Fix #1705822 [Impossible de récupérer des actualités Mediapart](https://bugs.launchpad.net/calibre/+bug/1705822)

This commit is contained in:
Kovid Goyal 2017-07-22 23:13:40 +05:30
parent 91f37ca343
commit 0f78497dc5
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C

View File

@ -9,7 +9,6 @@ Mediapart
'''
import re
from calibre.ebooks.BeautifulSoup import BeautifulSoup
from calibre.web.feeds.news import BasicNewsRecipe
from calibre.web.feeds import feeds_from_index
from datetime import date, timedelta
@ -54,7 +53,7 @@ class Mediapart(BasicNewsRecipe):
soup = self.index_to_soup(
'https://www.mediapart.fr/journal/fil-dactualites')
page = soup.find('section', {'class': 'global-wrapper'})
page = soup.find('main', {'class': 'global-wrapper'})
fils = page.find('ul', {'class': 'post-list universe-journal'})
for article in fils.findAll('li'):
@ -129,8 +128,7 @@ class Mediapart(BasicNewsRecipe):
'août', 'septembre', 'octobre', 'novembre', 'décembre'].index(date_arr[1]))
def print_version(self, url):
raw = self.browser.open(url).read()
soup = BeautifulSoup(raw.decode('utf8', 'replace'))
soup = self.index_to_soup(url)
# Filter old articles
# article_date = self.parse_french_date(self.tag_to_string(soup.find('span', 'article-date')))
@ -139,7 +137,6 @@ class Mediapart(BasicNewsRecipe):
tools = soup.find('li', {'class': 'print'})
link = tools.find('a', {'href': re.compile(r'\/print\/.*')})
print(link['href'])
# if link is None:
# print 'Error: print link not found'
# return None