Backup_Repos/calibre
1
0
Fork 0
mirror of git://github.com/kovidgoyal/calibre.git synced 2026-04-30 11:13:02 +02:00
Code Issues Projects Releases Wiki Activity

pdftohtml preprocess rules work

Browse source
This commit is contained in:
John Schember 2009-04-15 20:11:00 -04:00
parent aafc6d9764
commit 575b021f48
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/calibre/ebooks/conversion/preprocess.py
View file

@ -73,7 +73,7 @@ class HTMLPreProcessor(object):
(re.compile(r'<br.*?>'), lambda match : '<p>'),
# Un wrap lines
(re.compile(r'(?<=\w)\s*</i>\s*<p.*?>\s*<i>\s*(?=\w)'), lambda match: ' '),
(re.compile(r'(?<=\w)\s*</(i|b|u)>\s*<p.*?>\s*<(i|b|u)>\s*(?=\w)'), lambda match: ' '),
(re.compile(r'(?<=\w)\s*<p.*?>\s*(?=\w)', re.UNICODE), lambda match: ' '),
# Clean up spaces
(re.compile(u'(?<=\.|,|:|;|\?|!|”|"|\')[\s^ ]*(?=<)'), lambda match: ' '),