Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update Le Devoir

Browse Source
This commit is contained in:
Kovid Goyal 2013-03-07 22:18:05 +05:30
parent 3a58a0fab7
commit 55030cc0b0
1 changed files with 40 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

65
recipes/ledevoir.recipe
View File

@ -2,7 +2,7 @@ __license__ = 'GPL v3'
__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__ = '12, February 2012'
__date__ = '22, December 2012'
__description__ = 'Canadian Paper '
'''
@ -32,41 +32,50 @@ class ledevoir(BasicNewsRecipe):
recursion = 10
needs_subscription = 'optional'
filterDuplicates = False
url_list = []
remove_javascript = True
no_stylesheets = True
auto_cleanup = True
preprocess_regexps = [(re.compile(r'(title|alt)=".*?>.*?"', re.DOTALL), lambda m: '')]
#keep_only_tags = [
keep_only_tags = [
#dict(name='div', attrs={'id':'article_detail'}),
#dict(name='div', attrs={'id':'colonne_principale'})
#]
#dict(name='div', attrs={'id':'colonne_principale'}),
dict(name='article', attrs={'id':'article', 'class':'clearfix'}),
dict(name='article', attrs={'id':'article', 'class':'clearfix portrait'})
]
#remove_tags = [
#dict(name='div', attrs={'id':'dialog'}),
#dict(name='div', attrs={'class':['interesse_actions','reactions','taille_du_texte right clearfix','partage_sociaux clearfix']}),
#dict(name='aside', attrs={'class':['article_actions clearfix','reactions','partage_sociaux_wrapper']}),
#dict(name='ul', attrs={'class':'mots_cles'}),
#dict(name='ul', attrs={'id':'commentaires'}),
#dict(name='a', attrs={'class':'haut'}),
#dict(name='h5', attrs={'class':'interesse_actions'})
#]
remove_tags = [
dict(name='div', attrs={'id':'prive'}),
dict(name='div', attrs={'class':'acheter_article'}),
dict(name='div', attrs={'id':'col_complement'}),
dict(name='div', attrs={'id':'raccourcis','class':'clearfix'}),
dict(name='div', attrs={'id':'dialog'}),
dict(name='div', attrs={'id':'liste_photos_article','class':'clearfix'}),
dict(name='script', attrs={'type':'text/javascript'}),
dict(name='div', attrs={'class':['interesse_actions','reactions','taille_du_texte right clearfix','partage_sociaux clearfix']}),
dict(name='aside', attrs={'class':['article_actions clearfix','partage_sociaux_wrapper']}),
dict(name='aside', attrs={'class':'reactions', 'id':'reactions'}),
dict(name='ul', attrs={'class':'mots_cles'}),
dict(name='ul', attrs={'id':'commentaires'}),
dict(name='a', attrs={'class':'haut'}),
dict(name='h5', attrs={'class':'interesse_actions'})
]
feeds = [
(u'A la une', 'http://www.ledevoir.com/rss/manchettes.xml'),
(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'),
(u'International', 'http://www.ledevoir.com/rss/section/international.xml?id=76'),
(u'Culture', 'http://www.ledevoir.com/rss/section/culture.xml?id=48'),
(u'Environnement', 'http://www.ledevoir.com/rss/section/environnement.xml?id=78'),
(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'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'),
# (u'International', 'http://www.ledevoir.com/rss/section/international.xml?id=76'),
# (u'Culture', 'http://www.ledevoir.com/rss/section/culture.xml?id=48'),
# (u'Environnement', 'http://www.ledevoir.com/rss/section/environnement.xml?id=78'),
# (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'Art de vivre', 'http://www.ledevoir.com/rss/section/art-de-vivre.xml?id=50')
]
@ -88,7 +97,7 @@ class ledevoir(BasicNewsRecipe):
.texte {font-size:1.15em;line-height:1.4em;margin-bottom:17px;}
'''
def get_browser(self):
br = BasicNewsRecipe.get_browser(self)
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=0)
@ -97,4 +106,10 @@ class ledevoir(BasicNewsRecipe):
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