#!/usr/bin/env python __license__ = 'GPL v3' __copyright__ = '2010, Tomasz Dlugosz ' ''' nczas.com ''' from calibre.web.feeds.news import BasicNewsRecipe # class NCzas(BasicNewsRecipe): title = u'Najwy\u017cszy Czas!' description = u'Najwy\u017cszy Czas!\nwydanie internetowe' __author__ = u'Tomasz D\u0142ugosz' language = 'pl' oldest_article = 7 max_articles_per_feed = 100 no_stylesheets = True cover_url = 'http://nczas.com/wp-content/themes/default/grafika/logo.png' keep_only_tags = [dict(name='div', attrs={'class':'trescartykulu'})] feeds = [(u'Najwy\u017cszy Czas!', u'http://nczas.com/feed/')] def postprocess_html(self, soup, first): for tag in soup.findAll(name= 'img', alt=""): tag.extract() for item in soup.findAll(align = "right"): del item['align'] return soup