calibre/recipes/libertad_digital.recipe
Kovid Goyal 567040ee1e Perform PEP8 compliance checks on the entire codebase
Some bits of PEP 8 are turned off via setup.cfg
2016-07-29 21:25:17 +05:30

64 lines
2.4 KiB
Plaintext

__license__ = 'GPL v3'
__copyright__ = '2013, Darko Miletic <darko.miletic at gmail.com>'
'''
www.libertaddigital.com
'''
from calibre.web.feeds.news import BasicNewsRecipe
class LibertadDigital(BasicNewsRecipe):
title = 'Libertad Digital'
__author__ = 'Darko Miletic'
description = 'En Libertad Digital encontraras noticias y opinion sobre: España, el Mundo, Internet, sociedad, economia y deportes'
publisher = 'Libertad Digital S.A.'
category = 'noticias, ultima hora, españa, internet, mundo, economia, sociedad, Libertad Digital'
oldest_article = 2
max_articles_per_feed = 200
no_stylesheets = True
encoding = 'cp1252'
use_embedded_content = False
language = 'es'
remove_empty_feeds = True
publication_type = 'website'
masthead_url = 'http://s.libertaddigital.com/images/logo.gif'
extra_css = """
body{font-family: Verdana,sans-serif }
img{margin-bottom: 0.4em; display:block}
"""
conversion_options = {
'comment': description, 'tags': category, 'publisher': publisher, 'language': language
}
remove_tags = [
dict(name=['meta', 'link', 'iframe', 'embed', 'object']), dict(
name='p', attrs={'class': 'copyright'})
]
remove_attributes = ['lang']
feeds = [
(u'Portada', u'http://feeds2.feedburner.com/libertaddigital/deportes'),
(u'Opinion', u'http://feeds2.feedburner.com/libertaddigital/opinion'),
(u'España', u'http://feeds2.feedburner.com/libertaddigital/nacional'),
(u'Internacional', u'http://feeds2.feedburner.com/libertaddigital/internacional'),
(u'Libre Mercado', u'http://feeds2.feedburner.com/libertaddigital/economia'),
(u'Chic', u'http://feeds2.feedburner.com/libertaddigital/el-candelabro'),
(u'Internet', u'http://feeds2.feedburner.com/libertaddigital/internet'),
(u'Deportes', u'http://feeds2.feedburner.com/libertaddigital/deportes')
]
def get_article_url(self, article):
return article.get('guid', None)
def print_version(self, url):
art, sep, rest = url.rpartition('/')
aart, asep, artid = art.rpartition('-')
return 'http://www.libertaddigital.com/c.php?op=imprimir&id=' + artid
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
return soup