from calibre.web.feeds.news import BasicNewsRecipe class AdvancedUserRecipe1305547242(BasicNewsRecipe): title = u'Good House Keeping' language = 'en' __author__ = 'Anonymous' oldest_article = 7 max_articles_per_feed = 100 no_stylesheets = True use_embedded_content = False remove_javascript = True def print_version(self, url): if '/tips-for-making-desserts?' in url: return None segments = url.split('/') segments[-1] = segments[-1].split('?')[0] segments[-1] += '?page=all' printURL = '/'.join(segments[0:3]) + '/print-this/' + segments[-1] return printURL def preprocess_html(self, soup): for alink in soup.findAll('a'): if alink.string is not None: tstr = alink.string alink.replaceWith(tstr) return soup feeds = [ (u'Food and Recipes', u'http://www.goodhousekeeping.com/rss/recipes/'), (u'Home and Organizing', u'http://www.goodhousekeeping.com/rss/home/'), (u'Diet and Health', u'http://www.goodhousekeeping.com/rss/health/'), (u'Beauty and Anti-Aging', u'http://www.goodhousekeeping.com/rss/beauty/'), (u'Family and Relationships', u'http://www.goodhousekeeping.com/rss/family/'), (u'Holidays', u'http://www.goodhousekeeping.com/rss/holidays/'), (u'In the Test Kitchen', 'http://www.goodhousekeeping.com/rss/test-kitchen-blog/'), ]