Backup_Repos/calibre
1
0
Fork 0
mirror of git://github.com/kovidgoyal/calibre.git synced 2025-12-21 19:36:03 +01:00
Code Issues Projects Releases Wiki Activity

FB2 Output: Remove section regexp what do nothing and rearrange regexps

Browse source
This commit is contained in:
Andrey Efremov 2019-10-08 16:10:33 +07:00
parent d7a4e10c60
commit e7175e802b
1 changed files with 7 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
src/calibre/ebooks/fb2/fb2ml.py
View file

@ -85,6 +85,9 @@ def clean_text(self, text):
# Put the paragraph following a paragraph on a separate line.
text = re.sub(r'(?mu)</p>\s*<p>', '</p>\n<p>', text)
if self.opts.insert_blank_line:
text = re.sub(r'(?mu)</p>', '</p><empty-line/>', text)
# Clean up title endings.
text = re.sub(r'(?mu)\s+</title>', '</title>', text)
# Remove empty title elements.
@ -92,6 +95,10 @@ def clean_text(self, text):
# Put the paragraph following a title on a separate line.
text = re.sub(r'(?mu)</title>\s*<p>', '</title>\n<p>', text)
# Put line breaks between paragraphs on a separate line.
text = re.sub(r'(?mu)</(p|title)>\s*<empty-line/>', r'</\1>\n<empty-line/>', text)
text = re.sub(r'(?mu)<empty-line/>\s*<p>', '<empty-line/>\n<p>', text)
# Remove empty sections.
text = re.sub(r'(?mu)<section>\s*</section>', '', text)
# Clean up sections starts and ends.
@ -99,15 +106,6 @@ def clean_text(self, text):
text = re.sub(r'(?mu)<section>\s*', '<section>\n', text)
text = re.sub(r'(?mu)\s*</section>', '\n</section>', text)
text = re.sub(r'(?mu)</section>\s*', '</section>\n', text)
# Put the section following a section on a separate line.
text = re.sub(r'(?mu)</section>\s*<section>', '</section>\n<section>', text)
if self.opts.insert_blank_line:
text = re.sub(r'(?mu)</p>', '</p><empty-line/>', text)
# Put line breaks between paragraphs on a separate line.
text = re.sub(r'(?mu)</(p|title)>\s*<empty-line/>', r'</\1>\n<empty-line/>', text)
text = re.sub(r'(?mu)<empty-line/>\s*<p>', '<empty-line/>\n<p>', text)
return text