calibre/recipes/deutsche_welle_en.recipe
2023-06-18 19:47:15 +05:30

44 lines
1.5 KiB
Python

from calibre.web.feeds.news import BasicNewsRecipe, classes
class DeutscheWelle_en(BasicNewsRecipe):
title = 'Deutsche Welle'
__author__ = 'unkn0wn'
description = 'News from Germany and the world'
publisher = 'Deutsche Welle'
language = 'en'
oldest_article = 2
max_articles_per_feed = 50
no_stylesheets = True
remove_javascript = True
remove_empty_feeds = True
ignore_duplicate_articles = {'title', 'url'}
remove_attributes = ['height', 'width', 'style']
keep_only_tags = [
dict(name='article')
]
remove_tags = [
dict(name=['footer', 'source']),
dict(attrs={'data-tracking-name':'sharing-icons-inline'}),
classes('kicker advertisement vjs-wrapper')
]
feeds = [
('World', 'http://rss.dw.de/rdf/rss-en-world'),
('Germany', 'http://rss.dw.de/rdf/rss-en-ger'),
('Europe', 'http://rss.dw.de/rdf/rss-en-eu'),
('Business', 'http://rss.dw.de/rdf/rss-en-bus'),
('Culture & Lifestyle', 'http://rss.dw.de/rdf/rss-en-cul'),
('Sports', 'http://rss.dw.de/rdf/rss-en-sports'),
('Visit Germany', 'http://rss.dw.de/rdf/rss-en-visitgermany'),
('Asia', 'http://rss.dw.de/rdf/rss-en-asia'),
('Top Stories', 'http://rss.dw-world.de/rdf/rss-en-top'),
]
def preprocess_html(self, soup):
for img in soup.findAll('img', srcset=True):
img['src'] = img['srcset'].split()[6]
return soup