Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Improved recipe for El Pais

Browse Source
This commit is contained in:
Kovid Goyal 2010-01-10 09:16:18 -07:00
parent 8caf640ca9
commit fa49e42680
1 changed files with 48 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
resources/recipes/el_pais.recipe
View File

@ -1,6 +1,8 @@
#!/usr/bin/env python
__license__ = 'GPL v3'
__author__ = 'Lorenzo Vigentini, based on earlier version by Kovid Goyal'
__copyright__ = '2008, Kovid Goyal kovid@kovidgoyal.net'
description = 'Main daily newspaper from Spain - v1.02 (10, January 2010)'
__docformat__ = 'restructuredtext en'
'''
@ -10,18 +12,54 @@ elpais.es
from calibre.web.feeds.news import BasicNewsRecipe
class ElPais(BasicNewsRecipe):
title = u'EL PAIS'
language = 'es'
__author__ = 'Kovid Goyal & Lorenzo Vigentini'
description = 'Main daily newspaper from Spain'
oldest_article = 7
max_articles_per_feed = 100
cover_url = 'http://www.elpais.com/im/tit_logo_global.gif'
title = u'El Pais'
publisher = 'Ediciones El Pais SL'
category = 'News, politics, culture, economy, general interest'
language = 'es'
timefmt = '[%a, %d %b, %Y]'
oldest_article = 2
max_articles_per_feed = 15
use_embedded_content = False
recursion = 5
remove_javascript = True
no_stylesheets = True
remove_tags = [dict(name='div', attrs={'class':'zona_superior'}), dict(name='div', attrs={'class':'limpiar'}), dict(name='div', attrs={'id':'pie'})]
extra_css = 'h1 {font: sans-serif large;} \n h2 {font: sans-serif medium;} \n h3 {font: sans-serif small;} \n h4 {font: sans-serif bold small;} \n p{ font:10pt serif}'
keep_only_tags = [ dict(name='div', attrs={'class':['cabecera_noticia','cabecera_noticia_reportaje','contenido_noticia','caja_despiece','presentacion']})]
extra_css = '''
p{style:normal size:12 serif}
feeds = [(u'Internacional', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporint'), (u'Espana', u'http://www.elpais.es/rss/rss_section.html?anchor=elppornac'), (u'Deportes', u'http://www.elpais.es/rss/rss_section.html?anchor=elppordep'), (u'Economia', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporeco'), (u'Tecnologia', u'http://www.elpais.es/rss/rss_section.html?anchor=elpportec'), (u'Cultura', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporcul'), (u'Gente', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporgen'), (u'Sociedad', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporsoc'), (u'Opinion', u'http://www.elpais.es/rss/rss_section.html?anchor=elpporopi')]
'''
def print_version(self, url):
url = url+'?print=1'
return url
remove_tags = [
dict(name='div', attrs={'class':['zona_superior','pie_enlaces_inferiores','contorno_f','ampliar']}),
dict(name='div', attrs={'class':['limpiar','mod_apoyo','borde_sup','votos','info_complementa','info_relacionada']}),
dict(name='div', attrs={'id':['suscribirse suscrito','google_noticia','utilidades','coment','foros_not','pie','lomas']})
]
feeds = [
(u'Titulares de portada', u'http://www.elpais.com/rss/feed.html?feedId=1022'),
(u'Internacional', u'http://www.elpais.com/rss/feed.html?feedId=1001'),
(u'Espana', u'http://www.elpais.com/rss/feed.html?feedId=1002'),
(u'Deportes', u'http://www.elpais.com/rss/feed.html?feedId=1007'),
(u'Economia', u'http://www.elpais.com/rss/feed.html?feedId=1006'),
(u'Politica', u'http://www.elpais.com/rss/feed.html?feedId=17073'),
(u'Tecnologia', u'http://www.elpais.com/rss/feed.html?feedId=1005'),
(u'Cultura', u'http://www.elpais.com/rss/feed.html?feedId=1008'),
(u'Gente', u'http://www.elpais.com/rss/feed.html?feedId=1009'),
(u'Sociedad', u'http://www.elpais.com/rss/feed.html?feedId=1004'),
(u'Opinion', u'http://www.elpais.com/rss/feed.html?feedId=1003'),
(u'Ciencia', u'http://www.elpais.com/rss/feed.html?feedId=17068'),
(u'Justicia y leyes', u'http://www.elpais.com/rss/feed.html?feedId=17069'),
]
def print_version(self, url):
url = url+'?print=1'
return url