Correctly remove "partner" query param and some more fail proof in "La Jornada (Mexico)" recipe.

The recipe stopped working since July 11th, 2015 because "partner" query param changed position.
This commit is contained in:
Rogelio Domínguez Hernández 2015-07-17 09:22:05 -05:00
parent b876c31ef5
commit 68662d65d6

View File

@ -5,6 +5,8 @@ www.jornada.unam.mx
'''
import re
from urllib import urlencode
from urlparse import urlparse, urlunparse, parse_qs
from calibre import strftime
from calibre.web.feeds.news import BasicNewsRecipe
@ -86,6 +88,14 @@ class LaJornada_mx(BasicNewsRecipe):
return soup
def get_article_url(self, article):
# Get link to original article URL
rurl = article.get('guid', None)
return rurl.rpartition('&partner=')[0]
if not rurl:
# Use the "link" attribute as failover
return article.get('link', None)
# Remove "partner" query param
u = urlparse(rurl)
query = parse_qs(u.query)
query.pop('partner', None)
u = u._replace(query=urlencode(query, True))
return urlunparse(u)