Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-06-23 15:30:45 -04:00
Code Issues Packages Projects Releases Wiki Activity

Il Messaggero and ADN Kronos by Gabriele Marini

Browse Source
This commit is contained in:
Kovid Goyal 2010-05-03 08:31:59 -06:00
parent 32f8611a76
commit c9bcd9a35b
2 changed files with 115 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
resources/recipes/adnkronos.recipe Normal file
View File

@ -0,0 +1,59 @@
#!/usr/bin/env python
__license__ = 'GPL v3'
__author__ = 'Gabriele Marini, based on Darko Miletic'
__copyright__ = '2009-2010, Darko Miletic <darko.miletic at gmail.com>'
description = 'Italian daily newspaper - 02-05-2010'
'''
http://www.adnkronos.com/
'''
from calibre.web.feeds.news import BasicNewsRecipe
class Adnkronos(BasicNewsRecipe):
__author__ = 'Gabriele Marini'
description = 'News agency'
cover_url = 'http://www.adnkronos.com/IGN6/img/popup_ign.jpg'
title = u'Adnkronos'
publisher = 'Adnkronos Group - ews agency'
category = 'News, politics, culture, economy, general interest'
language = 'it'
timefmt = '[%a, %d %b, %Y]'
oldest_article = 7
max_articles_per_feed = 80
use_embedded_content = False
recursion = 10
remove_javascript = True
def get_article_url(self, article):
link = article.get('id', article.get('guid', None))
return link
extra_css = ' .newsAbstract{font-style: italic} '
keep_only_tags = [dict(name='div', attrs={'class':['breadCrumbs','newsTop','newsText']})
]
remove_tags = [
dict(name='div', attrs={'class':['leogoo','leogoo2']})
]
feeds = [
(u'Prima Pagina', u'http://rss.adnkronos.com/RSS_PrimaPagina.xml'),
(u'Ultima Ora', u'http://rss.adnkronos.com/RSS_Ultimora.xml'),
(u'Politica', u'http://rss.adnkronos.com/RSS_Politica.xml'),
(u'Esteri', u'http://rss.adnkronos.com/RSS_Esteri.xml'),
(u'Cronoca', u'http://rss.adnkronos.com/RSS_Cronaca.xml'),
(u'Economia', u'http://rss.adnkronos.com/RSS_Economia.xml'),
(u'Finanza', u'http://rss.adnkronos.com/RSS_Finanza.xml'),
(u'CyberNews', u'http://rss.adnkronos.com/RSS_CyberNews.xml'),
(u'Spettacolo', u'http://rss.adnkronos.com/RSS_Spettacolo.xml'),
(u'Cultura', u'http://rss.adnkronos.com/RSS_Cultura.xml'),
(u'Sport', u'http://rss.adnkronos.com/RSS_Sport.xml'),
(u'Sostenibilita', u'http://rss.adnkronos.com/RSS_Sostenibilita.xml'),
(u'Salute', u'http://rss.adnkronos.com/RSS_Salute.xml')
]

56
resources/recipes/il_messaggero.recipe Normal file
View File

@ -0,0 +1,56 @@
#!/usr/bin/env python
__license__ = 'GPL v3'
__author__ = 'Gabriele Marini, based on Darko Miletic'
__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>'
description = 'Italian daily newspaper - v1.01 (04, January 2010)'
'''
http://www.messaggero.it/
'''
from calibre.web.feeds.news import BasicNewsRecipe
class IlMessaggero(BasicNewsRecipe):
__author__ = 'Gabriele Marini'
description = 'Italian News'
cover_url = 'http://www.ilmessaggero.it/img_tst/logomsgr.gif'
title = u'Il Messaggero'
publisher = 'Caltagirone Editore'
category = 'News, politics, culture, economy, general interest'
language = 'it'
timefmt = '[%a, %d %b, %Y]'
oldest_article = 5
max_articles_per_feed = 100
use_embedded_content = False
recursion = 10
remove_javascript = True
keep_only_tags = [dict(name='h1', attrs={'class':'titoloLettura2'}),
dict(name='h2', attrs={'class':'sottotitLettura'}),
dict(name='span', attrs={'class':'testoArticoloG'})
]
feeds = [
(u'HomePage', u'http://www.ilmessaggero.it/rss/home.xml'),
(u'Primo Piano', u'http://www.ilmessaggero.it/rss/initalia_primopiano.xml'),
(u'Cronaca Bianca', u'http://www.ilmessaggero.it/rss/initalia_cronacabianca.xml'),
(u'Cronaca Nera', u'http://www.ilmessaggero.it/rss/initalia_cronacanera.xml'),
(u'Economia e Finanza', u'http://www.ilmessaggero.it/rss/economia.xml'),
(u'Politica', u'http://www.ilmessaggero.it/rss/initalia_politica.xml'),
(u'Scienza e Tecnologia', u'http://www.ilmessaggero.it/rss/scienza.xml'),
(u'Cinema', u'http://www.ilmessaggero.it/rss.php?refresh_ce#'),
(u'Viaggi', u'http://www.ilmessaggero.it/rss.php?refresh_ce#'),
(u'Roma', u'http://www.ilmessaggero.it/rss/roma.xml'),
(u'Cultura e Tendenze', u'http://www.ilmessaggero.it/rss/roma_culturaspet.xml'),
(u'Sport', u'http://www.ilmessaggero.it/rss/sport.xml'),
(u'Calcio', u'http://www.ilmessaggero.it/rss/sport_calcio.xml'),
(u'Motori', u'http://www.ilmessaggero.it/rss/sport_motori.xml')
]