This commit is contained in:
Kovid Goyal 2011-06-02 11:08:17 -06:00
parent b42f012764
commit 4264660c51
2 changed files with 10 additions and 64 deletions

View File

@ -4,75 +4,28 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
Profile to download CNN
'''
from calibre.web.feeds.news import BasicNewsRecipe
from calibre.ebooks.BeautifulSoup import BeautifulSoup
class CNN(BasicNewsRecipe):
title = 'CNN'
description = 'Global news'
timefmt = ' [%d %b %Y]'
__author__ = 'Krittika Goyal and Sujata Raman'
__author__ = 'Kovid Goyal'
language = 'en'
no_stylesheets = True
use_embedded_content = False
oldest_article = 15
recursions = 1
match_regexps = [r'http://sportsillustrated.cnn.com/.*/[1-9].html']
#recursions = 1
#match_regexps = [r'http://sportsillustrated.cnn.com/.*/[1-9].html']
max_articles_per_feed = 25
extra_css = '''
.cnn_strycntntlft{font-family :Arial,Helvetica,sans-serif;}
h2{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
.cnnTxtCmpnt{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
.cnnTMcontent{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#575757}
.storytext{font-family :Arial,Helvetica,sans-serif; font-size:small}
.storybyline{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#575757}
.credit{font-family :Arial,Helvetica,sans-serif; font-size:xx-small; color:#575757}
.storyBrandingBanner{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#575757}
.storytimestamp{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#575757}
.timestamp{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#575757}
.cnn_strytmstmp{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.cnn_stryimg640caption{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.cnn_strylccimg300cntr{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.cnn_stryichgfcpt{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.cnnByline{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.cnn_bulletbin cnnStryHghLght{ font-size:xx-small;}
.subhead p{font-family :Arial,Helvetica,sans-serif; font-size:x-small;}
.cnnStoryContent{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
.cnnContentContainer{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
.col1{font-family :Arial,Helvetica,sans-serif; font-size:x-small; color:#666666;}
.col3{color:#333333; font-family :Arial,Helvetica,sans-serif; font-size:x-small;font-weight:bold;}
.cnnInlineT1Caption{font-family :Arial,Helvetica,sans-serif; font-size:x-small;font-weight:bold;}
.cnnInlineT1Credit{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#333333;}
.col10{color:#5A637E;}
.cnnInlineRailBulletList{color:black;}
.cnnLine0{font-family :Arial,Helvetica,sans-serif; color:#666666;font-weight:bold;}
.cnnTimeStamp{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#333333;}
.galleryhedDek{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#575757;}
.galleryWidgetHeader{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#004276;}
.article-content{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
.cnnRecapStory{font-family :Arial,Helvetica,sans-serif; font-size:x-small}
h1{font-family :Arial,Helvetica,sans-serif; font-size:x-large}
.captionname{font-family :Arial,Helvetica,sans-serif; font-size:x-small;color:#575757;}
inStoryIE{{font-family :Arial,Helvetica,sans-serif; font-size:x-small;}
'''
#remove_tags_before = dict(name='h1', attrs={'class':'heading'})
#remove_tags_after = dict(name='td', attrs={'class':'newptool1'})
remove_tags = [
dict(name='iframe'),
dict(name='div', attrs={'class':['cnnEndOfStory', 'cnnShareThisItem', 'cnn_strylctcntr cnn_strylctcqrelt', 'cnnShareBoxContent', 'cnn_strybtmcntnt', 'cnn_strycntntrgt']}),
dict(name='div', attrs={'id':['IEContainer', 'clickIncludeBox']}),
#dict(name='ul', attrs={'class':'article-tools'}),
#dict(name='ul', attrs={'class':'articleTools'}),
]
feeds = [
('Top News', 'http://rss.cnn.com/rss/cnn_topstories.rss'),
('World', 'http://rss.cnn.com/rss/cnn_world.rss'),
('U.S.', 'http://rss.cnn.com/rss/cnn_us.rss'),
#('Sports', 'http://rss.cnn.com/rss/si_topstories.rss'),
('Sports', 'http://rss.cnn.com/rss/si_topstories.rss'),
('Business', 'http://rss.cnn.com/rss/money_latest.rss'),
('Politics', 'http://rss.cnn.com/rss/cnn_allpolitics.rss'),
('Law', 'http://rss.cnn.com/rss/cnn_law.rss'),
@ -84,15 +37,8 @@ class CNN(BasicNewsRecipe):
('Offbeat', 'http://rss.cnn.com/rss/cnn_offbeat.rss'),
('Most Popular', 'http://rss.cnn.com/rss/cnn_mostpopular.rss')
]
def preprocess_html(self, soup):
story = soup.find(name='div', attrs={'class':'cnnBody_Left'})
if story is None:
story = soup.find(name='div', attrs={'id':'cnnContentContainer'})
soup = BeautifulSoup('<html><head><title>t</title></head><body></body></html>')
body = soup.find(name='body')
body.insert(0, story)
else:
soup = BeautifulSoup('<html><head><title>t</title></head><body></body></html>')
body = soup.find(name='body')
body.insert(0, story)
return soup
def get_article_url(self, article):
ans = BasicNewsRecipe.get_article_url(self, article)
return ans.partition('?')[0]

View File

@ -593,7 +593,7 @@ if __name__ == '__main__':
db = db()
if True:
if False:
d = RuleEditor(db.field_metadata)
d.add_blank_condition()
d.exec_()