Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-08-05 08:40:13 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update Spektrum der Wissenschaft

Browse Source
This commit is contained in:
Kovid Goyal 2012-10-17 19:12:29 +05:30
parent aadbe9b01e
commit eac0c5ddb4
1 changed files with 43 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
recipes/spektrum.recipe
View File

@ -1,28 +1,53 @@
__license__ = 'GPL v3'
__copyright__ = ''
'''
Fetch RSS-Feeds spektrum.de
'''
from calibre.web.feeds.recipes import BasicNewsRecipe from calibre.web.feeds.recipes import BasicNewsRecipe
class AdvancedUserRecipe1303841067(BasicNewsRecipe): class AdvancedUserRecipe1303841067(BasicNewsRecipe):
title = u'Spektrum (der Wissenschaft)' title = u'Spektrum (der Wissenschaft)'
__author__ = 'schuster' __author__ = 'Armin Geller, Bratzzo, Rainer Zenz' # Update Bratzzo & AGE 2012-10-12
oldest_article = 7 oldest_article = 7
max_articles_per_feed = 100 max_articles_per_feed = 100 #100
language = 'de' no_stylesheets = True
cover_url = 'http://upload.wikimedia.org/wikipedia/de/3/3b/Spektrum_der_Wissenschaft_Logo.svg' remove_javascript = True
remove_empty_feeds = True
language = 'de_DE'
# conversion_options = {'base_font_size': 20}
remove_tags = [dict(attrs={'class':['hauptnaviPkt gainlayout', 'hauptnaviButton', 'suchButton', 'suchbegriffKasten', 'loginButton', 'subnavigation', 'artikelInfoLeiste gainlayout', 'artikelTools', 'nurLetzteSeite', 'link', 'boxUnterArtikel', 'leserbriefeBlock', 'boxTitel', 'boxInhalt', 'sehrklein', 'boxabstand', 'werbeboxinhalt', 'rbabstand', 'bildlinks', 'rechtebox', 'denkmalbox', 'denkmalfrage']}), # cover_url = 'http://upload.wikimedia.org/wikipedia/de/3/3b/Spektrum_der_Wissenschaft_Logo.svg' # old logo
dict(id=['pflip', 'verlagsleiste', 'bereich', 'bannerVertikal', 'headerLogoLink', 'kopf', 'topNavi', 'headerSchnellsuche', 'headerSchnellsucheWarten', 'navigation', 'navigationL', 'navigationR', 'inhalt', 'rechtespalte', 'sdwboxenshop', 'shopboxen', 'fuss']), # cover_url = 'http://spektrumverlag.de/sixcms/media.php/776/thumbnails/sdwv_faecher.jpg.459370.jpg'
dict(name=['naservice'])] # cover_url = 'http://spektrumverlag.de/sixcms/media.php/776/thumbnails/sdwv_logo.jpg.432645.jpg'
cover_url = 'http://upload.wikimedia.org/wikipedia/de/5/59/Spektrum-cover.jpg' # from Rainer Zenz
def print_version(self,url): extra_css = '''
newurl = url.replace('artikel/', 'sixcms/detail.php?id=') h1 {font-size: 2em;}
return newurl + '&_druckversion=1' h2 {font-size: 1em; font-style: italic; font-weight: bold;}
h3 {font-size: 1em;}
h4, h5, h6, .heading, .hgroup {font-size: 1em;}
'''
feeds = [
(u'Spektrum.de', u'http://www.spektrum.de/alias/rss/spektrum-de-rss-feed/996406'),
(u'Gehirn & Geist', u'http://www.spektrum.de/alias/rss/gehirn-geist-rss-feed/982626'),
(u'epoc', u'http://www.spektrum.de/alias/rss/epoc-rss-feed/982625'),
(u'Spektrum der Wissenschaft', u'http://www.spektrum.de/alias/rss/spektrum-der-wissenschaft-rss-feed/982623'),
(u'Sterne und Weltraum', u'http://www.spektrum.de/alias/rss/sterne-und-weltraum-rss-feed/865248'),
(u'Editional', u'http://www.spektrum.de/alias/rss/spektrum-de-editorial/996044'),
(u'Pressemitteilungen', u'http://www.spektrum.de/alias/rss/pressemitteilungen/995265'),
]
keep_only_tags = [
dict(name='div', attrs={'class':'border-h clearfix article-top'}),
dict(name='div', attrs={'class':'clearfix'}),
dict(name='div', attrs={'class':'bilderrahmenlinks'}),
dict(name='div', attrs={'class':'relcontainer'}),
]
feeds = [(u'Spektrum der Wissenschaft', u'http://www.spektrum.de/artikel/982623'), remove_tags_after=dict(name='div', attrs={'class':['sidebar-box-head']})
(u'SpektrumDirekt', u'http://www.spektrumdirekt.de/artikel/996406'),
(u'Sterne und Weltraum', u'http://www.astronomie-heute.de/artikel/865248'),
(u'Gehirn & Geist', u'http://www.gehirn-und-geist.de/artikel/982626'),
(u'epoc', u'http://www.epoc.de/artikel/982625')
] remove_tags = [
dict(attrs={'id':['recommend-article', 'dossierbox', 'cover', 'toc']}),
filter_regexps = [r'ads\.doubleclick\.net'] dict(attrs={'class':['sidebar-box-full clearfix', 'linktotop' ]}),
]