Ok I think I have it working now. More work than I thought!
Let me know if you have any problems.
Spoiler:
Code:
__license__ = 'GPL v3'
__copyright__ = '2011, Pat Stapleton <pat.stapleton at gmail.com>'
'''
abc.net.au/news
'''
import re
from calibre.web.feeds.recipes import BasicNewsRecipe
class ABCNews(BasicNewsRecipe):
title = 'ABC News'
__author__ = 'Pat Stapleton, Dean Cording'
description = 'News from Australia'
masthead_url = 'http://www.abc.net.au/news/assets/v5/images/common/logo-news.png'
cover_url = 'http://www.abc.net.au/news/assets/v5/images/common/logo-news.png'
oldest_article = 2
max_articles_per_feed = 100
no_stylesheets = False
#delay = 1
use_embedded_content = False
encoding = 'utf8'
publisher = 'ABC News'
category = 'News, Australia, World'
language = 'en_AU'
publication_type = 'newsportal'
# preprocess_regexps = [(re.compile(r'<!--.*?-->', re.DOTALL), lambda m: '')]
#Remove annoying map links (inline-caption class is also used for some image captions! hence regex to match maps.google)
preprocess_regexps = [(re.compile(r'<a class="inline-caption" href="http://maps\.google\.com.*?/a>', re.DOTALL), lambda m: '')]
conversion_options = {
'comments' : description
,'tags' : category
,'language' : language
,'publisher' : publisher
,'linearize_tables': False
}
keep_only_tags = [dict(attrs={'class':['article section']})]
remove_tags = [dict(attrs={'class':['related', 'tags', 'tools', 'attached-content ready',
'inline-content story left', 'inline-content map left contracted', 'published',
'story-map', 'statepromo', 'topics', ]})]
remove_attributes = ['width','height']
feeds = [
('Top Stories', 'http://www.abc.net.au/news/feed/45910/rss.xml'),
('Canberra', 'http://www.abc.net.au/news/feed/6910/rss.xml'),
('Sydney', 'http://www.abc.net.au/news/feed/10232/rss.xml'),
('Melbourne', 'http://www.abc.net.au/news/feed/21708/rss.xml'),
('Brisbane', 'http://www.abc.net.au/news/feed/12858/rss.xml'),
('Perth', 'feed://www.abc.net.au/news/feed/24886/rss.xml'),
('Australia', 'http://www.abc.net.au/news/feed/46182/rss.xml'),
('World', 'http://www.abc.net.au/news/feed/52278/rss.xml'),
('Business', 'http://www.abc.net.au/news/feed/51892/rss.xml'),
('Science and Technology', 'http://www.abc.net.au/news/feed/2298/rss.xml'),
]
Enjoy!
-Pat