mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-07-09 03:04:10 -04:00
Update The Hindu
This commit is contained in:
parent
68bca92457
commit
67cc77eb8a
@ -16,10 +16,14 @@ class TheHindu(BasicNewsRecipe):
|
||||
|
||||
keep_only_tags = [dict(id='content')]
|
||||
remove_tags = [dict(attrs={'class':['article-links', 'breadcr']}),
|
||||
dict(id=['email-section', 'right-column', 'printfooter'])]
|
||||
dict(id=['email-section', 'right-column', 'printfooter', 'topover',
|
||||
'slidebox', 'th_footer'])]
|
||||
|
||||
extra_css = '.photo-caption { font-size: smaller }'
|
||||
|
||||
def preprocess_raw_html(self, raw, url):
|
||||
return raw.replace('<body><p>', '<p>').replace('</p></body>', '</p>')
|
||||
|
||||
def postprocess_html(self, soup, first_fetch):
|
||||
for t in soup.findAll(['table', 'tr', 'td','center']):
|
||||
t.name = 'div'
|
||||
|
Loading…
x
Reference in New Issue
Block a user