split by more punctuations to avode crazy long lines.

git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@44494 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5/bootstrap/erp5_core/ExtensionTemplateItem/DocumentExtraction.py

@@ -104,9 +104,11 @@ def getExcerptText(context, txt, sw, tags, trail, maxlines):
   txt = re.sub(r,'',txt)
   r = re.compile('<([^>]+)>',re.DOTALL|re.IGNORECASE)
   txt = re.sub(r,'',txt)
+  txt = txt.replace('-',' - ') # to find hyphenated occurrences
+  txt = txt.replace(',',', ')
+  txt = txt.replace(';','; ')
   r = re.compile('\s+')
   txt = re.sub(r,' ',txt)
-  txt = txt.replace('-',' - ') # to find hyphenated occurrences
   text = ' '.join(txt.split('\n')).split(' ') # very rough tokenization
   return [p for p in generateParts(context,text,sw,tags,trail,maxlines)]
 

product/ERP5/bootstrap/erp5_core/bt/revision

-40937
+40938
\ No newline at end of file