nojhan/weboob-devel
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

vimeo: Fix for no-embed videos

Browse source 
Some videos like http://vimeo.com/61275290 have restrictions
on embedding, which makes the usual JSON url fail.
So we first try to parse the JSON data directly from the
script element in the page.

Signed-off-by: François Revol <revol@free.fr>
This commit is contained in:
François Revol 2013-03-14 03:59:20 +01:00 • committed by Florent
commit 1d0faa1af0
1 changed files with 17 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

21
modules/vimeo/pages.py
View file

@ -27,6 +27,7 @@ from urllib2 import HTTPError
from weboob.tools.browser import BasePage from weboob.tools.browser import BasePage
from weboob.tools.json import json from weboob.tools.json import json
import re
import datetime import datetime
from dateutil.parser import parse as parse_dt from dateutil.parser import parse as parse_dt
@ -65,12 +66,24 @@ class VideoPage(BasePage):
if len(obj) > 0: if len(obj) > 0:
v.thumbnail = Thumbnail(unicode(obj[0].attrib['content'])) v.thumbnail = Thumbnail(unicode(obj[0].attrib['content']))
# for the rest, use the JSON config descriptor data = None
json_data = self.browser.openurl('http://%s/config/%s?type=%s&referrer=%s' % ("player.vimeo.com", int(v.id), "html5_desktop_local", ""))
data = json.load(json_data) # First try to find the JSON data in the page itself.
# it's the only location in case the video is not allowed to be embeded
for script in self.parser.select(self.document.getroot(), 'script'):
m = re.match('.* = {config:({.*}),assets:.*', unicode(script.text), re.DOTALL)
if m:
data = json.loads(m.group(1))
break
# Else fall back to the API
if data is None:
# for the rest, use the JSON config descriptor
json_data = self.browser.openurl('http://%s/config/%s?type=%s&referrer=%s' % ("player.vimeo.com", int(v.id), "html5_desktop_local", ""))
data = json.load(json_data)
if data is None: if data is None:
raise BrokenPageError('Unable to get JSON config for id: %r' % int(v.id)) raise BrokenPageError('Unable to get JSON config for id: %r' % int(v.id))
#print data
if v.title is None: if v.title is None:
v.title = unicode(data['video']['title']) v.title = unicode(data['video']['title'])