lxbfYeaa/BookConverter
Archived
Template
1
0
Fork 0
forked from LiveCarta/BookConverter
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Optimize heading cleaning

Browse Source
This commit is contained in:
Kiryl
2022-06-23 17:44:48 +03:00
parent 536c1d23e4
commit 7e380ef431
1 changed files with 8 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
src/epub_converter/html_epub_preprocessor.py
View File

@@ -126,7 +126,7 @@ def _remove_headings_content(content_tag, title_of_chapter: str):
title_of_chapter = title_of_chapter.lower() title_of_chapter = title_of_chapter.lower()
for tag in content_tag.contents: for tag in content_tag.contents:
text = tag if isinstance(tag, NavigableString) else tag.text text = tag if isinstance(tag, NavigableString) else tag.text
if re.sub(r'([\s\xa0])', '', text): if re.sub(r"[\s\xa0]", "", text):
text = re.sub(r"[\s\xa0]", " ", text).lower() text = re.sub(r"[\s\xa0]", " ", text).lower()
text = text.strip() # delete extra spaces text = text.strip() # delete extra spaces
if title_of_chapter == text or \ if title_of_chapter == text or \
@@ -134,9 +134,10 @@ def _remove_headings_content(content_tag, title_of_chapter: str):
re.findall(r"^h[1-3]$", tag.name or content_tag.name)): re.findall(r"^h[1-3]$", tag.name or content_tag.name)):
_add_span_to_save_ids_for_links(tag, content_tag) _add_span_to_save_ids_for_links(tag, content_tag)
tag.extract() tag.extract()
return
elif not isinstance(tag, NavigableString): elif not isinstance(tag, NavigableString):
_remove_headings_content(tag, title_of_chapter) if not _remove_headings_content(tag, title_of_chapter):
break break
def _tags_to_correspond_livecarta_tag(chapter_tag): def _tags_to_correspond_livecarta_tag(chapter_tag):
@@ -275,13 +276,13 @@ def _preprocess_div_tags(chapter_tag):
Function replace <div> with <table>: Function replace <div> with <table>:
""" """
for div in chapter_tag.find_all("div"): for div in chapter_tag.find_all("div"):
if any(attr in ['width', 'border', 'bgcolor'] for attr in div.attrs): if any(attr in ["width", "border", "bgcolor"] for attr in div.attrs):
_wrap_tag_with_table( _wrap_tag_with_table(
chapter_tag, chapter_tag,
tag_to_be_wrapped=div, tag_to_be_wrapped=div,
width=div.attrs['width'] if div.attrs.get('width') else '100', width=div.attrs["width"] if div.attrs.get("width") else "100",
border=div.attrs['border'] if div.attrs.get('border') else None, border=div.attrs["border"] if div.attrs.get("border") else None,
bg_color=div.attrs['bgcolor'] if div.attrs.get('bgcolor') else None) bg_color=div.attrs["bgcolor"] if div.attrs.get("bgcolor") else None)
else: else:
div.name = "p" div.name = "p"
continue continue