mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-09-29 15:31:08 -04:00
87 lines
4.1 KiB
Python
87 lines
4.1 KiB
Python
#!/usr/bin/env python
|
|
# vim:fileencoding=utf-8
|
|
|
|
__license__ = 'GPL v3'
|
|
__copyright__ = '2009, Mathieu Godlewski <mathieu at godlewski.fr>\
|
|
2015, Rémi Vanicat <vanicat at debian.org'
|
|
'''
|
|
Courrier International
|
|
'''
|
|
|
|
import re
|
|
from calibre.web.feeds.news import BasicNewsRecipe
|
|
|
|
|
|
class CourrierInternational(BasicNewsRecipe):
|
|
title = 'Courrier International'
|
|
__author__ = 'Mathieu Godlewski <mathieu at godlewski.fr>'
|
|
description = 'Global news in french from international newspapers'
|
|
oldest_article = 7
|
|
language = 'fr'
|
|
|
|
max_articles_per_feed = 50
|
|
no_stylesheets = True
|
|
|
|
html2lrf_options = ['--base-font-size', '10']
|
|
|
|
keep_only_tags = [
|
|
dict(name='div', attrs={'class': 'article-content'}),
|
|
]
|
|
|
|
remove_tags = [
|
|
dict(name='aside', attrs={'class': re.compile(
|
|
'article-sitesocial|article-ad|article-tertiary|view-ci-service-article')}),
|
|
dict(name='button', attrs={'id': re.compile(
|
|
'action-zen-off-(top|bottom)')}),
|
|
]
|
|
|
|
needs_subscription = "optional"
|
|
login_url = 'http://www.courrierinternational.com/login'
|
|
|
|
def get_browser(self):
|
|
def is_form_login(form):
|
|
return "id" in form.attrs and form.attrs['id'] == "user-login-form"
|
|
br = BasicNewsRecipe.get_browser(self)
|
|
if self.username:
|
|
br.open(self.login_url)
|
|
br.select_form(predicate=is_form_login)
|
|
br['name'] = self.username
|
|
br['pass'] = self.password
|
|
br.submit()
|
|
return br
|
|
|
|
def preprocess_html(self, soup):
|
|
for link in soup.findAll("a", href=re.compile('^/')):
|
|
link["href"] = 'http://www.courrierinternational.com' + link["href"]
|
|
return soup
|
|
|
|
feeds = [
|
|
# Some articles requiring subscription fails on download.
|
|
('Tous', 'http://www.courrierinternational.com/feed/all/rss.xml'),
|
|
('France', 'https://www.courrierinternational.com/feed/category/6260/rss.xml'),
|
|
('Europe', 'https://www.courrierinternational.com/feed/category/6261/rss.xml'),
|
|
('Ameriques', 'https://www.courrierinternational.com/feed/category/6262/rss.xml'),
|
|
('Asie', 'https://www.courrierinternational.com/feed/category/6263/rss.xml'),
|
|
('Afrique', 'https://www.courrierinternational.com/feed/category/6264/rss.xml'),
|
|
('Moyen-Orient', 'https://www.courrierinternational.com/feed/category/6265/rss.xml'),
|
|
('Expat', 'https://www.courrierinternational.com/feed/category/6994/rss.xml'),
|
|
('Business', 'https://www.courrierinternational.com/feed/category/6991/rss.xml'),
|
|
('Culture', 'https://www.courrierinternational.com/feed/category/6270/rss.xml'),
|
|
('Economie', 'https://www.courrierinternational.com/feed/category/6266/rss.xml'),
|
|
('Écrans', 'https://www.courrierinternational.com/feed/category/7342/rss.xml'),
|
|
('Éducation', 'https://www.courrierinternational.com/feed/category/6997/rss.xml'),
|
|
('Environnement', 'https://www.courrierinternational.com/feed/category/6267/rss.xml'),
|
|
('Infographie', 'https://www.courrierinternational.com/feed/category/6281/rss.xml'),
|
|
('Lifestyle', 'https://www.courrierinternational.com/feed/category/7426/rss.xml'),
|
|
('Planète buzz', 'https://www.courrierinternational.com/feed/category/6680/rss.xml'),
|
|
('Politique', 'https://www.courrierinternational.com/feed/category/6681/rss.xml'),
|
|
('Santé', 'https://www.courrierinternational.com/feed/category/6682/rss.xml'),
|
|
('Sport', 'https://www.courrierinternational.com/feed/category/6273/rss.xml'),
|
|
('Tendances', 'https://www.courrierinternational.com/feed/category/6282/rss.xml'),
|
|
('Réveil', 'https://www.courrierinternational.com/feed/category/8224/rss.xml'),
|
|
('Enquêtes', 'https://www.courrierinternational.com/feed/category/6678/rss.xml'),
|
|
('Histoire', 'https://www.courrierinternational.com/feed/category/6679/rss.xml'),
|
|
('Sciences & Techno', 'https://www.courrierinternational.com/feed/category/6268/rss.xml'),
|
|
('Voyage', 'https://www.courrierinternational.com/feed/category/6271/rss.xml'),
|
|
]
|