from calibre.web.feeds.news import BasicNewsRecipe import re class PublicDomainReview(BasicNewsRecipe): title = 'The Public Domain Review' __author__ = 'fenuks' description = u'Online journal dedicated to showcasing the most interesting and unusual out-of-copyright works available on the web' cover_url = 'http://publicdomainreview.org/wp-content/themes/pdr/assets/img/pdr-logo.gif' category = 'culture' language = 'en' encoding = 'utf-8' oldest_article = 14 max_articles_per_feed = 100 no_stylesheets = True remove_empty_feeds = True ignore_duplicate_articles = {'title', 'url'} remove_javascript = True use_embedded_content = False # extra_css = '' # preprocess_regexps = [] # remove_attributes = ['style',] keep_only_tags = [dict(name='article', attrs={'class': re.compile(r'post-\d+')})] remove_tags = [dict(attrs={'class': 'synved-social-container synved-social-container-share'})] # remove_tags_before = dict() remove_tags_after = dict(name='div', attrs={'class': 'entry-content'}) feeds = [('Posts', 'http://publicdomainreview.org/feed/')]