From 2cdfe977d7e006ae93c427868ec1c5701d49cf1a Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Thu, 23 Apr 2020 21:44:13 +0200 Subject: [PATCH] [prosiebensat1] Improve extraction and remove 7tv.de support (#24948) --- youtube_dl/extractor/prosiebensat1.py | 29 +++++++++------------------ 1 file changed, 9 insertions(+), 20 deletions(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index 1bc4f9b6b..74074606e 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -175,7 +175,7 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): (?: (?:beta\.)? (?: - prosieben(?:maxx)?|sixx|sat1(?:gold)?|kabeleins(?:doku)?|the-voice-of-germany|7tv|advopedia + prosieben(?:maxx)?|sixx|sat1(?:gold)?|kabeleins(?:doku)?|the-voice-of-germany|advopedia )\.(?:de|at|ch)| ran\.de|fem\.com|advopedia\.de|galileo\.tv/video ) @@ -193,7 +193,7 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): 'info_dict': { 'id': '2104602', 'ext': 'mp4', - 'title': 'Episode 18 - Staffel 2', + 'title': 'CIRCUS HALLIGALLI - Episode 18 - Staffel 2', 'description': 'md5:8733c81b702ea472e069bc48bb658fc1', 'upload_date': '20131231', 'duration': 5845.04, @@ -300,7 +300,7 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): 'info_dict': { 'id': '2572814', 'ext': 'mp4', - 'title': 'Andreas Kümmert: Rocket Man', + 'title': 'The Voice of Germany - Andreas Kümmert: Rocket Man', 'description': 'md5:6ddb02b0781c6adf778afea606652e38', 'upload_date': '20131017', 'duration': 469.88, @@ -310,7 +310,7 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): }, }, { - 'url': 'http://www.fem.com/wellness/videos/wellness-video-clip-kurztripps-zum-valentinstag.html', + 'url': 'http://www.fem.com/videos/beauty-lifestyle/kurztrips-zum-valentinstag', 'info_dict': { 'id': '2156342', 'ext': 'mp4', @@ -332,19 +332,6 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): 'playlist_count': 2, 'skip': 'This video is unavailable', }, - { - 'url': 'http://www.7tv.de/circus-halligalli/615-best-of-circus-halligalli-ganze-folge', - 'info_dict': { - 'id': '4187506', - 'ext': 'mp4', - 'title': 'Best of Circus HalliGalli', - 'description': 'md5:8849752efd90b9772c9db6fdf87fb9e9', - 'upload_date': '20151229', - }, - 'params': { - 'skip_download': True, - }, - }, { # title in

'url': 'http://www.prosieben.de/stars/oscar-award/videos/jetzt-erst-enthuellt-das-geheimnis-von-emma-stones-oscar-robe-clip', @@ -421,7 +408,6 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): r']+id="veeseoDescription"[^>]*>(.+?)', ] _UPLOAD_DATE_REGEXES = [ - r'', r'\s*(\d{2}\.\d{2}\.\d{4} \d{2}:\d{2}) \|\s*\s*(\d{2}\.\d{2}\.\d{4}) \d{2}:\d{2} Uhr', r'(\d{2}\.\d{2}\.\d{4})', @@ -451,8 +437,11 @@ def _extract_clip(self, url, webpage): if description is None: description = self._og_search_description(webpage) thumbnail = self._og_search_thumbnail(webpage) - upload_date = unified_strdate(self._html_search_regex( - self._UPLOAD_DATE_REGEXES, webpage, 'upload date', default=None)) + upload_date = unified_strdate( + self._html_search_meta('og:published_time', webpage, + 'upload date', default=None) + or self._html_search_regex(self._UPLOAD_DATE_REGEXES, + webpage, 'upload date', default=None)) info.update({ 'id': clip_id,