Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-07 10:14:46 -04:00
Code Issues Packages Projects Releases Wiki Activity

Fixed recipe for The Australian

Browse Source
This commit is contained in:
Kovid Goyal 2009-11-29 09:07:43 -07:00
parent 9729c55413
commit 10ff0e8966
1 changed files with 41 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

71
resources/recipes/the_oz.recipe
View File

@ -27,14 +27,19 @@ class DailyTelegraph(BasicNewsRecipe):
, '--publisher' , title
]
keep_only_tags = [
dict(name='h1', attrs={'class':'section-heading'})
,dict(name='div', attrs={'id':'article'})
]
keep_only_tags = [dict(name='div', attrs={'id': 'story'})]
remove_tags = [dict(name=['object','link'])]
#remove_tags = [dict(name=['object','link'])]
remove_tags = [dict(name ='div', attrs = {'class': 'story-info'}),
dict(name ='div', attrs = {'class': 'story-header-tools'}),
dict(name ='div', attrs = {'class': 'story-sidebar'}),
dict(name ='div', attrs = {'class': 'story-footer'}),
dict(name ='div', attrs = {'id': 'comments'}),
dict(name ='div', attrs = {'class': 'story-extras story-extras-2'}),
dict(name ='div', attrs = {'class': 'group item-count-1 story-related'})
]
extra_css = '''
extra_css = '''
h1{font-family :Georgia,"Times New Roman",Times,serif; font-size:large; }
#article{font-family :Georgia,"Times New Roman",Times,serif; font-size: x-small;}
.module-subheader{font-family :Tahoma,Geneva,Arial,Helvetica,sans-serif; color:#666666; font-size: xx-small;}
@ -43,34 +48,40 @@ class DailyTelegraph(BasicNewsRecipe):
.caption{font-family:Trebuchet MS,Trebuchet,Helvetica,sans-serif; font-size: xx-small;}
'''
feeds = [
(u'News', u'http://feeds.news.com.au/public/rss/2.0/aus_news_807.xml'),
(u'World News', u'http://feeds.news.com.au/public/rss/2.0/aus_world_808.xml'),
(u'Opinion', u'http://feeds.news.com.au/public/rss/2.0/aus_opinion_58.xml'),
(u'Business', u'http://feeds.news.com.au/public/rss/2.0/aus_business_811.xml'),
(u'Media', u'http://feeds.news.com.au/public/rss/2.0/aus_media_57.xml'),
(u'Higher Education', u'http://feeds.news.com.au/public/rss/2.0/aus_higher_education_56.xml'),
(u'The Arts', u'http://feeds.news.com.au/public/rss/2.0/aus_arts_51.xml'),
(u'Commercial Property', u'http://feeds.news.com.au/public/rss/2.0/aus_business_commercial_property_708.xml'),
(u'The Nation', u'http://feeds.news.com.au/public/rss/2.0/aus_the_nation_62.xml'),
(u'Sport', u'http://feeds.news.com.au/public/rss/2.0/aus_sport_61.xml'),
(u'Travel', u'http://feeds.news.com.au/public/rss/2.0/aus_travel_and_indulgence_63.xml'),
(u'Defence', u'http://feeds.news.com.au/public/rss/2.0/aus_defence_54.xml'),
(u'Aviation', u'http://feeds.news.com.au/public/rss/2.0/aus_business_aviation_706.xml'),
(u'Mining', u'http://feeds.news.com.au/public/rss/2.0/aus_business_mining_704.xml'),
(u'Climate', u'http://feeds.news.com.au/public/rss/2.0/aus_climate_809.xml'),
(u'Property', u'http://feeds.news.com.au/public/rss/2.0/aus_property_59.xml'),
(u'US Election', u'http://feeds.news.com.au/public/rss/2.0/aus_uselection_687.xml')
]
feeds = [(u'News', u'http://feeds.news.com.au/public/rss/2.0/aus_news_807.xml'),
(u'Opinion', u'http://feeds.news.com.au/public/rss/2.0/aus_opinion_58.xml'),
(u'Business', u'http://feeds.news.com.au/public/rss/2.0/aus_business_811.xml'),
(u'Media', u'http://feeds.news.com.au/public/rss/2.0/aus_media_57.xml'),
(u'Higher Education', u'http://feeds.news.com.au/public/rss/2.0/aus_higher_education_56.xml'),
(u'The Arts', u'http://feeds.news.com.au/public/rss/2.0/aus_arts_51.xml'),
(u'Commercial Property', u'http://feeds.news.com.au/public/rss/2.0/aus_business_commercial_property_708.xml'),
(u'The Nation', u'http://feeds.news.com.au/public/rss/2.0/aus_the_nation_62.xml'),
(u'Sport', u'http://feeds.news.com.au/public/rss/2.0/aus_sport_61.xml'),
(u'Travel', u'http://feeds.news.com.au/public/rss/2.0/aus_travel_and_indulgence_63.xml'),
(u'Defence', u'http://feeds.news.com.au/public/rss/2.0/aus_defence_54.xml'),
(u'Aviation', u'http://feeds.news.com.au/public/rss/2.0/aus_business_aviation_706.xml'),
(u'Mining', u'http://feeds.news.com.au/public/rss/2.0/aus_business_mining_704.xml'),
(u'Climate', u'http://feeds.news.com.au/public/rss/2.0/aus_climate_809.xml'),
(u'Property', u'http://feeds.news.com.au/public/rss/2.0/aus_property_59.xml'),
(u'US Election', u'http://feeds.news.com.au/public/rss/2.0/aus_uselection_687.xml')]
def get_article_url(self, article):
return article.id
#br = self.get_browser()
#br.open(article.link).read()
#print br.geturl()
#return br.geturl()
def get_cover_url(self):
href = 'http://www.theaustralian.news.com.au/'
href = 'http://www.theaustralian.news.com.au/'
soup = self.index_to_soup(href)
img = soup.find('img',alt ="Digital editions of The Australian")
img = soup.find('img',alt ="AUS HP promo digital2")
print img
if img :
cover_url = img['src']
return cover_url