Fixing multiplicity analysis workflow

machine_learning_hep/analysis/analyzerdhadrons_mult.py

@@ -62,10 +62,13 @@ def __init__(self, datap, case, typean, period):
                                   [None for _ in range(len(self.lvar2_binmin))])
         self.inel0_var = datap["analysis"][self.typean].get("inel0_var", "n_tracklets")
 
-        self.d_resultsallpmc = datap["analysis"][typean]["mc"]["results"][period] \
-                if period is not None else datap["analysis"][typean]["mc"]["resultsallp"]
-        self.d_resultsallpdata = datap["analysis"][typean]["data"]["results"][period] \
-                if period is not None else datap["analysis"][typean]["data"]["resultsallp"]
+        self.dp = datap["analysis"][typean]
+        self.d_prefix_mc = self.dp["mc"].get("prefix_dir_res")
+        self.d_prefix_data = self.dp["data"].get("prefix_dir_res")
+        self.d_resultsallpmc = self.d_prefix_mc + self.dp["mc"]["results"][period] \
+                if period is not None else self.d_prefix_mc + self.dp["mc"]["resultsallp"]
+        self.d_resultsallpdata = self.d_prefix_data + dp["data"]["results"][period] \
+                if period is not None else self.d_prefix_data + self.dp["data"]["resultsallp"]
 
         self.p_corrmb_typean = datap["analysis"][self.typean]["corresp_mb_typean"]
         if self.p_corrmb_typean is not None:
@@ -116,8 +119,7 @@ def __init__(self, datap, case, typean, period):
         if not isinstance(self.p_includesecpeaks[0], list):
             self.p_inculdesecpeaks = [self.p_includesecpeaks for _ in range(self.p_nbin2)]
 
-        self.p_masssecpeak = datap["analysis"][self.typean]["masssecpeak"] \
-                if self.p_includesecpeaks else None
+        self.p_masssecpeak = datap["analysis"][self.typean].get("masssecpeak", None)
 
         self.p_fix_masssecpeaks = datap["analysis"][self.typean].get("fix_masssecpeak", None)
         if self.p_fix_masssecpeaks is None:
@@ -127,10 +129,8 @@ def __init__(self, datap, case, typean, period):
         if not isinstance(self.p_fix_masssecpeaks[0], list):
             self.p_fix_masssecpeaks = [self.p_fix_masssecpeaks for _ in range(self.p_nbin2)]
 
-        self.p_widthsecpeak = datap["analysis"][self.typean]["widthsecpeak"] \
-                if self.p_includesecpeaks else None
-        self.p_fix_widthsecpeak = datap["analysis"][self.typean]["fix_widthsecpeak"] \
-                if self.p_includesecpeaks else None
+        self.p_widthsecpeak = datap["analysis"][self.typean].get("widthsecpeak", None)
+        self.p_fix_widthsecpeak = datap["analysis"][self.typean].get("fix_widthsecpeak", None)
         self.p_fixedmean = datap["analysis"][self.typean]["FixedMean"]
         self.p_use_user_gauss_sigma = datap["analysis"][self.typean]["SetInitialGaussianSigma"]
         self.p_max_perc_sigma_diff = datap["analysis"][self.typean]["MaxPercSigmaDeviation"]

machine_learning_hep/bitwise.py

@@ -32,18 +32,18 @@ def tag_bit_df(dfin, namebitmap, activatedbit):
     bitson = activatedbit[0]
     bitsoff = activatedbit[1]
     array_cand_type = dfin.loc[:, namebitmap].values.astype("int")
-    res_on = pd.Series([True]*len(array_cand_type))
-    res_off = pd.Series([True]*len(array_cand_type))
-    res = pd.Series(dtype = 'int')
+    res_on = pd.Series([True]*len(array_cand_type), dtype=int)
+    res_off = pd.Series([True]*len(array_cand_type), dtype=int)
+    res = pd.Series(dtype=int)
 
     if bitson:
         mask = reduce(operator.or_, ((1 << bit) for bit in bitson), 0)
         bitmapon = selectbiton(array_cand_type, mask)
-        res_on = pd.Series(bitmapon)
+        res_on = pd.Series(bitmapon, dtype=int)
     if bitsoff:
         mask = reduce(operator.or_, ((1 << bit) for bit in bitsoff), 0)
         bitmapoff = selectbitoff(array_cand_type, mask)
-        res_off = pd.Series(bitmapoff)
+        res_off = pd.Series(bitmapoff, dtype=object)
     res = res_on & res_off
     return res
 

machine_learning_hep/data/data_run3/database_ml_parameters_LcToPKPi.yml

@@ -18,7 +18,6 @@ LcpKpi:
   doml: true
   mass: 2.286
   sel_reco_unp: "fPt>0"
-  sel_reco_singletrac_unp : null
   sel_gen_unp: "fPt>0"
   sel_cen_unp: null
   sel_good_evt_unp: "fIsEventReject == 0"
@@ -60,8 +59,8 @@ LcpKpi:
     var_jet_match: [df, fIndexHfCand2Prong]
     var_jetsub_match: [df, fIndexD0ChargedJets]
     var_evt:
-       data: [fIndexCollisions, fIsEventReject, fNumContrib]
-       mc:   [fIndexCollisions, fIndexMcCollisions, fIsEventReject, fNumContrib]
+       data: [fIndexCollisions, fIsEventReject, fNumContrib, fMultZeqNTracksPV, fMultZeqFT0A, fMultZeqFT0C, fMultFT0M, fMultZeqFV0A]
+       mc:   [fIndexCollisions, fIndexMcCollisions, fIsEventReject, fNumContrib, fMultZeqNTracksPV, fMultZeqFT0A, fMultZeqFT0C, fMultFT0M, fMultZeqFV0A]
     var_gen: [fIndexMcCollisions, fPt, fY, fFlagMc, fOriginMcGen]
     var_evt_match: [df, fIndexCollisions]
     var_evt_match_mc: [df, fIndexMcCollisions]
@@ -202,7 +201,7 @@ LcpKpi:
       chunksizeskim: [100] #list of periods
       fracmerge : [1.0] #list of periods
       seedmerge: [12] #list of periods
-      period: [test] #list of periods
+      period: [LHC22b1b] #list of periods
       select_period: [1]
       prefix_dir: /data2/MLhep/
       unmerged_tree_dir: [sim/alice/cern.ch/user/a/alihyperloop/jobs/0024] #list of periods
@@ -286,7 +285,7 @@ LcpKpi:
 
     Run3analysis:
       proc_type: Dhadrons
-      useperiod: [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+      useperiod: [1]
       plotbin: [1]
       usesinglebineff: 0
       sel_binmin2: [1,2,4,6,8,12] #list of nbins
@@ -295,12 +294,12 @@ LcpKpi:
       triggerbit: ''
       use_cuts: False
       cuts:
-        - "fCPA > 0.95"
-        - "fCPA > 0.95"
-        - "fCPA > 0.95"
-        - "fCPA > 0.95"
-        - "fCPA > 0.95"
-        - "fCPA > 0.95"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
+        - "fDecayLength > 0.02 and fCpa > 0.9"
 
 
       # To initialize the individual fits in pT bins
@@ -318,7 +317,7 @@ LcpKpi:
         weighttrig: false
 
       data:
-        runselection: [null, null] #FIXME
+        runselection: [null] #FIXME
         prefix_dir_res: /data2/MLhep/
         results: [LHC22pp/Results/prod_LHC22o/resultsdata] #list of periods
         resultsallp: LHC22pp/Results/resultsdatatot
@@ -339,10 +338,9 @@ LcpKpi:
       rebin: [6,6,6,6,6,6]
       fix_mean: [false,false,false,false,false,false]
       fix_sigma: [false,false,false,false,false,false]
-      masssecpeak: 0.
       # Fix mean and/or sigma
       FixedMean: False
-      SetFixGaussianSigma: [true,true,true,true,true,true]
+      SetFixGaussianSigma: [false,false,false,false,false,false]
       # Use value set for "masspeak" for initializing total fit, otherwise what is derived from MC fit is used
       SetInitialGaussianMean: true
       # Use values set for "sigmaarray" for initializing total fit (per pT bin),