From 8af58a4af2c258594878e3122496a31868671cc5 Mon Sep 17 00:00:00 2001 From: Kovid Goyal Date: Tue, 26 Sep 2017 08:35:49 +0530 Subject: [PATCH] Update El Periodico de Aragon --- recipes/el_periodico.recipe | 84 +++++++++++++++++++++++++++++-------- 1 file changed, 66 insertions(+), 18 deletions(-) diff --git a/recipes/el_periodico.recipe b/recipes/el_periodico.recipe index 966976ce9d..53d5af0fe4 100644 --- a/recipes/el_periodico.recipe +++ b/recipes/el_periodico.recipe @@ -1,12 +1,12 @@ -#!/usr/bin/env python2 +#!/usr/bin/env python2 # -*- coding: utf-8 -*- __license__ = 'GPL v3' __copyright__ = '04 December 2010, desUBIKado' __author__ = 'desUBIKado' __description__ = 'Daily newspaper from Aragon' -__version__ = 'v0.09' -__date__ = '07, August 2013' +__version__ = 'v0.10' +__date__ = '09, September 2017' ''' elperiodicodearagon.com ''' @@ -20,7 +20,7 @@ class elperiodicodearagon(BasicNewsRecipe): publisher = u'elperiodicodearagon.com' category = u'news, politics, Spain, Aragon' oldest_article = 1 - delay = 0 + delay = 1 max_articles_per_feed = 100 no_stylesheets = True use_embedded_content = False @@ -30,10 +30,6 @@ class elperiodicodearagon(BasicNewsRecipe): remove_empty_feeds = True remove_javascript = True - conversion_options = { - 'comments': description, 'tags': category, 'language': language, 'publisher': publisher - } - feeds = [ (u'Portada', u'http://zetaestaticos.com/aragon/rss/portada_es.xml'), (u'Arag\xf3n', u'http://zetaestaticos.com/aragon/rss/2_es.xml'), @@ -42,7 +38,7 @@ class elperiodicodearagon(BasicNewsRecipe): (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'CAI Zaragoza', u'http://zetaestaticos.com/aragon/rss/91_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'), @@ -50,15 +46,66 @@ class elperiodicodearagon(BasicNewsRecipe): (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'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_attributes = ['height', 'width'] + 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={'id': 'Noticia'})] + 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) - # 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) @@ -67,7 +114,8 @@ class elperiodicodearagon(BasicNewsRecipe): return 'http://pdf.elperiodicodearagon.com' + image['src'] return None - # Usamos la versión para móviles - - def print_version(self, url): - return url.replace('http://www.elperiodicodearagon.com/', 'http://www.elperiodicodearagon.com/m/') + 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;} + '''