From dcf45ce7975818cedb58391aba27601a8af7a0bd Mon Sep 17 00:00:00 2001 From: Kovid Goyal Date: Sat, 15 Oct 2011 06:13:04 +0530 Subject: [PATCH] Update various German recipes --- recipes/brand_eins.recipe | 5 +++-- recipes/ftd.recipe | 28 ++++++++++++++-------------- recipes/sueddeutsche.recipe | 6 +++--- recipes/telepolis.recipe | 9 ++------- 4 files changed, 22 insertions(+), 26 deletions(-) diff --git a/recipes/brand_eins.recipe b/recipes/brand_eins.recipe index e6fe57b334..277af4d789 100644 --- a/recipes/brand_eins.recipe +++ b/recipes/brand_eins.recipe @@ -110,8 +110,10 @@ class BrandEins(BasicNewsRecipe): selected_issue = issue_map[selected_issue_key] url = selected_issue.get('href', False) # Get the title for the magazin - build it out of the title of the cover - take the issue and year; - self.title = "brand eins " + selected_issue_key[4:] + "/" + selected_issue_key[0:4] + # self.title = "brand eins " + selected_issue_key[4:] + "/" + selected_issue_key[0:4] + # Get the alternative title for the magazin - build it out of the title of the cover - without the issue and year; url = 'http://brandeins.de/'+url + self.timefmt = ' ' + selected_issue_key[4:] + '/' + selected_issue_key[:4] # url = "http://www.brandeins.de/archiv/magazin/tierisch.html" titles_and_articles = self.brand_eins_parse_issue(url) @@ -163,4 +165,3 @@ class BrandEins(BasicNewsRecipe): current_articles.append({'title': title, 'url': url, 'description': description, 'date':''}) titles_and_articles.append([chapter_title, current_articles]) return titles_and_articles - diff --git a/recipes/ftd.recipe b/recipes/ftd.recipe index bdd084c329..1bda3dacfb 100644 --- a/recipes/ftd.recipe +++ b/recipes/ftd.recipe @@ -16,7 +16,7 @@ class FTDe(BasicNewsRecipe): use_embedded_content = False timefmt = ' [%d %b %Y]' language = 'de' - max_articles_per_feed = 40 + max_articles_per_feed = 30 no_stylesheets = True remove_tags = [dict(id='navi_top'), @@ -84,19 +84,19 @@ class FTDe(BasicNewsRecipe): dict(name='div', attrs={'class':'artikelsplitfaq'})] #remove_tags_after = [dict(name='a', attrs={'class':'more'})] - feeds = [ ('Finanzen', 'http://www.ftd.de/rss2/finanzen/maerkte'), - ('Meinungshungrige', 'http://www.ftd.de/rss2/meinungshungrige'), - ('Unternehmen', 'http://www.ftd.de/rss2/unternehmen'), - ('Politik', 'http://www.ftd.de/rss2/politik'), - ('Karriere_Management', 'http://www.ftd.de/rss2/karriere-management'), - ('IT_Medien', 'http://www.ftd.de/rss2/it-medien'), - ('Wissen', 'http://www.ftd.de/rss2/wissen'), - ('Sport', 'http://www.ftd.de/rss2/sport'), - ('Auto', 'http://www.ftd.de/rss2/auto'), - ('Lifestyle', 'http://www.ftd.de/rss2/lifestyle') - - ] + feeds = [ + ('Unternehmen', 'http://www.ftd.de/rss2/unternehmen'), + ('Finanzen', 'http://www.ftd.de/rss2/finanzen/maerkte'), + ('Meinungen', 'http://www.ftd.de/rss2/meinungshungrige'), + ('Politik', 'http://www.ftd.de/rss2/politik'), + ('Management & Karriere', 'http://www.ftd.de/rss2/karriere-management'), + ('IT & Medien', 'http://www.ftd.de/rss2/it-medien'), + ('Wissen', 'http://www.ftd.de/rss2/wissen'), + ('Sport', 'http://www.ftd.de/rss2/sport'), + ('Auto', 'http://www.ftd.de/rss2/auto'), + ('Lifestyle', 'http://www.ftd.de/rss2/lifestyle') + ] def print_version(self, url): - return url.replace('.html', '.html?mode=print') + return url.replace('.html', '.html?mode=print') \ No newline at end of file diff --git a/recipes/sueddeutsche.recipe b/recipes/sueddeutsche.recipe index c3863c6584..e0dc6742a1 100644 --- a/recipes/sueddeutsche.recipe +++ b/recipes/sueddeutsche.recipe @@ -10,7 +10,7 @@ from calibre.web.feeds.news import BasicNewsRecipe class Sueddeutsche(BasicNewsRecipe): - title = u'Süddeutsche' + title = u'sueddeutsche.de' description = 'News from Germany' __author__ = 'Oliver Niesner and Armin Geller' use_embedded_content = False @@ -62,7 +62,7 @@ class Sueddeutsche(BasicNewsRecipe): (u'Sport', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5ESport%24?output=rss'), (u'Leben', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5ELeben%24?output=rss'), (u'Karriere', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EKarriere%24?output=rss'), - (u'München&Region', u'http://www.sueddeutsche.de/app/service/rss/ressort/muenchen/rss.xml'), + (u'München & Region', u'http://www.sueddeutsche.de/app/service/rss/ressort/muenchen/rss.xml'), (u'Bayern', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EBayern%24?output=rss'), (u'Medien', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EMedien%24?output=rss'), (u'Digital', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EDigital%24?output=rss'), @@ -75,7 +75,7 @@ class Sueddeutsche(BasicNewsRecipe): (u'Job', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EJob%24?output=rss'), # sometimes only (u'Service', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EService%24?output=rss'), # sometimes only (u'Verlag', u'http://suche.sueddeutsche.de/query/%23/sort/-docdatetime/drilldown/%C2%A7ressort%3A%5EVerlag%24?output=rss'), # sometimes only - ] + ] def print_version(self, url): main, sep, id = url.rpartition('/') diff --git a/recipes/telepolis.recipe b/recipes/telepolis.recipe index 3611ebc642..3bc39014d5 100644 --- a/recipes/telepolis.recipe +++ b/recipes/telepolis.recipe @@ -3,7 +3,7 @@ from calibre.web.feeds.news import BasicNewsRecipe class TelepolisNews(BasicNewsRecipe): - title = u'Telepolis (News+Artikel)' + title = u'Telepolis' __author__ = 'syntaxis' publisher = 'Heise Zeitschriften Verlag GmbH & Co KG' description = 'News from Telepolis' @@ -15,11 +15,8 @@ class TelepolisNews(BasicNewsRecipe): encoding = "utf-8" language = 'de' - remove_empty_feeds = True - - keep_only_tags = [dict(name = 'div',attrs={'class':'head'}),dict(name = 'div',attrs={'class':'leftbox'}),dict(name='td',attrs={'class':'strict'})] remove_tags = [ dict(name='td',attrs={'class':'blogbottom'}), dict(name='div',attrs={'class':'forum'}), dict(name='div',attrs={'class':'social'}),dict(name='div',attrs={'class':'blog-letter p-news'}), @@ -28,7 +25,6 @@ class TelepolisNews(BasicNewsRecipe): remove_tags_after = [dict(name='span', attrs={'class':['breadcrumb']})] - feeds = [(u'News', u'http://www.heise.de/tp/news-atom.xml')] html2lrf_options = [ @@ -39,8 +35,7 @@ class TelepolisNews(BasicNewsRecipe): html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"' - def preprocess_html(self, soup): mtag = '' soup.head.insert(0,mtag) - return soup + return soup \ No newline at end of file