Backup_Repos/calibre
1
0
Fork 0
mirror of git://github.com/kovidgoyal/calibre.git synced 2026-05-09 05:32:57 +02:00
Code Issues Projects Releases Wiki Activity

Change line length amount to work with duplicates removed lengths. Enhance the unwrapping regex to account for more cases.

Browse source
This commit is contained in:
John Schember 2009-06-22 21:12:17 -04:00
parent e426c9e60d
commit 3a87f0d065
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
src/calibre/ebooks/conversion/preprocess.py
View file

@ -183,12 +183,12 @@ def __call__(self, html, remove_special_chars=None):
elif self.is_book_designer(html):
rules = self.BOOK_DESIGNER
elif self.is_pdftohtml(html):
length = line_length(html, .3)
length = line_length(html, .5)
line_length_rules = []
if length:
line_length_rules = [
# Un wrap using punctuation
(re.compile(r'(?<=.{%i}[a-z,;:-IA])\s*(?P<ital></(i|b|u)>)?\s*(<p.*?>)\s*(?=(<(i|b|u)>)?[\w\d])' % length, re.UNICODE), wrap_lines),
(re.compile(r'(?<=.{%i}[a-z\.,;:)-IA])\s*(?P<ital></(i|b|u)>)?\s*(<p.*?>)\s*(?=(<(i|b|u)>)?\s*[\w\d(])' % length, re.UNICODE), wrap_lines),
]
rules = self.PDFTOHTML + line_length_rules