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

Update The Australian

Browse Source
This commit is contained in:
Kovid Goyal 2013-05-19 09:26:47 +05:30
parent 3da5f5a2d5
commit a9440f99b3
1 changed files with 33 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

64
recipes/the_oz.recipe
View File

@ -26,14 +26,14 @@ class DailyTelegraph(BasicNewsRecipe):
keep_only_tags = [dict(name='div', attrs={'id': 'story'})] 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'}), remove_tags = [dict(name='div', attrs={'class': 'story-info'}),
dict(name ='div', attrs = {'class': 'story-header-tools'}), dict(name='div', attrs={'class': 'story-header-tools'}),
dict(name ='div', attrs = {'class': 'story-sidebar'}), dict(name='div', attrs={'class': 'story-sidebar'}),
dict(name ='div', attrs = {'class': 'story-footer'}), dict(name='div', attrs={'class': 'story-footer'}),
dict(name ='div', attrs = {'id': 'comments'}), dict(name='div', attrs={'id': 'comments'}),
dict(name ='div', attrs = {'class': 'story-extras story-extras-2'}), dict(name='div', attrs={'class': 'story-extras story-extras-2'}),
dict(name ='div', attrs = {'class': 'group item-count-1 story-related'}) dict(name='div', attrs={'class': 'group item-count-1 story-related'})
] ]
extra_css = ''' extra_css = '''
@ -45,30 +45,31 @@ class DailyTelegraph(BasicNewsRecipe):
.caption{font-family:Trebuchet MS,Trebuchet,Helvetica,sans-serif; font-size: xx-small;} .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'), feeds = [
(u'Opinion', u'http://feeds.news.com.au/public/rss/2.0/aus_opinion_58.xml'), (u'News', u'http://feeds.news.com.au/public/rss/2.0/aus_news_807.xml'),
(u'The Nation', u'http://feeds.news.com.au/public/rss/2.0/aus_the_nation_62.xml'), (u'Opinion', u'http://feeds.news.com.au/public/rss/2.0/aus_opinion_58.xml'),
(u'World News', u'http://feeds.news.com.au/public/rss/2.0/aus_world_808.xml'), (u'The Nation', u'http://feeds.news.com.au/public/rss/2.0/aus_the_nation_62.xml'),
(u'US Election', u'http://feeds.news.com.au/public/rss/2.0/aus_uselection_687.xml'), (u'World News', u'http://feeds.news.com.au/public/rss/2.0/aus_world_808.xml'),
(u'Climate', u'http://feeds.news.com.au/public/rss/2.0/aus_climate_809.xml'), (u'US Election', u'http://feeds.news.com.au/public/rss/2.0/aus_uselection_687.xml'),
(u'Media', u'http://feeds.news.com.au/public/rss/2.0/aus_media_57.xml'), (u'Climate', u'http://feeds.news.com.au/public/rss/2.0/aus_climate_809.xml'),
(u'IT', u'http://feeds.news.com.au/public/rss/2.0/ausit_itnews_topstories_367.xml'), (u'Media', u'http://feeds.news.com.au/public/rss/2.0/aus_media_57.xml'),
(u'Exec Tech', u'http://feeds.news.com.au/public/rss/2.0/ausit_exec_topstories_385.xml'), (u'IT', u'http://feeds.news.com.au/public/rss/2.0/ausit_itnews_topstories_367.xml'),
(u'Higher Education', u'http://feeds.news.com.au/public/rss/2.0/aus_higher_education_56.xml'), (u'Exec Tech', u'http://feeds.news.com.au/public/rss/2.0/ausit_exec_topstories_385.xml'),
(u'Arts', u'http://feeds.news.com.au/public/rss/2.0/aus_arts_51.xml'), (u'Higher Education', u'http://feeds.news.com.au/public/rss/2.0/aus_higher_education_56.xml'),
(u'Travel', u'http://feeds.news.com.au/public/rss/2.0/aus_travel_and_indulgence_63.xml'), (u'Arts', u'http://feeds.news.com.au/public/rss/2.0/aus_arts_51.xml'),
(u'Property', u'http://feeds.news.com.au/public/rss/2.0/aus_property_59.xml'), (u'Travel', u'http://feeds.news.com.au/public/rss/2.0/aus_travel_and_indulgence_63.xml'),
(u'Sport', u'http://feeds.news.com.au/public/rss/2.0/aus_sport_61.xml'), (u'Property', u'http://feeds.news.com.au/public/rss/2.0/aus_property_59.xml'),
(u'Business', u'http://feeds.news.com.au/public/rss/2.0/aus_business_811.xml'), (u'Sport', u'http://feeds.news.com.au/public/rss/2.0/aus_sport_61.xml'),
(u'Aviation', u'http://feeds.news.com.au/public/rss/2.0/aus_business_aviation_706.xml'), (u'Business', u'http://feeds.news.com.au/public/rss/2.0/aus_business_811.xml'),
(u'Commercial Property', u'http://feeds.news.com.au/public/rss/2.0/aus_business_commercial_property_708.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'Commercial Property', u'http://feeds.news.com.au/public/rss/2.0/aus_business_commercial_property_708.xml'),
(u'Mining', u'http://feeds.news.com.au/public/rss/2.0/aus_business_mining_704.xml')]
def get_browser(self): def get_browser(self):
br = BasicNewsRecipe.get_browser(self) br = BasicNewsRecipe.get_browser(self)
if self.username and self.password: if self.username and self.password:
br.open('http://www.theaustralian.com.au') br.open('http://www.theaustralian.com.au')
br.select_form(nr=0) br.select_form(nr=1)
br['username'] = self.username br['username'] = self.username
br['password'] = self.password br['password'] = self.password
raw = br.submit().read() raw = br.submit().read()
@ -80,10 +81,11 @@ class DailyTelegraph(BasicNewsRecipe):
def get_article_url(self, article): def get_article_url(self, article):
return article.id return article.id
#br = self.get_browser() # br = self.get_browser()
#br.open(article.link).read() # br.open(article.link).read()
#print br.geturl() # print br.geturl()
# return br.geturl()
#return br.geturl()