calibre/recipes/catavencii.recipe
2011-11-04 18:44:44 +05:30

52 lines
1.5 KiB
Plaintext

# -*- coding: utf-8 -*-
#!/usr/bin/env python
__license__ = 'GPL v3'
__copyright__ = u'2011, Silviu Cotoar\u0103'
'''
catavencii.ro
'''
from calibre.web.feeds.news import BasicNewsRecipe
class Catavencii(BasicNewsRecipe):
title = u'Ca\u0163avencii'
__author__ = u'Silviu Cotoar\u0103'
publisher = u'Ca\u0163avencii'
description = u'Ca\u0163avencii'
oldest_article = 5
language = 'ro'
max_articles_per_feed = 100
no_stylesheets = True
use_embedded_content = False
category = 'Ziare,Romania'
encoding = 'utf-8'
cover_url = 'http://www.simonatache.ro/wp-content/uploads/2011/06/catavencii-logo.png'
conversion_options = {
'comments' : description
,'tags' : category
,'language' : language
,'publisher' : publisher
}
keep_only_tags = [
dict(name='div', attrs={'id':'content'})
]
remove_tags = [
dict(name='div', attrs={'id':'breadcrumbs'})
, dict(name='span', attrs={'class':'info'})
, dict(name='div', attrs={'id':'social-media-article'})
]
remove_tags_after = [
dict(name='div', attrs={'id':'social-media-article'})
]
feeds = [
(u'\u0218tiri', u'http://www.catavencii.ro/rss')
]
def preprocess_html(self, soup):
return self.adeify_images(soup)