From ed2672e38c7203a02b505d2ed0bfe89323b65106 Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Mon, 20 Apr 2020 22:19:39 +0200 Subject: [PATCH 1/6] [prosiebensat1] remove 7tv as it does not longer exist to be precise: it is now called Joyn, but it works completely different now. Might even need it's own extractor. --- youtube_dl/extractor/prosiebensat1.py | 15 +-------------- 1 file changed, 1 insertion(+), 14 deletions(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index 1bc4f9b6b..46ddeb69d 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 ) @@ -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', From d16e75e406b163e749f8780cfb06e50af8fe2509 Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Mon, 20 Apr 2020 22:24:58 +0200 Subject: [PATCH 2/6] [prosiebensat1] fix unit test as titles are now prefixed with show name --- youtube_dl/extractor/prosiebensat1.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index 46ddeb69d..d39215b0c 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -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, From 08c51879ed94565bca3066350f1341b73ecb9b26 Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Mon, 20 Apr 2020 22:27:14 +0200 Subject: [PATCH 3/6] [prosiebensat1] fix unittest: fem.com url has changed --- youtube_dl/extractor/prosiebensat1.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index d39215b0c..a6bf121b0 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -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', From d1e0dd9aadaa317369fdeebb963bc4c73e316160 Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Mon, 20 Apr 2020 22:28:50 +0200 Subject: [PATCH 4/6] [prosiebensat1] fix regex for published_time meta --- youtube_dl/extractor/prosiebensat1.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index a6bf121b0..db3061673 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -408,7 +408,7 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): r']+id="veeseoDescription"[^>]*>(.+?)', ] _UPLOAD_DATE_REGEXES = [ - r'', + 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})', From d19a1d5ef9eae7c4f198565ca87bf86d39d7f381 Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Wed, 22 Apr 2020 20:14:34 +0200 Subject: [PATCH 5/6] [prosiebensat1] use _html_search_meta for published_time --- youtube_dl/extractor/prosiebensat1.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index db3061673..08c0f67be 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -408,7 +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})', @@ -438,8 +437,11 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): 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, From 548b87050bf62dec7dac180d0ed63fd8d9b0b19c Mon Sep 17 00:00:00 2001 From: Philipp Stehle Date: Wed, 22 Apr 2020 21:12:07 +0200 Subject: [PATCH 6/6] [prosiebensat1] fix flake8 issues --- youtube_dl/extractor/prosiebensat1.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/prosiebensat1.py b/youtube_dl/extractor/prosiebensat1.py index 08c0f67be..74074606e 100644 --- a/youtube_dl/extractor/prosiebensat1.py +++ b/youtube_dl/extractor/prosiebensat1.py @@ -439,9 +439,9 @@ class ProSiebenSat1IE(ProSiebenSat1BaseIE): thumbnail = self._og_search_thumbnail(webpage) 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)) + 'upload date', default=None) + or self._html_search_regex(self._UPLOAD_DATE_REGEXES, + webpage, 'upload date', default=None)) info.update({ 'id': clip_id,