mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-06-23 15:30:45 -04:00
45 lines
2.1 KiB
Plaintext
45 lines
2.1 KiB
Plaintext
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
|
|
class DziennikBaltycki(BasicNewsRecipe):
|
|
title = u'Dziennik Ba\u0142tycki'
|
|
__author__ = 'fenuks'
|
|
description = u'Gazeta Regionalna Dziennik Bałtycki. Najnowsze Wiadomości Trójmiasto i Wiadomości Pomorskie. Czytaj!'
|
|
category = 'newspaper'
|
|
language = 'pl'
|
|
masthead_url = 'http://s.polskatimes.pl/g/logo_naglowek/dziennikbaltycki.png?24'
|
|
oldest_article = 7
|
|
max_articles_per_feed = 100
|
|
remove_empty_feeds = True
|
|
no_stylesheets = True
|
|
use_embedded_content = False
|
|
ignore_duplicate_articles = {'title', 'url'}
|
|
keep_only_tags = [dict(name='section', attrs={'class': 'zajawka'}),
|
|
dict(name='section', attrs={'id': 'tresc'})
|
|
]
|
|
remove_tags = [dict(name='div', attrs={'class':['materialyZKategorii','materialyPodobne']}), dict(name='a', attrs={
|
|
'class': 'czytajDalej'}), dict(attrs={'src': 'http://nm.dz.com.pl/dz.png'})]
|
|
|
|
feeds = [
|
|
(u'Wiadomo\u015bci', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_wiadomosci.xml?201302'),
|
|
(u'Sport', u'http://dziennikbaltycki.feedsportal.com/c/32980/f/533756/index.rss?201302'),
|
|
(u'Rejsy', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_rejsy.xml?201302'),
|
|
(u'Biznes na Pomorzu', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_biznesnapomorzu.xml?201302'),
|
|
(u'GOM', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_gom.xml?201302'),
|
|
(u'Opinie', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_opinie.xml?201302'),
|
|
(u'Pitawal Pomorski', u'http://www.dziennikbaltycki.pl/rss/dziennikbaltycki_pitawalpomorski.xml?201302')]
|
|
|
|
def print_version(self, url):
|
|
return url.replace('artykul', 'drukuj')
|
|
|
|
def skip_ad_pages(self, soup):
|
|
if 'Advertisement' in soup.title:
|
|
nexturl = soup.find('a')['href']
|
|
return self.index_to_soup(nexturl, raw=True)
|
|
|
|
def get_cover_url(self):
|
|
soup = self.index_to_soup(
|
|
'http://www.prasa24.pl/gazeta/dziennik-baltycki/')
|
|
self.cover_url = soup.find(id='pojemnik').img['src']
|
|
return getattr(self, 'cover_url', self.cover_url)
|