Arizona Republic by Jim Olo

This commit is contained in:
Kovid Goyal 2011-06-29 09:03:24 -06:00
parent 9ab9d20add
commit 5798d82a26

View File

@ -0,0 +1,68 @@
__license__ = 'GPL v3'
__copyright__ = '2010, jolo'
'''
azrepublic.com
'''
from calibre.web.feeds.recipes import BasicNewsRecipe
class AdvancedUserRecipe1307301031(BasicNewsRecipe):
title = u'AZRepublic'
__author__ = 'Jim Olo'
language = 'en'
description = "The Arizona Republic is Arizona's leading provider of news and information, and has published a daily newspaper in Phoenix for more than 110 years"
publisher = 'AZRepublic/AZCentral'
masthead_url = 'http://freedom2t.com/wp-content/uploads/press_az_republic_v2.gif'
cover_url = 'http://www.valleyleadership.org/Common/Img/2line4c_AZRepublic%20with%20azcentral%20logo.jpg'
category = 'news, politics, USA, AZ, Arizona'
oldest_article = 7
max_articles_per_feed = 100
remove_empty_feeds = True
no_stylesheets = True
remove_javascript = True
# extra_css = '.headline {font-size: medium;} \n .fact { padding-top: 10pt }'
extra_css = ' body{ font-family: Verdana,Helvetica,Arial,sans-serif } .headline {font-size: medium} .introduction{font-weight: bold} .story-feature{display: block; padding: 0; border: 1px solid; width: 40%; font-size: small} .story-feature h2{text-align: center; text-transform: uppercase} '
remove_attributes = ['width','height','h2','subHeadline','style']
remove_tags = [
dict(name='div', attrs={'id':['slidingBillboard', 'top728x90', 'subindex-header', 'topSearch']}),
dict(name='div', attrs={'id':['simplesearch', 'azcLoginBox', 'azcLoginBoxInner', 'topNav']}),
dict(name='div', attrs={'id':['carsDrop', 'homesDrop', 'rentalsDrop', 'classifiedDrop']}),
dict(name='div', attrs={'id':['nav', 'mp', 'subnav', 'jobsDrop']}),
dict(name='h6', attrs={'class':['section-header']}),
dict(name='a', attrs={'href':['#comments']}),
dict(name='div', attrs={'class':['articletools clearfix', 'floatRight']}),
dict(name='div', attrs={'id':['fbFrame', 'ob', 'storyComments', 'storyGoogleAdBox']}),
dict(name='div', attrs={'id':['storyTopHomes', 'openRight', 'footerwrap', 'copyright']}),
dict(name='div', attrs={'id':['blogsHed', 'blog_comments', 'blogByline','blogTopics']}),
dict(name='div', attrs={'id':['membersRightMain', 'dealsfooter', 'azrTopHed', 'azrRightCol']}),
dict(name='div', attrs={'id':['ttdHeader', 'ttdTimeWeather']}),
dict(name='div', attrs={'id':['membersRightMain', 'deals-header-wrap']}),
dict(name='div', attrs={'id':['todoTopSearchBar', 'byline clearfix', 'subdex-topnav']}),
dict(name='h1', attrs={'id':['SEOtext']}),
dict(name='table', attrs={'class':['ap-mediabox-table']}),
dict(name='p', attrs={'class':['ap_para']}),
dict(name='span', attrs={'class':['source-org vcard', 'org fn']}),
dict(name='a', attrs={'href':['http://hosted2.ap.org/APDEFAULT/privacy']}),
dict(name='a', attrs={'href':['http://hosted2.ap.org/APDEFAULT/terms']}),
dict(name='div', attrs={'id':['onespot_nextclick']}),
]
feeds = [
(u'FrontPage', u'http://www.azcentral.com/rss/feeds/republicfront.xml'),
(u'TopUS-News', u'http://hosted.ap.org/lineups/USHEADS.rss?SITE=AZPHG&SECTION=HOME'),
(u'WorldNews', u'http://hosted.ap.org/lineups/WORLDHEADS.rss?SITE=AZPHG&SECTION=HOME'),
(u'TopBusiness', u'http://hosted.ap.org/lineups/BUSINESSHEADS.rss?SITE=AZPHG&SECTION=HOME'),
(u'Entertainment', u'http://hosted.ap.org/lineups/ENTERTAINMENT.rss?SITE=AZPHG&SECTION=HOME'),
(u'ArizonaNews', u'http://www.azcentral.com/rss/feeds/news.xml'),
(u'Gilbert', u'http://www.azcentral.com/rss/feeds/gilbert.xml'),
(u'Chandler', u'http://www.azcentral.com/rss/feeds/chandler.xml'),
(u'DiningReviews', u'http://www.azcentral.com/rss/feeds/diningreviews.xml'),
(u'AZBusiness', u'http://www.azcentral.com/rss/feeds/business.xml'),
(u'ArizonaDeals', u'http://www.azcentral.com/members/Blog%7E/RealDealsblog'),
(u'GroceryDeals', u'http://www.azcentral.com/members/Blog%7E/RealDealsblog/tag/2646')
]