Backup_Repos/calibre
1
0
Fork 0
mirror of git://github.com/kovidgoyal/calibre.git synced 2026-05-08 22:06:02 +02:00
Code Issues Projects Releases Wiki Activity

Allow saving index html easily

Browse source
This commit is contained in:
Kovid Goyal 2018-11-05 12:43:20 +05:30
parent 661c47501a
commit 6ce808c499
No known key found for this signature in database
GPG key ID: 06BC317B515ACE7C
1 changed files with 4 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
src/calibre/web/feeds/news.py
View file

@ -657,7 +657,7 @@ def canonicalize_internal_url(self, url, is_link=True):
return frozenset()
return frozenset([(parts.netloc, (parts.path or '').rstrip('/'))])
def index_to_soup(self, url_or_raw, raw=False, as_tree=False):
def index_to_soup(self, url_or_raw, raw=False, as_tree=False, save_raw=None):
'''
Convenience method that takes an URL to the index page and returns
a `BeautifulSoup <https://www.crummy.com/software/BeautifulSoup/bs3/documentation.html>`_
@ -692,6 +692,9 @@ def index_to_soup(self, url_or_raw, raw=False, as_tree=False):
else:
_raw = xml_to_unicode(_raw, strip_encoding_pats=True, resolve_entities=True)[0]
_raw = clean_xml_chars(_raw)
if save_raw:
with lopen(save_raw, 'wb') as f:
f.write(_raw.encode('utf-8'))
if as_tree:
from html5_parser import parse
return parse(_raw)