Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-07 18:24:30 -04:00
Code Issues Packages Projects Releases Wiki Activity

Add profile for Associated Press

Browse Source
This commit is contained in:
Kovid Goyal 2008-01-31 17:24:29 +00:00
parent f623f07af7
commit 4d6e1fef72
2 changed files with 39 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
src/libprs500/ebooks/lrf/web/profiles/ap.py Normal file
View File

@ -0,0 +1,38 @@
import re
from libprs500.ebooks.lrf.web.profiles import DefaultProfile
class AssociatedPress(DefaultProfile):
title = 'Associated Press'
max_recursions = 2
max_articles_per_feed = 15
html2lrf_options = ['--force-page-break-before-tag="chapter"']
preprocess_regexps = [ (re.compile(i[0], re.IGNORECASE | re.DOTALL), i[1]) for i in
[
(r'<HEAD>.*?</HEAD>' , lambda match : '<HEAD></HEAD>'),
(r'<body class="apple-rss-no-unread-mode" onLoad="setup(null)">.*?<!-- start Entries -->', lambda match : '<body>'),
(r'<!-- end apple-rss-content-area -->.*?</body>', lambda match : '</body>'),
(r'<script.*?>.*?</script>', lambda match : ''),
(r'<body.*?>.*?<span class="headline">', lambda match : '<body><span class="headline"><chapter>'),
(r'<tr><td><div class="body">.*?<p class="ap-story-p">', lambda match : '<p class="ap-story-p">'),
(r'<p class="ap-story-p">', lambda match : '<p>'),
(r'Learn more about our <a href="http://apdigitalnews.com/privacy.html">Privacy Policy</a>.*?</body>', lambda match : '</body>'),
]
]
def get_feeds(self):
return [ ('AP Headlines', 'http://hosted.ap.org/lineups/TOPHEADS-rss_2.0.xml?SITE=ORAST&SECTION=HOME'),
('AP US News', 'http://hosted.ap.org/lineups/USHEADS-rss_2.0.xml?SITE=CAVIC&SECTION=HOME'),
('AP World News', 'http://hosted.ap.org/lineups/WORLDHEADS-rss_2.0.xml?SITE=SCAND&SECTION=HOME'),
('AP Political News', 'http://hosted.ap.org/lineups/POLITICSHEADS-rss_2.0.xml?SITE=ORMED&SECTION=HOME'),
('AP Washington News', 'http://hosted.ap.org/lineups/WASHINGTONHEADS-rss_2.0.xml?SITE=NYPLA&SECTION=HOME'),
('AP Technology News', 'http://hosted.ap.org/lineups/TECHHEADS-rss_2.0.xml?SITE=NYWNE&SECTION=HOME'),
('AP Health News', 'http://hosted.ap.org/lineups/HEALTHHEADS-rss_2.0.xml?SITE=FLDAY&SECTION=HOME'),
('AP Science News', 'http://hosted.ap.org/lineups/SCIENCEHEADS-rss_2.0.xml?SITE=OHCIN&SECTION=HOME'),
('AP Strange News', 'http://hosted.ap.org/lineups/STRANGEHEADS-rss_2.0.xml?SITE=WCNC&SECTION=HOME'),
]

2
src/libprs500/ebooks/lrf/web/profiles/reuters.py
View File

@ -30,7 +30,7 @@ class Reuters(DefaultProfile):
('World News', 'http://feeds.reuters.com/reuters/worldNews?format=xml'), ('World News', 'http://feeds.reuters.com/reuters/worldNews?format=xml'),
('Politics News', 'http://feeds.reuters.com/reuters/politicsNews?format=xml'), ('Politics News', 'http://feeds.reuters.com/reuters/politicsNews?format=xml'),
('Science News', 'http://feeds.reuters.com/reuters/scienceNews?format=xml'), ('Science News', 'http://feeds.reuters.com/reuters/scienceNews?format=xml'),
('Emviroment News', 'http://feeds.reuters.com/reuters/Environment?format=xml'), ('Environment News', 'http://feeds.reuters.com/reuters/Environment?format=xml'),
('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')
] ]