Backup_Repos/calibre
1
0
Fork 0
mirror of git://github.com/kovidgoyal/calibre.git synced 2026-05-05 04:13:47 +02:00
Code Issues Projects Releases Wiki Activity

Update NYTimes

Browse source
This commit is contained in:
Kovid Goyal 2017-01-28 09:25:31 +05:30
parent d025bc327d
commit 1af85ad42e
2 changed files with 8 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
recipes/nytimes.recipe
View file

@ -240,6 +240,7 @@ def decode_url_date(self, url):
dict(attrs={'class': lambda x: x and 'skip-to-text-link' in x.split()}),
dict(attrs={'class': lambda x: x and 'sharetools' in x.split()}),
dict(attrs={'class': lambda x: x and 'ad' in x.split()}),
dict(attrs={'class': lambda x: x and 'visually-hidden' in x.split()}),
dict(name='div', attrs={'class': re.compile('toolsList')}), # bits
dict(name='div', attrs={
'class': re.compile('postNavigation')}), # bits
@ -1061,6 +1062,9 @@ def postprocess_html(self, soup, first_fetch):
aside.extract()
soup = self.strip_anchors(soup, True)
for t in soup.findAll('time', attrs={'class':'dateline'}):
t.name = 'div'
if soup.find('div', attrs={'id': 'blogcontent'}) is None:
if first_fetch:
aside = soup.find('div', 'aside')

4
recipes/nytimes_sub.recipe
View file

@ -240,6 +240,7 @@ def decode_url_date(self, url):
dict(attrs={'class': lambda x: x and 'skip-to-text-link' in x.split()}),
dict(attrs={'class': lambda x: x and 'sharetools' in x.split()}),
dict(attrs={'class': lambda x: x and 'ad' in x.split()}),
dict(attrs={'class': lambda x: x and 'visually-hidden' in x.split()}),
dict(name='div', attrs={'class': re.compile('toolsList')}), # bits
dict(name='div', attrs={
'class': re.compile('postNavigation')}), # bits
@ -1069,6 +1070,9 @@ def postprocess_html(self, soup, first_fetch):
aside.extract()
soup = self.strip_anchors(soup, True)
for t in soup.findAll('time', attrs={'class':'dateline'}):
t.name = 'div'
if soup.find('div', attrs={'id': 'blogcontent'}) is None:
if first_fetch:
aside = soup.find('div', 'aside')