diff --git a/resources/recipes/nzz_ger.recipe b/resources/recipes/nzz_ger.recipe index 66e22f50ef..09e011a1dc 100644 --- a/resources/recipes/nzz_ger.recipe +++ b/resources/recipes/nzz_ger.recipe @@ -1,6 +1,6 @@ __license__ = 'GPL v3' -__copyright__ = '2009, Darko Miletic ' +__copyright__ = '2009-2010, Darko Miletic ' ''' www.nzz.ch @@ -20,6 +20,19 @@ class Nzz(BasicNewsRecipe): encoding = 'utf-8' use_embedded_content = False language = 'de' + extra_css = """ + body{font-family: Georgia,"Times New Roman",Times,serif } + .artikel h3,.artikel h4,.bildLegende,.question,.autor{font-family: Arial,Verdana,Helvetica,sans-serif} + .bildLegende{font-size: small} + .autor{font-size: 0.9375em; color: #666666} + .quote{font-size: large !important; + font-style: italic; + font-weight: normal !important; + border-bottom: 1px dotted #BFBFBF; + border-top: 1px dotted #BFBFBF; + line-height: 1.25em} + .quelle{color: #666666; font-style: italic; white-space: nowrap} + """ conversion_options = { 'comments' : description @@ -28,12 +41,14 @@ class Nzz(BasicNewsRecipe): ,'publisher' : publisher } - keep_only_tags = [dict(name='div', attrs={'class':'article'})] - + keep_only_tags = [dict(name='div', attrs={'class':'zone'})] + remove_tags_before = dict(name='p', attrs={'class':'dachzeile'}) + remove_tags_after=dict(name='p', attrs={'class':'fussnote'}) + remove_attributes=['width','height','lang'] remove_tags = [ - dict(name=['object','link','base']) - ,dict(name='div',attrs={'class':['more','teaser','advXertXoriXals','legal']}) - ,dict(name='div',attrs={'id':['popup-src','readercomments','google-ad','advXertXoriXals']}) + dict(name=['object','link','base','meta','iframe']) + ,dict(attrs={'id':'content_rectangle_1'}) + ,dict(attrs={'class':['weiterfuehrendeLinks','fussnote','video']}) ] feeds = [ @@ -50,7 +65,7 @@ class Nzz(BasicNewsRecipe): ,(u'Reisen' , u'http://www.nzz.ch/magazin/reisen?rss=true') ] - def print_version(self, url): - return url + '?printview=true' - - + def preprocess_html(self, soup): + for item in soup.findAll(style=True): + del item['style'] + return self.adeify_images(soup)