnt1%=$rEW1xhAmb2D25L79^K0jen?($o)#DNA38V
z7?-BPNJ2_F>LB!PG?t|+#B0FCy#q`AwOo41*3rmk^_y0dE21V$$vI4v^>>~fIvS2d
zMe{wl#o3whL{5Aq->_y@k58(I+}&f4W^8P23}Y*B-`&v^=sARX2O&mg$Q><{iPl)^
zT}9#blV51Syd^b7aFv=(&7BPLeDk!Rc<1_QSIA-x#ezyU1+_`eW*{9Sa^PH*>__h5
zr43ISw^k&!&~chfbbI@3;oN(lsY-*THWW(ubXSH(JSa+8fBuCx@V2oqz9YP}DB=t-
z+2mXnAA1T+li^R}mFaG8FR!-p)q$x+b0z9)0an}2exip^D2@*T&YwYpWY6}OnhumX
zK09+%_$3AF(>I6(B)xc%fs<2JLqo#}G)Mo6G={ETM3gnly`@bYu7xA#6f$z;`1kzl
z{x`Qt+9U5l8gT+ig1ESpnQ>~xd^>=u1V<d=MkK=)7sua1p=5XP4K^qmCC?WX5`8t&
zXl*h%Iw&OM;>2Cll^7UzJV+3JWc?Q3fuU3Na;TKZ@cNc_vh1+*QoMPAkS{pkcLcSS
z4!HyHi~Vw2m4YHs*QiG6Q5}BCqlOan<kS8ckU^CHdFYM$fMA7G_i>b=yxT$HnGR*Q
z_RF+;W6e3`g=ErEs7IIoXNLKH37bTRtmCV;o~;)QyozbR*SPhQxKywg^zQ9V0s-vI
z%erpmm@ui&DKSMIEO$6qWO1GbMiTfA)PueI9y`h&`Q{2w#0hq*og>M#FK9_GhMN4^
zI$htbd@kn4-jj!)!(DD>-wiK*`qt;|>NL)PsXK1rgc@Yy;TVE(5kE2CsbY|0RW_g{
zy%1_*zH@K)@L}Y~V|=va*F@WhS0!KfG#*91@nA&0v3B{GxOf2hJaoeA^|AM11Inph
zR@Pi^5@ie@ME{V+5L$AOUr~==!?WM`<r;6ao{~%pFw3co5@Z)?Q<zmG-uu17M^uF9
zayy4gp7^S=_5BRoQqXtqChC4Ak&>h%H(8N$go{!xO`wL*@J$-pbFAc1jiO|G*}N-z
zjNW;<y)#g^t4K~0r&udqWh=iWxi>BUIeWkNNYrhSxV@pH)L$cV3djHBBK@Z_!2jQF
z^oh3Z4<x>fKp7y?59J4NL$pU&8C5gnLHUtVP^6!s;x&R@sQU8Y-rx`Z2oU}DNw2m8
z%xwAEuea1q(b3V_pvqK3b?m~Rc)_k;f1%&5*L(wvbH=f~wMP+22r#NzFL$a<5z@0)
z9CmnY#%>A-=(O-J`p~9AVh~b^!BNh$Jg2Goe|o<|p;8JKPQ(|iA63C7DA&zf)k|a2
zCT_x3g&`a}9zsiA790_ge(%K*9vKQS4@uHtiZj;W=|W6lQ^s`b*1VVl>YVCLzVg7l
zRAIJIGlhFrsd}0wPd+A{ZEAg&m3qMElt-lkvf+pl*nU7Ip}?r^tFb2JN0k*7<G}6}
z=`cY`7qq=!C3yR2&l9Mt{<%1t|Kn~UI*4p;SqWcW7V{o<(o2{4c0f&#OuR!~nNYBE
zN-FzmG!}4&m1!VPE&?`FJ)?Hhfh&}MM=&Y+@S!*2ptG9~R=7gUArC2;KBh}B_dkEC
zqdkwJ9>2=Pv6(KORdo!AWxbvnQ=|+Z7}8zYb`Kx9^5~|}ff&1k3!);bsv#lF=vj7i
zuwUREoe<@QnVfjDNh@ML2J-qw8D`FpS|F{gj2}{#lDup#`jB6>n;W<@*{Ej-M%b#+
z9r8&NpxLvL(g`>yGDzsRv)<xPkaOP%Bprr*TL8*NYW$!3RqSk-E>E<H2skahqeV&u
z@JFl2Spix}hMPAnX{l&$!8-45;4AaZx>;?fx@3?w*L9xZc>S3OMI>$Dq?CB0rmp`=
z2xsJF*K2I$sX}{;%oDQIW;vM{T2HnbsRyR1Wt0_3x{^j5Ma4A7#hZnLG9Vepypype
zTxgGS`S5SL;oW_+j^ikli?IY=8Bqvz$D1Q76KQLVS0<v!+c`Ld-@_(;>F43YdjKVz
z1K+4@^RnX)vxPhi81tQrKR^}e5CCU<5#a}O_C<q%o|F|A5%3g*|Ifkysge<!cT8%O
z^Z~3$+Jr=Ez{rTvLAZQ~9w`StrZ-f2hG@;_zZ7fpJHB(v(1ch0Ks1=(1Ke?xusWin
zy%xE+=`Pcz<`gNzE6Fz}Jx7UwE;@G^Q6OI*fR=Bh(|W31W*E(BQ+F{~ttcE(KB`xc
zCsX`)sn}q?3A*+eqN&btdixd0v33{;!s!nFpt7!*(bm>#@Y2BNGV#4L;MEwwG|V3P
zlt0-xK@v78vwis@s7}$Vdv|YB;Dixhl~f5hy@Ch(ag(=RhXK(8wqnTTs;{v!ozras
z!5_~}u?x_#$bvHNK=zmSDHQ4oYsX0}oT_dP3$Z-as(q<Zok233tHS>XR{GbTwiOzJ
z2^CgP`=|_>|KZ-O%l+SJ&2L(JTFFFN_i;YwZj85I5Ce~(pF-r&#r>rcF0ke#MfM)U
zO(t7uDH$K!QtV|!3OwFlC2Pa(T|A`PoODp;V`bH?x+{{esuRPx-Cw{}L^oN97|)+R
z_6Af@U*yrHdjq<ov-~j&Cy;Fo$H&C|U2UzRRfz2Blgp0Wv6DnwezjSsFC0A;2T{+P
zz&xu&JxytpfX8>jhkBF2M)zk|Q{EeSpu1^GdHDv7-KR}JFli}JqO<`yD33?V*zc|Z
zv7C!nSzECO-M;Cmp@zuK(Fagf@M6vF>xfi*!IU{ppFmwZEUAtXzS{uG5hSbF^oJoX
zjQ$U-=my(uBA)Z7|4>ku9`s0)_zEM@zy`^{`bj7TKYg*i40Et_iwGoih>PFuFiPe9
z%fl$5(yKa39L`udI{NRtGIEkW@%|h|qK5DOP;Ba{!K^UR5PzPOeq0R$x^^6S@J+J+
zymgr7MGtMghDnHFKqle7*r*e)$lZ8ec$ymWv}R;L9+n`XZd?hbX@Zfwt@BO;Yl51#
zgQK*UQKGkD8c<h0wu&D*_#Bza%gEac7a`&<{O(vat=Yf3Jw@-lC{U=iGstSoGa~vK
zwg~BkUDhRoeRUsG3>JvEv0ac{*JWM|$qzy9gV0-;4Q1mgy8(z>>J0LDwyUdIu$NG%
z*ekGKZ!`}M1wDP2@b#A>NX0$4>A?$^VXv__oIj0Z0Fq=Q(<X7rrVx-2x-uh>#08`j
z3ma&mCL9y<Wxmi9&y@ZZbbz3u?qg&;moLqACpT&5ij9(Hje+D0H|L6e%WDx#hZKko
z6y4MwJa`RJay+=<1e)D7&t847!w_gux<m3Uhk+8b$@TxZ&(E>nmSq3P@2Ci2DL&@j
z3pWib<sis<B|wB<*gz%VqzrMGf~ID8R_QFSIn+zMEkQY)C8e^8ewxzc0@GZk-hTNv
zg#H;|xEbVVp(uMF!Z({@X(S`<@4h8+N!KCP`|aCh2&Sigy=CqiYl^U)Egmn?PRhuT
z29xWDhqy=e<yq8Y-&_(^DZkYWIXb!VL(fU~yxqjedrsFOuL-tA0kU2kP>x0tln)<1
zY=@HKv|ZTrv>D`id3T`5w?)Q+dU>N>R{aiG0Y#FL-xlYjBAR^!!AI<fetaM#ev8PX
z2mF;*$hwG$i<^S2l*O$=Ej1|Uf`69+#P*0C;<KB!<ZFRC4`gSj*Loo(7y0XmhcILI
z-THn<VVd7FDznXp1Sk3ZVYF}mqkAcA6tnlFgh^2-Svf5e5=R?CO`+lukN_FRJLn2+
zsM<pL08D64kySjB@-AG2SQ5gLF7I~-t_=OvII@uhFl~tcdU`V4q}%Q$-|$IrMUmYe
zzB6nv!?-@$ws3()`<rmgp6{<7VDv5_RVmp7ig|?o(n3}?wq?lJs6eELgA6>KBr=JR
z6A6lpyd#O!8*SUfE~P>ifETJ+QBH(g2pXfX-!4Ms>H*<aaN`lfqI{V;XWy>0f}So`
z5NW0O&$||EhM`8><F!A-NJ)X(Y2qqGT4|7Thun=7_FS>^s)d-?cDv#TMEoV&E4?b}
z>M4*YNKwny)hu@^U}a^^IB^{HTyfu8rO{qaVEcBkhZd1~HG>4pZ?9w2*2$ZsG4&iL
zQK$?lc8PWXF!XZ<tABd(3O5BcZrK)r)sMjuNAExU_x;VyFhTqj!iwav^M-5zC%mN!
zK+EU$a4n)Af%MYK`-|q}vA3bVu<e*c_z(S~I}e~de%uo{AA98y*oXJ^1Dw0l==>S;
zMVmdXz9#iI3i0?M3@1GoJv1cFQP33HoZqxsea`b3HCVyma56Q0bz#1U<}KF00*Las
z@&HbMiJU%=-z2`~y(d0wRD)5G`{Bup&mrJ3JaeMGf86WH7k{IOv;U=2x&2$B0?2Ax
z;f$CQg!jkjut$!s5t}b6bc4Ml;A4xKK73LRV8^)7oSyjbJrL~Y{%PQZAg5V<u9rOg
zI_aRt56~$=rHdYdKYG)_Dj8|{n<)5;-W*0AIBmX2-#mCF;&(^bS#yZ8-cX)WIEMM>
z=N1?bO8BYXoBJIXmnB}Fh%!PQ^!U7vLWw4YzrG3P|J*BBDS2iOsgCQ!CmiAq96Q$o
zZsR#evbb}qG1Oj<5RU<zAd=TNiOwB20`CbF?{Sd}2WNxT$TFLRQsS;BKJ@b+G4?(8
z;P086=UyR}W79ZiQo7ddl)@W{ZxXkMUnP;ir5@2gJ#PlNT@GmVyLy5+mCTE7YacGP
z_2%k{XXp<0^KY>Rv>`~ic}0ESQvYFtM>qe0Z!=b!wl0woU@hQI=TPvPch5^*S&PTX
z?8nL;9*<}*?k=b$M<D(g9LJ7rO5C1xlPjET^u2riQ%9r_Uq9_XKRtLdf7ac;L~?J`
zsB?FHzW;I_z?K*v-q^!$om<#ytD<u8e6Q}(Fc@hOu>WQIRwsP<7QU>7FWdKA3E>|L
zhDYs%Qx@!#Blgi?Yub*l6hPx_T)|{p+f(Nx;CS5aEyo!vxH)&Ug`P)O^7SW75+Ngu
zUQodu!SEI}m<(og_0H3wxi(`oj!#cVTFCBN=B)$=@gya$4LKnm6C8pg4N0eWW`~+h
zDlAs+<(>bCI9aZ|wlO2i5aH$%DjM#{@xDh!deyVm<5}y_)c)xYD+D=*+cSu|7R8ez
z*dvkOUm?E_!|y9+W>378(ixEN#g2WlOy@1i3&{4D;bvz%&T;C~#z|ymUp|va?fXe{
zNO1G@*B!Dypi(3`ulyof93RBXP)t+xA@-89F4gRh*in1ch2>Aj#|-GIT$;KrN?IBs
z6LQSTHnVRn_p9M(iK)}6pGj=~2hd^f&o7Xweo?A&vl;P+QBDsvBcrV9$w(+MT?tqW
z*dl+cBJo@e@i;xssp9K>nK6nPx?KJT4mIB&9;-piP=Fh%QhBDyw#p|lpc!k)6{Dvp
zPfz9QYWDn-hNKLB8ZC}p)w->E_4q7#<h_Hab8&>8woqp0>#NlzFV<NeeHpx;7%$|E
zJ2DxXbjnVZ%lwb{{_y9=Kdru`$D6HG7OvA>WONK9@~7Q<4F%iOzO{uL)oN2`x?uW*
zk0nM>IOAUbeyN^3dCSY{-h7AHmGzCEM7U3@zr*j`K3*=$B58?7_RJfJ;Z|Z59R{un
zv)4oEwns<wz6i&>>}is@$!Y%Q5}GFB5zk7M^&L%L+9FfrQN;8=`Q#RrkqgQs=Q~S8
zitQc0n8Z5RkeJ9&Y}MgwZ;<(w3uTMB?|cwG-^URyi$Y!VX5OUx!{t0$qtjWJ#CH6G
zS}+BAAYskAziNr>?l}qD^{qqChUz#X0_0bpih}RE9K9U&pkgC<)`?#3I(qxzwxO%)
z%&PPMCb@sojL|vEuZV9W{`s{BDcGp7)DmS27stR~Qxq}aIxbIkJoCdZ)~vRrZ8Fi<
zdA#}BbUH&ml)?<-Xx$`Rnq(9CSs{Xk#3b+oH-+S%wA<#q^WaYaU@uDX@#9#)i6num
zaW5<^%wpVqu6%WA=*3wApo^5j{Bb~CQ^%%>L-jQg>dy-TwL-?^ycqh<ly|c0(TqXc
zn^!iqNYomfc-_j<^##md51BVcaaz4g0QG2Mn_=Q~ikmE3(j86l%&%K*QweW-^m+EL
z<)y_rWB!4f$*ILhog0}wX!q<=&a!R`+oogLjS|Oya5&dFO6q22FYIt)E~H<kQ59<Z
z?Q@okQ825*O;9rlLSQOhJH@~OYoWfex0N%oH+-K6@Sf>}31W!mVblmQa#-!R=ddLx
zH-)i@BV`qU_NYO2Lmg`OYSz})!Ruf+Gv8kqXMj9I<wj@r5&!`JI`X&&Qr;zYGbSqP
zC=^Y4R@}AJo{DoW1sSFF1<9@lr-c;9Rs`H@r|axY2@~lh`lcn4spgFwDmUm-cQabH
z+%girXb0MA*g>ApgfX;=%tod`H91L-*u*tM`BHWvmr9mHZ(DS!gIoU{6&g3Kp?Rz2
zWyc(oZPl6VO1uumfVSxhdPk@K(IAWa3v#ZE>~FPlokdYSDYM}MZVUOfQ&t6)ojh!5
zVqp);Ki9C_sZTPAw*I)a!j-o5PF!ZZP31hlvY?G>-etYpj)F;;!3v~KkJK@w0-)>)
zDbysPt>LpJ-~y&M3ISe_iqB4(#&zkvUXR!%odhJ`+x(G-+B^?3r;vPMU}91QC>Uue
zF|l7Xf2NP8;CotgM~58pnzc%~)<<uywsIeNN3f&n_<SlqcN(jvz_Y0~oi`Yma6B|&
zDwuiU4q7$9z<)MXs#z+{+RVnDy7h-zWtv@d0$XsDl?|Tf4&%Cod-~hrq#4am8Ko!C
zI-gDL6PA>NZJB!&)n}YsZ8Hwq%5*LFe6W-(>8(#mc-K2fpuV0p8NT?t=TE}dne2tU
zr0w;oH?t*Ek^GXj=_iXenRxi_jn3k=?E9$H<m!*8jaT93_|v;~t5-L5vTCU?S=a4Z
zy6pM~x{mVqtEu68YA&lyrj1wpSv%6-@8n|D)U`}LUm&dWqF`IPXvBXbna-Y*UM{L*
zV|fbdkL<Pi0Q)#i&Mr}0T+9?&2iQx1Wke<+a`pBS^XF@{g+45-^s;MMimI$%zm2P-
z1Q(|itLI13xMy}L`j81b=z6N2(Whd-R$|jXh~bu?Y<VyH@gDK)uNtFMU*gKL<%$^i
z8W??Y2X`N^B=_%bpd+~i2a`T%&RW|x^w8>8?)v8Lv}V^EX4t>wE-*I2Rh(ZH5Gj;+
zqvZelCqu)A^*_@&?HL?0(|=9&b2|yIisuY(liDQt9DF9N^~XC<!pxfdN;l2TYG~dx
zi=a)lZWt6fT93t-X>(Ypqgj&Wigk1RdY7wnR~GHXdWk}tbWCi_gT!io{4~p*Y*e}c
zehN|bYM4^=-I|-9w-RuhoRTdre=_N852)FC>ua4@UaKrj&G{V~QFk+<Ob}xj%_i4C
zsy6;Ju*6?wO2~FvJ+uA=p=-|Ar$D2yaQQxFWqI1ve0wDt`$zcG6P;!q*R6^&gY*bp
z4V{LKl7&?c_0X1=d^;)bnhT=e<b<{pR5LXZlr-entfecVR9Rp>`Y=-?FR7lDg@wgw
zdl}DR{ad0oA3;w+nX3b8$=Vjld+|~w<xoJt`NYIT$6((qZa3{-=amI49`yy&UZELp
zL$0;k5KgK;o*G}9A8(KG*E@uAzr2!OtE@T~s+ClTsg%%|iBKu^5vp&gmy@rKV#59M
zr~dQQ9`z;LMP>I?ssep(ULMqEC)T8-q!yj12J$x#pxhT!7Q;$HA|Z~&uCp<^oTosk
zH{mMO`4vJJaP{1%FKTdY{?kmzvZ-OQ*kF2M$VDPSRcJ#(LV^QFB`qKen^KEk38YV-
z)_QS-90o)yIkn4r^U*fRlG75-qvqV?wPng6HL}xAc8NH>px|yAid`&0xRT$Fd}8t3
z#7uNkSIjfHF0Ig9T-%LSGV$(PDuZYE)Eg<HQ!6@8MNgP7#&WD5<*=C5M86~qF&M}d
zQb;$b)~1c9hkhV-SgbK5JXbv~tN*8_oUVY^UB8V*Hal=+<hGYWn%24ok6-QtSu$%r
zs~SGHbdFxGqR?fV$6?*Rd})&l@62CGKbRf7+R(dPxmBu5z*jg1FZ6nk49)&@riiDX
zvahYTv8e2tUcP7FbWA|UpWcf47GGjxZHSb$dzMMAnN%$GHeQuf1*c#G20f2^xqfkJ
zI(M#I^9J2j+1<JDW^mmdeewR;gy@itG{9hVq^mI_F(M=f0<lSKKakjQ;ddSE*82K(
zz;$pycDyzp5Mw+X<=k3;MSP-HZZ7k?5t0zp*>6pL$}1cw`gB(znUck&JUhQ9>e(=b
z!m{kqE#6}<Irz>R;jONB+U)Z5*!E3Kxh8T5+L^o(j`kGB45Cf6?05^9vy|;rFX{*S
zr0TR+sWtRq4W7PeWXimtBGXMycWY|v0X2p#PwEhdc}U9FUp*>~CKNU@y{BvR(?{y8
zmO4MBN#50Mskrdw17mb$E^bs--_jt|!8GeM^LaE!s#HJQeqi1_?f;fFzr?h=e>ISP
zP)jWNO!4h((}tbvFSo_Lbt5&?jPs~ELrw>eum0XFu<kFqc0*Ps{D!JC38j9Z9<7HJ
z24mg5=w{yVqTUN@Ut5u+C*NOEF2EV}(FW5nbyE^&F%b)qtHi>VjdnlB6k*4pJLJ9$
z_`rH4r8X*&A7u`Q7g<h)kqMM#wNQ?$3)tA$Z@c76H&Ij*W>+hg*TSurO9&gLg8sJa
zvT0@8(=GP5(%dq79kjb$`zHKeX)WW4`wyw&3o<6ZgZ5m8f&>A<+;nAXaajElMJy7e
zWI|~X7+Y%IkJ8x^DIOjkr?D`@eDhw;rVlJ4X!q@c-tVGCKOn@u*OG-Tqvo=Y+-TX5
zMlwZYq}N_0`&%s((bF13ZH(MGA?t3XkBm-OQ`7DWo>Q{$tc3lhn@(t?Ap5rkA`5I#
zQ(Hr&klxDE(yGs9wkDd{+5RdzpQq;ZO`imI6g_sWb$WhvEIcIG#C_-Ba(DI}wCih=
zC!-`(D?gHSaNR6Tf!k^P?lR3;6oU&Jcm@HT*@e>T*=dTb{p3|us-TX}$5{tnv`*Jd
z)2Df5^DK_dbKh@yHRGj5UN4tx`v2b8Ny5iBU~9Kd%YLJf%4!MblNEAs_(9$6gO1L2
zqY9rYna%9T*0X6^vt8D27!+S=?JYznjEiIO$p0(-r!~yG33ZMcO}0<$S2$*60`6_l
zIR<Udnd@3~^M=rozXn()R9iLA3M{=Vv7gIAiVwekBRq+K1Rxuv#|(v@Bp{h3ElhbL
z0%Pniu|?3D2vW^<rl&LQ-E+Tf;@>4MLmnl6@)sWK(;bd&z`exX+3rzIO-X>felsLg
zM&ix~29PCp$2*(!a9n0$<gwN1>b)9OHk|cjZ0bXLuSE`%R%CeBRJtB%ifQiWhNK%c
zldGW#YFgc{zh0kw4k{-kvw`DMN{gHMI)P1@YBTUcXaCQSk-a%+(^QLJWvub9a7*DU
zQ`77HR%%>Kf^G)1!QXdxlkhryb}JESNSU>)ohq!9JH2<X(r;o`VH!*eS8a_}L2(z7
zBnO|^zNAWbH<Yn_W2=PE)6>6{`gCo7v2}GI4S(|(6JQEjI4*rwR^2ox$#Q1YFTf<3
zGUbnvudeW>-kO_`(QTKoh>+TRK^HGe^w7LUIIH19tgV9X;|}*OLkpqXoRP(rYW?<(
ztxtb_(}>A>A)JTy<Pgvt&z`#5I)1gaH=7ade+XS%yF@5+-4ugz8i8Ha8EP-J9M~1^
z+2FzS$?cNcQ&P{oc`k%zHaNMG+Qekl^)N;3m(Q%yL_Th>S>;qu*X4EwS5@pxZ<cKF
zM-<MN@Nq`DVY)FpwBxnEX0%-93k5n76wHDh`l=@_N(Y+VS6EV0EwtC?SEK!dCf~c8
z+d4N+Yy0y4Vv<$EHOPiK+fJ8j6eiauv0}yx)5=z!3UEYz3-4gR-+0u~*%r^|$dP5Z
zpg+_3NvuqKb82S(6M^5!I?}<(SIEKq5{JKk=h;m^!d&!Rij^B*I-w*g<VCDX`kBWZ
z`oDb=pyC2aiHXq9^KYbA#ET%QE1$?#`N(DZ!n4tiil5b+E8l~*8`^ccRr;Md%F&_v
zuD9e_Jb&fdYGh|WIn%+R(W~zX3m1I(R;Q}pi#v#bY?+Qh4qw()_3)Ho<>xJ`>}|D<
zJQC$w9QppMoXsWEq4c&RUjtahy3LR^bbVKdk6$ivB5%?8sQU2|2Cvg^i#eR0#Ua<F
z6hVK6A$_t=QZ71Y27k3~#C2tyW46wTL2h-4979c$K2=hv2w4MNI`S=F(HGIJoYYlu
zpM%eeQlD26{&TE=Vk-Q%8AI#RSuQ?Z|A^A&DcN800&?e-pZzxP=WcBh%9J?wvey=k
z-wX+F4(1YaczFnuZ6)U?H!91x-IOaxE?+Ku$xm26tLJxS$=t%bt$-Jp!*Q-G9)BK0
ztL*5>$#?fg9+e8Xm$oevTq*#$o4{-6KCEV2m@~KTYC$<>eo6aqnn<Vc{0eztiJofu
zd!s*r!WaEK7^}3`g)OhL9R9-4xqNL-hTiI!YeiA&LbXgX>G9{5<+B2l4ohYOQRxyJ
zRp08{L|ppUQ|OI&hw1sBQ);{P3!XW~HN5hZ)8k7_OIpcs<?ZTAy_quoKL?wHOdp>6
ziM?^pQvbojg4*la*_+|bv29_)tc?D`3K2G>(@h>3$`PhzB_lRlzm45r&%G3VLDR-5
z$o*1l6KbIM54);%>c$Y5EBdlJxl~lODyQDj!GJ1eO?f?4Baz|6bK)KDQq!D2?RHV}
za`B_5?prg{jrrU2g_h1lC!s2HKU$@LBxU2R&}WK9E*0>dJNW}MC<Lx5BSU$d(~t6e
z{z#hA;jO3$>gBgrB;FZLTaYg4kWE;=hIu92i6-RgU?Mb$O#8WS33kP=lL=gPXt{dO
zv-t`!OPMhZlqfPT8JGXStBz7EucK#9NNgv5UcSNj!06b;@*eh232gRR&Z5`OpMFC&
z+$YGS9J&2gJZ1UwDbycb+U09#3d6yamiq45HfvQp+C&L+qVOtZrt~?>{_E{fLn6N3
z*IWCFoOhJS<A+$ZN>=d{Nfim4SeD84%9KQfk<VtlM4T2Fif@RI&y3tZ9BF+JWrUl_
zD4|#;tflSzGB3ZS?P(d5A%omd5xL=u6|1V)Bs6~lTcty*59CuMO-dB(t3#JBVWwn;
z2Shugt`zMnu?MciDF?ich+gQaJxq<<SYE~4;)|<}`+U-XD2WOx7!rHLo-0<*NM|b@
zpn^jYHQ@1gZ!c?YSL2*WWRwpmLpqVz3r_mROpQV*k52k-6uC{h?OJ^gbP6m_t^U4F
z@ig$`%Cj43!oQ&ahpNm~OCa#x;DAC4E#*c)oLAa1*$K*6Z+ddn8Y%Laq3~dlTSrdF
zC})0?A3CNFX_hq|Zsi2XZN9QBQb_hP>1VmXeV+<4Z3l?Yklun?@-sj;s^a8_ph$gU
ztKR1VLFSkyTt*}v!>DLj=uFwl;>BzRnMnD8Q-@{mpQ!R~zH&e`0bCXGrRj$dHQV5k
z@H{9l^@@iAWU(Y_;kifQ2BMVCrM0ef`m(s08EQ#R$chR-zG?aL5aimfJ(QWDPGp;*
zzAOkFTcU#^&#qI7YR5hMWc2LEm`tfSr0(4F&K_n;eVF?!_6sq?HHJ3f2s2c^Gh`F1
znguU7JdCqhd;_4E*Plft-h5D?CwqI*Lln90962%_w?UkI#7=%`9+>@<Q`Dg!_6Qe-
zK54Xi423#)r&Z+Hp#oB38G56@(_j4T{NZ697%|bfXPn(AhWP&L@Og-jxDERpJ~uzu
z8_XB3YuENhkkvX%48P~-in2$>$iO?nu|rJkb;Ju!MvgTR_$4R)?@tw^3jg+#;`hCz
zB%D>TpKgZw@xQ<}K6GBM0bmF`Sp@_o0oYL6smwwGVCp-WY!I*jQpxMLs-;qWiR6Xh
zMZO0#Ul9iu>htFZ0pW-gwc4SSf-o2n;$bI~{pL*J6u{%{$}E7kgTNc$@Ia;7#66zs
z%0ytRkeKQywZ}-d1H?JTG^_Xl#DDpkpUo!ix}jXhKu4FDb#?!$D+=RDBoLSbLG{yh
zs3Zf3Ls3~d2;gu^Gr0%}4FUcZn5scEB?Ba@Y*Wxoa9r^qY>A3M3Jf62k!qURQQ@5C
zX#hznY}j5YtW=5MH17Z)vn!;NN}%|d@49KX2k{TR#mJ}xSmhC008Z+BJL-`m4k)q$
zk#+db$coqN&AH0Thlaj}y8z<I?4pRe`}jFhs65G3Bpm<1w;dt(Ak2aB=17asy5?q8
z;9ZR{-?@|7)qB)K;@*GVinGlU9&HM<(YXU89@$&BZb2r;T5|<(yu1#7V$`;omu7m^
z0`m20&lE07!@B`Q+^<r=dBvg34In3ge3R$fgVI?<$kNi$!S$&%q)+wrCxM=$S?LB?
zY{0c{Oy{*I14~ST+Tj{Avs^$BP<`y80R2*3Q{z+yIZb_l*IeAg*hKQZm~Kyei{O}%
z@+shBI{{3a3V>Bp0LuFj)340-r#DCPnL;@%AR&^^o&cEKHb|f>kN67_zMU27_`srv
zt)mN__<E8#!&G8_5rhdJjovl;P`4G^80uJ6lOlZOj!Jj;@`ZXvzg4+N-bA3A@;wNC
z1BgWg5}lq^@(}Lb1q7AM!L2ba&D^J6|0O<)DI@_M`Nm3!kXa03U<G_w{P2JU)GW46
zB#qe1;G^z+4*-p{chLTW@Vk$seBt>228AZUkbT#07e7XWbvpgUW`8K6W&Z~)7V9O7
zY%-5GA>bqeZl$#%fHiq?A3kxY?#LIZzt?8~d;;9#A1c?6Z|1`vU@-o{(GZ(LWe;$J
zC9Z@203>!!o;YUp&y{i4X9G_agq=n0*F1=C?;#*i`(>hYRe!N7oLqCgv`}fiwGSS;
zPyY98M&A0QNf9<$P>@wSPEdvw6EBt$Lvd65{T->8Aow5%hMKM`o*!Wc>;qJ4$zy(a
zv??>VBIPZVa14Bt#>l4lFL23@w9WCTr~B|3I02h<SqyykXJ-x_bNokh0oxzB7WMxU
zUQp`%X@7y04@?znPk^r@?@Ph8Vsu&02PxquCMBf-N)BNcxUCHzW&FDLQm6w@B;$sm
z{ED;*V27C3KhzQE7R!Z|+cH3>YzyZw1r2*0$tpskO7-Plc?xwfCW@3XjK?O4))>%$
z)J+(lHYBILJ<9CfW{cOMmvkEO<?4tNJ9$U9O#S);pPXyA_a<S;@BolUgoMzqUcEvF
zqNfR}2I)Mohfx4lZfZip_sO#;d4YOaR0<WqX=5|WzH;xNBivgj3$hQ>w#loj-;^|I
zi-QR4;vG%y0UNr1!{$ZT97hd%<dUFFJQqlWQ;@Rve~-zxa$O7e6e5~4BD1Fw6yiFa
zL@AKIQh^dTUkIT5N(&m)ePTq%oEfF}kOw>c-zEA5I^eX$rcu%6Y$=CDUy;cr3w)MN
zCTDm+z%|L9LQJ%x5>V+7FaqSTOLdq4q5SF7r-GgvC{c!gz1Ck8pmnFjSBYFn1XP>V
zD4_ru07I)H8TNhMSVvDGBPXW@fItMvCviI{D98vHz`DNL`%BlX9^o#3xO!IHn*HCa
z;Tn{Hm&<_~J%AF70cZitOErM@0jLAJG@($-kEp6RvA496z&Te+l5UIl7fb_ssyqdk
z1)>fS3v(Knya@CXjyE^O#QW$ko3NqHHCW3=0nL{Hbcp`>5^;a(fCPjng7DC3jRC_B
zDSEp`8>rV1etyym2P1=!5rJ30NKY>gK%2k9BP=TEw>s_CgIhOg0Q6{?<03n-cTL7w
zYKnv3G#TJPD`x<t{HOQ1z32Za&hzvIpQ<i-o66z(e$q)<U7L^_eIz0QJQ{#HFB?-H
zxIAzkIo(5^Wa9C4)ALmN2(~MrnU<6x4*?5*8}}mqt5-X&kiGU@`VN)$aDkF!#q*I7
z;z;6P9}w8t6wvkQZxI$X%%BcuA{e$3*6U*b!t}MUhyGDB`>UK&9^{;&nDhle>X#>o
zzlj{3$3h-o7dgDIT2N`kx4%7jed6!amA?s^C6rR1F$+-KzH9b01wk^wA<Xey|K|_+
zd;3*q4|1Y!IqzeiE70%#_(wR#)AtnV-zWZ&zk+=Fzfv$^@2G9Rl3kYs9_MjOTGWvH
zN^XnpMJMlE65)hBD9wES)~QL+Z%~7xl?2mtz>f%Za2!zZ52dA-Ox(O+OYH$mO&EmC
zYkHL}&ZPpS7D;}utvNs&6aP@gTelEs5dwz?<`vBYvNG-?l!v_ZuDn`dx%=;UddzY0
z?n!4##ulp#{@PCq+A_8EgQLmxEUzLYdIlwW)V#J&+p7jgU)>fJ0e=bD&wZNA(eK`!
zBfcbM`_C<BS^kWLa`V;8?hdOv`st1G_yv!r@u@z$DP-z>i&Ct8=RteL!)Z)Wy6L*4
z;l31q`LBj7*I3)m0h$&-Ta28XoUO4$r)oX)u2`L-^?Sy}tr|7*uUkH!S3j&LU2#Ej
zlR|Qn;Zx;sbK3UlNpR-(7R=J?$6t%>f%&)AE_(}xx=r%l>ou_96hObI05cR(ao{ik
z^`R2+uc8HAFVN6rLw5++XwpyxJd0I7uneSsXaZwgA9U;4nd)rBb3xjuKw}Yv9m`ZS
zbkG%w3D3?nU=ZRpR74p7ObhiNq@Mvow|Zk?n!sGKb{`=JH-y|lx@sW&&Q)I)r-@ik
zaH7%x3})0Cb8iED+=C=$9*&AhAt50q?Fo`5&=sJWXPk0-jH|K@3KRz~SPeBohY1Al
zrKdfck#XA<_!WMuJhEBE8-Bmkx1Ru|AGoQq?x0KT%I4<uNX_QgL}Zq|EH|2s05cZ6
zXqwl}4oo4*VOuwWRQgxA;@|{e85_aNA-z~Y_s6_tu22V0(x!~2e$!N}e-@bIz+nl3
zE*n5P-|K+E>$H^8YyLnxZWy8dLVlLRbpvbuGv^+X6@=4Ol$9?`v~E3t#v(uj4M3V}
z{NvcGsOcmE&GRsCzK43DMWn9Vy1ovOY>|lt2zvtX|5D)}!mTUNZ@4J1*vN*=LyX8`
z1k5+WM%maL+|A0=Duk)ga6$|sIEX}WH-MCV*sD(i)p2wap~LThXKoCYN`$!%FOX36
z;%It)en-puBT@AFgy@|m`a2rAqM#L{RI{{Gfn_QQO$0<x@n?L1IHD%UKmW2T8=ak9
zxuXfLYw=%jp2qpwI#T;V1xiLkXqr<21WQVHfO}b&U3#-O*8%i<bB(+ku2|=@Sr{#t
zZ;OBny5smjCq?c^Ey)t#mLo9h^7XNZJ!l&6B#~+%ptyoB^dK{C0fxP&%rOu6KS}<A
z8xm-j`B%`qXUHx7#@D9?2I0`ahetOwG&BNE2%nywMm-?q784VD5TH?Ptsf&YUkVHQ
z0eElWP{0U7h)0kfXKcN?-URW9A#gBhChi_2^q6LI?6sfB2c}$=4><>eF!5n}NrSdH
zF+V&hDJjCe;$QAi@-qg?J}218G2k4Y5q?=UI?cKaQ9%+c(-_c#Y#I?JDD<5e8#}_#
z7}J&@NwNs{w)&`Txk^?2$(j>H0|zXG(7D6E)DLU%SEwqvUhAu5gK{hruENzBsC>kB
z<Aq61ibavVkkgNVRzizy;xs%mbI?V`5vX)uP25eD3VWe##rvm_{~!M_4lx2JXGGI<
ze<^;{?i#`CHDLWrp!LkY=K|_!A6?mpdT@~&0-b;aXuENQ`}}SM2xC_Hie~cERYe!T
z<{<z+i;~G3boBJ;U^Sh9`f!LU1dX(L(yVYPuGrmimTWH^bW2xBSDl0s#2L${Rd@EC
zB-!YQ@83nCy$#oshPL`<CM7(`ic+#H4;An@`O*PHPwZNT=!0gyxe2hkP=)-(T|6|)
z=dGvG2}V5z)-Da<TEWgju>H3zYZ-q2nokTTkPL4{S}pvRM981W0s;H!XXIcwL2}s6
z0vd;80tp)4iAM$l$_GlEq;{KHVf33JK(zpRC(^IL&k=cjs6j)qvu4z3YMpa>XaTrq
z>1Acw09CaN9{~Rk^xqv}g~PJY5CVu9amc=dAnn<i+Ux<xaVpr$RjK_dwCi*3yXoC|
zhNjS2V8eZPZr2!g_{Ip$ZVK=Np`!q2>+pK@_Ud3ZluK=A`!s+gi?ox_^A|``h<^Z^
zJVq693z-@eHl5$stAuF+9uckt=j^BPfP{c9cs9zwhj=I@WeaK!dJT!y|M(%Vt4m#*
z8bGLsFG~6dGpq(7Im~P%!Z-z5vI+1sdmwBR_EX8J9x>(Ur2o3kr#Jtr8bMgX6wu$=
z5{%&J4E#FNS5sTd2)I^+<AyXv0Kaqsgjm(WA`8f)=k_NK2ABf+K6!PjGd&sPq@;iM
z%|rGt-3gL`NH3)1&oACT0*(zrqlaCa%b?L90o>VG;Nwrt*V2U2B5+m+|Jy(E*j&5O
z_NT>tM+s>JgVaf2&e1@04rkT(hFrvDmReMocL!@H1&}+DmN@{}EwCKEk2Ge0jy<Qa
z>i1sF-JDZ)J23+)O&Nf*&=8Dbdog-f8Ily&1gX()e{X!>-D=sL08vy0OLB<HTTb+e
zz$7%}>9|wa7YEXCi<eguqWTW-5DH*Tg*`6@TuYl2g&8EJqS8rR#RJOa7GQy4>1!ai
zScZ)@3B$w#!<Ayfatql+kT_6;7govE6+lS5Ftm=rofR(RZ}3~6ooO2fH#+$J`(gG>
z-~{>A@?Ecr-rX|nci-JcS|T9!a2ptvI7Gqf8yX%3AuwQOyd@r4%&4TKsE7s823G%1
z;~iLkD8v(5V_zWfUBDRc5l@76`C;+HE8;ctjjBGSJwWnDAd<#_FuG@@0}4bnJyno6
zYw$8qQrnRrzt^I)iRDgTGe%}P!IHlaG64AcU`AW;sF4mBNag|Hx*gDVD5nq2C%1A3
zl0o%D$CKrvyTBmy&$^4!kptE9Z$-F|B{MGWenTj;nyRWQ>iQy$p3e=rs0>KLOac^K
z$;rtnMr5HBY2by3E-d0mHtaZL6N49jfr^T1;+$(1y3$SM_0Z<d<{Z+}4JG#59WqI3
z%0V{7WK>kpVZs|Fc7YrkB0%S-U&?+V`@sXTlAo%oQdkdErhzv+IOza1Bb`$FIlmtx
zwx0pg3^AD1M~2i%KBRFAWx5F^;KupvXO&G%OysrIsxJC>nxx6SoU*U!E4B&#_%Q>x
zUS0p17J<S5Bv^z<6Q)}P;e28HZSk*`cbsxDK+;5Z4)kY;c3laB)$b_8pz(|=Vsar>
z(~<TAETH=hAeb}&i8J{6@^}(7gGtKCx!W4MUJd<7WFSz~0C?~%AQEj+3AypN#vWEq
z@a6qvH>0TpGPL+p&p(XDZLJDgfP7&$``OwUEhGfjHi1%<0n6DaPUHG?VR5kua$%To
zE{p|mng&2_EyevPSr9f6m`RKe?IPr!!Z8)=HPbx>4P<&q+nvY2Q#BJ@8#=TIUPu`1
z&*p^WWP}e*#BII1m>^v~;>RZqjttTU$pB$M1qKJ}=jP@fB(r_g2hwT)?#T@yp+Of&
zosGd_ae+y>ty9uL_RT7-xf@J;I=T-mra`qg>E3D3Ak6^0G=xQ+l#=p;Y-D%O;)f#^
zf3q1{&2WKR!U;D+2d{pe+;UQsJ5gv85g4e4G!9!t8pj}aM@Z0+i^BsUpq-3l-`@Xt
zcQpy@j{(Gx2GDf}=^tb8`^S5W!pl}(NA!Fgfr*d22y`>?scURxhu(22Acho@OuNto
zfnuWdAVd54(=}vyPwUR(Aa_YZL179#EX)x0x3K4LHQ4WC;J&aT!-H@y3Czyev!m2^
zff(@Y%$*^@-7TBY(eJPxzAT2X?>Rxw1{omPAuX7Se|a-D0?_0T1=iz#m2$2zO`c&G
zW-*G(hAtynFeomg8l#z!!$eSlIfitTIs^~#P^Lu$jNl9i)bp}Tod^kthfYuwTWSRq
zB*;;m3Y+uD(J^5<2ZD7XT9$!Oc0F}V_IKGILP*;0Yk9xtxu5%e?<;AMK3j6x$&a%1
zC=fl!hfZ)0UL!}2-1UZ!e{5ODJ0$a+9UUDt5RN2%{vgQxU*F}M6MgUH>h(HFmXn?K
z;9yxXc4Xa4tM3w)uU})>vC*YHDJZuqL7U3PK8UyJ)1S-4R1M}~U%;owMRm)V9PExH
z+5C?@t=Tm~zARg?olvg?-32cdTN>?A>om=!?3M7}ljr-lX%}ZZaLZr+j|249cPVZ%
zNMi#}1vpe$S?Rw!VjRDb#BPb6ARUXs_5NP=jFR-~<Kn@U9LmTz$=Zo#J;k&XWXW3p
zL0h-q*#KOur9bfEbi`yMZD;2OqF;@+u*|%>h>1)l1T^`q+c_7gIMcaE-jtq+Cl)A`
zy&nc}l-+h>)qjtqwyn>0`$gd_Fap4m$yXRDW)K%Y!K|Q%NEj5%%I$M2;TFVJ^)0D;
zNh_+S{`Vy9U~rH#QGJ)}Co=31w~_l?&Vx<L1-szMVJ|FSj0wRl<c8~}ihh!2%_>2^
z<dgAlLpB4A$`q~5tP;cPo1TVu=pE%Hyi5Cz;ccrn|Fkv+6JL}%){<c@NfiU4&ctde
zw+@&S?$_*^Y93;U5>5*`$H-#U;b)PDf0d`O0Yf@O#mRd?kjm>mz_&}BiBr;MfGta9
z?7B9sFbb*iO={{v1x;wZb+BLOu`bnT^OHOo#Xw65(Qnk4@p+V1VNz(B0|fb)udZ4y
zgbN^8f(?|<)Cv-YEQ}{q808|CcnQeAv|nxHjvKmDE$S`7Jh!3H%%oE==X6mtLnxNS
zRo`<M=N=%Pg6=i9{gJ&(lE3Us$Zg2)f_FlvhfyZcYJsKgS`$aZgy+NIi>Luhx9Z1X
zzzKISe=Oqr3Hpi!Av+_=daV7vnh|5C;NeJ~1aXrqgRN<4VKb6-g~>p8dQT88_E%>J
zZrnu7)b^G$|40o!qpoU=u12FYrHuEk|6NN)BtjrYv(#x5W;d@)Y)eqR*b4h2FkOfY
znpQAI8pV;p!L6DUH@Hq8q%WZLX|nm>96_!ZC-$ffePYN*BFthn=@HyaM2eeK%^>%}
zsy}oOhd8(9tc;|H)jDYcQ^!Y3Ajgm(H9<kLXu}J%RHE(?O2*S@G^V!D**trDdplvJ
zcrYBzKvHdVb|ok)`-66)U77Z7gZmUJrcbb6^7C961`m>k<gBt}vt&im2)d9skK3y%
zCb?Y!D1xeruWTm_ge(_H;eCT9uczoHmpr{@4h{~^R514#1+~b3yx^i$Yv?a7Dmsvn
z@i1u!-T0U=R3p;1^Y9Y2x}d&(5#4SjK0z48$s)-s!7a)`8st)``s!#q|6Es0B!4Gw
z#$`pfU;{}|cUaUsGKy{%$sJ6TlTa%IX~DpdTHw!};uLjQMeLS_*EA`bR)ovx>qN#z
zVC5JNLtjMGY3IKfQvo=oZeLXfsE?jMf2nqe8&J*S8j&}T8X4wtu5F>prw$9!i3paN
z^VyA<DvJwU7EiXbn{&04pR02>*{6W+i}$`I`!X`}3m-|-%JRzU<Wz46C8FVcZmvRk
z^RlY5qoc_wG|go$^+)J!fVpbEZR`C649sc8r!w9fT10iyDVlOVvM)6!Q?vVMsq3i$
z<uq`yL81uH`R*M@s4ze>0yEymXyL(|FGx&wm2&VSO&2aK3*X*<3ah~r{#L@1yCHFv
zy<hwG|K<q4(=0ld!f_xw`{}ZUm>D+69Z$QsB!EHX(pcfqQpw=BSeQ%9N7iK<-Sp6P
z7Rr8_JV)0(ZspNyB|n9ETfm0lXFS6BlU`cu{H!79-v)R&`KSriv8~Vdh8DlB6qm+g
z;>11uvw6Q)J0)kzHnt>L>lwm<grYQbY_jU8pk|E#Jan@Y3%2)}!ewG?=EQV_@+c4C
z?DB=9?B2aT=lS=p(qS{Rvht^|DvzS|6T7x{1StkNfZgKYb?-QQDHDgXeKy|T+0fvS
z9?zvr%s(=$(pW7Whr&-I?X~=wpyW((PuL4MD0h;!My}20KUvXJy+UgH+0~cNL)U?@
z(^@oC-F0H0g`0XTRkAd;O-MGfGw_9O-Lw0b1|ay@*f#Lr1l1-<|B~p2z*;9ZD^wi$
z*Ud>T(paI|##p-Uq(-vieq1{<v@B<mt;70HcxKyb`p>w9{$aZbXV(<QHI6+eZ>0QK
L6;Lqm^DqAbAc;3z

diff --git a/_images/sphx_glr_plot_zero_weights_lasso_thumb.png b/_images/sphx_glr_plot_zero_weights_lasso_thumb.png
index 2c66c22c61318381ec398fb4f35563f3c8749255..1c402e97da4418a0f0ea8d650985afddb1ce2a44 100644
GIT binary patch
literal 17954
zcmeHvcQh9M`?m<CNcPAsv+UJv?-WWB*<^2JZy}qI>=A_!A)8ys-em8+_ukKSe}Cus
z{qsBL`Tu#&(>b5h=icvejn{QuuUA1Yl;m)6C~(lw&~W9SNWVlwyY>nGBfEhG&%D!0
zb3sFsS&^5Pc;%e5G2!A)yfATfHPrE#iK(>}85!p1ho+Bv#`;0^h4l5cmygV?LSAXP
zS&UVMIMMv9ahuw>sPZa1uwMIFVp<aT51qBhOoC*DPQ3ohC;#=clwzB`3kF;B&kyds
zd*a=_x3iq#{LXk0@xHUeZ84!@@g=XK)D1enaxHURQt=yfe@o=irQxyeeQ`g0$zz%S
z|KR`bGuZbY9P-DPykWNy$%aqohwoXTk1=21b;0)Y=g&Ny8j6CUu?(|G*R@CE)@8cw
zY>yr-kk%e!8W|ahq%6+2Mq5r)#$?Dv6un>SqJBO!KHh;%<8J$=y^_M*!ouZf!s#N#
z^I&!0cS*@km(W(4nNNkSPKh}&E-tS7b{+PxWhu4dUTp0-W=P}g(C+Hm+Q|ZkS8EU<
z6Z>h+K(5-J*VU1ivoLk-m-nK4e0+J@RbNI&M|BdH{E0|Oi}f3@_^rpEsHmuv)s~c%
z?G0+$IGp9>=i_-l7rXd8*BpWUO@Tc&#A(VS76Tjam)F%Lwc^{dF_*KAvJ0I>-OJrr
zwggf9bw{DKp#mljj;*4sjk0Tf!^43imQB`WD!RJQJv>DG{r%^cmN0ty`XrT=?^aY)
zAd^>HA{mHCNYG!rdDFTx*UZAoik?<d5>w-L^zzM{n}2?&WZ70t+@=vtDk|bWTpwvb
zdYt{z_^JMo1RhAeeanDGOia8!Rh#;@Iov_WwxqPQ$fTPZ_I^#!ei{G!_wQ(e$x4S?
zQ_odZi4!t+U{`mEh@>qoAD*0?%r7j=I9Ayql94B739=Fr*Rb*K{n>ErA02%vBZC<h
z7WPU_P1ja?;NL%3FuMH)Q9<=C4oQgr)YawD@3Jyo7o&ZG+<~V$I@AGoXb!2Q+uGYb
zYKP2A-cA3x791SB?OKDZG`F$I`uf#peccib2M5R0+&sOag4{bNhxHy_z*g0R2M_KE
zJNF!oz|K9ezpk#DD<vZuTceq-Ew8T|P9Z&zu51q<{xvU&N={aU^Q&^&GKOct0A}nF
zBV{%;aF60c66^v3vFGRK)=VVE#qlC;hsSMR`^(BxjyJWL2m|n_>KEP>zFT;j_nNJ>
zwKYlDC2D+J>mHt5*+zP4=@-@9*IcgqrbIF`Jzu4g9>2%R>X(<2!hmh=?!HS>T9Ro<
ztFBJ!>Qd6xuHJ298@A-v@jNkFcRX8ATQ)Mau*mrK&99?FL4lDFKH`gg=jF?nlA4;7
z&EZsYaI^mg2Mx-`PF9|KwJoK3DZ!<~u_7JJ_!B+O5{K5l!&g^V3&X)(+jbV7IBbC5
z+}+#r@DeNYs?cKRRh&ECnfu$+6crUEX=cXacDSy8b$L;)#SVr$Tx=3yLxU*OP9AgY
zGBY*R>rN5*{rmUyj<fsOMa4n2)0PeeB_+$FM@AQ?2cc9#U%q@Haon2RxYfonST<n(
z{rB(Ba1ICrLeOzNu0W^O*_|`7^YfQ4@3yBU;HXqpRl!%_mTXRz`g?nOH@0l5cd$xI
zO7heT>z%6CSuTt`&R9>k8}5)les_OoXUB4~IzBuc&yt_gu_`%$+BGmeou1!jvICB_
zwY!_?;lmE6U&&%#qBgY`{0ws8-^$D5QqSj^z>rkb)mzTCYP<Kk#h%;Ra$1)qg15dO
z(sojQ{kkd92HDD1P+c!a;~{Nt&r9RJeJzBP!}w%(!O+ME-tbXps-z8$;~3@Zzw{e8
z$!_QA)~1Y(Y9u5i9CtY(8xA+dde;u}G)jU^QawnDM-&VV@6&jlb1^9;>Gge+mzI+&
zww^$wc%GW2Sun6Zd^o?q(hpu59UbkflqIFAYLCz^J5W(mYtBmb4DMf<pNkM%^ueaq
z+nH_T<>#M$SK)j-6JQ3G{p;5!FkRh)Qz07PI&pn{y0fD#6Np#B$J4&&7h7IT2t>lg
z+0pxMk)s!|<=zZA7x?6x8dpBCvytxOv8hX0V`Ij%!;QDCQ4h$%WX;XZhku(9a#U}l
zqdc$nxRHaCB5ZxMB&2N0Gm%9l^B1`KM?9)F_@+XQpWv-~D_N-oq@;3IR*%}hJZdoI
zJuM9}YX5SVfZ#(qGb-Rce!uuaMWrG2>fFThAXm50?a2CrudmD1#nH)N8M0B+rl#=|
zofPxKhx=BfuXAH!bzFmf|K`<VCxunb4Hv#u|M_~eYD=`fzTO!=bMwv}CIJD`kj4*n
zded-wCSdgniTq)I|4P3u(COsq7WU1_iR(1Jhlj0G5WCPG`?h!_J1=h@qR_(MAi~>J
zh|jeb8#WKMD?dj?5g_4z&jwxIydf8I+Wc!!2$zYK;=U`tS-q#k%F24<)~&a2<Y1vM
zb#%JmMfBH4IXMJGM1McxQkGv`p1JIG3Iv6O%=JGPTR5A#5*-*AKn2G7dP&doIhy{-
z$<15l8NDqKs!3CvXK#w0%o7?I7?kgI2|)-vZ?8SSRczduo}C?-dbyuvIbMDjVjIN4
zJHp#&wPzDsaa?b)FflQ~WW{zibRl-5mi_4yemJ#Aa5}T0vGGf|Q@5>}=G+2go9@-A
z8UZ2U$S-}2qw)85wyO8=e*E|`Az4{jnJ{&}C=g1)7aSOPW76}?3Jy0WCPsL#g9mj%
zEiF$5^R?T5mQ4l^tu<Qm@AW<7hy+V)HLTxoEVi4I{lcz|BMxz-_I!b(V&RMC@x+#A
zb4NVSW4qbwC=02iqA4maK53x2`~wLU6(iBg#^2piR8&M|Im1>0-N%neZlUa-KQZj5
zW*J1*&Iqsbw`psBiwm*3izw>`m%UBS{cY*MwNX<RZRNmq%7=%UIk1`9EmS6iMET`Q
zu%>Y^<MaJ2u?HL+5quVd5O#Xj4!{HnNlBaG20Gz>f8^xU8x-~o4$4C4hE#;4np}9t
zrC)y?;@XHw>H}fnB<(84ugo}L98^M%oJMU#DQ=sOvGDFKSB*Y)*}c!g!tzkFG*mvC
zF)}u`x!d!QXR~sR{c<P5>(|?6JYL0ywS&IiUI`^7Jk%B$WJCYV?M&Mv<exm*w&GoC
z>+9)36SuZzXOF!lquf<xO(aui%3I{Pp^>SS{8U4OqQ?O%GbhI;*<+_8vjD=aYy@>@
zkx@HmyaL>Q<j0R#_6VxZ&d%SZrG0&U9~uq5LdfKL+q4a9SB7j2JG86NLMxxJmJPob
z)W2fbpQ&hQXsE{>@0~y81-M|Z{uajzRn^9-Eo4ahLcJd63vKPNRzri8zVG9%Yd=xR
zCb3i2!h&sMV}po@$bO8yDV!>#)o^wufW~;a=d1haN|tHD$8_fP$(qExyoV{St1KUH
zlh0P6|H#h%W6Ha`-~XJ)V(=l}J;6WVeciK-A+4RAMoV4EkO$KGS4?eeOoo5ycO>wI
zC}+ri9a<|Kv6PmTEwUWp))mS8{+-LXgN*TcN=K$*VkKWir4~CX`A*mQp1}7)rd%oQ
zB9^Js<NW1%V7}Ou{7g$L$%ZOrtn4<HugwHjLQW1ZAD=#dq6@4Nk}f}9fG^}hlxG>d
zeQPIpFO-by12(lYT6C+QPJzklzaQ?qtxRBpJXWK;X}S3-FJC%uRxHIme$Oc6xURQ0
z_|v&Y=auSraFBohUZJ*HR8$1PVqd$2TGXAaY;0q!R+#z@z???wvc1_54ue8<LHA?M
z;B>h0ky0zly~WPClcm(7pHj=_2Q96wGj3C7)C~c5vLHdzTrA6t+xIJ4j{a^P;$5^A
z3=9m^TOTg$h~x5w_y&2Biz4hk6O+>|kFv4O&nzm=t3R?(0SB(iFCd`5r>E!VIixjB
z>k9v+6u0ojMdQrpsa;*kLXT~yQU2AgHt6x>$&Im6D+Yf4FE<Gp)0m%@k;J5>rL|Au
zomzQIOH13X{$s(T7Jg!3k-HU8Ibl5|alYJ}u|8H70f7LnwSIf1!DabtKsh+IyNTrz
z5<=P*kK?*}dV0DoV6<oIs;Vc;84-f-=5MHEDoAAYM}GOz1i8xPe5aXQ%u^VG6CgUU
z%$2bpgaia|cHdy@nVF1$P27$)ABc&i?(f@!DaQ%g(MO0~3XY77KyXF~3CTZyzVtPK
zdY0zu7#juUv^6l0nK7^nINBX&VdSpY$z({TS>BuCk24Jc(Vsp|TlhwVg*6%%+pQ0C
z=|)R<?JQtSd0)eL^6c5$uB6BAOF~mK84;oe(|>&KGcv+W#qn8?6x1A26c!dr%gD^w
z3ij~gX7vYwVT_mCp>Wl><CqolBqyg0kYH<jyHr*`m;13T>R3FECYsMr_B?!CH>V4Y
zJ1M{{0wW@tXKas4!L5gjjBZg1Sff&kOh0d5e}6o;5zf@brc=DA2g0N)X$FEIq>NL%
zCS_%1UIBskr~77rj)TI(Q9yEjaS@d+pFYJgGcy~vYo>DFOPu<<VsF5oh{`09a?~hP
zh|0KtzNh%N4qPG6BDv@1=eLL7M%Y2Z$;e^NQc5;BKi*j%uZaFC72v43L>&=;LX{4y
zuOy|UJdPXyy_s5C&I6SDH!#3uJN?Id=%Rvg3~~n~Mz}TbefRIxIt8Rv6H0b=TzCN(
z0S$Gb<U9Zy?m#dCfP#VYW^?mdFqkBSLoi$;1j#25_ohw<%i@GxinhG7nQ<WdW&j3(
zd($f$>*d9bZTD}U13TI2m7`Gx{1SMV;RYTaqlidyJderl@mvIt!QUI<R6<Rc7iS)d
zrf5MyK}n(>34MJpH8eE=G9h;%t3@RwwEZ^gKb|~7%0bLbPEH2+N%BbJF9Zsgjglc-
z!Fv$MGRG<rq*_?M*Oo`8o+k-Pz@M@?NdE9%D2d)Sh0egzVfV>SwNr3oh~#jHNPsGR
zGR??tW6{&o4_7&HK>4xFM|kwA)~Y^1z=pa#mV*hh4mpqU^nq(oXsC2QZ+$}p3p;z0
zc}bDSnIjUCvH^FzdZ7VkV`C#KA3>zMm8{eZxLf||)7d&~8bdgeP->Bc(9qkEW*MV{
zpqv5+kOUyf_BwC6u$g*JXzBzH4Cp==S0oghpMryNO06f_K49K*H+lHx&6|!?vDBkU
zw}^`OOQ>+WcQ|Gf9T(S<E)zn=XHEoBi{atJP>6f)-o1OGsQ5i6Cs;9&pUdsg66+3i
z$KX#*kJ!)**)Vz`p+vwT*0mRoQPI&{WVgWy?h+FATU~&A>_RZ(HtvYCP7V0@aTh#%
z7vev+*QGnWwLr=LK>XD!B9$y9Og%8NjueqZTuT0509a80$oGzDO8|}M69WT!$bPi2
zhmm3v9;Zz$G>GXRA=RAjub_oe3WWdt``0>lJSQjJj3SZWY8DK_2z;~p;&2R=>M(DS
z*!C+5+`V_N18x<TcbkGY5Ps%UN{W)BBY$LMBm_YouxE%BK@ky4HAj<k5S*6cO;Yrs
zxBvs)Yhj2u?&4nRPTd;oB_t$71u_(%Z2~iPo(*DBN4h_T<A+R0S)=pZ%}oeS?k!ZX
zkR;FnE32pwQu15UGBE`}wO(Y_&j?Qbf(WxlryGzBgl6}vvndpkUOsT8_SjcOWg|e!
z(M-=5;iwoNKFk|Kbhor@^ToO=Lum_*0vc4Ea7Xl*7S?_(*J&9sLm}eI3n}XJXCi91
zbuJY61B-xHk-5QuFW@p7;LeB3?HC$9;xasb9N*LP0!le|hdamNn5eMM!xO#IpB0^y
z1V(0Y0|&>TH(h4jeXmn|2<foNS_6J3v{g+AD~9X=o++zLw7a`|<K|8M#$e)yZ;Jc?
z8OA!S4MxVtx7PdL0x&oOM-A3BrS=gV_Jy)?9SS>wT{%NOeFI$$sQ|o%{HjfQSW9Q3
z00iAX6HZf|iUJeY|A5azG%$cF!QRm^DmHfM)|H>%^`<b&0Hv(AiyiT33Ms<zu&dtj
zaRn$tpqujG(W9_mzc>NV07z-j<NOp8BV%s9s;0pX$>F)k0ny6Jil3jK{n;4czph|n
zHZp`gGa-{wb)Kv0Z*h?>0zDLjm6w-?`^YK?p+EJWuwDYVXguE%ISvsT**#RC8_{Sm
zLpgOUOY=oZt8H_l3c$$sqM|4$HkQDCbob9tiGHprta`ui8I(6^x!PB1wWq|jC!IF=
zvXN&s=jZN|?t6-!mwTyR3w}H=@<-4L)gLa~BcMb%t7J~U0|)^M0Wx0G7iHZWjzRz%
zAu%Ia;+7*14;^!Ia#~timX);u_ur+UcnT@EtXoq<BY(86s8;BRNZsB2TvgM!026vV
z(No>^XLI8<fhmVG_%*2Wh>Kg+yy|ff`<kDhuXe_)1&|(oSeW{G8KqB=_Vw2`16LJH
zx4I|x%0It9ibIy)mQjA8dXQu7bTM)=sBClg?&7FT1ixUY9ne2TPcxd#l!_{<m9@KG
z+|8}`2W@I8>v`(p#nF_>zij0q9^K||mo}D5xmP+j@fG{YIQ6`L3r(R_>ujk}AG}8-
z11^!;l$$_9y-7F!O2;kqh@vKjo;|vG?xc;7Iyn9KBAC?9YA5S*ERUT6LFL1qP>zGt
zy1FQ(3B;crJ#tZI*wYzl9lsi>t2%4r5b-I`eBynB4&C@9pF5bM<DbK@i<KP9>#T8x
zij@Dnz{u0(m&5E_O|eq=_4>-fr}rW&Wg@-)_(RC8DSp2`Mji=7jF%=SdX4L)Q<b)&
z+tUQ~WK9_*bE1%}%%c!)?Oas<_8Q?`13d-v`wh|b<YZKFUCBBHO&{k<J`22gBgh}`
z!$qp2WK|Vr@|yZ<^+>SNPg?H3f70z(uCWY-O@A1Q1`xbHQB|h7^ymH7udNkvA;}vv
zhvwzNK5N4jh_c}3Dbc^(Vu&u4E$XJ(@@+zWDt>{VubJp15L!AMd{#qCg?ZU^9oPZ6
zKk&A<s`R{}QWZU(#sEl*X<s9900kAQv?FGFeT;P4e3ZCLwiVX8cJOqmkW$?%`>J}?
zQHwh4MMg^hfw9I<QqKo2uO-Q*QX+kFSjWGJc)R5{<VL8JP>o-WQd3{8(NSOSd9Dhf
z3#bocVnCIGN-L)y>I@3g%qSv64#`kGo^D8kb-N4g90>F;<)SV%I)eqe-J>NI3j=fl
zX_qgrj>4bOF|ZRQqaQt2c)+q-zt!CT1Bau7sx7pxu=->}{=SUA)^^oUP6P4JSEcG)
z<OfMP8)?5L@BK2;-lh&5+sK5j8UROVa=44&M?ei)r?>6ENDfgBGQo71z~l{eI+?3)
zIjqpJ6aeo=_Rjqc%c%W#b-ZtyQe$<)DD1UPwT0j+CYHWX^W4CVz8_@vLra}|t>lG<
z5ByXnE%miDh!~>SFEi)Z?CRJ)H8sgX^oPPDt-PEptG^W>*+_{6Dda|*KREaO@MGJv
z)q-yE=x_^n^?I+B40t?|B&)4);?-?enx1%|0xw}%z9FqfM`h*!v2*>h?w2YmwcOs3
z_>%j4iWpl52{$|gad7GR5PoDTFT`URYFz#BXAyG)yuy=zuyd{lLRYyKWvhPMA^*n*
z*M66Mi!XCb{hHo5IwW&9eW-P9-2cWWsw_&E^-hf6W7-M|5sMdFiVUp0zo@GG$`j^s
zqyy0&znpv;Fd-AC{9Iq#<d5#pM}G$xuhC0;cTmMYc`YeuUC@+t7lZ5`&ESv>LAshe
z;TQaC-1Lt>kZw_QWi_uIZ(+aVTy^l!zE9^Gqvdnsu6QP`NKj-Q0rhO9eJ|Q$0j!qC
z37PsO8b(hv>0Fo1YyK$(B5ms!ABHs6gN<+{c0SapjssW+E$9}2*cB&>iSc|EU$&|c
z$#|+UA@P)|FFpysv1f=rt+mtM&h~;@Vs>`6d@a9f80jH|=2}A?(JW<Rl4bK`3_2-;
zd2gh&wI@_rel3gfD!v`c*GAP|+}}2EpRDR?af(;#Z~y{JOQW4?;sUUs<Gv$<M}?4v
zwkH}F7Z)^}??t!TJ3M`O8Ust>=H_;nx)(wj^6x5a<Ab6)R1NNzCrbeOK0<FUeq*eR
z@b29@fRIpj(;^TQ-pk9T=@}U(=X>1+8ee6u(wK3eVg9YACb_Mx4apu}-uCUX5!{BJ
z{)0)13##3!u1*G+*Bc5ks2r=3?V#_d8~vpmDvuN-`-f$zmsbM$A%_?@ad37gXn82Y
zCN6bhaulXWolF2Rg7oHp1`UyDcTo>qNWXwVy_aLvIEe_$LOWhbZXk~(nP`5nN`H;s
z5C0-gt^asqQBss`dDSgTMfJ6)C}s)GeoaLthT(HN!o=tQxEyRB?#Vvr9n0;{HFLKu
zPdHz7+$4{!ANEnMsrI{+$Ka<bFw&LhaSBy1bbYO*ODZ2onV_TE^8<(4W~lZo#DZV^
zDMjbQ3k&v=4adu(tWecwS==7BThz<^RJn62N8!`<ZpZV^RBI_I5|Yxk1?(AOqndqx
z3B5WSXSaVH+Vateh)6XR;ipGr2iHH}x~HU?8#BzwAXf9<xj@UqG(lD|G1>111D5vD
z*bMCQu~u(Wr_?MNH|_a{y>erWZ}|-p#Unv1>JcNG=|zn=zHUzBCJ7ya6V#Hv`w;id
zN(F5K(SBGK8rYR{k>m0vCjWB~gy6j_#Sq01_n0^3BqN$E@}6DeUU7x+xwC_4YEs@?
zcURu4vuVNE`pVPe1=Z4>9?CPb<=;HKbV+>Am6#r|&R6eM7^l`#2yh0=%tiHfg&!W4
ztY51^hE_Faa+<xkjZoF!`ON$A{x}(SZ-LRB=`Z>>y_;$XUy?cb@=o0BO?9cR5#aq3
z8O%Q^_R4=`Mt^SyQJPXyi*8$;ncSo#G*gGwmhk8^zXrQts>LDBecmU{-CGnA0TYdy
z5sHjYrq15Pkr1}sdLK}@U`SPqG_fz}uB{EeBe2_l(%|yegr(NYRQO_AC%Igy<c;?7
zEB|f1D?Ayi*Q8j<eFdAhv~K41`%Wy-H4zP~%`^E?262&_7FKd%u2gC3kAEiaj{C=q
zvxbO&d_TPRM<i)dH=i5kJ0;Uk?6fm>;W?7dzrHq49Xa#i#gy2lXENv+RZ<XPD##}L
z1>A^pkh3Via1x3?YEyDdvfg()mGK(yG{M9r+xI9gtQ4WZ?)v@Wg`3gYd1UZwlJm>=
zi`CV>wVRLDlIyj_A^lukh~xPhNVz!dI2*gqT-v!>GQ;MG-pk#d-q=#0W1X*WuC+jj
z^V0Tg+5Y>TpH=9pkuqh?t~W#DR6Ko@x_9~2;t2P90?Dxho3)lwM68%K+oD3W|A$JY
zOB=VFvBec0cBj!LROcId#HTv<{&lVHzZlnRTBz6;WUAHh`XE!VX<SF+xMQaiHr1|$
zT)QpDHul51crU=}f&H*{Y|1hY1X6uS_qX0f)xJw7qxi-oI{w78r!l|)nTheOwh%F%
z*PScT|7YE{cr*KoGP>N+rJBO3G(Y)D7U{kAt76-^WRIU?Mb&>k8=t5$1?y=TUPYe6
zcNL<jj1S#PcMA0-dKzvZj;{Cnrx==Z3mq>~p2eN*;-zRshBVHIZXKDAT2@!pekcjZ
z?PbA%>r4zk6Z<uprcQ_;7S`ctRj%`+tV%u6Rk9F&j$GSy=5T&4Bem8Ww1-Ja&8Ik5
z+%1|>Q4#Ndi`0OZVr6+*Z=%uxCJIos|9ck8+4G86@un?S4y(qUZcE89*Gyc+G<)60
zh?<o61_E@=Qt_9nzu8P{%}WJYtcxHKpqDkbuON7&u>E=s9D4QNb0PRXw2@E(K#?M-
zXEEa?wEh=OP1+WefHy_{LSLK7cTUj{Yvu9liLSKa7l1rQp-6toU{)vyo*9))B%vBY
zjN77gJ0pAY!`RvbkwIw*Z6^81y8C{Vtp(OhbmLC~CFIPazpJ%76_sLO9$$&F%-ee~
zMRa!<E_-vnsOp+k`WHV#zwj(f_$o0!(aJh_wbPL4BO&p5H&1;Z38j^fh8Hf$xb>Zk
zbNAV$<Qk)MPtBX;JNkaulr2i(tA|bn>Dy~(@3MJ>$dd_LYrEQ;FziP~hz6aaD5xZ*
z0(@IuV8nCjdZ0haVy)gkxMwbutZ<J=>ZKtk-quOkP`&Ieyx3S%6?}&c{+Ah2Ju42s
z9}6e_*res@lCJ)-rst8_*yM)II;=YSp*QXA#EvU<?uwzT?C*lhdfx*gzBreaIl`pu
z723aP7^)A3vk6G=N_r|XF+??RxcF8ytb9)^^#zkj=)kk4NmkI$z)b9K;3>nKu~ik~
zj}1_4UL%Vp2%5k^)P#>JF%4MpM=?lcTNg(WkxLejNMsjU1o3BK$7ZH2+Y45I(jAa}
zY1pDx#-WO-+K78EcawJgS%1Ip+ciwDx6ov7j-WAtIvF+E0z*Nyj-e(WDSV4fPELNC
ztmj}pS}9BDD$p>Q%K2UwwQfRA@P{1cayL%l%yeewF-x*B&23)3184HeViPG5mjnrA
zVpHq-p25(p@8la%{mkd?8t4b#RRLqP2oyEg3h*6dD^N^t%n*_U_YM4L+2~zveG{9*
zoQ=Tzoj>LAqfnDbM#93p(X{L^TNt5z)IPpz%r@omp2i{OK;cP5{mkLJqD)^2JH}Bf
z_6|n@0i#Ac2%!AAA)$hV3!SeOWU1a_%NFLA2jTmMLrVD`X8aNN1{3|?KXpYUFVm%L
z-FTstQ0NZ*n$xLd6cW?nOb98ASI^b>;aNBD+>ZG-Y5$jfhK!H7{oo6$LJ|XUZ_y8@
zWu!N@9sizW)3tVC!%7YwSdi`4TWEUyrK971{px&%RkQQTTr}DKs!Ka5jq$wG6j^)A
zKW1T*L`8dq|EZ%;_fLeRr>YQPf5ypFK+|0_?xKS7Aex_D4XgXtEiVQvWG{^-Z=_T{
zSSwns9hseX(hOd<w45=t5>bk1l8UEu*2K;JQ<*W4hfr_*-J#y(CwSso896kxo>ktf
zP}M0T?;B61g1^B}`c(pPd4=7?&BcIKJvuVy`%x#0^M%h%eQIh-<5_7%Y`&k4)sG^x
zamqz>_Fbt`sS6gSL=oYcKCRhVjkAW58oXdzG|1AJ@!i)S@ovn8$uYglDUJ5hj+V=s
zP=qG5t#cp_*Qc*<-V}e2x*B%I@l|RxePLEyxouAIBlyGEcm034?vI3gepV@M_YMjp
z-Bhr;C;2}wDhZ9@P_~BhaCyGqHPt>sKy~K+t~%V;=+WPg<XmGD(*6`=M3ikqSk1f5
z17|I{mQt#f3H|(J_NW(Z-tF}vam>dyv<2!ODWyyLa*j=q9k5bN+q#j(k|;`+M^2p|
zyuVD*ls7QE;pJ?4Pl}MiJO5zWs*qrGq%Al1sL*|gyuM3kM{u0FDN{C*ce-An+oKt0
z16?k9qROeiudg0vekb$M&((_zZ(i+*T{ZN5e}<YCqPnX91n-G?qDDB%FyyGcI;(X`
zY`*lmz_dFczK)5leJYscJ+hfOu}J>_@l8%GLr1Z}(uvS6GErDpM^-9;swRhkaIl#B
zv9^vDHAkkNM?~AV9Gql+(wVx`l!*-GwEPLBUlI>ke`6b^vt4J^$otm(QKv{MGTnP(
zVaC3OJVVE5Sq6dW5!t^eF5p1no1!(ku@}Rp#tjn@=-l-1hCv4lnyuF|72@lmTDj)D
zB>te*Wsc^2dEYh3S|H8fU#_5tjy<@ny+{D{H#Nv2>=7R&vZe<r`6u;!@f0y41`jIQ
zGiDCV$#7Y%BR;0v`{z$A$P}<HJK(T84M|{DdfNQ@x|X(hXiB!Pt-p3?ex34PF&z$g
z!Ni7y#7fDoOO3W7hK?cwdq+!L@$+BvswMqqPG4EuT-ulH-N^6JeC$nQ;A6cWF!n5W
zAV?z1Ham|vL!xn^_esdj2M-W_IV6-Zk4SDK_-NNSUS(~&9tztt$QCV(kdt+bX7@#(
z3J8rqJf5Rwv^r&L=qS0Z6ihGD<=32{JZShebK>K~f&>P)fns)!q`Ys%nzUTw(40n=
z_gJnz?{XY3Glj3x_kr2_@GI%AJS%oz)$dtL-bRd@ti4htULCdkmWn$t%1^mbn$J{I
zc{y3qMaba$!>a6U&kw-~9Rm7L0HOdhb+Xb_h>D`=Zw%)C9@BbNI{FJCl8u8rajwx`
za~AiDlqh7((XjgWd!*}&?ps@TGAn*2hYd%`5BB%{u;1eJ*=3IOSxvarXrKLP&b>=)
zPD>{o_n{K+!@l+yp$P^TL%zlEHjg3+>T#jc#nS!~%(_ReaG!@ygTu0Fuk{-K<FR|e
zix*3p*e)e0&1pr#<f3X-UTeFXgQgP@{U%~kaZqKdYWngHX$88xqN35!<^+uFVKCbO
zgL>4|qv4L&HzYE}&o6&59yls6KX6$}c0zJTh#7UyefOers8eFrErpPYd%}0VRP)w#
zG->3(f@7|V#+rqmT{tix^BcA?x#wrIg-rA!!UpJbe3y#`cyxX~(Qv-f8-&E@^>^kA
zgD}w@>_L2eP4#jO$u%x!r`O%)7$nC)ETa$OlE2N(;A*pH!b~jpYzzp*xJ4v26qv0e
zWUT^{mtSBw#*$`@NkI_`xy(}Oj`CM5i0>clS<Se<5(y;q=Hk#vxN7`V!b-Aw$7)eX
za5t|x*Qap?Iy%z112VY-^1pwqJ-n@FTOHKLcnfTiepo!WQC>RG=a`eBG0a%f*M8Z7
zb>lTD##>3(pYA3JkNyT=7Sl4_tQvdP_4n0BqedSrue0OpwO4s@dcK?03%B@}kLPB0
z&Noc?t~|W(zK4+vZWic`tk+eaHcz}N8zGA%>Y5K#x27ftQ1jKz6AsP1`|H5O!R5qS
z<Gw`Yo`8{7_GLq;s}9oZwwJtyti`JvgeFF_SSh*=ejEN_%um(yMzyGQyqwW-lks%U
zx7F3QdmX&-26F|+b?7B46J`mvDC89r!(y;mkIMkxTO5(=Xbby%qSY>ASc;du#l}y%
zJ(@vyvJtE`+YnY=SWFa_Pl%P<-w8Ng$5ZcNt%pD?Kw!Wn@}a!k7N)`l)@*f)yW0ml
zEW*1j(&N+|Q<{;_6io5A)*m&eZB?F@Aw(3Yi5iN_@bq83yU$Yib>35_t=7Z#_6xEo
zd|9oy$n=j&qurBt`g2*Mc?^!Lvm<GfM;-1+_CrGe86AFw6brbv!X>=3YMs!bL{MMl
z8k8lCy?5F`$db1;FE)wrZs>R5Hgs~~TCG}i$aiR5+_$~*dxi9zBD<XHJ{!^QQ*e)q
z%zlC4<!X<v`p9uNckFn`r9pnr^<*wsOH@#6qJ+I$=?a$beij)ND=ZE5J!%XF0~9MO
zD-`divJyeUuGIoGn0SN(D-`(JJUwe%@d~;Xp_ngT5z7n>hz=75D+Q$YJC)REEA5>v
zjBE9+DDV4dH!Z}<52<ES=VztSiXeQcGj{sabg1}AHJ=M)tUAO^=!ibDoV{#a5#bm|
z@l_-xD^qFbZ-*Kz%=HVt9BG5P)ULc{U97+3?>^b;=mxLmy>Wy)T650W51TR`I*|kh
zVynM#A0SmVmVUshm%4pB^yb2JJV{fQ%w2K=JxzTEhG+&`*Xqx_>g86>=&DxyIOale
zNp~AoDwV4LbkNzYKQe1t+2Q7X9ED+`L3YoEo<}tF5!0GHM9+un&vjHa3u0|vJ-r@{
z-Ds~Xk8`d!c|#U2R6@B+7ug|5$Wt_Xd=^R3v+@G^yv&C8{q}~w_Y2x)`j0fsj=1Jr
zW5PL7e9}lnAT5=}ptsh&xof+${4HeilZfsGMkp0o;}$i=*4dWRk=n0KbrHr4YDM3U
zzc;|@yQ`Ck?pbqWCFXq=(2N&j45Ce*70M{97|#>c>MRf-ec5lE)jzYi;}j7Uga2_d
zZ3}U5DS1Maj2D~vVrV9<y0b2CCaC&y*zh6Vott-MyLdI?_#URfg4=6_V!A^p$O&kg
z*fc|)gxpy=XlCp;uzLsV%0I)<*=A-HtA2mI|D$6vuiSOYZ1WN>r+Ks}hN~0fU^!#M
zC^D<`hD}CM;lKkI$Q7Z}H5kOV$?;_kZSng9npDPCzWr{kuTHve-4@KmFewc#)`E}m
zUiP+l0c)N*BAFrPrvlAAIs0s%Kzxj^SX3;9jFWE=$wiE};&&K>DLW1S#%Eg<6WWzl
zN#u66CTBzi-r(C_Eo0nU2||2ilKHi1imRk3)9YlV%dCAQ@9B0WXpH;(KH`4%*uMFl
za5>VNZXVZtF`;h%04F{V)HGQ4WKw4AAIH+07e7s`PSrBSL{nW|j7C+)I|_67Cx_p(
zm6_Z=ZF93kGS_FD9QFvlUeBeKE_kQuxHUB_rh5_bJjB3V<1^68CPF(;B(uI<(`A0<
zI4hPoh~!#HReO?AYfE@E{y_LoM=^)9LKvRDx$Mw2H|ajHS~n-9qOGZhTFx!GtP37t
zrp&9onBf|2B*!h>KD(`BFCG1q0(a_y!|tW5+Fu-@@<E0QPt__W#nj~?iMP)gbp?7M
zJ?K8IZM|f7c#ca;uRb-9vSmw1X>53sLD-5oGY;sXrQ#6#<eTYnu<{k4?vkDvkE|~q
zCTDh!Woh(1^3cm?x}}2S=f-m^jzs*NS25jY%l5Y4(*4DRc~h_%GtoKSOslufqzLQD
z67ZXNwy2Dq<O$E7CY)@Yy;m20un{7L*O{x2)3<YAM`}aYlZylWEOgu7Ik?F<S0@L(
z!rXje4GTZ3tG@>lyZsb?EEHn5q8zFaUag1S!m%Z$YrpqN(3-kW)k{EoRwj!jvd-P@
zls0W`vOl)$kvHU{5T>kTFRWPG&O_;1<$>QL&NCOntBCdL;~Bfz`VYsa$Nz45>Z%fZ
z)|m3b52HN!oookYE!%0?orqV(fH>FL?*PoJ)!;pS`7`vQX@i0Ar13-KJ!XrU$(<j_
z?c?-4?G^PT>nS#2KqLSbP(j1$HB%=^z}AH#+FAGS?Yc<w54ea`A*y8}%#i+u!+I<d
z?dj<$V~4`3a`K50;}R=>Tw&gYT<rdoLrft#ufvoO9aXYuibjsXWayBEXVz6#Q&1;z
zEf*~g>B@_JvQWs%Sd+j|y;e5#%p+hfs^9PWN*s@Yw_h`#bwi;B!it@U2>N~*4J5<i
z#x*QY3U=ooy*XbPi+*EEt|l;5>u|#<T*=_6?Ck;(?v#V>ap^c^IqNrUQ_dbV4{(jA
zoCxc~{IJ&w5hB_GQ8D@5B=cIwZFDgHs4r|~4wejM$!<QP<A6}M+Z$;v%4VuJ;Ddwq
zGF{TPvWmJzJ{)J$D~#5&awcE=bWZ5KeNNw;CgYnO{f9M=qgAolAm7%B#Fpog3`M+=
znj4AJ4QbfWZ<<!GouYAgM9xi{MEoohAT70si1t)#vr9?ZpUAe?{PRP*>sC@|WHAJ-
zxgpJa5z)Bgy!XR>dg-99XsK>V{6>talO*;5^KI14JN%r6e`0Dm9W*5{00odG=WNfF
zG?Vdg2B&h6NxXYCbP=^R{@HG#nw~jLGcAXi)GD=hv@PMGdy=jcwA7PYTdWvduW*sE
zYtW)dj2I}K)O==`?UNYr7e;~z)tTnKBA4?;g>=c=NG6WI2R6_u?0YL^R#JJa;aHEW
zdbC8X7Quzd#fkU~^{5}kz9Ex?*F%rZ7U{)sh3Y()wO9T-y5%&E-`Cl@N1DG^QPV)i
z{u@Af(wf}hW<`D>^9TL!tZl!kBNY2hKO4?2#c{WrHy%J*`k<rgY}O#3i5uQ%xIR`~
zu2xE7_pJyw8G0WFS#js~59>ZY3$V8Io*HC1rIAsNa|t30UDJk)FM#Cp!96JneHw1q
zW%RLez;<o>+Qazxx|As6%`SOkOY>K{q%3;q@AmE8s?VplT-MpF7ENU+3xoy3tC|n1
zCwN}$n+P2zogKNjb?X`-xsx}>2kr#lm#+PiOeWzt8*?Xzh*@7&q_)CIp$so$DA)V?
zlaR*NB^kx#SH#5)or51vJ@3+XyFUt1stPeKj?Sw6;?OFfsDs<xO&;8^@V$R?w&Jmr
zhlnT4?@HH1glWaE5fB<19{zHM{;It5FXY(Xst3zsRzETK_vhbhor%9UCpd3J-t^GE
z4ayA5Yd7(0a<+Ak7HqrAiDv6V=UYBy1EYW9<}lTKPcfAz2sN})ztw5~Z`sBymR-J0
zN`5T$(P!wv2U2?HF#lG+XDO=O9w28v{4VFRZOP@U7|d>g6aH&#1KADi5*B9WG~{I)
z=|h<DtlsQn$s5<LYr7yO;&a?>s9jQ-(y!h3BeknbT+UT1$jr>d0#--#bI~BBthr>T
z$vW$@V%r(XDcjN>zHdg)C#g8{pJPegk@F)Odhp}QB4A|Z&_S!TN?33rmRj`Oz+LVF
zXw`?C6YY~VZa_uO0_O3S*VP#i<EKpD!i=`bklSv)<?a+28jOxDr;y2ui;HZmf#ru#
z<am0D0*iO6Hx7UHl>WfW%FczCI!XAGThn*W`*cY(lcJ$MTwm3ApHC470>#gu(B30^
zD5HNR8>o*f_UL-Rd;)?S(6wfUYA^XvRM5J*I%X(2b1r5}XlGYdR8{juZS-DV4dF#(
zMiP<NUx^4`OYEPctUr8ujDDYuEd+RfQ#3FDe*-iZ_@25M+ZP%d!7zOXx)==O1E#0*
z)d-oGn1Gp}AR$plOe(7dZ6bY`GH2&Jb#ZagC6&`Mg6~gxzlFXYw(T7<jsZhL`!YH!
z6_qO4nnd7ks@+=HnLf~YqnS|dSfMr4^+$FH+7}zKZ-<g1VCeHKU&Mdegcmn_==b~I
zMak7|r7*?8C%q5jF)@N~LxnJIBNY=?(ov<1sB!QvGPckqeHOSEg?2ZcMwyt|oU`BD
zj^LXb6ptzctk>vInP6UXdN)>X&SK<dM@z7jD$!!tQ^UOKPK|x^n$a_<(4!6Ke90TB
z`CMIm+;a-mN5`+A_y>4#BTAy_FEOIuQ2#2P1fKdE5OJTO>mS=*%p;XR&vfVt*1Phs
z9mW!VQzQRE#y&%rj~ez2oJQaKVY)C^O>yaQAzxkdZ%3t6iSE=#sVdQAQb<zSgGuac
znrv&cV0n!`Xw47Hept)7Bp)dREu)%)Ni?&1f^?eYUsZsnzN!!>v2!%yE+W~Z%9WOh
z{Jr9NKI2tUdc)tphjf<MQ3vvgh`a-U$?VSwqubAJFV)FooK-%VMQ}5)g+pGr-jmyz
zPCzTUkCBtUogq|oviRAaUh1|Vc6&aHs~YWQR_kk4X)$QB_6^D4r#H_tR{oEvC%}7F
zHnx%9X3;j{EXx$H%z;qxIVQ&8EdMs6!&{gS7HW4N6tLptq^AvO*VgVl{jYgnqJYgN
z|ErB0)m%`IRKwI3*?UWyofPO?FlGA_`+8r<v1+31Q6OdafzB0`BWaj1HfBk(hxTXB
zR#IBI2x_B+wwOGvil|4LrT@hWax5zjftm1s*x(Hf9&_JFNl1#!2UtO=Q{cHGliJ4{
z2Jig^Tpscl>sa<RL$%PhIm-=ymtE~2Fge`3Ih%=Oe_1={7AgST6req)SqN$ZIUjT?
zATo$kU<6KoU~sSjaBuAG?T?W{=Kq*w!zdcp3Th_>YQ^`-Wt36lPauQf1AW><`<`8l
z_SS*!fce|!@`S)vXi{Q=&RTpV@z7A?s+zkRkfXSOo%<;=QsU)H0tW|&+xG-Nfer=q
z3Q#AZDEhBdRSWg%(1DER&I7YY)0+Kv?4V;>Y^NMM(70kzrc-DQY+T*Px4<zNd~q{^
zqe@S%9=sE%O72(5pFW}eVTceWIDUDPPMYS{&AuJ42n9VSWkMiSggy;gU!y+|>LY}h
z$xh4q_s=<(yoYwE=d)9v|1r;gR*M7SH%;w(a~8zeTc@htNoJ)m2=gt<Tq`rlh37a<
z6}{IvtQMDOl2Jb}<+_I!mK+X7i7y!#vMg{dN_Rl1c*LhSEdyGhE+HiaAuyhwrH4<A
zM<xH+{&oXu+_KxxXoa|{|NAA_i65t^^E)h~P7+y0IRc7rYism3AsYnh-1;<0Mo0^z
z9ryA=(x(jQ3z7%!A`v&~F6CFRnhqCSH%m|~9gyTWq*mzAY?D4(J#ZbjcKWkez>I@P
zW>0H>p-L>HjDGqOk*pMuX{NS6Y9*t*yu5t;)Wfw#XLM`~hA+$fSK>pRu|INF*}4b%
zNNc%z#@b?&$0B^%^RL7=u6z=lXoqZQ%ep6rYdPi(4yM?yrT4D@YX;3?FmLC<=bRjr
zss~gDFn7ivp_dE4rw<I8WqTj9-J-D#P*-s~t`B>kW63CI7ZfyG@<+$S5P!lIS5?Ki
zbqh$3bnNV1TT^z{&4?q`tM@3{J`fvF-`qJixAQ4AwY%JIF2DR-<ZBKqh^@qjdakRn
zNu$?XctoZ%0|&Yn+N2;y$ZxZGr5Xcj9$@~3wi>S5U0}bvQX}$*1jB7P%n2df%>n77
zQR_1<>Sl3;?5XC^xY9uTsZ*noo1J}Qp8ofcu9x<nGlYLH8043_pX_>f>h|+)KnHeo
zu`>~)G8E`zW->A|*R-2RU4*GO4u*7V&D;BV6~{mcRBF<ll3Up!ZsVo<`qis%K<wYl
zdGs7UzEL(csG*_3?p7(jc?FcpAz*R1c_4kXPz%AR@Y2n@Bv!nobp~t<LZHKu-To;4
z=)>u$D**uk`XJVtV-+Y{=)rer_X6@4#mdGmVtA@1|A9p8$&)7qDTUghnP&g~{reOZ
z^-lE*cF*J_9)8cku@NLvwF80C4aW{+9y>cV8_QtmLd?$1ah$Q{0&xXy+<W~B#9Kgr
zJWcR-4QZH3D=y}tT`Q}pp^{Oq4Ez316PPVPe)aS7+dr+Qq@)ZA4o29BVC{41Wb_VO
zDI81=0w=$J<@T}2m-2GKsX7*XAdkbRW7~;;8G%a6l3(-BqE~qPLV8gV0gw!U;|src
zP1~u*A&r%TgNd6v3W5%VQVfJr8N4emFBGWK&^QJYxwGu(89rd%IJbij>|kJlif{P>
zegTFVGYbnHuq;6iwz|5?{y4c|W*U+(M2^T$pTvPbk(^v3km3w-2h`q}m_T$gnZbP^
z4SF1)sDjjqkdksNcdhKlkD!DE3VH^Ho4B~30jnMge*=;D)$7-%=mi3@AM*+dj&R>t
z^CyD70O+wm*8qj&M7YW|x6m?}Sqe-#V|cslErKqjf2~~*80ZiIuVZ6lTc)uo_B$`c
zsJ$Ta2e}z?XlZJ2un8shfun!(+f-S+E5&JZyfu<R?$?N=``Or(@gHy0oc=vn$T3PX
z0@TDHpvD8KlR3B^di}tY-aJ?U_7cFSXATa0KqLYx={E4}0pvIXsRZMl?BK|VJisK-
zHUio6BRBx8H$wE386*TXNa5v5g3C@D&j#R81Azm$>>uE|kOx0)Fy~jRz!Y>ZR@KHm
z&o-`OVxmNPKv$Wm_Xi5f`G@`uU{pUBJ@Etm=o%<Xih=VB$Y!L-$S*+m%~U>!$WZA8
z6bzw&3p)j-q?JEE`|^Vcu1V1ITo613B|!jF23ARd^98uY_N3dEJ||fms1+ckfg;5p
z_$`~YSC?$8dlhfxKo^E$s(?2pb~gFlBrHD|)VxT@8d?Jh)gvI0@A}f5dk^Wj;R4MM
zxbT4SP*wvZtS=AWfFj1H3LVs&1|ZxR_qw{c2Lwvm_VT%Kp*Vgk4e@UNGRqMZ@0c-~
zg;os#ykrJ3F`AIZJ_{2SU1Lr%;Bx|cOdG;3UF!xSeH@Po9;i)0phv=?gCgZV_Bgc{
z|B{u%dW)1D3=2D?Q4*9bur3Bx)?i3M-Vb>g8JnE2+yTgLBG0VhR1`ctJ>et&Vf-5O
zq)7lH4%m~&b$2|)6%}z`Tj)&=;}Z}_0*f3HE^5RHjL|~VUV0EIIQOo=pp#tCmKH^B
z2SEvn!8$*0`1I-1{Bi{v&tp2JtmwA3r@$Ke3kRiMZaed!{U2d8qar;d^Q5GE;2q90
z{-mx(7my7=V~7$;78geYO*6&o(gg~E{9_!T@B>d>UO{30xz{Nw_6YC%B?JY-iAeYR
zGZks+>tOWI8b@iCM6b@@KYjMh1-{bM+B!2MBV+AMZT}WKy`W$mXbw@Kf&!zzYz`zj
zsEyco?j>G}7ZV4`Bzy-TdJ$ShabP?_(;rlQQr6bi1>H0#$L&aZ9DRIz?Ct-Lbi|T>
z+*&|31Eu?c1*5xeE&TDp0D(Xelo}9PKnWbuNbt)DlpAo|Egc=-3JXDiK{6$hDGo|k
zlyDNH1>hzjoyHkIe%ve=qWqOr3qm>dQcH3OA*fLqgbk3Ia8xJ-;8He$I!;PX4jaVF
z|D~gG`;Ow`!u9^0vF#ZhT@^y^H@P}hXswSF2LREEJholtEf{j%6hsazTnf~=XM+rA
z?ck-ZZbUk>n)`wv@Ju&rj!?4C)|Qq(Aie78v2A1u`j4Q%r`PnaB|l1sY--A!lY>$w
zH_z=nV$i$B4MGc0)>Jue7;KJLTzk{hOyd<gwt+Y7LpqJp(0~ixzaz4b|MlzF8qq7e
zxh9Yk)`QXk+{h8m5v8L+rQhjkJrE_3eKJbw2jO;|FZKvf`a!N^1h&o`{H{^)HYu>r
zX*~Atg1{4~zKh+d-9V)VEeDf`hzM<a-p`*XkO}0kFOM{v;QK){VF>!nYj0A(y(^Yd
z=60HCmQc)fSY`F$s0Caej~?AfZ*Lu7LX>a<C6ol}B`AtNjagT1r1ujO5pAR9k|>h>
zKb7Z&z?z36hRaaX$^>bWcem&n%6&;5s*68+b_<lkyHyp7AzdPsiJw3J1@myZI2dv>
zx-d301R)Fy2ZsrYu?`(|fhqTBP*Q5`V4wj9d3M$SKxDVrmGI4*H$mi{07US;;KeG3
zwbu}eOwG*bL`72|cNc;o8F#02g1nFkW|ALT4bf@SfNuCmr=V;owF!TSM6thzcO%CN
zJ0Vm1sHWiJB21vZa)krEFrULDWk?3aeJnqn|E)d#33$l`$Pa(#cC-eDLT>cZJn;^s
zR01K{)q(DnjG|(v%T>M^1;{DN^M({5J8S*oXhE^ldzdi&m)R~B9X2Jk9s#h<NJ~S5
zPzS0{e;|$np&fLRS82f1dsA$@lX`WMDttKn782iP`5ay;gf9q?*mtN6K~#Wj=G+2H
z1FW3=^Jn9KX~dG0m{2mjgv>&#F@8W+6`(FS0ex`c*ak`z3bGQwKrn@SAS6^(pcaJh
z3I9ZZT|;YI+ibYdCa^3pK7am<5=kkg2nV$<%tFEdVGJioDNwQr;PLurnt^;wMMI+v
zv?OpFOaFeTpq@}yzhUnnoeR+zJQB7UZ8cU}2#1dZUU?K;Bn&<tJ8#o2EG&SM52YOi
z7OvaE%6+=i;Ol|qGoY|*o<m*HH4H3HP+fo~2T})Qx#JYyf*(In0yv{~5(sEO`lqL(
z^Ep0t25b~3<YZlj21DlTY-Y$WJFYc=RBR#mQczF?0;B|x0LK~wnDimZEl@{;5d3f^
z0e`lwMMZt~Bj5tmb%E#zq<xeT%?;qpE$!_Mb32^x=AY<dm&51a804W+0of81LjY_+
z*>ML13_vv|0Cqd%;+!0iDak*3mH{05xHvLU0iu*55QAqy00Vd&qy`|cViFW216>g$
zc-;a5y7T|AAnQ5${2ut>*K@dvSBB!CdOq(kfgD-S0j24U|9;Q^fAIgrfn41gCZ|p{
VVf=mz{_Phuc^M_?d`Z0z{|l^UAbbD-

literal 19010
zcmeIa<zJL-7dDDFw_uQhbO<OQBHbwp2+|_m-6@?aNQX#wNk}6~hm=S+NSAcS&@k_s
z``LfM{<`<?{a`T6Tyvi5T5+s(9>*1|DE|`s4%r<vG&F2!De>26Xg5B=ACg-b@QR~$
zhC3SCtJ~7zVsG41wr5<u2*(?c$nhRki_uNm_y0Z*idY*ZxkW-u;kR8qUcHp`r{*l@
zbcI-um$#;dciiQ3@}G$E@TKaTyw)LiGu3FGWl85}%+`G`+&;W|VQMsSt1V07ok)KN
zUF1VU*Ojf5MECv_{<jfX7G%$F(X=JLzNkZk|FFp3djS9ZLwkePACLCI|NY_r+cwC;
z{C>^ikM}(2Zshd)TQvT+e!s*1)O59N$HmKAVL8g=C8Af_csOZQy0Gu|kEWi31AEiu
zGV9~VkJk~d5fuC_0odex_VZ|3mDcNs6Ebr0v%OZ_esi7G=g!XDzQ~I?^y}*al|pk1
zi-mCB*=NHa-Y#AC3GOPYsdWTW`$(CYnZec1H%pBwCgAEeHc$J74zU7psYRAoRf{#-
z^{x+y>vip(ZSp4;6flXL&Z8F;7G78<3O{CMHttR0F4C!s3kV3XMUG8Qc3G7+jBVR-
zb91A4U7xK|2skk^nmE?YTwd+^n)Rm%Z_PH;sm(d?Q50#{aLhY5UY~E8A+Gf8`6#^3
z*Yp1V{hPAy+CPQ%wX`%ZI~zT)RSZ`1<HwJ`tE<t;$#<VAD-+zjdDGa;>;kvo>f$6i
zJstO%yE}g%4rNwG2HLxo6;qd;X=!)`?PVlY`_ev}$C3G~SFiq*8)4SDY`54}!4^?J
zAP@+&nAq5I&l6i5Dxp8yZKKoEJrR_G`NhRI^>=3*2#JZ&=N)VQG<fpfzklC6=@)D(
zV9@*_4^Q0mw6>$8<DQ#PwKfa#fJ=^!l{He>^LT#W&D_>Dx2P!W{M>`-$rI0MRTUK#
zE^h7{I+*QuS?k~Rnl459UeXm87n|vkdbinDMMp=sHhN!NGDHmxy+-c-nA+Z&s0jS<
z;U6!N^rjL`Ev>JGg+?R!N@#zVmj?$1o*5a@p}}JKQst3mp&=o^ySfgL+XI;|uA>!A
zOdi30mw!jmp?$Bas<0Yo@oiXMT{WtjBIULu1p8UfHKyv+xeN{sedt~`952y%qoQ&{
zhoNR`%i8hn+d3U}Cnu*eqYfP3Gh7P3`8n3iO+!O4L-ma9tjFflEf+y)X=w_6M}K&K
zio99XRaR|nlIyH{;QmraHsc-7lbz+${UtES#>Ph3gmp|@+@fQRJSj#_P7bW+Vsr5$
zv2I31Mf7T#Cr)6i!^(XR3{6c<MkXfOr%$m%+duT0(mi<+nw3QhukyL=J-9g8&8w{?
z3v6xf_c`XQw4He+CUyhv*|p8Y;NXJjtd|HiTpGMZQB(8IyYbC!7mB#JxX;PSo5&M_
zj`*o<$zd)WN&#_eYv#_*PPmyBus4_8+3>0D+?<?`l!7j&k(mS+qru**{<XCNE?a-Z
zz_WBq$9<>M|MUsE?tE!%Oj}$u6uCN5rk78Vf?rNm*_~i6k8e)dRw-+1%euJm*x1-m
z2)PaDmFkI#g83BTRaI3Ym%7xDkURCdtnzuoP1n&DR#uzGyS`VOdS6RQR(g5oj?!5@
zWqc*mMFK_6*XVvlP|A3D32ly-<`ox*_w~tB;9yowJKhB+FVbsD3k$=7+jINw-NSV2
zcxH{3uGlA}f-cM-KYqMC$`j4`@gw%~(reapS2)e<K)T12lPWS$rBKxYF{-Bg`gKcT
zX=4|7-p>wdUs)$7E^41snyMKW@+YcA4}^tNKYzY=bY3DiHZ;`2*)P-zzQay}g$lwq
zYHGs+13!gNrdxC6k_R>s^@n|eLf5BDut_{p(oTkSua<?@z(gJ!vcbW@BHf0hm(<>b
zYm>nThll(wTO>6#HO8i<S;Om!D)&?6$89b~Mn?9|R@0XpYfk@0D^&0QBCpx5oXRgP
z?bvpCd$C;w-}cPMM`SopL39=2%7Nkc9(DzHs~2)s_1Q}DS%>Jg7;K=<b$11F;3HAd
zbaYG{d-x<OH~L&3rXTLq?S77n`&FV-SH0VK!Imfs{#|Z6Bar5?@mSYw;ik`7x1J$K
z;-5+z@+WE~#=pWyxBis-fq8K#_}<Of=6(D2xnC5SQeDjlsZFENi=RHavq_IxuQBD*
zCp<!l1wUMo`ET7U*#!k5YP#+p^{zJ5AfzZOCVD(~8#;7cXWAh_{0hRS``zDvvP%q3
zV3zJ37ZXD`ZOaRZXK}vS@6WbN<4MCVJ`vF`_{>W;H@=C9i9gmnSQzM9RkkLFE4@8Q
z+;<B{yI|LDTjd>!DzOm}I4SZBBI%L$pQsK@OrY}GU1FBMdkd|`#>RWF=YyZ<xQcJy
zw4F2|y^%an5DdEGp29W9iZohVaYcJn)N~{?G)N5$45nSCY#<JlH8s27PW!2z6ciY8
zk!e(zV)IVhqn9?GGdVaopfdL5Jf>&{m}92S6|!r;g3kc~q@pV^8;(S`E5x|_T4vzs
zRz-hMaPUGwVM)uf<9;jk^(Xhk6(vQ*pQvaZE71uBi$WqKx2um@P4`XXf4lN2KE6Z2
z_gopdADf!mi;gSwOiJn&CJx0jCBmVhAu!tNcGcWUzwb4t@5R(ynPzd3MnwRaG>&)H
z&9B)H)AjCpSJx{a_#KDsG#@Q3t$K~)BipLvg9Aq@&n<RTO`@fvJDc~zRj;<E0gHj@
z(!7r~I5;`W)tTLds9;sy@U>{qAd#S|+g&2&dGJZIo#`415w8R&aq%;KxY?57i?B=r
za78mnl4#88W$yw4Fz(#FTLW%&x%_>?+UuZ`!O6wtn?WmvWH?zn%03yJA69Bs)0UzY
zL<}HlhLNymogh5MJSFkg5gy3kVRJ6WlNr|v>j{pKkPt<}kEng6Uu~H|{@S~~*j+++
z3~nOcsH)y0Cho|N6zLlIs{A!S|EFmMr0eHU|JvKzuhD0>T{um;2)X2z_B*H<goILT
zrYbWT@4!}jlX*Ffeqp7#E#3~mB3*2n;&=Re3(_9Sdo6}@qoB-yVc(WCo?AAaF5cy_
zn+?mgIQ0^loSgjK(_`HH;bs^qCr0UbX!kNL4NY^?a0;W0jt(UnxF_U)LJ9u$N+rUZ
z(VcCW)LuV(5;>t1+^8-2`9=j{S5c}ral~6S)eG4l6*WYmAAeYlzgANVl+Vj9FOPzH
zoRpLV6#-n2)1c+w>$85}*SflWutSHs88o;rkO1L(4QFbdvB-I_Lfijwy1MRWZA!^Y
z^8jPrxxPMWT5_p_%mnr^ELV4%e~*c}!yS>-c~DQ_`rw?M-QAm&da@~e_@V7B;79V3
zk{Ic}*ZfJjU?9lb43O<!Qn~x%(@CE0KI`i0g8U$KzLwpS%!{hrdyteIr);XCz%%2X
zYMEPFKF=E-1SkW6lE`aEwe&0eMd|o3#GkS8!|r(2lX0V0ukH}g%j%yWtruxkm1sp5
zPHnHO4Q69ei=?2^NKa3X?Uc3>A%u3JYO&$|VtZSA`&rdOuu-=@+i8_e+1G+sFJ7Qi
zQd0i6jC4`zykTigO)^~JBfRqRa;Wf-pjrW{1+^QBm5y`1U6B>J*y02?OyssA8c-RB
zyDPNcM$nTg7!P;90(=oFVnM-c4UJC7H12<+UP>t_<ki<t`;OPz&4~z6MIuiRJQ`3Y
zf|@$4AR69lYBbmA%@~4$1F%7D_vKijv%k^THn((&sS*<tg-+-F!l;CkpeFlVE>XYj
zd=y^>`P6qfVR-!^E$xS30tODVK8n|`U!(H4F-KHP%*fLtVJcis)1LXLmsB*&%wd>R
zZtwQ?>;V(h#?CF+m+3eEgK|>9r-{OzJOEY=01!fm&1+4GJbdUMPR@IeoV+{cv9cmz
zj#g;Mpt;b})qFknr=Xx)5mdsjpt$z-_FhwO?7Lku>_Y`81Dpe?fP~%f-|Ux>D{)%3
z2H?!sPbvx0eRms968Icwm6ViFFzsnkL4nuhe#h2aQ#vW9X~yunl(e+L@z%t&FY-z@
zO(?OxUI;)PBP;6(^~mkpw`1btmEXJ(hkFDbohs-G$U?@_k_iB}Y^p%)#cor&Qnn-%
zKQT{F!LMJxPCIr#_1bSI-kPfF_qD=4O;*j7O$5ufY$DFPk-cWDNxAzF=$rsX@7=q1
zX|o468FGl@-n^J^bo1F_GDBlqn>g6Adh1(vHGrhk)1^pK4&z&@h4+(tf7QDmy1~_h
z!^0mjG2O`<4$ifJO&xg%N7Kupkd^*mmc;dPK|xpMi@@KVonm>z?C?#JVI<8bjVHo!
z%o>ZhBLobQCv(2(M@LSJOG^#{sn=J}l?WRz4`fh@vE}34L9XzTkvFD23C#^Bv)uqh
z`F$?=UPe%4R#*S=<%kFg$%1e;>WEO%(vm@yVwY`NLBV9Gm!G4fAt5Y;JSmwU)3R${
zc0`#)Tbsgl+5r=Ap~+OGO-xM8>4|%Kf8dPmNjwcFXLQ|eW3miu1EGovzsC_{G=qXE
zlr1P}PE#worST|SnaF8&kBn@?w}jJt;K9b`Cco1f9`qDgq#O^|oPD!8c^*Fe7h*<N
zm&)DUeU7QLf3<bNUQt~=Xn4JC-3jHHBA0tFn2GMfYS7l41jxw9Pz@4*<VW1xu{19N
zJ|!l0qRRHs`tTKlxP%f4b)KD_$tH5#3~YT52A;763@4Bp5h;3|kd#CQ0qHkkO%ETg
zdAsU}ygC6tCxO=C_s&e+m7F=6YGEwOI)DM+zUB02nAJ=B{+$a0m#PN}ncH47CfF#m
zp@9-o<u{XV5)m&x*2Ys>`Ml_*j!4%jRfZ3c7|?V&E&x}dvYCqyV5h37@YV(W7XRB6
zeD?1kjYd<Q==q#If;BlT{Tdt_3&^z~!NPz5_*3(i!Si@)pO@fd%zet=c{x_?n_>H%
zFiE<%D?QDthxnwV;4=P!fy!W@8C$dAT)An0AnMFSvtEaCsN!}!X@$ykaOrZR4h}H6
zZR43K6p9|m>!du^B(_!LkXj-4TUc7s3J4_e&UyWJ6Fqlo0wLiW{?M!eoc8$16QkAu
zY_F5LIfP#Mm%O|R<Ia1tJ{LAbthzm@!UpiQT(g*2t1=9BQ0{@SY-wqcl98bUmw+0t
ztfbUj>%5sD>MPP3h-(H{y4LyyZ)CR|VFYMdb9vA${pwYVJ{#Ws`x3IUvQ^tgKQ!F|
zuqLt@Jg;gB3<z*sO?5pOd`Vpn-~0LV=RUcQi05zK5OSLJ{i}cQx3ec*G@aXOtPH-`
zv_GvEz<3WJSN~B<lb!^2^<veP{`CIE_Rt#%kYpk5Co0U)*(`=07O7fQY1yi%K!`YR
zjyb@0a2Wr_Q!UZ%ocv*Rw&U8={U~1h!-Tcz>D~g|kN_OYt|V?NDLFY4Xx_vnCGSyD
z^`T5>XU8BSl2$e6lT2<~{l0_BGfYABDrRlXe5ZN|2dT31`H{>eX6>luG9<Ev8P_?J
zz7&3;qk$Ji>gE4{^|{Bj{d7z70P~L)s&V>Wd!byiSwAB)^F3?|<<C%6HL7eW*Voq}
zuj`hguuXz2z0aN>bs6Af$QKY<457rV^iQ9LuMg*;pf>>#(U?s#E;&!noc4h;q$RIi
z&$%Ulj(=B@?dUl;ILr|^I>g%Gr5)=|pHotLU|qccZ3CvZ;kQG+YV0<XT!w~*rzkUk
zIY{{(AA+m<2L!;Iw@N*K`m+3~cA&Gcu;@t>PX6-c;jAyx8ypkjeQ8LB^=55EF@ftk
z1fI6@sLC3`0}_zf>(}=}vOYyezW|76*LX$`*ok^v&sVrmtrWHL9GUpnDk>6C4P@e&
z_u~z^Vje>$zkfVsheB9X-bYHa?yHor^y>4CLU0(vt8>J|hY#(0xyS9HilLRcADTis
zK=ldca7!2%I;gK9<;x~<-A<M7n67zyBSFMl0Ad6MV4x8J=aq$e4;UKqh-q)q$7X$0
z`veZ_Te#d}Sibk0F7i`S(km$`sVd((<n@L7VJ~mz@89|xBVWzll!g!$?J$cS|01_z
zgScX;+xm9R60(Mg6X*wd9u;nmu&`6)^?F{qVZ}uCdT#Q0xznOkrQMu7^q9-b%a@D9
zO%H^GE@RT?>fJf5#!Hqb%+lM9Id14|pEsS4mY#0g`FgU1c+GbA_AZaA=`9mCowir8
z@cAu&z{X4N4Uo@kwdOf-RT3B7JM#idu{*5|*!sHiQ2+#d{^G?&o_>}ZYS@sUum7ja
zprN1;vK_o092B&6e3F}+`_|=~&|K=;O(JS)-@dU4vl~ML19<j&r_WN{hz6%z`NZeg
zNU$J!kjH@~tG6LZd!I+Npga)1coX5`iN|;$VUJL`N?+4(TTJX&PjUNogm8x67Zj{I
zH3_DtT9YGw*p2z73!(3yd0hfR#IfXkHi&$qvi$QsI;!t{NKb!`^x68uj_P!Og_AR2
z9!kh=`05L4T%Ab2Gf$E<G|c)gh>IqYzN@8krOiSNX)kJ33H{v&^5=2y`EG@6{nb9B
z#4s0AvY>-!PR*kBzDL8{#zH3JtDi6GgYJg@?|q_nb^B@Zc~jjAs>=rxQDpRal81rG
z4EV&R8@q<LY1-a7%r;8730!n*y4l&?q?Bjf#2qc8DJ^LVedv#e-gT4w!R-ZlLuQP6
z5$sSU1;;nP*PGxMS^e5&H$JD~{X$MBbPoJcy!M(hUcrhKyo!f;-%;Y-W71ncM@C5K
zu_FG~C9d|{Q$D}NuBct+`T)(g6t{HugIAO;jC<XPo!-G*a8308<)l{_<9BfIc2lQq
z73Pm;sChA#Fa|)L+oETk3Vcv=2e_TiQJw3qFVd68`p>|2g+OY_6qbphK08_Xs#hwN
z`}?E%L23kRAFgm3A`qJ#k`1T1l@)phj45C+Lgjnu03G@^Qrp0Q26@^cnqN|a<73h8
z4py$To@jwQ-Zw$`Eu4ozp}r#hk?)@_LC>zfPJ#AL#LKJDLzze;gU6^Gb<$tqY1qQp
zpgU)CkPm2rUN!;ddG4P=nfiD`ve>*%?ZGvhxH}cg>Kp1(nzotSGt|yT^IldPiwpT>
zn><{(9zZRU6?P{lhYeUG8~RS@{h;#70-){}KB0g}KJ;Z_VhYL6XXLS&bYkgx(#Ht^
z8b+(>6?bnn$cw4me<LB5$CU3lyP>vPd)<DP(}4zH3l(P)N`$iX1Rf1(0Gyy-3PJ^i
zM(OwztMX*Xq2KtXu-Rs<v&Wr_?j(N2*{=ThFnZ~g_v@~!-~M+M=v$xw_-yXZVd2vB
z<pklOchh6j3$v$K8I$<ZoX0w)-dinhcqDgex~+757QP-bqnkdxf5(BUI)fqR9&zas
zHqVYuO?-%{AigR3t`J`+X{+IRo%-{j(kMYZ2i*ee^)HpZ(g_pKtn!i<&m*o`KI1Kt
zKYZ5aZZY(UF)plX5n<6oPTKz=>h_h+Lz*^TDS|jYRt;H?G#vq=x~c9$S2?=2{LNk3
z+cb@0JMb|XmHRES0=HqU%J&{D*5_F0An_Ls73W|cHZw8TzNU=M@~mDYTt2@3^r1hW
ztB8nV=i(|)#c54%k%%I^VVx75>DfWS)ah)fVD;k1r(WNYi(kbDE3!vzE<JjcR87?8
zWAVKtMEB0QbSFGgx`;4ipW<BCB2FbLSaI?L*H~>Io4)lC#%So;73nq`ToZ5ULANBp
zV(fJx?w;$RuljaBTVCQuNjgV$Q|KcM5)~yWg<D1^`}DA;BY(D$P~*MM8_vgJci1zO
z?{i_(_INnehGySu@8q7Ee#7Nz$i7itBSbjTS@&>pTA9$zqtm=Ac^fMdcc~*j@-J;?
z<oi*7#G31nh7f&}Myt{3HiF<5yVP!4sD`CAW0kyun3Dd%C-aizl<7H%d+2Yh6M_f*
zXisNjn%e>r&+GWSS7KZr=#~;kcKCIUApE%HT1_i53z#x9GgrU7^M1_I(>5LEw%8IT
zOL1>>t{q=2SNXnAqcK;<hJf33u53}%`|#<hg`>UxdV#DL*xH||7~bjyy>s&>W9e__
z-dByhF*G>X{G_!0!;I}JAZ?SiflSnp#QWly9l!uf!_jM`nX|ym&9{o~^+~dUlO6K1
z0xpdl<ZoJf1Ku@mA}lVCuC!<Sm^!{FpOx;)%^xRz_112=J9Vauq;>PKZt3dorekN1
zvMQ+wfSwXvpCfUAD?Bca2x*?by}b<$JaoYBwy&7`o;W)@zj^Z}r7C;L(y2ja@ot&d
zHV*DoC>cAsQ+2%v4|U=0)IYlOeJAI{&7($;p92aCpmcgQmMl~dKh<o5_Oln}-uZ=_
zWQKJxNB-Yj+OedIx$pg=2YLG6H%h}BYx|r@xelj$1~C6*og|PF4|@NkLjxcf#E{G{
z+>Kq@R9=kdSE0(A<+rT>XIanG#sW&=w3`(I+$syxHJBFuvixz6k`ntF%mYYCN$UYX
zt`-f@Rv%bS)*wpj7_Gi>I@mw76J`tb-?I}vRVMs+^DdG84^3$E_)l&MW4^iR<KC+t
zLT)`R>Pj+48vAEXLck8o|6G((ZkQ!u<wAdr{aH(Et2+$q?CWL#^T5y>4e;z?V8?Cn
zgY?}+4?e<to%V2{lXmu&UJM_fQS*Ryt@Ao5Gv>PQ!9$Yd2BzqN1(UxEEkrHerPv$J
zRjZRaF%4!)N<Q?5irpte?0J<QrZB8tj^LwSeZ$Tm96`-O_TC-sctgm9mD#WPUGo59
zZ_u2}<-%^?Mm{pqOoacrs`U8X7nzcjzSSnLeTg^2<jBchXnK(>%eg>}F+O_0?{Lnw
z?nQI`s46Q~ubBTs_mYvT104azOPt>;y*Ch4ziToq$UIcUlxnHwb%QwWwVsu3c5wO{
z3)#(xMxTb}&(-psFeM2-uKlgMDCA(h>8z7C@ctZ2>Yp#TiAPLG9|Q@;%fN4LVX@TA
z6oYdTchqGQf@2aC7at0>X{fuFB|S--Ygz8MTR`#<9lJW5_xS&nN~>k|V~b8>APU!C
zSbKkFE&I=x$X@A4BY6+zeiq8^J^!+67WtoEBHLM{dvZtVRmU+(Q!77L4i3h4IIM{~
zrTz%0EL0|bg_!9cmicwGo=-&lTKVh0Qh29a{AYJ=usKK(3aB(MA=p{?|EaiM<X=8Q
zXf3xRD5tNxY)j6TctixXhx58|;%;y!WZ^ANdl-ueNFaEghYR?4EHgT7{M<h95b&LS
z(>Cchy5`J;EMH-a|I%T=j+lE~JF!D6a5R#d`|z1Ro|qBOnv_Bro2lYo3GN_$wuB?5
zf`SY*$%aY0-T3b=wP&3dHCrOHZJ+o3PEQIu^tTrJ_7`i?37@CL3yEOIS0QH!tiILA
z^Jlf3Ee`I6>HA0u5h!P!BEC8JlU=Mfd}fUtJD8nyw46#U`s0#vwft1rM|B6jmzX(k
zDzt!!#EGGaXf{Q!zp1z@XLx(zB_y>HjTh*;<H2TB!aCutO_I~tx99(gG7j}*u?M#W
zkR~P)k=yh4tIFo%SBM3!4408Oi~Ao5;eACNLb6C9g)}=pi&`74r&mU+3^@(;*Viub
zwPRC`R0+D9$gm%ydA_%KVm-ehYP$q~j64lmrA3E{l}$bCujX#74u!d0H$9R3e2={A
zZk4Rz_aCdo!H=G`z5BawA!adVzU_2)h>PnulkQq3kXO(j?^|y1tp;9A-#?kwrEWsf
zrbBsfp|)GZCBGwozPJ<Q&*l{JVLXqWxzDkyE-=lFJAAc??^okD)8km~!)fA<0j@?w
z*kW7Eggf{4<xYi{07aT^X?A9-bY6F8W2C8cZ?XAuc<Lj<@$|F1Dc~@I+4pLk#3|HL
zSGGMYTJE9CXZN<_&I$jV{Y+m%Vzn$<N*+0pkv+PzlJitz#FekcwU12_J7n!xxS_zl
ze)gA)R>&=0ayN6`$to(WriQ$*CTphWrOJ#&R%Vu#T+&MK<&2%nFYJ9EHR--3^pjr@
zEv4-Lpme+8x+IHHJOMkO@pKQL)bOuy-Yf%AL`L;&fOP4>(Wub?=D=Oq8?-O5@?WVD
zs2B%jTD_XItM{f_i<v036YV=Bzatfb)l6n#+UZ|Geh(GH$$GQWHpzZb!3#y2A{SvN
zV)5hSg@QD<@f^hV+`I`#r%3J*kPS%BZ}LzSo5<PTr;!x4GmCxO`kkidRp>I3=l;Q5
z`f@gt4&A*{4|^W^7vuK}pR19dHu-u5Ap$K2-pN<RBh`6IJo&xJ3bHcBzq)x(^7gYY
zK4B@1ge_;)cSa}-yE(|PAHIB)!u7AbWGRzSg%k6SZJeQ{*3-^dI)(EZ`-=Ato%mr?
zgwl@h#7da%DCh|Xgv7~XNF=$RZ?ONY3S{y)5?m)7sT^K7*xTY>@<2EmB)9ga^Aal!
zuK@JuOuI+jO`kS1(vA5<E8pfVo+*#ww0?H}EYbJS1kR)(A~;(lsF)DS2xKE#4Jcq_
z40#SPK6;l*e<qWR9w&J58+-8D=q$HH%2tD4W;s!dEQYo&wGs^mGB~Ha^`q{#z!|R-
z-+E=}lli0bd>hwuhIq@Z(__Z(bLqk{0MGu$R2@V=dHI>*sVnt{bl-ge6$2@!Tl~vU
zPr0PRZ*M!hrF@Da=^Gg_-$~$(X&w|-DBs?<aO?X<{HKiXE@wcyJ;MKCKzpi}|8_MS
z!J#BQ9<k}ew^D@Rlzq0Y`3!<4Nk;;ov!63sea~iOaxkF(9?kXZ{)qElQ~E0U-pFFq
z7X~M6m)gU^_Y^NRkj_;?W^NQqpK1A3$D8ZJ+;YE2`dYbrFNfuR@9pedfHMdhm@>d%
zrbG0#ov)llVM1>;wi%B=*|>QN@n8Ip!WeW3Jz_d92NHYTMrLO+(kx#oQ|~P-U{q2R
znz?vT#P+5_d{?`?=BSI{RT>@Y+*&B#VzS*xs_+q~pr*`!*fPx>6-{PH-!##S`I2mq
z5PgbCV2?~rn@8XVZW%{mQmaaF4<&y8LV>CmnKt)q*J_~Ex|3&(x8$9qPXAeJeLWin
zyB(%S!Zaxiabc@|-uExhCmps5xBu1@o3ynDM2kGcsq1bPo|9h9nTd_zo~W@pR`kLk
zz*jO5au{{<If|tCoYnJCEBkbz!@Wm7^wFf@-{=&km589F$;4Jt;iA;!+?+j2Rqj9!
z#`Nw@qs32*l|)=@9*TTM7!rHUjjsaQHG;6%KdB70#V!>))z(}06q*et{5!PRTE;Jo
z75-!IIE}+{l7eBz2OK1zJ+QnuBaOdxDX5a<jH@as_(hhzNBv)ws<*>OG(EB#0;H;c
zyf2+M9h?dlS<^cf{?1N(&BJbv9`*H+BY$%IhMk!vg+4OGaK}ezjqoF~+MVfsss<py
zcI8k`^KpG69|B}gp-~KX%4bQzV0`+W^yKHa_QtNszP(D)5MU4f*h+iDej}_Jq0>=;
z;q`gEjrgTyq9{k2RmZQ)e?+oL^ZI_*VQx<7O#f!I;ye4X9Y#smxjl^GuxLH7^---d
zZotqrnu4;We9{G#(?ald81y^F$8PGZQl}5L4Vn^{+OZngJKcIJzeh6v{sgC@xD57c
z6pX{oPXG7=j~udiTYsqVa+i{``f!vPkZ)M)c8>CuT4|n+3eRkRiN2=+U0{Lc4Zc6~
z#-^Ig$&t>&cBJ<xTUrKhfBB_Zee(HDz%Px{ey!Yt#W=&A24fOrW8<fo1UyYaVF0!#
z{gl7v6=M-Mdy9nCTZ@j=ZBsrytJ2(B-k)1yke9U(^tp6H4IzQ!Fae_v*Ew%qYTrwC
zn2CJJ$r+rTmG}0(lFs|?Y&-qT%!FIhu9WEh%~bjIQM?;>=F935%1S?*;+$BXg%o^p
z$`7!wNQ=vMFmEtl8h(7IpkRJy`F>ayW7veohf<dJaciH3c?=Sc1h!aksY)e{^KS&!
zxi*N)dfZ<n#mdniDXebLES1DoA-tJW8M`lzz9fG(k(SGa;No36TO}U(;q>h+)X0fX
z$a=(;d-IelT!m;!P4Ii9v&a3am)T+;lkm{Pd5A}!Ee7$?^ED)p6KKTwVpa7AerU;$
zvk!Qq&gb|bOqe`$@6^kyM0x+Twq)kofPKjL?ZW=qpCb+C=xN=g=^C%b3jf?B7e{P?
zmB_GFpUB9gaEZ(DbD>Fr8Gp%-_I4>Bn+G%f?fAPwwmbIExEfEHLk6Fj$q?PmdhTA-
zm=L^TF5+2{v=S#@nQ*kG5_>Gqt0n6J*@LsCh~Y$$q)tIe1Up~lJ|xBok65AXH2S(w
zo*$ci-P`-+BI1$#n3VFk<UJmh8!GUQ$Ag&^g*F|V!}7IP8-yRTO5+;Z^Iwz}e0Q#$
z_^Ld(wqbdGmS+&A&PGw8)c<Si&a23=tnsgxt%hG}T04K84pP{a?RTX{4<hW}yL~&a
zsu2>pw#5n9PA+p7yv0r)oRlE0^@@v02luC(M=Tq52`i&XW74yo-KGbnxv5#TIFl5%
z-|pW`&sEB6em~P1gv(ie)osK#t;-|+@sqm-hDSr!Z=D>*`wrh30tPMLmhXoeeV@{{
z#$ji3-j`aW?ftS{*tT-~S!Xh=cbR6h&}3Itg}$G<Z@}%YWNKH*r9i~i)0cVqBai37
z{CK8@LyJ)fvL2FKk6RV-Bs^dL%QSE9%5nYdk=~Z}3mXg0ryi8-M05uD^dm_wE05<V
z3J3Eoj7gpR$(;RT@&?jKBA$<bdI=YPF#jvduq2YACv8bCA)G1oX0V%(ZsX~<cjvV@
z^dwHkJ}NgS9n2@Tj6&p10t)7zj;>8_E|uj*{Sf}BLAMy3Z)LVpj&W>_7*L;)OqQ3(
zE`02ELG&a(y!@NNjMC&}sx{|&qK^hf=U^rb_Ug$PRi1{a$@^ZOkUU^W9#S3VauT2{
zC^&ab*x0Z<-kI?ndTgUN@u@%dHr{<xk?nyM0hP?MM!$wfz0*Q;_ai$zCwIjQ_U6Nk
zQzB+ha%+Ty3#?vDDu@QwWie6|S!P#f`aL4}>>+G?IHQcQ0L4yFL5D>`$?dHtzs4iE
z{E(JU35rzpPJ9ACRx{`C*G|sr^)s!;kv=AAnq-%+p_>08Y{E<xeLDK|U>*C%8VM);
zdQAG+c}y;xp&*bAmA##jdQ*Pu>4NO}cLY>uo~!It3zMiQDwX}1VEd}~v7*~gLS?{<
zjM9ds#fb^#WM(fJTzh4kw?|Lsa<a|W?C;&Td{$cE_P+ScIiH1L0RzrbUaYD0s%#gt
zJa=Z&mCLTFNqtH5W!TzvV0v22f3bbrDf?|rH4A63pNX8?R;u`hil(MO0uRmm@BVBe
zT_r#NI%JPO9^w7I-zq^ju=#V;VCiD`GowY>;N+0lmx6g`K8H;FOl5~ZI$zWc6FG#E
zieD<|h@eGk(P`(+J9|R+8$3|x)}SLskj3P-r6cT4>vTsq5VNWvLzysyNkF0IiVNGz
z7P~^ed1>#odlljGlvP}%tZM3y_jkZayuI7<%t@FHxh}B^H>2e<izXt*^HPuA+bbaE
z)(>6}Z|H{=yr1bD$yAb+!lG&V649z+qs{GuSGvozJVyF>TA;J*!n1oqX=KfpKya}z
za1@%Mz+Zw|cS$U;JLVk>#vXs?sygg&4s>%86L6LN_95$2s#@>PQux{oOO}%eVSF@F
zBefN}7IhB#_rf2;x4uu^=3Tq&cBz^1YAlVpV3Au2FLH#SWKJSZU3_t9kzeqkVWy9I
zjB2s`<NCA^%?;Kv-ln8U9?bYi>NcI|Vj-l-jfD{6$nMta5qndt(9G`zx8pkkm6X`l
z2q{uoRwAWng@3hOe;{K@+76QonV%_gEJ@}{@XUL(U|*-f#{5sD!V>EyPL$rTe2~)1
z?_f0t_R~Va=8zTno`MhkOQ>O<2zGW>#^b%~2wE&E-ssoa8Ug}wt4%_ZUp8JGx&xmf
zRo7r@;S~<y*y7=7(3l0IE^ia^VLVA_6b*YOR%8{MiPS?Q$<*`HK+c3A7nmWU4v&C_
zg9j(npJ-mZQ&zC3GaIokxtJ9<^cYMUUi4d?>oh1VejgX(<ijQUW86E;kAJ1Op*ptZ
zgNGrD0&dqq6^WxdrRGOnmETfWt6wix>}ypIaUMM33Jl!+{iVw9#a0!CVdv+;Hh;v-
zRI#v}MJ;<%+^7jo4?*ekVXlB*n&z8MoKTf(T^L{-HkbOTZD&bt<OhEO8(+u^m16Oa
zB-IavBNT8|jYhssJH4Pc_=?={yY7u!-FKn(y`B;Y*$_;KxhJ9IjVXd%LSnYuP4~lz
zPx$g;)aVv9wE_vAV6DGeB^3j|fPaT=s;paB)S!sD?~_1EiSQn3QiaLhrQ4Xl^mgW#
zt4@^0cm6J%t+(lMe3_~ZL~o3tD-zaj4a;N~*9t4*f8G7(%Pq+np2S-fj*|y9-XfCp
z0ToR}AHScn{ldDlIy`L&Shp<Uh~lg5l3@1Yx|93F+j>%~FNyU@JOVl*ETs4{c~dgp
zb<wvw=VpBj=!H86X($-&3@l%94HD5=MUH$n!0y;aC~4a*&B_EAP?5>GOLGnfw3jW;
zpp!TiG<p4!$BgWgY5mhG;~_litDO+Md^qGLHi%D--*QzYOUuz5R#ER)@24{Dqk^H4
zOF>mbM?j~TuKA0cnzF5E9N%&X`Z8n1FNt7i=_u317FqwWkh}$RN>7<26!gl*;Z<^#
zYe9>d^1eJtp7?9AZB9mlt>dbtD`yn0+kuoz@A*rf_4Q1vvZ*t#pk2&3=>4w*MIWzm
zamSWOHdeX_1s<Zo6?+15(=Xp9nlGdFDJtrM>a0E^jVF@E+x9n9vxu?KAGljTLPBlG
z$3u6W6ACtv&Xi5jd7&yzU|SlAeG~CTJdk4};_Dwdf$I~y3BnD5GDeSxJd?c|mYAxN
zL}^LmaWWnXyQQX3m+VKD?><aqa4@l_1llFt`iVO${LH~GoIDR_1sdQ9y|JFcDVZ|K
zgfht_iv*#B{Bj3zH2l_3_qS?~{b+lgYQ0A)`RzSwTf}d`IK-NTH3>r?$I2&q)h<cu
zXl=IP_n*zjHmULq^wGg^-VMh_bw50hxpIn|HznUXeK<S9y(c=O+%+lu$e+ad(qe|D
zL^jg!a{jmORgnJc6r3X#vw=9@Lw~np%EDjk#|~DeEf%fl&)!zESEriWvK~2-Solrd
z#=-PQToKmIr3`U#k3IL-P<R=b`j?T@w@x@(sv_;wo|7jYGq77myZUx=x8KbABavzG
z7l8y7x-Szn`lI7>4F_DEhlOHe6k~CBQ)Q!PK61CnhBg#xFba(PZt9h3-y9Aqclx36
z<Y(k6?($BjDh%q7$t{CS^*)>bZcL`C)Ad|q)({Z%6#1G9xTvX}H9r3Lv}v<Lv4S!D
z)Sdv!rl}HHxq&2>o=)*erPmJ|9R|({u@1QMqiO4(`D$P7Tn)pyT)CIO3kGwM|Iz*P
zEPAH>{!$Uin^>H@)h|>1`D(X$H@{jqJ3Z0UPQWk4c<MqNjBlzt=a6!BlY5$<Kk8{@
ze@mVP^dn}kk!ZNR<b8-}OajZ1Gt7U59GES1^{TE;(jKc?Zff$)4m6*b;TmlWm>V)n
zYHr!K^m<&^fBp2Zq9wV??DgD2o8id*>D0x3Z379GfU0ee*?1Jw^Du$MR@uvwA4X?Y
zD*Rkd)#g>68tFPMv2<By9W&A9TlV<isS_(LD+hbYKitu(ZCvOd`R}#24zJsL1hJ8u
zSJtgLxx3HXO;`(Pc4w8YTq>sGIOz#82=UP+Zg%`Ss}lV=qg*sC+<s@E*avwMYGf^}
z9?||KPr^@8Wx-7XUBf-rxJT+{MsW^hd~ZZ&MwT+6Zml!ZT}k@WIfJv)#>X%@lMY*}
z<<R3}{!quDRi1EEG_CPr+;*3VVK20M$jx5z9-NLqV9XVD8?{$1C15krU{Lxt8vN7m
z<>-Go>B6Pbyf8e#@97Iu@R&*l55tdvm}VZs5L*RzaH`^~*_4|io}W2;sNCK-y;M_>
zGW2pJgx90;+?EcXb3Az(*EGu?t0NJvEy5Fpua`{xsOJmAR$=WA5`=n~4)S7?K&?Fh
z(HQNgRBCv&@hs$g@^@yGFjmtiy#4JdgVX{x#)B!}s)VDw8QfyMi!{W;?lrlydrhjj
zrSZFM2X7?yS&ar0b_j@5Z?&XV&UJR>AuUt2B1R7j6(WZg_Uy+CJ>7iItZ83_%`xd^
z$P?Ay(Rtf*!PPsrPyfsCZ&MR*X2<Tu_Y&JCfge_aPJGi|SD(VlNd^b(tLxQe1#)r1
zS6~u1JF6^Nt<i6yXT#|ykFfLezs1R2jPpjtpuN(NXL;P=du_PrG)L#0)#DcH`RiEE
z&SLY0+3^auv%as|B4Zvya~=i3<Bhsa4F625?;9xIPoWAoa4?W1CtE`HWDN#Svp@OI
zK&kve5=Qsxg#9?ko#v8!g?9>VmiK)ww4DvFXx6If-{Mj46lM(W7|sT{twhTHTshji
zV2y~Pxc|JU-A!^SE|uEMV*bEE_1b8{?U=%D@w~J1OZ&-H>_MXCw;J{GZN7;donA}`
z0jii;)34>iu3}rkEuD2Y(^!cjrH%I!-t)^$eKgdw;DiDRh2xBS8|DY3$%OTFJ`M6W
zBn;Ze3|)wL4<H22`Zs^2AF!+{nmkFT$Zy$3;2RD&oT)#-Xb8!=w302TGtC)Va8iaE
zvorA`_zw3$@+Q5#oasH7&Ts^Zr+O{C8fU*gh%*$J(LJ!5)F|Js7JbH~CBXk?G(>FN
zIj^efOHBvbdU0|tc4qqyEp#o92rvxBHlgy<aUbig)<HLlN}++nOHF?^o>T;M>@U#0
zMIA_3xiQ0RGAX!Io99<ZXU4(7ilGpLh1^D`JHZ2t)GqY+-2I<97FRliO(zvk)IG~E
zD#Vsg(y5oev8xXC`)Ob}>mf=UJh}U#$+;bFXbO#W32H3I@J`hjro=AnH_ucRYEo=5
zTN;nf-bt~g-ZhEme8in}q2SPibTFa`ESwtndI#Acl72J$-YHjJQbSHJ>ewz#Q|Ml(
zbbll(_2{-|rsRx8Ug9WMxMFq$&0r>>54V};<oN8qRPgQPE(gEUwP%l74WDXrUvIUb
zKWNKX!VKl`Kuk2^L=s5ihUg#hN^0;qFnHz`KXp9k|HxJ`i^Onr%EupeXG=~R&)dJ~
zk1TJ<kMh**Ui_HMz#$dTg?ED_ny~q0WKU*RO@&qEPmz>3MDlSQ4{CZmG`23OQuizO
zwbErjmd8EYAzy@~uNv>k0li?8x_1y!UGi!Zv%Y%D$0@U1j5qXhn;SUDZgP&Zsgl=l
z4mW-FW6&eqBX-p`>9=q^mXYEYjesxww{L$^Moqs|dQ&O#%@^X<A7cgcALiyZs;ZAS
zTj@2V@0sX6rQRw1yKu94ora#pXH4o)I(_Eohj&h|PODV{7PS8st%eEl`29bk<%O*~
zqT`))imL>th058`{R$QjoWj*T7ro;u77!L#;T3zkde{mFa*_?_#k@7n1&(8_>)Ol@
zpUF2{5A2UDGf*4lHHrJL!9eEqq<Ke1fuqVpmRpA$i_!|WvcC`Z<<Kp<BQAzLM#ZBE
z@3$bQKdmE?z%?9x@^b%V#za;s;%{xIpP$B7bb*EQCqT;;y%J8W9$$5O9zJ7X^P_!n
zI$Rum`qJfQqF;8kz~K?=#WOVps>Ddmqxh&Xdl4@;oOT2c`1{T2#!|*RC)ylgQ)W5G
zr#`RPZ2C%<n#Ps~H|&QET@NQu@=m6D-mm&)U<4H$2`MMsnl4&!=pmyT-|V-8@s^nq
zor161Cm18A#h`h3-*%|{ob~*CE22iU1zq3DfDVQuM9~pNpG(T8`vN1luO6BI(06sW
zDC6h997WU5^8J%S<#X0)08=@yB&}Yg4bv7JSvxdq@-d~g-6FuLimahqylM@_dLveY
zzJJQ)Q_mEbDj`R~*u98bCA+wByTQs5-&;nB@Qnl<0u6GreZZqtM8(`7#GKuRv81M~
zqGyTDtq+cvLn8~u%09wS)n^tKLg{`qtB0?B1Wz|BA?*gJHqB*oEAmz3Q<}b4l7@VD
zUo!l=EgXJVTCwt3So`m3e1{@v6%B_f>|^Sqrqfw^gWwS4^qM6AZ;_TD25oTj&Gl-4
zY@GZndXXN&5z=xL0VGlx11Nb6Tt|AsE<iYRsiTBby$P?4_(*1YVYQUyYIb06^yvYg
z1GE@_zM0$wb{C3n${2zYgIHKw8w2?hxJ#YLiR+{x)f8%Z?wsQ4Q|T*Q?d!lbkf?wV
zLJ8<nK<E~9*_!{FmzGsup8~vTGyrQ3HQF<#)Yf)(dDYdYUx3mOs;2;dzZU~?&Eed8
zLPA3OwE-G5V2nb;)Sw%g#d!VL#;$gqUO-Ur3{mRa3kT3_$-Q|gL#o=-9!W!Qu&`bu
zTc{r>=gOjJS=YX7WZ{><3^M{|9H<gNUlq`?M8m)kIGFv?h|4AX#@m9&Htp@jd5`7V
ze?CC4dU#=cGasbQRY;c?7ylQipg<S{qX5<V0TWZfl$j^&5DN>&8E>#CUL(=RAA1@|
zh&#y%fYxMS-Kk;2`;kAMq5h{RoR4_V{}um7{1%P)Gt|M{|9)V9K#E{0s7LcwMZJa|
zk=vX!ZHYQzYVVb(ZWs#X!UJ+m(2Z4lXL1y92*6=LB0B;n67<IBBx)Z>1OEmiMLwQ5
zz9SR^hL<jRRZXEzVxQ6QpJVY61d1O3ae$P_YGH@}g$9@A3iBO**{^_l?Xd)6)3y1N
zI*3j2EMo@!(I2^@zC@;FX-f*l=(!d2ZxHDI-u>?fsmL3jL39BG_%PtgB8y8(BB19i
zvzu$QO^<zSgUORP<fI>c-sHx3$Lz?(C-Mdmre%O{SCxn{2iz!>AmJRbX<=h?)A-fY
zd-+^0%0FMOSy^DD8QO4iSa({tzcbsA48l3|PcboVz$knH%r=j!CXaI$Cnq*w%^vU0
zDQIbtM@2<(<ZPqSD<|PtzeW<rQ+~uQns!gY2_%Nj8(P1-s|Zm_5Jvq5(tF)?{<yN9
zUVjwbOA=uJ0f%&QQWIFg?m#bGUt4>?!$b55md^3NrHdlh-rV$KDSJ8PJEii<U6}#6
z9?*eQ+ek7+(F7>_bTonTMjY-wK0ZF`MBPcCW-;Um5cXk)0v=lf;-TQk$W>tbF>rB>
zG~p!<x#&kfmc!)viahk|$Wt0KHd1q->ops|a&D2b;!Db1u=gaTvZHCUKR447ZVVQ2
z+tUYf&=wNOup*}S-F%RgSQgyGzIa-xg<MbuN+;mvYsU-1Bg<XtAW^q<dpL^&R9tOx
zju&=5JlFYKk^cED2mC)e(?Q)!t@&!5EH=5V>ogl&a*BGyfUmFkTl>sY52)a-YX4i3
zER;~qMh6wWR=QI>$&;Gu{bxZu8vp8B+U$@*7IVG}PKQxbbk?|Wm-hc7uKSRJ#G}CM
zD-t3Qc^OA!cYEWYNu%GJ=mAgiw$|<kAV`7^Q&Ep&XcOTkl$P;SPl?b;Ku2*LXeJ?9
z^9T&P*W{`4fw>lDqgwJxgg{toR7nSg$k>=F{J-|8=pW}pIT!+Vzb!fHu){iTJdk6E
zFq#kB6jkr=EsLY<*(UF~s$h-SO-m`bzIgZyMux$`!S%pty!L;s5=-|o0@xA3K(Epa
zLNV!Y6VU%_!v}n6F;H(2Fvw$d*@3zR?!JJ40O${JyKpiJn1I{a*4BncOiaAD%jcjk
zAt~u{Q3sS;;A;JMVH3{=UEp{ma#;*%#5eK2QOPMP!g>dk@wY^@&9lfx*axsRL6l+m
z_jlClU*3X(SMu^WKoIUQEI+Ef_H(~{64LS>9TVkyrWG{`@#IDS?V=QR5$D}?KK{Xc
zv_M5uLqmg2EwkZhNG54~6MOPtYHI2Sgx0?{XGyshn7C9UW2Ji3BhH{^(|p3ncxywT
zv2PP$$qFCf*GkW|n1tN*9WrRjurr1y@*)>hnl^}y(N$}2_&6xaMu4B)ck0<JbgfpT
zp82Dy>g3GPug2T2#uwx8R7B<KNfJ^0aSlj{K&8Y^@?o4jqp~v2-`_vuN3(|z6>JbB
zPhuJ(PTWX$p)@<42}BO0=;zO$pC5&6zu}TY>CQkc^!A&`Pf-yqwyGsw6;;)=;kh+?
zzBLFj7M7iq;)>kdfYZ}c&c<$#u+hOSVC?W!QMs>LI6B8Xq6RuLVDX4eG=lyVjz=NF
z(pl>=GjD{pKL_F}{2v65`T6<LV@8T~`$@9&Aky%wyv)s|*Je>0NH0g}R}|!%;eNvB
zt7mpHinA*!4pXq))?Mlzv$A$9J4U6X$b#Tiv5L|uHUE{2OmJA(Lk^D5Akebr(Ha<A
zh4(#v3|y%9fBza=T3Pwkrl;#yO#u-TF0BNV-;vb^;He<_0oe~217xGh%F1sicllH0
zBdA5?B_wVFNq1y)6wZBS&6W#_i*vvxP#=IF=&6!B3;e^}va(O$>Bwi_ynlF;lab{Y
z6wty71_l&iITu?gqi|y6?p`Ch2TZ4ak3>+#{-@4eqOP5>1*HyrJ8;HrSHEUBFWFdI
z|1I^sdV&UGzek{k@otO&p^&1kZp4K3N`&Cf+Trqx@z^7jtT&4hrJM0?yacM0uKV(R
zD3%(P)*6Z>X;r|3yuP|v-`e_nF^3d^NP^1{fTNrar(d@X<;VcVb{@!3z%it9o4=8j
zoqgfu4XXgX2?%+lL2ts&&JMgL5K(pjBby@T$+zQU7l@V4{(du%1%N}$AD&f%I1^aT
z8;1z7@zgy#-wQ3EK11CFrY}C|$w1?9Kc;RMM6N!FLOqAG<#;15GWX5mDj;_Ollr`)
zM=<fJR^=XW9UW$z$CW@q1V97Ac)2^CfQrgT)nu~CS9Ei%*a*Z<kb`x-5h@@Fvq|*G
z&dU1mA`k~9`%7Rq0^#%{9-iZ(Dm%~{+BIGAfQs|%aP{r{&l|8~AZIQIklXSVa-{RQ
z0wWSdpucnH&f3<NG2jGd?P|6RlXEW-lpZA{H1upR(x)wywK1rF)o8Zf9k^zSuU~`9
z{)=LQhh+V`i}NKT;~j`MZL1eB;4FhY##!5;{XWPipX940qcjqr`cP+%%a~pVf$k>|
zY=aA?>&SUW|3{a<k=#3G<Fu?VB`f<EG@Y&6F43U;vZ<Q+1;R=Yu&OJL<rOf=#Iro&
z<RnP$l>{0#sJ<AOn3&AbFWcMNPD|gN?9TN=^=t+8NM~;^cGo$VoDoXGMkzp8HzN-g
z5IrA^1pnLM(Noa7bNe<*^aAHDkYoD4TMVmLT9ZOP%DdDBzkd;cwE$Jrdg9y1i=Dbi
z@Ms8tzFJvo?<0Jc_)v%+*da<3rC#f_ZqYqL4zp;wM~^=C^vGITTDom~%{w=~*g$wd
zf(Ly+XiQQe22gTfuy5Zv4$$@&?fD>}Q0zCb{K3ErTu=lO>~SbiC_s&cM2kmF-S2%e
z(=k@8^_Yca1qBg-`CeIBS<;_A%F1`MVnP-m0X8;vf^2R^(GnO5B~8|{Yv5+=#$;q+
z$^G)hUoM&FX3lLXP#$@oJAvr3v#)RGa2cIK)F%b>GWV#dslR_ZuLXUx5EU-W3c(#=
zy&ulEOh5q!YW@nR_1BR3b8|70dx^d~LrNSP9+sAo$%d>1IS9#_s}JOL;2^#Q7C1`b
zZ)U~-3p({`%s^a&fW_$SU>Wo#Zvm}NPEEnX14OrZD4`%k3LldI21ca~&=|jb`7(Xv
zW4ocj`cMvN?2?j`uU6w5|LlN{6jHnWdQSYJOC5Z8H+ox3MWr2%ofA1s?wFXEz=0a1
zEG5F_qFo9C0)iTNXaT&<4CwNJYC+!CcTPJ9c6JhAPXwX0lOkxO%*`1=kEf>>IlhS~
zr8B?50g5buvQYJmKn^IT(E%#z4;@RlCxeOSn=z^XE(B7e1Z?X(|2XJDExboVL$ftg
z7Y`wHa(0kum4A&YzC6hw3X@VCb9fHV0ss%%ny!fkA>KDoG(b#l(g&o$lL|oX9R?_y
zmpA^AOdLvIN&w@1a23Td4IQ0cxNRTDH%-7^sLR8n3Lx5vVi4<20w`(TP7~5+r~XhL
zCD?%g{0(XdKcdfI!Fnj1ghWIKBgzU&dU^_~s$&+I%Dvs)OSrzL=%9T<$tyvYVpOg!
z1)`SedwFk{!g&usb{Y-hy~c}~-Jy|@w%G<xbl}_766m;MVw1JO#ofW?;J!E@{fzIH
zu&{W1xY}>F*cMz5(hY5w2@|uK8$VYMmDJVY85Il)B4-bw1{7)j2q`S7-6)Ey1KlGk
zDnXM3lo$AVlpCQ$Jb<5jfUjPrUaPB)WyxCfwte*w;ckcJU@*vqEjQdZ;gex8DS<q8
z&S-LYV4Zv@!B&Y#WIoKg1<(v{HBGxR^hW0<;&>hJAu!Zaa9C;E&Yt}vj>QM^LsUBh
z)p+9zvc!T6L|6RT3-kpA1s-0+e?hN#%Ths%<~$T@l?5XO<cy!=<6pwUOB;^K02Z5c
z$AOAZZvd2~py)g8wex8OHBET`>Jmyk2Wv*v%c3G?&?ar~2DpRRngAbP44&WtsT5TG
z3ywAFW%@Uga?2f-B>x{h)Y)2gBs@9;GG8c=^5FphewetXz=pkQkcQ4;`hCz1l=k&W
zN%aR*`2wLhG+romFnD}}=LsyY<MQl3VHHY;mz0Y#G(1}bc(i%nzsG>4@-@5w*lqwp
zs|f`d9>9Q-=~z!!M`4rm*h3LPNyUIuFAicHcpd{V>QVY~kO>bkG(`KavEF-N_!BH%
zI*xkA0n|Tcjc+K`9!RjEvR(V^0bT&X6e{b3e?rXynGPf)h3C)z0hQ7Y5n56%K1jXb
zAr4R#L5GiOPN0v4&w|{9h($YS)Uw^Q0v|pBjAT*U+cbR<M<^nnmDLltQ(wM(Df7NS
zHPz6LLeq&^I^GSB{<vcOsZp$odT2_9m|svx2q<{)C@FgZ-`NYK_N4Gf173i~cPRa*
zz6VJrXhIq2=>x!iAUkV^?EMGo;nSy2fr64bzR8~4%SV9&Z#(ylzJ-r>8|6SZqUf1X
x{|urZbmc>D$4L(@-~WA9&;KuNaDYTB<VJpAID4fJ&xS&imXH@Od9MHA{{T+DGAIB5

diff --git a/_sources/auto_examples/plot_compare_time.rst.txt b/_sources/auto_examples/plot_compare_time.rst.txt
index 3b5f128ce..e7cc7a330 100644
--- a/_sources/auto_examples/plot_compare_time.rst.txt
+++ b/_sources/auto_examples/plot_compare_time.rst.txt
@@ -37,7 +37,7 @@ Compare time to solve large scale Lasso problems with scikit-learn.
 
  .. code-block:: none
 
-
    file_sizes:   0%|                                   | 0.00/26.8M [00:00<?, ?B/s]
    file_sizes:   0%|                           | 24.6k/26.8M [00:00<03:31, 127kB/s]
    file_sizes:   0%|                           | 49.2k/26.8M [00:00<03:31, 126kB/s]
    file_sizes:   0%|                            | 106k/26.8M [00:00<02:11, 203kB/s]
    file_sizes:   1%|▏                           | 221k/26.8M [00:00<01:14, 355kB/s]
    file_sizes:   2%|▍                           | 451k/26.8M [00:00<00:40, 652kB/s]
    file_sizes:   3%|▉                          | 909k/26.8M [00:01<00:21, 1.23MB/s]
    file_sizes:   7%|█▊                        | 1.83M/26.8M [00:01<00:10, 2.37MB/s]
    file_sizes:  14%|███▌                      | 3.66M/26.8M [00:01<00:05, 4.59MB/s]
    file_sizes:  20%|█████                     | 5.23M/26.8M [00:01<00:03, 5.69MB/s]
    file_sizes:  29%|███████▋                  | 7.86M/26.8M [00:01<00:02, 7.99MB/s]
    file_sizes:  35%|█████████▏                | 9.43M/26.8M [00:02<00:02, 8.06MB/s]
    file_sizes:  45%|███████████▋              | 12.1M/26.8M [00:02<00:01, 9.63MB/s]
    file_sizes:  51%|█████████████▏            | 13.6M/26.8M [00:02<00:01, 9.24MB/s]
    file_sizes:  61%|███████████████▊          | 16.2M/26.8M [00:02<00:01, 10.5MB/s]
    file_sizes:  70%|██████████████████▎       | 18.9M/26.8M [00:02<00:00, 11.4MB/s]
    file_sizes:  80%|████████████████████▊     | 21.5M/26.8M [00:03<00:00, 12.0MB/s]
    file_sizes:  90%|███████████████████████▍  | 24.1M/26.8M [00:03<00:00, 12.5MB/s]
    file_sizes: 100%|██████████████████████████| 26.8M/26.8M [00:03<00:00, 7.99MB/s]
+
    file_sizes:   0%|                                   | 0.00/26.8M [00:00<?, ?B/s]
    file_sizes:   0%|                           | 24.6k/26.8M [00:00<03:34, 125kB/s]
    file_sizes:   0%|                           | 49.2k/26.8M [00:00<03:35, 124kB/s]
    file_sizes:   0%|                            | 106k/26.8M [00:00<02:13, 200kB/s]
    file_sizes:   1%|▏                           | 221k/26.8M [00:00<01:16, 349kB/s]
    file_sizes:   2%|▍                           | 451k/26.8M [00:00<00:41, 640kB/s]
    file_sizes:   3%|▉                          | 909k/26.8M [00:01<00:21, 1.21MB/s]
    file_sizes:   7%|█▊                        | 1.83M/26.8M [00:01<00:10, 2.32MB/s]
    file_sizes:  14%|███▌                      | 3.66M/26.8M [00:01<00:05, 4.51MB/s]
    file_sizes:  20%|█████                     | 5.23M/26.8M [00:01<00:03, 5.57MB/s]
    file_sizes:  25%|██████▌                   | 6.81M/26.8M [00:01<00:03, 6.31MB/s]
    file_sizes:  35%|█████████▏                | 9.43M/26.8M [00:02<00:02, 8.31MB/s]
    file_sizes:  41%|██████████▋               | 11.0M/26.8M [00:02<00:01, 8.22MB/s]
    file_sizes:  47%|████████████▏             | 12.6M/26.8M [00:02<00:01, 8.16MB/s]
    file_sizes:  53%|█████████████▋            | 14.1M/26.8M [00:02<00:01, 8.13MB/s]
    file_sizes:  63%|████████████████▎         | 16.8M/26.8M [00:02<00:01, 9.60MB/s]
    file_sizes:  72%|██████████████████▊       | 19.4M/26.8M [00:03<00:00, 10.6MB/s]
    file_sizes:  78%|████████████████████▎     | 21.0M/26.8M [00:03<00:00, 9.90MB/s]
    file_sizes:  88%|██████████████████████▉   | 23.6M/26.8M [00:03<00:00, 10.9MB/s]
    file_sizes:  98%|█████████████████████████▍| 26.2M/26.8M [00:03<00:00, 11.5MB/s]
    file_sizes: 100%|██████████████████████████| 26.8M/26.8M [00:03<00:00, 7.08MB/s]
 
 
 
@@ -138,7 +138,7 @@ Compare time to solve large scale Lasso problems with scikit-learn.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 45.604 seconds)
+   **Total running time of the script:** (0 minutes 47.620 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_compare_time.py:
diff --git a/_sources/auto_examples/plot_group_logistic_regression.rst.txt b/_sources/auto_examples/plot_group_logistic_regression.rst.txt
index 29d5a261f..7be35fcfa 100644
--- a/_sources/auto_examples/plot_group_logistic_regression.rst.txt
+++ b/_sources/auto_examples/plot_group_logistic_regression.rst.txt
@@ -158,7 +158,7 @@ Fit check that groups are either all 0 or all non zero
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 14.785 seconds)
+   **Total running time of the script:** (0 minutes 15.201 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_group_logistic_regression.py:
diff --git a/_sources/auto_examples/plot_lasso_vs_weighted.rst.txt b/_sources/auto_examples/plot_lasso_vs_weighted.rst.txt
index 4d6aa7250..6ec3467c9 100644
--- a/_sources/auto_examples/plot_lasso_vs_weighted.rst.txt
+++ b/_sources/auto_examples/plot_lasso_vs_weighted.rst.txt
@@ -84,7 +84,7 @@ Illustrate the importance of feature normalization when penalizing.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 3.731 seconds)
+   **Total running time of the script:** (0 minutes 3.903 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_lasso_vs_weighted.py:
diff --git a/_sources/auto_examples/plot_logreg_various_penalties.rst.txt b/_sources/auto_examples/plot_logreg_various_penalties.rst.txt
index 6a4723389..2f6e3ecd1 100644
--- a/_sources/auto_examples/plot_logreg_various_penalties.rst.txt
+++ b/_sources/auto_examples/plot_logreg_various_penalties.rst.txt
@@ -112,7 +112,7 @@ Illustrate the modularity of ``skglm`` by using ``GeneralizedLinearEstimator`` w
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 8.793 seconds)
+   **Total running time of the script:** (0 minutes 8.967 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_logreg_various_penalties.py:
diff --git a/_sources/auto_examples/plot_pen_prox.rst.txt b/_sources/auto_examples/plot_pen_prox.rst.txt
index 1d96cd61f..06d288cbc 100644
--- a/_sources/auto_examples/plot_pen_prox.rst.txt
+++ b/_sources/auto_examples/plot_pen_prox.rst.txt
@@ -77,7 +77,7 @@ Illustrate the value and proximal operators of some sparse penalties.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 2.314 seconds)
+   **Total running time of the script:** (0 minutes 2.487 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_pen_prox.py:
diff --git a/_sources/auto_examples/plot_reweighted_l1.rst.txt b/_sources/auto_examples/plot_reweighted_l1.rst.txt
index d7494839a..ff0337d83 100644
--- a/_sources/auto_examples/plot_reweighted_l1.rst.txt
+++ b/_sources/auto_examples/plot_reweighted_l1.rst.txt
@@ -131,7 +131,7 @@ direct proximal computation and iterative reweighting.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 5.021 seconds)
+   **Total running time of the script:** (0 minutes 5.326 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_reweighted_l1.py:
diff --git a/_sources/auto_examples/plot_sparse_group_lasso.rst.txt b/_sources/auto_examples/plot_sparse_group_lasso.rst.txt
index a11fca7f3..9b6c5b908 100644
--- a/_sources/auto_examples/plot_sparse_group_lasso.rst.txt
+++ b/_sources/auto_examples/plot_sparse_group_lasso.rst.txt
@@ -173,14 +173,14 @@ some values are 0 too
  .. code-block:: none
 
 
-    ([<matplotlib.axis.YTick object at 0x7fb9b50f1970>, <matplotlib.axis.YTick object at 0x7fb9b4011e80>, <matplotlib.axis.YTick object at 0x7fb9b4e21490>], [Text(0, 0, '0'), Text(0, 1, '1'), Text(0, 2, '2')])
+    ([<matplotlib.axis.YTick object at 0x7fef611c0cd0>, <matplotlib.axis.YTick object at 0x7fef60f50df0>, <matplotlib.axis.YTick object at 0x7fef6867a6d0>], [Text(0, 0, '0'), Text(0, 1, '1'), Text(0, 2, '2')])
 
 
 
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 5.547 seconds)
+   **Total running time of the script:** (0 minutes 5.796 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_sparse_group_lasso.py:
diff --git a/_sources/auto_examples/plot_sparse_recovery.rst.txt b/_sources/auto_examples/plot_sparse_recovery.rst.txt
index 46ed1d92b..d88c2cacf 100644
--- a/_sources/auto_examples/plot_sparse_recovery.rst.txt
+++ b/_sources/auto_examples/plot_sparse_recovery.rst.txt
@@ -190,7 +190,7 @@ Illustrate the superior performance of penalties for sparse recovery.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 8.910 seconds)
+   **Total running time of the script:** (0 minutes 8.881 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_sparse_recovery.py:
diff --git a/_sources/auto_examples/plot_survival_analysis.rst.txt b/_sources/auto_examples/plot_survival_analysis.rst.txt
index 57a44a028..9f17a742c 100644
--- a/_sources/auto_examples/plot_survival_analysis.rst.txt
+++ b/_sources/auto_examples/plot_survival_analysis.rst.txt
@@ -382,7 +382,7 @@ with more than x100 less time!
 
  .. code-block:: none
 
-    speed up ratio: 879
+    speed up ratio: 930
 
 
 
@@ -566,7 +566,7 @@ Finally, let's compare the timings of both solvers
 
  .. code-block:: none
 
-    speed up ratio: 787
+    speed up ratio: 1042
 
 
 
@@ -579,7 +579,7 @@ even for the Efron estimate.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 52.615 seconds)
+   **Total running time of the script:** (0 minutes 57.445 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_survival_analysis.py:
diff --git a/_sources/auto_examples/plot_ucurve.rst.txt b/_sources/auto_examples/plot_ucurve.rst.txt
index 2f7cbff60..f20272f7c 100644
--- a/_sources/auto_examples/plot_ucurve.rst.txt
+++ b/_sources/auto_examples/plot_ucurve.rst.txt
@@ -69,7 +69,7 @@ We also retrain 2000 samples in training dataset.
 
  .. code-block:: none
 
-
    file_sizes:   0%|                                   | 0.00/13.7M [00:00<?, ?B/s]
    file_sizes:   0%|                           | 24.6k/13.7M [00:00<01:51, 123kB/s]
    file_sizes:   0%|                           | 49.2k/13.7M [00:00<01:51, 122kB/s]
    file_sizes:   1%|▏                           | 106k/13.7M [00:00<01:09, 196kB/s]
    file_sizes:   2%|▍                           | 221k/13.7M [00:00<00:39, 343kB/s]
    file_sizes:   3%|▉                           | 451k/13.7M [00:01<00:21, 630kB/s]
    file_sizes:   7%|█▊                         | 909k/13.7M [00:01<00:10, 1.18MB/s]
    file_sizes:  13%|███▍                      | 1.83M/13.7M [00:01<00:05, 2.28MB/s]
    file_sizes:  27%|██████▉                   | 3.66M/13.7M [00:01<00:02, 4.41MB/s]
    file_sizes:  38%|█████████▉                | 5.23M/13.7M [00:01<00:01, 5.46MB/s]
    file_sizes:  50%|████████████▉             | 6.81M/13.7M [00:02<00:01, 6.18MB/s]
    file_sizes:  61%|███████████████▊          | 8.38M/13.7M [00:02<00:00, 6.69MB/s]
    file_sizes:  80%|████████████████████▊     | 11.0M/13.7M [00:02<00:00, 8.51MB/s]
    file_sizes:  92%|███████████████████████▊  | 12.6M/13.7M [00:02<00:00, 8.36MB/s]
    file_sizes: 100%|██████████████████████████| 13.7M/13.7M [00:02<00:00, 5.20MB/s]
+
    file_sizes:   0%|                                   | 0.00/13.7M [00:00<?, ?B/s]
    file_sizes:   0%|                           | 24.6k/13.7M [00:00<01:49, 125kB/s]
    file_sizes:   0%|                           | 49.2k/13.7M [00:00<01:50, 124kB/s]
    file_sizes:   1%|▏                           | 106k/13.7M [00:00<01:08, 199kB/s]
    file_sizes:   2%|▍                           | 221k/13.7M [00:00<00:38, 349kB/s]
    file_sizes:   3%|▉                           | 451k/13.7M [00:00<00:20, 639kB/s]
    file_sizes:   7%|█▊                         | 909k/13.7M [00:01<00:10, 1.21MB/s]
    file_sizes:  13%|███▍                      | 1.83M/13.7M [00:01<00:05, 2.32MB/s]
    file_sizes:  27%|██████▉                   | 3.66M/13.7M [00:01<00:02, 4.50MB/s]
    file_sizes:  38%|█████████▉                | 5.23M/13.7M [00:01<00:01, 5.56MB/s]
    file_sizes:  50%|████████████▉             | 6.81M/13.7M [00:01<00:01, 6.29MB/s]
    file_sizes:  61%|███████████████▊          | 8.38M/13.7M [00:02<00:00, 6.81MB/s]
    file_sizes:  80%|████████████████████▊     | 11.0M/13.7M [00:02<00:00, 8.65MB/s]
    file_sizes:  92%|███████████████████████▊  | 12.6M/13.7M [00:02<00:00, 8.48MB/s]
    file_sizes: 100%|██████████████████████████| 13.7M/13.7M [00:02<00:00, 5.29MB/s]
 
 
 
@@ -150,7 +150,7 @@ Notice the "sweet spot" at around ``1e-4``, which sits at the boundary between u
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 13.278 seconds)
+   **Total running time of the script:** (0 minutes 13.957 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_ucurve.py:
diff --git a/_sources/auto_examples/plot_zero_weights_lasso.rst.txt b/_sources/auto_examples/plot_zero_weights_lasso.rst.txt
index 675313287..ea55daa56 100644
--- a/_sources/auto_examples/plot_zero_weights_lasso.rst.txt
+++ b/_sources/auto_examples/plot_zero_weights_lasso.rst.txt
@@ -84,7 +84,7 @@ allowing it to not compute the dual and handle 0 weights.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** (0 minutes 0.342 seconds)
+   **Total running time of the script:** (0 minutes 0.374 seconds)
 
 
 .. _sphx_glr_download_auto_examples_plot_zero_weights_lasso.py:
diff --git a/_sources/auto_examples/sg_execution_times.rst.txt b/_sources/auto_examples/sg_execution_times.rst.txt
index 8a7de6b28..59af21c17 100644
--- a/_sources/auto_examples/sg_execution_times.rst.txt
+++ b/_sources/auto_examples/sg_execution_times.rst.txt
@@ -6,7 +6,7 @@
 
 Computation times
 =================
-**02:40.942** total execution time for 11 files **from auto_examples**:
+**02:49.958** total execution time for 11 files **from auto_examples**:
 
 .. container::
 
@@ -33,35 +33,35 @@ Computation times
      - Time
      - Mem (MB)
    * - :ref:`sphx_glr_auto_examples_plot_survival_analysis.py` (``plot_survival_analysis.py``)
-     - 00:52.615
+     - 00:57.445
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_compare_time.py` (``plot_compare_time.py``)
-     - 00:45.604
+     - 00:47.620
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_group_logistic_regression.py` (``plot_group_logistic_regression.py``)
-     - 00:14.785
+     - 00:15.201
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_ucurve.py` (``plot_ucurve.py``)
-     - 00:13.278
-     - 0.0
-   * - :ref:`sphx_glr_auto_examples_plot_sparse_recovery.py` (``plot_sparse_recovery.py``)
-     - 00:08.910
+     - 00:13.957
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_logreg_various_penalties.py` (``plot_logreg_various_penalties.py``)
-     - 00:08.793
+     - 00:08.967
+     - 0.0
+   * - :ref:`sphx_glr_auto_examples_plot_sparse_recovery.py` (``plot_sparse_recovery.py``)
+     - 00:08.881
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_sparse_group_lasso.py` (``plot_sparse_group_lasso.py``)
-     - 00:05.547
+     - 00:05.796
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_reweighted_l1.py` (``plot_reweighted_l1.py``)
-     - 00:05.021
+     - 00:05.326
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_lasso_vs_weighted.py` (``plot_lasso_vs_weighted.py``)
-     - 00:03.731
+     - 00:03.903
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_pen_prox.py` (``plot_pen_prox.py``)
-     - 00:02.314
+     - 00:02.487
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_zero_weights_lasso.py` (``plot_zero_weights_lasso.py``)
-     - 00:00.342
+     - 00:00.374
      - 0.0
diff --git a/_sources/sg_execution_times.rst.txt b/_sources/sg_execution_times.rst.txt
index 1ad37a49c..cf2ddef90 100644
--- a/_sources/sg_execution_times.rst.txt
+++ b/_sources/sg_execution_times.rst.txt
@@ -6,7 +6,7 @@
 
 Computation times
 =================
-**02:40.942** total execution time for 11 files **from all galleries**:
+**02:49.958** total execution time for 11 files **from all galleries**:
 
 .. container::
 
@@ -33,35 +33,35 @@ Computation times
      - Time
      - Mem (MB)
    * - :ref:`sphx_glr_auto_examples_plot_survival_analysis.py` (``../examples/plot_survival_analysis.py``)
-     - 00:52.615
+     - 00:57.445
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_compare_time.py` (``../examples/plot_compare_time.py``)
-     - 00:45.604
+     - 00:47.620
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_group_logistic_regression.py` (``../examples/plot_group_logistic_regression.py``)
-     - 00:14.785
+     - 00:15.201
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_ucurve.py` (``../examples/plot_ucurve.py``)
-     - 00:13.278
-     - 0.0
-   * - :ref:`sphx_glr_auto_examples_plot_sparse_recovery.py` (``../examples/plot_sparse_recovery.py``)
-     - 00:08.910
+     - 00:13.957
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_logreg_various_penalties.py` (``../examples/plot_logreg_various_penalties.py``)
-     - 00:08.793
+     - 00:08.967
+     - 0.0
+   * - :ref:`sphx_glr_auto_examples_plot_sparse_recovery.py` (``../examples/plot_sparse_recovery.py``)
+     - 00:08.881
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_sparse_group_lasso.py` (``../examples/plot_sparse_group_lasso.py``)
-     - 00:05.547
+     - 00:05.796
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_reweighted_l1.py` (``../examples/plot_reweighted_l1.py``)
-     - 00:05.021
+     - 00:05.326
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_lasso_vs_weighted.py` (``../examples/plot_lasso_vs_weighted.py``)
-     - 00:03.731
+     - 00:03.903
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_pen_prox.py` (``../examples/plot_pen_prox.py``)
-     - 00:02.314
+     - 00:02.487
      - 0.0
    * - :ref:`sphx_glr_auto_examples_plot_zero_weights_lasso.py` (``../examples/plot_zero_weights_lasso.py``)
-     - 00:00.342
+     - 00:00.374
      - 0.0
diff --git a/auto_examples/plot_compare_time.html b/auto_examples/plot_compare_time.html
index 38e2d5b60..e98b7a3e7 100644
--- a/auto_examples/plot_compare_time.html
+++ b/auto_examples/plot_compare_time.html
@@ -354,24 +354,26 @@
 <span id="sphx-glr-auto-examples-plot-compare-time-py"></span><h1>Timing comparison with scikit-learn for Lasso<a class="headerlink" href="#timing-comparison-with-scikit-learn-for-lasso" title="Permalink to this heading">¶</a></h1>
 <p>Compare time to solve large scale Lasso problems with scikit-learn.</p>
 <img src="../_images/sphx_glr_plot_compare_time_001.png" srcset="../_images/sphx_glr_plot_compare_time_001.png" alt="lasso, enet" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>file_sizes:   0%|                                   | 0.00/26.8M [00:00&lt;?, ?B/s]
-file_sizes:   0%|                           | 24.6k/26.8M [00:00&lt;03:31, 127kB/s]
-file_sizes:   0%|                           | 49.2k/26.8M [00:00&lt;03:31, 126kB/s]
-file_sizes:   0%|                            | 106k/26.8M [00:00&lt;02:11, 203kB/s]
-file_sizes:   1%|▏                           | 221k/26.8M [00:00&lt;01:14, 355kB/s]
-file_sizes:   2%|▍                           | 451k/26.8M [00:00&lt;00:40, 652kB/s]
-file_sizes:   3%|▉                          | 909k/26.8M [00:01&lt;00:21, 1.23MB/s]
-file_sizes:   7%|█▊                        | 1.83M/26.8M [00:01&lt;00:10, 2.37MB/s]
-file_sizes:  14%|███▌                      | 3.66M/26.8M [00:01&lt;00:05, 4.59MB/s]
-file_sizes:  20%|█████                     | 5.23M/26.8M [00:01&lt;00:03, 5.69MB/s]
-file_sizes:  29%|███████▋                  | 7.86M/26.8M [00:01&lt;00:02, 7.99MB/s]
-file_sizes:  35%|█████████▏                | 9.43M/26.8M [00:02&lt;00:02, 8.06MB/s]
-file_sizes:  45%|███████████▋              | 12.1M/26.8M [00:02&lt;00:01, 9.63MB/s]
-file_sizes:  51%|█████████████▏            | 13.6M/26.8M [00:02&lt;00:01, 9.24MB/s]
-file_sizes:  61%|███████████████▊          | 16.2M/26.8M [00:02&lt;00:01, 10.5MB/s]
-file_sizes:  70%|██████████████████▎       | 18.9M/26.8M [00:02&lt;00:00, 11.4MB/s]
-file_sizes:  80%|████████████████████▊     | 21.5M/26.8M [00:03&lt;00:00, 12.0MB/s]
-file_sizes:  90%|███████████████████████▍  | 24.1M/26.8M [00:03&lt;00:00, 12.5MB/s]
-file_sizes: 100%|██████████████████████████| 26.8M/26.8M [00:03&lt;00:00, 7.99MB/s]
+file_sizes:   0%|                           | 24.6k/26.8M [00:00&lt;03:34, 125kB/s]
+file_sizes:   0%|                           | 49.2k/26.8M [00:00&lt;03:35, 124kB/s]
+file_sizes:   0%|                            | 106k/26.8M [00:00&lt;02:13, 200kB/s]
+file_sizes:   1%|▏                           | 221k/26.8M [00:00&lt;01:16, 349kB/s]
+file_sizes:   2%|▍                           | 451k/26.8M [00:00&lt;00:41, 640kB/s]
+file_sizes:   3%|▉                          | 909k/26.8M [00:01&lt;00:21, 1.21MB/s]
+file_sizes:   7%|█▊                        | 1.83M/26.8M [00:01&lt;00:10, 2.32MB/s]
+file_sizes:  14%|███▌                      | 3.66M/26.8M [00:01&lt;00:05, 4.51MB/s]
+file_sizes:  20%|█████                     | 5.23M/26.8M [00:01&lt;00:03, 5.57MB/s]
+file_sizes:  25%|██████▌                   | 6.81M/26.8M [00:01&lt;00:03, 6.31MB/s]
+file_sizes:  35%|█████████▏                | 9.43M/26.8M [00:02&lt;00:02, 8.31MB/s]
+file_sizes:  41%|██████████▋               | 11.0M/26.8M [00:02&lt;00:01, 8.22MB/s]
+file_sizes:  47%|████████████▏             | 12.6M/26.8M [00:02&lt;00:01, 8.16MB/s]
+file_sizes:  53%|█████████████▋            | 14.1M/26.8M [00:02&lt;00:01, 8.13MB/s]
+file_sizes:  63%|████████████████▎         | 16.8M/26.8M [00:02&lt;00:01, 9.60MB/s]
+file_sizes:  72%|██████████████████▊       | 19.4M/26.8M [00:03&lt;00:00, 10.6MB/s]
+file_sizes:  78%|████████████████████▎     | 21.0M/26.8M [00:03&lt;00:00, 9.90MB/s]
+file_sizes:  88%|██████████████████████▉   | 23.6M/26.8M [00:03&lt;00:00, 10.9MB/s]
+file_sizes:  98%|█████████████████████████▍| 26.2M/26.8M [00:03&lt;00:00, 11.5MB/s]
+file_sizes: 100%|██████████████████████████| 26.8M/26.8M [00:03&lt;00:00, 7.08MB/s]
 </pre></div>
 </div>
 <div class="line-block">
@@ -463,7 +465,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 45.604 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 47.620 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-compare-time-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/2c9537288a512ce6fe5c8f920df56f33/plot_compare_time.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_compare_time.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_group_logistic_regression.html b/auto_examples/plot_group_logistic_regression.html
index bad543565..77b22e097 100644
--- a/auto_examples/plot_group_logistic_regression.html
+++ b/auto_examples/plot_group_logistic_regression.html
@@ -426,7 +426,7 @@
  [ 0.57737623 -0.25394772  0.15405176]]
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 14.785 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 15.201 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-group-logistic-regression-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/382fdeae60d02217e9c69837e7c101fe/plot_group_logistic_regression.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_group_logistic_regression.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_lasso_vs_weighted.html b/auto_examples/plot_lasso_vs_weighted.html
index d7b2de748..e1f2aa18c 100644
--- a/auto_examples/plot_lasso_vs_weighted.html
+++ b/auto_examples/plot_lasso_vs_weighted.html
@@ -396,7 +396,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 3.731 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 3.903 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-lasso-vs-weighted-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/35ec8c6c8060b614de1bf3cd99484dea/plot_lasso_vs_weighted.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_lasso_vs_weighted.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_logreg_various_penalties.html b/auto_examples/plot_logreg_various_penalties.html
index bea0d4aee..eb1f8a32f 100644
--- a/auto_examples/plot_logreg_various_penalties.html
+++ b/auto_examples/plot_logreg_various_penalties.html
@@ -424,7 +424,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">()</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 8.793 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 8.967 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-logreg-various-penalties-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/c2764ad753df1646865bb3aa9db52704/plot_logreg_various_penalties.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_logreg_various_penalties.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_pen_prox.html b/auto_examples/plot_pen_prox.html
index 84cd68d36..6064dfa55 100644
--- a/auto_examples/plot_pen_prox.html
+++ b/auto_examples/plot_pen_prox.html
@@ -390,7 +390,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 2.314 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 2.487 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-pen-prox-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/ae9288a5d61917f221d6822700f520c2/plot_pen_prox.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_pen_prox.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_reweighted_l1.html b/auto_examples/plot_reweighted_l1.html
index 51ed242b1..1f5740d69 100644
--- a/auto_examples/plot_reweighted_l1.html
+++ b/auto_examples/plot_reweighted_l1.html
@@ -444,7 +444,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 5.021 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 5.326 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-reweighted-l1-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/8de292cb5edb12244bb3c6652c3a58f0/plot_reweighted_l1.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_reweighted_l1.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_sparse_group_lasso.html b/auto_examples/plot_sparse_group_lasso.html
index 658f0c346..8e89fc329 100644
--- a/auto_examples/plot_sparse_group_lasso.html
+++ b/auto_examples/plot_sparse_group_lasso.html
@@ -419,10 +419,10 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">yticks</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">n_groups</span><span class="p">));</span>
 </pre></div>
 </div>
-<img src="../_images/sphx_glr_plot_sparse_group_lasso_001.png" srcset="../_images/sphx_glr_plot_sparse_group_lasso_001.png" alt="Non zero values (in black) in model coefficients" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>([&lt;matplotlib.axis.YTick object at 0x7fb9b50f1970&gt;, &lt;matplotlib.axis.YTick object at 0x7fb9b4011e80&gt;, &lt;matplotlib.axis.YTick object at 0x7fb9b4e21490&gt;], [Text(0, 0, &#39;0&#39;), Text(0, 1, &#39;1&#39;), Text(0, 2, &#39;2&#39;)])
+<img src="../_images/sphx_glr_plot_sparse_group_lasso_001.png" srcset="../_images/sphx_glr_plot_sparse_group_lasso_001.png" alt="Non zero values (in black) in model coefficients" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>([&lt;matplotlib.axis.YTick object at 0x7fef611c0cd0&gt;, &lt;matplotlib.axis.YTick object at 0x7fef60f50df0&gt;, &lt;matplotlib.axis.YTick object at 0x7fef6867a6d0&gt;], [Text(0, 0, &#39;0&#39;), Text(0, 1, &#39;1&#39;), Text(0, 2, &#39;2&#39;)])
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 5.547 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 5.796 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-sparse-group-lasso-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/5c80f1f9f9b119b18c74f23b9d54985c/plot_sparse_group_lasso.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_sparse_group_lasso.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_sparse_recovery.html b/auto_examples/plot_sparse_recovery.html
index 3ebe0075a..4c460a201 100644
--- a/auto_examples/plot_sparse_recovery.html
+++ b/auto_examples/plot_sparse_recovery.html
@@ -498,7 +498,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 8.910 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 8.881 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-sparse-recovery-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/bd97045a660ad082c4a4a4abf0767b0e/plot_sparse_recovery.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_sparse_recovery.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_survival_analysis.html b/auto_examples/plot_survival_analysis.html
index ac4fa1919..f778cd587 100644
--- a/auto_examples/plot_survival_analysis.html
+++ b/auto_examples/plot_survival_analysis.html
@@ -561,7 +561,7 @@ <h2>Results<a class="headerlink" href="#results" title="Permalink to this headin
 <span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;speed up ratio: </span><span class="si">{</span><span class="n">speed_up</span><span class="si">:</span><span class="s2">.0f</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>speed up ratio: 879
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>speed up ratio: 930
 </pre></div>
 </div>
 </section>
@@ -664,12 +664,12 @@ <h2>Efron estimate<a class="headerlink" href="#efron-estimate" title="Permalink
 <span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;speed up ratio: </span><span class="si">{</span><span class="n">speed_up</span><span class="si">:</span><span class="s2">.0f</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>speed up ratio: 787
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>speed up ratio: 1042
 </pre></div>
 </div>
 <p>As shown by the last print, we still preserve the x100 ratio speed up
 even for the Efron estimate.</p>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 52.615 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 57.445 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-survival-analysis-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/8fea211f7305fcd7e5b7ad33d1c6449f/plot_survival_analysis.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_survival_analysis.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_ucurve.html b/auto_examples/plot_ucurve.html
index 24945ddcd..8053c1ec1 100644
--- a/auto_examples/plot_ucurve.html
+++ b/auto_examples/plot_ucurve.html
@@ -375,20 +375,20 @@
 </pre></div>
 </div>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>file_sizes:   0%|                                   | 0.00/13.7M [00:00&lt;?, ?B/s]
-file_sizes:   0%|                           | 24.6k/13.7M [00:00&lt;01:51, 123kB/s]
-file_sizes:   0%|                           | 49.2k/13.7M [00:00&lt;01:51, 122kB/s]
-file_sizes:   1%|▏                           | 106k/13.7M [00:00&lt;01:09, 196kB/s]
-file_sizes:   2%|▍                           | 221k/13.7M [00:00&lt;00:39, 343kB/s]
-file_sizes:   3%|▉                           | 451k/13.7M [00:01&lt;00:21, 630kB/s]
-file_sizes:   7%|█▊                         | 909k/13.7M [00:01&lt;00:10, 1.18MB/s]
-file_sizes:  13%|███▍                      | 1.83M/13.7M [00:01&lt;00:05, 2.28MB/s]
-file_sizes:  27%|██████▉                   | 3.66M/13.7M [00:01&lt;00:02, 4.41MB/s]
-file_sizes:  38%|█████████▉                | 5.23M/13.7M [00:01&lt;00:01, 5.46MB/s]
-file_sizes:  50%|████████████▉             | 6.81M/13.7M [00:02&lt;00:01, 6.18MB/s]
-file_sizes:  61%|███████████████▊          | 8.38M/13.7M [00:02&lt;00:00, 6.69MB/s]
-file_sizes:  80%|████████████████████▊     | 11.0M/13.7M [00:02&lt;00:00, 8.51MB/s]
-file_sizes:  92%|███████████████████████▊  | 12.6M/13.7M [00:02&lt;00:00, 8.36MB/s]
-file_sizes: 100%|██████████████████████████| 13.7M/13.7M [00:02&lt;00:00, 5.20MB/s]
+file_sizes:   0%|                           | 24.6k/13.7M [00:00&lt;01:49, 125kB/s]
+file_sizes:   0%|                           | 49.2k/13.7M [00:00&lt;01:50, 124kB/s]
+file_sizes:   1%|▏                           | 106k/13.7M [00:00&lt;01:08, 199kB/s]
+file_sizes:   2%|▍                           | 221k/13.7M [00:00&lt;00:38, 349kB/s]
+file_sizes:   3%|▉                           | 451k/13.7M [00:00&lt;00:20, 639kB/s]
+file_sizes:   7%|█▊                         | 909k/13.7M [00:01&lt;00:10, 1.21MB/s]
+file_sizes:  13%|███▍                      | 1.83M/13.7M [00:01&lt;00:05, 2.32MB/s]
+file_sizes:  27%|██████▉                   | 3.66M/13.7M [00:01&lt;00:02, 4.50MB/s]
+file_sizes:  38%|█████████▉                | 5.23M/13.7M [00:01&lt;00:01, 5.56MB/s]
+file_sizes:  50%|████████████▉             | 6.81M/13.7M [00:01&lt;00:01, 6.29MB/s]
+file_sizes:  61%|███████████████▊          | 8.38M/13.7M [00:02&lt;00:00, 6.81MB/s]
+file_sizes:  80%|████████████████████▊     | 11.0M/13.7M [00:02&lt;00:00, 8.65MB/s]
+file_sizes:  92%|███████████████████████▊  | 12.6M/13.7M [00:02&lt;00:00, 8.48MB/s]
+file_sizes: 100%|██████████████████████████| 13.7M/13.7M [00:02&lt;00:00, 5.29MB/s]
 </pre></div>
 </div>
 <p>Next, we define the regularization path.
@@ -421,7 +421,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<img src="../_images/sphx_glr_plot_ucurve_001.png" srcset="../_images/sphx_glr_plot_ucurve_001.png" alt="Mean squared error" class = "sphx-glr-single-img"/><p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 13.278 seconds)</p>
+<img src="../_images/sphx_glr_plot_ucurve_001.png" srcset="../_images/sphx_glr_plot_ucurve_001.png" alt="Mean squared error" class = "sphx-glr-single-img"/><p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 13.957 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-ucurve-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/c561e76c25a76938aecd426c0cbcdd39/plot_ucurve.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_ucurve.ipynb</span></code></a></p>
diff --git a/auto_examples/plot_zero_weights_lasso.html b/auto_examples/plot_zero_weights_lasso.html
index 16d7338fb..b0d6e30e6 100644
--- a/auto_examples/plot_zero_weights_lasso.html
+++ b/auto_examples/plot_zero_weights_lasso.html
@@ -395,7 +395,7 @@
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">(</span><span class="n">block</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 0.342 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> (0 minutes 0.374 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-zero-weights-lasso-py">
 <div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
 <p><a class="reference download internal" download="" href="../_downloads/761049011e8bc99f97372c6ab12d9dc1/plot_zero_weights_lasso.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_zero_weights_lasso.ipynb</span></code></a></p>
diff --git a/auto_examples/sg_execution_times.html b/auto_examples/sg_execution_times.html
index db877e9ba..c92127657 100644
--- a/auto_examples/sg_execution_times.html
+++ b/auto_examples/sg_execution_times.html
@@ -347,7 +347,7 @@
         <article role="main" id="furo-main-content">
           <section id="computation-times">
 <span id="sphx-glr-auto-examples-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this heading">¶</a></h1>
-<p><strong>02:40.942</strong> total execution time for 11 files <strong>from auto_examples</strong>:</p>
+<p><strong>02:49.958</strong> total execution time for 11 files <strong>from auto_examples</strong>:</p>
 <div class="docutils container">
 <style scoped>
 <link href="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/5.3.0/css/bootstrap.min.css" rel="stylesheet" />
@@ -370,47 +370,47 @@
 </thead>
 <tbody>
 <tr class="row-even"><td><p><a class="reference internal" href="plot_survival_analysis.html#sphx-glr-auto-examples-plot-survival-analysis-py"><span class="std std-ref">Comparison of lifelines with skglm for survival analysis</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_survival_analysis.py</span></code>)</p></td>
-<td><p>00:52.615</p></td>
+<td><p>00:57.445</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="plot_compare_time.html#sphx-glr-auto-examples-plot-compare-time-py"><span class="std std-ref">Timing comparison with scikit-learn for Lasso</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_compare_time.py</span></code>)</p></td>
-<td><p>00:45.604</p></td>
+<td><p>00:47.620</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="plot_group_logistic_regression.html#sphx-glr-auto-examples-plot-group-logistic-regression-py"><span class="std std-ref">Group Logistic regression in python</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_group_logistic_regression.py</span></code>)</p></td>
-<td><p>00:14.785</p></td>
+<td><p>00:15.201</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="plot_ucurve.html#sphx-glr-auto-examples-plot-ucurve-py"><span class="std std-ref">Show U-curve of regularization</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_ucurve.py</span></code>)</p></td>
-<td><p>00:13.278</p></td>
+<td><p>00:13.957</p></td>
 <td><p>0.0</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="plot_sparse_recovery.html#sphx-glr-auto-examples-plot-sparse-recovery-py"><span class="std std-ref">Sparse recovery with non-convex penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_sparse_recovery.py</span></code>)</p></td>
-<td><p>00:08.910</p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="plot_logreg_various_penalties.html#sphx-glr-auto-examples-plot-logreg-various-penalties-py"><span class="std std-ref">Logistic regression with Elastic net and minimax concave penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_logreg_various_penalties.py</span></code>)</p></td>
+<td><p>00:08.967</p></td>
 <td><p>0.0</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="plot_logreg_various_penalties.html#sphx-glr-auto-examples-plot-logreg-various-penalties-py"><span class="std std-ref">Logistic regression with Elastic net and minimax concave penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_logreg_various_penalties.py</span></code>)</p></td>
-<td><p>00:08.793</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="plot_sparse_recovery.html#sphx-glr-auto-examples-plot-sparse-recovery-py"><span class="std std-ref">Sparse recovery with non-convex penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_sparse_recovery.py</span></code>)</p></td>
+<td><p>00:08.881</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="plot_sparse_group_lasso.html#sphx-glr-auto-examples-plot-sparse-group-lasso-py"><span class="std std-ref">Fast Sparse Group Lasso in python</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_sparse_group_lasso.py</span></code>)</p></td>
-<td><p>00:05.547</p></td>
+<td><p>00:05.796</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="plot_reweighted_l1.html#sphx-glr-auto-examples-plot-reweighted-l1-py"><span class="std std-ref">Timing comparison between direct prox computation and reweighting</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_reweighted_l1.py</span></code>)</p></td>
-<td><p>00:05.021</p></td>
+<td><p>00:05.326</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="plot_lasso_vs_weighted.html#sphx-glr-auto-examples-plot-lasso-vs-weighted-py"><span class="std std-ref">Comparison of Lasso and Weighted Lasso</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_lasso_vs_weighted.py</span></code>)</p></td>
-<td><p>00:03.731</p></td>
+<td><p>00:03.903</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="plot_pen_prox.html#sphx-glr-auto-examples-plot-pen-prox-py"><span class="std std-ref">Value and proximal operators of penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_pen_prox.py</span></code>)</p></td>
-<td><p>00:02.314</p></td>
+<td><p>00:02.487</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="plot_zero_weights_lasso.html#sphx-glr-auto-examples-plot-zero-weights-lasso-py"><span class="std std-ref">Weighted Lasso with some zero weights</span></a> (<code class="docutils literal notranslate"><span class="pre">plot_zero_weights_lasso.py</span></code>)</p></td>
-<td><p>00:00.342</p></td>
+<td><p>00:00.374</p></td>
 <td><p>0.0</p></td>
 </tr>
 </tbody>
diff --git a/generated/skglm.CoxEstimator.html b/generated/skglm.CoxEstimator.html
index ab77f449a..3adcbc9c8 100644
--- a/generated/skglm.CoxEstimator.html
+++ b/generated/skglm.CoxEstimator.html
@@ -349,7 +349,7 @@
 <h1>skglm.CoxEstimator<a class="headerlink" href="#skglm-coxestimator" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.CoxEstimator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">CoxEstimator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.7</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'efron'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1205"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.CoxEstimator" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">CoxEstimator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.7</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'efron'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1205"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.CoxEstimator" title="Permalink to this definition">¶</a></dt>
 <dd><p>Elastic Cox estimator with Efron and Breslow estimate.</p>
 <p>Refer to <a class="reference internal" href="../tutorials/cox_datafit.html#maths-cox-datafit"><span class="std std-ref">Mathematics behind Cox datafit</span></a>
 for details about the datafit expression. The data convention for the estimator is</p>
@@ -392,7 +392,7 @@ <h1>skglm.CoxEstimator<a class="headerlink" href="#skglm-coxestimator" title="Pe
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.CoxEstimator.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.7</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'efron'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.CoxEstimator.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.7</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'efron'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.CoxEstimator.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.ElasticNet.html b/generated/skglm.ElasticNet.html
index 65f45ec4d..50d7a0961 100644
--- a/generated/skglm.ElasticNet.html
+++ b/generated/skglm.ElasticNet.html
@@ -349,7 +349,7 @@
 <h1>skglm.ElasticNet<a class="headerlink" href="#skglm-elasticnet" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.ElasticNet">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">ElasticNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L615"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.ElasticNet" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">ElasticNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L615"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.ElasticNet" title="Permalink to this definition">¶</a></dt>
 <dd><p>Elastic net estimator.</p>
 <p>The optimization objective for Elastic net is:</p>
 <div class="math-wrapper docutils container">
@@ -413,7 +413,7 @@ <h1>skglm.ElasticNet<a class="headerlink" href="#skglm-elasticnet" title="Permal
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.ElasticNet.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L683"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.ElasticNet.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L683"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.ElasticNet.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.GeneralizedLinearEstimator.html b/generated/skglm.GeneralizedLinearEstimator.html
index fd0d46cda..d1121d7ae 100644
--- a/generated/skglm.GeneralizedLinearEstimator.html
+++ b/generated/skglm.GeneralizedLinearEstimator.html
@@ -349,7 +349,7 @@
 <h1>skglm.GeneralizedLinearEstimator<a class="headerlink" href="#skglm-generalizedlinearestimator" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.GeneralizedLinearEstimator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">GeneralizedLinearEstimator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GeneralizedLinearEstimator" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">GeneralizedLinearEstimator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GeneralizedLinearEstimator" title="Permalink to this definition">¶</a></dt>
 <dd><p>Generic generalized linear estimator.</p>
 <p>This estimator takes a penalty and a datafit and runs a coordinate descent solver
 to solve the optimization problem. It handles classification and regression tasks.</p>
@@ -383,7 +383,7 @@ <h1>skglm.GeneralizedLinearEstimator<a class="headerlink" href="#skglm-generaliz
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.GeneralizedLinearEstimator.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GeneralizedLinearEstimator.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GeneralizedLinearEstimator.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.GroupLasso.html b/generated/skglm.GroupLasso.html
index 48097a1b9..fe4e77034 100644
--- a/generated/skglm.GroupLasso.html
+++ b/generated/skglm.GroupLasso.html
@@ -349,7 +349,7 @@
 <h1>skglm.GroupLasso<a class="headerlink" href="#skglm-grouplasso" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.GroupLasso">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">GroupLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">groups</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GroupLasso" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">GroupLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">groups</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GroupLasso" title="Permalink to this definition">¶</a></dt>
 <dd><p>GroupLasso estimator based on Celer solver and primal extrapolation.</p>
 <p>The optimization objective for GroupLasso is:</p>
 <div class="math-wrapper docutils container">
@@ -412,7 +412,7 @@ <h1>skglm.GroupLasso<a class="headerlink" href="#skglm-grouplasso" title="Permal
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.GroupLasso.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">groups</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1617"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GroupLasso.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">groups</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1617"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.GroupLasso.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.Lasso.html b/generated/skglm.Lasso.html
index 1f4f146ce..46cd218fb 100644
--- a/generated/skglm.Lasso.html
+++ b/generated/skglm.Lasso.html
@@ -349,7 +349,7 @@
 <h1>skglm.Lasso<a class="headerlink" href="#skglm-lasso" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.Lasso">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">Lasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.Lasso" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">Lasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.Lasso" title="Permalink to this definition">¶</a></dt>
 <dd><p>Lasso estimator based on Celer solver and primal extrapolation.</p>
 <p>The optimization objective for Lasso is:</p>
 <div class="math-wrapper docutils container">
@@ -408,7 +408,7 @@ <h1>skglm.Lasso<a class="headerlink" href="#skglm-lasso" title="Permalink to thi
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.Lasso.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L366"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.Lasso.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L366"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.Lasso.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.LinearSVC.html b/generated/skglm.LinearSVC.html
index 2c7749515..6dba7d19d 100644
--- a/generated/skglm.LinearSVC.html
+++ b/generated/skglm.LinearSVC.html
@@ -349,7 +349,7 @@
 <h1>skglm.LinearSVC<a class="headerlink" href="#skglm-linearsvc" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.LinearSVC">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">LinearSVC</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1094"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.LinearSVC" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">LinearSVC</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1094"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.LinearSVC" title="Permalink to this definition">¶</a></dt>
 <dd><p>LinearSVC estimator, with hinge loss.</p>
 <p>The optimization objective for LinearSVC is:</p>
 <div class="math-wrapper docutils container">
@@ -414,7 +414,7 @@ <h1>skglm.LinearSVC<a class="headerlink" href="#skglm-linearsvc" title="Permalin
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.LinearSVC.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.LinearSVC.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.LinearSVC.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.MCPRegression.html b/generated/skglm.MCPRegression.html
index 63585c34d..a51cb4df8 100644
--- a/generated/skglm.MCPRegression.html
+++ b/generated/skglm.MCPRegression.html
@@ -349,7 +349,7 @@
 <h1>skglm.MCPRegression<a class="headerlink" href="#skglm-mcpregression" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.MCPRegression">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">MCPRegression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L769"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MCPRegression" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">MCPRegression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L769"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MCPRegression" title="Permalink to this definition">¶</a></dt>
 <dd><p>Linear regression with MCP penalty estimator.</p>
 <p>The optimization objective for MCPRegression is, with <span class="math notranslate nohighlight">`x &gt;= 0`</span>:</p>
 <div class="math-wrapper docutils container">
@@ -422,7 +422,7 @@ <h1>skglm.MCPRegression<a class="headerlink" href="#skglm-mcpregression" title="
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.MCPRegression.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L848"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MCPRegression.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L848"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MCPRegression.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.MultiTaskLasso.html b/generated/skglm.MultiTaskLasso.html
index f701d3e37..329ea6b09 100644
--- a/generated/skglm.MultiTaskLasso.html
+++ b/generated/skglm.MultiTaskLasso.html
@@ -349,7 +349,7 @@
 <h1>skglm.MultiTaskLasso<a class="headerlink" href="#skglm-multitasklasso" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.MultiTaskLasso">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">MultiTaskLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy_X</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1368"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MultiTaskLasso" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">MultiTaskLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy_X</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1368"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MultiTaskLasso" title="Permalink to this definition">¶</a></dt>
 <dd><p>MultiTaskLasso estimator.</p>
 <p>The optimization objective for MultiTaskLasso is:</p>
 <div class="math-wrapper docutils container">
@@ -397,7 +397,7 @@ <h1>skglm.MultiTaskLasso<a class="headerlink" href="#skglm-multitasklasso" title
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.MultiTaskLasso.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy_X</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1423"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MultiTaskLasso.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy_X</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1423"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.MultiTaskLasso.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.SparseLogisticRegression.html b/generated/skglm.SparseLogisticRegression.html
index 12cffa4d8..97b004e8c 100644
--- a/generated/skglm.SparseLogisticRegression.html
+++ b/generated/skglm.SparseLogisticRegression.html
@@ -349,7 +349,7 @@
 <h1>skglm.SparseLogisticRegression<a class="headerlink" href="#skglm-sparselogisticregression" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.SparseLogisticRegression">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">SparseLogisticRegression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L957"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.SparseLogisticRegression" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">SparseLogisticRegression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L957"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.SparseLogisticRegression" title="Permalink to this definition">¶</a></dt>
 <dd><p>Sparse Logistic regression estimator.</p>
 <p>The optimization objective for sparse Logistic regression is:</p>
 <div class="math-wrapper docutils container">
@@ -393,7 +393,7 @@ <h1>skglm.SparseLogisticRegression<a class="headerlink" href="#skglm-sparselogis
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.SparseLogisticRegression.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L1006"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.SparseLogisticRegression.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L1006"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.SparseLogisticRegression.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.WeightedLasso.html b/generated/skglm.WeightedLasso.html
index a6d99c8c5..3e9c684ae 100644
--- a/generated/skglm.WeightedLasso.html
+++ b/generated/skglm.WeightedLasso.html
@@ -349,7 +349,7 @@
 <h1>skglm.WeightedLasso<a class="headerlink" href="#skglm-weightedlasso" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.WeightedLasso">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">WeightedLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L452"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.WeightedLasso" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.</span></span><span class="sig-name descname"><span class="pre">WeightedLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L452"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.WeightedLasso" title="Permalink to this definition">¶</a></dt>
 <dd><p>WeightedLasso estimator based on Celer solver and primal extrapolation.</p>
 <p>The optimization objective for WeightedLasso is:</p>
 <div class="math-wrapper docutils container">
@@ -413,7 +413,7 @@ <h1>skglm.WeightedLasso<a class="headerlink" href="#skglm-weightedlasso" title="
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.WeightedLasso.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/estimators.py#L521"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.WeightedLasso.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/estimators.py#L521"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.WeightedLasso.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Cox.html b/generated/skglm.datafits.Cox.html
index c13db9ff2..3ad9230cf 100644
--- a/generated/skglm.datafits.Cox.html
+++ b/generated/skglm.datafits.Cox.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Cox<a class="headerlink" href="#skglm-datafits-cox" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Cox">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Cox</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">use_efron</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L668"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Cox" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Cox</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">use_efron</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L668"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Cox" title="Permalink to this definition">¶</a></dt>
 <dd><p>Cox datafit for survival analysis.</p>
 <p>Refer to <a class="reference internal" href="../tutorials/cox_datafit.html#maths-cox-datafit"><span class="std std-ref">Mathematics behind Cox datafit</span></a> for details.</p>
 <dl class="field-list simple">
@@ -381,7 +381,7 @@ <h1>skglm.datafits.Cox<a class="headerlink" href="#skglm-datafits-cox" title="Pe
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Cox.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">use_efron</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L700"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Cox.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">use_efron</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L700"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Cox.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Gamma.html b/generated/skglm.datafits.Gamma.html
index 0c453bf90..96d6adcdd 100644
--- a/generated/skglm.datafits.Gamma.html
+++ b/generated/skglm.datafits.Gamma.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Gamma<a class="headerlink" href="#skglm-datafits-gamma" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Gamma">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Gamma</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L614"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Gamma" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Gamma</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L614"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Gamma" title="Permalink to this definition">¶</a></dt>
 <dd><p>Gamma datafit.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -362,7 +362,7 @@ <h1>skglm.datafits.Gamma<a class="headerlink" href="#skglm-datafits-gamma" title
 This allows for faster computations.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Gamma.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L629"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Gamma.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L629"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Gamma.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Huber.html b/generated/skglm.datafits.Huber.html
index 61a7a92f4..468b7685c 100644
--- a/generated/skglm.datafits.Huber.html
+++ b/generated/skglm.datafits.Huber.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Huber<a class="headerlink" href="#skglm-datafits-huber" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Huber">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Huber</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">delta</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L418"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Huber" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Huber</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">delta</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L418"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Huber" title="Permalink to this definition">¶</a></dt>
 <dd><p>Huber datafit.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -372,7 +372,7 @@ <h1>skglm.datafits.Huber<a class="headerlink" href="#skglm-datafits-huber" title
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Huber.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">delta</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Huber.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">delta</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Huber.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Logistic.html b/generated/skglm.datafits.Logistic.html
index ef30bf25b..af91524f4 100644
--- a/generated/skglm.datafits.Logistic.html
+++ b/generated/skglm.datafits.Logistic.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Logistic<a class="headerlink" href="#skglm-datafits-logistic" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Logistic">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Logistic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L249"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Logistic" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Logistic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L249"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Logistic" title="Permalink to this definition">¶</a></dt>
 <dd><p>Logistic datafit with labels in {-1, 1}.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -361,7 +361,7 @@ <h1>skglm.datafits.Logistic<a class="headerlink" href="#skglm-datafits-logistic"
 This allows for faster computations.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Logistic.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L262"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Logistic.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L262"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Logistic.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.LogisticGroup.html b/generated/skglm.datafits.LogisticGroup.html
index 93416b133..017a3abbb 100644
--- a/generated/skglm.datafits.LogisticGroup.html
+++ b/generated/skglm.datafits.LogisticGroup.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.LogisticGroup<a class="headerlink" href="#skglm-datafits-logisticgroup" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.LogisticGroup">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">LogisticGroup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/group.py#L106"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.LogisticGroup" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">LogisticGroup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/group.py#L106"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.LogisticGroup" title="Permalink to this definition">¶</a></dt>
 <dd><p>Logistic datafit used with group penalties.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -372,7 +372,7 @@ <h1>skglm.datafits.LogisticGroup<a class="headerlink" href="#skglm-datafits-logi
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.LogisticGroup.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/group.py#L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.LogisticGroup.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/group.py#L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.LogisticGroup.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Poisson.html b/generated/skglm.datafits.Poisson.html
index d9af4b9e0..c89ff3657 100644
--- a/generated/skglm.datafits.Poisson.html
+++ b/generated/skglm.datafits.Poisson.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Poisson<a class="headerlink" href="#skglm-datafits-poisson" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Poisson">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Poisson</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L542"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Poisson" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Poisson</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L542"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Poisson" title="Permalink to this definition">¶</a></dt>
 <dd><p>Poisson datafit.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -361,7 +361,7 @@ <h1>skglm.datafits.Poisson<a class="headerlink" href="#skglm-datafits-poisson" t
 This allows for faster computations.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Poisson.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L555"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Poisson.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L555"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Poisson.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.Quadratic.html b/generated/skglm.datafits.Quadratic.html
index d59d50bd7..22d26e837 100644
--- a/generated/skglm.datafits.Quadratic.html
+++ b/generated/skglm.datafits.Quadratic.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.Quadratic<a class="headerlink" href="#skglm-datafits-quadratic" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.Quadratic">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Quadratic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Quadratic" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">Quadratic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Quadratic" title="Permalink to this definition">¶</a></dt>
 <dd><p>Quadratic datafit.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -367,7 +367,7 @@ <h1>skglm.datafits.Quadratic<a class="headerlink" href="#skglm-datafits-quadrati
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.Quadratic.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L29"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Quadratic.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L29"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.Quadratic.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.QuadraticGroup.html b/generated/skglm.datafits.QuadraticGroup.html
index 4da4c5312..1fe382cbf 100644
--- a/generated/skglm.datafits.QuadraticGroup.html
+++ b/generated/skglm.datafits.QuadraticGroup.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.QuadraticGroup<a class="headerlink" href="#skglm-datafits-quadraticgroup" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.QuadraticGroup">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">QuadraticGroup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/group.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticGroup" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">QuadraticGroup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/group.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticGroup" title="Permalink to this definition">¶</a></dt>
 <dd><p>Quadratic datafit used with group penalties.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -370,7 +370,7 @@ <h1>skglm.datafits.QuadraticGroup<a class="headerlink" href="#skglm-datafits-qua
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.QuadraticGroup.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/group.py#L28"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticGroup.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/group.py#L28"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticGroup.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.QuadraticSVC.html b/generated/skglm.datafits.QuadraticSVC.html
index 960ed3e60..2fd25219d 100644
--- a/generated/skglm.datafits.QuadraticSVC.html
+++ b/generated/skglm.datafits.QuadraticSVC.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.QuadraticSVC<a class="headerlink" href="#skglm-datafits-quadraticsvc" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.QuadraticSVC">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">QuadraticSVC</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L341"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticSVC" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">QuadraticSVC</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L341"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticSVC" title="Permalink to this definition">¶</a></dt>
 <dd><p>A Quadratic SVC datafit used for classification tasks.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -361,7 +361,7 @@ <h1>skglm.datafits.QuadraticSVC<a class="headerlink" href="#skglm-datafits-quadr
 This allows for faster computations.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.QuadraticSVC.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L354"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticSVC.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L354"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.QuadraticSVC.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.datafits.WeightedQuadratic.html b/generated/skglm.datafits.WeightedQuadratic.html
index 45be24501..f7bc01ad9 100644
--- a/generated/skglm.datafits.WeightedQuadratic.html
+++ b/generated/skglm.datafits.WeightedQuadratic.html
@@ -349,7 +349,7 @@
 <h1>skglm.datafits.WeightedQuadratic<a class="headerlink" href="#skglm-datafits-weightedquadratic" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.datafits.WeightedQuadratic">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">WeightedQuadratic</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_weights</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.WeightedQuadratic" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.datafits.</span></span><span class="sig-name descname"><span class="pre">WeightedQuadratic</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_weights</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.WeightedQuadratic" title="Permalink to this definition">¶</a></dt>
 <dd><p>Weighted Quadratic datafit to handle sample weights.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -370,7 +370,7 @@ <h1>skglm.datafits.WeightedQuadratic<a class="headerlink" href="#skglm-datafits-
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.datafits.WeightedQuadratic.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_weights</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/datafits/single_task.py#L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.WeightedQuadratic.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sample_weights</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/datafits/single_task.py#L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.datafits.WeightedQuadratic.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.experimental.IterativeReweightedL1.html b/generated/skglm.experimental.IterativeReweightedL1.html
index c96538ab1..0baadcb9e 100644
--- a/generated/skglm.experimental.IterativeReweightedL1.html
+++ b/generated/skglm.experimental.IterativeReweightedL1.html
@@ -349,7 +349,7 @@
 <h1>skglm.experimental.IterativeReweightedL1<a class="headerlink" href="#skglm-experimental-iterativereweightedl1" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.experimental.IterativeReweightedL1">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">IterativeReweightedL1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit=&lt;skglm.datafits.single_task.Quadratic</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty=&lt;skglm.penalties.separable.L0_5</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_reweights=5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/reweighted.py#L8"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.IterativeReweightedL1" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">IterativeReweightedL1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit=&lt;skglm.datafits.single_task.Quadratic</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty=&lt;skglm.penalties.separable.L0_5</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_reweights=5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/reweighted.py#L8"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.IterativeReweightedL1" title="Permalink to this definition">¶</a></dt>
 <dd><p>Reweighted L1-norm estimator.</p>
 <p>This estimator solves a non-convex problems by iteratively solving
 convex surrogates involving weighted L1 norms.</p>
@@ -389,7 +389,7 @@ <h1>skglm.experimental.IterativeReweightedL1<a class="headerlink" href="#skglm-e
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.experimental.IterativeReweightedL1.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit=&lt;skglm.datafits.single_task.Quadratic</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty=&lt;skglm.penalties.separable.L0_5</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_reweights=5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/reweighted.py#L44"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.IterativeReweightedL1.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">datafit=&lt;skglm.datafits.single_task.Quadratic</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">penalty=&lt;skglm.penalties.separable.L0_5</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">solver=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_reweights=5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/reweighted.py#L44"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.IterativeReweightedL1.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.experimental.PDCD_WS.html b/generated/skglm.experimental.PDCD_WS.html
index d7f0f55a1..bc6f7bb52 100644
--- a/generated/skglm.experimental.PDCD_WS.html
+++ b/generated/skglm.experimental.PDCD_WS.html
@@ -349,7 +349,7 @@
 <h1>skglm.experimental.PDCD_WS<a class="headerlink" href="#skglm-experimental-pdcd-ws" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.experimental.PDCD_WS">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">PDCD_WS</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dual_init</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/pdcd_ws.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.PDCD_WS" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">PDCD_WS</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dual_init</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/pdcd_ws.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.PDCD_WS" title="Permalink to this definition">¶</a></dt>
 <dd><p>Primal-Dual Coordinate Descent solver with working sets.</p>
 <p>It solves</p>
 <div class="math-wrapper docutils container">
@@ -414,7 +414,7 @@ <h1>skglm.experimental.PDCD_WS<a class="headerlink" href="#skglm-experimental-pd
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.experimental.PDCD_WS.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dual_init</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/pdcd_ws.py#L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.PDCD_WS.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dual_init</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/pdcd_ws.py#L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.PDCD_WS.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.experimental.Pinball.html b/generated/skglm.experimental.Pinball.html
index 70e8d951c..36e134de0 100644
--- a/generated/skglm.experimental.Pinball.html
+++ b/generated/skglm.experimental.Pinball.html
@@ -349,7 +349,7 @@
 <h1>skglm.experimental.Pinball<a class="headerlink" href="#skglm-experimental-pinball" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.experimental.Pinball">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">Pinball</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">quantile_level</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/quantile_regression.py#L7"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.Pinball" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">Pinball</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">quantile_level</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/quantile_regression.py#L7"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.Pinball" title="Permalink to this definition">¶</a></dt>
 <dd><p>Pinball datafit.</p>
 <p>The datafit reads:</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">sum_i</span> <span class="n">quantile_level</span> <span class="o">*</span> <span class="nb">max</span><span class="p">(</span><span class="n">y_i</span> <span class="o">-</span> <span class="n">Xw_i</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span> <span class="o">+</span>
@@ -368,7 +368,7 @@ <h1>skglm.experimental.Pinball<a class="headerlink" href="#skglm-experimental-pi
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.experimental.Pinball.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">quantile_level</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/quantile_regression.py#L24"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.Pinball.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">quantile_level</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/quantile_regression.py#L24"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.Pinball.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.experimental.SqrtLasso.html b/generated/skglm.experimental.SqrtLasso.html
index 6d611d1e4..e3d507049 100644
--- a/generated/skglm.experimental.SqrtLasso.html
+++ b/generated/skglm.experimental.SqrtLasso.html
@@ -349,7 +349,7 @@
 <h1>skglm.experimental.SqrtLasso<a class="headerlink" href="#skglm-experimental-sqrtlasso" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.experimental.SqrtLasso">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">SqrtLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/sqrt_lasso.py#L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtLasso" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">SqrtLasso</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/sqrt_lasso.py#L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtLasso" title="Permalink to this definition">¶</a></dt>
 <dd><p>Square root Lasso estimator based on Prox Newton solver.</p>
 <p>The optimization objective for square root Lasso is:</p>
 <div class="math-wrapper docutils container">
@@ -376,7 +376,7 @@ <h1>skglm.experimental.SqrtLasso<a class="headerlink" href="#skglm-experimental-
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.experimental.SqrtLasso.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/sqrt_lasso.py#L106"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtLasso.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/sqrt_lasso.py#L106"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtLasso.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.experimental.SqrtQuadratic.html b/generated/skglm.experimental.SqrtQuadratic.html
index 62c841c00..0f8a98a82 100644
--- a/generated/skglm.experimental.SqrtQuadratic.html
+++ b/generated/skglm.experimental.SqrtQuadratic.html
@@ -349,7 +349,7 @@
 <h1>skglm.experimental.SqrtQuadratic<a class="headerlink" href="#skglm-experimental-sqrtquadratic" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.experimental.SqrtQuadratic">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">SqrtQuadratic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/sqrt_lasso.py#L14"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtQuadratic" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.experimental.</span></span><span class="sig-name descname"><span class="pre">SqrtQuadratic</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/sqrt_lasso.py#L14"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtQuadratic" title="Permalink to this definition">¶</a></dt>
 <dd><p>Unnormalized square root quadratic datafit.</p>
 <p>The datafit reads:</p>
 <div class="math-wrapper docutils container">
@@ -358,7 +358,7 @@ <h1>skglm.experimental.SqrtQuadratic<a class="headerlink" href="#skglm-experimen
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.experimental.SqrtQuadratic.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/experimental/sqrt_lasso.py#L24"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtQuadratic.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/experimental/sqrt_lasso.py#L24"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.experimental.SqrtQuadratic.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.BlockSCAD.html b/generated/skglm.penalties.BlockSCAD.html
index 1e0b871ba..623a53ee2 100644
--- a/generated/skglm.penalties.BlockSCAD.html
+++ b/generated/skglm.penalties.BlockSCAD.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.BlockSCAD<a class="headerlink" href="#skglm-penalties-blockscad" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.BlockSCAD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">BlockSCAD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/block_separable.py#L161"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.BlockSCAD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">BlockSCAD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/block_separable.py#L161"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.BlockSCAD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Block Smoothly Clipped Absolute Deviation.</p>
 <p class="rubric">Notes</p>
 <p>With <span class="math notranslate nohighlight">`W_j`</span> the j-th row of math:<cite>W</cite>, the penalty is:</p>
@@ -368,7 +368,7 @@ <h1>skglm.penalties.BlockSCAD<a class="headerlink" href="#skglm-penalties-blocks
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.BlockSCAD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/block_separable.py#L180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.BlockSCAD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/block_separable.py#L180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.BlockSCAD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.IndicatorBox.html b/generated/skglm.penalties.IndicatorBox.html
index 85163fdef..56c1a485b 100644
--- a/generated/skglm.penalties.IndicatorBox.html
+++ b/generated/skglm.penalties.IndicatorBox.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.IndicatorBox<a class="headerlink" href="#skglm-penalties-indicatorbox" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.IndicatorBox">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">IndicatorBox</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.IndicatorBox" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">IndicatorBox</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.IndicatorBox" title="Permalink to this definition">¶</a></dt>
 <dd><p>Box constraint penalty.</p>
 <p class="rubric">Notes</p>
 <div class="math-wrapper docutils container">
@@ -360,7 +360,7 @@ <h1>skglm.penalties.IndicatorBox<a class="headerlink" href="#skglm-penalties-ind
 <span class="math notranslate nohighlight">`[0, alpha]^(n_&quot;samples&quot;)`</span></p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.IndicatorBox.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L453"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.IndicatorBox.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L453"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.IndicatorBox.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.L0_5.html b/generated/skglm.penalties.L0_5.html
index 86154d1c7..3a23f3dff 100644
--- a/generated/skglm.penalties.L0_5.html
+++ b/generated/skglm.penalties.L0_5.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.L0_5<a class="headerlink" href="#skglm-penalties-l0-5" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.L0_5">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L0_5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L503"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L0_5" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L0_5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L503"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L0_5" title="Permalink to this definition">¶</a></dt>
 <dd><p><span class="math notranslate nohighlight">`ell_(0.5)`</span> non-convex quasi-norm penalty.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.L0_5.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L506"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L0_5.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L506"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L0_5.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.L1.html b/generated/skglm.penalties.L1.html
index ac8b2e0cf..ff507796f 100644
--- a/generated/skglm.penalties.L1.html
+++ b/generated/skglm.penalties.L1.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.L1<a class="headerlink" href="#skglm-penalties-l1" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.L1">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L11"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L11"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1" title="Permalink to this definition">¶</a></dt>
 <dd><p><span class="math notranslate nohighlight">`ell_1`</span> penalty.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.L1.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L14"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L14"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.L1_plus_L2.html b/generated/skglm.penalties.L1_plus_L2.html
index 6841e12d3..6da146a0b 100644
--- a/generated/skglm.penalties.L1_plus_L2.html
+++ b/generated/skglm.penalties.L1_plus_L2.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.L1_plus_L2<a class="headerlink" href="#skglm-penalties-l1-plus-l2" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.L1_plus_L2">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L1_plus_L2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1_plus_L2" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L1_plus_L2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1_plus_L2" title="Permalink to this definition">¶</a></dt>
 <dd><p><span class="math notranslate nohighlight">`ell_1 + ell_2`</span> penalty (aka ElasticNet penalty).</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.L1_plus_L2.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L74"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1_plus_L2.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l1_ratio</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L74"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L1_plus_L2.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.L2.html b/generated/skglm.penalties.L2.html
index 1267eda78..b0834be70 100644
--- a/generated/skglm.penalties.L2.html
+++ b/generated/skglm.penalties.L2.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.L2<a class="headerlink" href="#skglm-penalties-l2" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.L2">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L721"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L721"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2" title="Permalink to this definition">¶</a></dt>
 <dd><p><span class="math notranslate nohighlight">`ell_2`</span> penalty.</p>
 <p>The penalty reads</p>
 <div class="math-wrapper docutils container">
@@ -358,7 +358,7 @@ <h1>skglm.penalties.L2<a class="headerlink" href="#skglm-penalties-l2" title="Pe
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.L2.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L731"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L731"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.L2_3.html b/generated/skglm.penalties.L2_3.html
index 4f74e888f..5d5667b7e 100644
--- a/generated/skglm.penalties.L2_3.html
+++ b/generated/skglm.penalties.L2_3.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.L2_3<a class="headerlink" href="#skglm-penalties-l2-3" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.L2_3">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L2_3</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L562"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2_3" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">L2_3</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L562"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2_3" title="Permalink to this definition">¶</a></dt>
 <dd><p><span class="math notranslate nohighlight">`ell_(2/3)`</span> quasi-norm non-convex penalty.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.L2_3.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L565"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2_3.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L565"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.L2_3.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.LogSumPenalty.html b/generated/skglm.penalties.LogSumPenalty.html
index 91ab87fce..0b84805e7 100644
--- a/generated/skglm.penalties.LogSumPenalty.html
+++ b/generated/skglm.penalties.LogSumPenalty.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.LogSumPenalty<a class="headerlink" href="#skglm-penalties-logsumpenalty" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.LogSumPenalty">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">LogSumPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L616"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.LogSumPenalty" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">LogSumPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L616"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.LogSumPenalty" title="Permalink to this definition">¶</a></dt>
 <dd><p>Log sum penalty.</p>
 <p>The penalty value reads</p>
 <div class="math-wrapper docutils container">
@@ -358,7 +358,7 @@ <h1>skglm.penalties.LogSumPenalty<a class="headerlink" href="#skglm-penalties-lo
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.LogSumPenalty.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L626"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.LogSumPenalty.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">eps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L626"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.LogSumPenalty.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.MCPenalty.html b/generated/skglm.penalties.MCPenalty.html
index 54ff993f4..a0efbd49e 100644
--- a/generated/skglm.penalties.MCPenalty.html
+++ b/generated/skglm.penalties.MCPenalty.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.MCPenalty<a class="headerlink" href="#skglm-penalties-mcpenalty" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.MCPenalty">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">MCPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L212"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.MCPenalty" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">MCPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L212"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.MCPenalty" title="Permalink to this definition">¶</a></dt>
 <dd><p>Minimax Concave Penalty (MCP), a non-convex sparse penalty.</p>
 <p class="rubric">Notes</p>
 <p>With <span class="math notranslate nohighlight">`x &gt;= 0`</span>:</p>
@@ -364,7 +364,7 @@ <h1>skglm.penalties.MCPenalty<a class="headerlink" href="#skglm-penalties-mcpena
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.MCPenalty.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L226"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.MCPenalty.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L226"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.MCPenalty.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.PositiveConstraint.html b/generated/skglm.penalties.PositiveConstraint.html
index ee5f194f4..34e8f760f 100644
--- a/generated/skglm.penalties.PositiveConstraint.html
+++ b/generated/skglm.penalties.PositiveConstraint.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.PositiveConstraint<a class="headerlink" href="#skglm-penalties-positiveconstraint" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.PositiveConstraint">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">PositiveConstraint</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.PositiveConstraint" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">PositiveConstraint</span></span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.PositiveConstraint" title="Permalink to this definition">¶</a></dt>
 <dd><p>Positivity constraint penalty.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.PositiveConstraint.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L679"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.PositiveConstraint.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L679"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.PositiveConstraint.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.SCAD.html b/generated/skglm.penalties.SCAD.html
index a29128b01..49d17e508 100644
--- a/generated/skglm.penalties.SCAD.html
+++ b/generated/skglm.penalties.SCAD.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.SCAD<a class="headerlink" href="#skglm-penalties-scad" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.SCAD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">SCAD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L369"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SCAD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">SCAD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L369"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SCAD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Smoothly Clipped Absolute Deviation.</p>
 <p class="rubric">Notes</p>
 <p>With <span class="math notranslate nohighlight">`x &gt;= 0`</span>:</p>
@@ -368,7 +368,7 @@ <h1>skglm.penalties.SCAD<a class="headerlink" href="#skglm-penalties-scad" title
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.SCAD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L388"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SCAD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L388"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SCAD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.SLOPE.html b/generated/skglm.penalties.SLOPE.html
index 30c4286e3..15c45dd3d 100644
--- a/generated/skglm.penalties.SLOPE.html
+++ b/generated/skglm.penalties.SLOPE.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.SLOPE<a class="headerlink" href="#skglm-penalties-slope" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.SLOPE">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">SLOPE</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alphas</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/non_separable.py#L8"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SLOPE" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">SLOPE</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alphas</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/non_separable.py#L8"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SLOPE" title="Permalink to this definition">¶</a></dt>
 <dd><p>Sorted L-One Penalized Estimation (SLOPE) penalty.</p>
 <p class="rubric">References</p>
 <div role="list" class="citation-list">
@@ -373,7 +373,7 @@ <h1>skglm.penalties.SLOPE<a class="headerlink" href="#skglm-penalties-slope" tit
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.SLOPE.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alphas</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/non_separable.py#L26"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SLOPE.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alphas</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/non_separable.py#L26"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.SLOPE.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.WeightedGroupL2.html b/generated/skglm.penalties.WeightedGroupL2.html
index 0f29551be..fd010f50b 100644
--- a/generated/skglm.penalties.WeightedGroupL2.html
+++ b/generated/skglm.penalties.WeightedGroupL2.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.WeightedGroupL2<a class="headerlink" href="#skglm-penalties-weightedgroupl2" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedGroupL2">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedGroupL2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/block_separable.py#L233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedGroupL2" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedGroupL2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/block_separable.py#L233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedGroupL2" title="Permalink to this definition">¶</a></dt>
 <dd><p>Weighted Group L2 penalty.</p>
 <p>The penalty reads</p>
 <div class="math-wrapper docutils container">
@@ -385,7 +385,7 @@ <h1>skglm.penalties.WeightedGroupL2<a class="headerlink" href="#skglm-penalties-
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedGroupL2.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/block_separable.py#L273"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedGroupL2.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_ptr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grp_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/block_separable.py#L273"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedGroupL2.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.WeightedL1.html b/generated/skglm.penalties.WeightedL1.html
index 33d466895..9647342a3 100644
--- a/generated/skglm.penalties.WeightedL1.html
+++ b/generated/skglm.penalties.WeightedL1.html
@@ -349,11 +349,11 @@
 <h1>skglm.penalties.WeightedL1<a class="headerlink" href="#skglm-penalties-weightedl1" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedL1">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedL1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L145"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedL1" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedL1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L145"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedL1" title="Permalink to this definition">¶</a></dt>
 <dd><p>Weighted L1 penalty.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedL1.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedL1.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedL1.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.penalties.WeightedMCPenalty.html b/generated/skglm.penalties.WeightedMCPenalty.html
index c9918ea2f..57e39f58e 100644
--- a/generated/skglm.penalties.WeightedMCPenalty.html
+++ b/generated/skglm.penalties.WeightedMCPenalty.html
@@ -349,7 +349,7 @@
 <h1>skglm.penalties.WeightedMCPenalty<a class="headerlink" href="#skglm-penalties-weightedmcpenalty" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedMCPenalty">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedMCPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L286"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedMCPenalty" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.penalties.</span></span><span class="sig-name descname"><span class="pre">WeightedMCPenalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L286"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedMCPenalty" title="Permalink to this definition">¶</a></dt>
 <dd><p>Weighted Minimax Concave Penalty (MCP), a non-convex sparse penalty.</p>
 <p class="rubric">Notes</p>
 <p>With <span class="math notranslate nohighlight">`x &gt;= 0`</span>:</p>
@@ -364,7 +364,7 @@ <h1>skglm.penalties.WeightedMCPenalty<a class="headerlink" href="#skglm-penaltie
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.penalties.WeightedMCPenalty.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/penalties/separable.py#L300"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedMCPenalty.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gamma</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positive</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/penalties/separable.py#L300"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.penalties.WeightedMCPenalty.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.AndersonCD.html b/generated/skglm.solvers.AndersonCD.html
index 01c759be0..81c9cf516 100644
--- a/generated/skglm.solvers.AndersonCD.html
+++ b/generated/skglm.solvers.AndersonCD.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.AndersonCD<a class="headerlink" href="#skglm-solvers-andersoncd" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.AndersonCD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">AndersonCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/anderson_cd.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.AndersonCD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">AndersonCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/anderson_cd.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.AndersonCD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Coordinate descent solver with working sets and Anderson acceleration.</p>
 <dl class="simple">
 <dt>fit_intercept<span class="classifier">bool</span></dt><dd><p>Whether or not to fit an intercept.</p>
@@ -387,7 +387,7 @@ <h1>skglm.solvers.AndersonCD<a class="headerlink" href="#skglm-solvers-andersonc
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.AndersonCD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/anderson_cd.py#L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.AndersonCD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/anderson_cd.py#L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.AndersonCD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.FISTA.html b/generated/skglm.solvers.FISTA.html
index e3fc9d11d..3ee89e509 100644
--- a/generated/skglm.solvers.FISTA.html
+++ b/generated/skglm.solvers.FISTA.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.FISTA<a class="headerlink" href="#skglm-solvers-fista" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.FISTA">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">FISTA</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">opt_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/fista.py#L9"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.FISTA" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">FISTA</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">opt_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/fista.py#L9"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.FISTA" title="Permalink to this definition">¶</a></dt>
 <dd><p>ISTA solver with Nesterov acceleration (FISTA).</p>
 <p class="rubric">References</p>
 <div role="list" class="citation-list">
@@ -375,7 +375,7 @@ <h1>skglm.solvers.FISTA<a class="headerlink" href="#skglm-solvers-fista" title="
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.FISTA.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">opt_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/fista.py#L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.FISTA.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">opt_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/fista.py#L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.FISTA.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.GramCD.html b/generated/skglm.solvers.GramCD.html
index 59a251b25..e76e8ae35 100644
--- a/generated/skglm.solvers.GramCD.html
+++ b/generated/skglm.solvers.GramCD.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.GramCD<a class="headerlink" href="#skglm-solvers-gramcd" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.GramCD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GramCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">greedy_cd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/gram_cd.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GramCD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GramCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">greedy_cd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/gram_cd.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GramCD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Coordinate descent solver keeping the gradients up-to-date with Gram updates.</p>
 <p>This solver should be used when <code class="docutils literal notranslate"><span class="pre">n_features</span></code> &lt; <code class="docutils literal notranslate"><span class="pre">n_samples</span></code>, and computes the
 (<code class="docutils literal notranslate"><span class="pre">n_features</span></code>, <code class="docutils literal notranslate"><span class="pre">n_features</span></code>) Gram matrix which comes with an overhead. It is
@@ -392,7 +392,7 @@ <h1>skglm.solvers.GramCD<a class="headerlink" href="#skglm-solvers-gramcd" title
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.GramCD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">greedy_cd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/gram_cd.py#L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GramCD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">greedy_cd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/gram_cd.py#L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GramCD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.GroupBCD.html b/generated/skglm.solvers.GroupBCD.html
index 5ec160dad..4ef410ed5 100644
--- a/generated/skglm.solvers.GroupBCD.html
+++ b/generated/skglm.solvers.GroupBCD.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.GroupBCD<a class="headerlink" href="#skglm-solvers-groupbcd" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.GroupBCD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GroupBCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/group_bcd.py#L11"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupBCD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GroupBCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/group_bcd.py#L11"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupBCD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Block coordinate descent solver for group problems.</p>
 <dl class="field-list simple">
 <dt class="field-odd">Attributes<span class="colon">:</span></dt>
@@ -375,7 +375,7 @@ <h1>skglm.solvers.GroupBCD<a class="headerlink" href="#skglm-solvers-groupbcd" t
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.GroupBCD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/group_bcd.py#L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupBCD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/group_bcd.py#L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupBCD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.GroupProxNewton.html b/generated/skglm.solvers.GroupProxNewton.html
index 14d6b5726..4f9f3d20a 100644
--- a/generated/skglm.solvers.GroupProxNewton.html
+++ b/generated/skglm.solvers.GroupProxNewton.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.GroupProxNewton<a class="headerlink" href="#skglm-solvers-groupproxnewton" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.GroupProxNewton">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GroupProxNewton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/group_prox_newton.py#L15"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupProxNewton" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">GroupProxNewton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/group_prox_newton.py#L15"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupProxNewton" title="Permalink to this definition">¶</a></dt>
 <dd><p>Group Prox Newton solver combined with working sets.</p>
 <dl class="simple">
 <dt>p0<span class="classifier">int, default 10</span></dt><dd><p>Minimum number of features to be included in the working set.</p>
@@ -383,7 +383,7 @@ <h1>skglm.solvers.GroupProxNewton<a class="headerlink" href="#skglm-solvers-grou
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.GroupProxNewton.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/group_prox_newton.py#L50"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupProxNewton.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/group_prox_newton.py#L50"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.GroupProxNewton.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.LBFGS.html b/generated/skglm.solvers.LBFGS.html
index 61219fec8..4e9c2e478 100644
--- a/generated/skglm.solvers.LBFGS.html
+++ b/generated/skglm.solvers.LBFGS.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.LBFGS<a class="headerlink" href="#skglm-solvers-lbfgs" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.LBFGS">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">LBFGS</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/lbfgs.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.LBFGS" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">LBFGS</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/lbfgs.py#L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.LBFGS" title="Permalink to this definition">¶</a></dt>
 <dd><p>A wrapper for scipy L-BFGS solver.</p>
 <p>Refer to <a class="reference external" href="https://docs.scipy.org/doc/scipy/reference/optimize.minimize-lbfgsb.html#optimize-minimize-lbfgsb">scipy L-BFGS-B</a> documentation for details.</p>
 <dl class="field-list simple">
@@ -366,7 +366,7 @@ <h1>skglm.solvers.LBFGS<a class="headerlink" href="#skglm-solvers-lbfgs" title="
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.LBFGS.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/lbfgs.py#L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.LBFGS.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/lbfgs.py#L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.LBFGS.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.MultiTaskBCD.html b/generated/skglm.solvers.MultiTaskBCD.html
index 20bde1fa8..481aee557 100644
--- a/generated/skglm.solvers.MultiTaskBCD.html
+++ b/generated/skglm.solvers.MultiTaskBCD.html
@@ -349,11 +349,11 @@
 <h1>skglm.solvers.MultiTaskBCD<a class="headerlink" href="#skglm-solvers-multitaskbcd" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.MultiTaskBCD">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">MultiTaskBCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/multitask_bcd.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.MultiTaskBCD" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">MultiTaskBCD</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/multitask_bcd.py#L10"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.MultiTaskBCD" title="Permalink to this definition">¶</a></dt>
 <dd><p>Block coordinate descent solver for multi-task problems.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.MultiTaskBCD.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/multitask_bcd.py#L16"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.MultiTaskBCD.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_epochs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">50000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_acc</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/multitask_bcd.py#L16"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.MultiTaskBCD.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/generated/skglm.solvers.ProxNewton.html b/generated/skglm.solvers.ProxNewton.html
index db0b4160e..6c2582ccb 100644
--- a/generated/skglm.solvers.ProxNewton.html
+++ b/generated/skglm.solvers.ProxNewton.html
@@ -349,7 +349,7 @@
 <h1>skglm.solvers.ProxNewton<a class="headerlink" href="#skglm-solvers-proxnewton" title="Permalink to this heading">¶</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="skglm.solvers.ProxNewton">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">ProxNewton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/prox_newton.py#L17"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.ProxNewton" title="Permalink to this definition">¶</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">skglm.solvers.</span></span><span class="sig-name descname"><span class="pre">ProxNewton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/prox_newton.py#L17"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.ProxNewton" title="Permalink to this definition">¶</a></dt>
 <dd><p>Prox Newton solver combined with working sets.</p>
 <dl class="simple">
 <dt>p0<span class="classifier">int, default 10</span></dt><dd><p>Minimum number of features to be included in the working set.</p>
@@ -387,7 +387,7 @@ <h1>skglm.solvers.ProxNewton<a class="headerlink" href="#skglm-solvers-proxnewto
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="skglm.solvers.ProxNewton.__init__">
-<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/ed7bf2d/skglm/solvers/prox_newton.py#L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.ProxNewton.__init__" title="Permalink to this definition">¶</a></dt>
+<span class="sig-name descname"><span class="pre">__init__</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p0</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_pn_iter</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tol</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ws_strategy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'subdiff'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fit_intercept</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warm_start</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/scikit-learn-contrib/skglm/blob/6ac303c/skglm/solvers/prox_newton.py#L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#skglm.solvers.ProxNewton.__init__" title="Permalink to this definition">¶</a></dt>
 <dd></dd></dl>
 
 <p class="rubric">Methods</p>
diff --git a/searchindex.js b/searchindex.js
index c5db2fdc5..d6c8f28f8 100644
--- a/searchindex.js
+++ b/searchindex.js
@@ -1 +1 @@
-Search.setIndex({"docnames": ["api", "auto_examples/index", "auto_examples/plot_compare_time", "auto_examples/plot_group_logistic_regression", "auto_examples/plot_lasso_vs_weighted", "auto_examples/plot_logreg_various_penalties", "auto_examples/plot_pen_prox", "auto_examples/plot_reweighted_l1", "auto_examples/plot_sparse_group_lasso", "auto_examples/plot_sparse_recovery", "auto_examples/plot_survival_analysis", "auto_examples/plot_ucurve", "auto_examples/plot_zero_weights_lasso", "auto_examples/sg_execution_times", "changes/0.1", "changes/0.2", "changes/0.3", "changes/0.4", "changes/whats_new", "contribute", "generated/skglm.CoxEstimator", "generated/skglm.ElasticNet", "generated/skglm.GeneralizedLinearEstimator", "generated/skglm.GroupLasso", "generated/skglm.Lasso", "generated/skglm.LinearSVC", "generated/skglm.MCPRegression", "generated/skglm.MultiTaskLasso", "generated/skglm.SparseLogisticRegression", "generated/skglm.WeightedLasso", "generated/skglm.datafits.Cox", "generated/skglm.datafits.Gamma", "generated/skglm.datafits.Huber", "generated/skglm.datafits.Logistic", "generated/skglm.datafits.LogisticGroup", "generated/skglm.datafits.Poisson", "generated/skglm.datafits.Quadratic", "generated/skglm.datafits.QuadraticGroup", "generated/skglm.datafits.QuadraticSVC", "generated/skglm.datafits.WeightedQuadratic", "generated/skglm.experimental.IterativeReweightedL1", "generated/skglm.experimental.PDCD_WS", "generated/skglm.experimental.Pinball", "generated/skglm.experimental.SqrtLasso", "generated/skglm.experimental.SqrtQuadratic", "generated/skglm.penalties.BlockSCAD", "generated/skglm.penalties.IndicatorBox", "generated/skglm.penalties.L0_5", "generated/skglm.penalties.L1", "generated/skglm.penalties.L1_plus_L2", "generated/skglm.penalties.L2", "generated/skglm.penalties.L2_3", "generated/skglm.penalties.LogSumPenalty", "generated/skglm.penalties.MCPenalty", "generated/skglm.penalties.PositiveConstraint", "generated/skglm.penalties.SCAD", "generated/skglm.penalties.SLOPE", "generated/skglm.penalties.WeightedGroupL2", "generated/skglm.penalties.WeightedL1", "generated/skglm.penalties.WeightedMCPenalty", "generated/skglm.solvers.AndersonCD", "generated/skglm.solvers.FISTA", "generated/skglm.solvers.GramCD", "generated/skglm.solvers.GroupBCD", "generated/skglm.solvers.GroupProxNewton", "generated/skglm.solvers.LBFGS", "generated/skglm.solvers.MultiTaskBCD", "generated/skglm.solvers.ProxNewton", "getting_started", "index", "sg_execution_times", "tutorials/add_datafit", "tutorials/add_penalty", "tutorials/alpha_max", "tutorials/cox_datafit", "tutorials/intercept", "tutorials/prox_nn_group_lasso", "tutorials/tutorials"], "filenames": ["api.rst", "auto_examples/index.rst", "auto_examples/plot_compare_time.rst", "auto_examples/plot_group_logistic_regression.rst", "auto_examples/plot_lasso_vs_weighted.rst", "auto_examples/plot_logreg_various_penalties.rst", "auto_examples/plot_pen_prox.rst", "auto_examples/plot_reweighted_l1.rst", "auto_examples/plot_sparse_group_lasso.rst", "auto_examples/plot_sparse_recovery.rst", "auto_examples/plot_survival_analysis.rst", "auto_examples/plot_ucurve.rst", "auto_examples/plot_zero_weights_lasso.rst", "auto_examples/sg_execution_times.rst", "changes/0.1.rst", "changes/0.2.rst", "changes/0.3.rst", "changes/0.4.rst", "changes/whats_new.rst", "contribute.rst", "generated/skglm.CoxEstimator.rst", "generated/skglm.ElasticNet.rst", "generated/skglm.GeneralizedLinearEstimator.rst", "generated/skglm.GroupLasso.rst", "generated/skglm.Lasso.rst", "generated/skglm.LinearSVC.rst", "generated/skglm.MCPRegression.rst", "generated/skglm.MultiTaskLasso.rst", "generated/skglm.SparseLogisticRegression.rst", "generated/skglm.WeightedLasso.rst", "generated/skglm.datafits.Cox.rst", "generated/skglm.datafits.Gamma.rst", "generated/skglm.datafits.Huber.rst", "generated/skglm.datafits.Logistic.rst", "generated/skglm.datafits.LogisticGroup.rst", "generated/skglm.datafits.Poisson.rst", "generated/skglm.datafits.Quadratic.rst", "generated/skglm.datafits.QuadraticGroup.rst", "generated/skglm.datafits.QuadraticSVC.rst", "generated/skglm.datafits.WeightedQuadratic.rst", "generated/skglm.experimental.IterativeReweightedL1.rst", "generated/skglm.experimental.PDCD_WS.rst", "generated/skglm.experimental.Pinball.rst", "generated/skglm.experimental.SqrtLasso.rst", "generated/skglm.experimental.SqrtQuadratic.rst", "generated/skglm.penalties.BlockSCAD.rst", "generated/skglm.penalties.IndicatorBox.rst", "generated/skglm.penalties.L0_5.rst", "generated/skglm.penalties.L1.rst", "generated/skglm.penalties.L1_plus_L2.rst", "generated/skglm.penalties.L2.rst", "generated/skglm.penalties.L2_3.rst", "generated/skglm.penalties.LogSumPenalty.rst", "generated/skglm.penalties.MCPenalty.rst", "generated/skglm.penalties.PositiveConstraint.rst", "generated/skglm.penalties.SCAD.rst", "generated/skglm.penalties.SLOPE.rst", "generated/skglm.penalties.WeightedGroupL2.rst", "generated/skglm.penalties.WeightedL1.rst", "generated/skglm.penalties.WeightedMCPenalty.rst", "generated/skglm.solvers.AndersonCD.rst", "generated/skglm.solvers.FISTA.rst", "generated/skglm.solvers.GramCD.rst", "generated/skglm.solvers.GroupBCD.rst", "generated/skglm.solvers.GroupProxNewton.rst", "generated/skglm.solvers.LBFGS.rst", "generated/skglm.solvers.MultiTaskBCD.rst", "generated/skglm.solvers.ProxNewton.rst", "getting_started.rst", "index.rst", "sg_execution_times.rst", "tutorials/add_datafit.rst", "tutorials/add_penalty.rst", "tutorials/alpha_max.rst", "tutorials/cox_datafit.rst", "tutorials/intercept.rst", "tutorials/prox_nn_group_lasso.rst", "tutorials/tutorials.rst"], "titles": ["API Documentation", "Examples Gallery", "Timing comparison with scikit-learn for Lasso", "Group Logistic regression in python", "Comparison of Lasso and Weighted Lasso", "Logistic regression with Elastic net and minimax concave penalties", "Value and proximal operators of penalties", "Timing comparison between direct prox computation and reweighting", "Fast Sparse Group Lasso in python", "Sparse recovery with non-convex penalties", "Comparison of lifelines with skglm for survival analysis", "Show U-curve of regularization", "Weighted Lasso with some zero weights", "Computation times", "Version 0.1 (2022/04/25)", "Version 0.2 (2022/10/18)", "Version 0.3 (2023/07/01)", "Version 0.4 (in progress)", "What\u2019s new", "Contribute to <code class=\"docutils literal notranslate\"><span class=\"pre\">skglm</span></code>", "skglm.CoxEstimator", "skglm.ElasticNet", "skglm.GeneralizedLinearEstimator", "skglm.GroupLasso", "skglm.Lasso", "skglm.LinearSVC", "skglm.MCPRegression", "skglm.MultiTaskLasso", "skglm.SparseLogisticRegression", "skglm.WeightedLasso", "skglm.datafits.Cox", "skglm.datafits.Gamma", "skglm.datafits.Huber", "skglm.datafits.Logistic", "skglm.datafits.LogisticGroup", "skglm.datafits.Poisson", "skglm.datafits.Quadratic", "skglm.datafits.QuadraticGroup", "skglm.datafits.QuadraticSVC", "skglm.datafits.WeightedQuadratic", "skglm.experimental.IterativeReweightedL1", "skglm.experimental.PDCD_WS", "skglm.experimental.Pinball", "skglm.experimental.SqrtLasso", "skglm.experimental.SqrtQuadratic", "skglm.penalties.BlockSCAD", "skglm.penalties.IndicatorBox", "skglm.penalties.L0_5", "skglm.penalties.L1", "skglm.penalties.L1_plus_L2", "skglm.penalties.L2", "skglm.penalties.L2_3", "skglm.penalties.LogSumPenalty", "skglm.penalties.MCPenalty", "skglm.penalties.PositiveConstraint", "skglm.penalties.SCAD", "skglm.penalties.SLOPE", "skglm.penalties.WeightedGroupL2", "skglm.penalties.WeightedL1", "skglm.penalties.WeightedMCPenalty", "skglm.solvers.AndersonCD", "skglm.solvers.FISTA", "skglm.solvers.GramCD", "skglm.solvers.GroupBCD", "skglm.solvers.GroupProxNewton", "skglm.solvers.LBFGS", "skglm.solvers.MultiTaskBCD", "skglm.solvers.ProxNewton", "Getting started", "<code class=\"docutils literal notranslate\"><span class=\"pre\">skglm</span></code>", "Computation times", "How to add a custom datafit", "How to add a custom penalty", "Critical regularization strength above which solution is 0", "Mathematic behind Cox datafit", "Computation of the intercept", "Details on the Positive Group Lasso", "Tutorials"], "terms": {"group": [1, 13, 15, 17, 18, 23, 30, 34, 37, 57, 63, 64, 70, 74], "logist": [1, 13, 15, 18, 28, 34, 70], "regress": [1, 8, 13, 20, 22, 26, 28, 70], "python": [1, 2, 4, 5, 6, 7, 9, 10, 11, 12, 13, 69, 70, 72], "valu": [1, 7, 8, 9, 10, 13, 20, 45, 52, 53, 55, 56, 59, 62, 63, 68, 70, 71, 72, 73, 75], "proxim": [1, 7, 10, 13, 15, 18, 41, 57, 70, 71, 72], "oper": [1, 13, 57, 70, 71, 72, 74], "penalti": [1, 2, 3, 7, 8, 10, 12, 13, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 29, 34, 37, 40, 41, 43, 62, 68, 69, 70, 71, 73], "show": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 68, 70, 71, 73, 76], "u": [1, 2, 13, 19, 68, 69, 70, 73, 74, 76], "curv": [1, 13, 70], "regular": [1, 7, 10, 13, 16, 18, 20, 21, 24, 25, 26, 27, 28, 29, 56, 57, 68, 69, 70, 72], "comparison": [1, 13, 70], "lasso": [1, 9, 11, 13, 15, 16, 17, 18, 21, 23, 26, 29, 43, 57, 70], "weight": [1, 3, 6, 13, 17, 18, 23, 24, 26, 29, 39, 40, 57, 58, 59, 70], "some": [1, 4, 6, 8, 13, 68, 70, 71, 72, 74, 76], "zero": [1, 3, 4, 7, 8, 9, 11, 13, 41, 62, 63, 70, 71, 72, 76], "fast": [1, 12, 13, 61, 69, 70], "spars": [1, 6, 10, 13, 21, 22, 24, 25, 26, 27, 28, 29, 30, 41, 53, 59, 60, 64, 67, 69, 70, 71], "elast": [1, 13, 16, 18, 20, 21, 70], "net": [1, 13, 21, 70], "minimax": [1, 13, 53, 59, 70], "concav": [1, 13, 53, 59, 70], "time": [1, 3, 4, 5, 6, 8, 9, 11, 12, 16, 18, 20, 30, 31, 33, 35, 38, 69, 71, 74, 75, 76], "between": [1, 10, 11, 13, 70], "direct": [1, 13, 70], "prox": [1, 13, 26, 28, 41, 43, 64, 67, 70, 76], "comput": [1, 9, 11, 12, 17, 18, 31, 33, 35, 36, 38, 39, 62, 68, 71, 72, 74, 76], "reweight": [1, 13, 40, 70], "scikit": [1, 3, 8, 13, 19, 68, 69, 70], "learn": [1, 3, 8, 13, 19, 68, 69, 70, 73, 77], "recoveri": [1, 13, 70], "non": [1, 3, 8, 13, 15, 16, 18, 25, 40, 47, 51, 53, 59, 70, 72, 76], "convex": [1, 13, 40, 41, 46, 47, 51, 53, 56, 59, 70, 73, 74], "lifelin": [1, 13, 70], "skglm": [1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 68, 70, 71, 72, 73, 75, 77], "surviv": [1, 13, 16, 18, 30, 70, 74], "analysi": [1, 13, 16, 18, 30, 70, 74], "download": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "all": [1, 3, 9, 11, 12, 70, 72, 75, 76], "sourc": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67], "code": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 41, 60, 64, 67, 69], "auto_examples_python": 1, "zip": [1, 2, 10], "jupyt": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "notebook": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "auto_examples_jupyt": 1, "gener": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 22, 64, 67, 69, 71, 72, 75], "sphinx": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "go": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "end": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 73], "full": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "exampl": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 19, 68, 70, 71, 77], "compar": [2, 7, 10], "solv": [2, 10, 16, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 69, 71, 75], "larg": [2, 12, 41, 68, 69], "scale": [2, 64, 67, 72], "problem": [2, 7, 10, 15, 16, 18, 20, 22, 25, 28, 40, 41, 60, 61, 63, 66, 71], "file_s": [2, 11], "0": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 42, 43, 46, 47, 53, 55, 57, 59, 60, 61, 62, 63, 64, 65, 66, 67, 70, 71, 72, 74, 75], "00": [2, 11, 13, 70], "26": [2, 15, 18], "8m": 2, "b": [2, 11, 64, 65, 67, 73, 74, 76], "": [2, 5, 7, 9, 10, 11, 20, 30, 64, 67, 68, 69, 71, 73, 74, 76], "24": [2, 7, 11], "6k": [2, 11], "03": [2, 3, 13, 70], "31": 2, "127kb": 2, "49": [2, 11], "2k": [2, 11], "126kb": 2, "106k": [2, 11], "02": [2, 3, 8, 9, 11, 13, 70], "11": [2, 8, 11, 13, 70], "203kb": 2, "1": [2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 45, 46, 52, 53, 55, 56, 57, 59, 60, 61, 62, 64, 67, 68, 71, 72, 73, 74, 75, 76], "221k": [2, 11], "01": [2, 3, 8, 11, 73], "14": [2, 3, 13, 15, 18, 70], "355kb": 2, "2": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 29, 32, 36, 37, 38, 39, 41, 45, 50, 51, 53, 55, 59, 60, 62, 64, 67, 71, 73, 74, 75, 76], "451k": [2, 11], "40": [2, 13, 56, 70], "652kb": 2, "3": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 26, 51, 56, 69, 73, 74, 76], "909k": [2, 11], "21": [2, 11, 27], "23mb": 2, "7": [2, 6, 8, 11, 20, 74], "83m": [2, 11], "10": [2, 3, 4, 7, 8, 10, 11, 12, 21, 23, 24, 25, 26, 27, 29, 41, 43, 56, 60, 61, 63, 64, 66, 67, 68], "37mb": 2, "66m": [2, 11], "05": [2, 3, 8, 11, 13, 70], "4": [2, 3, 4, 6, 8, 9, 10, 11, 23, 43, 61, 62, 63, 64, 65, 67, 73, 75, 76], "59mb": 2, "20": [2, 10, 14, 18, 28, 43, 64, 65, 67], "5": [2, 3, 4, 7, 8, 9, 10, 11, 12, 20, 21, 40, 42, 47, 62, 76], "23m": [2, 11], "69mb": [2, 11], "29": [2, 15, 18], "86m": 2, "99mb": 2, "35": 2, "9": [2, 8, 56], "43m": 2, "8": [2, 5, 6, 7, 8, 9, 11], "06mb": 2, "45": [2, 13, 70], "12": [2, 9, 11, 15], "1m": 2, "63mb": 2, "51": [2, 10, 11, 15, 18], "13": [2, 11, 13, 70, 74], "6m": [2, 11], "24mb": 2, "61": [2, 11], "16": 2, "2m": 2, "5mb": 2, "70": 2, "18": [2, 73], "9m": 2, "4mb": 2, "80": [2, 11], "5m": 2, "0mb": 2, "90": 2, "100": [2, 5, 7, 10, 11, 12, 41, 43, 61, 62, 63, 66, 68, 69], "import": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "warn": [2, 10], "numpi": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "np": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 26, 30, 41, 68, 71, 72, 73], "from": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 17, 18, 21, 22, 24, 25, 26, 27, 29, 68, 69, 70, 71, 72, 74], "linalg": [2, 4, 5, 7, 9, 10, 11], "norm": [2, 4, 5, 7, 9, 10, 11, 24, 29, 40, 47, 51, 73, 75, 76], "matplotlib": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "pyplot": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "plt": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "libsvmdata": [2, 11], "fetch_libsvm": [2, 11], "sklearn": [2, 5, 9, 11, 68], "except": [2, 68], "convergencewarn": 2, "linear_model": 2, "lasso_sklearn": 2, "elasticnet": [2, 20, 49], "enet_sklearn": 2, "filterwarn": [2, 10], "ignor": [2, 10], "categori": 2, "def": [2, 7, 71, 72], "compute_obj": 2, "x": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 23, 25, 26, 27, 30, 32, 36, 39, 53, 55, 59, 62, 68, 71, 73, 74, 75, 76], "y": [2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 26, 27, 29, 36, 37, 39, 43, 44, 62, 68, 71, 74, 75], "w": [2, 7, 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, 38, 40, 41, 43, 45, 50, 52, 56, 62, 71, 72], "alpha": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 26, 27, 28, 29, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 55, 56, 57, 58, 59, 68, 72, 73], "l1_ratio": [2, 5, 6, 10, 20, 21, 49], "loss": [2, 25, 73], "len": [2, 4, 5, 7, 8, 9, 10, 11, 12, 71], "sum": [2, 7, 10, 52, 71, 72, 74], "ab": [2, 4, 7, 12, 32, 41, 52, 53, 55, 59, 60, 64, 67, 72, 73], "return": [2, 7, 71, 72], "news20": 2, "binari": [2, 11, 28, 72, 74], "max": [2, 4, 7, 9, 12, 25, 42, 68, 72, 73, 76], "t": [2, 4, 7, 9, 11, 12, 25, 28, 36, 38, 39, 62, 64, 67, 68, 71, 73], "dict_sklearn": 2, "fit_intercept": [2, 4, 7, 8, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 28, 29, 60, 62, 63, 64, 66, 67], "fals": [2, 4, 6, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 48, 49, 53, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 72], "tol": [2, 7, 8, 11, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "1e": [2, 7, 8, 9, 10, 11, 23, 41, 43, 61, 62, 63, 64, 65, 66, 67], "enet": 2, "dict_our": 2, "model": [2, 3, 8, 10, 41, 60, 63, 64, 67, 68, 69, 71, 75], "fig": [2, 4, 6, 7, 9, 10, 12], "axarr": [2, 4, 6, 7, 9, 12], "subplot": [2, 4, 6, 7, 9, 10, 12], "constrained_layout": [2, 6, 7, 9, 12], "true": [2, 4, 5, 6, 7, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 28, 29, 30, 57, 60, 62, 66, 67, 72], "ax": [2, 7, 10, 73], "pobj_dict": 2, "list": [2, 10, 23, 28, 40, 68], "time_dict": 2, "remov": [2, 17, 18], "compil": [2, 7, 22, 31, 33, 35, 38, 40, 68, 71, 72], "max_it": [2, 10, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "10_000": 2, "w_star": [2, 5], "fit": [2, 3, 4, 5, 7, 8, 11, 12, 15, 16, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 33, 35, 38, 40, 60, 63, 67, 71, 77], "coef_": [2, 3, 4, 5, 7, 8, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40], "pobj_star": 2, "n_iter_sklearn": 2, "uniqu": [2, 10, 30, 56, 74], "geomspac": [2, 9, 11], "50": [2, 4, 5, 10, 11, 12, 20, 21, 23, 24, 25, 26, 27, 29, 60, 65], "num": [2, 6, 9], "15": [2, 56], "astyp": 2, "int": [2, 9, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67], "t_start": 2, "w_sklearn": 2, "append": [2, 10, 11], "n_iter_u": 2, "rang": [2, 9, 10, 69, 71], "semilogi": [2, 10], "label": [2, 5, 6, 9, 10, 11, 28, 33], "set_ylim": [2, 9], "set_titl": [2, 4, 6, 7, 10, 12], "legend": [2, 5, 6, 9, 10, 11], "set_ylabel": [2, 7, 9, 10], "object": [2, 7, 8, 10, 21, 23, 24, 25, 26, 27, 28, 29, 40, 43, 71, 72, 76], "suboptim": [2, 10], "set_xlabel": [2, 7, 9, 10, 12], "block": [2, 4, 6, 7, 9, 11, 12, 15, 18, 23, 45, 60, 63, 66, 74, 76], "total": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 70, 74], "run": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 22, 69], "script": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "minut": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "604": [2, 13, 70], "second": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 72, 74], "plot_compare_tim": [2, 13, 70], "ipynb": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "py": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 70], "galleri": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 70], "i": [3, 7, 8, 10, 11, 12, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 38, 39, 40, 41, 43, 45, 46, 56, 57, 60, 61, 62, 63, 64, 65, 67, 68, 69, 71, 72, 74, 75, 76], "miss": [3, 8, 69], "estim": [3, 7, 8, 9, 11, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 43, 56, 69], "we": [3, 8, 10, 11, 19, 25, 68, 71, 72, 73, 74, 75, 76], "how": [3, 8, 10, 12, 19, 41, 68, 76], "implement": [3, 8, 68, 69, 71, 72, 75, 77], "one": [3, 5, 8, 76], "author": [3, 4, 5, 6, 7, 8, 9, 69], "mathurin": [3, 4, 6, 8, 9], "massia": [3, 4, 6, 8, 9, 41, 60, 64, 67, 69], "generalizedlinearestim": [3, 5, 7, 8, 68, 71], "datafit": [3, 5, 8, 9, 10, 15, 16, 17, 18, 20, 22, 25, 40, 41, 42, 44, 62, 68, 69, 72], "logisticgroup": 3, "weightedgroupl2": 3, "solver": [3, 7, 8, 9, 10, 12, 15, 16, 17, 18, 20, 22, 23, 24, 29, 40, 41, 43, 69, 71, 72, 75], "groupproxnewton": 3, "util": [3, 4, 5, 7, 8, 9, 10, 12], "data": [3, 4, 5, 7, 8, 9, 11, 12, 20, 71, 74], "make_correlated_data": [3, 4, 5, 7, 8, 9, 12], "grp_convert": [3, 8], "n_featur": [3, 4, 5, 7, 8, 9, 10, 12, 15, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 34, 36, 37, 39, 40, 56, 57, 62, 63, 68, 71, 72], "30": [3, 4, 8, 9], "_": [3, 4, 5, 8, 10, 27, 41, 46, 71, 73, 74, 75], "n_sampl": [3, 4, 5, 7, 8, 9, 10, 12, 15, 18, 25, 30, 39, 41, 62, 63, 68, 71], "random_st": [3, 4, 5, 7, 8, 9, 10, 12], "sign": [3, 5, 72, 75], "classifi": [3, 28], "creation": [3, 8], "combin": [3, 8, 10, 20, 21, 64, 67, 68, 69, 74, 76], "grp_size": [3, 8], "ar": [3, 8, 10, 19, 23, 26, 29, 41, 68, 69, 71, 72, 74, 75], "made": [3, 69], "consecut": [3, 8, 30, 34, 37, 57], "featur": [3, 4, 8, 11, 12, 19, 23, 26, 28, 29, 41, 43, 45, 52, 53, 55, 56, 59, 60, 62, 64, 67, 68, 69, 71, 72, 75], "n_group": [3, 8, 23, 34, 37, 57], "grp_indic": [3, 8, 34, 37, 57], "grp_ptr": [3, 8, 34, 37, 57], "ones": [3, 8, 72, 75], "verbos": [3, 8, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "train": [3, 5, 8, 11, 68], "clf": [3, 8, 11, 12, 71], "iter": [3, 7, 8, 10, 20, 21, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67, 72], "6931471806": 3, "stop": [3, 8, 20, 21, 23, 24, 25, 26, 27, 28, 29, 72], "crit": [3, 8], "93e": 3, "pn": 3, "1818203070": 3, "38e": 3, "earli": 3, "exit": 3, "1130022685": 3, "59e": [3, 8, 10], "0920956247": 3, "58e": 3, "0846061796": 3, "6": [3, 8, 9, 10, 11, 74], "29e": 3, "0799359080": 3, "83e": 3, "0794149142": 3, "96e": 3, "04": [3, 10], "0794078275": 3, "61e": 3, "In": [3, 8, 68, 73, 74, 75], "environ": [3, 8], "pleas": [3, 8, 69], "rerun": [3, 8], "thi": [3, 8, 10, 12, 19, 22, 25, 31, 33, 35, 38, 40, 62, 68, 71, 72, 73, 74, 75, 76], "cell": [3, 8], "html": [3, 8, 19, 60, 64, 67], "represent": [3, 8], "trust": [3, 8], "On": [3, 8, 74], "github": [3, 8, 19, 41, 60, 64, 67], "unabl": [3, 8], "render": [3, 8], "try": [3, 8], "load": [3, 8, 11], "page": [3, 8, 68, 69], "nbviewer": [3, 8], "org": [3, 8, 41, 56, 60, 61, 64, 67], "generalizedlinearestimatorgeneralizedlinearestim": [3, 8], "check": [3, 10, 17, 18, 19, 68], "either": [3, 69], "print": [3, 9, 10, 68], "reshap": [3, 8], "10333799": 3, "8924748": 3, "77600881": 3, "68695901": 3, "26460422": 3, "06153908": 3, "13524978": 3, "01904796": 3, "02993834": 3, "02193371": 3, "02044038": 3, "00869443": 3, "90363392": 3, "96337048": 3, "49341585": 3, "57737623": 3, "25394772": 3, "15405176": 3, "785": [3, 13, 70], "plot_group_logistic_regress": [3, 13, 70], "illustr": [4, 5, 6, 9, 11, 68, 74], "normal": [4, 10], "when": [4, 16, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 42, 56, 57, 62, 71, 72], "penal": [4, 10, 12, 26, 56, 72, 75], "quentin": [4, 9], "bertrand": [4, 9, 41, 60, 69], "weightedlasso": [4, 12, 24], "w_true": [4, 7, 9, 12], "nnz": 4, "assum": [4, 23, 74], "reason": [4, 69], "have": [4, 10, 19, 30, 68, 71, 72, 73, 74, 75], "smaller": 4, "than": [4, 10, 24, 26, 29, 68, 69], "other": [4, 68, 69, 71, 74, 76], "nois": 4, "random": [4, 9, 12, 68], "randn": 4, "shape": [4, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 34, 36, 37, 39, 40, 41, 56, 57, 62, 63, 71], "us": [4, 5, 10, 12, 19, 20, 21, 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 36, 37, 38, 39, 41, 60, 62, 63, 67, 68, 69, 71, 72, 73, 74, 75, 76], "signal": 4, "ratio": [4, 10], "doe": 4, "select": [4, 26, 56, 62, 68], "small": [4, 12, 19], "while": [4, 68], "alpha_max": [4, 7, 9, 11, 12, 72, 73], "la": 4, "wei": 4, "axi": [4, 5, 8, 9], "sharei": [4, 9, 12], "figsiz": [4, 6, 7, 9, 10, 12], "stem": [4, 5, 12], "coeff": 4, "731": [4, 13, 70], "plot_lasso_vs_weight": [4, 13, 70], "modular": [5, 68, 69], "pierr": [5, 7], "antoin": [5, 7], "bannier": [5, 7, 41, 60, 69], "metric": [5, 9, 10, 11], "f1_score": [5, 9], "l1_plus_l2": [5, 6, 16, 18], "mcpenalti": [5, 6, 9], "y_ind": 5, "standard": [5, 9, 10], "mcp": [5, 9, 26, 53, 59], "sqrt": [5, 7, 9, 76], "split": [5, 68], "set": [5, 10, 15, 17, 18, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 46, 57, 60, 62, 63, 64, 67, 68, 72, 74, 76], "test": [5, 11, 19, 68], "x_train": [5, 11, 68], "y_train": [5, 11, 68], "x_test": [5, 9, 11, 68], "y_test": [5, 9, 11, 68], "005": 5, "gamma": [5, 6, 9, 16, 18, 26, 45, 53, 55, 59], "clf_enet": 5, "y_pred_enet": 5, "predict": [5, 11, 68], "f1_score_enet": 5, "clf_mcp": 5, "y_pred_mcp": 5, "f1_score_mcp": 5, "m": [5, 41, 56, 60, 61, 64, 67, 69, 74], "where": [5, 10, 20, 30, 32, 41, 46, 57, 62, 73, 74, 75, 76], "ravel": 5, "markerfmt": 5, "coeffici": [5, 8, 10, 17, 18, 21, 23, 24, 26, 28, 29, 41, 57, 62, 63, 68, 71, 72, 74, 75], "setp": 5, "color": [5, 9], "2ca02c": 5, "ff7f0e": 5, "bx": 5, "loc": [5, 9], "best": 5, "titl": [5, 8, 11, 69], "f1": [5, 9], "3f": 5, "793": [5, 13, 70], "plot_logreg_various_penalti": [5, 13, 70], "gmail": [6, 7], "com": [6, 7, 19, 41, 60, 64, 67, 76], "weightedl1": [6, 16, 18], "l1": [6, 9, 10, 16, 18, 20, 21, 22, 23, 24, 26, 29, 40, 41, 56, 58, 60, 68, 69, 73, 77], "scad": [6, 9, 15, 18], "l0_5": [6, 7, 9, 40], "l2_3": [6, 9], "arrai": [6, 7, 10, 20, 21, 22, 23, 24, 25, 26, 27, 29, 30, 34, 36, 37, 39, 40, 41, 56, 57, 62, 63, 71], "x_rang": 6, "linspac": 6, "300": 6, "pen": [6, 26, 45, 53, 55, 59], "plot": [6, 7, 9, 10, 11, 19], "__class__": 6, "__name__": 6, "prox_1d": [6, 72], "314": [6, 13, 70], "plot_pen_prox": [6, 13, 70], "pierreantoin": 7, "panda": [7, 10], "pd": [7, 10], "separ": [7, 17, 18, 40, 41, 76], "experiment": [7, 15, 16, 18], "iterativereweightedl1": 7, "andersoncd": [7, 9, 22, 40, 71], "200": 7, "500": [7, 10], "ord": [7, 9, 11], "inf": [7, 9, 11, 26, 72], "1000": [7, 9, 28, 41, 43, 63, 64, 67, 68], "_obj": 7, "fit_l05": 7, "start": [7, 10, 69, 74], "iterative_l05": 7, "iterative_tim": 7, "subdiff": [7, 21, 23, 24, 25, 26, 27, 29, 60, 61, 63, 66, 67, 72], "strategi": [7, 41, 62, 76], "uninform": 7, "direct_l05": 7, "ws_strategi": [7, 8, 9, 21, 23, 24, 25, 26, 27, 29, 60, 63, 66, 67, 72], "fixpoint": [7, 8, 9, 21, 23, 24, 25, 26, 27, 29, 60, 67, 72], "direct_tim": 7, "result": [7, 69, 74], "cach": [7, 68], "numba": [7, 31, 33, 35, 38, 68, 71, 72], "time_result": 7, "obj_result": 7, "actual": 7, "enumer": [7, 9, 10, 11, 72], "iterative_obj": 7, "direct_obj": 7, "time_df": 7, "datafram": [7, 10], "column": [7, 20, 74, 75], "obj_df": 7, "index": [7, 8, 12, 74], "bar": 7, "rot": 7, "r": [7, 9, 11, 12, 71, 72, 73, 74], "lambda": [7, 9, 11, 72, 73, 76], "lambda_": [7, 9, 73], "obj": [7, 10, 26], "solut": [7, 9, 10, 11, 21, 23, 24, 25, 26, 27, 28, 29, 72, 75], "021": [7, 13, 70], "plot_reweighted_l1": [7, 13, 70], "groupbcd": [8, 15, 18], "quadraticgroup": 8, "weightedl1groupl2": 8, "take": [8, 22, 68, 71, 76], "weights_g": [8, 23], "dtype": [8, 71, 72], "float64": [8, 72], "weights_f": 8, "weights_group": 8, "weights_featur": 8, "9165123923": 8, "04e": 8, "0255867844": 8, "6875753707": 8, "40e": 8, "6804572587": 8, "23e": 8, "6804570767": 8, "22e": 8, "6804570709": 8, "64e": 8, "07": 8, "17e": 8, "08": [8, 13, 70], "37e": 8, "fulli": [8, 69], "insid": 8, "too": [8, 11], "imshow": 8, "cmap": [8, 9], "grei": 8, "black": 8, "ylabel": 8, "xlabel": [8, 11], "xtick": 8, "arang": [8, 12], "ytick": 8, "0x7fb9b50f1970": 8, "0x7fb9b4011e80": 8, "0x7fb9b4e21490": 8, "text": [8, 73, 74, 76], "547": [8, 13, 70], "plot_sparse_group_lasso": [8, 13, 70], "superior": 9, "perform": [9, 40, 41, 69, 74], "l05": 9, "l23": 9, "klopfenstein": [9, 41, 60, 69], "model_select": [9, 11, 68], "train_test_split": [9, 11, 68], "mean_squared_error": [9, 11], "quadrat": [9, 15, 17, 18, 22, 37, 38, 39, 40, 44, 62, 71], "jit_compil": [9, 10], "compiled_clon": [9, 10], "get_cmap": 9, "tab10": 9, "simul": 9, "densiti": 9, "seed": [9, 12], "supp": 9, "choic": [9, 12, 68], "size": [9, 21, 23, 24, 25, 26, 27, 29, 41, 60, 74, 75], "replac": [9, 12, 19, 22, 40, 69], "x_": [9, 71, 75, 76], "y_": [9, 74], "snr": 9, "rho": 9, "test_siz": 9, "which": [9, 10, 11, 62, 69, 72, 74], "lead": [9, 75, 76], "defin": [9, 11, 74, 75], "n_alpha": 9, "estimation_error": 9, "prediction_error": 9, "l0": 9, "mse_ref": 9, "zeros_lik": [9, 72], "idx": [9, 10, 11, 72], "kei": [9, 19], "f": [9, 10, 32, 41, 71, 72, 73, 74, 75], "estimator_path": 9, "path": [9, 11], "f1_temp": 9, "prediction_error_temp": 9, "j": [9, 26, 41, 45, 52, 53, 55, 59, 61, 64, 67, 71, 72, 73, 74, 75, 76], "name_estim": 9, "ell_": [9, 47, 51], "close": [9, 10, 11], "sharex": 9, "semilogx": [9, 11], "c": [9, 25, 56, 64, 67, 68, 69, 76], "max_f1": 9, "argmax": 9, "vline": 9, "ymin": 9, "ymax": 9, "linestyl": 9, "line1": 9, "clip_on": 9, "marker": [9, 10], "markers": 9, "min_error": 9, "argmin": [9, 73, 75], "lim": 9, "get_ylim": 9, "min": [9, 10, 73, 76], "line2": 9, "mathrm": 9, "score": [9, 21, 23, 24, 25, 26, 27, 29, 41, 60, 67, 68], "pred": 9, "rmse": 9, "left": [9, 71], "out": [9, 10, 19], "bbox_to_anchor": 9, "lower": [9, 68], "mode": [9, 19], "expand": 9, "borderaxespad": 9, "ncol": 9, "910": [9, 13, 70], "plot_sparse_recoveri": [9, 13, 70], "exactli": [10, 76], "x100": 10, "less": 10, "let": [10, 11, 68, 71, 73, 74, 75, 76], "first": [10, 11, 12, 19, 20, 21, 23, 24, 25, 26, 27, 29, 41, 60, 68, 69, 71, 72, 74], "synthet": 10, "make_dummy_survival_data": 10, "tm": [10, 20, 30], "The": [10, 12, 20, 21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 38, 39, 41, 42, 43, 44, 50, 52, 56, 57, 60, 67, 68, 71, 72, 73, 74, 76], "ha": [10, 68, 74, 76], "follow": [10, 19, 71, 72, 75], "properti": [10, 21, 22, 24, 25, 26, 27, 29], "matrix": [10, 20, 21, 22, 24, 25, 26, 27, 29, 62, 71, 74, 75], "predictor": [10, 20, 74], "distribut": 10, "toeplitz": 10, "covari": 10, "vector": [10, 11, 20, 21, 23, 24, 25, 26, 27, 29, 40, 41, 57, 62, 63, 71, 72, 74, 75], "occurr": [10, 20, 30, 74], "weibul": 10, "indic": [10, 23, 30, 34, 37, 46, 57, 71, 75], "observ": [10, 20, 30, 74], "censorship": [10, 74], "bernoulli": 10, "inspect": 10, "quickli": 10, "tight_layout": 10, "dist": [10, 72], "axes_titl": 10, "fifth": 10, "name": [10, 72], "hist": 10, "bin": 10, "auto": 10, "count": 10, "after": [10, 40, 68, 75], "can": [10, 11, 19, 21, 23, 24, 25, 26, 27, 29, 62, 68, 69, 71, 72, 74, 75, 76], "now": [10, 68, 71, 74, 75], "todo": 10, "so": [10, 19, 69], "need": [10, 19, 71, 72, 76], "ell_1": [10, 48, 49, 72, 77], "newton": [10, 15, 18, 28, 43, 64, 67], "proxnewton": [10, 15, 16, 17, 18, 20, 71], "intens": 10, "intern": 10, "init": [10, 68], "initi": [10, 14, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 62, 63, 71, 72], "w_sk": 10, "For": [10, 11, 20, 21, 26, 68, 71, 72, 73, 74, 76], "rel": 10, "found": [10, 71, 72], "number": [10, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67, 68, 73, 75], "nonzero": 10, "66": 10, "through": [10, 68, 69, 74, 77], "its": [10, 41, 68, 69, 71, 72, 75, 76], "coxphfitt": 10, "two": [10, 20, 30, 34, 37, 57, 71, 74], "packag": [10, 69, 75], "format": 10, "stacked_y_x": 10, "hstack": 10, "df": 10, "lifelines_estim": 10, "duration_col": 10, "event_col": 10, "w_ll": 10, "params_": 10, "both": [10, 71, 74], "find": [10, 68], "same": [10, 30, 68, 74], "obj_sk": 10, "obj_ll": 10, "6f": 10, "differ": [10, 71, 75], "2e": 10, "494576": 10, "47e": 10, "do": [10, 19, 69, 75], "euclidean": 10, "distanc": [10, 17, 18, 41, 57, 72, 76], "3e": 10, "373e": 10, "yield": [10, 71, 76], "execut": [10, 13, 70], "To": [10, 19, 25, 71, 72, 74], "get": [10, 69, 71, 72, 74, 77], "evolut": 10, "optim": [10, 11, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 56, 60, 64, 67, 71, 72, 73, 75, 76], "increas": [10, 76], "save": 10, "record": [10, 74], "max_run": 10, "n_iter": 10, "perf_count": 10, "fit_opt": 10, "max_step": 10, "cast": [10, 74], "asarrai": 10, "optimal_obj": 10, "o": [10, 74], "accord": [10, 73, 74], "more": [10, 26, 68, 73], "speed_up": 10, "speed": [10, 68, 69], "up": [10, 62, 69], "0f": 10, "879": 10, "previou": [10, 21, 23, 24, 25, 26, 27, 28, 29], "extend": [10, 19, 74], "case": [10, 19, 69, 74, 75], "handl": [10, 12, 20, 22, 28, 30, 39, 74, 75], "ti": [10, 20, 30, 74], "achiev": [10, 68], "pass": [10, 23, 68, 71, 72], "with_ti": 10, "function": [10, 15, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 40, 41, 46, 57, 73, 74, 75, 76], "It": [10, 20, 22, 30, 41, 62, 68, 69, 72, 74], "straightforward": 10, "an": [10, 11, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 60, 62, 67, 68, 71, 72, 74, 75, 77], "onli": [10, 62, 71, 72, 73, 74, 75, 76], "use_efron": [10, 30], "ensur": [10, 69], "again": [10, 74], "82": 10, "stacked_tm_s_x": 10, "none": [10, 22, 23, 26, 29, 40, 41, 62, 63], "new": [10, 19, 69, 72], "501289": 10, "09": [10, 11], "046e": 10, "final": [10, 11, 72, 74, 75], "total_time_skglm": 10, "total_time_lifelin": 10, "deduc": [10, 74], "787": 10, "As": [10, 19, 76], "shown": [10, 73], "last": [10, 12, 19, 68], "still": 10, "preserv": 10, "even": [10, 19], "52": [10, 13, 70], "615": [10, 13, 70], "plot_survival_analysi": [10, 13, 70], "sweet": 11, "spot": 11, "much": [11, 41], "littl": 11, "showcas": 11, "rcv1": 11, "dataset": [11, 68, 69, 71], "keep": [11, 62], "2000": 11, "also": [11, 41, 69, 74], "retrain": 11, "sampl": [11, 17, 18, 21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 39, 46, 62, 68, 71, 73, 74, 75], "7m": 11, "123kb": 11, "122kb": 11, "196kb": 11, "39": 11, "343kb": 11, "630kb": 11, "18mb": 11, "28mb": 11, "27": 11, "41mb": 11, "38": 11, "46mb": 11, "81m": 11, "38m": 11, "0m": 11, "51mb": 11, "92": 11, "36mb": 11, "20mb": 11, "next": 11, "well": [11, 68, 72], "know": 11, "abov": 11, "along": 11, "mse": 11, "mse_train": 11, "mse_test": 11, "warm_start": [11, 21, 23, 24, 25, 26, 27, 28, 29, 60, 62, 63, 64, 66, 67], "notic": [11, 74], "around": 11, "sit": 11, "boundari": 11, "underfit": 11, "overfit": 11, "mean": [11, 72, 74], "squar": [11, 15, 18, 43, 44, 73, 74, 76], "error": [11, 68], "strength": [11, 20, 21, 23, 24, 25, 26, 27, 28, 29, 43], "278": [11, 13, 70], "plot_ucurv": [11, 13, 70], "demonstr": 12, "vanish": 12, "adapt": [12, 41, 56, 68], "primal": [12, 16, 18, 23, 24, 25, 29, 41], "anderson": [12, 60], "acceler": [12, 15, 18, 60, 61], "allow": [12, 31, 33, 35, 38, 68, 71, 72], "dual": [12, 16, 18, 25, 41, 64, 67], "empti": [12, 76], "unpen": [12, 23, 29, 67, 77], "put": [12, 72], "coef": 12, "neq": 12, "heavili": 12, "few": [12, 69], "lightli": 12, "mani": [12, 69, 75], "342": [12, 13, 70], "plot_zero_weights_lasso": [12, 13, 70], "942": [13, 70], "file": [13, 70], "auto_exampl": 13, "mem": [13, 70], "mb": [13, 70], "releas": [14, 18], "commit": [14, 18], "root": [15, 18, 43, 44], "class": [15, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72, 75], "chambol": [15, 18], "pock": [15, 18], "pr": [15, 16, 17, 18], "57": [15, 18], "coordin": [15, 16, 18, 22, 26, 41, 60, 62, 63, 66, 71, 72, 76], "descent": [15, 16, 18, 22, 26, 41, 60, 62, 63, 66, 75], "work": [15, 17, 18, 21, 23, 24, 25, 26, 27, 29, 41, 43, 60, 63, 64, 67, 76], "28": [15, 18], "huber": [15, 18], "ad": [15, 18, 19, 68, 75], "blockscad": [15, 18], "intercept": [15, 18, 21, 23, 24, 25, 26, 27, 28, 29, 60, 67], "55": [15, 18], "gramcd": [15, 18], "gram": [15, 18, 62], "base": [15, 18, 23, 24, 29, 41, 43, 62, 71, 72], "ll": [15, 18], "59": [15, 18], "transform": [15, 18], "63": [15, 18], "add": [16, 17, 18, 19, 68], "cox": [16, 18, 20], "support": [16, 17, 18, 23, 29, 68, 69, 71, 72], "l2": [16, 18, 20, 21, 25, 57], "171": [16, 18], "reduc": [16, 18, 74, 76], "overhead": [16, 18, 62, 68], "129": [16, 18], "180": [16, 18], "168": [16, 18], "159": [16, 18], "157": [16, 18], "pinbal": [16, 18], "134": [16, 18], "113": [16, 18], "posit": [16, 17, 18, 20, 21, 23, 24, 25, 26, 28, 29, 48, 49, 53, 54, 57, 58, 59, 71, 72, 74], "constraint": [16, 18, 46, 54, 77], "110": [16, 18], "positiveconstraint": [16, 18], "126": [16, 18], "l": [16, 18, 56, 65, 73, 74, 76], "bfg": [16, 18, 65], "smooth": [16, 18, 25, 71], "165": [16, 18], "173": [16, 18], "131": [16, 18], "float32": [16, 18], "grouplasso": [17, 18], "228": [17, 18], "tutori": [17, 18, 68, 69, 71, 73, 74, 76], "221": [17, 18], "compat": [17, 18, 68, 69], "137": [17, 18], "258": [17, 18], "mcpregress": [17, 18, 24, 29], "184": [17, 18], "move": [17, 18], "specif": [17, 18, 69, 71], "method": [17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72, 75], "eas": [17, 18], "192": [17, 18], "logsumpenalti": [17, 18], "127": [17, 18], "abstract": [17, 18], "basedatafit": [17, 18, 22, 40, 71], "basepenalti": [17, 18, 22, 40, 72], "make": [17, 18, 19, 68, 69], "easier": [17, 18], "205": [17, 18], "fix": [17, 18, 19, 41], "point": [17, 18, 41, 76], "build": [17, 18, 19, 21, 23, 24, 25, 26, 27, 29, 60, 67, 68], "138": [17, 18], "continu": 19, "endeavour": 19, "reli": 19, "commun": 19, "effort": 19, "evolv": 19, "welcom": 19, "highli": [19, 69], "valuabl": 19, "you": [19, 68, 69, 71, 72], "help": 19, "bug": 19, "report": 19, "unit": [19, 73, 76], "codebas": [19, 72], "prevent": 19, "tighten": 19, "ani": [19, 68, 69, 71, 72, 76], "encount": 19, "issu": 19, "section": [19, 68, 69, 73, 74], "request": 19, "constantli": 19, "improv": 19, "would": [19, 68, 74], "like": [19, 30, 68, 71, 74], "align": [19, 73], "user": [19, 69, 72], "appreci": 19, "suggest": 19, "pull": 19, "mai": [19, 27, 75], "typo": 19, "document": [19, 65, 71, 72], "submit": 19, "integr": 19, "chang": 19, "reach": [19, 21, 22, 23, 24, 25, 26, 27, 28, 29, 76], "shortli": 19, "If": [19, 22, 23, 26, 27, 29, 30, 40, 41, 62, 63, 67, 68, 72, 76], "refer": [19, 20, 30, 40, 41, 56, 57, 60, 61, 64, 65, 67], "guid": 19, "part": [19, 23, 26, 29], "contrib": 19, "organ": 19, "adopt": 19, "conduct": 19, "willing": 19, "below": 19, "instal": [19, 68], "develop": 19, "version": [19, 69], "here": [19, 68, 71, 74], "step": [19, 41, 69], "want": 19, "fork": 19, "repositori": 19, "command": 19, "clone": 19, "sure": [19, 68], "your_github_usernam": 19, "usernam": 19, "git": 19, "http": [19, 40, 41, 56, 60, 61, 64, 67, 76], "cd": [19, 21, 23, 24, 25, 26, 27, 29, 41, 60], "directori": 19, "edit": 19, "pip": [19, 68, 69], "e": [19, 23, 25, 29, 56, 72, 73, 74], "doc": 19, "without": 19, "noplot": 19, "efron": [20, 30], "0001": [20, 21, 23, 24, 25, 26, 27, 28, 29, 43, 60, 61, 62, 63, 64, 65, 67], "breslow": 20, "mathemat": [20, 30], "behind": [20, 30], "detail": [20, 26, 30, 57, 65, 72, 73], "about": [20, 68, 74, 77], "express": 20, "convent": 20, "design": [20, 68, 69, 71, 75], "event": [20, 74], "censor": [20, 74], "lbfg": 20, "otherwis": [20, 21, 23, 24, 25, 26, 27, 28, 29, 62, 74, 76], "paramet": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 42, 43, 57, 65, 71, 72, 73], "float": [20, 21, 23, 24, 25, 26, 27, 28, 29, 32, 41, 42, 43, 57, 60, 61, 62, 63, 64, 65, 67, 71], "option": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 57, 60, 67, 71, 72], "must": [20, 25, 28, 42, 71, 72, 76], "strictli": [20, 25, 76], "default": [20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 61, 62, 63, 64, 65, 67, 72], "mix": [20, 21], "criterion": [20, 21, 23, 24, 25, 26, 27, 28, 29, 72, 75], "maximum": [20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "bool": [20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 57, 60, 61, 62, 63, 64, 65, 67], "amount": [20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "attribut": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 34, 36, 37, 39, 40, 56, 57, 61, 62, 63, 71, 72], "stop_crit_": 20, "converg": [20, 43, 61, 62, 63, 64, 65, 67, 75], "__init__": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72], "max_epoch": [21, 23, 24, 25, 26, 27, 28, 29, 41, 60, 63, 66], "50000": [21, 23, 24, 25, 26, 27, 29, 60, 66], "p0": [21, 23, 24, 25, 26, 27, 29, 41, 43, 60, 63, 64, 66, 67], "xx": [21, 23, 24, 25, 26, 27, 29, 36, 37, 39, 57, 59, 62], "n_": [21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 39, 45, 46, 52, 53, 55, 57, 59, 62, 71], "xw": [21, 24, 26, 27, 29, 31, 32, 33, 34, 35, 36, 37, 39, 41, 43, 44, 62, 71], "_2": [21, 23, 24, 25, 26, 27, 29, 36, 37, 38, 43, 44, 50, 73, 75, 76], "tt": 21, "_1": [21, 24, 28, 29, 43, 72, 73], "subproblem": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 64, 67], "definit": [21, 23, 24, 25, 26, 27, 28, 29, 60, 71, 74], "epoch": [21, 23, 24, 25, 26, 27, 29, 41, 60, 62, 63], "each": [21, 23, 24, 25, 26, 27, 28, 29, 34, 39, 41, 43, 60, 64, 67, 74, 75], "forc": [21, 23, 24, 26, 29, 57], "whether": [21, 23, 24, 25, 26, 27, 28, 29, 60], "reus": [21, 23, 24, 25, 26, 27, 28, 29], "call": [21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40], "just": [21, 23, 24, 25, 26, 27, 28, 29, 68, 73], "eras": [21, 23, 24, 25, 26, 27, 28, 29], "str": [21, 23, 24, 25, 26, 27, 29], "cost": [21, 22, 23, 24, 25, 26, 27, 29, 40, 74], "formula": [21, 22, 23, 24, 25, 26, 27, 29, 40, 74, 76], "sparse_coef_": [21, 22, 24, 25, 26, 27, 29], "scipi": [21, 22, 24, 25, 26, 27, 29, 65], "readonli": [21, 22, 24, 25, 26, 27, 29], "deriv": [21, 22, 24, 25, 26, 27, 29, 57, 71, 74, 75, 76], "intercept_": [21, 22, 23, 24, 25, 26, 27, 28, 29], "constant": [21, 22, 23, 24, 25, 26, 27, 28, 29, 34, 75], "term": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 74], "decis": [21, 22, 23, 24, 25, 26, 27, 28, 29], "n_iter_": [21, 22, 23, 24, 25, 26, 27, 28, 29], "specifi": [21, 22, 23, 24, 25, 26, 27, 28, 29, 71, 72], "toler": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "linear": [22, 26, 61, 64, 67, 68, 69, 71, 74, 75], "classif": [22, 38], "task": [22, 38, 66], "instanc": [22, 40, 69, 71, 72, 74], "jit": [22, 31, 33, 35, 38, 40, 71, 72], "basesolv": [22, 40], "n_task": 22, "celer": [23, 24, 27, 29, 64, 67], "extrapol": [23, 24, 29, 62, 64, 67], "sum_g": 23, "w_": [23, 57], "g": [23, 41, 57, 60, 69, 72, 73], "th": [23, 41, 45, 57, 71, 72], "partit": 23, "contigu": [23, 34, 37, 57], "being": [23, 57, 69, 72], "contain": [23, 56, 68], "equal": [23, 26, 29, 36, 39, 74, 75], "50_000": 23, "defautl": 23, "note": [23, 29, 31, 33, 35, 38, 45, 46, 53, 55, 59, 71, 72, 75], "sparser": [24, 29], "hing": 25, "sum_": [25, 26, 28, 31, 32, 33, 34, 35, 39, 45, 52, 53, 55, 57, 59, 71, 72, 74, 75, 76], "y_i": [25, 28, 31, 32, 33, 34, 35, 39, 42, 71, 74, 75], "beta": [25, 71, 72, 74, 75], "stai": 25, "our": [25, 71], "framework": 25, "svc": [25, 38], "yx": [25, 38], "w_i": 25, "relat": 25, "given": [25, 28, 68, 72], "invers": [25, 61], "proport": 25, "dual_": 25, "w_j": [26, 41, 45, 52, 53, 55, 59, 76], "see": 26, "algorithm": [26, 41, 61, 64, 67, 74], "nonconvex": 26, "applic": 26, "biolog": 26, "breheni": 26, "huang": 26, "hard": 26, "threshold": [26, 32, 61, 72], "soft": [26, 72], "should": [26, 62, 71, 72, 75], "larger": 26, "copy_x": 27, "multipli": 27, "l21": 27, "copi": 27, "els": [27, 72], "overwritten": 27, "log": [28, 31, 33, 34, 52, 74, 75], "exp": [28, 31, 33, 34, 35, 71, 75], "x_i": [28, 74], "outer": [28, 43, 64, 67], "classes_": 28, "ndarrai": 28, "n_class": 28, "A": [28, 38, 41, 60, 61, 62, 64, 65, 67, 69, 73, 74, 76], "known": [28, 72], "Not": 28, "yet": 28, "unweight": 29, "t_indic": 30, "stack": [30, 34, 37, 57], "horizont": 30, "group_1": 30, "group_2": 30, "ascend": 30, "order": [30, 74], "initialize_spars": [30, 71], "t_indptr": 30, "element": [30, 34, 37, 57, 72, 74], "delimit": [30, 34, 37, 57], "h_indic": 30, "uncensor": [30, 74], "h_indptr": 30, "read": [31, 32, 33, 34, 35, 36, 37, 38, 39, 42, 44, 50, 52, 57, 71, 72, 74], "_i": [31, 32, 33, 34, 35, 39, 71, 72], "faster": [31, 33, 35, 38, 41, 60, 69, 71], "delta": [32, 75], "hyperparamet": [32, 72], "grp1_indic": [34, 37, 57], "grp2_indic": [34, 37, 57], "pointer": [34, 37, 57], "lipschitz": [34, 75], "xty": 36, "pre": [36, 39, 71], "quantiti": [36, 39, 71, 76], "dure": [36, 39], "gradient": [36, 39, 62, 71, 72, 73, 75], "evalu": [36, 39, 71, 74], "sample_weight": 39, "weights_i": 39, "xtwy": 39, "samples_weight": 39, "single_task": 40, "n_reweight": 40, "surrog": 40, "involv": [40, 74], "cand\u00e8": 40, "et": 40, "al": 40, "2007": [40, 74], "enhanc": 40, "sparsiti": [40, 73], "minim": [40, 62, 73, 76], "web": 40, "stanford": 40, "edu": 40, "boyd": 40, "paper": 40, "pdf": 40, "rwl1": 40, "loss_history_": 40, "histori": 40, "everi": [40, 56], "dual_init": 41, "06": [41, 66], "min_w": 41, "saddl": 41, "max_z": 41, "z": [41, 71, 75], "fenchel": [41, 76], "conjug": [41, 76], "requir": [41, 71, 72], "built": 41, "assign": 41, "vari": 41, "updat": [41, 62, 75], "_j": [41, 59], "tau_j": 41, "g_j": [41, 75], "x_j": [41, 74, 76], "associ": [41, 74, 75], "equival": [41, 56], "variabl": [41, 56, 76], "thei": [41, 71], "silent": [41, 43, 60, 61, 62, 63, 64, 65, 67], "olivi": [41, 73], "fercoq": [41, 73], "pascal": 41, "bianchi": 41, "possibli": 41, "nonsepar": 41, "siam": [41, 61], "journal": [41, 74], "2020": [41, 64, 67], "epub": [41, 61], "doi": [41, 56, 61], "1137": [41, 61], "18m1168480": 41, "badr": 41, "moufad": 41, "q": [41, 60, 62, 69], "p": [41, 60, 69, 71, 72, 74, 75, 76], "gidel": [41, 60, 69], "beyond": [41, 60, 69], "better": [41, 60, 69], "neurip": [41, 69], "2022": [41, 60, 69], "arxiv": [41, 60, 64, 67], "2204": [41, 60], "07826": [41, 60], "quantile_level": 42, "sum_i": 42, "xw_i": 42, "quantil": 42, "level": [42, 56, 72], "becom": 42, "least": [42, 72, 73], "absolut": [42, 45, 55, 75], "deviat": [42, 45, 55], "lad": 42, "max_pn_it": [43, 64, 67], "minimum": [43, 63, 64, 67], "includ": [43, 63, 64, 67], "unnorm": 44, "smoothli": [45, 55], "clip": [45, 55], "With": [45, 53, 55, 59, 69, 71], "row": [45, 74], "math": [45, 71, 72, 73, 76], "box": 46, "bb": [46, 72, 75], "quasi": [47, 51], "ell_2": [49, 50], "aka": 49, "ep": 52, "epsilon": 52, "sort": [56, 74], "One": 56, "bogdan": 56, "van": 56, "den": 56, "berg": 56, "sabatti": 56, "su": 56, "cand": 56, "via": [56, 69], "annal": 56, "appli": [56, 71, 72, 74], "statist": [56, 74], "1103": 56, "1214": 56, "aoas842": 56, "singl": 56, "_g": 57, "i_": [57, 76], "geq": [57, 73, 74, 76], "orthant": 57, "subdifferenti": [57, 72, 73], "resolut": 60, "restrict": [60, 76], "aistat": 60, "2021": 60, "proceed": [60, 64, 67], "mlr": [60, 64, 67], "press": [60, 64, 67], "v130": 60, "bertrand21a": 60, "mathurinm": [60, 64, 67], "opt_strategi": 61, "ista": 61, "nesterov": 61, "beck": 61, "teboul": 61, "shrinkag": 61, "2009": 61, "imag": 61, "sci": 61, "080716542": 61, "use_acc": [62, 66], "greedy_cd": 62, "date": 62, "come": 62, "suit": 62, "rewritten": [62, 74], "w_init": [62, 63], "instead": [62, 63, 74], "past": 62, "greedi": 62, "cyclic": 62, "xw_init": 63, "vaiter": [64, 67], "gramfort": [64, 67, 73], "salmon": [64, 67, 73], "jmlr": [64, 67], "1907": [64, 67], "05830": [64, 67], "johnson": [64, 67], "guestrin": [64, 67], "blitz": [64, 67], "principl": [64, 67], "meta": [64, 67], "icml": [64, 67], "2015": [64, 67], "v37": [64, 67], "johnson15": [64, 67], "tbjohn": [64, 67], "blitzl1": [64, 67], "wrapper": 65, "multi": 66, "provid": [68, 74], "starter": 68, "familiar": 68, "explor": [68, 77], "high": [68, 69], "dimension": [68, 75], "toi": 68, "greater": 68, "don": 68, "unseen": 68, "By": [68, 74, 76], "overcom": [68, 74], "drawback": 68, "creat": 68, "custom": [68, 69], "sequel": 68, "perfectli": 68, "experi": 68, "thank": [68, 76], "beforehand": 68, "alreadi": [68, 69], "pypi": [68, 69], "conda": [68, 69], "forg": [68, 69], "make_regress": 68, "Then": 68, "r\u00b2": 68, "subsequ": 68, "top": [68, 71, 74, 75], "sinc": [68, 72, 75, 76], "sever": [68, 69], "api": [68, 69], "inform": 68, "avail": [68, 69], "suppos": 68, "latter": [68, 74], "outlier": 68, "mitig": 68, "effect": 68, "ideal": [68, 74], "regressor": 68, "offer": [68, 69], "flexibl": [68, 69], "compos": [68, 69, 75], "simpl": [68, 69], "possibl": 68, "huberregressor": 68, "befor": [68, 71, 76], "wai": 68, "glm": [68, 69], "argument": 68, "arbitrari": [68, 71, 72], "sake": 68, "modul": 68, "appropri": 68, "your": [68, 69, 72], "own": [68, 69, 72], "further": 68, "advanc": [68, 69], "topic": [68, 69], "hand": [68, 69, 74, 76, 77], "wide": 69, "choos": [69, 73, 77], "glimps": 69, "conceiv": 69, "There": 69, "opt": 69, "among": 69, "abl": 69, "tackl": 69, "dens": 69, "million": 69, "friendli": 69, "enabl": [69, 74], "exist": 69, "extens": 69, "easi": 69, "matter": 69, "line": 69, "drop": 69, "latest": 69, "cover": 69, "persever": 69, "research": 69, "licens": 69, "under": [69, 73], "bsd": 69, "claus": 69, "free": 69, "inproceed": 69, "booktitl": 69, "year": 69, "motiv": 71, "limit": [71, 74], "form": [71, 74], "hat": 71, "arg": 71, "min_": [71, 76], "mathbb": [71, 72, 73, 74], "omega": 71, "n": [71, 73, 74, 75, 76], "f_i": 71, "omega_j": 71, "beta_j": [71, 72, 75], "denot": [71, 75, 76], "whose": 71, "explicitli": 71, "mydatafit": 71, "mypenalti": 71, "jitclass": [71, 72], "inherit": [71, 72], "get_spec": [71, 72], "self": [71, 72], "type": [71, 72], "spec": [71, 72], "tupl": [71, 72], "attribute_nam": [71, 72], "params_to_dict": [71, 72], "dict_of_param": [71, 72], "dict": [71, 72], "instanti": [71, 72], "target": [71, 72], "x_data": 71, "x_indptr": 71, "x_indic": 71, "n_element": 71, "csc": 71, "indptr": 71, "array_lik": 71, "These": 71, "overload": 71, "suffix": 71, "_spars": 71, "hessian": 71, "frac": [71, 73, 74, 76], "simplifi": 71, "z_i": 71, "qquad": 71, "nabla": [71, 73, 74, 75], "underbrac": [71, 75], "raw": 71, "grad": [71, 72], "besid": [71, 72], "directli": [71, 75], "leq": [71, 73, 76], "diag": [71, 74], "y_iz_i": 71, "right": [71, 74], "therefor": [71, 74, 76], "raw_grad": 71, "raw_hessian": 71, "But": 71, "gradient_scalar": 71, "respect": [71, 74, 75, 76], "partial": [71, 72, 73, 76], "give": [71, 72, 75], "rais": 71, "valueerror": 71, "full_grad_spars": 71, "gradient_scalar_spars": 71, "idx_i": 71, "intercept_update_step": [71, 75], "usual": [71, 75], "serv": 71, "effici": [71, 74], "subclass": 72, "is_pen": 72, "mask": 72, "generalized_support": 72, "Theses": 72, "control": 72, "constructor": 72, "strongli": 72, "thu": [72, 73, 76], "bool_": 72, "addition": 72, "helper": 72, "dictionari": 72, "st": 72, "sgn": 72, "expect": 72, "stepsiz": 72, "turn": 72, "condit": [72, 73, 75, 76], "subdiff_dist": 72, "neg": [72, 74, 76], "nabla_j": 72, "typic": [72, 74], "might": 72, "smallest": 72, "null": 72, "reserv": 72, "keyword": 72, "correspond": 72, "togeth": [72, 76], "elif": 72, "grad_j": 72, "infti": [72, 73, 76], "gradient0": [72, 73], "frequent": 73, "fraction": 73, "consid": [73, 74], "min_x": 73, "d": [73, 76], "differenti": 73, "aim": 73, "determin": 73, "fermat": 73, "rule": [73, 74, 75], "infin": 73, "ball": [73, 76], "cdot": [73, 76], "begin": 73, "equat": [73, 74, 75, 76, 77], "leftrightarrow": 73, "ordinari": 73, "2n": [73, 75], "At": [73, 76], "proposit": 73, "particular": [73, 74, 77], "eugen": 73, "ndiay": 73, "alexandr": 73, "joseph": 73, "2017": 73, "gap": 73, "safe": 73, "screen": 73, "enforc": 73, "mach": 73, "re": 73, "januari": 73, "4671": 73, "4703": 73, "present": [74, 76], "mathbf": 74, "occur": 74, "interest": 74, "establish": 74, "likelihood": 74, "s_i": 74, "langl": [74, 76], "rangl": [74, 76], "y_j": 74, "compact": 74, "introduc": [74, 76], "wherea": 74, "write": [74, 75, 76], "wise": 74, "reformul": 74, "depend": 74, "henc": [74, 75], "focu": 74, "chain": 74, "divis": 74, "costli": [74, 75], "becaus": 74, "side": 74, "mathcal": [74, 76], "diagon": 74, "upper": 74, "bound": 74, "construct": 74, "semi": 74, "preceq": 74, "inequ": [74, 75], "loewner": 74, "matric": 74, "refin": 74, "ident": 74, "i_1": 74, "ldot": [74, 76], "i_m": 74, "h_": 74, "i_l": 74, "stand": 74, "cardin": 74, "rewrit": 74, "leverag": 74, "closer": [74, 76], "look": 74, "reveal": 74, "what": 74, "distinguish": 74, "presenc": 74, "doubl": 74, "cup_": 74, "fuse": 74, "minu": 74, "within": [74, 75], "a_": 74, "a_i": 74, "chosen": 74, "accordingli": 74, "preform": 74, "procedur": 74, "v": [74, 76], "similar": [74, 76], "subtract": 74, "complex": 74, "despit": 74, "multipl": 74, "due": 74, "special": 74, "structur": 74, "dy": 74, "lin": 74, "lifetim": 74, "471": 74, "480": 74, "bradlei": 74, "american": 74, "72": 74, "359": 74, "557": 74, "565": 74, "1977": 74, "insight": 75, "guidanc": 75, "rr": 75, "beta_0": 75, "liblinear": 75, "extra": 75, "memori": 75, "rout": 75, "star": [75, 76], "underset": 75, "phi": 75, "triangleq": 75, "_n": 75, "k": 75, "l_0": 75, "nabla_": 75, "local": 75, "statisfi": 75, "foral": [75, 76], "x_0": 75, "h": [75, 76], "moreov": 75, "nabla_beta": 75, "three": 75, "eq": 75, "4n": 75, "f_": 75, "f_delta": 75, "r_i": 75, "bbb": 75, "classic": 75, "l_2": 76, "nonneg": [76, 77], "mapsto": 76, "legendr": 76, "proj": 76, "moreau": 76, "decomposit": 76, "subdiff_diff": 76, "compon": 76, "quad": 76, "iota_": 76, "over": 76, "intuit": 76, "clear": 76, "v_i": 76, "cancel": 76, "n_i": 76, "u_i": 76, "bring": 76, "rigor": 76, "lagrangian": 76, "nu": 76, "mu": 76, "down": 76, "treat": 76, "complementari": 76, "slack": 76, "conclus": 76, "c_1": 76, "c_g": 76, "c_j": 76, "project": 76, "onto": 76, "p_j": 76, "v_j": 76, "stackexchang": 76, "2887332": 76, "167258": 76, "poisson": 77}, "objects": {"skglm": [[20, 0, 1, "", "CoxEstimator"], [21, 0, 1, "", "ElasticNet"], [22, 0, 1, "", "GeneralizedLinearEstimator"], [23, 0, 1, "", "GroupLasso"], [24, 0, 1, "", "Lasso"], [25, 0, 1, "", "LinearSVC"], [26, 0, 1, "", "MCPRegression"], [27, 0, 1, "", "MultiTaskLasso"], [28, 0, 1, "", "SparseLogisticRegression"], [29, 0, 1, "", "WeightedLasso"]], "skglm.CoxEstimator": [[20, 1, 1, "", "__init__"]], "skglm.ElasticNet": [[21, 1, 1, "", "__init__"]], "skglm.GeneralizedLinearEstimator": [[22, 1, 1, "", "__init__"]], "skglm.GroupLasso": [[23, 1, 1, "", "__init__"]], "skglm.Lasso": [[24, 1, 1, "", "__init__"]], "skglm.LinearSVC": [[25, 1, 1, "", "__init__"]], "skglm.MCPRegression": [[26, 1, 1, "", "__init__"]], "skglm.MultiTaskLasso": [[27, 1, 1, "", "__init__"]], "skglm.SparseLogisticRegression": [[28, 1, 1, "", "__init__"]], "skglm.WeightedLasso": [[29, 1, 1, "", "__init__"]], "skglm.datafits": [[30, 0, 1, "", "Cox"], [31, 0, 1, "", "Gamma"], [32, 0, 1, "", "Huber"], [33, 0, 1, "", "Logistic"], [34, 0, 1, "", "LogisticGroup"], [35, 0, 1, "", "Poisson"], [36, 0, 1, "", "Quadratic"], [37, 0, 1, "", "QuadraticGroup"], [38, 0, 1, "", "QuadraticSVC"], [39, 0, 1, "", "WeightedQuadratic"]], "skglm.datafits.Cox": [[30, 1, 1, "", "__init__"]], "skglm.datafits.Gamma": [[31, 1, 1, "", "__init__"]], "skglm.datafits.Huber": [[32, 1, 1, "", "__init__"]], "skglm.datafits.Logistic": [[33, 1, 1, "", "__init__"]], "skglm.datafits.LogisticGroup": [[34, 1, 1, "", "__init__"]], "skglm.datafits.Poisson": [[35, 1, 1, "", "__init__"]], "skglm.datafits.Quadratic": [[36, 1, 1, "", "__init__"]], "skglm.datafits.QuadraticGroup": [[37, 1, 1, "", "__init__"]], "skglm.datafits.QuadraticSVC": [[38, 1, 1, "", "__init__"]], "skglm.datafits.WeightedQuadratic": [[39, 1, 1, "", "__init__"]], "skglm.experimental": [[40, 0, 1, "", "IterativeReweightedL1"], [41, 0, 1, "", "PDCD_WS"], [42, 0, 1, "", "Pinball"], [43, 0, 1, "", "SqrtLasso"], [44, 0, 1, "", "SqrtQuadratic"]], "skglm.experimental.IterativeReweightedL1": [[40, 1, 1, "", "__init__"]], "skglm.experimental.PDCD_WS": [[41, 1, 1, "", "__init__"]], "skglm.experimental.Pinball": [[42, 1, 1, "", "__init__"]], "skglm.experimental.SqrtLasso": [[43, 1, 1, "", "__init__"]], "skglm.experimental.SqrtQuadratic": [[44, 1, 1, "", "__init__"]], "skglm.penalties": [[45, 0, 1, "", "BlockSCAD"], [46, 0, 1, "", "IndicatorBox"], [47, 0, 1, "", "L0_5"], [48, 0, 1, "", "L1"], [49, 0, 1, "", "L1_plus_L2"], [50, 0, 1, "", "L2"], [51, 0, 1, "", "L2_3"], [52, 0, 1, "", "LogSumPenalty"], [53, 0, 1, "", "MCPenalty"], [54, 0, 1, "", "PositiveConstraint"], [55, 0, 1, "", "SCAD"], [56, 0, 1, "", "SLOPE"], [57, 0, 1, "", "WeightedGroupL2"], [58, 0, 1, "", "WeightedL1"], [59, 0, 1, "", "WeightedMCPenalty"]], "skglm.penalties.BlockSCAD": [[45, 1, 1, "", "__init__"]], "skglm.penalties.IndicatorBox": [[46, 1, 1, "", "__init__"]], "skglm.penalties.L0_5": [[47, 1, 1, "", "__init__"]], "skglm.penalties.L1": [[48, 1, 1, "", "__init__"]], "skglm.penalties.L1_plus_L2": [[49, 1, 1, "", "__init__"]], "skglm.penalties.L2": [[50, 1, 1, "", "__init__"]], "skglm.penalties.L2_3": [[51, 1, 1, "", "__init__"]], "skglm.penalties.LogSumPenalty": [[52, 1, 1, "", "__init__"]], "skglm.penalties.MCPenalty": [[53, 1, 1, "", "__init__"]], "skglm.penalties.PositiveConstraint": [[54, 1, 1, "", "__init__"]], "skglm.penalties.SCAD": [[55, 1, 1, "", "__init__"]], "skglm.penalties.SLOPE": [[56, 1, 1, "", "__init__"]], "skglm.penalties.WeightedGroupL2": [[57, 1, 1, "", "__init__"]], "skglm.penalties.WeightedL1": [[58, 1, 1, "", "__init__"]], "skglm.penalties.WeightedMCPenalty": [[59, 1, 1, "", "__init__"]], "skglm.solvers": [[60, 0, 1, "", "AndersonCD"], [61, 0, 1, "", "FISTA"], [62, 0, 1, "", "GramCD"], [63, 0, 1, "", "GroupBCD"], [64, 0, 1, "", "GroupProxNewton"], [65, 0, 1, "", "LBFGS"], [66, 0, 1, "", "MultiTaskBCD"], [67, 0, 1, "", "ProxNewton"]], "skglm.solvers.AndersonCD": [[60, 1, 1, "", "__init__"]], "skglm.solvers.FISTA": [[61, 1, 1, "", "__init__"]], "skglm.solvers.GramCD": [[62, 1, 1, "", "__init__"]], "skglm.solvers.GroupBCD": [[63, 1, 1, "", "__init__"]], "skglm.solvers.GroupProxNewton": [[64, 1, 1, "", "__init__"]], "skglm.solvers.LBFGS": [[65, 1, 1, "", "__init__"]], "skglm.solvers.MultiTaskBCD": [[66, 1, 1, "", "__init__"]], "skglm.solvers.ProxNewton": [[67, 1, 1, "", "__init__"]]}, "objtypes": {"0": "py:class", "1": "py:method"}, "objnames": {"0": ["py", "class", "Python class"], "1": ["py", "method", "Python method"]}, "titleterms": {"api": 0, "document": 0, "estim": [0, 10, 68, 74], "penalti": [0, 5, 6, 9, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 72, 76, 77], "datafit": [0, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 71, 74, 75, 77], "solver": [0, 60, 61, 62, 63, 64, 65, 66, 67], "experiment": [0, 40, 41, 42, 43, 44], "exampl": [1, 73], "galleri": 1, "time": [2, 7, 10, 13, 70], "comparison": [2, 4, 7, 10], "scikit": 2, "learn": 2, "lasso": [2, 4, 8, 12, 24, 68, 76, 77], "group": [3, 8, 76, 77], "logist": [3, 5, 33, 75], "regress": [3, 5, 68], "python": [3, 8], "weight": [4, 12], "elast": 5, "net": 5, "minimax": 5, "concav": 5, "valu": 6, "proxim": [6, 76], "oper": [6, 76], "between": 7, "direct": 7, "prox": 7, "comput": [7, 13, 70, 75, 77], "reweight": 7, "fast": 8, "spars": [8, 9], "recoveri": 9, "non": 9, "convex": 9, "lifelin": 10, "skglm": [10, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 69], "surviv": 10, "analysi": 10, "data": [10, 68], "fit": [10, 68], "cox": [10, 30, 74, 77], "result": 10, "efron": [10, 74], "show": 11, "u": 11, "curv": 11, "regular": [11, 73, 77], "some": 12, "zero": 12, "version": [14, 15, 16, 17, 18], "0": [14, 15, 16, 17, 18, 73, 76, 77], "1": [14, 17, 18], "2022": [14, 15, 18], "04": [14, 18], "25": [14, 18], "2": [15, 18], "10": [15, 18], "18": [15, 18], "3": [16, 17, 18], "2023": [16, 17, 18], "07": [16, 18], "01": [16, 18], "4": [17, 18], "progress": [17, 18], "12": [17, 18], "21": [17, 18], "what": 18, "": 18, "new": 18, "contribut": 19, "setup": [19, 73, 74], "your": 19, "local": 19, "machin": 19, "coxestim": 20, "elasticnet": 21, "generalizedlinearestim": 22, "grouplasso": 23, "linearsvc": 25, "mcpregress": 26, "multitasklasso": 27, "sparselogisticregress": 28, "weightedlasso": 29, "gamma": 31, "huber": [32, 68, 75], "logisticgroup": 34, "poisson": [35, 71], "quadrat": [36, 75], "quadraticgroup": 37, "quadraticsvc": 38, "weightedquadrat": 39, "iterativereweightedl1": 40, "pdcd_w": 41, "pinbal": 42, "sqrtlasso": 43, "sqrtquadrat": 44, "blockscad": 45, "indicatorbox": 46, "l0_5": 47, "l1": [48, 72], "l1_plus_l2": 49, "l2": 50, "l2_3": 51, "logsumpenalti": 52, "mcpenalti": 53, "positiveconstraint": 54, "scad": 55, "slope": 56, "weightedgroupl2": 57, "weightedl1": 58, "weightedmcpenalti": 59, "andersoncd": 60, "fista": 61, "gramcd": 62, "groupbcd": 63, "groupproxnewton": 64, "lbfg": 65, "multitaskbcd": 66, "proxnewton": 67, "get": 68, "start": 68, "ell_1": 68, "gener": 68, "corrupt": 68, "why": 69, "instal": 69, "cite": 69, "how": [71, 72, 77], "add": [71, 72, 77], "custom": [71, 72, 77], "A": [71, 72], "case": [71, 72, 76], "point": [71, 72], "defin": [71, 72], "critic": [73, 77], "strength": [73, 77], "abov": [73, 77], "which": [73, 77], "solut": [73, 77], "i": [73, 77], "problem": [73, 74], "theoret": 73, "background": 73, "refer": [73, 74, 76], "mathemat": [74, 77], "behind": [74, 77], "breslow": 74, "express": 74, "gradient": 74, "hessian": 74, "intercept": [75, 77], "The": 75, "detail": [76, 77], "posit": 76, "constraint": 76, "subdifferenti": 76, "w": 76, "notin": 76, "mathbb": 76, "r": 76, "_": 76, "g": 76, "ne": 76, "tutori": 77}, "envversion": {"sphinx.domains.c": 3, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 9, "sphinx.domains.index": 1, "sphinx.domains.javascript": 3, "sphinx.domains.math": 2, "sphinx.domains.python": 4, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.intersphinx": 1, "sphinx": 58}, "alltitles": {"API Documentation": [[0, "api-documentation"]], "Estimators": [[0, "estimators"]], "Penalties": [[0, "penalties"]], "Datafits": [[0, "datafits"]], "Solvers": [[0, "solvers"]], "Experimental": [[0, "experimental"]], "Examples Gallery": [[1, "examples-gallery"]], "Timing comparison with scikit-learn for Lasso": [[2, "timing-comparison-with-scikit-learn-for-lasso"]], "Group Logistic regression in python": [[3, "group-logistic-regression-in-python"]], "Comparison of Lasso and Weighted Lasso": [[4, "comparison-of-lasso-and-weighted-lasso"]], "Logistic regression with Elastic net and minimax concave penalties": [[5, "logistic-regression-with-elastic-net-and-minimax-concave-penalties"]], "Value and proximal operators of penalties": [[6, "value-and-proximal-operators-of-penalties"]], "Timing comparison between direct prox computation and reweighting": [[7, "timing-comparison-between-direct-prox-computation-and-reweighting"]], "Fast Sparse Group Lasso in python": [[8, "fast-sparse-group-lasso-in-python"]], "Sparse recovery with non-convex penalties": [[9, "sparse-recovery-with-non-convex-penalties"]], "Comparison of lifelines with skglm for survival analysis": [[10, "comparison-of-lifelines-with-skglm-for-survival-analysis"]], "Data": [[10, "data"]], "Fitting the Cox Estimator": [[10, "fitting-the-cox-estimator"]], "Timing comparison": [[10, "timing-comparison"]], "Results": [[10, "results"]], "Efron estimate": [[10, "efron-estimate"], [74, "efron-estimate"]], "Show U-curve of regularization": [[11, "show-u-curve-of-regularization"]], "Weighted Lasso with some zero weights": [[12, "weighted-lasso-with-some-zero-weights"]], "Computation times": [[13, "computation-times"], [70, "computation-times"]], "Version 0.1 (2022/04/25)": [[14, "version-0-1-2022-04-25"], [18, "version-0-1-2022-04-25"]], "Version 0.2 (2022/10/18)": [[15, "version-0-2-2022-10-18"], [18, "version-0-2-2022-10-18"]], "Version 0.3 (2023/07/01)": [[16, "version-0-3-2023-07-01"], [18, "version-0-3-2023-07-01"]], "Version 0.4 (in progress)": [[17, "version-0-4-in-progress"], [18, "version-0-4-in-progress"]], "Version 0.3.1 (2023/12/21)": [[17, "version-0-3-1-2023-12-21"], [18, "version-0-3-1-2023-12-21"]], "What\u2019s new": [[18, "what-s-new"]], "Contribute to skglm": [[19, "contribute-to-skglm"]], "Setup skglm on your local machine": [[19, "setup-skglm-on-your-local-machine"]], "skglm.CoxEstimator": [[20, "skglm-coxestimator"]], "skglm.ElasticNet": [[21, "skglm-elasticnet"]], "skglm.GeneralizedLinearEstimator": [[22, "skglm-generalizedlinearestimator"]], "skglm.GroupLasso": [[23, "skglm-grouplasso"]], "skglm.Lasso": [[24, "skglm-lasso"]], "skglm.LinearSVC": [[25, "skglm-linearsvc"]], "skglm.MCPRegression": [[26, "skglm-mcpregression"]], "skglm.MultiTaskLasso": [[27, "skglm-multitasklasso"]], "skglm.SparseLogisticRegression": [[28, "skglm-sparselogisticregression"]], "skglm.WeightedLasso": [[29, "skglm-weightedlasso"]], "skglm.datafits.Cox": [[30, "skglm-datafits-cox"]], "skglm.datafits.Gamma": [[31, "skglm-datafits-gamma"]], "skglm.datafits.Huber": [[32, "skglm-datafits-huber"]], "skglm.datafits.Logistic": [[33, "skglm-datafits-logistic"]], "skglm.datafits.LogisticGroup": [[34, "skglm-datafits-logisticgroup"]], "skglm.datafits.Poisson": [[35, "skglm-datafits-poisson"]], "skglm.datafits.Quadratic": [[36, "skglm-datafits-quadratic"]], "skglm.datafits.QuadraticGroup": [[37, "skglm-datafits-quadraticgroup"]], "skglm.datafits.QuadraticSVC": [[38, "skglm-datafits-quadraticsvc"]], "skglm.datafits.WeightedQuadratic": [[39, "skglm-datafits-weightedquadratic"]], "skglm.experimental.IterativeReweightedL1": [[40, "skglm-experimental-iterativereweightedl1"]], "skglm.experimental.PDCD_WS": [[41, "skglm-experimental-pdcd-ws"]], "skglm.experimental.Pinball": [[42, "skglm-experimental-pinball"]], "skglm.experimental.SqrtLasso": [[43, "skglm-experimental-sqrtlasso"]], "skglm.experimental.SqrtQuadratic": [[44, "skglm-experimental-sqrtquadratic"]], "skglm.penalties.BlockSCAD": [[45, "skglm-penalties-blockscad"]], "skglm.penalties.IndicatorBox": [[46, "skglm-penalties-indicatorbox"]], "skglm.penalties.L0_5": [[47, "skglm-penalties-l0-5"]], "skglm.penalties.L1": [[48, "skglm-penalties-l1"]], "skglm.penalties.L1_plus_L2": [[49, "skglm-penalties-l1-plus-l2"]], "skglm.penalties.L2": [[50, "skglm-penalties-l2"]], "skglm.penalties.L2_3": [[51, "skglm-penalties-l2-3"]], "skglm.penalties.LogSumPenalty": [[52, "skglm-penalties-logsumpenalty"]], "skglm.penalties.MCPenalty": [[53, "skglm-penalties-mcpenalty"]], "skglm.penalties.PositiveConstraint": [[54, "skglm-penalties-positiveconstraint"]], "skglm.penalties.SCAD": [[55, "skglm-penalties-scad"]], "skglm.penalties.SLOPE": [[56, "skglm-penalties-slope"]], "skglm.penalties.WeightedGroupL2": [[57, "skglm-penalties-weightedgroupl2"]], "skglm.penalties.WeightedL1": [[58, "skglm-penalties-weightedl1"]], "skglm.penalties.WeightedMCPenalty": [[59, "skglm-penalties-weightedmcpenalty"]], "skglm.solvers.AndersonCD": [[60, "skglm-solvers-andersoncd"]], "skglm.solvers.FISTA": [[61, "skglm-solvers-fista"]], "skglm.solvers.GramCD": [[62, "skglm-solvers-gramcd"]], "skglm.solvers.GroupBCD": [[63, "skglm-solvers-groupbcd"]], "skglm.solvers.GroupProxNewton": [[64, "skglm-solvers-groupproxnewton"]], "skglm.solvers.LBFGS": [[65, "skglm-solvers-lbfgs"]], "skglm.solvers.MultiTaskBCD": [[66, "skglm-solvers-multitaskbcd"]], "skglm.solvers.ProxNewton": [[67, "skglm-solvers-proxnewton"]], "Getting started": [[68, "getting-started"]], "Fitting a Lasso estimator": [[68, "fitting-a-lasso-estimator"]], "Fitting \\ell_1 Huber regression": [[68, "fitting-ell-1-huber-regression"]], "Generate corrupt data": [[68, "generate-corrupt-data"]], "skglm": [[69, "skglm"]], "Why skglm?": [[69, "why-skglm"]], "Installing skglm": [[69, "installing-skglm"]], "Cite": [[69, "cite"]], "How to add a custom datafit": [[71, "how-to-add-a-custom-datafit"], [77, "how-to-add-a-custom-datafit"]], "A case in point: defining Poisson datafit": [[71, "a-case-in-point-defining-poisson-datafit"]], "How to add a custom penalty": [[72, "how-to-add-a-custom-penalty"], [77, "how-to-add-a-custom-penalty"]], "A case in point: defining L1 penalty": [[72, "a-case-in-point-defining-l1-penalty"]], "Critical regularization strength above which solution is 0": [[73, "critical-regularization-strength-above-which-solution-is-0"], [77, "critical-regularization-strength-above-which-solution-is-0"]], "Problem setup": [[73, "problem-setup"], [74, "problem-setup"]], "Theoretical background": [[73, "theoretical-background"]], "Example": [[73, "example"]], "References": [[73, "references"], [76, "references"]], "Mathematic behind Cox datafit": [[74, "mathematic-behind-cox-datafit"]], "Breslow estimate": [[74, "breslow-estimate"]], "Datafit expression": [[74, "datafit-expression"], [74, "id1"]], "Gradient and Hessian": [[74, "gradient-and-hessian"], [74, "id2"]], "Reference": [[74, "reference"]], "Computation of the intercept": [[75, "computation-of-the-intercept"], [77, "computation-of-the-intercept"]], "The Quadratic datafit": [[75, "the-quadratic-datafit"]], "The Logistic datafit": [[75, "the-logistic-datafit"]], "The Huber datafit": [[75, "the-huber-datafit"]], "Details on the Positive Group Lasso": [[76, "details-on-the-positive-group-lasso"]], "Proximity operator of the group Lasso": [[76, "proximity-operator-of-the-group-lasso"]], "Proximity operator of the group Lasso with positivity constraints": [[76, "proximity-operator-of-the-group-lasso-with-positivity-constraints"]], "Subdifferential of the positive Group Lasso penalty": [[76, "subdifferential-of-the-positive-group-lasso-penalty"]], "Case w \\notin \\mathbb{R}_+^g": [[76, "case-w-notin-mathbb-r-g"]], "Case w = 0": [[76, "case-w-0"]], "Case || w  || \\ne 0": [[76, "case-w-ne-0"]], "Tutorials": [[77, "tutorials"]], "Mathematics behind Cox datafit": [[77, "mathematics-behind-cox-datafit"]], "Details on the group Lasso": [[77, "details-on-the-group-lasso"]]}, "indexentries": {"coxestimator (class in skglm)": [[20, "skglm.CoxEstimator"]], "__init__() (skglm.coxestimator method)": [[20, "skglm.CoxEstimator.__init__"]], "elasticnet (class in skglm)": [[21, "skglm.ElasticNet"]], "__init__() (skglm.elasticnet method)": [[21, "skglm.ElasticNet.__init__"]], "generalizedlinearestimator (class in skglm)": [[22, "skglm.GeneralizedLinearEstimator"]], "__init__() (skglm.generalizedlinearestimator method)": [[22, "skglm.GeneralizedLinearEstimator.__init__"]], "grouplasso (class in skglm)": [[23, "skglm.GroupLasso"]], "__init__() (skglm.grouplasso method)": [[23, "skglm.GroupLasso.__init__"]], "lasso (class in skglm)": [[24, "skglm.Lasso"]], "__init__() (skglm.lasso method)": [[24, "skglm.Lasso.__init__"]], "linearsvc (class in skglm)": [[25, "skglm.LinearSVC"]], "__init__() (skglm.linearsvc method)": [[25, "skglm.LinearSVC.__init__"]], "mcpregression (class in skglm)": [[26, "skglm.MCPRegression"]], "__init__() (skglm.mcpregression method)": [[26, "skglm.MCPRegression.__init__"]], "multitasklasso (class in skglm)": [[27, "skglm.MultiTaskLasso"]], "__init__() (skglm.multitasklasso method)": [[27, "skglm.MultiTaskLasso.__init__"]], "sparselogisticregression (class in skglm)": [[28, "skglm.SparseLogisticRegression"]], "__init__() (skglm.sparselogisticregression method)": [[28, "skglm.SparseLogisticRegression.__init__"]], "weightedlasso (class in skglm)": [[29, "skglm.WeightedLasso"]], "__init__() (skglm.weightedlasso method)": [[29, "skglm.WeightedLasso.__init__"]], "cox (class in skglm.datafits)": [[30, "skglm.datafits.Cox"]], "__init__() (skglm.datafits.cox method)": [[30, "skglm.datafits.Cox.__init__"]], "gamma (class in skglm.datafits)": [[31, "skglm.datafits.Gamma"]], "__init__() (skglm.datafits.gamma method)": [[31, "skglm.datafits.Gamma.__init__"]], "huber (class in skglm.datafits)": [[32, "skglm.datafits.Huber"]], "__init__() (skglm.datafits.huber method)": [[32, "skglm.datafits.Huber.__init__"]], "logistic (class in skglm.datafits)": [[33, "skglm.datafits.Logistic"]], "__init__() (skglm.datafits.logistic method)": [[33, "skglm.datafits.Logistic.__init__"]], "logisticgroup (class in skglm.datafits)": [[34, "skglm.datafits.LogisticGroup"]], "__init__() (skglm.datafits.logisticgroup method)": [[34, "skglm.datafits.LogisticGroup.__init__"]], "poisson (class in skglm.datafits)": [[35, "skglm.datafits.Poisson"]], "__init__() (skglm.datafits.poisson method)": [[35, "skglm.datafits.Poisson.__init__"]], "quadratic (class in skglm.datafits)": [[36, "skglm.datafits.Quadratic"]], "__init__() (skglm.datafits.quadratic method)": [[36, "skglm.datafits.Quadratic.__init__"]], "quadraticgroup (class in skglm.datafits)": [[37, "skglm.datafits.QuadraticGroup"]], "__init__() (skglm.datafits.quadraticgroup method)": [[37, "skglm.datafits.QuadraticGroup.__init__"]], "quadraticsvc (class in skglm.datafits)": [[38, "skglm.datafits.QuadraticSVC"]], "__init__() (skglm.datafits.quadraticsvc method)": [[38, "skglm.datafits.QuadraticSVC.__init__"]], "weightedquadratic (class in skglm.datafits)": [[39, "skglm.datafits.WeightedQuadratic"]], "__init__() (skglm.datafits.weightedquadratic method)": [[39, "skglm.datafits.WeightedQuadratic.__init__"]], "iterativereweightedl1 (class in skglm.experimental)": [[40, "skglm.experimental.IterativeReweightedL1"]], "__init__() (skglm.experimental.iterativereweightedl1 method)": [[40, "skglm.experimental.IterativeReweightedL1.__init__"]], "pdcd_ws (class in skglm.experimental)": [[41, "skglm.experimental.PDCD_WS"]], "__init__() (skglm.experimental.pdcd_ws method)": [[41, "skglm.experimental.PDCD_WS.__init__"]], "pinball (class in skglm.experimental)": [[42, "skglm.experimental.Pinball"]], "__init__() (skglm.experimental.pinball method)": [[42, "skglm.experimental.Pinball.__init__"]], "sqrtlasso (class in skglm.experimental)": [[43, "skglm.experimental.SqrtLasso"]], "__init__() (skglm.experimental.sqrtlasso method)": [[43, "skglm.experimental.SqrtLasso.__init__"]], "sqrtquadratic (class in skglm.experimental)": [[44, "skglm.experimental.SqrtQuadratic"]], "__init__() (skglm.experimental.sqrtquadratic method)": [[44, "skglm.experimental.SqrtQuadratic.__init__"]], "blockscad (class in skglm.penalties)": [[45, "skglm.penalties.BlockSCAD"]], "__init__() (skglm.penalties.blockscad method)": [[45, "skglm.penalties.BlockSCAD.__init__"]], "indicatorbox (class in skglm.penalties)": [[46, "skglm.penalties.IndicatorBox"]], "__init__() (skglm.penalties.indicatorbox method)": [[46, "skglm.penalties.IndicatorBox.__init__"]], "l0_5 (class in skglm.penalties)": [[47, "skglm.penalties.L0_5"]], "__init__() (skglm.penalties.l0_5 method)": [[47, "skglm.penalties.L0_5.__init__"]], "l1 (class in skglm.penalties)": [[48, "skglm.penalties.L1"]], "__init__() (skglm.penalties.l1 method)": [[48, "skglm.penalties.L1.__init__"]], "l1_plus_l2 (class in skglm.penalties)": [[49, "skglm.penalties.L1_plus_L2"]], "__init__() (skglm.penalties.l1_plus_l2 method)": [[49, "skglm.penalties.L1_plus_L2.__init__"]], "l2 (class in skglm.penalties)": [[50, "skglm.penalties.L2"]], "__init__() (skglm.penalties.l2 method)": [[50, "skglm.penalties.L2.__init__"]], "l2_3 (class in skglm.penalties)": [[51, "skglm.penalties.L2_3"]], "__init__() (skglm.penalties.l2_3 method)": [[51, "skglm.penalties.L2_3.__init__"]], "logsumpenalty (class in skglm.penalties)": [[52, "skglm.penalties.LogSumPenalty"]], "__init__() (skglm.penalties.logsumpenalty method)": [[52, "skglm.penalties.LogSumPenalty.__init__"]], "mcpenalty (class in skglm.penalties)": [[53, "skglm.penalties.MCPenalty"]], "__init__() (skglm.penalties.mcpenalty method)": [[53, "skglm.penalties.MCPenalty.__init__"]], "positiveconstraint (class in skglm.penalties)": [[54, "skglm.penalties.PositiveConstraint"]], "__init__() (skglm.penalties.positiveconstraint method)": [[54, "skglm.penalties.PositiveConstraint.__init__"]], "scad (class in skglm.penalties)": [[55, "skglm.penalties.SCAD"]], "__init__() (skglm.penalties.scad method)": [[55, "skglm.penalties.SCAD.__init__"]], "slope (class in skglm.penalties)": [[56, "skglm.penalties.SLOPE"]], "__init__() (skglm.penalties.slope method)": [[56, "skglm.penalties.SLOPE.__init__"]], "weightedgroupl2 (class in skglm.penalties)": [[57, "skglm.penalties.WeightedGroupL2"]], "__init__() (skglm.penalties.weightedgroupl2 method)": [[57, "skglm.penalties.WeightedGroupL2.__init__"]], "weightedl1 (class in skglm.penalties)": [[58, "skglm.penalties.WeightedL1"]], "__init__() (skglm.penalties.weightedl1 method)": [[58, "skglm.penalties.WeightedL1.__init__"]], "weightedmcpenalty (class in skglm.penalties)": [[59, "skglm.penalties.WeightedMCPenalty"]], "__init__() (skglm.penalties.weightedmcpenalty method)": [[59, "skglm.penalties.WeightedMCPenalty.__init__"]], "andersoncd (class in skglm.solvers)": [[60, "skglm.solvers.AndersonCD"]], "__init__() (skglm.solvers.andersoncd method)": [[60, "skglm.solvers.AndersonCD.__init__"]], "fista (class in skglm.solvers)": [[61, "skglm.solvers.FISTA"]], "__init__() (skglm.solvers.fista method)": [[61, "skglm.solvers.FISTA.__init__"]], "gramcd (class in skglm.solvers)": [[62, "skglm.solvers.GramCD"]], "__init__() (skglm.solvers.gramcd method)": [[62, "skglm.solvers.GramCD.__init__"]], "groupbcd (class in skglm.solvers)": [[63, "skglm.solvers.GroupBCD"]], "__init__() (skglm.solvers.groupbcd method)": [[63, "skglm.solvers.GroupBCD.__init__"]], "groupproxnewton (class in skglm.solvers)": [[64, "skglm.solvers.GroupProxNewton"]], "__init__() (skglm.solvers.groupproxnewton method)": [[64, "skglm.solvers.GroupProxNewton.__init__"]], "lbfgs (class in skglm.solvers)": [[65, "skglm.solvers.LBFGS"]], "__init__() (skglm.solvers.lbfgs method)": [[65, "skglm.solvers.LBFGS.__init__"]], "multitaskbcd (class in skglm.solvers)": [[66, "skglm.solvers.MultiTaskBCD"]], "__init__() (skglm.solvers.multitaskbcd method)": [[66, "skglm.solvers.MultiTaskBCD.__init__"]], "proxnewton (class in skglm.solvers)": [[67, "skglm.solvers.ProxNewton"]], "__init__() (skglm.solvers.proxnewton method)": [[67, "skglm.solvers.ProxNewton.__init__"]]}})
\ No newline at end of file
+Search.setIndex({"docnames": ["api", "auto_examples/index", "auto_examples/plot_compare_time", "auto_examples/plot_group_logistic_regression", "auto_examples/plot_lasso_vs_weighted", "auto_examples/plot_logreg_various_penalties", "auto_examples/plot_pen_prox", "auto_examples/plot_reweighted_l1", "auto_examples/plot_sparse_group_lasso", "auto_examples/plot_sparse_recovery", "auto_examples/plot_survival_analysis", "auto_examples/plot_ucurve", "auto_examples/plot_zero_weights_lasso", "auto_examples/sg_execution_times", "changes/0.1", "changes/0.2", "changes/0.3", "changes/0.4", "changes/whats_new", "contribute", "generated/skglm.CoxEstimator", "generated/skglm.ElasticNet", "generated/skglm.GeneralizedLinearEstimator", "generated/skglm.GroupLasso", "generated/skglm.Lasso", "generated/skglm.LinearSVC", "generated/skglm.MCPRegression", "generated/skglm.MultiTaskLasso", "generated/skglm.SparseLogisticRegression", "generated/skglm.WeightedLasso", "generated/skglm.datafits.Cox", "generated/skglm.datafits.Gamma", "generated/skglm.datafits.Huber", "generated/skglm.datafits.Logistic", "generated/skglm.datafits.LogisticGroup", "generated/skglm.datafits.Poisson", "generated/skglm.datafits.Quadratic", "generated/skglm.datafits.QuadraticGroup", "generated/skglm.datafits.QuadraticSVC", "generated/skglm.datafits.WeightedQuadratic", "generated/skglm.experimental.IterativeReweightedL1", "generated/skglm.experimental.PDCD_WS", "generated/skglm.experimental.Pinball", "generated/skglm.experimental.SqrtLasso", "generated/skglm.experimental.SqrtQuadratic", "generated/skglm.penalties.BlockSCAD", "generated/skglm.penalties.IndicatorBox", "generated/skglm.penalties.L0_5", "generated/skglm.penalties.L1", "generated/skglm.penalties.L1_plus_L2", "generated/skglm.penalties.L2", "generated/skglm.penalties.L2_3", "generated/skglm.penalties.LogSumPenalty", "generated/skglm.penalties.MCPenalty", "generated/skglm.penalties.PositiveConstraint", "generated/skglm.penalties.SCAD", "generated/skglm.penalties.SLOPE", "generated/skglm.penalties.WeightedGroupL2", "generated/skglm.penalties.WeightedL1", "generated/skglm.penalties.WeightedMCPenalty", "generated/skglm.solvers.AndersonCD", "generated/skglm.solvers.FISTA", "generated/skglm.solvers.GramCD", "generated/skglm.solvers.GroupBCD", "generated/skglm.solvers.GroupProxNewton", "generated/skglm.solvers.LBFGS", "generated/skglm.solvers.MultiTaskBCD", "generated/skglm.solvers.ProxNewton", "getting_started", "index", "sg_execution_times", "tutorials/add_datafit", "tutorials/add_penalty", "tutorials/alpha_max", "tutorials/cox_datafit", "tutorials/intercept", "tutorials/prox_nn_group_lasso", "tutorials/tutorials"], "filenames": ["api.rst", "auto_examples/index.rst", "auto_examples/plot_compare_time.rst", "auto_examples/plot_group_logistic_regression.rst", "auto_examples/plot_lasso_vs_weighted.rst", "auto_examples/plot_logreg_various_penalties.rst", "auto_examples/plot_pen_prox.rst", "auto_examples/plot_reweighted_l1.rst", "auto_examples/plot_sparse_group_lasso.rst", "auto_examples/plot_sparse_recovery.rst", "auto_examples/plot_survival_analysis.rst", "auto_examples/plot_ucurve.rst", "auto_examples/plot_zero_weights_lasso.rst", "auto_examples/sg_execution_times.rst", "changes/0.1.rst", "changes/0.2.rst", "changes/0.3.rst", "changes/0.4.rst", "changes/whats_new.rst", "contribute.rst", "generated/skglm.CoxEstimator.rst", "generated/skglm.ElasticNet.rst", "generated/skglm.GeneralizedLinearEstimator.rst", "generated/skglm.GroupLasso.rst", "generated/skglm.Lasso.rst", "generated/skglm.LinearSVC.rst", "generated/skglm.MCPRegression.rst", "generated/skglm.MultiTaskLasso.rst", "generated/skglm.SparseLogisticRegression.rst", "generated/skglm.WeightedLasso.rst", "generated/skglm.datafits.Cox.rst", "generated/skglm.datafits.Gamma.rst", "generated/skglm.datafits.Huber.rst", "generated/skglm.datafits.Logistic.rst", "generated/skglm.datafits.LogisticGroup.rst", "generated/skglm.datafits.Poisson.rst", "generated/skglm.datafits.Quadratic.rst", "generated/skglm.datafits.QuadraticGroup.rst", "generated/skglm.datafits.QuadraticSVC.rst", "generated/skglm.datafits.WeightedQuadratic.rst", "generated/skglm.experimental.IterativeReweightedL1.rst", "generated/skglm.experimental.PDCD_WS.rst", "generated/skglm.experimental.Pinball.rst", "generated/skglm.experimental.SqrtLasso.rst", "generated/skglm.experimental.SqrtQuadratic.rst", "generated/skglm.penalties.BlockSCAD.rst", "generated/skglm.penalties.IndicatorBox.rst", "generated/skglm.penalties.L0_5.rst", "generated/skglm.penalties.L1.rst", "generated/skglm.penalties.L1_plus_L2.rst", "generated/skglm.penalties.L2.rst", "generated/skglm.penalties.L2_3.rst", "generated/skglm.penalties.LogSumPenalty.rst", "generated/skglm.penalties.MCPenalty.rst", "generated/skglm.penalties.PositiveConstraint.rst", "generated/skglm.penalties.SCAD.rst", "generated/skglm.penalties.SLOPE.rst", "generated/skglm.penalties.WeightedGroupL2.rst", "generated/skglm.penalties.WeightedL1.rst", "generated/skglm.penalties.WeightedMCPenalty.rst", "generated/skglm.solvers.AndersonCD.rst", "generated/skglm.solvers.FISTA.rst", "generated/skglm.solvers.GramCD.rst", "generated/skglm.solvers.GroupBCD.rst", "generated/skglm.solvers.GroupProxNewton.rst", "generated/skglm.solvers.LBFGS.rst", "generated/skglm.solvers.MultiTaskBCD.rst", "generated/skglm.solvers.ProxNewton.rst", "getting_started.rst", "index.rst", "sg_execution_times.rst", "tutorials/add_datafit.rst", "tutorials/add_penalty.rst", "tutorials/alpha_max.rst", "tutorials/cox_datafit.rst", "tutorials/intercept.rst", "tutorials/prox_nn_group_lasso.rst", "tutorials/tutorials.rst"], "titles": ["API Documentation", "Examples Gallery", "Timing comparison with scikit-learn for Lasso", "Group Logistic regression in python", "Comparison of Lasso and Weighted Lasso", "Logistic regression with Elastic net and minimax concave penalties", "Value and proximal operators of penalties", "Timing comparison between direct prox computation and reweighting", "Fast Sparse Group Lasso in python", "Sparse recovery with non-convex penalties", "Comparison of lifelines with skglm for survival analysis", "Show U-curve of regularization", "Weighted Lasso with some zero weights", "Computation times", "Version 0.1 (2022/04/25)", "Version 0.2 (2022/10/18)", "Version 0.3 (2023/07/01)", "Version 0.4 (in progress)", "What\u2019s new", "Contribute to <code class=\"docutils literal notranslate\"><span class=\"pre\">skglm</span></code>", "skglm.CoxEstimator", "skglm.ElasticNet", "skglm.GeneralizedLinearEstimator", "skglm.GroupLasso", "skglm.Lasso", "skglm.LinearSVC", "skglm.MCPRegression", "skglm.MultiTaskLasso", "skglm.SparseLogisticRegression", "skglm.WeightedLasso", "skglm.datafits.Cox", "skglm.datafits.Gamma", "skglm.datafits.Huber", "skglm.datafits.Logistic", "skglm.datafits.LogisticGroup", "skglm.datafits.Poisson", "skglm.datafits.Quadratic", "skglm.datafits.QuadraticGroup", "skglm.datafits.QuadraticSVC", "skglm.datafits.WeightedQuadratic", "skglm.experimental.IterativeReweightedL1", "skglm.experimental.PDCD_WS", "skglm.experimental.Pinball", "skglm.experimental.SqrtLasso", "skglm.experimental.SqrtQuadratic", "skglm.penalties.BlockSCAD", "skglm.penalties.IndicatorBox", "skglm.penalties.L0_5", "skglm.penalties.L1", "skglm.penalties.L1_plus_L2", "skglm.penalties.L2", "skglm.penalties.L2_3", "skglm.penalties.LogSumPenalty", "skglm.penalties.MCPenalty", "skglm.penalties.PositiveConstraint", "skglm.penalties.SCAD", "skglm.penalties.SLOPE", "skglm.penalties.WeightedGroupL2", "skglm.penalties.WeightedL1", "skglm.penalties.WeightedMCPenalty", "skglm.solvers.AndersonCD", "skglm.solvers.FISTA", "skglm.solvers.GramCD", "skglm.solvers.GroupBCD", "skglm.solvers.GroupProxNewton", "skglm.solvers.LBFGS", "skglm.solvers.MultiTaskBCD", "skglm.solvers.ProxNewton", "Getting started", "<code class=\"docutils literal notranslate\"><span class=\"pre\">skglm</span></code>", "Computation times", "How to add a custom datafit", "How to add a custom penalty", "Critical regularization strength above which solution is 0", "Mathematic behind Cox datafit", "Computation of the intercept", "Details on the Positive Group Lasso", "Tutorials"], "terms": {"group": [1, 13, 15, 17, 18, 23, 30, 34, 37, 57, 63, 64, 70, 74], "logist": [1, 13, 15, 18, 28, 34, 70], "regress": [1, 8, 13, 20, 22, 26, 28, 70], "python": [1, 2, 4, 5, 6, 7, 9, 10, 11, 12, 13, 69, 70, 72], "valu": [1, 7, 8, 9, 10, 13, 20, 45, 52, 53, 55, 56, 59, 62, 63, 68, 70, 71, 72, 73, 75], "proxim": [1, 7, 10, 13, 15, 18, 41, 57, 70, 71, 72], "oper": [1, 13, 57, 70, 71, 72, 74], "penalti": [1, 2, 3, 7, 8, 10, 12, 13, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 29, 34, 37, 40, 41, 43, 62, 68, 69, 70, 71, 73], "show": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 68, 70, 71, 73, 76], "u": [1, 2, 13, 19, 68, 69, 70, 73, 74, 76], "curv": [1, 13, 70], "regular": [1, 7, 10, 13, 16, 18, 20, 21, 24, 25, 26, 27, 28, 29, 56, 57, 68, 69, 70, 72], "comparison": [1, 13, 70], "lasso": [1, 9, 11, 13, 15, 16, 17, 18, 21, 23, 26, 29, 43, 57, 70], "weight": [1, 3, 6, 13, 17, 18, 23, 24, 26, 29, 39, 40, 57, 58, 59, 70], "some": [1, 4, 6, 8, 13, 68, 70, 71, 72, 74, 76], "zero": [1, 3, 4, 7, 8, 9, 11, 13, 41, 62, 63, 70, 71, 72, 76], "fast": [1, 12, 13, 61, 69, 70], "spars": [1, 6, 10, 13, 21, 22, 24, 25, 26, 27, 28, 29, 30, 41, 53, 59, 60, 64, 67, 69, 70, 71], "elast": [1, 13, 16, 18, 20, 21, 70], "net": [1, 13, 21, 70], "minimax": [1, 13, 53, 59, 70], "concav": [1, 13, 53, 59, 70], "time": [1, 3, 4, 5, 6, 8, 9, 11, 12, 16, 18, 20, 30, 31, 33, 35, 38, 69, 71, 74, 75, 76], "between": [1, 10, 11, 13, 70], "direct": [1, 13, 70], "prox": [1, 13, 26, 28, 41, 43, 64, 67, 70, 76], "comput": [1, 9, 11, 12, 17, 18, 31, 33, 35, 36, 38, 39, 62, 68, 71, 72, 74, 76], "reweight": [1, 13, 40, 70], "scikit": [1, 3, 8, 13, 19, 68, 69, 70], "learn": [1, 3, 8, 13, 19, 68, 69, 70, 73, 77], "recoveri": [1, 13, 70], "non": [1, 3, 8, 13, 15, 16, 18, 25, 40, 47, 51, 53, 59, 70, 72, 76], "convex": [1, 13, 40, 41, 46, 47, 51, 53, 56, 59, 70, 73, 74], "lifelin": [1, 13, 70], "skglm": [1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 68, 70, 71, 72, 73, 75, 77], "surviv": [1, 13, 16, 18, 30, 70, 74], "analysi": [1, 13, 16, 18, 30, 70, 74], "download": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "all": [1, 3, 9, 11, 12, 70, 72, 75, 76], "sourc": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67], "code": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 41, 60, 64, 67, 69], "auto_examples_python": 1, "zip": [1, 2, 10], "jupyt": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "notebook": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "auto_examples_jupyt": 1, "gener": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 22, 64, 67, 69, 71, 72, 75], "sphinx": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "go": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "end": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 73], "full": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "exampl": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 19, 68, 70, 71, 77], "compar": [2, 7, 10], "solv": [2, 10, 16, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 69, 71, 75], "larg": [2, 12, 41, 68, 69], "scale": [2, 64, 67, 72], "problem": [2, 7, 10, 15, 16, 18, 20, 22, 25, 28, 40, 41, 60, 61, 63, 66, 71], "file_s": [2, 11], "0": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 42, 43, 46, 47, 53, 55, 57, 59, 60, 61, 62, 63, 64, 65, 66, 67, 70, 71, 72, 74, 75], "00": [2, 11, 13, 70], "26": [2, 15, 18], "8m": 2, "b": [2, 11, 64, 65, 67, 73, 74, 76], "": [2, 5, 7, 9, 10, 11, 20, 30, 64, 67, 68, 69, 71, 73, 74, 76], "24": [2, 7, 11], "6k": [2, 11], "03": [2, 3, 13, 70], "34": 2, "125kb": [2, 11], "49": [2, 11, 13, 70], "2k": [2, 11], "35": 2, "124kb": [2, 11], "106k": [2, 11], "02": [2, 3, 8, 9, 11, 13, 70], "13": [2, 11, 13, 70, 74], "200kb": 2, "1": [2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 45, 46, 52, 53, 55, 56, 57, 59, 60, 61, 62, 64, 67, 68, 71, 72, 73, 74, 75, 76], "221k": [2, 11], "01": [2, 3, 8, 11, 73], "16": 2, "349kb": [2, 11], "2": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 29, 32, 36, 37, 38, 39, 41, 45, 50, 51, 53, 55, 59, 60, 62, 64, 67, 71, 73, 74, 75, 76], "451k": [2, 11], "41": 2, "640kb": 2, "3": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 26, 51, 56, 69, 73, 74, 76], "909k": [2, 11], "21": [2, 27], "21mb": [2, 11], "7": [2, 6, 8, 11, 20, 74], "83m": [2, 11], "10": [2, 3, 4, 7, 8, 10, 11, 12, 21, 23, 24, 25, 26, 27, 29, 41, 43, 56, 60, 61, 63, 64, 66, 67, 68], "32mb": [2, 11], "14": [2, 15, 18], "66m": [2, 11], "05": [2, 3, 8, 11, 13, 70], "4": [2, 3, 4, 6, 8, 9, 10, 11, 23, 43, 61, 62, 63, 64, 65, 67, 73, 75, 76], "51mb": 2, "20": [2, 10, 11, 14, 18, 28, 43, 64, 65, 67], "5": [2, 3, 4, 7, 8, 9, 10, 11, 12, 20, 21, 40, 42, 47, 62, 76], "23m": [2, 11], "57mb": 2, "25": 2, "6": [2, 3, 8, 9, 10, 11, 74], "81m": [2, 11], "31mb": 2, "9": [2, 8, 56], "43m": 2, "8": [2, 5, 6, 7, 8, 9, 11], "11": [2, 8, 11, 13, 70], "0m": [2, 11], "22mb": 2, "47": [2, 13, 70], "12": [2, 9, 11, 15], "6m": [2, 11], "16mb": 2, "53": 2, "1m": 2, "13mb": 2, "63": [2, 15, 18], "60mb": 2, "72": [2, 74], "19": 2, "4m": 2, "6mb": 2, "78": 2, "90mb": 2, "88": 2, "23": 2, "9mb": 2, "98": 2, "2m": 2, "5mb": 2, "100": [2, 5, 7, 10, 11, 12, 41, 43, 61, 62, 63, 66, 68, 69], "08mb": 2, "import": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "warn": [2, 10], "numpi": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "np": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 26, 30, 41, 68, 71, 72, 73], "from": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 17, 18, 21, 22, 24, 25, 26, 27, 29, 68, 69, 70, 71, 72, 74], "linalg": [2, 4, 5, 7, 9, 10, 11], "norm": [2, 4, 5, 7, 9, 10, 11, 24, 29, 40, 47, 51, 73, 75, 76], "matplotlib": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "pyplot": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "plt": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "libsvmdata": [2, 11], "fetch_libsvm": [2, 11], "sklearn": [2, 5, 9, 11, 68], "except": [2, 68], "convergencewarn": 2, "linear_model": 2, "lasso_sklearn": 2, "elasticnet": [2, 20, 49], "enet_sklearn": 2, "filterwarn": [2, 10], "ignor": [2, 10], "categori": 2, "def": [2, 7, 71, 72], "compute_obj": 2, "x": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 23, 25, 26, 27, 30, 32, 36, 39, 53, 55, 59, 62, 68, 71, 73, 74, 75, 76], "y": [2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 26, 27, 29, 36, 37, 39, 43, 44, 62, 68, 71, 74, 75], "w": [2, 7, 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, 38, 40, 41, 43, 45, 50, 52, 56, 62, 71, 72], "alpha": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 26, 27, 28, 29, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 55, 56, 57, 58, 59, 68, 72, 73], "l1_ratio": [2, 5, 6, 10, 20, 21, 49], "loss": [2, 25, 73], "len": [2, 4, 5, 7, 8, 9, 10, 11, 12, 71], "sum": [2, 7, 10, 52, 71, 72, 74], "ab": [2, 4, 7, 12, 32, 41, 52, 53, 55, 59, 60, 64, 67, 72, 73], "return": [2, 7, 71, 72], "news20": 2, "binari": [2, 11, 28, 72, 74], "max": [2, 4, 7, 9, 12, 25, 42, 68, 72, 73, 76], "t": [2, 4, 7, 9, 11, 12, 25, 28, 36, 38, 39, 62, 64, 67, 68, 71, 73], "dict_sklearn": 2, "fit_intercept": [2, 4, 7, 8, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 28, 29, 60, 62, 63, 64, 66, 67], "fals": [2, 4, 6, 7, 8, 9, 10, 11, 12, 20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 48, 49, 53, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 72], "tol": [2, 7, 8, 11, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "1e": [2, 7, 8, 9, 10, 11, 23, 41, 43, 61, 62, 63, 64, 65, 66, 67], "enet": 2, "dict_our": 2, "model": [2, 3, 8, 10, 41, 60, 63, 64, 67, 68, 69, 71, 75], "fig": [2, 4, 6, 7, 9, 10, 12], "axarr": [2, 4, 6, 7, 9, 12], "subplot": [2, 4, 6, 7, 9, 10, 12], "constrained_layout": [2, 6, 7, 9, 12], "true": [2, 4, 5, 6, 7, 9, 10, 11, 12, 21, 23, 24, 25, 26, 27, 28, 29, 30, 57, 60, 62, 66, 67, 72], "ax": [2, 7, 10, 73], "pobj_dict": 2, "list": [2, 10, 23, 28, 40, 68], "time_dict": 2, "remov": [2, 17, 18], "compil": [2, 7, 22, 31, 33, 35, 38, 40, 68, 71, 72], "max_it": [2, 10, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "10_000": 2, "w_star": [2, 5], "fit": [2, 3, 4, 5, 7, 8, 11, 12, 15, 16, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 33, 35, 38, 40, 60, 63, 67, 71, 77], "coef_": [2, 3, 4, 5, 7, 8, 12, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40], "pobj_star": 2, "n_iter_sklearn": 2, "uniqu": [2, 10, 30, 56, 74], "geomspac": [2, 9, 11], "50": [2, 4, 5, 10, 11, 12, 20, 21, 23, 24, 25, 26, 27, 29, 60, 65], "num": [2, 6, 9], "15": [2, 3, 13, 56, 70], "astyp": 2, "int": [2, 9, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67], "t_start": 2, "w_sklearn": 2, "append": [2, 10, 11], "n_iter_u": 2, "rang": [2, 9, 10, 69, 71], "semilogi": [2, 10], "label": [2, 5, 6, 9, 10, 11, 28, 33], "set_ylim": [2, 9], "set_titl": [2, 4, 6, 7, 10, 12], "legend": [2, 5, 6, 9, 10, 11], "set_ylabel": [2, 7, 9, 10], "object": [2, 7, 8, 10, 21, 23, 24, 25, 26, 27, 28, 29, 40, 43, 71, 72, 76], "suboptim": [2, 10], "set_xlabel": [2, 7, 9, 10, 12], "block": [2, 4, 6, 7, 9, 11, 12, 15, 18, 23, 45, 60, 63, 66, 74, 76], "total": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 70, 74], "run": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 22, 69], "script": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 68], "minut": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "620": [2, 13, 70], "second": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 20, 72, 74], "plot_compare_tim": [2, 13, 70], "ipynb": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], "py": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 70], "galleri": [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 19, 70], "i": [3, 7, 8, 10, 11, 12, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 38, 39, 40, 41, 43, 45, 46, 56, 57, 60, 61, 62, 63, 64, 65, 67, 68, 69, 71, 72, 74, 75, 76], "miss": [3, 8, 69], "estim": [3, 7, 8, 9, 11, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 43, 56, 69], "we": [3, 8, 10, 11, 19, 25, 68, 71, 72, 73, 74, 75, 76], "how": [3, 8, 10, 12, 19, 41, 68, 76], "implement": [3, 8, 68, 69, 71, 72, 75, 77], "one": [3, 5, 8, 76], "author": [3, 4, 5, 6, 7, 8, 9, 69], "mathurin": [3, 4, 6, 8, 9], "massia": [3, 4, 6, 8, 9, 41, 60, 64, 67, 69], "generalizedlinearestim": [3, 5, 7, 8, 68, 71], "datafit": [3, 5, 8, 9, 10, 15, 16, 17, 18, 20, 22, 25, 40, 41, 42, 44, 62, 68, 69, 72], "logisticgroup": 3, "weightedgroupl2": 3, "solver": [3, 7, 8, 9, 10, 12, 15, 16, 17, 18, 20, 22, 23, 24, 29, 40, 41, 43, 69, 71, 72, 75], "groupproxnewton": 3, "util": [3, 4, 5, 7, 8, 9, 10, 12], "data": [3, 4, 5, 7, 8, 9, 11, 12, 20, 71, 74], "make_correlated_data": [3, 4, 5, 7, 8, 9, 12], "grp_convert": [3, 8], "n_featur": [3, 4, 5, 7, 8, 9, 10, 12, 15, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 34, 36, 37, 39, 40, 56, 57, 62, 63, 68, 71, 72], "30": [3, 4, 8, 9], "_": [3, 4, 5, 8, 10, 27, 41, 46, 71, 73, 74, 75], "n_sampl": [3, 4, 5, 7, 8, 9, 10, 12, 15, 18, 25, 30, 39, 41, 62, 63, 68, 71], "random_st": [3, 4, 5, 7, 8, 9, 10, 12], "sign": [3, 5, 72, 75], "classifi": [3, 28], "creation": [3, 8], "combin": [3, 8, 10, 20, 21, 64, 67, 68, 69, 74, 76], "grp_size": [3, 8], "ar": [3, 8, 10, 19, 23, 26, 29, 41, 68, 69, 71, 72, 74, 75], "made": [3, 69], "consecut": [3, 8, 30, 34, 37, 57], "featur": [3, 4, 8, 11, 12, 19, 23, 26, 28, 29, 41, 43, 45, 52, 53, 55, 56, 59, 60, 62, 64, 67, 68, 69, 71, 72, 75], "n_group": [3, 8, 23, 34, 37, 57], "grp_indic": [3, 8, 34, 37, 57], "grp_ptr": [3, 8, 34, 37, 57], "ones": [3, 8, 72, 75], "verbos": [3, 8, 20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 66, 67], "train": [3, 5, 8, 11, 68], "clf": [3, 8, 11, 12, 71], "iter": [3, 7, 8, 10, 20, 21, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67, 72], "6931471806": 3, "stop": [3, 8, 20, 21, 23, 24, 25, 26, 27, 28, 29, 72], "crit": [3, 8], "93e": 3, "pn": 3, "1818203070": 3, "38e": 3, "earli": 3, "exit": 3, "1130022685": 3, "59e": [3, 8, 10], "0920956247": 3, "58e": 3, "0846061796": 3, "29e": 3, "0799359080": 3, "83e": 3, "0794149142": 3, "96e": 3, "04": [3, 10], "0794078275": 3, "61e": 3, "In": [3, 8, 68, 73, 74, 75], "environ": [3, 8], "pleas": [3, 8, 69], "rerun": [3, 8], "thi": [3, 8, 10, 12, 19, 22, 25, 31, 33, 35, 38, 40, 62, 68, 71, 72, 73, 74, 75, 76], "cell": [3, 8], "html": [3, 8, 19, 60, 64, 67], "represent": [3, 8], "trust": [3, 8], "On": [3, 8, 74], "github": [3, 8, 19, 41, 60, 64, 67], "unabl": [3, 8], "render": [3, 8], "try": [3, 8], "load": [3, 8, 11], "page": [3, 8, 68, 69], "nbviewer": [3, 8], "org": [3, 8, 41, 56, 60, 61, 64, 67], "generalizedlinearestimatorgeneralizedlinearestim": [3, 8], "check": [3, 10, 17, 18, 19, 68], "either": [3, 69], "print": [3, 9, 10, 68], "reshap": [3, 8], "10333799": 3, "8924748": 3, "77600881": 3, "68695901": 3, "26460422": 3, "06153908": 3, "13524978": 3, "01904796": 3, "02993834": 3, "02193371": 3, "02044038": 3, "00869443": 3, "90363392": 3, "96337048": 3, "49341585": 3, "57737623": 3, "25394772": 3, "15405176": 3, "201": [3, 13, 70], "plot_group_logistic_regress": [3, 13, 70], "illustr": [4, 5, 6, 9, 11, 68, 74], "normal": [4, 10], "when": [4, 16, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 42, 56, 57, 62, 71, 72], "penal": [4, 10, 12, 26, 56, 72, 75], "quentin": [4, 9], "bertrand": [4, 9, 41, 60, 69], "weightedlasso": [4, 12, 24], "w_true": [4, 7, 9, 12], "nnz": 4, "assum": [4, 23, 74], "reason": [4, 69], "have": [4, 10, 19, 30, 68, 71, 72, 73, 74, 75], "smaller": 4, "than": [4, 10, 24, 26, 29, 68, 69], "other": [4, 68, 69, 71, 74, 76], "nois": 4, "random": [4, 9, 12, 68], "randn": 4, "shape": [4, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 34, 36, 37, 39, 40, 41, 56, 57, 62, 63, 71], "us": [4, 5, 10, 12, 19, 20, 21, 23, 24, 25, 26, 27, 29, 30, 31, 33, 34, 35, 36, 37, 38, 39, 41, 60, 62, 63, 67, 68, 69, 71, 72, 73, 74, 75, 76], "signal": 4, "ratio": [4, 10], "doe": 4, "select": [4, 26, 56, 62, 68], "small": [4, 12, 19], "while": [4, 68], "alpha_max": [4, 7, 9, 11, 12, 72, 73], "la": 4, "wei": 4, "axi": [4, 5, 8, 9], "sharei": [4, 9, 12], "figsiz": [4, 6, 7, 9, 10, 12], "stem": [4, 5, 12], "coeff": 4, "903": [4, 13, 70], "plot_lasso_vs_weight": [4, 13, 70], "modular": [5, 68, 69], "pierr": [5, 7], "antoin": [5, 7], "bannier": [5, 7, 41, 60, 69], "metric": [5, 9, 10, 11], "f1_score": [5, 9], "l1_plus_l2": [5, 6, 16, 18], "mcpenalti": [5, 6, 9], "y_ind": 5, "standard": [5, 9, 10], "mcp": [5, 9, 26, 53, 59], "sqrt": [5, 7, 9, 76], "split": [5, 68], "set": [5, 10, 15, 17, 18, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 46, 57, 60, 62, 63, 64, 67, 68, 72, 74, 76], "test": [5, 11, 19, 68], "x_train": [5, 11, 68], "y_train": [5, 11, 68], "x_test": [5, 9, 11, 68], "y_test": [5, 9, 11, 68], "005": 5, "gamma": [5, 6, 9, 16, 18, 26, 45, 53, 55, 59], "clf_enet": 5, "y_pred_enet": 5, "predict": [5, 11, 68], "f1_score_enet": 5, "clf_mcp": 5, "y_pred_mcp": 5, "f1_score_mcp": 5, "m": [5, 41, 56, 60, 61, 64, 67, 69, 74], "where": [5, 10, 20, 30, 32, 41, 46, 57, 62, 73, 74, 75, 76], "ravel": 5, "markerfmt": 5, "coeffici": [5, 8, 10, 17, 18, 21, 23, 24, 26, 28, 29, 41, 57, 62, 63, 68, 71, 72, 74, 75], "setp": 5, "color": [5, 9], "2ca02c": 5, "ff7f0e": 5, "bx": 5, "loc": [5, 9], "best": 5, "titl": [5, 8, 11, 69], "f1": [5, 9], "3f": 5, "967": [5, 13, 70], "plot_logreg_various_penalti": [5, 13, 70], "gmail": [6, 7], "com": [6, 7, 19, 41, 60, 64, 67, 76], "weightedl1": [6, 16, 18], "l1": [6, 9, 10, 16, 18, 20, 21, 22, 23, 24, 26, 29, 40, 41, 56, 58, 60, 68, 69, 73, 77], "scad": [6, 9, 15, 18], "l0_5": [6, 7, 9, 40], "l2_3": [6, 9], "arrai": [6, 7, 10, 20, 21, 22, 23, 24, 25, 26, 27, 29, 30, 34, 36, 37, 39, 40, 41, 56, 57, 62, 63, 71], "x_rang": 6, "linspac": 6, "300": 6, "pen": [6, 26, 45, 53, 55, 59], "plot": [6, 7, 9, 10, 11, 19], "__class__": 6, "__name__": 6, "prox_1d": [6, 72], "487": [6, 13, 70], "plot_pen_prox": [6, 13, 70], "pierreantoin": 7, "panda": [7, 10], "pd": [7, 10], "separ": [7, 17, 18, 40, 41, 76], "experiment": [7, 15, 16, 18], "iterativereweightedl1": 7, "andersoncd": [7, 9, 22, 40, 71], "200": 7, "500": [7, 10], "ord": [7, 9, 11], "inf": [7, 9, 11, 26, 72], "1000": [7, 9, 28, 41, 43, 63, 64, 67, 68], "_obj": 7, "fit_l05": 7, "start": [7, 10, 69, 74], "iterative_l05": 7, "iterative_tim": 7, "subdiff": [7, 21, 23, 24, 25, 26, 27, 29, 60, 61, 63, 66, 67, 72], "strategi": [7, 41, 62, 76], "uninform": 7, "direct_l05": 7, "ws_strategi": [7, 8, 9, 21, 23, 24, 25, 26, 27, 29, 60, 63, 66, 67, 72], "fixpoint": [7, 8, 9, 21, 23, 24, 25, 26, 27, 29, 60, 67, 72], "direct_tim": 7, "result": [7, 69, 74], "cach": [7, 68], "numba": [7, 31, 33, 35, 38, 68, 71, 72], "time_result": 7, "obj_result": 7, "actual": 7, "enumer": [7, 9, 10, 11, 72], "iterative_obj": 7, "direct_obj": 7, "time_df": 7, "datafram": [7, 10], "column": [7, 20, 74, 75], "obj_df": 7, "index": [7, 8, 12, 74], "bar": 7, "rot": 7, "r": [7, 9, 11, 12, 71, 72, 73, 74], "lambda": [7, 9, 11, 72, 73, 76], "lambda_": [7, 9, 73], "obj": [7, 10, 26], "solut": [7, 9, 10, 11, 21, 23, 24, 25, 26, 27, 28, 29, 72, 75], "326": [7, 13, 70], "plot_reweighted_l1": [7, 13, 70], "groupbcd": [8, 15, 18], "quadraticgroup": 8, "weightedl1groupl2": 8, "take": [8, 22, 68, 71, 76], "weights_g": [8, 23], "dtype": [8, 71, 72], "float64": [8, 72], "weights_f": 8, "weights_group": 8, "weights_featur": 8, "9165123923": 8, "04e": 8, "0255867844": 8, "6875753707": 8, "40e": 8, "6804572587": 8, "23e": 8, "6804570767": 8, "22e": 8, "6804570709": 8, "64e": 8, "07": 8, "17e": 8, "08": [8, 11, 13, 70], "37e": 8, "fulli": [8, 69], "insid": 8, "too": [8, 11], "imshow": 8, "cmap": [8, 9], "grei": 8, "black": 8, "ylabel": 8, "xlabel": [8, 11], "xtick": 8, "arang": [8, 12], "ytick": 8, "0x7fef611c0cd0": 8, "0x7fef60f50df0": 8, "0x7fef6867a6d0": 8, "text": [8, 73, 74, 76], "796": [8, 13, 70], "plot_sparse_group_lasso": [8, 13, 70], "superior": 9, "perform": [9, 40, 41, 69, 74], "l05": 9, "l23": 9, "klopfenstein": [9, 41, 60, 69], "model_select": [9, 11, 68], "train_test_split": [9, 11, 68], "mean_squared_error": [9, 11], "quadrat": [9, 15, 17, 18, 22, 37, 38, 39, 40, 44, 62, 71], "jit_compil": [9, 10], "compiled_clon": [9, 10], "get_cmap": 9, "tab10": 9, "simul": 9, "densiti": 9, "seed": [9, 12], "supp": 9, "choic": [9, 12, 68], "size": [9, 21, 23, 24, 25, 26, 27, 29, 41, 60, 74, 75], "replac": [9, 12, 19, 22, 40, 69], "x_": [9, 71, 75, 76], "y_": [9, 74], "snr": 9, "rho": 9, "test_siz": 9, "which": [9, 10, 11, 62, 69, 72, 74], "lead": [9, 75, 76], "defin": [9, 11, 74, 75], "n_alpha": 9, "estimation_error": 9, "prediction_error": 9, "l0": 9, "mse_ref": 9, "zeros_lik": [9, 72], "idx": [9, 10, 11, 72], "kei": [9, 19], "f": [9, 10, 32, 41, 71, 72, 73, 74, 75], "estimator_path": 9, "path": [9, 11], "f1_temp": 9, "prediction_error_temp": 9, "j": [9, 26, 41, 45, 52, 53, 55, 59, 61, 64, 67, 71, 72, 73, 74, 75, 76], "name_estim": 9, "ell_": [9, 47, 51], "close": [9, 10, 11], "sharex": 9, "semilogx": [9, 11], "c": [9, 25, 56, 64, 67, 68, 69, 76], "max_f1": 9, "argmax": 9, "vline": 9, "ymin": 9, "ymax": 9, "linestyl": 9, "line1": 9, "clip_on": 9, "marker": [9, 10], "markers": 9, "min_error": 9, "argmin": [9, 73, 75], "lim": 9, "get_ylim": 9, "min": [9, 10, 73, 76], "line2": 9, "mathrm": 9, "score": [9, 21, 23, 24, 25, 26, 27, 29, 41, 60, 67, 68], "pred": 9, "rmse": 9, "left": [9, 71], "out": [9, 10, 19], "bbox_to_anchor": 9, "lower": [9, 68], "mode": [9, 19], "expand": 9, "borderaxespad": 9, "ncol": 9, "881": [9, 13, 70], "plot_sparse_recoveri": [9, 13, 70], "exactli": [10, 76], "x100": 10, "less": 10, "let": [10, 11, 68, 71, 73, 74, 75, 76], "first": [10, 11, 12, 19, 20, 21, 23, 24, 25, 26, 27, 29, 41, 60, 68, 69, 71, 72, 74], "synthet": 10, "make_dummy_survival_data": 10, "tm": [10, 20, 30], "The": [10, 12, 20, 21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 38, 39, 41, 42, 43, 44, 50, 52, 56, 57, 60, 67, 68, 71, 72, 73, 74, 76], "ha": [10, 68, 74, 76], "follow": [10, 19, 71, 72, 75], "properti": [10, 21, 22, 24, 25, 26, 27, 29], "matrix": [10, 20, 21, 22, 24, 25, 26, 27, 29, 62, 71, 74, 75], "predictor": [10, 20, 74], "distribut": 10, "toeplitz": 10, "covari": 10, "vector": [10, 11, 20, 21, 23, 24, 25, 26, 27, 29, 40, 41, 57, 62, 63, 71, 72, 74, 75], "occurr": [10, 20, 30, 74], "weibul": 10, "indic": [10, 23, 30, 34, 37, 46, 57, 71, 75], "observ": [10, 20, 30, 74], "censorship": [10, 74], "bernoulli": 10, "inspect": 10, "quickli": 10, "tight_layout": 10, "dist": [10, 72], "axes_titl": 10, "fifth": 10, "name": [10, 72], "hist": 10, "bin": 10, "auto": 10, "count": 10, "after": [10, 40, 68, 75], "can": [10, 11, 19, 21, 23, 24, 25, 26, 27, 29, 62, 68, 69, 71, 72, 74, 75, 76], "now": [10, 68, 71, 74, 75], "todo": 10, "so": [10, 19, 69], "need": [10, 19, 71, 72, 76], "ell_1": [10, 48, 49, 72, 77], "newton": [10, 15, 18, 28, 43, 64, 67], "proxnewton": [10, 15, 16, 17, 18, 20, 71], "intens": 10, "intern": 10, "init": [10, 68], "initi": [10, 14, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 62, 63, 71, 72], "w_sk": 10, "For": [10, 11, 20, 21, 26, 68, 71, 72, 73, 74, 76], "rel": 10, "found": [10, 71, 72], "number": [10, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 43, 60, 61, 62, 63, 64, 65, 67, 68, 73, 75], "nonzero": 10, "66": 10, "through": [10, 68, 69, 74, 77], "its": [10, 41, 68, 69, 71, 72, 75, 76], "coxphfitt": 10, "two": [10, 20, 30, 34, 37, 57, 71, 74], "packag": [10, 69, 75], "format": 10, "stacked_y_x": 10, "hstack": 10, "df": 10, "lifelines_estim": 10, "duration_col": 10, "event_col": 10, "w_ll": 10, "params_": 10, "both": [10, 71, 74], "find": [10, 68], "same": [10, 30, 68, 74], "obj_sk": 10, "obj_ll": 10, "6f": 10, "differ": [10, 71, 75], "2e": 10, "494576": 10, "47e": 10, "do": [10, 19, 69, 75], "euclidean": 10, "distanc": [10, 17, 18, 41, 57, 72, 76], "3e": 10, "373e": 10, "yield": [10, 71, 76], "execut": [10, 13, 70], "To": [10, 19, 25, 71, 72, 74], "get": [10, 69, 71, 72, 74, 77], "evolut": 10, "optim": [10, 11, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 56, 60, 64, 67, 71, 72, 73, 75, 76], "increas": [10, 76], "save": 10, "record": [10, 74], "max_run": 10, "n_iter": 10, "perf_count": 10, "fit_opt": 10, "max_step": 10, "cast": [10, 74], "asarrai": 10, "optimal_obj": 10, "o": [10, 74], "accord": [10, 73, 74], "more": [10, 26, 68, 73], "speed_up": 10, "speed": [10, 68, 69], "up": [10, 62, 69], "0f": 10, "930": 10, "previou": [10, 21, 23, 24, 25, 26, 27, 28, 29], "extend": [10, 19, 74], "case": [10, 19, 69, 74, 75], "handl": [10, 12, 20, 22, 28, 30, 39, 74, 75], "ti": [10, 20, 30, 74], "achiev": [10, 68], "pass": [10, 23, 68, 71, 72], "with_ti": 10, "function": [10, 15, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 40, 41, 46, 57, 73, 74, 75, 76], "51": [10, 15, 18], "It": [10, 20, 22, 30, 41, 62, 68, 69, 72, 74], "straightforward": 10, "an": [10, 11, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 60, 62, 67, 68, 71, 72, 74, 75, 77], "onli": [10, 62, 71, 72, 73, 74, 75, 76], "use_efron": [10, 30], "ensur": [10, 69], "again": [10, 74], "82": 10, "stacked_tm_s_x": 10, "none": [10, 22, 23, 26, 29, 40, 41, 62, 63], "new": [10, 19, 69, 72], "501289": 10, "09": 10, "046e": 10, "final": [10, 11, 72, 74, 75], "total_time_skglm": 10, "total_time_lifelin": 10, "deduc": [10, 74], "1042": 10, "As": [10, 19, 76], "shown": [10, 73], "last": [10, 12, 19, 68], "still": 10, "preserv": 10, "even": [10, 19], "57": [10, 13, 15, 18, 70], "445": [10, 13, 70], "plot_survival_analysi": [10, 13, 70], "sweet": 11, "spot": 11, "much": [11, 41], "littl": 11, "showcas": 11, "rcv1": 11, "dataset": [11, 68, 69, 71], "keep": [11, 62], "2000": 11, "also": [11, 41, 69, 74], "retrain": 11, "sampl": [11, 17, 18, 21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 39, 46, 62, 68, 71, 73, 74, 75], "7m": 11, "199kb": 11, "38": 11, "639kb": 11, "27": 11, "50mb": 11, "56mb": 11, "29mb": 11, "61": 11, "38m": 11, "81mb": 11, "80": 11, "65mb": 11, "92": 11, "48mb": 11, "next": 11, "well": [11, 68, 72], "know": 11, "abov": 11, "along": 11, "mse": 11, "mse_train": 11, "mse_test": 11, "warm_start": [11, 21, 23, 24, 25, 26, 27, 28, 29, 60, 62, 63, 64, 66, 67], "notic": [11, 74], "around": 11, "sit": 11, "boundari": 11, "underfit": 11, "overfit": 11, "mean": [11, 72, 74], "squar": [11, 15, 18, 43, 44, 73, 74, 76], "error": [11, 68], "strength": [11, 20, 21, 23, 24, 25, 26, 27, 28, 29, 43], "957": [11, 13, 70], "plot_ucurv": [11, 13, 70], "demonstr": 12, "vanish": 12, "adapt": [12, 41, 56, 68], "primal": [12, 16, 18, 23, 24, 25, 29, 41], "anderson": [12, 60], "acceler": [12, 15, 18, 60, 61], "allow": [12, 31, 33, 35, 38, 68, 71, 72], "dual": [12, 16, 18, 25, 41, 64, 67], "empti": [12, 76], "unpen": [12, 23, 29, 67, 77], "put": [12, 72], "coef": 12, "neq": 12, "heavili": 12, "few": [12, 69], "lightli": 12, "mani": [12, 69, 75], "374": [12, 13, 70], "plot_zero_weights_lasso": [12, 13, 70], "958": [13, 70], "file": [13, 70], "auto_exampl": 13, "mem": [13, 70], "mb": [13, 70], "releas": [14, 18], "commit": [14, 18], "root": [15, 18, 43, 44], "class": [15, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72, 75], "chambol": [15, 18], "pock": [15, 18], "pr": [15, 16, 17, 18], "coordin": [15, 16, 18, 22, 26, 41, 60, 62, 63, 66, 71, 72, 76], "descent": [15, 16, 18, 22, 26, 41, 60, 62, 63, 66, 75], "work": [15, 17, 18, 21, 23, 24, 25, 26, 27, 29, 41, 43, 60, 63, 64, 67, 76], "29": [15, 18], "28": [15, 18], "huber": [15, 18], "ad": [15, 18, 19, 68, 75], "blockscad": [15, 18], "intercept": [15, 18, 21, 23, 24, 25, 26, 27, 28, 29, 60, 67], "55": [15, 18], "gramcd": [15, 18], "gram": [15, 18, 62], "base": [15, 18, 23, 24, 29, 41, 43, 62, 71, 72], "ll": [15, 18], "59": [15, 18], "transform": [15, 18], "add": [16, 17, 18, 19, 68], "cox": [16, 18, 20], "support": [16, 17, 18, 23, 29, 68, 69, 71, 72], "l2": [16, 18, 20, 21, 25, 57], "171": [16, 18], "reduc": [16, 18, 74, 76], "overhead": [16, 18, 62, 68], "129": [16, 18], "180": [16, 18], "168": [16, 18], "159": [16, 18], "157": [16, 18], "pinbal": [16, 18], "134": [16, 18], "113": [16, 18], "posit": [16, 17, 18, 20, 21, 23, 24, 25, 26, 28, 29, 48, 49, 53, 54, 57, 58, 59, 71, 72, 74], "constraint": [16, 18, 46, 54, 77], "110": [16, 18], "positiveconstraint": [16, 18], "126": [16, 18], "l": [16, 18, 56, 65, 73, 74, 76], "bfg": [16, 18, 65], "smooth": [16, 18, 25, 71], "165": [16, 18], "173": [16, 18], "131": [16, 18], "float32": [16, 18], "grouplasso": [17, 18], "228": [17, 18], "tutori": [17, 18, 68, 69, 71, 73, 74, 76], "221": [17, 18], "compat": [17, 18, 68, 69], "137": [17, 18], "258": [17, 18], "mcpregress": [17, 18, 24, 29], "184": [17, 18], "move": [17, 18], "specif": [17, 18, 69, 71], "method": [17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72, 75], "eas": [17, 18], "192": [17, 18], "logsumpenalti": [17, 18], "127": [17, 18], "abstract": [17, 18], "basedatafit": [17, 18, 22, 40, 71], "basepenalti": [17, 18, 22, 40, 72], "make": [17, 18, 19, 68, 69], "easier": [17, 18], "205": [17, 18], "fix": [17, 18, 19, 41], "point": [17, 18, 41, 76], "build": [17, 18, 19, 21, 23, 24, 25, 26, 27, 29, 60, 67, 68], "138": [17, 18], "continu": 19, "endeavour": 19, "reli": 19, "commun": 19, "effort": 19, "evolv": 19, "welcom": 19, "highli": [19, 69], "valuabl": 19, "you": [19, 68, 69, 71, 72], "help": 19, "bug": 19, "report": 19, "unit": [19, 73, 76], "codebas": [19, 72], "prevent": 19, "tighten": 19, "ani": [19, 68, 69, 71, 72, 76], "encount": 19, "issu": 19, "section": [19, 68, 69, 73, 74], "request": 19, "constantli": 19, "improv": 19, "would": [19, 68, 74], "like": [19, 30, 68, 71, 74], "align": [19, 73], "user": [19, 69, 72], "appreci": 19, "suggest": 19, "pull": 19, "mai": [19, 27, 75], "typo": 19, "document": [19, 65, 71, 72], "submit": 19, "integr": 19, "chang": 19, "reach": [19, 21, 22, 23, 24, 25, 26, 27, 28, 29, 76], "shortli": 19, "If": [19, 22, 23, 26, 27, 29, 30, 40, 41, 62, 63, 67, 68, 72, 76], "refer": [19, 20, 30, 40, 41, 56, 57, 60, 61, 64, 65, 67], "guid": 19, "part": [19, 23, 26, 29], "contrib": 19, "organ": 19, "adopt": 19, "conduct": 19, "willing": 19, "below": 19, "instal": [19, 68], "develop": 19, "version": [19, 69], "here": [19, 68, 71, 74], "step": [19, 41, 69], "want": 19, "fork": 19, "repositori": 19, "command": 19, "clone": 19, "sure": [19, 68], "your_github_usernam": 19, "usernam": 19, "git": 19, "http": [19, 40, 41, 56, 60, 61, 64, 67, 76], "cd": [19, 21, 23, 24, 25, 26, 27, 29, 41, 60], "directori": 19, "edit": 19, "pip": [19, 68, 69], "e": [19, 23, 25, 29, 56, 72, 73, 74], "doc": 19, "without": 19, "noplot": 19, "efron": [20, 30], "0001": [20, 21, 23, 24, 25, 26, 27, 28, 29, 43, 60, 61, 62, 63, 64, 65, 67], "breslow": 20, "mathemat": [20, 30], "behind": [20, 30], "detail": [20, 26, 30, 57, 65, 72, 73], "about": [20, 68, 74, 77], "express": 20, "convent": 20, "design": [20, 68, 69, 71, 75], "event": [20, 74], "censor": [20, 74], "lbfg": 20, "otherwis": [20, 21, 23, 24, 25, 26, 27, 28, 29, 62, 74, 76], "paramet": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40, 41, 42, 43, 57, 65, 71, 72, 73], "float": [20, 21, 23, 24, 25, 26, 27, 28, 29, 32, 41, 42, 43, 57, 60, 61, 62, 63, 64, 65, 67, 71], "option": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 40, 41, 57, 60, 67, 71, 72], "must": [20, 25, 28, 42, 71, 72, 76], "strictli": [20, 25, 76], "default": [20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 61, 62, 63, 64, 65, 67, 72], "mix": [20, 21], "criterion": [20, 21, 23, 24, 25, 26, 27, 28, 29, 72, 75], "maximum": [20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "bool": [20, 21, 23, 24, 25, 26, 27, 28, 29, 30, 41, 43, 57, 60, 61, 62, 63, 64, 65, 67], "amount": [20, 21, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "attribut": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 34, 36, 37, 39, 40, 56, 57, 61, 62, 63, 71, 72], "stop_crit_": 20, "converg": [20, 43, 61, 62, 63, 64, 65, 67, 75], "__init__": [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 71, 72], "max_epoch": [21, 23, 24, 25, 26, 27, 28, 29, 41, 60, 63, 66], "50000": [21, 23, 24, 25, 26, 27, 29, 60, 66], "p0": [21, 23, 24, 25, 26, 27, 29, 41, 43, 60, 63, 64, 66, 67], "xx": [21, 23, 24, 25, 26, 27, 29, 36, 37, 39, 57, 59, 62], "n_": [21, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 39, 45, 46, 52, 53, 55, 57, 59, 62, 71], "xw": [21, 24, 26, 27, 29, 31, 32, 33, 34, 35, 36, 37, 39, 41, 43, 44, 62, 71], "_2": [21, 23, 24, 25, 26, 27, 29, 36, 37, 38, 43, 44, 50, 73, 75, 76], "tt": 21, "_1": [21, 24, 28, 29, 43, 72, 73], "subproblem": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 64, 67], "definit": [21, 23, 24, 25, 26, 27, 28, 29, 60, 71, 74], "epoch": [21, 23, 24, 25, 26, 27, 29, 41, 60, 62, 63], "each": [21, 23, 24, 25, 26, 27, 28, 29, 34, 39, 41, 43, 60, 64, 67, 74, 75], "forc": [21, 23, 24, 26, 29, 57], "whether": [21, 23, 24, 25, 26, 27, 28, 29, 60], "reus": [21, 23, 24, 25, 26, 27, 28, 29], "call": [21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 40], "just": [21, 23, 24, 25, 26, 27, 28, 29, 68, 73], "eras": [21, 23, 24, 25, 26, 27, 28, 29], "str": [21, 23, 24, 25, 26, 27, 29], "cost": [21, 22, 23, 24, 25, 26, 27, 29, 40, 74], "formula": [21, 22, 23, 24, 25, 26, 27, 29, 40, 74, 76], "sparse_coef_": [21, 22, 24, 25, 26, 27, 29], "scipi": [21, 22, 24, 25, 26, 27, 29, 65], "readonli": [21, 22, 24, 25, 26, 27, 29], "deriv": [21, 22, 24, 25, 26, 27, 29, 57, 71, 74, 75, 76], "intercept_": [21, 22, 23, 24, 25, 26, 27, 28, 29], "constant": [21, 22, 23, 24, 25, 26, 27, 28, 29, 34, 75], "term": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 74], "decis": [21, 22, 23, 24, 25, 26, 27, 28, 29], "n_iter_": [21, 22, 23, 24, 25, 26, 27, 28, 29], "specifi": [21, 22, 23, 24, 25, 26, 27, 28, 29, 71, 72], "toler": [21, 22, 23, 24, 25, 26, 27, 28, 29, 41, 43, 60, 61, 62, 63, 64, 65, 67], "linear": [22, 26, 61, 64, 67, 68, 69, 71, 74, 75], "classif": [22, 38], "task": [22, 38, 66], "instanc": [22, 40, 69, 71, 72, 74], "jit": [22, 31, 33, 35, 38, 40, 71, 72], "basesolv": [22, 40], "n_task": 22, "celer": [23, 24, 27, 29, 64, 67], "extrapol": [23, 24, 29, 62, 64, 67], "sum_g": 23, "w_": [23, 57], "g": [23, 41, 57, 60, 69, 72, 73], "th": [23, 41, 45, 57, 71, 72], "partit": 23, "contigu": [23, 34, 37, 57], "being": [23, 57, 69, 72], "contain": [23, 56, 68], "equal": [23, 26, 29, 36, 39, 74, 75], "50_000": 23, "defautl": 23, "note": [23, 29, 31, 33, 35, 38, 45, 46, 53, 55, 59, 71, 72, 75], "sparser": [24, 29], "hing": 25, "sum_": [25, 26, 28, 31, 32, 33, 34, 35, 39, 45, 52, 53, 55, 57, 59, 71, 72, 74, 75, 76], "y_i": [25, 28, 31, 32, 33, 34, 35, 39, 42, 71, 74, 75], "beta": [25, 71, 72, 74, 75], "stai": 25, "our": [25, 71], "framework": 25, "svc": [25, 38], "yx": [25, 38], "w_i": 25, "relat": 25, "given": [25, 28, 68, 72], "invers": [25, 61], "proport": 25, "dual_": 25, "w_j": [26, 41, 45, 52, 53, 55, 59, 76], "see": 26, "algorithm": [26, 41, 61, 64, 67, 74], "nonconvex": 26, "applic": 26, "biolog": 26, "breheni": 26, "huang": 26, "hard": 26, "threshold": [26, 32, 61, 72], "soft": [26, 72], "should": [26, 62, 71, 72, 75], "larger": 26, "copy_x": 27, "multipli": 27, "l21": 27, "copi": 27, "els": [27, 72], "overwritten": 27, "log": [28, 31, 33, 34, 52, 74, 75], "exp": [28, 31, 33, 34, 35, 71, 75], "x_i": [28, 74], "outer": [28, 43, 64, 67], "classes_": 28, "ndarrai": 28, "n_class": 28, "A": [28, 38, 41, 60, 61, 62, 64, 65, 67, 69, 73, 74, 76], "known": [28, 72], "Not": 28, "yet": 28, "unweight": 29, "t_indic": 30, "stack": [30, 34, 37, 57], "horizont": 30, "group_1": 30, "group_2": 30, "ascend": 30, "order": [30, 74], "initialize_spars": [30, 71], "t_indptr": 30, "element": [30, 34, 37, 57, 72, 74], "delimit": [30, 34, 37, 57], "h_indic": 30, "uncensor": [30, 74], "h_indptr": 30, "read": [31, 32, 33, 34, 35, 36, 37, 38, 39, 42, 44, 50, 52, 57, 71, 72, 74], "_i": [31, 32, 33, 34, 35, 39, 71, 72], "faster": [31, 33, 35, 38, 41, 60, 69, 71], "delta": [32, 75], "hyperparamet": [32, 72], "grp1_indic": [34, 37, 57], "grp2_indic": [34, 37, 57], "pointer": [34, 37, 57], "lipschitz": [34, 75], "xty": 36, "pre": [36, 39, 71], "quantiti": [36, 39, 71, 76], "dure": [36, 39], "gradient": [36, 39, 62, 71, 72, 73, 75], "evalu": [36, 39, 71, 74], "sample_weight": 39, "weights_i": 39, "xtwy": 39, "samples_weight": 39, "single_task": 40, "n_reweight": 40, "surrog": 40, "involv": [40, 74], "cand\u00e8": 40, "et": 40, "al": 40, "2007": [40, 74], "enhanc": 40, "sparsiti": [40, 73], "minim": [40, 62, 73, 76], "web": 40, "stanford": 40, "edu": 40, "boyd": 40, "paper": 40, "pdf": 40, "rwl1": 40, "loss_history_": 40, "histori": 40, "everi": [40, 56], "dual_init": 41, "06": [41, 66], "min_w": 41, "saddl": 41, "max_z": 41, "z": [41, 71, 75], "fenchel": [41, 76], "conjug": [41, 76], "requir": [41, 71, 72], "built": 41, "assign": 41, "vari": 41, "updat": [41, 62, 75], "_j": [41, 59], "tau_j": 41, "g_j": [41, 75], "x_j": [41, 74, 76], "associ": [41, 74, 75], "equival": [41, 56], "variabl": [41, 56, 76], "thei": [41, 71], "silent": [41, 43, 60, 61, 62, 63, 64, 65, 67], "olivi": [41, 73], "fercoq": [41, 73], "pascal": 41, "bianchi": 41, "possibli": 41, "nonsepar": 41, "siam": [41, 61], "journal": [41, 74], "2020": [41, 64, 67], "epub": [41, 61], "doi": [41, 56, 61], "1137": [41, 61], "18m1168480": 41, "badr": 41, "moufad": 41, "q": [41, 60, 62, 69], "p": [41, 60, 69, 71, 72, 74, 75, 76], "gidel": [41, 60, 69], "beyond": [41, 60, 69], "better": [41, 60, 69], "neurip": [41, 69], "2022": [41, 60, 69], "arxiv": [41, 60, 64, 67], "2204": [41, 60], "07826": [41, 60], "quantile_level": 42, "sum_i": 42, "xw_i": 42, "quantil": 42, "level": [42, 56, 72], "becom": 42, "least": [42, 72, 73], "absolut": [42, 45, 55, 75], "deviat": [42, 45, 55], "lad": 42, "max_pn_it": [43, 64, 67], "minimum": [43, 63, 64, 67], "includ": [43, 63, 64, 67], "unnorm": 44, "smoothli": [45, 55], "clip": [45, 55], "With": [45, 53, 55, 59, 69, 71], "row": [45, 74], "math": [45, 71, 72, 73, 76], "box": 46, "bb": [46, 72, 75], "quasi": [47, 51], "ell_2": [49, 50], "aka": 49, "ep": 52, "epsilon": 52, "sort": [56, 74], "One": 56, "bogdan": 56, "van": 56, "den": 56, "berg": 56, "sabatti": 56, "su": 56, "cand": 56, "via": [56, 69], "annal": 56, "appli": [56, 71, 72, 74], "statist": [56, 74], "1103": 56, "40": 56, "1214": 56, "aoas842": 56, "singl": 56, "_g": 57, "i_": [57, 76], "geq": [57, 73, 74, 76], "orthant": 57, "subdifferenti": [57, 72, 73], "resolut": 60, "restrict": [60, 76], "aistat": 60, "2021": 60, "proceed": [60, 64, 67], "mlr": [60, 64, 67], "press": [60, 64, 67], "v130": 60, "bertrand21a": 60, "mathurinm": [60, 64, 67], "opt_strategi": 61, "ista": 61, "nesterov": 61, "beck": 61, "teboul": 61, "shrinkag": 61, "2009": 61, "imag": 61, "sci": 61, "080716542": 61, "use_acc": [62, 66], "greedy_cd": 62, "date": 62, "come": 62, "suit": 62, "rewritten": [62, 74], "w_init": [62, 63], "instead": [62, 63, 74], "past": 62, "greedi": 62, "cyclic": 62, "xw_init": 63, "vaiter": [64, 67], "gramfort": [64, 67, 73], "salmon": [64, 67, 73], "jmlr": [64, 67], "1907": [64, 67], "05830": [64, 67], "johnson": [64, 67], "guestrin": [64, 67], "blitz": [64, 67], "principl": [64, 67], "meta": [64, 67], "icml": [64, 67], "2015": [64, 67], "v37": [64, 67], "johnson15": [64, 67], "tbjohn": [64, 67], "blitzl1": [64, 67], "wrapper": 65, "multi": 66, "provid": [68, 74], "starter": 68, "familiar": 68, "explor": [68, 77], "high": [68, 69], "dimension": [68, 75], "toi": 68, "greater": 68, "don": 68, "unseen": 68, "By": [68, 74, 76], "overcom": [68, 74], "drawback": 68, "creat": 68, "custom": [68, 69], "sequel": 68, "perfectli": 68, "experi": 68, "thank": [68, 76], "beforehand": 68, "alreadi": [68, 69], "pypi": [68, 69], "conda": [68, 69], "forg": [68, 69], "make_regress": 68, "Then": 68, "r\u00b2": 68, "subsequ": 68, "top": [68, 71, 74, 75], "sinc": [68, 72, 75, 76], "sever": [68, 69], "api": [68, 69], "inform": 68, "avail": [68, 69], "suppos": 68, "latter": [68, 74], "outlier": 68, "mitig": 68, "effect": 68, "ideal": [68, 74], "regressor": 68, "offer": [68, 69], "flexibl": [68, 69], "compos": [68, 69, 75], "simpl": [68, 69], "possibl": 68, "huberregressor": 68, "befor": [68, 71, 76], "wai": 68, "glm": [68, 69], "argument": 68, "arbitrari": [68, 71, 72], "sake": 68, "modul": 68, "appropri": 68, "your": [68, 69, 72], "own": [68, 69, 72], "further": 68, "advanc": [68, 69], "topic": [68, 69], "hand": [68, 69, 74, 76, 77], "wide": 69, "choos": [69, 73, 77], "glimps": 69, "conceiv": 69, "There": 69, "opt": 69, "among": 69, "abl": 69, "tackl": 69, "dens": 69, "million": 69, "friendli": 69, "enabl": [69, 74], "exist": 69, "extens": 69, "easi": 69, "matter": 69, "line": 69, "drop": 69, "latest": 69, "cover": 69, "persever": 69, "research": 69, "licens": 69, "under": [69, 73], "bsd": 69, "claus": 69, "free": 69, "inproceed": 69, "booktitl": 69, "year": 69, "motiv": 71, "limit": [71, 74], "form": [71, 74], "hat": 71, "arg": 71, "min_": [71, 76], "mathbb": [71, 72, 73, 74], "omega": 71, "n": [71, 73, 74, 75, 76], "f_i": 71, "omega_j": 71, "beta_j": [71, 72, 75], "denot": [71, 75, 76], "whose": 71, "explicitli": 71, "mydatafit": 71, "mypenalti": 71, "jitclass": [71, 72], "inherit": [71, 72], "get_spec": [71, 72], "self": [71, 72], "type": [71, 72], "spec": [71, 72], "tupl": [71, 72], "attribute_nam": [71, 72], "params_to_dict": [71, 72], "dict_of_param": [71, 72], "dict": [71, 72], "instanti": [71, 72], "target": [71, 72], "x_data": 71, "x_indptr": 71, "x_indic": 71, "n_element": 71, "csc": 71, "indptr": 71, "array_lik": 71, "These": 71, "overload": 71, "suffix": 71, "_spars": 71, "hessian": 71, "frac": [71, 73, 74, 76], "simplifi": 71, "z_i": 71, "qquad": 71, "nabla": [71, 73, 74, 75], "underbrac": [71, 75], "raw": 71, "grad": [71, 72], "besid": [71, 72], "directli": [71, 75], "leq": [71, 73, 76], "diag": [71, 74], "y_iz_i": 71, "right": [71, 74], "therefor": [71, 74, 76], "raw_grad": 71, "raw_hessian": 71, "But": 71, "gradient_scalar": 71, "respect": [71, 74, 75, 76], "partial": [71, 72, 73, 76], "give": [71, 72, 75], "rais": 71, "valueerror": 71, "full_grad_spars": 71, "gradient_scalar_spars": 71, "idx_i": 71, "intercept_update_step": [71, 75], "usual": [71, 75], "serv": 71, "effici": [71, 74], "subclass": 72, "is_pen": 72, "mask": 72, "generalized_support": 72, "Theses": 72, "control": 72, "constructor": 72, "strongli": 72, "thu": [72, 73, 76], "bool_": 72, "addition": 72, "helper": 72, "dictionari": 72, "st": 72, "sgn": 72, "expect": 72, "stepsiz": 72, "turn": 72, "condit": [72, 73, 75, 76], "subdiff_dist": 72, "neg": [72, 74, 76], "nabla_j": 72, "typic": [72, 74], "might": 72, "smallest": 72, "null": 72, "reserv": 72, "keyword": 72, "correspond": 72, "togeth": [72, 76], "elif": 72, "grad_j": 72, "infti": [72, 73, 76], "gradient0": [72, 73], "frequent": 73, "fraction": 73, "consid": [73, 74], "min_x": 73, "d": [73, 76], "differenti": 73, "aim": 73, "determin": 73, "fermat": 73, "rule": [73, 74, 75], "infin": 73, "ball": [73, 76], "cdot": [73, 76], "begin": 73, "equat": [73, 74, 75, 76, 77], "leftrightarrow": 73, "ordinari": 73, "2n": [73, 75], "At": [73, 76], "proposit": 73, "particular": [73, 74, 77], "eugen": 73, "ndiay": 73, "alexandr": 73, "joseph": 73, "2017": 73, "gap": 73, "safe": 73, "screen": 73, "enforc": 73, "mach": 73, "re": 73, "18": 73, "januari": 73, "4671": 73, "4703": 73, "present": [74, 76], "mathbf": 74, "occur": 74, "interest": 74, "establish": 74, "likelihood": 74, "s_i": 74, "langl": [74, 76], "rangl": [74, 76], "y_j": 74, "compact": 74, "introduc": [74, 76], "wherea": 74, "write": [74, 75, 76], "wise": 74, "reformul": 74, "depend": 74, "henc": [74, 75], "focu": 74, "chain": 74, "divis": 74, "costli": [74, 75], "becaus": 74, "side": 74, "mathcal": [74, 76], "diagon": 74, "upper": 74, "bound": 74, "construct": 74, "semi": 74, "preceq": 74, "inequ": [74, 75], "loewner": 74, "matric": 74, "refin": 74, "ident": 74, "i_1": 74, "ldot": [74, 76], "i_m": 74, "h_": 74, "i_l": 74, "stand": 74, "cardin": 74, "rewrit": 74, "leverag": 74, "closer": [74, 76], "look": 74, "reveal": 74, "what": 74, "distinguish": 74, "presenc": 74, "doubl": 74, "cup_": 74, "fuse": 74, "minu": 74, "within": [74, 75], "a_": 74, "a_i": 74, "chosen": 74, "accordingli": 74, "preform": 74, "procedur": 74, "v": [74, 76], "similar": [74, 76], "subtract": 74, "complex": 74, "despit": 74, "multipl": 74, "due": 74, "special": 74, "structur": 74, "dy": 74, "lin": 74, "lifetim": 74, "471": 74, "480": 74, "bradlei": 74, "american": 74, "359": 74, "557": 74, "565": 74, "1977": 74, "insight": 75, "guidanc": 75, "rr": 75, "beta_0": 75, "liblinear": 75, "extra": 75, "memori": 75, "rout": 75, "star": [75, 76], "underset": 75, "phi": 75, "triangleq": 75, "_n": 75, "k": 75, "l_0": 75, "nabla_": 75, "local": 75, "statisfi": 75, "foral": [75, 76], "x_0": 75, "h": [75, 76], "moreov": 75, "nabla_beta": 75, "three": 75, "eq": 75, "4n": 75, "f_": 75, "f_delta": 75, "r_i": 75, "bbb": 75, "classic": 75, "l_2": 76, "nonneg": [76, 77], "mapsto": 76, "legendr": 76, "proj": 76, "moreau": 76, "decomposit": 76, "subdiff_diff": 76, "compon": 76, "quad": 76, "iota_": 76, "over": 76, "intuit": 76, "clear": 76, "v_i": 76, "cancel": 76, "n_i": 76, "u_i": 76, "bring": 76, "rigor": 76, "lagrangian": 76, "nu": 76, "mu": 76, "down": 76, "treat": 76, "complementari": 76, "slack": 76, "conclus": 76, "c_1": 76, "c_g": 76, "c_j": 76, "project": 76, "onto": 76, "p_j": 76, "v_j": 76, "stackexchang": 76, "2887332": 76, "167258": 76, "poisson": 77}, "objects": {"skglm": [[20, 0, 1, "", "CoxEstimator"], [21, 0, 1, "", "ElasticNet"], [22, 0, 1, "", "GeneralizedLinearEstimator"], [23, 0, 1, "", "GroupLasso"], [24, 0, 1, "", "Lasso"], [25, 0, 1, "", "LinearSVC"], [26, 0, 1, "", "MCPRegression"], [27, 0, 1, "", "MultiTaskLasso"], [28, 0, 1, "", "SparseLogisticRegression"], [29, 0, 1, "", "WeightedLasso"]], "skglm.CoxEstimator": [[20, 1, 1, "", "__init__"]], "skglm.ElasticNet": [[21, 1, 1, "", "__init__"]], "skglm.GeneralizedLinearEstimator": [[22, 1, 1, "", "__init__"]], "skglm.GroupLasso": [[23, 1, 1, "", "__init__"]], "skglm.Lasso": [[24, 1, 1, "", "__init__"]], "skglm.LinearSVC": [[25, 1, 1, "", "__init__"]], "skglm.MCPRegression": [[26, 1, 1, "", "__init__"]], "skglm.MultiTaskLasso": [[27, 1, 1, "", "__init__"]], "skglm.SparseLogisticRegression": [[28, 1, 1, "", "__init__"]], "skglm.WeightedLasso": [[29, 1, 1, "", "__init__"]], "skglm.datafits": [[30, 0, 1, "", "Cox"], [31, 0, 1, "", "Gamma"], [32, 0, 1, "", "Huber"], [33, 0, 1, "", "Logistic"], [34, 0, 1, "", "LogisticGroup"], [35, 0, 1, "", "Poisson"], [36, 0, 1, "", "Quadratic"], [37, 0, 1, "", "QuadraticGroup"], [38, 0, 1, "", "QuadraticSVC"], [39, 0, 1, "", "WeightedQuadratic"]], "skglm.datafits.Cox": [[30, 1, 1, "", "__init__"]], "skglm.datafits.Gamma": [[31, 1, 1, "", "__init__"]], "skglm.datafits.Huber": [[32, 1, 1, "", "__init__"]], "skglm.datafits.Logistic": [[33, 1, 1, "", "__init__"]], "skglm.datafits.LogisticGroup": [[34, 1, 1, "", "__init__"]], "skglm.datafits.Poisson": [[35, 1, 1, "", "__init__"]], "skglm.datafits.Quadratic": [[36, 1, 1, "", "__init__"]], "skglm.datafits.QuadraticGroup": [[37, 1, 1, "", "__init__"]], "skglm.datafits.QuadraticSVC": [[38, 1, 1, "", "__init__"]], "skglm.datafits.WeightedQuadratic": [[39, 1, 1, "", "__init__"]], "skglm.experimental": [[40, 0, 1, "", "IterativeReweightedL1"], [41, 0, 1, "", "PDCD_WS"], [42, 0, 1, "", "Pinball"], [43, 0, 1, "", "SqrtLasso"], [44, 0, 1, "", "SqrtQuadratic"]], "skglm.experimental.IterativeReweightedL1": [[40, 1, 1, "", "__init__"]], "skglm.experimental.PDCD_WS": [[41, 1, 1, "", "__init__"]], "skglm.experimental.Pinball": [[42, 1, 1, "", "__init__"]], "skglm.experimental.SqrtLasso": [[43, 1, 1, "", "__init__"]], "skglm.experimental.SqrtQuadratic": [[44, 1, 1, "", "__init__"]], "skglm.penalties": [[45, 0, 1, "", "BlockSCAD"], [46, 0, 1, "", "IndicatorBox"], [47, 0, 1, "", "L0_5"], [48, 0, 1, "", "L1"], [49, 0, 1, "", "L1_plus_L2"], [50, 0, 1, "", "L2"], [51, 0, 1, "", "L2_3"], [52, 0, 1, "", "LogSumPenalty"], [53, 0, 1, "", "MCPenalty"], [54, 0, 1, "", "PositiveConstraint"], [55, 0, 1, "", "SCAD"], [56, 0, 1, "", "SLOPE"], [57, 0, 1, "", "WeightedGroupL2"], [58, 0, 1, "", "WeightedL1"], [59, 0, 1, "", "WeightedMCPenalty"]], "skglm.penalties.BlockSCAD": [[45, 1, 1, "", "__init__"]], "skglm.penalties.IndicatorBox": [[46, 1, 1, "", "__init__"]], "skglm.penalties.L0_5": [[47, 1, 1, "", "__init__"]], "skglm.penalties.L1": [[48, 1, 1, "", "__init__"]], "skglm.penalties.L1_plus_L2": [[49, 1, 1, "", "__init__"]], "skglm.penalties.L2": [[50, 1, 1, "", "__init__"]], "skglm.penalties.L2_3": [[51, 1, 1, "", "__init__"]], "skglm.penalties.LogSumPenalty": [[52, 1, 1, "", "__init__"]], "skglm.penalties.MCPenalty": [[53, 1, 1, "", "__init__"]], "skglm.penalties.PositiveConstraint": [[54, 1, 1, "", "__init__"]], "skglm.penalties.SCAD": [[55, 1, 1, "", "__init__"]], "skglm.penalties.SLOPE": [[56, 1, 1, "", "__init__"]], "skglm.penalties.WeightedGroupL2": [[57, 1, 1, "", "__init__"]], "skglm.penalties.WeightedL1": [[58, 1, 1, "", "__init__"]], "skglm.penalties.WeightedMCPenalty": [[59, 1, 1, "", "__init__"]], "skglm.solvers": [[60, 0, 1, "", "AndersonCD"], [61, 0, 1, "", "FISTA"], [62, 0, 1, "", "GramCD"], [63, 0, 1, "", "GroupBCD"], [64, 0, 1, "", "GroupProxNewton"], [65, 0, 1, "", "LBFGS"], [66, 0, 1, "", "MultiTaskBCD"], [67, 0, 1, "", "ProxNewton"]], "skglm.solvers.AndersonCD": [[60, 1, 1, "", "__init__"]], "skglm.solvers.FISTA": [[61, 1, 1, "", "__init__"]], "skglm.solvers.GramCD": [[62, 1, 1, "", "__init__"]], "skglm.solvers.GroupBCD": [[63, 1, 1, "", "__init__"]], "skglm.solvers.GroupProxNewton": [[64, 1, 1, "", "__init__"]], "skglm.solvers.LBFGS": [[65, 1, 1, "", "__init__"]], "skglm.solvers.MultiTaskBCD": [[66, 1, 1, "", "__init__"]], "skglm.solvers.ProxNewton": [[67, 1, 1, "", "__init__"]]}, "objtypes": {"0": "py:class", "1": "py:method"}, "objnames": {"0": ["py", "class", "Python class"], "1": ["py", "method", "Python method"]}, "titleterms": {"api": 0, "document": 0, "estim": [0, 10, 68, 74], "penalti": [0, 5, 6, 9, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 72, 76, 77], "datafit": [0, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 71, 74, 75, 77], "solver": [0, 60, 61, 62, 63, 64, 65, 66, 67], "experiment": [0, 40, 41, 42, 43, 44], "exampl": [1, 73], "galleri": 1, "time": [2, 7, 10, 13, 70], "comparison": [2, 4, 7, 10], "scikit": 2, "learn": 2, "lasso": [2, 4, 8, 12, 24, 68, 76, 77], "group": [3, 8, 76, 77], "logist": [3, 5, 33, 75], "regress": [3, 5, 68], "python": [3, 8], "weight": [4, 12], "elast": 5, "net": 5, "minimax": 5, "concav": 5, "valu": 6, "proxim": [6, 76], "oper": [6, 76], "between": 7, "direct": 7, "prox": 7, "comput": [7, 13, 70, 75, 77], "reweight": 7, "fast": 8, "spars": [8, 9], "recoveri": 9, "non": 9, "convex": 9, "lifelin": 10, "skglm": [10, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 69], "surviv": 10, "analysi": 10, "data": [10, 68], "fit": [10, 68], "cox": [10, 30, 74, 77], "result": 10, "efron": [10, 74], "show": 11, "u": 11, "curv": 11, "regular": [11, 73, 77], "some": 12, "zero": 12, "version": [14, 15, 16, 17, 18], "0": [14, 15, 16, 17, 18, 73, 76, 77], "1": [14, 17, 18], "2022": [14, 15, 18], "04": [14, 18], "25": [14, 18], "2": [15, 18], "10": [15, 18], "18": [15, 18], "3": [16, 17, 18], "2023": [16, 17, 18], "07": [16, 18], "01": [16, 18], "4": [17, 18], "progress": [17, 18], "12": [17, 18], "21": [17, 18], "what": 18, "": 18, "new": 18, "contribut": 19, "setup": [19, 73, 74], "your": 19, "local": 19, "machin": 19, "coxestim": 20, "elasticnet": 21, "generalizedlinearestim": 22, "grouplasso": 23, "linearsvc": 25, "mcpregress": 26, "multitasklasso": 27, "sparselogisticregress": 28, "weightedlasso": 29, "gamma": 31, "huber": [32, 68, 75], "logisticgroup": 34, "poisson": [35, 71], "quadrat": [36, 75], "quadraticgroup": 37, "quadraticsvc": 38, "weightedquadrat": 39, "iterativereweightedl1": 40, "pdcd_w": 41, "pinbal": 42, "sqrtlasso": 43, "sqrtquadrat": 44, "blockscad": 45, "indicatorbox": 46, "l0_5": 47, "l1": [48, 72], "l1_plus_l2": 49, "l2": 50, "l2_3": 51, "logsumpenalti": 52, "mcpenalti": 53, "positiveconstraint": 54, "scad": 55, "slope": 56, "weightedgroupl2": 57, "weightedl1": 58, "weightedmcpenalti": 59, "andersoncd": 60, "fista": 61, "gramcd": 62, "groupbcd": 63, "groupproxnewton": 64, "lbfg": 65, "multitaskbcd": 66, "proxnewton": 67, "get": 68, "start": 68, "ell_1": 68, "gener": 68, "corrupt": 68, "why": 69, "instal": 69, "cite": 69, "how": [71, 72, 77], "add": [71, 72, 77], "custom": [71, 72, 77], "A": [71, 72], "case": [71, 72, 76], "point": [71, 72], "defin": [71, 72], "critic": [73, 77], "strength": [73, 77], "abov": [73, 77], "which": [73, 77], "solut": [73, 77], "i": [73, 77], "problem": [73, 74], "theoret": 73, "background": 73, "refer": [73, 74, 76], "mathemat": [74, 77], "behind": [74, 77], "breslow": 74, "express": 74, "gradient": 74, "hessian": 74, "intercept": [75, 77], "The": 75, "detail": [76, 77], "posit": 76, "constraint": 76, "subdifferenti": 76, "w": 76, "notin": 76, "mathbb": 76, "r": 76, "_": 76, "g": 76, "ne": 76, "tutori": 77}, "envversion": {"sphinx.domains.c": 3, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 9, "sphinx.domains.index": 1, "sphinx.domains.javascript": 3, "sphinx.domains.math": 2, "sphinx.domains.python": 4, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.intersphinx": 1, "sphinx": 58}, "alltitles": {"API Documentation": [[0, "api-documentation"]], "Estimators": [[0, "estimators"]], "Penalties": [[0, "penalties"]], "Datafits": [[0, "datafits"]], "Solvers": [[0, "solvers"]], "Experimental": [[0, "experimental"]], "Examples Gallery": [[1, "examples-gallery"]], "Timing comparison with scikit-learn for Lasso": [[2, "timing-comparison-with-scikit-learn-for-lasso"]], "Group Logistic regression in python": [[3, "group-logistic-regression-in-python"]], "Comparison of Lasso and Weighted Lasso": [[4, "comparison-of-lasso-and-weighted-lasso"]], "Logistic regression with Elastic net and minimax concave penalties": [[5, "logistic-regression-with-elastic-net-and-minimax-concave-penalties"]], "Value and proximal operators of penalties": [[6, "value-and-proximal-operators-of-penalties"]], "Timing comparison between direct prox computation and reweighting": [[7, "timing-comparison-between-direct-prox-computation-and-reweighting"]], "Fast Sparse Group Lasso in python": [[8, "fast-sparse-group-lasso-in-python"]], "Sparse recovery with non-convex penalties": [[9, "sparse-recovery-with-non-convex-penalties"]], "Comparison of lifelines with skglm for survival analysis": [[10, "comparison-of-lifelines-with-skglm-for-survival-analysis"]], "Data": [[10, "data"]], "Fitting the Cox Estimator": [[10, "fitting-the-cox-estimator"]], "Timing comparison": [[10, "timing-comparison"]], "Results": [[10, "results"]], "Efron estimate": [[10, "efron-estimate"], [74, "efron-estimate"]], "Show U-curve of regularization": [[11, "show-u-curve-of-regularization"]], "Weighted Lasso with some zero weights": [[12, "weighted-lasso-with-some-zero-weights"]], "Computation times": [[13, "computation-times"], [70, "computation-times"]], "Version 0.1 (2022/04/25)": [[14, "version-0-1-2022-04-25"], [18, "version-0-1-2022-04-25"]], "Version 0.2 (2022/10/18)": [[15, "version-0-2-2022-10-18"], [18, "version-0-2-2022-10-18"]], "Version 0.3 (2023/07/01)": [[16, "version-0-3-2023-07-01"], [18, "version-0-3-2023-07-01"]], "Version 0.4 (in progress)": [[17, "version-0-4-in-progress"], [18, "version-0-4-in-progress"]], "Version 0.3.1 (2023/12/21)": [[17, "version-0-3-1-2023-12-21"], [18, "version-0-3-1-2023-12-21"]], "What\u2019s new": [[18, "what-s-new"]], "Contribute to skglm": [[19, "contribute-to-skglm"]], "Setup skglm on your local machine": [[19, "setup-skglm-on-your-local-machine"]], "skglm.CoxEstimator": [[20, "skglm-coxestimator"]], "skglm.ElasticNet": [[21, "skglm-elasticnet"]], "skglm.GeneralizedLinearEstimator": [[22, "skglm-generalizedlinearestimator"]], "skglm.GroupLasso": [[23, "skglm-grouplasso"]], "skglm.Lasso": [[24, "skglm-lasso"]], "skglm.LinearSVC": [[25, "skglm-linearsvc"]], "skglm.MCPRegression": [[26, "skglm-mcpregression"]], "skglm.MultiTaskLasso": [[27, "skglm-multitasklasso"]], "skglm.SparseLogisticRegression": [[28, "skglm-sparselogisticregression"]], "skglm.WeightedLasso": [[29, "skglm-weightedlasso"]], "skglm.datafits.Cox": [[30, "skglm-datafits-cox"]], "skglm.datafits.Gamma": [[31, "skglm-datafits-gamma"]], "skglm.datafits.Huber": [[32, "skglm-datafits-huber"]], "skglm.datafits.Logistic": [[33, "skglm-datafits-logistic"]], "skglm.datafits.LogisticGroup": [[34, "skglm-datafits-logisticgroup"]], "skglm.datafits.Poisson": [[35, "skglm-datafits-poisson"]], "skglm.datafits.Quadratic": [[36, "skglm-datafits-quadratic"]], "skglm.datafits.QuadraticGroup": [[37, "skglm-datafits-quadraticgroup"]], "skglm.datafits.QuadraticSVC": [[38, "skglm-datafits-quadraticsvc"]], "skglm.datafits.WeightedQuadratic": [[39, "skglm-datafits-weightedquadratic"]], "skglm.experimental.IterativeReweightedL1": [[40, "skglm-experimental-iterativereweightedl1"]], "skglm.experimental.PDCD_WS": [[41, "skglm-experimental-pdcd-ws"]], "skglm.experimental.Pinball": [[42, "skglm-experimental-pinball"]], "skglm.experimental.SqrtLasso": [[43, "skglm-experimental-sqrtlasso"]], "skglm.experimental.SqrtQuadratic": [[44, "skglm-experimental-sqrtquadratic"]], "skglm.penalties.BlockSCAD": [[45, "skglm-penalties-blockscad"]], "skglm.penalties.IndicatorBox": [[46, "skglm-penalties-indicatorbox"]], "skglm.penalties.L0_5": [[47, "skglm-penalties-l0-5"]], "skglm.penalties.L1": [[48, "skglm-penalties-l1"]], "skglm.penalties.L1_plus_L2": [[49, "skglm-penalties-l1-plus-l2"]], "skglm.penalties.L2": [[50, "skglm-penalties-l2"]], "skglm.penalties.L2_3": [[51, "skglm-penalties-l2-3"]], "skglm.penalties.LogSumPenalty": [[52, "skglm-penalties-logsumpenalty"]], "skglm.penalties.MCPenalty": [[53, "skglm-penalties-mcpenalty"]], "skglm.penalties.PositiveConstraint": [[54, "skglm-penalties-positiveconstraint"]], "skglm.penalties.SCAD": [[55, "skglm-penalties-scad"]], "skglm.penalties.SLOPE": [[56, "skglm-penalties-slope"]], "skglm.penalties.WeightedGroupL2": [[57, "skglm-penalties-weightedgroupl2"]], "skglm.penalties.WeightedL1": [[58, "skglm-penalties-weightedl1"]], "skglm.penalties.WeightedMCPenalty": [[59, "skglm-penalties-weightedmcpenalty"]], "skglm.solvers.AndersonCD": [[60, "skglm-solvers-andersoncd"]], "skglm.solvers.FISTA": [[61, "skglm-solvers-fista"]], "skglm.solvers.GramCD": [[62, "skglm-solvers-gramcd"]], "skglm.solvers.GroupBCD": [[63, "skglm-solvers-groupbcd"]], "skglm.solvers.GroupProxNewton": [[64, "skglm-solvers-groupproxnewton"]], "skglm.solvers.LBFGS": [[65, "skglm-solvers-lbfgs"]], "skglm.solvers.MultiTaskBCD": [[66, "skglm-solvers-multitaskbcd"]], "skglm.solvers.ProxNewton": [[67, "skglm-solvers-proxnewton"]], "Getting started": [[68, "getting-started"]], "Fitting a Lasso estimator": [[68, "fitting-a-lasso-estimator"]], "Fitting \\ell_1 Huber regression": [[68, "fitting-ell-1-huber-regression"]], "Generate corrupt data": [[68, "generate-corrupt-data"]], "skglm": [[69, "skglm"]], "Why skglm?": [[69, "why-skglm"]], "Installing skglm": [[69, "installing-skglm"]], "Cite": [[69, "cite"]], "How to add a custom datafit": [[71, "how-to-add-a-custom-datafit"], [77, "how-to-add-a-custom-datafit"]], "A case in point: defining Poisson datafit": [[71, "a-case-in-point-defining-poisson-datafit"]], "How to add a custom penalty": [[72, "how-to-add-a-custom-penalty"], [77, "how-to-add-a-custom-penalty"]], "A case in point: defining L1 penalty": [[72, "a-case-in-point-defining-l1-penalty"]], "Critical regularization strength above which solution is 0": [[73, "critical-regularization-strength-above-which-solution-is-0"], [77, "critical-regularization-strength-above-which-solution-is-0"]], "Problem setup": [[73, "problem-setup"], [74, "problem-setup"]], "Theoretical background": [[73, "theoretical-background"]], "Example": [[73, "example"]], "References": [[73, "references"], [76, "references"]], "Mathematic behind Cox datafit": [[74, "mathematic-behind-cox-datafit"]], "Breslow estimate": [[74, "breslow-estimate"]], "Datafit expression": [[74, "datafit-expression"], [74, "id1"]], "Gradient and Hessian": [[74, "gradient-and-hessian"], [74, "id2"]], "Reference": [[74, "reference"]], "Computation of the intercept": [[75, "computation-of-the-intercept"], [77, "computation-of-the-intercept"]], "The Quadratic datafit": [[75, "the-quadratic-datafit"]], "The Logistic datafit": [[75, "the-logistic-datafit"]], "The Huber datafit": [[75, "the-huber-datafit"]], "Details on the Positive Group Lasso": [[76, "details-on-the-positive-group-lasso"]], "Proximity operator of the group Lasso": [[76, "proximity-operator-of-the-group-lasso"]], "Proximity operator of the group Lasso with positivity constraints": [[76, "proximity-operator-of-the-group-lasso-with-positivity-constraints"]], "Subdifferential of the positive Group Lasso penalty": [[76, "subdifferential-of-the-positive-group-lasso-penalty"]], "Case w \\notin \\mathbb{R}_+^g": [[76, "case-w-notin-mathbb-r-g"]], "Case w = 0": [[76, "case-w-0"]], "Case || w  || \\ne 0": [[76, "case-w-ne-0"]], "Tutorials": [[77, "tutorials"]], "Mathematics behind Cox datafit": [[77, "mathematics-behind-cox-datafit"]], "Details on the group Lasso": [[77, "details-on-the-group-lasso"]]}, "indexentries": {"coxestimator (class in skglm)": [[20, "skglm.CoxEstimator"]], "__init__() (skglm.coxestimator method)": [[20, "skglm.CoxEstimator.__init__"]], "elasticnet (class in skglm)": [[21, "skglm.ElasticNet"]], "__init__() (skglm.elasticnet method)": [[21, "skglm.ElasticNet.__init__"]], "generalizedlinearestimator (class in skglm)": [[22, "skglm.GeneralizedLinearEstimator"]], "__init__() (skglm.generalizedlinearestimator method)": [[22, "skglm.GeneralizedLinearEstimator.__init__"]], "grouplasso (class in skglm)": [[23, "skglm.GroupLasso"]], "__init__() (skglm.grouplasso method)": [[23, "skglm.GroupLasso.__init__"]], "lasso (class in skglm)": [[24, "skglm.Lasso"]], "__init__() (skglm.lasso method)": [[24, "skglm.Lasso.__init__"]], "linearsvc (class in skglm)": [[25, "skglm.LinearSVC"]], "__init__() (skglm.linearsvc method)": [[25, "skglm.LinearSVC.__init__"]], "mcpregression (class in skglm)": [[26, "skglm.MCPRegression"]], "__init__() (skglm.mcpregression method)": [[26, "skglm.MCPRegression.__init__"]], "multitasklasso (class in skglm)": [[27, "skglm.MultiTaskLasso"]], "__init__() (skglm.multitasklasso method)": [[27, "skglm.MultiTaskLasso.__init__"]], "sparselogisticregression (class in skglm)": [[28, "skglm.SparseLogisticRegression"]], "__init__() (skglm.sparselogisticregression method)": [[28, "skglm.SparseLogisticRegression.__init__"]], "weightedlasso (class in skglm)": [[29, "skglm.WeightedLasso"]], "__init__() (skglm.weightedlasso method)": [[29, "skglm.WeightedLasso.__init__"]], "cox (class in skglm.datafits)": [[30, "skglm.datafits.Cox"]], "__init__() (skglm.datafits.cox method)": [[30, "skglm.datafits.Cox.__init__"]], "gamma (class in skglm.datafits)": [[31, "skglm.datafits.Gamma"]], "__init__() (skglm.datafits.gamma method)": [[31, "skglm.datafits.Gamma.__init__"]], "huber (class in skglm.datafits)": [[32, "skglm.datafits.Huber"]], "__init__() (skglm.datafits.huber method)": [[32, "skglm.datafits.Huber.__init__"]], "logistic (class in skglm.datafits)": [[33, "skglm.datafits.Logistic"]], "__init__() (skglm.datafits.logistic method)": [[33, "skglm.datafits.Logistic.__init__"]], "logisticgroup (class in skglm.datafits)": [[34, "skglm.datafits.LogisticGroup"]], "__init__() (skglm.datafits.logisticgroup method)": [[34, "skglm.datafits.LogisticGroup.__init__"]], "poisson (class in skglm.datafits)": [[35, "skglm.datafits.Poisson"]], "__init__() (skglm.datafits.poisson method)": [[35, "skglm.datafits.Poisson.__init__"]], "quadratic (class in skglm.datafits)": [[36, "skglm.datafits.Quadratic"]], "__init__() (skglm.datafits.quadratic method)": [[36, "skglm.datafits.Quadratic.__init__"]], "quadraticgroup (class in skglm.datafits)": [[37, "skglm.datafits.QuadraticGroup"]], "__init__() (skglm.datafits.quadraticgroup method)": [[37, "skglm.datafits.QuadraticGroup.__init__"]], "quadraticsvc (class in skglm.datafits)": [[38, "skglm.datafits.QuadraticSVC"]], "__init__() (skglm.datafits.quadraticsvc method)": [[38, "skglm.datafits.QuadraticSVC.__init__"]], "weightedquadratic (class in skglm.datafits)": [[39, "skglm.datafits.WeightedQuadratic"]], "__init__() (skglm.datafits.weightedquadratic method)": [[39, "skglm.datafits.WeightedQuadratic.__init__"]], "iterativereweightedl1 (class in skglm.experimental)": [[40, "skglm.experimental.IterativeReweightedL1"]], "__init__() (skglm.experimental.iterativereweightedl1 method)": [[40, "skglm.experimental.IterativeReweightedL1.__init__"]], "pdcd_ws (class in skglm.experimental)": [[41, "skglm.experimental.PDCD_WS"]], "__init__() (skglm.experimental.pdcd_ws method)": [[41, "skglm.experimental.PDCD_WS.__init__"]], "pinball (class in skglm.experimental)": [[42, "skglm.experimental.Pinball"]], "__init__() (skglm.experimental.pinball method)": [[42, "skglm.experimental.Pinball.__init__"]], "sqrtlasso (class in skglm.experimental)": [[43, "skglm.experimental.SqrtLasso"]], "__init__() (skglm.experimental.sqrtlasso method)": [[43, "skglm.experimental.SqrtLasso.__init__"]], "sqrtquadratic (class in skglm.experimental)": [[44, "skglm.experimental.SqrtQuadratic"]], "__init__() (skglm.experimental.sqrtquadratic method)": [[44, "skglm.experimental.SqrtQuadratic.__init__"]], "blockscad (class in skglm.penalties)": [[45, "skglm.penalties.BlockSCAD"]], "__init__() (skglm.penalties.blockscad method)": [[45, "skglm.penalties.BlockSCAD.__init__"]], "indicatorbox (class in skglm.penalties)": [[46, "skglm.penalties.IndicatorBox"]], "__init__() (skglm.penalties.indicatorbox method)": [[46, "skglm.penalties.IndicatorBox.__init__"]], "l0_5 (class in skglm.penalties)": [[47, "skglm.penalties.L0_5"]], "__init__() (skglm.penalties.l0_5 method)": [[47, "skglm.penalties.L0_5.__init__"]], "l1 (class in skglm.penalties)": [[48, "skglm.penalties.L1"]], "__init__() (skglm.penalties.l1 method)": [[48, "skglm.penalties.L1.__init__"]], "l1_plus_l2 (class in skglm.penalties)": [[49, "skglm.penalties.L1_plus_L2"]], "__init__() (skglm.penalties.l1_plus_l2 method)": [[49, "skglm.penalties.L1_plus_L2.__init__"]], "l2 (class in skglm.penalties)": [[50, "skglm.penalties.L2"]], "__init__() (skglm.penalties.l2 method)": [[50, "skglm.penalties.L2.__init__"]], "l2_3 (class in skglm.penalties)": [[51, "skglm.penalties.L2_3"]], "__init__() (skglm.penalties.l2_3 method)": [[51, "skglm.penalties.L2_3.__init__"]], "logsumpenalty (class in skglm.penalties)": [[52, "skglm.penalties.LogSumPenalty"]], "__init__() (skglm.penalties.logsumpenalty method)": [[52, "skglm.penalties.LogSumPenalty.__init__"]], "mcpenalty (class in skglm.penalties)": [[53, "skglm.penalties.MCPenalty"]], "__init__() (skglm.penalties.mcpenalty method)": [[53, "skglm.penalties.MCPenalty.__init__"]], "positiveconstraint (class in skglm.penalties)": [[54, "skglm.penalties.PositiveConstraint"]], "__init__() (skglm.penalties.positiveconstraint method)": [[54, "skglm.penalties.PositiveConstraint.__init__"]], "scad (class in skglm.penalties)": [[55, "skglm.penalties.SCAD"]], "__init__() (skglm.penalties.scad method)": [[55, "skglm.penalties.SCAD.__init__"]], "slope (class in skglm.penalties)": [[56, "skglm.penalties.SLOPE"]], "__init__() (skglm.penalties.slope method)": [[56, "skglm.penalties.SLOPE.__init__"]], "weightedgroupl2 (class in skglm.penalties)": [[57, "skglm.penalties.WeightedGroupL2"]], "__init__() (skglm.penalties.weightedgroupl2 method)": [[57, "skglm.penalties.WeightedGroupL2.__init__"]], "weightedl1 (class in skglm.penalties)": [[58, "skglm.penalties.WeightedL1"]], "__init__() (skglm.penalties.weightedl1 method)": [[58, "skglm.penalties.WeightedL1.__init__"]], "weightedmcpenalty (class in skglm.penalties)": [[59, "skglm.penalties.WeightedMCPenalty"]], "__init__() (skglm.penalties.weightedmcpenalty method)": [[59, "skglm.penalties.WeightedMCPenalty.__init__"]], "andersoncd (class in skglm.solvers)": [[60, "skglm.solvers.AndersonCD"]], "__init__() (skglm.solvers.andersoncd method)": [[60, "skglm.solvers.AndersonCD.__init__"]], "fista (class in skglm.solvers)": [[61, "skglm.solvers.FISTA"]], "__init__() (skglm.solvers.fista method)": [[61, "skglm.solvers.FISTA.__init__"]], "gramcd (class in skglm.solvers)": [[62, "skglm.solvers.GramCD"]], "__init__() (skglm.solvers.gramcd method)": [[62, "skglm.solvers.GramCD.__init__"]], "groupbcd (class in skglm.solvers)": [[63, "skglm.solvers.GroupBCD"]], "__init__() (skglm.solvers.groupbcd method)": [[63, "skglm.solvers.GroupBCD.__init__"]], "groupproxnewton (class in skglm.solvers)": [[64, "skglm.solvers.GroupProxNewton"]], "__init__() (skglm.solvers.groupproxnewton method)": [[64, "skglm.solvers.GroupProxNewton.__init__"]], "lbfgs (class in skglm.solvers)": [[65, "skglm.solvers.LBFGS"]], "__init__() (skglm.solvers.lbfgs method)": [[65, "skglm.solvers.LBFGS.__init__"]], "multitaskbcd (class in skglm.solvers)": [[66, "skglm.solvers.MultiTaskBCD"]], "__init__() (skglm.solvers.multitaskbcd method)": [[66, "skglm.solvers.MultiTaskBCD.__init__"]], "proxnewton (class in skglm.solvers)": [[67, "skglm.solvers.ProxNewton"]], "__init__() (skglm.solvers.proxnewton method)": [[67, "skglm.solvers.ProxNewton.__init__"]]}})
\ No newline at end of file
diff --git a/sg_execution_times.html b/sg_execution_times.html
index c2f40ad72..bd084a22c 100644
--- a/sg_execution_times.html
+++ b/sg_execution_times.html
@@ -347,7 +347,7 @@
         <article role="main" id="furo-main-content">
           <section id="computation-times">
 <span id="sphx-glr-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this heading">¶</a></h1>
-<p><strong>02:40.942</strong> total execution time for 11 files <strong>from all galleries</strong>:</p>
+<p><strong>02:49.958</strong> total execution time for 11 files <strong>from all galleries</strong>:</p>
 <div class="docutils container">
 <style scoped>
 <link href="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/5.3.0/css/bootstrap.min.css" rel="stylesheet" />
@@ -370,47 +370,47 @@
 </thead>
 <tbody>
 <tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_survival_analysis.html#sphx-glr-auto-examples-plot-survival-analysis-py"><span class="std std-ref">Comparison of lifelines with skglm for survival analysis</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_survival_analysis.py</span></code>)</p></td>
-<td><p>00:52.615</p></td>
+<td><p>00:57.445</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_compare_time.html#sphx-glr-auto-examples-plot-compare-time-py"><span class="std std-ref">Timing comparison with scikit-learn for Lasso</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_compare_time.py</span></code>)</p></td>
-<td><p>00:45.604</p></td>
+<td><p>00:47.620</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_group_logistic_regression.html#sphx-glr-auto-examples-plot-group-logistic-regression-py"><span class="std std-ref">Group Logistic regression in python</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_group_logistic_regression.py</span></code>)</p></td>
-<td><p>00:14.785</p></td>
+<td><p>00:15.201</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_ucurve.html#sphx-glr-auto-examples-plot-ucurve-py"><span class="std std-ref">Show U-curve of regularization</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_ucurve.py</span></code>)</p></td>
-<td><p>00:13.278</p></td>
+<td><p>00:13.957</p></td>
 <td><p>0.0</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_sparse_recovery.html#sphx-glr-auto-examples-plot-sparse-recovery-py"><span class="std std-ref">Sparse recovery with non-convex penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_sparse_recovery.py</span></code>)</p></td>
-<td><p>00:08.910</p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_logreg_various_penalties.html#sphx-glr-auto-examples-plot-logreg-various-penalties-py"><span class="std std-ref">Logistic regression with Elastic net and minimax concave penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_logreg_various_penalties.py</span></code>)</p></td>
+<td><p>00:08.967</p></td>
 <td><p>0.0</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_logreg_various_penalties.html#sphx-glr-auto-examples-plot-logreg-various-penalties-py"><span class="std std-ref">Logistic regression with Elastic net and minimax concave penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_logreg_various_penalties.py</span></code>)</p></td>
-<td><p>00:08.793</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_sparse_recovery.html#sphx-glr-auto-examples-plot-sparse-recovery-py"><span class="std std-ref">Sparse recovery with non-convex penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_sparse_recovery.py</span></code>)</p></td>
+<td><p>00:08.881</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_sparse_group_lasso.html#sphx-glr-auto-examples-plot-sparse-group-lasso-py"><span class="std std-ref">Fast Sparse Group Lasso in python</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_sparse_group_lasso.py</span></code>)</p></td>
-<td><p>00:05.547</p></td>
+<td><p>00:05.796</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_reweighted_l1.html#sphx-glr-auto-examples-plot-reweighted-l1-py"><span class="std std-ref">Timing comparison between direct prox computation and reweighting</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_reweighted_l1.py</span></code>)</p></td>
-<td><p>00:05.021</p></td>
+<td><p>00:05.326</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_lasso_vs_weighted.html#sphx-glr-auto-examples-plot-lasso-vs-weighted-py"><span class="std std-ref">Comparison of Lasso and Weighted Lasso</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_lasso_vs_weighted.py</span></code>)</p></td>
-<td><p>00:03.731</p></td>
+<td><p>00:03.903</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="auto_examples/plot_pen_prox.html#sphx-glr-auto-examples-plot-pen-prox-py"><span class="std std-ref">Value and proximal operators of penalties</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_pen_prox.py</span></code>)</p></td>
-<td><p>00:02.314</p></td>
+<td><p>00:02.487</p></td>
 <td><p>0.0</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="auto_examples/plot_zero_weights_lasso.html#sphx-glr-auto-examples-plot-zero-weights-lasso-py"><span class="std std-ref">Weighted Lasso with some zero weights</span></a> (<code class="docutils literal notranslate"><span class="pre">../examples/plot_zero_weights_lasso.py</span></code>)</p></td>
-<td><p>00:00.342</p></td>
+<td><p>00:00.374</p></td>
 <td><p>0.0</p></td>
 </tr>
 </tbody>