Updated Le Devoir

This commit is contained in:
Kovid Goyal 2012-02-13 09:56:06 +05:30
parent 4451b82992
commit 2eab5b4486

View File

@ -1,8 +1,8 @@
__license__ = 'GPL v3'
__author__ = 'Lorenzo Vigentini'
__copyright__ = '2009, Lorenzo Vigentini <l.vigentini at gmail.com>'
__author__ = 'Lorenzo Vigentini and Olivier Daigle'
__copyright__ = '2012, Lorenzo Vigentini <l.vigentini at gmail.com>, Olivier Daigle <odaigle _at nuvucameras __dot__ com>'
__version__ = 'v1.01'
__date__ = '14, January 2010'
__date__ = '12, February 2012'
__description__ = 'Canadian Paper '
'''
@ -12,6 +12,7 @@ http://www.ledevoir.com/
import re
from calibre.web.feeds.news import BasicNewsRecipe
from datetime import date
class ledevoir(BasicNewsRecipe):
author = 'Lorenzo Vigentini'
@ -26,11 +27,15 @@ class ledevoir(BasicNewsRecipe):
encoding = 'utf-8'
timefmt = '[%a, %d %b, %Y]'
max_articles_per_feed = 50
oldest_article = 1
max_articles_per_feed = 200
use_embedded_content = False
recursion = 10
needs_subscription = 'optional'
filterDuplicates = False
url_list = []
remove_javascript = True
no_stylesheets = True
@ -38,7 +43,7 @@ class ledevoir(BasicNewsRecipe):
keep_only_tags = [
dict(name='div', attrs={'id':'article'}),
dict(name='ul', attrs={'id':'ariane'})
dict(name='div', attrs={'id':'colonne_principale'})
]
remove_tags = [
@ -51,7 +56,7 @@ class ledevoir(BasicNewsRecipe):
feeds = [
(u'A la une', 'http://www.ledevoir.com/rss/manchettes.xml'),
(u'Edition complete', 'http://feeds2.feedburner.com/fluxdudevoir'),
(u'Édition complete', 'http://feeds2.feedburner.com/fluxdudevoir'),
(u'Opinions', 'http://www.ledevoir.com/rss/opinions.xml'),
(u'Chroniques', 'http://www.ledevoir.com/rss/chroniques.xml'),
(u'Politique', 'http://www.ledevoir.com/rss/section/politique.xml?id=51'),
@ -61,7 +66,7 @@ class ledevoir(BasicNewsRecipe):
(u'Societe', 'http://www.ledevoir.com/rss/section/societe.xml?id=52'),
(u'Economie', 'http://www.ledevoir.com/rss/section/economie.xml?id=49'),
(u'Sports', 'http://www.ledevoir.com/rss/section/sports.xml?id=85'),
(u'Loisirs', 'http://www.ledevoir.com/rss/section/loisirs.xml?id=50')
(u'Art de vivre', 'http://www.ledevoir.com/rss/section/art-de-vivre.xml?id=50')
]
extra_css = '''
@ -85,8 +90,16 @@ class ledevoir(BasicNewsRecipe):
br = BasicNewsRecipe.get_browser()
if self.username is not None and self.password is not None:
br.open('http://www.ledevoir.com')
br.select_form(nr=1)
br['login[courriel]'] = self.username
br['login[password]'] = self.password
br.select_form(nr=0)
br['login_popup[courriel]'] = self.username
br['login_popup[password]'] = self.password
br.submit()
return br
def print_version(self, url):
if self.filterDuplicates:
if url in self.url_list:
return
self.url_list.append(url)
return url