mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-09-29 15:31:08 -04:00
92 lines
3.7 KiB
Python
92 lines
3.7 KiB
Python
#!/usr/bin/env python
|
|
# -*- coding: utf-8 -*-
|
|
from __future__ import absolute_import, division, print_function, unicode_literals
|
|
|
|
__license__ = 'GPL v3'
|
|
__copyright__ = '2009, Gerhard Aigner <gerhard.aigner at gmail.com>'
|
|
|
|
''' http://www.derstandard.at - Austrian Newspaper '''
|
|
|
|
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
|
|
def classes(classes):
|
|
q = frozenset(classes.split(' '))
|
|
return dict(attrs={
|
|
'class': lambda x: x and frozenset(x.split()).intersection(q)})
|
|
|
|
|
|
class DerStandardRecipe(BasicNewsRecipe):
|
|
title = u'derStandard'
|
|
__author__ = 'Gerhard Aigner and Sujata Raman and Marcel Jira and Peter Reschenhofer'
|
|
description = u'Nachrichten aus Österreich'
|
|
publisher = 'derStandard.at'
|
|
category = 'news, politics, nachrichten, Austria'
|
|
use_embedded_content = False
|
|
remove_empty_feeds = True
|
|
no_stylesheets = True
|
|
encoding = 'utf-8'
|
|
language = 'de_AT'
|
|
|
|
oldest_article = 1
|
|
max_articles_per_feed = 100
|
|
ignore_duplicate_articles = {'title', 'url'}
|
|
|
|
masthead_url = 'http://images.derstandard.at/2012/06/19/derStandardat_1417x274.gif'
|
|
|
|
feeds = [
|
|
(u'International', u'https://www.derstandard.at/rss/international'),
|
|
(u'Inland', u'https://www.derstandard.at/rss/inland'),
|
|
(u'Wirtschaft', u'https://www.derstandard.at/rss/wirtschaft'),
|
|
(u'Web', u'https://www.derstandard.at/rss/web'),
|
|
(u'Sport', u'https://derStandard.at/rss/Sport'),
|
|
(u'Panorama', u'https://derStandard.at/rss/Panorama'),
|
|
(u'Etat', u'https://derStandard.at/rss/Etat'),
|
|
(u'Kultur', u'https://derStandard.at/rss/Kultur'),
|
|
(u'Wissenschaft', u'https://derStandard.at/rss/Wissenschaft'),
|
|
(u'Gesundheit', u'https://derStandard.at/rss/Gesundheit'),
|
|
# (u'Bildung', u'https://derStandard.at/rss/Bildung'),
|
|
# (u'Meinung', u'https://derStandard.at/rss/Meinung'),
|
|
(u'Lifestyle', u'https://derStandard.at/rss/Lifestyle'),
|
|
# (u'Reisen', u'https://derStandard.at/rss/Reisen'),
|
|
# (u'Familie', u'https://derstandard.at/rss/Familie'),
|
|
# (u'Meinung', u'https://derStandard.at/rss/Meinung'),
|
|
# (u'User', u'https://derStandard.at/rss/User'),
|
|
(u'Karriere', u'https://derStandard.at/rss/Karriere'),
|
|
# (u'Immobilien', u'https://derstandard.at/rss/Immobilien'),
|
|
# (u'Automobil', u'https://derstandard.at/rss/Automobil'),
|
|
(u'dieStandard', u'https://derStandard.at/rss/diestandard'),
|
|
(u'Newsroom', u'https://www.derstandard.at/rss'),
|
|
]
|
|
|
|
def get_browser(self):
|
|
br = BasicNewsRecipe.get_browser(self)
|
|
br.set_simple_cookie('DSGVO_Check', '', '.derstandard.at')
|
|
headers = {
|
|
'X-Requested-With': 'XMLHttpRequest',
|
|
'Content-Type': 'application/json; charset=UTF-8',
|
|
'DNT': '1',
|
|
'Pragma': 'no-cache',
|
|
'Cache-Control': 'no-cache'
|
|
}
|
|
import mechanize
|
|
req = mechanize.Request(url='https://apps.derstandard.at/privacyprotection/api/agree', data=None, headers=headers, method='POST')
|
|
br.open(req)
|
|
return br
|
|
|
|
keep_only_tags = [
|
|
classes('article-header article-body article-origins article-subtitle article-pubdate'),
|
|
]
|
|
|
|
remove_tags = [
|
|
dict(name=['link', 'iframe', 'style', 'hr']),
|
|
dict(attrs={'class': ['lookup-links', 'media-list']}),
|
|
dict(name='form', attrs={'name': 'sitesearch'}),
|
|
dict(name='div', attrs={'class': ['socialsharing', 'block video',
|
|
'blog-browsing section',
|
|
'diashow', 'supplemental']}),
|
|
dict(name='div', attrs={'id': 'highlighted'})
|
|
]
|
|
|
|
remove_attributes = ['width', 'height']
|