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

Fix #7151 (The Miami Herald)

Browse Source
This commit is contained in:
Kovid Goyal 2010-10-12 13:24:22 -06:00
parent 63dabd9789
commit 4c49384670
1 changed files with 28 additions and 45 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

73
resources/recipes/miami_herald.recipe
View File

@ -1,6 +1,6 @@
__license__ = 'GPL v3'
__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
__copyright__ = '2009-2010, Darko Miletic <darko.miletic at gmail.com>'
'''
miamiherald.com
'''
@ -15,13 +15,11 @@ class TheMiamiHerald(BasicNewsRecipe):
max_articles_per_feed = 100
publisher = u'The Miami Herald'
category = u'miami herald, weather, dolphins, news, miami news, local news, miamiherald, miami newspaper, miamiherald.com, miami, the miami herald, broward, miami-dade'
language = 'en'
language = 'en'
no_stylesheets = True
use_embedded_content = False
encoding = 'cp1252'
remove_javascript = True
extra_css = '''
h1{font-family:Arial,Helvetica,sans-serif; font-size:large; color:#1A272F; }
.subheadline{font-family:Arial,Helvetica,sans-serif; font-size:30%; color: #666666;}
@ -33,50 +31,35 @@ class TheMiamiHerald(BasicNewsRecipe):
.imageCaption{font-family:Arial,Helvetica,sans-serif; font-size:30%; color:#666666; }
'''
keep_only_tags = [dict(name='div', attrs={'id':['storyBody','storyPhotoContentArea']}),
]
conversion_options = {
'comment' : description
, 'tags' : category
, 'publisher' : publisher
, 'language' : language
}
remove_tags = [dict(name=['object','link','embed']),
dict(name='div', attrs={'class':["imageBuyButton","shareLinksArea","storyTools","spill_navigation pagination","circPromoArea","storyTools_footer","storyYahooContentMatch"]}) ,
dict(name='div', attrs={'id':["pluck","mlt","storyAssets"]}) ]
keep_only_tags = [dict(name='div', attrs={'id':'wide'}),]
remove_tags = [dict(name=['object','link','embed','iframe','meta'])]
feeds = [
(u'Breaking News' , u'http://www.miamiherald.com/416/index.xml' )
,(u'Miami-Dade' , u'http://www.miamiherald.com/460/index.xml' )
,(u'Broward' , u'http://www.miamiherald.com/467/index.xml' )
,(u'Florida Keys' , u'http://www.miamiherald.com/505/index.xml' )
,(u'Florida' , u'http://www.miamiherald.com/569/index.xml' )
,(u'Nation' , u'http://www.miamiherald.com/509/index.xml' )
,(u'World' , u'http://www.miamiherald.com/578/index.xml' )
,(u'Americas' , u'http://www.miamiherald.com/579/index.xml' )
,(u'Cuba' , u'http://www.miamiherald.com/581/index.xml' )
,(u'Haiti' , u'http://www.miamiherald.com/582/index.xml' )
,(u'Politics' , u'http://www.miamiherald.com/515/index.xml' )
,(u'Education' , u'http://www.miamiherald.com/295/index.xml' )
,(u'Environment' , u'http://www.miamiherald.com/573/index.xml' )
(u'Breaking News' , u'http://www.miamiherald.com/news/breaking-news/index.xml' )
,(u'Miami-Dade' , u'http://www.miamiherald.com/news/miami-dade/index.xml' )
,(u'Broward' , u'http://www.miamiherald.com/news/broward/index.xml' )
,(u'Florida Keys' , u'http://www.miamiherald.com/news/florida-keys/index.xml' )
,(u'Florida' , u'http://www.miamiherald.com/news/florida/index.xml' )
,(u'Nation' , u'http://www.miamiherald.com/news/nation/index.xml' )
,(u'World' , u'http://www.miamiherald.com/news/world/index.xml' )
,(u'Americas' , u'http://www.miamiherald.com/news/americas/index.xml' )
,(u'Cuba' , u'http://www.miamiherald.com/news/americas/cuba/index.xml' )
,(u'Haiti' , u'http://www.miamiherald.com/news/americas/haiti/index.xml' )
,(u'Politics' , u'http://www.miamiherald.com/news/politics/index.xml' )
,(u'Education' , u'http://www.miamiherald.com/news/education/index.xml' )
,(u'Environment' , u'http://www.miamiherald.com/news/environment/index.xml' )
]
def get_article_url(self, article):
ans = article.get('guid', None)
print ans
try:
self.log('Looking for full story link in', ans)
soup = self.index_to_soup(ans)
x = soup.find(text="Full Story")
if x is not None:
a = x.parent
if a and a.has_key('href'):
ans = 'http://www.miamiherald.com'+a['href']
self.log('Found full story link', ans)
except:
pass
return ans
def print_version(self, url):
art, sep, rest = url.rpartition('/')
art2, sep2, rest2 = art.rpartition('/')
return art2 + '/v-print/' + rest2 + '/' + rest