2 from __future__ import unicode_literals
4 from .common import InfoExtractor
5 from ..compat import compat_str
6 from ..utils import month_by_name
9 class FranceInterIE(InfoExtractor):
10 _VALID_URL = r'https?://(?:www\.)?franceinter\.fr/emissions/(?P<id>[^?#]+)'
13 'url': 'https://www.franceinter.fr/emissions/la-tete-au-carre/la-tete-au-carre-14-septembre-2016',
14 'md5': '4e3aeb58fe0e83d7b0581fa213c409d0',
16 'id': 'la-tete-au-carre/la-tete-au-carre-14-septembre-2016',
18 'title': 'Et si les rêves pouvaient nous aider à agir dans notre vie quotidienne ?',
19 'description': 'md5:a245dd62cf5bf51de915f8d9956d180a',
20 'upload_date': '20160914',
24 def _real_extract(self, url):
25 video_id = self._match_id(url)
27 webpage = self._download_webpage(url, video_id)
29 video_url = self._search_regex(
30 r'(?s)<div[^>]+class=["\']page-diffusion["\'][^>]*>.*?<button[^>]+data-url=(["\'])(?P<url>(?:(?!\1).)+)\1',
31 webpage, 'video url', group='url')
33 title = self._og_search_title(webpage)
34 description = self._og_search_description(webpage)
36 upload_date_str = self._search_regex(
37 r'class=["\']cover-emission-period["\'][^>]*>[^<]+\s+(\d{1,2}\s+[^\s]+\s+\d{4})<',
38 webpage, 'upload date', fatal=False)
40 upload_date_list = upload_date_str.split()
41 upload_date_list.reverse()
42 upload_date_list[1] = '%02d' % (month_by_name(upload_date_list[1], lang='fr') or 0)
43 upload_date = ''.join(upload_date_list)
50 'description': description,
51 'upload_date': upload_date,