#!/usr/bin/env python __license__ = 'GPL v3' __copyright__ = '2008-2011, Lionel Bergeret ' ''' cinebel.be ''' from calibre.web.feeds.news import BasicNewsRecipe class Cinebel(BasicNewsRecipe): title = u'Cinebel' __author__ = u'Lionel Bergeret' description = u'Cinema news from Belgium in French' publisher = u'cinebel.be' category = 'news, cinema, movie, Belgium' oldest_article = 15 language = 'fr' max_articles_per_feed = 20 no_stylesheets = True use_embedded_content = False timefmt = ' [%d %b %Y]' filterDuplicates = True keep_only_tags = [ dict(name='span', attrs={'class': 'movieMainTitle'}), dict(name='div', attrs={'id': 'filmPoster'}), dict( name='div', attrs={'id': 'filmDefinition'}), dict(name='div', attrs={'id': 'synopsis'}) ] feeds = [ (u'Les sorties de la semaine', u'http://www.cinebel.be/Servlets/RssServlet?languageCode=fr&rssType=0'), (u'Top 10', u'http://www.cinebel.be/Servlets/RssServlet?languageCode=fr&rssType=2') ] def preprocess_html(self, soup): for alink in soup.findAll('a', href=True): tstr = "Site officiel: " + alink['href'] alink.replaceWith(tstr) return soup def get_cover_url(self): cover_url = 'http://www.cinebel.be/portal/resources/common/logo_index.gif' return cover_url