Remplacement de la table de conversion \IeC => ASCII par un traitement par PlasTeX

Le code est plus propre, et tous les caractères sont gérés.

index.py

@@ -9,6 +9,8 @@
 #         src is the .sxd file generated by latex
 #
 
+from plasTeX.TeX import TeX
+from unidecode import unidecode
 import sys
 import re
 import locale
@@ -17,47 +19,16 @@ import warnings
 # Pattern set to ignore latex command in title prefix
 keywordPattern = re.compile(r"^%(\w+)\s?(.*)$")
 firstLetterPattern = re.compile(r"^(?:\{?\\\w+\}?)*[^\w]*(\w)")
-iecPattern = re.compile(r"\IeC {\\(.*?)}")
-replacePattern = {
-     '`A': 'À',
-     '`a': 'à',
-     '^a': 'â',
-     "'a": 'á',
-     "~a": 'ã',
-     'oe': 'œ',
-     "'e" : 'é',
-     "`e" : 'è',
-     "^e" : 'ê',
-     '"e' : 'ë',
-     "'E" : 'É',
-     "`E" : 'È',
-     "'o" : 'ó',
-     "^o" : 'ô',
-     r'"\i' : 'i',
-     r'^\i' : 'i',
-     '"u' : 'ü',
-     '`u' : 'ù',
-     '`u' : 'ù',
-     '~n' : 'ñ',
-     "c C" : 'Ç',
-     "c c" : 'ç',
-     "textquoteright" : "'",
-}
 
 def sortkey(value):
     '''
     From a title, return something usable for sorting. It handles locale (but
-    don't forget to call locale.setlocale(locale.LC_ALL, '')). It also try to
+    don't forget to call locale.setlocale(locale.LC_ALL, '')). It also handles
-    handle the sort with crappy latex escape sequences. Some chars may not be
+    the sort with  latex escape sequences.
-    handled by this function, so add them to *replacePattern* dictionnary.
     '''
-    def repl(match):
+    tex = TeX()
-        try:
+    tex.input(value)
-            return replacePattern[match.group(1).strip()]
+    return locale.strxfrm(unidecode(tex.parse().textContent.replace(' ', 'A')))
-        except KeyError:
-            warnings.warn("Error, no match to replace %s in %s. You should add it in the coresponding table in title_sort.py" % (match.group(0), match.group(1)))
-
-    return locale.strxfrm(iecPattern.sub(repl, value).replace(' ', 'A'))
 
 def processSXDEntry(tab):
     return (tab[0], tab[1], tab[2])