From: remitamine Date: Mon, 23 May 2016 08:30:26 +0000 (+0100) Subject: [cbs] allow to pass content id to the extractor(closes #9589) X-Git-Url: http://git.bitcoin.ninja/?a=commitdiff_plain;h=42a7439717610530b0f7c630ef0eecf1b0638475;p=youtube-dl [cbs] allow to pass content id to the extractor(closes #9589) --- diff --git a/youtube_dl/extractor/cbs.py b/youtube_dl/extractor/cbs.py index 051d783a2..ac2c7dced 100644 --- a/youtube_dl/extractor/cbs.py +++ b/youtube_dl/extractor/cbs.py @@ -1,5 +1,7 @@ from __future__ import unicode_literals +import re + from .theplatform import ThePlatformIE from ..utils import ( xpath_text, @@ -21,7 +23,7 @@ class CBSBaseIE(ThePlatformIE): class CBSIE(CBSBaseIE): - _VALID_URL = r'https?://(?:www\.)?(?:cbs\.com/shows/[^/]+/(?:video|artist)|colbertlateshow\.com/(?:video|podcasts))/[^/]+/(?P[^/]+)' + _VALID_URL = r'(?:cbs:(?P\w+)|https?://(?:www\.)?(?:cbs\.com/shows/[^/]+/(?:video|artist)|colbertlateshow\.com/(?:video|podcasts))/[^/]+/(?P[^/]+))' _TESTS = [{ 'url': 'http://www.cbs.com/shows/garth-brooks/video/_u7W953k6la293J7EPTd9oHkSPs6Xn6_/connect-chat-feat-garth-brooks/', @@ -66,11 +68,12 @@ class CBSIE(CBSBaseIE): TP_RELEASE_URL_TEMPLATE = 'http://link.theplatform.com/s/dJ5BDC/%s?mbr=true' def _real_extract(self, url): - display_id = self._match_id(url) - webpage = self._download_webpage(url, display_id) - content_id = self._search_regex( - [r"video\.settings\.content_id\s*=\s*'([^']+)';", r"cbsplayer\.contentId\s*=\s*'([^']+)';"], - webpage, 'content id') + content_id, display_id = re.match(self._VALID_URL, url).groups() + if not content_id: + webpage = self._download_webpage(url, display_id) + content_id = self._search_regex( + [r"video\.settings\.content_id\s*=\s*'([^']+)';", r"cbsplayer\.contentId\s*=\s*'([^']+)';"], + webpage, 'content id') items_data = self._download_xml( 'http://can.cbs.com/thunder/player/videoPlayerService.php', content_id, query={'partner': 'cbs', 'contentId': content_id})