[ie/youtube] Avoid poToken experiment player responses (#10456)

Closes #10397
Authored by: seproDev
This commit is contained in:
sepro 2024-07-14 01:19:17 +02:00 committed by GitHub
parent 644d84d778
commit 8b8b442cb0
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -1294,6 +1294,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'401': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'vcodec': 'av01.0.12M.08'}, '401': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'vcodec': 'av01.0.12M.08'},
} }
_SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt') _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt')
_POTOKEN_EXPERIMENTS = ('51217476', '51217102')
_GEO_BYPASS = False _GEO_BYPASS = False
@ -3703,6 +3704,13 @@ def _invalid_player_response(self, pr, video_id):
def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, smuggled_data): def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, smuggled_data):
initial_pr = None initial_pr = None
if webpage: if webpage:
experiments = traverse_obj(master_ytcfg, (
'WEB_PLAYER_CONTEXT_CONFIGS', ..., 'serializedExperimentIds', {str}, {lambda x: x.split(',')}, ..., {str}))
if all(x in experiments for x in self._POTOKEN_EXPERIMENTS):
self.report_warning(
'Webpage contains broken formats (poToken experiment detected). Ignoring initial player response')
master_ytcfg = self._get_default_ytcfg()
else:
initial_pr = self._search_json( initial_pr = self._search_json(
self._YT_INITIAL_PLAYER_RESPONSE_RE, webpage, 'initial player response', video_id, fatal=False) self._YT_INITIAL_PLAYER_RESPONSE_RE, webpage, 'initial player response', video_id, fatal=False)
@ -3746,11 +3754,22 @@ def append_client(*client_names):
player_url = self._download_player_url(video_id) player_url = self._download_player_url(video_id)
tried_iframe_fallback = True tried_iframe_fallback = True
pr = initial_pr if client == 'web' and initial_pr else None
for retry in self.RetryManager(fatal=False):
try: try:
pr = initial_pr if client == 'web' and initial_pr else self._extract_player_response( pr = pr or self._extract_player_response(
client, video_id, player_ytcfg or master_ytcfg, player_ytcfg, player_url if require_js_player else None, initial_pr, smuggled_data) client, video_id, player_ytcfg or master_ytcfg, player_ytcfg,
player_url if require_js_player else None, initial_pr, smuggled_data)
except ExtractorError as e: except ExtractorError as e:
self.report_warning(e) self.report_warning(e)
break
experiments = traverse_obj(pr, (
'responseContext', 'serviceTrackingParams', lambda _, v: v['service'] == 'GFEEDBACK',
'params', lambda _, v: v['key'] == 'e', 'value', {lambda x: x.split(',')}, ..., {str}))
if all(x in experiments for x in self._POTOKEN_EXPERIMENTS):
pr = None
retry.error = ExtractorError('API returned broken formats (poToken experiment detected)', expected=True)
if not pr:
continue continue
if pr_id := self._invalid_player_response(pr, video_id): if pr_id := self._invalid_player_response(pr, video_id):