Code:
import re
from calibre.web.feeds.news import BasicNewsRecipe
class AdvancedUserRecipe1306097511(BasicNewsRecipe):
title = u'Metro UK'
description = 'News as provide by The Metro -UK'
__author__ = 'Dave Asbury'
#last update 3/12/11
cover_url = 'http://profile.ak.fbcdn.net/hprofile-ak-snc4/276636_117118184990145_2132092232_n.jpg'
no_stylesheets = True
oldest_article = 1
max_articles_per_feed = 20
remove_empty_feeds = True
remove_javascript = True
#preprocess_regexps = [(re.compile(r'Tweet'), lambda a : '')]
preprocess_regexps = [
(re.compile(r'<span class="img-cap legend">', re.IGNORECASE | re.DOTALL), lambda match: '<p></p><span class="img-cap legend"> ')]
preprocess_regexps = [
(re.compile(r'tweet', re.IGNORECASE | re.DOTALL), lambda match: '')]
language = 'en_GB'
masthead_url = 'http://e-edition.metro.co.uk/images/metro_logo.gif'
keep_only_tags = [
dict(name='h1'),dict(name='h2', attrs={'class':'h2'}),
dict(attrs={'class':['img-cnt figure']}),
dict(attrs={'class':['art-img']}),
dict(name='div', attrs={'class':'art-lft'}),
dict(name='p')
]
remove_tags = [
dict(name = 'div',attrs={'id' : ['comments-news','formSubmission']}),
dict(name='div', attrs={'class':[ 'news m12 clrd clr-b p5t shareBtm', 'commentForm', 'metroCommentInnerWrap',
'art-rgt','pluck-app pluck-comm','news m12 clrd clr-l p5t', 'flt-r','username','clrd' ]}),
dict(attrs={'class':['username', 'metroCommentFormWrap','commentText','commentsNav','avatar','submDateAndTime','addYourComment','displayName']})
,dict(name='div', attrs={'class' : 'clrd art-fd fd-gr1-b'})
]
feeds = [
(u'News', u'http://www.metro.co.uk/rss/news/'), (u'Money', u'http://www.metro.co.uk/rss/money/'), (u'Sport', u'http://www.metro.co.uk/rss/sport/'), (u'Film', u'http://www.metro.co.uk/rss/metrolife/film/'), (u'Music', u'http://www.metro.co.uk/rss/metrolife/music/'), (u'TV', u'http://www.metro.co.uk/rss/tv/'), (u'Showbiz', u'http://www.metro.co.uk/rss/showbiz/'), (u'Weird News', u'http://www.metro.co.uk/rss/weird/'), (u'Travel', u'http://www.metro.co.uk/rss/travel/'), (u'Lifestyle', u'http://www.metro.co.uk/rss/lifestyle/'), (u'Books', u'http://www.metro.co.uk/rss/lifestyle/books/'), (u'Food', u'http://www.metro.co.uk/rss/lifestyle/restaurants/')]
extra_css = '''
body {font: sans-serif medium;}'
h1 {text-align : center; font-family:Arial,Helvetica,sans-serif; font-size:20px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold;}
h2 {text-align : center;color:#4D4D4D;font-family:Arial,Helvetica,sans-serif; font-size:15px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:bold; }
span{ font-size:9.5px; font-weight:bold;font-style:italic}
p { text-align: justify; font-family:Arial,Helvetica,sans-serif; font-size:11px; font-size-adjust:none; font-stretch:normal; font-style:normal; font-variant:normal; font-weight:normal;}
'''