You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

36 lines
1.2 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
from .common import InfoExtractor
from ..utils import remove_end
class TelegraafIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?telegraaf\.nl/tv/(?:[^/]+/)+(?P<id>\d+)/[^/]+\.html'
_TEST = {
'url': 'http://www.telegraaf.nl/tv/nieuws/binnenland/24353229/__Tikibad_ontruimd_wegens_brand__.html',
'md5': '83245a9779bcc4a24454bfd53c65b6dc',
'info_dict': {
'id': '24353229',
'ext': 'mp4',
'title': 'Tikibad ontruimd wegens brand',
'description': 'md5:05ca046ff47b931f9b04855015e163a4',
'thumbnail': 're:^https?://.*\.jpg$',
'duration': 33,
},
}
def _real_extract(self, url):
playlist_id = self._match_id(url)
webpage = self._download_webpage(url, playlist_id)
playlist_url = self._search_regex(
r"iframe\.loadPlayer\('([^']+)'", webpage, 'player')
entries = self._extract_xspf_playlist(playlist_url, playlist_id)
title = remove_end(self._og_search_title(webpage), ' - VIDEO')
description = self._og_search_description(webpage)
return self.playlist_result(entries, playlist_id, title, description)