calibre/recipes/opinion_bo.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

80 lines
3.0 KiB
Plaintext

# -*- coding: utf-8 -*-
__license__ = 'GPL v3'
__copyright__ = '2011, Piet van Oostrum <piet@vanoostrum.org>'
'''
www.opinion.com.bo
'''
from calibre import strftime
from calibre.web.feeds.news import BasicNewsRecipe
class Opinion_Bol(BasicNewsRecipe):
title = u'Opinión - Bolivia'
__author__ = 'Piet van Oostrum'
description = u'Opinión diario de circulación nacional, Cochabamba, Bolivia'
publisher = 'Coboce Ltda - Editora Opinión'
category = 'news, politics, Bolivia'
version = 1
oldest_article = 1
max_articles_per_feed = 20
no_stylesheets = True
encoding = 'utf-8'
use_embedded_content = False
language = 'es_BO'
publication_type = 'newspaper'
delay = 1
remove_empty_feeds = True
cover_url = strftime(
'http://www.opinion.com.bo/opinion/articulos/%Y/%m%d/fotos/portada_650.jpg')
masthead_url = 'http://opinion.com.bo/opinion/articulos/imagenes/logo_opinion.gif'
extra_css = """body{font-family: Helvetica,Arial,sans-serif}
.seccion_encabezado_nota_inte{font-size: 1.1em;
font-weight: bold;}
.autor_nota_inte{color: #999999; font-size: 0.8em;
margin-bottom: 0.5em; text-align: right;}
.pie{font-size: 0.8em;}"""
conversion_options = {
'comment': description, 'tags': category, 'publisher': publisher, 'language': language
}
keep_only_tags = [
dict(name='div', attrs={'class': 'columna_izq_nota_intererior'})]
remove_tags = [dict(name=['meta', 'link', 'form', 'iframe', 'embed', 'object', 'style']),
dict(name='div', attrs={'class': 'ocultar'})]
remove_attributes = ['width', 'height']
feeds = [
(u'El País', u'http://www.opinion.com.bo/opinion/rss/el_pais_rss.xml'),
(u'Cochabamba', u'http://www.opinion.com.bo/opinion/rss/cochabamba_rss.xml'),
(u'Economía', u'http://www.opinion.com.bo/opinion/rss/economia_rss.xml'),
(u'Cultura', u'http://www.opinion.com.bo/opinion/rss/cultura_rss.xml'),
(u'Mundo', u'http://www.opinion.com.bo/opinion/rss/mundo_rss.xml'),
(u'Ciencia y Tecnología', u'http://www.opinion.com.bo/opinion/rss/ciencia_tecnologia_rss.xml'),
(u'Policial', u'http://www.opinion.com.bo/opinion/rss/policial_rss.xml'),
(u'Editorial', u'http://www.opinion.com.bo/opinion/rss/editorial_rss.xml'),
(u'Subeditorial', u'http://www.opinion.com.bo/opinion/rss/subeditorial_rss.xml'),
(u'Opinión', u'http://www.opinion.com.bo/opinion/rss/opinion_rss.xml'),
(u'Deportes', u'http://www.opinion.com.bo/opinion/rss/deportes_rss.xml'),
(u' Vida de hoy', u'http://www.opinion.com.bo/opinion/rss/vidadehoy_rss.xml')
]
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
return soup
# Filter out today's articles
# maybe should take timezone into account
today = strftime('/%Y/%m%d/')
def get_article_url(self, article):
link = article.link
if self.today in link:
return link