Mirrors/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-09-06 02:49:35 +02:00
Code Issues Releases Activity

[youtube] Improve age-gate detection (#577)

Browse Source 
Authored by: colethedj
This commit is contained in:
colethedj 2021-07-29 17:23:56 +12:00 committed by pukkandan
parent 6552469433
commit 9275f62cf8
No known key found for this signature in database
GPG Key ID: 0F00D95A001F4698
1 changed files with 17 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
yt_dlp/extractor/youtube.py
View File

@ -1121,7 +1121,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
_AGE_GATE_REASONS = ( _AGE_GATE_REASONS = (
'Sign in to confirm your age', 'Sign in to confirm your age',
'This video may be inappropriate for some users.', 'This video may be inappropriate for some users.',
'Sorry, this content is age-restricted.') 'Sorry, this content is age-restricted.',
'Please confirm your age.')
_AGE_GATE_STATUS_REASONS = (
'AGE_VERIFICATION_REQUIRED',
'AGE_CHECK_REQUIRED'
)
_GEO_BYPASS = False _GEO_BYPASS = False
@ -2430,6 +2436,15 @@ def _generate_player_context(sts=None):
'racyCheckOk': True 'racyCheckOk': True
} }
def _is_agegated(self, player_response):
reasons = traverse_obj(player_response, ('playabilityStatus', ('status', 'reason')), default=[])
for reason in reasons:
if reason in self._AGE_GATE_REASONS + self._AGE_GATE_STATUS_REASONS:
return True
if traverse_obj(player_response, ('playabilityStatus', 'desktopLegacyAgeGateReason')) is not None:
return True
return False
def _extract_player_response(self, client, video_id, master_ytcfg, player_ytcfg, identity_token, player_url, initial_pr): def _extract_player_response(self, client, video_id, master_ytcfg, player_ytcfg, identity_token, player_url, initial_pr):
session_index = self._extract_session_index(player_ytcfg, master_ytcfg) session_index = self._extract_session_index(player_ytcfg, master_ytcfg)
@ -2491,15 +2506,6 @@ def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, pl
player_ytcfg = master_ytcfg if client == 'web' else {} player_ytcfg = master_ytcfg if client == 'web' else {}
if 'configs' not in self._configuration_arg('player_skip'): if 'configs' not in self._configuration_arg('player_skip'):
player_ytcfg = self._extract_player_ytcfg(client, video_id) or player_ytcfg player_ytcfg = self._extract_player_ytcfg(client, video_id) or player_ytcfg
if client == 'web_embedded':
# If we extracted the embed webpage, it'll tell us if we can view the video
embedded_pr = self._parse_json(
traverse_obj(player_ytcfg, ('PLAYER_VARS', 'embedded_player_response'), expected_type=str) or '{}',
video_id=video_id)
embedded_ps_reason = traverse_obj(embedded_pr, ('playabilityStatus', 'reason'), expected_type=str) or ''
if embedded_ps_reason in self._AGE_GATE_REASONS:
self.report_warning(f'Youtube said: {embedded_ps_reason}')
continue
pr = ( pr = (
initial_pr if client == 'web' and initial_pr initial_pr if client == 'web' and initial_pr
@ -2508,7 +2514,7 @@ def _extract_player_responses(self, clients, video_id, webpage, master_ytcfg, pl
if pr: if pr:
yield pr yield pr
if traverse_obj(pr, ('playabilityStatus', 'reason')) in self._AGE_GATE_REASONS: if self._is_agegated(pr):
client = f'{client}_agegate' client = f'{client}_agegate'
if client in self._YT_CLIENTS and client not in original_clients: if client in self._YT_CLIENTS and client not in original_clients:
clients.append(client) clients.append(client)