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 Honolulu Star Advertiser

Browse Source
This commit is contained in:
Kovid Goyal 2011-09-08 09:00:00 -06:00
parent 6ac2cea0d6
commit 7b4c73c111
1 changed files with 33 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

53
recipes/staradvertiser.recipe
View File

@ -1,5 +1,5 @@
__license__ = 'GPL v3' __license__ = 'GPL v3'
__copyright__ = '2009-2011, Darko Miletic <darko.miletic at gmail.com>' __copyright__ = '2011, M. Ching modified from work 2009-2011 Darko Miletic <darko.miletic at gmail.com>'
''' '''
staradvertiser.com staradvertiser.com
''' '''
@ -7,12 +7,13 @@ staradvertiser.com
from calibre.web.feeds.news import BasicNewsRecipe from calibre.web.feeds.news import BasicNewsRecipe
class Starbulletin(BasicNewsRecipe): class Starbulletin(BasicNewsRecipe):
title = 'Honolulu Star Advertiser' title = 'Honolulu Star-Advertiser'
__author__ = 'Darko Miletic' __author__ = 'Darko Miletic'
description = 'Latest national and local Hawaii sports news' description = 'Latest national and local Hawaii sports news'
publisher = 'Honolulu Star-Advertiser' publisher = 'Honolulu Star-Advertiser'
category = 'news, Honolulu, Hawaii' category = 'news, Honolulu, Hawaii'
oldest_article = 2 oldest_article = 2
needs_subscription = True
max_articles_per_feed = 100 max_articles_per_feed = 100
language = 'en' language = 'en'
no_stylesheets = True no_stylesheets = True
@ -20,12 +21,12 @@ class Starbulletin(BasicNewsRecipe):
encoding = 'utf8' encoding = 'utf8'
publication_type = 'newspaper' publication_type = 'newspaper'
masthead_url = 'http://media.staradvertiser.com/designimages/star-advertiser-logo-small.gif' masthead_url = 'http://media.staradvertiser.com/designimages/star-advertiser-logo-small.gif'
extra_css = """ # extra_css = """
body{font-family: Verdana,Arial,Helvetica,sans-serif} # body{font-family: Verdana,Arial,Helvetica,sans-serif}
h1,.brown,.postCredit{color: #663300} # h1,.brown,.hsa_postCredit{color: #663300}
.storyDeck{font-size: 1.2em; font-weight: bold} # .storyDeck{font-size: 1.2em; font-weight: bold}
img{display: block} # img{display: block}
""" # """
conversion_options = { conversion_options = {
'comment' : description 'comment' : description
@ -35,26 +36,36 @@ class Starbulletin(BasicNewsRecipe):
, 'linearize_tables' : True , 'linearize_tables' : True
} }
keep_only_tags = [ keep_only_tags = [
dict(attrs={'id':'storyTitle'}) dict(attrs={'id':'hsa_storyTitle'})
,dict(attrs={'class':['storyDeck','postCredit']}) ,dict(attrs={'class':['hsa_dateStamp','hsa_postCredit','storyDeck']})
,dict(name='span',attrs={'class':'brown'}) ,dict(name='span',attrs={'class':['hsa_dateStamp','hsa_postCredit']})
,dict(name='div',attrs={'class':'storytext article-important'})
,dict(name='div',attrs={'class':'storytext'}) ,dict(name='div',attrs={'class':'storytext'})
] ]
remove_tags = [ remove_tags = [
dict(name=['object','link','script','span','meta','base','iframe']) dict(name=['object','link','script','meta','base','iframe'])
# removed 'span' from preceding list to permit keeping of author and timestamp
,dict(attrs={'class':['insideStoryImage','insideStoryAd']}) ,dict(attrs={'class':['insideStoryImage','insideStoryAd']})
,dict(attrs={'name':'fb_share'}) ,dict(attrs={'name':'fb_share'})
] ]
feeds = [ def get_browser(self):
(u'Headlines' , u'http://www.staradvertiser.com/staradvertiser_headlines.rss' ) br = BasicNewsRecipe.get_browser()
,(u'News' , u'http://www.staradvertiser.com/news/index.rss' ) if self.username is not None and self.password is not None:
,(u'Sports' , u'http://www.staradvertiser.com/sports/index.rss' ) br.open('http://www.staradvertiser.com/manage/Login/')
,(u'Features' , u'http://www.staradvertiser.com/features/index.rss' ) br.select_form(name='loginForm')
,(u'Editorials', u'http://www.staradvertiser.com/editorials/index.rss' ) br['email'] = self.username
,(u'Business' , u'http://www.staradvertiser.com/business/index.rss' ) br['password'] = self.password
,(u'Travel' , u'http://www.staradvertiser.com/travel/index.rss' ) br.submit()
] return br
feeds = [
(u'Breaking News', u'http://www.staradvertiser.com/news/breaking/index.rss')
,(u'News', u'http://www.staradvertiser.com/newspremium/index.rss')
,(u'Business', u'http://www.staradvertiser.com/businesspremium/index.rss')
,(u'Sports', u'http://www.staradvertiser.com/sportspremium/index.rss')
,(u'Features', u'http://www.staradvertiser.com/featurespremium/index.rss')
]
def preprocess_html(self, soup): def preprocess_html(self, soup):
for item in soup.findAll(style=True): for item in soup.findAll(style=True):