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

Fix #1768 (Jerusalem Post recipe error)

Browse Source
This commit is contained in:
Kovid Goyal 2009-02-04 15:31:52 -08:00
parent 6a6ce5a4f9
commit 4b609b699b
1 changed files with 16 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
src/calibre/web/feeds/recipes/recipe_jpost.py
View File

@ -1,37 +1,29 @@
import re
from calibre.web.feeds.news import BasicNewsRecipe from calibre.web.feeds.news import BasicNewsRecipe
class JerusalemPost(BasicNewsRecipe): class JerusalemPost(BasicNewsRecipe):
title = 'Jerusalem Post' title = 'Jerusalem Post'
description = 'News from Israel and the Middle East' description = 'News from Israel and the Middle East'
use_embedded_content = False use_embedded_content = False
__author__ = 'Kovid Goyal' __author__ = 'Kovid Goyal'
max_articles_per_feed = 10 max_articles_per_feed = 10
remove_tags_before = {'class':'byline'}
remove_tags = [
preprocess_regexps = [ (re.compile(i[0], re.IGNORECASE | re.DOTALL), i[1]) for i in {'class':['artAdBlock clearboth', 'tbartop', 'divdot_vrttbox',
[ 'slideshow']},
(r'<HEAD>.*?</HEAD>' , lambda match : '<HEAD></HEAD>'), dict(id=['artFontButtons', 'artRelatedBlock']),
(r'<BODY.*?>.*?<!-- start Entries -->', lambda match : '<BODY><!-- start Entries -->'), ]
(r'<!-- end Entries -->.*?</BODY>', lambda match : '</BODY>'), remove_tags_after = {'id':'artTxtBlock'}
(r'<script.*?>.*?</script>', lambda match : ''),
(r'<div class="apple-rss-article apple-rss-read" onclick=.*?<div class="apple-rss-article-body">', lambda match : ''),
(r'<img src=\'/images/logo_NWAnews.gif\' alt=\'NWAnews.com :: Northwest Arkansas\' News Source\'.*?>', lambda match : ''),
(r'<img src=\'/images/logo_adg.gif\'.*?>', lambda match : ''),
(r'<P CLASS="smallprint">.*?</body>', lambda match : '</body>'),
]
]
feeds = [ ('Front Page', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333346'), feeds = [ ('Front Page', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333346'),
('Israel News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1178443463156'), ('Israel News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1178443463156'),
('Middle East News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333498'), ('Middle East News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333498'),
('International News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1178443463144'), ('International News', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1178443463144'),
('Editorials', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333211'), ('Editorials', 'http://www.jpost.com/servlet/Satellite?pagename=JPost/Page/RSS&cid=1123495333211'),
] ]
def print_version(self, url): def postprocess_html(self, soup, first):
return ('http://www.jpost.com/servlet/Satellite?cid=' + url.rpartition('&')[2] + '&pagename=JPost%2FJPArticle%2FPrinter') for tag in soup.findAll(name=['table', 'tr', 'td']):
tag.name = 'div'
return soup