#!/usr/bin/env python __license__ = 'GPL v3' __copyright__ = '11 December 2010, desUBIKado' __author__ = 'desUBIKado' __description__ = 'Entertainment guide from Aragon' __version__ = 'v0.01' __date__ = '11, December 2010' ''' [url]http://www.redaragon.es/[/url] ''' from calibre.web.feeds.news import BasicNewsRecipe class heraldo(BasicNewsRecipe): __author__ = 'desUBIKado' description = u'Guia de ocio desde Aragon' title = u'RedAragon' publisher = 'Grupo Z' category = 'Concerts, Movies, Entertainment news' cover_url = 'http://www.redaragon.com/2008_img/logotipo.gif' language = 'es' timefmt = '[%a, %d %b, %Y]' oldest_article = 15 max_articles_per_feed = 100 encoding = 'iso-8859-1' use_embedded_content = False remove_javascript = True no_stylesheets = True feeds = [(u'Conciertos', u'http://redaragon.com/rss/agenda.asp?tid=1'), (u'Exposiciones', u'http://redaragon.com/rss/agenda.asp?tid=5'), (u'Teatro', u'http://redaragon.com/rss/agenda.asp?tid=10'), (u'Conferencias', u'http://redaragon.com/rss/agenda.asp?tid=2'), (u'Ferias', u'http://redaragon.com/rss/agenda.asp?tid=6'), (u'Filmotecas/Cineclubs', u'http://redaragon.com/rss/agenda.asp?tid=7'), (u'Presentaciones', u'http://redaragon.com/rss/agenda.asp?tid=9'), (u'Fiestas', u'http://redaragon.com/rss/agenda.asp?tid=11'), (u'Infantil', u'http://redaragon.com/rss/agenda.asp?tid=13'), (u'Otros', u'http://redaragon.com/rss/agenda.asp?tid=8')] keep_only_tags = [dict(name='div', attrs={'id': 'FichaEventoAgenda'})] remove_tags = [dict(name='div', attrs={ 'class': ['Comparte', 'CajaAgenda', 'Caja', 'Cintillo']})] remove_tags_before = dict(name='div', attrs={'id': 'FichaEventoAgenda'}) remove_tags_after = dict(name='div', attrs={'class': 'Cintillo'})