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

Improved Reuters

Browse Source
This commit is contained in:
Kovid Goyal 2010-04-29 08:16:59 -06:00
parent 41e696eba1
commit 6a516d5703
1 changed files with 24 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
resources/recipes/reuters.recipe
View File

@ -7,14 +7,31 @@ class Reuters(BasicNewsRecipe):
title = 'Reuters' title = 'Reuters'
description = 'Global news' description = 'Global news'
__author__ = 'Kovid Goyal' __author__ = 'Kovid Goyal and Sujata Raman'
use_embedded_content = False use_embedded_content = False
language = 'en' language = 'en'
max_articles_per_feed = 10 max_articles_per_feed = 10
no_stylesheets = True
remove_javascript = True
extra_css = '''
preprocess_regexps = [ (re.compile(i[0], re.IGNORECASE | re.DOTALL), i[1]) for i in body{font-family:arial,helvetica,sans;}
h1{ font-size:larger ; font-weight:bold; }
.byline{color:#006E97;font-size:x-small; font-weight:bold;}
.location{font-size:x-small; font-weight:bold;}
.timestamp{font-size:x-small; }
'''
keep_only_tags = [dict(name='div', attrs={'class':'column2 gridPanel grid8'})]
remove_tags = [dict(name='div', attrs={'id':['recommendedArticles','relatedNews','relatedVideo','relatedFactboxes']}),
dict(name='p', attrs={'class':['relatedTopics']}),
dict(name='a', attrs={'id':['fullSizeLink']}),
dict(name='div', attrs={'class':['photoNav','relatedTopicButtons','articleComments','gridPanel grid8','footerHalf gridPanel grid1','gridPanel grid2','gridPanel grid3']}),]
preprocess_regexps = [ (re.compile(i[0], re.IGNORECASE | re.DOTALL), i[1]) for i in
[ [
##(r'<HEAD>.*?</HEAD>' , lambda match : '<HEAD></HEAD>'), ##(r'<HEAD>.*?</HEAD>' , lambda match : '<HEAD></HEAD>'),
(r'<div id="apple-rss-sidebar-background">.*?<!-- start Entries -->', lambda match : ''), (r'<div id="apple-rss-sidebar-background">.*?<!-- start Entries -->', lambda match : ''),
@ -24,10 +41,10 @@ class Reuters(BasicNewsRecipe):
(r'<h3>Share:</h3>.*?</body>', lambda match : '<!-- END:: Shared Module id=36615 --></body>'), (r'<h3>Share:</h3>.*?</body>', lambda match : '<!-- END:: Shared Module id=36615 --></body>'),
(r'<div id="atools" class="articleTools">.*?<div class="linebreak">', lambda match : '<div class="linebreak">'), (r'<div id="atools" class="articleTools">.*?<div class="linebreak">', lambda match : '<div class="linebreak">'),
] ]
] ]
feeds = [ ('Top Stories', 'http://feeds.reuters.com/reuters/topNews?format=xml'), feeds = [ ('Top Stories', 'http://feeds.reuters.com/reuters/topNews?format=xml'),
('US News', 'http://feeds.reuters.com/reuters/domesticNews?format=xml'), ('US News', 'http://feeds.reuters.com/reuters/domesticNews?format=xml'),
('World News', 'http://feeds.reuters.com/reuters/worldNews?format=xml'), ('World News', 'http://feeds.reuters.com/reuters/worldNews?format=xml'),
@ -37,6 +54,4 @@ class Reuters(BasicNewsRecipe):
('Technology News', 'http://feeds.reuters.com/reuters/technologyNews?format=xml'), ('Technology News', 'http://feeds.reuters.com/reuters/technologyNews?format=xml'),
('Oddly Enough News', 'http://feeds.reuters.com/reuters/oddlyEnoughNews?format=xml') ('Oddly Enough News', 'http://feeds.reuters.com/reuters/oddlyEnoughNews?format=xml')
] ]
def print_version(self, url):
return ('http://www.reuters.com/article/id' + url + '?sp=true')