Fix #803147 (Updated recipe for El Universal from Venezuela)

This commit is contained in:
Kovid Goyal 2011-06-28 14:51:50 -06:00
parent ccb8d330c1
commit f07d04dc33

View File

@ -1,5 +1,5 @@
__license__ = 'GPL v3' __license__ = 'GPL v3'
__copyright__ = '2010, Darko Miletic <darko.miletic at gmail.com>' __copyright__ = '2010-2011, Darko Miletic <darko.miletic at gmail.com>'
''' '''
www.eluniversal.com www.eluniversal.com
''' '''
@ -15,12 +15,20 @@ class ElUniversal(BasicNewsRecipe):
max_articles_per_feed = 100 max_articles_per_feed = 100
no_stylesheets = True no_stylesheets = True
use_embedded_content = False use_embedded_content = False
remove_empty_feeds = True
encoding = 'cp1252' encoding = 'cp1252'
publisher = 'El Universal' publisher = 'El Universal'
category = 'news, Caracas, Venezuela, world' category = 'news, Caracas, Venezuela, world'
language = 'es_VE' language = 'es_VE'
publication_type = 'newspaper'
cover_url = strftime('http://static.eluniversal.com/%Y/%m/%d/portada.jpg') cover_url = strftime('http://static.eluniversal.com/%Y/%m/%d/portada.jpg')
extra_css = """
.txt60{font-family: Tahoma,Geneva,sans-serif; font-size: small}
.txt29{font-family: Tahoma,Geneva,sans-serif; font-size: small; color: gray}
.txt38{font-family: Georgia,"Times New Roman",Times,serif; font-size: xx-large}
.txt35{font-family: Georgia,"Times New Roman",Times,serif; font-size: large}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
"""
conversion_options = { conversion_options = {
'comments' : description 'comments' : description
,'tags' : category ,'tags' : category
@ -28,10 +36,11 @@ class ElUniversal(BasicNewsRecipe):
,'publisher' : publisher ,'publisher' : publisher
} }
keep_only_tags = [dict(name='div', attrs={'class':'Nota'})] remove_tags_before=dict(attrs={'class':'header-print MB10'})
remove_tags_after= dict(attrs={'id':'SizeText'})
remove_tags = [ remove_tags = [
dict(name=['object','link','script','iframe']) dict(name=['object','link','script','iframe','meta'])
,dict(name='div',attrs={'class':'Herramientas'}) ,dict(attrs={'class':'header-print MB10'})
] ]
feeds = [ feeds = [