__license__ = 'GPL v3' __copyright__ = '2010, Darko Miletic ' ''' www.alo.rs ''' import re from calibre.web.feeds.recipes import BasicNewsRecipe class Alo_Novine(BasicNewsRecipe): title = 'Alo!' __author__ = 'Darko Miletic' description = "News Portal from Serbia" publisher = 'Alo novine d.o.o.' category = 'news, politics, Serbia' oldest_article = 2 max_articles_per_feed = 100 delay = 4 no_stylesheets = True encoding = 'utf-8' use_embedded_content = False language = 'sr' extra_css = """ @font-face {font-family: "sans1";src:url(res:///opt/sony/ebook/FONT/tt0003m_.ttf)} .article_description,body{font-family: Arial,Helvetica,sans1,sans-serif} .lead {font-size: 1.3em} h1{color: #DB0700} .article_uvod{font-style: italic; font-size: 1.2em} img{margin-bottom: 0.8em} """ conversion_options = { 'comment' : description , 'tags' : category , 'publisher': publisher , 'language' : language } preprocess_regexps = [(re.compile(u'\u0110'), lambda match: u'\u00D0')] remove_tags = [dict(name=['object','link','embed'])] remove_attributes = ['height','width'] feeds = [ (u'Najnovije Vijesti', u'http://www.alo.rs/rss/danasnje_vesti') ,(u'Politika' , u'http://www.alo.rs/rss/politika') ,(u'Vesti' , u'http://www.alo.rs/rss/vesti') ,(u'Sport' , u'http://www.alo.rs/rss/sport') ,(u'Ljudi' , u'http://www.alo.rs/rss/ljudi') ,(u'Saveti' , u'http://www.alo.rs/rss/saveti') ] def preprocess_html(self, soup): for item in soup.findAll(style=True): del item['style'] return soup def print_version(self, url): artl = url.rpartition('/')[0] artid = artl.rpartition('/')[2] return 'http://www.alo.rs/resources/templates/tools/print.php?id=' + artid def image_url_processor(self, baseurl, url): return url.replace('alo.rs//','alo.rs/')