import random from calibre import browser from calibre.web.feeds.recipes import BasicNewsRecipe class AdvancedUserRecipe1325006965(BasicNewsRecipe): title = u'The Sun UK' description = 'Articles from The Sun tabloid UK' __author__ = 'Dave Asbury' # last updated 5/5/13 better cover fetch language = 'en_GB' oldest_article = 1 max_articles_per_feed = 15 remove_empty_feeds = True masthead_url = 'http://www.thesun.co.uk/sol/img/global/Sun-logo.gif' encoding = 'UTF-8' remove_javascript = True no_stylesheets = True ignore_duplicate_articles = {'title', 'url'} compress_news_images = True keep_only_tags = [ dict(name='div', attrs={'class': 'intro'}), dict(name='h3'), dict(name='div', attrs={'id': 'articlebody'}), ] remove_tags_after = [dict(id='bodyText')] remove_tags = [ dict(name='li'), dict(attrs={'class': 'grid-4 right-hand-column'}), ] feeds = [ (u'News', u'http://www.thesun.co.uk/sol/homepage/news/rss'), (u'Sport', u'http://www.thesun.co.uk/sol/homepage/sport/rss'), (u'Showbiz', u'http://www.thesun.co.uk/sol/homepage/showbiz/rss'), (u'Woman', u'http://www.thesun.co.uk/sol/homepage/woman/rss'), ] # starsons code def parse_feeds(self): feeds = BasicNewsRecipe.parse_feeds(self) for feed in feeds: for article in feed.articles[:]: if 'Try out The Sun' in article.title.upper() or 'Try-out-The-Suns' in article.url: feed.articles.remove(article) if 'Web porn harms kids' in article.title.upper() or 'Sun-says-Web-porn' in article.url: feed.articles.remove(article) return feeds def get_cover_url(self): br = browser() br.set_handle_redirect(False) cover_url = 'http://www.thepaperboy.com/frontpages/current/The_Sun_newspaper_front_page.jpg' try: br.open_novisit( 'http://www.thepaperboy.com/frontpages/current/The_Sun_newspaper_front_page.jpg') except: cover_url = random.choice([ 'http://img.thesun.co.uk/multimedia/archive/00905/errorpage6_677961a_905507a.jpg', 'http://img.thesun.co.uk/multimedia/archive/00905/errorpage7_677962a_905505a.jpg', 'http://img.thesun.co.uk/multimedia/archive/00905/errorpage5_677960a_905512a.jpg', 'http://img.thesun.co.uk/multimedia/archive/00905/errorpage2_677957a_905502a.jpg', 'http://img.thesun.co.uk/multimedia/archive/00905/errorpage3_677958a_905503a.jpg' # noqa ]) return cover_url