calibre/recipes/cronica.recipe
2018-01-10 23:46:09 +05:30

80 lines
2.7 KiB
Plaintext

# -*- mode: python -*-
# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
__copyright__ = '2018, Darko Miletic <darko.miletic at gmail.com>'
'''
www.cronica.com.ar
'''
from calibre.web.feeds.news import BasicNewsRecipe
class Veintitres(BasicNewsRecipe):
title = 'Cronica'
__author__ = 'Darko Miletic'
description = 'Últimas noticias'
publisher = 'Grupo Crónica'
category = 'politica, noticias generales, Argentina'
oldest_article = 15
max_articles_per_feed = 100
no_stylesheets = False
use_embedded_content = False
encoding = 'utf-8'
masthead_url = 'https://www.cronica.com.ar/export/sites/cronica/arte/logos/logoCronica.svg_799932565.svg'
language = 'es_AR'
remove_javascript = True
publication_type = 'magazine'
remove_empty_feeds = True
auto_cleanup = True
auto_cleanup_keep = '//h1'
resolve_internal_links = True
INDEX = "https://www.cronica.com.ar"
extra_css = """
img{margin-bottom: 0.8em}
"""
conversion_options = {
'comment': description,
'tags': category,
'publisher': publisher,
'language': language
}
feeds = [
(u'Policiales', u'https://www.cronica.com.ar/seccion/policiales/'),
(u'Politica', u'https://www.cronica.com.ar/seccion/politica/'),
(u'General', u'https://www.cronica.com.ar/seccion/info-general/'),
(u'Mundo', u'https://www.cronica.com.ar/seccion/mundo/'),
(u'Opinion', u'https://www.cronica.com.ar/seccion/opinion/'),
(u'Deportes', u'https://www.cronica.com.ar/seccion/deportes/'),
(u'Cosa de locos', u'https://www.cronica.com.ar/seccion/cosa-de-locos/'),
(u'Espectaculos', u'https://www.diarioshow.com/seccion/espectaculos/'),
]
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': 'cover-link'}):
url = self.INDEX + item['href']
if feedtitle == 'Espectaculos':
url = 'https://www.diarioshow.com' + item['href']
title = item['title']
articles.append({
'title': title,
'date': '',
'url': url,
'description': ''
})
totalfeeds.append((feedtitle, articles))
return totalfeeds