Fix #769692 (Updated recipe for Clarin)

This commit is contained in:
Kovid Goyal 2011-04-23 19:30:38 -06:00
parent 28922fd4e4
commit 2b6561b7aa

View File

@ -1,6 +1,6 @@
__license__ = 'GPL v3'
__copyright__ = '2008-2010, Darko Miletic <darko.miletic at gmail.com>'
__copyright__ = '2008-2011, Darko Miletic <darko.miletic at gmail.com>'
'''
clarin.com
'''
@ -18,11 +18,18 @@ class Clarin(BasicNewsRecipe):
use_embedded_content = False
no_stylesheets = True
encoding = 'utf8'
delay = 1
language = 'es_AR'
publication_type = 'newspaper'
INDEX = 'http://www.clarin.com'
masthead_url = 'http://www.clarin.com/static/CLAClarin/images/logo-clarin-print.jpg'
extra_css = ' body{font-family: Arial,Helvetica,sans-serif} h2{font-family: Georgia,serif; font-size: xx-large} .hora{font-weight:bold} .hd p{font-size: small} .nombre-autor{color: #0F325A} '
extra_css = """
body{font-family: Arial,Helvetica,sans-serif}
h2{font-family: Georgia,serif; font-size: xx-large}
.hora{font-weight:bold}
.hd p{font-size: small}
.nombre-autor{color: #0F325A}
"""
conversion_options = {
'comment' : description
@ -32,6 +39,8 @@ class Clarin(BasicNewsRecipe):
}
keep_only_tags = [dict(attrs={'class':['hd','mt']})]
remove_tags = [dict(name=['meta','base','link'])]
remove_attributes = ['lang','_mce_bogus']
feeds = [
(u'Pagina principal', u'http://www.clarin.com/rss/' )
@ -47,6 +56,10 @@ class Clarin(BasicNewsRecipe):
,(u'Ciudades' , u'http://www.clarin.com/rss/ciudades/' )
]
def get_article_url(self, article):
return article.get('guid', None)
def print_version(self, url):
return url + '?print=1'