(Imroved El Pais) Fix #8095 (Completely new recipe for El Pais)

This commit is contained in:
Kovid Goyal 2010-12-29 11:51:59 -07:00
parent 85bf7da1fb
commit a323051dcd

View File

@ -1,86 +1,95 @@
# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
__copyright__ = '2010, Darko Miletic <darko.miletic at gmail.com>'
'''
www.elpais.com/diario/
www.elpais.com
'''
from calibre import strftime
from calibre.web.feeds.news import BasicNewsRecipe
class ElPaisImpresa(BasicNewsRecipe):
title = u'El Pa\xeds - edicion impresa'
class ElPais_RSS(BasicNewsRecipe):
title = 'El Pais'
__author__ = 'Darko Miletic'
description = u'el periodico global en Espa\xf1ol'
description = 'el periodico global en Castellano'
publisher = 'EDICIONES EL PAIS, S.L.'
category = 'news, politics,Spain,actualidad,noticias,informacion,videos,fotografias,audios,graficos,nacional,internacional,deportes,economia,tecnologia,cultura,gente,television,sociedad,opinion,blogs,foros,chats,encuestas,entrevistas,participacion'
category = 'news, politics, finances, world, spain'
oldest_article = 2
max_articles_per_feed = 200
no_stylesheets = True
encoding = 'latin1'
encoding = 'cp1252'
use_embedded_content = False
language = 'es'
language = 'es_ES'
remove_empty_feeds = True
publication_type = 'newspaper'
masthead_url = 'http://www.elpais.com/im/tit_logo_global.gif'
index = 'http://www.elpais.com/diario/'
extra_css = ' p{text-align: justify} body{ text-align: left; font-family: Georgia,"Times New Roman",Times,serif } h2{font-family: Arial,Helvetica,sans-serif} img{margin-bottom: 0.4em} '
masthead_url = 'http://www.elpais.com/im/tit_logo.gif'
extra_css = """
body{font-family: Georgia,"Times New Roman",Times,serif }
h3{font-family: Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block}
"""
conversion_options = {
'comment' : description
, 'tags' : category
, 'publisher' : publisher
, 'language' : language
'comment' : description
, 'tags' : category
, 'publisher' : publisher
, 'language' : language
}
feeds = [
(u'Internacional' , index + u'internacional/' )
,(u'Espa\xf1a' , index + u'espana/' )
,(u'Economia' , index + u'economia/' )
,(u'Opinion' , index + u'opinion/' )
,(u'Vi\xf1etas' , index + u'vineta/' )
,(u'Sociedad' , index + u'sociedad/' )
,(u'Cultura' , index + u'cultura/' )
,(u'Tendencias' , index + u'tendencias/' )
,(u'Gente' , index + u'gente/' )
,(u'Obituarios' , index + u'obituarios/' )
,(u'Deportes' , index + u'deportes/' )
,(u'Pantallas' , index + u'radioytv/' )
,(u'Ultima' , index + u'ultima/' )
,(u'Educacion' , index + u'educacion/' )
,(u'Saludo' , index + u'salud/' )
,(u'Ciberpais' , index + u'ciberpais/' )
,(u'EP3' , index + u'ep3/' )
,(u'Cine' , index + u'cine/' )
,(u'Babelia' , index + u'babelia/' )
,(u'El viajero' , index + u'viajero/' )
,(u'Negocios' , index + u'negocios/' )
,(u'Domingo' , index + u'domingo/' )
,(u'El Pais semanal' , index + u'eps/' )
,(u'Quadern Catalunya' , index + u'quadern-catalunya/' )
]
keep_only_tags = [dict(attrs={'class':['cabecera_noticia estirar','cabecera_noticia','','contenido_noticia']})]
remove_tags = [
dict(name=['meta','link','base','iframe','embed','object'])
,dict(attrs={'class':['info_complementa','estructura_2col_der','votos estirar','votos']})
,dict(attrs={'id':'utilidades'})
]
remove_tags_after = dict(attrs={'id':'utilidades'})
remove_attributes = ['lang','border','width','height']
keep_only_tags=[dict(attrs={'class':['cabecera_noticia','contenido_noticia']})]
remove_attributes=['width','height']
remove_tags=[dict(name='link')]
def parse_index(self):
totalfeeds = []
lfeeds = self.get_feeds()
for feedobj in lfeeds:
feedtitle, feedurl = feedobj
self.report_progress(0, _('Fetching feed')+' %s...'%(feedtitle if feedtitle else feedurl))
articles = []
soup = self.index_to_soup(feedurl)
for item in soup.findAll('a',attrs={'class':['g19r003','g19i003','g17r003','g17i003']}):
url = 'http://www.elpais.com' + item['href'].rpartition('/')[0]
title = self.tag_to_string(item)
date = strftime(self.timefmt)
articles.append({
'title' :title
,'date' :date
,'url' :url
,'description':''
})
totalfeeds.append((feedtitle, articles))
return totalfeeds
feeds = [
(u'Lo ultimo' , u'http://www.elpais.com/rss/feed.html?feedId=17046')
,(u'America Latina' , u'http://www.elpais.com/rss/feed.html?feedId=17041')
,(u'Mexico' , u'http://www.elpais.com/rss/feed.html?feedId=17042')
,(u'Europa' , u'http://www.elpais.com/rss/feed.html?feedId=17043')
,(u'Estados Unidos' , u'http://www.elpais.com/rss/feed.html?feedId=17044')
,(u'Oriente proximo' , u'http://www.elpais.com/rss/feed.html?feedId=17045')
,(u'Espana' , u'http://www.elpais.com/rss/feed.html?feedId=1002' )
,(u'Andalucia' , u'http://www.elpais.com/rss/feed.html?feedId=17057')
,(u'Catalunia' , u'http://www.elpais.com/rss/feed.html?feedId=17059')
,(u'Comunidad Valenciana' , u'http://www.elpais.com/rss/feed.html?feedId=17061')
,(u'Madrid' , u'http://www.elpais.com/rss/feed.html?feedId=1016' )
,(u'Pais Vasco' , u'http://www.elpais.com/rss/feed.html?feedId=17062')
,(u'Galicia' , u'http://www.elpais.com/rss/feed.html?feedId=17063')
,(u'Opinion' , u'http://www.elpais.com/rss/feed.html?feedId=1003' )
,(u'Sociedad' , u'http://www.elpais.com/rss/feed.html?feedId=1004' )
,(u'Deportes' , u'http://www.elpais.com/rss/feed.html?feedId=1007' )
,(u'Cultura' , u'http://www.elpais.com/rss/feed.html?feedId=1008' )
,(u'Cine' , u'http://www.elpais.com/rss/feed.html?feedId=17052')
,(u'Literatura' , u'http://www.elpais.com/rss/feed.html?feedId=17053')
,(u'Musica' , u'http://www.elpais.com/rss/feed.html?feedId=17051')
,(u'Arte' , u'http://www.elpais.com/rss/feed.html?feedId=17060')
,(u'Tecnologia' , u'http://www.elpais.com/rss/feed.html?feedId=1005' )
,(u'Economia' , u'http://www.elpais.com/rss/feed.html?feedId=1006' )
,(u'Ciencia' , u'http://www.elpais.com/rss/feed.html?feedId=17068')
,(u'Salud' , u'http://www.elpais.com/rss/feed.html?feedId=17074')
,(u'Ocio' , u'http://www.elpais.com/rss/feed.html?feedId=17075')
,(u'Justicia y Leyes' , u'http://www.elpais.com/rss/feed.html?feedId=17069')
,(u'Guerras y conflictos' , u'http://www.elpais.com/rss/feed.html?feedId=17070')
,(u'Politica' , u'http://www.elpais.com/rss/feed.html?feedId=17073')
]
def print_version(self, url):
return url + '?print=1'
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
for item in soup.findAll('a'):
if item.string is not None:
tstr = item.string
item.replaceWith(tstr)
else:
item.name='span'
for atrs in ['href','target','alt','title']:
if item.has_key(atrs):
del item[atrs]
for item in soup.findAll('img',alt=False):
item['alt'] = 'image'
return soup