Backport PR matplotlib#26173: Synchronize mathtext docs and handling

doc/sphinxext/math_symbol_table.py

@@ -1,7 +1,11 @@
+import re
 from docutils.parsers.rst import Directive
 
 from matplotlib import _mathtext, _mathtext_data
 
+bb_pattern = re.compile("Bbb[A-Z]")
+scr_pattern = re.compile("scr[a-zA-Z]")
+frak_pattern = re.compile("frak[A-Z]")
 
 symbols = [
     ["Lower-case Greek",
@@ -18,6 +22,9 @@
     ["Hebrew",
      6,
      (r"\aleph", r"\beth", r"\gimel", r"\daleth")],
+    ["Latin named characters",
+     6,
+     r"""\aa \AA \ae \AE \oe \OE \O \o \thorn \Thorn \ss \eth \dh \DH""".split()],
     ["Delimiters",
      5,
      _mathtext.Parser._delims],
@@ -27,85 +34,56 @@
     ["Standard function names",
      5,
      {fr"\{fn}" for fn in _mathtext.Parser._function_names}],
-    ["Binary operation and relation symbols",
+    ["Binary operation symbols",
+     4,
+     _mathtext.Parser._binary_operators],
+    ["Relation symbols",
      4,
-     r"""\ast \pm \slash \cap \star \mp \cup \cdot \uplus
-     \triangleleft \circ \odot \sqcap \triangleright \bullet \ominus
-     \sqcup \bigcirc \oplus \wedge \diamond \oslash \vee
-     \bigtriangledown \times \otimes \dag \bigtriangleup \div \wr
-     \ddag \barwedge \veebar \boxplus \curlywedge \curlyvee \boxminus
-     \Cap \Cup \boxtimes \bot \top \dotplus \boxdot \intercal
-     \rightthreetimes \divideontimes \leftthreetimes \equiv \leq \geq
-     \perp \cong \prec \succ \mid \neq \preceq \succeq \parallel \sim
-     \ll \gg \bowtie \simeq \subset \supset \Join \approx \subseteq
-     \supseteq \ltimes \asymp \sqsubset \sqsupset \rtimes \doteq
-     \sqsubseteq \sqsupseteq \smile \propto \dashv \vdash \frown
-     \models \in \ni \notin \approxeq \leqq \geqq \lessgtr \leqslant
-     \geqslant \lesseqgtr \backsim \lessapprox \gtrapprox \lesseqqgtr
-     \backsimeq \lll \ggg \gtreqqless \triangleq \lessdot \gtrdot
-     \gtreqless \circeq \lesssim \gtrsim \gtrless \bumpeq \eqslantless
-     \eqslantgtr \backepsilon \Bumpeq \precsim \succsim \between
-     \doteqdot \precapprox \succapprox \pitchfork \Subset \Supset
-     \fallingdotseq \subseteqq \supseteqq \risingdotseq \sqsubset
-     \sqsupset \varpropto \preccurlyeq \succcurlyeq \Vdash \therefore
-     \curlyeqprec \curlyeqsucc \vDash \because \blacktriangleleft
-     \blacktriangleright \Vvdash \eqcirc \trianglelefteq
-     \trianglerighteq \neq \vartriangleleft \vartriangleright \ncong
-     \nleq \ngeq \nsubseteq \nmid \nsupseteq \nparallel \nless \ngtr
-     \nprec \nsucc \subsetneq \nsim \supsetneq \nVDash \precnapprox
-     \succnapprox \subsetneqq \nvDash \precnsim \succnsim \supsetneqq
-     \nvdash \lnapprox \gnapprox \ntriangleleft \ntrianglelefteq
-     \lneqq \gneqq \ntriangleright \lnsim \gnsim \ntrianglerighteq
-     \coloneq \eqsim \nequiv \napprox \nsupset \doublebarwedge \nVdash
-     \Doteq \nsubset \eqcolon \ne
-     """.split()],
+     _mathtext.Parser._relation_symbols],
     ["Arrow symbols",
      4,
-     r"""\leftarrow \longleftarrow \uparrow \Leftarrow \Longleftarrow
-     \Uparrow \rightarrow \longrightarrow \downarrow \Rightarrow
-     \Longrightarrow \Downarrow \leftrightarrow \updownarrow
-     \longleftrightarrow \updownarrow \Leftrightarrow
-     \Longleftrightarrow \Updownarrow \mapsto \longmapsto \nearrow
-     \hookleftarrow \hookrightarrow \searrow \leftharpoonup
-     \rightharpoonup \swarrow \leftharpoondown \rightharpoondown
-     \nwarrow \rightleftharpoons \leadsto \dashrightarrow
-     \dashleftarrow \leftleftarrows \leftrightarrows \Lleftarrow
-     \Rrightarrow \twoheadleftarrow \leftarrowtail \looparrowleft
-     \leftrightharpoons \curvearrowleft \circlearrowleft \Lsh
-     \upuparrows \upharpoonleft \downharpoonleft \multimap
-     \leftrightsquigarrow \rightrightarrows \rightleftarrows
-     \rightrightarrows \rightleftarrows \twoheadrightarrow
-     \rightarrowtail \looparrowright \rightleftharpoons
-     \curvearrowright \circlearrowright \Rsh \downdownarrows
-     \upharpoonright \downharpoonright \rightsquigarrow \nleftarrow
-     \nrightarrow \nLeftarrow \nRightarrow \nleftrightarrow
-     \nLeftrightarrow \to \Swarrow \Searrow \Nwarrow \Nearrow
-     \leftsquigarrow
-     """.split()],
+     _mathtext.Parser._arrow_symbols],
+    ["Dot symbols",
+     4,
+     r"""\cdots \vdots \ldots \ddots \adots \Colon \therefore \because""".split()],
+    ["Black-board characters",
+     6,
+     [fr"\{symbol}" for symbol in _mathtext_data.tex2uni
+      if re.match(bb_pattern, symbol)]],
+    ["Script characters",
+     6,
+     [fr"\{symbol}" for symbol in _mathtext_data.tex2uni
+      if re.match(scr_pattern, symbol)]],
+    ["Fraktur characters",
+     6,
+     [fr"\{symbol}" for symbol in _mathtext_data.tex2uni
+      if re.match(frak_pattern, symbol)]],
     ["Miscellaneous symbols",
      4,
      r"""\neg \infty \forall \wp \exists \bigstar \angle \partial
-     \nexists \measuredangle \eth \emptyset \sphericalangle \clubsuit
+     \nexists \measuredangle \emptyset \sphericalangle \clubsuit
      \varnothing \complement \diamondsuit \imath \Finv \triangledown
-     \heartsuit \jmath \Game \spadesuit \ell \hbar \vartriangle \cdots
-     \hslash \vdots \blacksquare \ldots \blacktriangle \ddots \sharp
+     \heartsuit \jmath \Game \spadesuit \ell \hbar \vartriangle
+     \hslash \blacksquare \blacktriangle \sharp \increment
      \prime \blacktriangledown \Im \flat \backprime \Re \natural
-     \circledS \P \copyright \ss \circledR \S \yen \AA \checkmark \$
-     \cent \triangle \QED \sinewave \nabla \mho""".split()]
+     \circledS \P \copyright \circledR \S \yen \checkmark \$
+     \cent \triangle \QED \sinewave \dag \ddag \perthousand \ac
+     \lambdabar \L \l \degree \danger \maltese \clubsuitopen
+     \i \hermitmatrix \sterling \nabla \mho""".split()],
 ]
 
 
 def run(state_machine):
 
     def render_symbol(sym, ignore_variant=False):
-        if ignore_variant and sym != r"\varnothing":
+        if ignore_variant and sym not in (r"\varnothing", r"\varlrtriangle"):
             sym = sym.replace(r"\var", "\\")
         if sym.startswith("\\"):
             sym = sym.lstrip("\\")
             if sym not in (_mathtext.Parser._overunder_functions |
                            _mathtext.Parser._function_names):
                 sym = chr(_mathtext_data.tex2uni[sym])
-        return f'\\{sym}' if sym in ('\\', '|') else sym
+        return f'\\{sym}' if sym in ('\\', '|', '+', '-', '*') else sym
 
     lines = []
     for category, columns, syms in symbols:
@@ -165,7 +143,10 @@ def setup(app):
                 if sym[1:] not in _mathtext_data.tex2uni:
                     print(sym)
 
+    # Add accents
+    all_symbols.update({v[1:]: k for k, v in _mathtext.Parser._accent_map.items()})
+    all_symbols.update({v: v for v in _mathtext.Parser._wide_accents})
     print("SYMBOLS NOT IN TABLE:")
-    for sym in _mathtext_data.tex2uni:
+    for sym, val in _mathtext_data.tex2uni.items():
         if sym not in all_symbols:
-            print(sym)
+            print(f"{sym} = {chr(val)}")

doc/users/next_whats_new/mathtext_documentation.rst

@@ -0,0 +1,5 @@
+``mathtext`` documentation improvements
+---------------------------------------
+
+The documentation is updated to take information directly from the parser. This
+means that (almost) all supported symbols, operators etc are shown at :ref:`mathtext`.

doc/users/next_whats_new/mathtext_relation_operators.rst

@@ -5,5 +5,3 @@ There has been a number of operators added and corrected when a Unicode font is
 In addition, correct spacing has been added to a number of the previous operators.
 Especially, the characters used for ``\gnapprox``, ``\lnapprox``, ``\leftangle``, and
 ``\rightangle`` have been corrected.
-
-All supported operators can be seen at :ref:`mathtext`.

doc/users/next_whats_new/mathtext_spacing.rst

@@ -0,0 +1,5 @@
+``mathtext`` spacing corrections
+--------------------------------
+
+As consequence of the updated documentation, the spacing on a number of relational and
+operator symbols were classified like that and therefore will be spaced properly.

lib/matplotlib/_mathtext.py

@@ -1733,7 +1733,17 @@ class _MathStyle(enum.Enum):
       \cap            \triangleleft            \dagger
       \cup            \triangleright           \ddagger
       \uplus          \lhd                     \amalg
-      \dotplus        \dotminus'''.split())
+      \dotplus        \dotminus                \Cap
+      \Cup            \barwedge                \boxdot
+      \boxminus       \boxplus                 \boxtimes
+      \curlyvee       \curlywedge              \divideontimes
+      \doublebarwedge \leftthreetimes          \rightthreetimes
+      \slash          \veebar                  \barvee
+      \cupdot         \intercal                \amalg
+      \circledcirc    \circleddash             \circledast
+      \boxbar         \obar                    \merge
+      \minuscolon     \dotsminusdots
+      '''.split())
 
     _relation_symbols = set(r'''
       = < > :
@@ -1768,20 +1778,53 @@ class _MathStyle(enum.Enum):
       \trianglelefteq    \ntrianglelefteq    \trianglerighteq
       \ntrianglerighteq  \blacktriangleleft  \blacktriangleright
       \equalparallel     \measuredrightangle \varlrtriangle
-      '''.split())
-
-    _arrow_symbols = set(r'''
-      \leftarrow              \longleftarrow           \uparrow
-      \Leftarrow              \Longleftarrow           \Uparrow
-      \rightarrow             \longrightarrow          \downarrow
-      \Rightarrow             \Longrightarrow          \Downarrow
-      \leftrightarrow         \longleftrightarrow      \updownarrow
-      \Leftrightarrow         \Longleftrightarrow      \Updownarrow
-      \mapsto                 \longmapsto              \nearrow
-      \hookleftarrow          \hookrightarrow          \searrow
-      \leftharpoonup          \rightharpoonup          \swarrow
-      \leftharpoondown        \rightharpoondown        \nwarrow
-      \rightleftharpoons      \leadsto'''.split())
+      \Doteq        \Bumpeq       \Subset      \Supset
+      \backepsilon  \because      \therefore   \bot
+      \top          \bumpeq       \circeq      \coloneq
+      \curlyeqprec  \curlyeqsucc  \eqcirc      \eqcolon
+      \eqsim        \fallingdotseq \gtrdot     \gtrless
+      \ltimes       \rtimes       \lessdot     \ne
+      \ncong        \nequiv       \ngeq        \ngtr
+      \nleq         \nless        \nmid        \notin
+      \nprec        \nsubset      \nsubseteq   \nsucc
+      \nsupset      \nsupseteq    \pitchfork   \preccurlyeq
+      \risingdotseq \subsetneq    \succcurlyeq \supsetneq
+      \varpropto    \vartriangleleft \scurel
+      \vartriangleright \rightangle \equal     \backcong
+      \eqdef        \wedgeq       \questeq     \between
+      \veeeq        \disin        \varisins    \isins
+      \isindot      \varisinobar  \isinobar    \isinvb
+      \isinE        \nisd         \varnis      \nis
+      \varniobar    \niobar       \bagmember   \ratio
+      \Equiv        \stareq       \measeq      \arceq
+      \rightassert  \rightModels  \smallin     \smallowns
+      \notsmallowns \nsimeq'''.split())
+
+    _arrow_symbols = set(r"""
+     \leftarrow \longleftarrow \uparrow \Leftarrow \Longleftarrow
+     \Uparrow \rightarrow \longrightarrow \downarrow \Rightarrow
+     \Longrightarrow \Downarrow \leftrightarrow \updownarrow
+     \longleftrightarrow \updownarrow \Leftrightarrow
+     \Longleftrightarrow \Updownarrow \mapsto \longmapsto \nearrow
+     \hookleftarrow \hookrightarrow \searrow \leftharpoonup
+     \rightharpoonup \swarrow \leftharpoondown \rightharpoondown
+     \nwarrow \rightleftharpoons \leadsto \dashrightarrow
+     \dashleftarrow \leftleftarrows \leftrightarrows \Lleftarrow
+     \Rrightarrow \twoheadleftarrow \leftarrowtail \looparrowleft
+     \leftrightharpoons \curvearrowleft \circlearrowleft \Lsh
+     \upuparrows \upharpoonleft \downharpoonleft \multimap
+     \leftrightsquigarrow \rightrightarrows \rightleftarrows
+     \rightrightarrows \rightleftarrows \twoheadrightarrow
+     \rightarrowtail \looparrowright \rightleftharpoons
+     \curvearrowright \circlearrowright \Rsh \downdownarrows
+     \upharpoonright \downharpoonright \rightsquigarrow \nleftarrow
+     \nrightarrow \nLeftarrow \nRightarrow \nleftrightarrow
+     \nLeftrightarrow \to \Swarrow \Searrow \Nwarrow \Nearrow
+     \leftsquigarrow \overleftarrow \overleftrightarrow \cwopencirclearrow
+     \downzigzagarrow \cupleftarrow \rightzigzagarrow \twoheaddownarrow
+     \updownarrowbar \twoheaduparrow \rightarrowbar \updownarrows
+     \barleftarrow \mapsfrom \mapsdown \mapsup \Ldsh \Rdsh
+     """.split())
 
     _spaced_symbols = _binary_operators | _relation_symbols | _arrow_symbols
 
@@ -2116,9 +2159,11 @@ def symbol(self, s, loc, toks):
             # such as ${ -2}$, $ -2$, or $   -2$.
             prev_char = next((c for c in s[:loc][::-1] if c != ' '), '')
             # Binary operators at start of string should not be spaced
-            if (c in self._binary_operators and
-                    (len(s[:loc].split()) == 0 or prev_char == '{' or
-                     prev_char in self._left_delims)):
+            # Also, operators in sub- or superscripts should not be spaced
+            if (self._in_subscript_or_superscript or (
+                    c in self._binary_operators and (
+                    len(s[:loc].split()) == 0 or prev_char == '{' or
+                    prev_char in self._left_delims))):
                 return [char]
             else:
                 return [Hlist([self._make_space(0.2),

lib/matplotlib/_mathtext_data.py

@@ -800,6 +800,8 @@
     'eta'                      : 951,
     'forall'                   : 8704,
     'eth'                      : 240,
+    'dh'                       : 240,
+    'DH'                       : 208,
     'colon'                    : 58,
     'sqcup'                    : 8852,
     'bigsqcup'                 : 10758,
@@ -908,7 +910,7 @@
     'O'                        : 216,
     'hookleftarrow'            : 8617,
     'trianglerighteq'          : 8885,
-    'nsime'                    : 8772,
+    'nsimeq'                   : 8772,
     'oe'                       : 339,
     'nwarrow'                  : 8598,
     'o'                        : 248,
@@ -1090,6 +1092,8 @@
     'bagmember'                : 8959,
     'triangle'                 : 9651,
     'iiiint'                   : 10764,
+    'amalg'                    : 10815,
+    'merge'                    : 10837,
 }
 
 # Each element is a 4-tuple of the form: