calibre/recipes/el_periodico.recipe
Kovid Goyal 29cd8d64ea
Change shebangs to python from python2
Also remove a few other miscellaneous references to python2
2020-08-22 18:47:51 +05:30

122 lines
4.8 KiB
Python

#!/usr/bin/env python
# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
__copyright__ = '04 December 2010, desUBIKado'
__author__ = 'desUBIKado'
__description__ = 'Daily newspaper from Aragon'
__version__ = 'v0.10'
__date__ = '09, September 2017'
'''
elperiodicodearagon.com
'''
from calibre.web.feeds.news import BasicNewsRecipe
class elperiodicodearagon(BasicNewsRecipe):
title = u'El Periodico de Aragon'
__author__ = u'desUBIKado'
description = u'Noticias desde Aragon'
publisher = u'elperiodicodearagon.com'
category = u'news, politics, Spain, Aragon'
oldest_article = 1
delay = 1
max_articles_per_feed = 100
no_stylesheets = True
use_embedded_content = False
language = 'es'
masthead_url = 'http://pdf.elperiodicodearagon.com/img/logotipo.gif'
encoding = 'iso-8859-1'
remove_empty_feeds = True
remove_javascript = True
feeds = [
(u'Portada', u'http://zetaestaticos.com/aragon/rss/portada_es.xml'),
(u'Arag\xf3n', u'http://zetaestaticos.com/aragon/rss/2_es.xml'),
(u'Internacional', u'http://zetaestaticos.com/aragon/rss/4_es.xml'),
(u'Espa\xf1a', u'http://zetaestaticos.com/aragon/rss/3_es.xml'),
(u'Econom\xeda', u'http://zetaestaticos.com/aragon/rss/5_es.xml'),
(u'Deportes', u'http://zetaestaticos.com/aragon/rss/7_es.xml'),
(u'Real Zaragoza', u'http://zetaestaticos.com/aragon/rss/10_es.xml'),
(u'Tecnyconta Zaragoza', u'http://zetaestaticos.com/aragon/rss/91_es.xml'),
(u'Monta\xf1ismo', u'http://zetaestaticos.com/aragon/rss/354_es.xml'),
(u'Opini\xf3n', u'http://zetaestaticos.com/aragon/rss/103_es.xml'),
(u'Tema del d\xeda', u'http://zetaestaticos.com/aragon/rss/102_es.xml'),
(u'Escenarios', u'http://zetaestaticos.com/aragon/rss/105_es.xml'),
(u'Sociedad', u'http://zetaestaticos.com/aragon/rss/104_es.xml'),
(u'Gente', u'http://zetaestaticos.com/aragon/rss/330_es.xml'),
(u'Espacio 3', u'http://zetaestaticos.com/aragon/rss/328_es.xml'),
(u'Fiestas del Pilar', u'http://zetaestaticos.com/aragon/rss/107_es.xml'),
(u'Semana Santa', u'http://zetaestaticos.com/aragon/rss/385_es.xml'), (
u'La crónica de Valdejal\xf3n',
u'http://zetaestaticos.com/aragon/rss/206_es.xml'
), (
u'La crónica de Campo de Borja',
u'http://zetaestaticos.com/aragon/rss/208_es.xml'
), (
u'La crónica de Ejea y sus pueblos',
u'http://zetaestaticos.com/aragon/rss/212_es.xml'
), (
u'La crónica del Bajo Gállego',
u'http://zetaestaticos.com/aragon/rss/205_es.xml'
), (
u'La crónica del Campo de Cariñena',
u'http://zetaestaticos.com/aragon/rss/207_es.xml'
), (
u'La crónica de la Ribera Alta del Ebro',
u'http://zetaestaticos.com/aragon/rss/211_es.xml'
), (
u'La crónica del Campo de Belchite',
u'http://zetaestaticos.com/aragon/rss/331_es.xml'
)
]
remove_tags_before = dict(name='div', attrs={'class': 'Pagina'})
remove_tags_after = dict(name='div', attrs={'class': 'ComentariosNew'})
keep_only_tags = [dict(name='div', attrs={'class': 'Pagina'})]
remove_tags = [
dict(
name='nav',
attrs={'class': ['Compartir', 'HerramientasConversacion Herramientas']}
),
dict(name='h5', attrs={'class': ['CintilloBox']}),
dict(
name='div',
attrs={
'class': [
'BoxMenu BoxMenuConFoto', 'BxGalerias', 'ConStick',
'HerramientasComentarioNew Herramientas', 'NumeroComentarioNew'
]
}
),
dict(
name='div',
attrs={
'class': [
'BoxPestanas', 'Box', 'ColumnaDerecha',
'NoticiasRelacionadasDeNoticia',
'CintilloNoticiasRelacionadasDeNoticia'
]
}
),
dict(name='a', attrs={'class': ['IrA BotonLink']})
]
# Recuperamos la portada de papel (la imagen format=1 tiene mayor resolucion)
def get_cover_url(self):
index = 'http://pdf.elperiodicodearagon.com/edicion.php'
soup = self.index_to_soup(index)
for image in soup.findAll('img', src=True):
if image['src'].startswith('/funciones/img-public.php?key='):
return 'http://pdf.elperiodicodearagon.com' + image['src']
return None
extra_css = '''
h1 {font-family:Arial,Helvetica,sans-serif; font-weight:bold;font-size:28px;}
h2 {font-family:Arial,Helvetica,sans-serif; font-style:italic;font-size:14px;color:#4D4D4D;}
h3 {font-family:Arial,Helvetica,sans-serif; font-weight:bold;font-size:18px;}
'''