calibre/resources/recipes/blic.recipe
2010-03-28 09:06:39 +05:30

48 lines
2.2 KiB
Plaintext

__license__ = 'GPL v3'
__copyright__ = '2008-2010, Darko Miletic <darko.miletic at gmail.com>'
'''
blic.rs
'''
import re
from calibre.web.feeds.news import BasicNewsRecipe
class Blic(BasicNewsRecipe):
title = 'Blic'
__author__ = 'Darko Miletic'
description = 'Blic.rs online verzija najtiraznije novine u Srbiji donosi najnovije vesti iz Srbije i sveta, komentare, politicke analize, poslovne i ekonomske vesti, vesti iz regiona, intervjue, informacije iz kulture, reportaze, pokriva sve sportske dogadjaje, detaljan tv program, nagradne igre, zabavu, fenomenalni Blic strip, dnevni horoskop, arhivu svih dogadjaja'
publisher = 'RINGIER d.o.o.'
category = 'news, politics, Serbia'
oldest_article = 2
max_articles_per_feed = 100
no_stylesheets = True
use_embedded_content = False
masthead_url = 'http://www.blic.rs/resources/images/header/header_back.png'
language = 'sr'
publication_type = 'newspaper'
extra_css = '@font-face {font-family: "serif1";src:url(res:///opt/sony/ebook/FONT/tt0011m_.ttf)} @font-face {font-family: "sans1";src:url(res:///opt/sony/ebook/FONT/tt0003m_.ttf)} body{font-family: Georgia, serif1, serif} .article_description{font-family: Arial, sans1, sans-serif} .img_full{float: none} img{margin-bottom: 0.8em} '
conversion_options = {
'comment' : description
, 'tags' : category
, 'publisher': publisher
, 'language' : language
}
preprocess_regexps = [(re.compile(u'\u0110'), lambda match: u'\u00D0')]
remove_tags_before = dict(name='div', attrs={'id':'article_info'})
remove_tags = [dict(name=['object','link'])]
remove_attributes = ['width','height']
feeds = [(u'Danasnje Vesti', u'http://www.blic.rs/rss/danasnje-vesti')]
def print_version(self, url):
return url + '/print'
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
return self.adeify_images(soup)