]> jfr.im git - yt-dlp.git/blobdiff - yt_dlp/extractor/youtube.py
[extractor] Fix format sorting of `channels`
[yt-dlp.git] / yt_dlp / extractor / youtube.py
index f20b7321ad7c7cec7350fe39443a0ad058a6365d..b59c8630aea1df11bb4d7a5fdbcc748e99498ab0 100644 (file)
@@ -28,7 +28,6 @@
     clean_html,
     datetime_from_str,
     dict_get,
-    error_to_compat_str,
     float_or_none,
     format_field,
     get_first,
@@ -45,7 +44,6 @@
     parse_iso8601,
     parse_qs,
     qualities,
-    remove_end,
     remove_start,
     smuggle_url,
     str_or_none,
@@ -70,7 +68,7 @@
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'WEB',
-                'clientVersion': '2.20211221.00.00',
+                'clientVersion': '2.20220801.00.00',
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 1
@@ -80,7 +78,7 @@
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'WEB_EMBEDDED_PLAYER',
-                'clientVersion': '1.20211215.00.01',
+                'clientVersion': '1.20220731.00.00',
             },
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 56
@@ -91,7 +89,7 @@
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'WEB_REMIX',
-                'clientVersion': '1.20211213.00.00',
+                'clientVersion': '1.20220727.01.00',
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 67,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'WEB_CREATOR',
-                'clientVersion': '1.20211220.02.00',
+                'clientVersion': '1.20220726.00.00',
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 62,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'ANDROID',
-                'clientVersion': '16.49',
+                'clientVersion': '17.29.34',
+                'androidSdkVersion': 30
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 3,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'ANDROID_EMBEDDED_PLAYER',
-                'clientVersion': '16.49',
+                'clientVersion': '17.29.34',
+                'androidSdkVersion': 30
             },
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 55,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'ANDROID_MUSIC',
-                'clientVersion': '4.57',
+                'clientVersion': '5.16.51',
+                'androidSdkVersion': 30
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 21,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'ANDROID_CREATOR',
-                'clientVersion': '21.47',
+                'clientVersion': '22.28.100',
+                'androidSdkVersion': 30
             },
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 14,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'IOS',
-                'clientVersion': '16.46',
+                'clientVersion': '17.30.1',
                 'deviceModel': 'iPhone14,3',
             }
         },
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'IOS_MESSAGES_EXTENSION',
-                'clientVersion': '16.46',
+                'clientVersion': '17.30.1',
                 'deviceModel': 'iPhone14,3',
             },
         },
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'IOS_MUSIC',
-                'clientVersion': '4.57',
+                'clientVersion': '5.18',
             },
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 26,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'IOS_CREATOR',
-                'clientVersion': '21.47',
+                'clientVersion': '22.29.101',
             },
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 15,
         'INNERTUBE_CONTEXT': {
             'client': {
                 'clientName': 'MWEB',
-                'clientVersion': '2.20211221.01.00',
+                'clientVersion': '2.20220801.00.00',
             }
         },
         'INNERTUBE_CONTEXT_CLIENT_NAME': 2
@@ -763,74 +765,54 @@ def _extract_time_text(self, renderer, *path_list):
     def _extract_response(self, item_id, query, note='Downloading API JSON', headers=None,
                           ytcfg=None, check_get_keys=None, ep='browse', fatal=True, api_hostname=None,
                           default_client='web'):
-        response = None
-        last_error = None
-        count = -1
-        retries = self.get_param('extractor_retries', 3)
-        if check_get_keys is None:
-            check_get_keys = []
-        while count < retries:
-            count += 1
-            if last_error:
-                self.report_warning('%s. Retrying ...' % remove_end(last_error, '.'))
+        for retry in self.RetryManager():
             try:
                 response = self._call_api(
                     ep=ep, fatal=True, headers=headers,
-                    video_id=item_id, query=query,
+                    video_id=item_id, query=query, note=note,
                     context=self._extract_context(ytcfg, default_client),
                     api_key=self._extract_api_key(ytcfg, default_client),
-                    api_hostname=api_hostname, default_client=default_client,
-                    note='%s%s' % (note, ' (retry #%d)' % count if count else ''))
+                    api_hostname=api_hostname, default_client=default_client)
             except ExtractorError as e:
-                if isinstance(e.cause, network_exceptions):
-                    if isinstance(e.cause, urllib.error.HTTPError):
-                        first_bytes = e.cause.read(512)
-                        if not is_html(first_bytes):
-                            yt_error = try_get(
-                                self._parse_json(
-                                    self._webpage_read_content(e.cause, None, item_id, prefix=first_bytes) or '{}', item_id, fatal=False),
-                                lambda x: x['error']['message'], str)
-                            if yt_error:
-                                self._report_alerts([('ERROR', yt_error)], fatal=False)
-                    # Downloading page may result in intermittent 5xx HTTP error
-                    # Sometimes a 404 is also received. See: https://github.com/ytdl-org/youtube-dl/issues/28289
-                    # We also want to catch all other network exceptions since errors in later pages can be troublesome
-                    # See https://github.com/yt-dlp/yt-dlp/issues/507#issuecomment-880188210
-                    if not isinstance(e.cause, urllib.error.HTTPError) or e.cause.code not in (403, 429):
-                        last_error = error_to_compat_str(e.cause or e.msg)
-                        if count < retries:
-                            continue
-                if fatal:
-                    raise
-                else:
-                    self.report_warning(error_to_compat_str(e))
-                    return
+                if not isinstance(e.cause, network_exceptions):
+                    return self._error_or_warning(e, fatal=fatal)
+                elif not isinstance(e.cause, urllib.error.HTTPError):
+                    retry.error = e
+                    continue
 
-            else:
-                try:
-                    self._extract_and_report_alerts(response, only_once=True)
-                except ExtractorError as e:
-                    # YouTube servers may return errors we want to retry on in a 200 OK response
-                    # See: https://github.com/yt-dlp/yt-dlp/issues/839
-                    if 'unknown error' in e.msg.lower():
-                        last_error = e.msg
-                        continue
-                    if fatal:
-                        raise
-                    self.report_warning(error_to_compat_str(e))
-                    return
-                if not check_get_keys or dict_get(response, check_get_keys):
-                    break
-                # Youtube sometimes sends incomplete data
-                # See: https://github.com/ytdl-org/youtube-dl/issues/28194
-                last_error = 'Incomplete data received'
-                if count >= retries:
-                    if fatal:
-                        raise ExtractorError(last_error)
-                    else:
-                        self.report_warning(last_error)
-                        return
-        return response
+                first_bytes = e.cause.read(512)
+                if not is_html(first_bytes):
+                    yt_error = try_get(
+                        self._parse_json(
+                            self._webpage_read_content(e.cause, None, item_id, prefix=first_bytes) or '{}', item_id, fatal=False),
+                        lambda x: x['error']['message'], str)
+                    if yt_error:
+                        self._report_alerts([('ERROR', yt_error)], fatal=False)
+                # Downloading page may result in intermittent 5xx HTTP error
+                # Sometimes a 404 is also recieved. See: https://github.com/ytdl-org/youtube-dl/issues/28289
+                # We also want to catch all other network exceptions since errors in later pages can be troublesome
+                # See https://github.com/yt-dlp/yt-dlp/issues/507#issuecomment-880188210
+                if e.cause.code not in (403, 429):
+                    retry.error = e
+                    continue
+                return self._error_or_warning(e, fatal=fatal)
+
+            try:
+                self._extract_and_report_alerts(response, only_once=True)
+            except ExtractorError as e:
+                # YouTube servers may return errors we want to retry on in a 200 OK response
+                # See: https://github.com/yt-dlp/yt-dlp/issues/839
+                if 'unknown error' in e.msg.lower():
+                    retry.error = e
+                    continue
+                return self._error_or_warning(e, fatal=fatal)
+            # Youtube sometimes sends incomplete data
+            # See: https://github.com/ytdl-org/youtube-dl/issues/28194
+            if not traverse_obj(response, *variadic(check_get_keys)):
+                retry.error = ExtractorError('Incomplete data received')
+                continue
+
+            return response
 
     @staticmethod
     def is_music_url(url):
@@ -2276,6 +2258,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'tags': [],
                 'uploader_url': 'http://www.youtube.com/user/nao20010128nao',
             }
+        }, {
+            'note': '6 channel audio',
+            'url': 'https://www.youtube.com/watch?v=zgdo7-RRjgo',
+            'only_matching': True,
         }
     ]
 
@@ -3155,7 +3141,14 @@ def append_client(*client_names):
                 continue
 
             if pr:
-                prs.append(pr)
+                # YouTube may return a different video player response than expected.
+                # See: https://github.com/TeamNewPipe/NewPipe/issues/8713
+                pr_video_id = traverse_obj(pr, ('videoDetails', 'videoId'))
+                if pr_video_id and pr_video_id != video_id:
+                    self.report_warning(
+                        f'Skipping player response from {client} client (got player response for video "{pr_video_id}" instead of "{video_id}")' + bug_reports_message())
+                else:
+                    prs.append(pr)
 
             # creator clients can bypass AGE_VERIFICATION_REQUIRED if logged in
             if variant == 'embedded' and self._is_unplayable(pr) and self.is_authenticated:
@@ -3268,10 +3261,13 @@ def _extract_formats_and_subtitles(self, streaming_data, video_id, player_url, i
                     '%s%s' % (audio_track.get('displayName') or '',
                               ' (default)' if language_preference > 0 else ''),
                     fmt.get('qualityLabel') or quality.replace('audio_quality_', ''),
+                    try_get(fmt, lambda x: x['projectionType'].replace('RECTANGULAR', '').lower()),
+                    try_get(fmt, lambda x: x['spatialAudioType'].replace('SPATIAL_AUDIO_TYPE_', '').lower()),
                     throttled and 'THROTTLED', is_damaged and 'DAMAGED', delim=', '),
                 # Format 22 is likely to be damaged. See https://github.com/yt-dlp/yt-dlp/issues/3372
                 'source_preference': -10 if throttled else -5 if itag == '22' else -1,
                 'fps': int_or_none(fmt.get('fps')) or None,
+                'audio_channels': fmt.get('audioChannels'),
                 'height': height,
                 'quality': q(quality),
                 'has_drm': bool(fmt.get('drmFamilies')),
@@ -3592,7 +3588,8 @@ def feed_entry(name):
         formats.extend(self._extract_storyboard(player_responses, duration))
 
         # source_preference is lower for throttled/potentially damaged formats
-        self._sort_formats(formats, ('quality', 'res', 'fps', 'hdr:12', 'source', 'codec:vp9.2', 'lang', 'proto'))
+        self._sort_formats(formats, (
+            'quality', 'res', 'fps', 'hdr:12', 'source', 'vcodec:vp9.2', 'channels', 'acodec', 'lang', 'proto'))
 
         info = {
             'id': video_id,
@@ -4522,48 +4519,30 @@ def skip_webpage(self):
         return 'webpage' in self._configuration_arg('skip', ie_key=YoutubeTabIE.ie_key())
 
     def _extract_webpage(self, url, item_id, fatal=True):
-        retries = self.get_param('extractor_retries', 3)
-        count = -1
-        webpage = data = last_error = None
-        while count < retries:
-            count += 1
-            # Sometimes youtube returns a webpage with incomplete ytInitialData
-            # See: https://github.com/yt-dlp/yt-dlp/issues/116
-            if last_error:
-                self.report_warning('%s. Retrying ...' % last_error)
+        webpage, data = None, None
+        for retry in self.RetryManager(fatal=fatal):
             try:
-                webpage = self._download_webpage(
-                    url, item_id,
-                    note='Downloading webpage%s' % (' (retry #%d)' % count if count else '',))
+                webpage = self._download_webpage(url, item_id, note='Downloading webpage')
                 data = self.extract_yt_initial_data(item_id, webpage or '', fatal=fatal) or {}
             except ExtractorError as e:
                 if isinstance(e.cause, network_exceptions):
                     if not isinstance(e.cause, urllib.error.HTTPError) or e.cause.code not in (403, 429):
-                        last_error = error_to_compat_str(e.cause or e.msg)
-                        if count < retries:
-                            continue
-                if fatal:
-                    raise
-                self.report_warning(error_to_compat_str(e))
+                        retry.error = e
+                        continue
+                self._error_or_warning(e, fatal=fatal)
                 break
-            else:
-                try:
-                    self._extract_and_report_alerts(data)
-                except ExtractorError as e:
-                    if fatal:
-                        raise
-                    self.report_warning(error_to_compat_str(e))
-                    break
 
-                if dict_get(data, ('contents', 'currentVideoEndpoint', 'onResponseReceivedActions')):
-                    break
+            try:
+                self._extract_and_report_alerts(data)
+            except ExtractorError as e:
+                self._error_or_warning(e, fatal=fatal)
+                break
 
-                last_error = 'Incomplete yt initial data received'
-                if count >= retries:
-                    if fatal:
-                        raise ExtractorError(last_error)
-                    self.report_warning(last_error)
-                    break
+            # Sometimes youtube returns a webpage with incomplete ytInitialData
+            # See: https://github.com/yt-dlp/yt-dlp/issues/116
+            if not traverse_obj(data, 'contents', 'currentVideoEndpoint', 'onResponseReceivedActions'):
+                retry.error = ExtractorError('Incomplete yt initial data received')
+                continue
 
         return webpage, data