Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Improved recipe for The Hindu

Browse Source
This commit is contained in:
Kovid Goyal 2009-12-09 10:42:23 -07:00
parent bddbefd121
commit 6914dd8b20
1 changed files with 10 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
resources/recipes/hindu.recipe
View File

@ -10,17 +10,21 @@ class TheHindu(BasicNewsRecipe):
language = 'en_IN' language = 'en_IN'
oldest_article = 7 oldest_article = 7
__author__ = _('Kovid Goyal') __author__ = 'Kovid Goyal and Sujata Raman'
max_articles_per_feed = 100 max_articles_per_feed = 100
no_stylesheets = True
remove_tags_before = {'name':'font', 'class':'storyhead'} remove_tags_before = {'name':'font', 'class':'storyhead'}
preprocess_regexps = [ preprocess_regexps = [
(re.compile(r'<!-- story ends -->.*', re.DOTALL), (re.compile(r'<!-- story ends -->.*', re.DOTALL),
lambda match: '</body></html>'), lambda match: '</body></html>'),
] ]
extra_css = '''
.storyhead{font-family:Arial,Helvetica,sans-serif; font-size:large; color:#000099;}
body{font-family:Verdana,Arial,Helvetica,sans-serif; font-size:x-small; text-align:left;}
'''
feeds = [ feeds = [
(u'Main - Font Page', u'http://www.hindu.com/rss/01hdline.xml'), (u'Main - Front Page', u'http://www.hindu.com/rss/01hdline.xml'),
(u'Main - National', u'http://www.hindu.com/rss/02hdline.xml'), (u'Main - National', u'http://www.hindu.com/rss/02hdline.xml'),
(u'Main - International', u'http://www.hindu.com/rss/03hdline.xml'), (u'Main - International', u'http://www.hindu.com/rss/03hdline.xml'),
(u'Main - Opinion', u'http://www.hindu.com/rss/05hdline.xml'), (u'Main - Opinion', u'http://www.hindu.com/rss/05hdline.xml'),
@ -43,6 +47,8 @@ class TheHindu(BasicNewsRecipe):
] ]
def postprocess_html(self, soup, first_fetch): def postprocess_html(self, soup, first_fetch):
for t in soup.findAll(['table', 'tr', 'td']): for t in soup.findAll(['table', 'tr', 'td','center']):
t.name = 'div' t.name = 'div'
return soup return soup