mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-09-29 15:31:08 -04:00
51 lines
1.9 KiB
Plaintext
51 lines
1.9 KiB
Plaintext
__license__ = 'GPL v3'
|
|
__copyright__ = '2009-2010, Darko Miletic <darko.miletic at gmail.com>'
|
|
'''
|
|
emol.com
|
|
'''
|
|
|
|
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
class ElMercurio(BasicNewsRecipe):
|
|
title = 'El Mercurio online'
|
|
__author__ = 'Darko Miletic'
|
|
description = 'El sitio de noticias online de Chile'
|
|
publisher = 'El Mercurio'
|
|
category = 'news, politics, Chile'
|
|
oldest_article = 2
|
|
max_articles_per_feed = 100
|
|
no_stylesheets = True
|
|
use_embedded_content = False
|
|
encoding = 'cp1252'
|
|
masthead_url = 'http://www.emol.com/especiales/logo_emol/logo_emol.gif'
|
|
remove_javascript = True
|
|
use_embedded_content = False
|
|
language = 'es_CL'
|
|
|
|
|
|
conversion_options = {
|
|
'comment' : description
|
|
, 'tags' : category
|
|
, 'publisher' : publisher
|
|
, 'language' : language
|
|
}
|
|
|
|
keep_only_tags = [dict(name='div', attrs={'id':['cont_iz_titulobajada','cont_iz_creditos_1_a','cont_iz_cuerpo']})]
|
|
remove_tags = [dict(name='div', attrs={'id':'cont_iz_cuerpo_relacionados'})]
|
|
remove_attributes = ['height','width']
|
|
|
|
feeds = [
|
|
(u'Noticias de ultima hora', u'http://rss.emol.com/rss.asp?canal=0')
|
|
,(u'Nacional', u'http://rss.emol.com/rss.asp?canal=1')
|
|
,(u'Mundo', u'http://rss.emol.com/rss.asp?canal=2')
|
|
,(u'Deportes', u'http://rss.emol.com/rss.asp?canal=4')
|
|
,(u'Magazine', u'http://rss.emol.com/rss.asp?canal=6')
|
|
,(u'Tecnologia', u'http://rss.emol.com/rss.asp?canal=5')
|
|
]
|
|
|
|
def preprocess_html(self, soup):
|
|
for item in soup.findAll(style=True):
|
|
del item['style']
|
|
return soup
|
|
|