yt_dlp/extractor/twitch.py

   1 import collections
   2 import itertools
   3 import json
   4 import random
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_parse_qs,
  10     compat_str,
  11     compat_urllib_parse_urlencode,
  12     compat_urllib_parse_urlparse,
  13 )
  14 from ..utils import (
  15     ExtractorError,
  16     UserNotLive,
  17     base_url,
  18     clean_html,
  19     dict_get,
  20     float_or_none,
  21     int_or_none,
  22     make_archive_id,
  23     parse_duration,
  24     parse_iso8601,
  25     parse_qs,
  26     qualities,
  27     str_or_none,
  28     traverse_obj,
  29     try_get,
  30     unified_timestamp,
  31     update_url_query,
  32     url_or_none,
  33     urljoin,
  34 )
  35
  36
  37 class TwitchBaseIE(InfoExtractor):
  38     _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv'
  39
  40     _API_BASE = 'https://api.twitch.tv'
  41     _USHER_BASE = 'https://usher.ttvnw.net'
  42     _LOGIN_FORM_URL = 'https://www.twitch.tv/login'
  43     _LOGIN_POST_URL = 'https://passport.twitch.tv/login'
  44     _CLIENT_ID = 'kimne78kx3ncx6brgo4mv6wki5h1ko'
  45     _NETRC_MACHINE = 'twitch'
  46
  47     _OPERATION_HASHES = {
  48         'CollectionSideBar': '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14',
  49         'FilterableVideoTower_Videos': 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb',
  50         'ClipsCards__User': 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777',
  51         'ChannelCollectionsContent': '447aec6a0cc1e8d0a8d7732d47eb0762c336a2294fdb009e9c9d854e49d484b9',
  52         'StreamMetadata': 'a647c2a13599e5991e175155f798ca7f1ecddde73f7f341f39009c14dbf59962',
  53         'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
  54         'VideoAccessToken_Clip': '36b89d2507fce29e5ca551df756d27c1cfe079e2609642b4390aa4c35796eb11',
  55         'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
  56         'VideoMetadata': '49b5b8f268cdeb259d75b58dcb0c1a748e3b575003448a2333dc5cdafd49adad',
  57         'VideoPlayer_ChapterSelectButtonVideo': '8d2793384aac3773beab5e59bd5d6f585aedb923d292800119e03d40cd0f9b41',
  58         'VideoPlayer_VODSeekbarPreviewVideo': '07e99e4d56c5a7c67117a154777b0baf85a5ffefa393b213f4bc712ccaf85dd6',
  59     }
  60
  61     def _perform_login(self, username, password):
  62         def fail(message):
  63             raise ExtractorError(
  64                 'Unable to login. Twitch said: %s' % message, expected=True)
  65
  66         def login_step(page, urlh, note, data):
  67             form = self._hidden_inputs(page)
  68             form.update(data)
  69
  70             page_url = urlh.geturl()
  71             post_url = self._search_regex(
  72                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  73                 'post url', default=self._LOGIN_POST_URL, group='url')
  74             post_url = urljoin(page_url, post_url)
  75
  76             headers = {
  77                 'Referer': page_url,
  78                 'Origin': 'https://www.twitch.tv',
  79                 'Content-Type': 'text/plain;charset=UTF-8',
  80             }
  81
  82             response = self._download_json(
  83                 post_url, None, note, data=json.dumps(form).encode(),
  84                 headers=headers, expected_status=400)
  85             error = dict_get(response, ('error', 'error_description', 'error_code'))
  86             if error:
  87                 fail(error)
  88
  89             if 'Authenticated successfully' in response.get('message', ''):
  90                 return None, None
  91
  92             redirect_url = urljoin(
  93                 post_url,
  94                 response.get('redirect') or response['redirect_path'])
  95             return self._download_webpage_handle(
  96                 redirect_url, None, 'Downloading login redirect page',
  97                 headers=headers)
  98
  99         login_page, handle = self._download_webpage_handle(
 100             self._LOGIN_FORM_URL, None, 'Downloading login page')
 101
 102         # Some TOR nodes and public proxies are blocked completely
 103         if 'blacklist_message' in login_page:
 104             fail(clean_html(login_page))
 105
 106         redirect_page, handle = login_step(
 107             login_page, handle, 'Logging in', {
 108                 'username': username,
 109                 'password': password,
 110                 'client_id': self._CLIENT_ID,
 111             })
 112
 113         # Successful login
 114         if not redirect_page:
 115             return
 116
 117         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 118             # TODO: Add mechanism to request an SMS or phone call
 119             tfa_token = self._get_tfa_info('two-factor authentication token')
 120             login_step(redirect_page, handle, 'Submitting TFA token', {
 121                 'authy_token': tfa_token,
 122                 'remember_2fa': 'true',
 123             })
 124
 125     def _prefer_source(self, formats):
 126         try:
 127             source = next(f for f in formats if f['format_id'] == 'Source')
 128             source['quality'] = 10
 129         except StopIteration:
 130             for f in formats:
 131                 if '/chunked/' in f['url']:
 132                     f.update({
 133                         'quality': 10,
 134                         'format_note': 'Source',
 135                     })
 136
 137     def _download_base_gql(self, video_id, ops, note, fatal=True):
 138         headers = {
 139             'Content-Type': 'text/plain;charset=UTF-8',
 140             'Client-ID': self._CLIENT_ID,
 141         }
 142         gql_auth = self._get_cookies('https://gql.twitch.tv').get('auth-token')
 143         if gql_auth:
 144             headers['Authorization'] = 'OAuth ' + gql_auth.value
 145         return self._download_json(
 146             'https://gql.twitch.tv/gql', video_id, note,
 147             data=json.dumps(ops).encode(),
 148             headers=headers, fatal=fatal)
 149
 150     def _download_gql(self, video_id, ops, note, fatal=True):
 151         for op in ops:
 152             op['extensions'] = {
 153                 'persistedQuery': {
 154                     'version': 1,
 155                     'sha256Hash': self._OPERATION_HASHES[op['operationName']],
 156                 }
 157             }
 158         return self._download_base_gql(video_id, ops, note)
 159
 160     def _download_access_token(self, video_id, token_kind, param_name):
 161         method = '%sPlaybackAccessToken' % token_kind
 162         ops = {
 163             'query': '''{
 164               %s(
 165                 %s: "%s",
 166                 params: {
 167                   platform: "web",
 168                   playerBackend: "mediaplayer",
 169                   playerType: "site"
 170                 }
 171               )
 172               {
 173                 value
 174                 signature
 175               }
 176             }''' % (method, param_name, video_id),
 177         }
 178         return self._download_base_gql(
 179             video_id, ops,
 180             'Downloading %s access token GraphQL' % token_kind)['data'][method]
 181
 182     def _get_thumbnails(self, thumbnail):
 183         return [{
 184             'url': re.sub(r'\d+x\d+(\.\w+)($|(?=[?#]))', r'0x0\g<1>', thumbnail),
 185             'preference': 1,
 186         }, {
 187             'url': thumbnail,
 188         }] if thumbnail else None
 189
 190
 191 class TwitchVodIE(TwitchBaseIE):
 192     IE_NAME = 'twitch:vod'
 193     _VALID_URL = r'''(?x)
 194                     https?://
 195                         (?:
 196                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 197                             player\.twitch\.tv/\?.*?\bvideo=v?|
 198                             www\.twitch\.tv/[^/]+/schedule\?vodID=
 199                         )
 200                         (?P<id>\d+)
 201                     '''
 202
 203     _TESTS = [{
 204         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 205         'info_dict': {
 206             'id': 'v6528877',
 207             'ext': 'mp4',
 208             'title': 'LCK Summer Split - Week 6 Day 1',
 209             'thumbnail': r're:^https?://.*\.jpg$',
 210             'duration': 17208,
 211             'timestamp': 1435131734,
 212             'upload_date': '20150624',
 213             'uploader': 'Riot Games',
 214             'uploader_id': 'riotgames',
 215             'view_count': int,
 216             'start_time': 310,
 217             'chapters': [
 218                 {
 219                     'start_time': 0,
 220                     'end_time': 17208,
 221                     'title': 'League of Legends'
 222                 }
 223             ],
 224             'live_status': 'was_live',
 225         },
 226         'params': {
 227             # m3u8 download
 228             'skip_download': True,
 229         },
 230     }, {
 231         # Untitled broadcast (title is None)
 232         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 233         'info_dict': {
 234             'id': 'v11230755',
 235             'ext': 'mp4',
 236             'title': 'Untitled Broadcast',
 237             'thumbnail': r're:^https?://.*\.jpg$',
 238             'duration': 1638,
 239             'timestamp': 1439746708,
 240             'upload_date': '20150816',
 241             'uploader': 'BelkAO_o',
 242             'uploader_id': 'belkao_o',
 243             'view_count': int,
 244         },
 245         'params': {
 246             # m3u8 download
 247             'skip_download': True,
 248         },
 249         'skip': 'HTTP Error 404: Not Found',
 250     }, {
 251         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 252         'only_matching': True,
 253     }, {
 254         'url': 'https://www.twitch.tv/videos/6528877',
 255         'only_matching': True,
 256     }, {
 257         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 258         'only_matching': True,
 259     }, {
 260         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 261         'only_matching': True,
 262     }, {
 263         'url': 'https://player.twitch.tv/?video=480452374',
 264         'only_matching': True,
 265     }, {
 266         'url': 'https://www.twitch.tv/videos/635475444',
 267         'info_dict': {
 268             'id': 'v635475444',
 269             'ext': 'mp4',
 270             'title': 'Riot Games',
 271             'duration': 11643,
 272             'uploader': 'Riot Games',
 273             'uploader_id': 'riotgames',
 274             'timestamp': 1590770569,
 275             'upload_date': '20200529',
 276             'chapters': [
 277                 {
 278                     'start_time': 0,
 279                     'end_time': 573,
 280                     'title': 'League of Legends'
 281                 },
 282                 {
 283                     'start_time': 573,
 284                     'end_time': 3922,
 285                     'title': 'Legends of Runeterra'
 286                 },
 287                 {
 288                     'start_time': 3922,
 289                     'end_time': 11643,
 290                     'title': 'Art'
 291                 }
 292             ],
 293             'live_status': 'was_live',
 294             'thumbnail': r're:^https?://.*\.jpg$',
 295             'view_count': int,
 296         },
 297         'params': {
 298             'skip_download': True
 299         },
 300     }, {
 301         'note': 'Storyboards',
 302         'url': 'https://www.twitch.tv/videos/635475444',
 303         'info_dict': {
 304             'id': 'v635475444',
 305             'format_id': 'sb0',
 306             'ext': 'mhtml',
 307             'title': 'Riot Games',
 308             'duration': 11643,
 309             'uploader': 'Riot Games',
 310             'uploader_id': 'riotgames',
 311             'timestamp': 1590770569,
 312             'upload_date': '20200529',
 313             'chapters': [
 314                 {
 315                     'start_time': 0,
 316                     'end_time': 573,
 317                     'title': 'League of Legends'
 318                 },
 319                 {
 320                     'start_time': 573,
 321                     'end_time': 3922,
 322                     'title': 'Legends of Runeterra'
 323                 },
 324                 {
 325                     'start_time': 3922,
 326                     'end_time': 11643,
 327                     'title': 'Art'
 328                 }
 329             ],
 330             'live_status': 'was_live',
 331             'thumbnail': r're:^https?://.*\.jpg$',
 332             'view_count': int,
 333             'columns': int,
 334             'rows': int,
 335         },
 336         'params': {
 337             'format': 'mhtml',
 338             'skip_download': True
 339         }
 340     }, {
 341         'note': 'VOD with single chapter',
 342         'url': 'https://www.twitch.tv/videos/1536751224',
 343         'info_dict': {
 344             'id': 'v1536751224',
 345             'ext': 'mp4',
 346             'title': 'Porter Robinson Star Guardian Stream Tour with LilyPichu',
 347             'duration': 8353,
 348             'uploader': 'Riot Games',
 349             'uploader_id': 'riotgames',
 350             'timestamp': 1658267731,
 351             'upload_date': '20220719',
 352             'chapters': [
 353                 {
 354                     'start_time': 0,
 355                     'end_time': 8353,
 356                     'title': 'League of Legends'
 357                 }
 358             ],
 359             'live_status': 'was_live',
 360             'thumbnail': r're:^https?://.*\.jpg$',
 361             'view_count': int,
 362         },
 363         'params': {
 364             'skip_download': True
 365         },
 366         'expected_warnings': ['Unable to download JSON metadata: HTTP Error 403: Forbidden']
 367     }, {
 368         'url': 'https://www.twitch.tv/tangotek/schedule?vodID=1822395420',
 369         'only_matching': True,
 370     }]
 371
 372     def _download_info(self, item_id):
 373         data = self._download_gql(
 374             item_id, [{
 375                 'operationName': 'VideoMetadata',
 376                 'variables': {
 377                     'channelLogin': '',
 378                     'videoID': item_id,
 379                 },
 380             }, {
 381                 'operationName': 'VideoPlayer_ChapterSelectButtonVideo',
 382                 'variables': {
 383                     'includePrivate': False,
 384                     'videoID': item_id,
 385                 },
 386             }, {
 387                 'operationName': 'VideoPlayer_VODSeekbarPreviewVideo',
 388                 'variables': {
 389                     'includePrivate': False,
 390                     'videoID': item_id,
 391                 },
 392             }],
 393             'Downloading stream metadata GraphQL')
 394
 395         video = traverse_obj(data, (..., 'data', 'video'), get_all=False)
 396         if video is None:
 397             raise ExtractorError(f'Video {item_id} does not exist', expected=True)
 398
 399         video['moments'] = traverse_obj(data, (..., 'data', 'video', 'moments', 'edges', ..., 'node'))
 400         video['storyboard'] = traverse_obj(
 401             data, (..., 'data', 'video', 'seekPreviewsURL', {url_or_none}), get_all=False)
 402
 403         return video
 404
 405     def _extract_info(self, info):
 406         status = info.get('status')
 407         if status == 'recording':
 408             is_live = True
 409         elif status == 'recorded':
 410             is_live = False
 411         else:
 412             is_live = None
 413         _QUALITIES = ('small', 'medium', 'large')
 414         quality_key = qualities(_QUALITIES)
 415         thumbnails = []
 416         preview = info.get('preview')
 417         if isinstance(preview, dict):
 418             for thumbnail_id, thumbnail_url in preview.items():
 419                 thumbnail_url = url_or_none(thumbnail_url)
 420                 if not thumbnail_url:
 421                     continue
 422                 if thumbnail_id not in _QUALITIES:
 423                     continue
 424                 thumbnails.append({
 425                     'url': thumbnail_url,
 426                     'preference': quality_key(thumbnail_id),
 427                 })
 428         return {
 429             'id': info['_id'],
 430             'title': info.get('title') or 'Untitled Broadcast',
 431             'description': info.get('description'),
 432             'duration': int_or_none(info.get('length')),
 433             'thumbnails': thumbnails,
 434             'uploader': info.get('channel', {}).get('display_name'),
 435             'uploader_id': info.get('channel', {}).get('name'),
 436             'timestamp': parse_iso8601(info.get('recorded_at')),
 437             'view_count': int_or_none(info.get('views')),
 438             'is_live': is_live,
 439             'was_live': True,
 440         }
 441
 442     def _extract_chapters(self, info, item_id):
 443         if not info.get('moments'):
 444             game = traverse_obj(info, ('game', 'displayName'))
 445             if game:
 446                 yield {'title': game}
 447             return
 448
 449         for moment in info['moments']:
 450             start_time = int_or_none(moment.get('positionMilliseconds'), 1000)
 451             duration = int_or_none(moment.get('durationMilliseconds'), 1000)
 452             name = str_or_none(moment.get('description'))
 453
 454             if start_time is None or duration is None:
 455                 self.report_warning(f'Important chapter information missing for chapter {name}', item_id)
 456                 continue
 457             yield {
 458                 'start_time': start_time,
 459                 'end_time': start_time + duration,
 460                 'title': name,
 461             }
 462
 463     def _extract_info_gql(self, info, item_id):
 464         vod_id = info.get('id') or item_id
 465         # id backward compatibility for download archives
 466         if vod_id[0] != 'v':
 467             vod_id = 'v%s' % vod_id
 468         thumbnail = url_or_none(info.get('previewThumbnailURL'))
 469         is_live = None
 470         if thumbnail:
 471             if re.findall(r'/404_processing_[^.?#]+\.png', thumbnail):
 472                 is_live, thumbnail = True, None
 473             else:
 474                 is_live = False
 475
 476         return {
 477             'id': vod_id,
 478             'title': info.get('title') or 'Untitled Broadcast',
 479             'description': info.get('description'),
 480             'duration': int_or_none(info.get('lengthSeconds')),
 481             'thumbnails': self._get_thumbnails(thumbnail),
 482             'uploader': try_get(info, lambda x: x['owner']['displayName'], compat_str),
 483             'uploader_id': try_get(info, lambda x: x['owner']['login'], compat_str),
 484             'timestamp': unified_timestamp(info.get('publishedAt')),
 485             'view_count': int_or_none(info.get('viewCount')),
 486             'chapters': list(self._extract_chapters(info, item_id)),
 487             'is_live': is_live,
 488             'was_live': True,
 489         }
 490
 491     def _extract_storyboard(self, item_id, storyboard_json_url, duration):
 492         if not duration or not storyboard_json_url:
 493             return
 494         spec = self._download_json(storyboard_json_url, item_id, 'Downloading storyboard metadata JSON', fatal=False) or []
 495         # sort from highest quality to lowest
 496         # This makes sb0 the highest-quality format, sb1 - lower, etc which is consistent with youtube sb ordering
 497         spec.sort(key=lambda x: int_or_none(x.get('width')) or 0, reverse=True)
 498         base = base_url(storyboard_json_url)
 499         for i, s in enumerate(spec):
 500             count = int_or_none(s.get('count'))
 501             images = s.get('images')
 502             if not (images and count):
 503                 continue
 504             fragment_duration = duration / len(images)
 505             yield {
 506                 'format_id': f'sb{i}',
 507                 'format_note': 'storyboard',
 508                 'ext': 'mhtml',
 509                 'protocol': 'mhtml',
 510                 'acodec': 'none',
 511                 'vcodec': 'none',
 512                 'url': urljoin(base, images[0]),
 513                 'width': int_or_none(s.get('width')),
 514                 'height': int_or_none(s.get('height')),
 515                 'fps': count / duration,
 516                 'rows': int_or_none(s.get('rows')),
 517                 'columns': int_or_none(s.get('cols')),
 518                 'fragments': [{
 519                     'url': urljoin(base, path),
 520                     'duration': fragment_duration,
 521                 } for path in images],
 522             }
 523
 524     def _real_extract(self, url):
 525         vod_id = self._match_id(url)
 526
 527         video = self._download_info(vod_id)
 528         info = self._extract_info_gql(video, vod_id)
 529         access_token = self._download_access_token(vod_id, 'video', 'id')
 530
 531         formats = self._extract_m3u8_formats(
 532             '%s/vod/%s.m3u8?%s' % (
 533                 self._USHER_BASE, vod_id,
 534                 compat_urllib_parse_urlencode({
 535                     'allow_source': 'true',
 536                     'allow_audio_only': 'true',
 537                     'allow_spectre': 'true',
 538                     'player': 'twitchweb',
 539                     'playlist_include_framerate': 'true',
 540                     'nauth': access_token['value'],
 541                     'nauthsig': access_token['signature'],
 542                 })),
 543             vod_id, 'mp4', entry_protocol='m3u8_native')
 544
 545         formats.extend(self._extract_storyboard(vod_id, video.get('storyboard'), info.get('duration')))
 546
 547         self._prefer_source(formats)
 548         info['formats'] = formats
 549
 550         parsed_url = compat_urllib_parse_urlparse(url)
 551         query = compat_parse_qs(parsed_url.query)
 552         if 't' in query:
 553             info['start_time'] = parse_duration(query['t'][0])
 554
 555         if info.get('timestamp') is not None:
 556             info['subtitles'] = {
 557                 'rechat': [{
 558                     'url': update_url_query(
 559                         'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, {
 560                             'client_id': self._CLIENT_ID,
 561                         }),
 562                     'ext': 'json',
 563                 }],
 564             }
 565
 566         return info
 567
 568
 569 def _make_video_result(node):
 570     assert isinstance(node, dict)
 571     video_id = node.get('id')
 572     if not video_id:
 573         return
 574     return {
 575         '_type': 'url_transparent',
 576         'ie_key': TwitchVodIE.ie_key(),
 577         'id': 'v' + video_id,
 578         'url': 'https://www.twitch.tv/videos/%s' % video_id,
 579         'title': node.get('title'),
 580         'thumbnail': node.get('previewThumbnailURL'),
 581         'duration': float_or_none(node.get('lengthSeconds')),
 582         'view_count': int_or_none(node.get('viewCount')),
 583     }
 584
 585
 586 class TwitchCollectionIE(TwitchBaseIE):
 587     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P<id>[^/]+)'
 588
 589     _TESTS = [{
 590         'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ',
 591         'info_dict': {
 592             'id': 'wlDCoH0zEBZZbQ',
 593             'title': 'Overthrow Nook, capitalism for children',
 594         },
 595         'playlist_mincount': 13,
 596     }]
 597
 598     _OPERATION_NAME = 'CollectionSideBar'
 599
 600     def _real_extract(self, url):
 601         collection_id = self._match_id(url)
 602         collection = self._download_gql(
 603             collection_id, [{
 604                 'operationName': self._OPERATION_NAME,
 605                 'variables': {'collectionID': collection_id},
 606             }],
 607             'Downloading collection GraphQL')[0]['data']['collection']
 608         title = collection.get('title')
 609         entries = []
 610         for edge in collection['items']['edges']:
 611             if not isinstance(edge, dict):
 612                 continue
 613             node = edge.get('node')
 614             if not isinstance(node, dict):
 615                 continue
 616             video = _make_video_result(node)
 617             if video:
 618                 entries.append(video)
 619         return self.playlist_result(
 620             entries, playlist_id=collection_id, playlist_title=title)
 621
 622
 623 class TwitchPlaylistBaseIE(TwitchBaseIE):
 624     _PAGE_LIMIT = 100
 625
 626     def _entries(self, channel_name, *args):
 627         cursor = None
 628         variables_common = self._make_variables(channel_name, *args)
 629         entries_key = '%ss' % self._ENTRY_KIND
 630         for page_num in itertools.count(1):
 631             variables = variables_common.copy()
 632             variables['limit'] = self._PAGE_LIMIT
 633             if cursor:
 634                 variables['cursor'] = cursor
 635             page = self._download_gql(
 636                 channel_name, [{
 637                     'operationName': self._OPERATION_NAME,
 638                     'variables': variables,
 639                 }],
 640                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
 641                 fatal=False)
 642             if not page:
 643                 break
 644             edges = try_get(
 645                 page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
 646             if not edges:
 647                 break
 648             for edge in edges:
 649                 if not isinstance(edge, dict):
 650                     continue
 651                 if edge.get('__typename') != self._EDGE_KIND:
 652                     continue
 653                 node = edge.get('node')
 654                 if not isinstance(node, dict):
 655                     continue
 656                 if node.get('__typename') != self._NODE_KIND:
 657                     continue
 658                 entry = self._extract_entry(node)
 659                 if entry:
 660                     cursor = edge.get('cursor')
 661                     yield entry
 662             if not cursor or not isinstance(cursor, compat_str):
 663                 break
 664
 665
 666 class TwitchVideosIE(TwitchPlaylistBaseIE):
 667     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:videos|profile)'
 668
 669     _TESTS = [{
 670         # All Videos sorted by Date
 671         'url': 'https://www.twitch.tv/spamfish/videos?filter=all',
 672         'info_dict': {
 673             'id': 'spamfish',
 674             'title': 'spamfish - All Videos sorted by Date',
 675         },
 676         'playlist_mincount': 924,
 677     }, {
 678         # All Videos sorted by Popular
 679         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views',
 680         'info_dict': {
 681             'id': 'spamfish',
 682             'title': 'spamfish - All Videos sorted by Popular',
 683         },
 684         'playlist_mincount': 931,
 685     }, {
 686         # Past Broadcasts sorted by Date
 687         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives',
 688         'info_dict': {
 689             'id': 'spamfish',
 690             'title': 'spamfish - Past Broadcasts sorted by Date',
 691         },
 692         'playlist_mincount': 27,
 693     }, {
 694         # Highlights sorted by Date
 695         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights',
 696         'info_dict': {
 697             'id': 'spamfish',
 698             'title': 'spamfish - Highlights sorted by Date',
 699         },
 700         'playlist_mincount': 901,
 701     }, {
 702         # Uploads sorted by Date
 703         'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time',
 704         'info_dict': {
 705             'id': 'esl_csgo',
 706             'title': 'esl_csgo - Uploads sorted by Date',
 707         },
 708         'playlist_mincount': 5,
 709     }, {
 710         # Past Premieres sorted by Date
 711         'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres',
 712         'info_dict': {
 713             'id': 'spamfish',
 714             'title': 'spamfish - Past Premieres sorted by Date',
 715         },
 716         'playlist_mincount': 1,
 717     }, {
 718         'url': 'https://www.twitch.tv/spamfish/videos/all',
 719         'only_matching': True,
 720     }, {
 721         'url': 'https://m.twitch.tv/spamfish/videos/all',
 722         'only_matching': True,
 723     }, {
 724         'url': 'https://www.twitch.tv/spamfish/videos',
 725         'only_matching': True,
 726     }]
 727
 728     Broadcast = collections.namedtuple('Broadcast', ['type', 'label'])
 729
 730     _DEFAULT_BROADCAST = Broadcast(None, 'All Videos')
 731     _BROADCASTS = {
 732         'archives': Broadcast('ARCHIVE', 'Past Broadcasts'),
 733         'highlights': Broadcast('HIGHLIGHT', 'Highlights'),
 734         'uploads': Broadcast('UPLOAD', 'Uploads'),
 735         'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'),
 736         'all': _DEFAULT_BROADCAST,
 737     }
 738
 739     _DEFAULT_SORTED_BY = 'Date'
 740     _SORTED_BY = {
 741         'time': _DEFAULT_SORTED_BY,
 742         'views': 'Popular',
 743     }
 744
 745     _OPERATION_NAME = 'FilterableVideoTower_Videos'
 746     _ENTRY_KIND = 'video'
 747     _EDGE_KIND = 'VideoEdge'
 748     _NODE_KIND = 'Video'
 749
 750     @classmethod
 751     def suitable(cls, url):
 752         return (False
 753                 if any(ie.suitable(url) for ie in (
 754                     TwitchVideosClipsIE,
 755                     TwitchVideosCollectionsIE))
 756                 else super(TwitchVideosIE, cls).suitable(url))
 757
 758     @staticmethod
 759     def _make_variables(channel_name, broadcast_type, sort):
 760         return {
 761             'channelOwnerLogin': channel_name,
 762             'broadcastType': broadcast_type,
 763             'videoSort': sort.upper(),
 764         }
 765
 766     @staticmethod
 767     def _extract_entry(node):
 768         return _make_video_result(node)
 769
 770     def _real_extract(self, url):
 771         channel_name = self._match_id(url)
 772         qs = parse_qs(url)
 773         filter = qs.get('filter', ['all'])[0]
 774         sort = qs.get('sort', ['time'])[0]
 775         broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST)
 776         return self.playlist_result(
 777             self._entries(channel_name, broadcast.type, sort),
 778             playlist_id=channel_name,
 779             playlist_title='%s - %s sorted by %s'
 780             % (channel_name, broadcast.label,
 781                self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY)))
 782
 783
 784 class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
 785     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)'
 786
 787     _TESTS = [{
 788         # Clips
 789         'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all',
 790         'info_dict': {
 791             'id': 'vanillatv',
 792             'title': 'vanillatv - Clips Top All',
 793         },
 794         'playlist_mincount': 1,
 795     }, {
 796         'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d',
 797         'only_matching': True,
 798     }]
 799
 800     Clip = collections.namedtuple('Clip', ['filter', 'label'])
 801
 802     _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D')
 803     _RANGE = {
 804         '24hr': Clip('LAST_DAY', 'Top 24H'),
 805         '7d': _DEFAULT_CLIP,
 806         '30d': Clip('LAST_MONTH', 'Top 30D'),
 807         'all': Clip('ALL_TIME', 'Top All'),
 808     }
 809
 810     # NB: values other than 20 result in skipped videos
 811     _PAGE_LIMIT = 20
 812
 813     _OPERATION_NAME = 'ClipsCards__User'
 814     _ENTRY_KIND = 'clip'
 815     _EDGE_KIND = 'ClipEdge'
 816     _NODE_KIND = 'Clip'
 817
 818     @staticmethod
 819     def _make_variables(channel_name, filter):
 820         return {
 821             'login': channel_name,
 822             'criteria': {
 823                 'filter': filter,
 824             },
 825         }
 826
 827     @staticmethod
 828     def _extract_entry(node):
 829         assert isinstance(node, dict)
 830         clip_url = url_or_none(node.get('url'))
 831         if not clip_url:
 832             return
 833         return {
 834             '_type': 'url_transparent',
 835             'ie_key': TwitchClipsIE.ie_key(),
 836             'id': node.get('id'),
 837             'url': clip_url,
 838             'title': node.get('title'),
 839             'thumbnail': node.get('thumbnailURL'),
 840             'duration': float_or_none(node.get('durationSeconds')),
 841             'timestamp': unified_timestamp(node.get('createdAt')),
 842             'view_count': int_or_none(node.get('viewCount')),
 843             'language': node.get('language'),
 844         }
 845
 846     def _real_extract(self, url):
 847         channel_name = self._match_id(url)
 848         qs = parse_qs(url)
 849         range = qs.get('range', ['7d'])[0]
 850         clip = self._RANGE.get(range, self._DEFAULT_CLIP)
 851         return self.playlist_result(
 852             self._entries(channel_name, clip.filter),
 853             playlist_id=channel_name,
 854             playlist_title='%s - Clips %s' % (channel_name, clip.label))
 855
 856
 857 class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
 858     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/videos/*?\?.*?\bfilter=collections'
 859
 860     _TESTS = [{
 861         # Collections
 862         'url': 'https://www.twitch.tv/spamfish/videos?filter=collections',
 863         'info_dict': {
 864             'id': 'spamfish',
 865             'title': 'spamfish - Collections',
 866         },
 867         'playlist_mincount': 3,
 868     }, {
 869         'url': 'https://www.twitch.tv/monstercat/videos?filter=collections',
 870         'info_dict': {
 871             'id': 'monstercat',
 872             'title': 'monstercat - Collections',
 873         },
 874         'playlist_mincount': 13,
 875     }]
 876
 877     _OPERATION_NAME = 'ChannelCollectionsContent'
 878     _ENTRY_KIND = 'collection'
 879     _EDGE_KIND = 'CollectionsItemEdge'
 880     _NODE_KIND = 'Collection'
 881
 882     @staticmethod
 883     def _make_variables(channel_name):
 884         return {
 885             'ownerLogin': channel_name,
 886         }
 887
 888     @staticmethod
 889     def _extract_entry(node):
 890         assert isinstance(node, dict)
 891         collection_id = node.get('id')
 892         if not collection_id:
 893             return
 894         return {
 895             '_type': 'url_transparent',
 896             'ie_key': TwitchCollectionIE.ie_key(),
 897             'id': collection_id,
 898             'url': 'https://www.twitch.tv/collections/%s' % collection_id,
 899             'title': node.get('title'),
 900             'thumbnail': node.get('thumbnailURL'),
 901             'duration': float_or_none(node.get('lengthSeconds')),
 902             'timestamp': unified_timestamp(node.get('updatedAt')),
 903             'view_count': int_or_none(node.get('viewCount')),
 904         }
 905
 906     def _real_extract(self, url):
 907         channel_name = self._match_id(url)
 908         return self.playlist_result(
 909             self._entries(channel_name), playlist_id=channel_name,
 910             playlist_title='%s - Collections' % channel_name)
 911
 912
 913 class TwitchStreamIE(TwitchBaseIE):
 914     IE_NAME = 'twitch:stream'
 915     _VALID_URL = r'''(?x)
 916                     https?://
 917                         (?:
 918                             (?:(?:www|go|m)\.)?twitch\.tv/|
 919                             player\.twitch\.tv/\?.*?\bchannel=
 920                         )
 921                         (?P<id>[^/#?]+)
 922                     '''
 923
 924     _TESTS = [{
 925         'url': 'http://www.twitch.tv/shroomztv',
 926         'info_dict': {
 927             'id': '12772022048',
 928             'display_id': 'shroomztv',
 929             'ext': 'mp4',
 930             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 931             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 932             'is_live': True,
 933             'timestamp': 1421928037,
 934             'upload_date': '20150122',
 935             'uploader': 'ShroomzTV',
 936             'uploader_id': 'shroomztv',
 937             'view_count': int,
 938         },
 939         'params': {
 940             # m3u8 download
 941             'skip_download': True,
 942         },
 943         'skip': 'User does not exist',
 944     }, {
 945         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 946         'only_matching': True,
 947     }, {
 948         'url': 'https://player.twitch.tv/?channel=lotsofs',
 949         'only_matching': True,
 950     }, {
 951         'url': 'https://go.twitch.tv/food',
 952         'only_matching': True,
 953     }, {
 954         'url': 'https://m.twitch.tv/food',
 955         'only_matching': True,
 956     }, {
 957         'url': 'https://www.twitch.tv/monstercat',
 958         'info_dict': {
 959             'id': '40500071752',
 960             'display_id': 'monstercat',
 961             'title': 're:Monstercat',
 962             'description': 'md5:0945ad625e615bc8f0469396537d87d9',
 963             'is_live': True,
 964             'timestamp': 1677107190,
 965             'upload_date': '20230222',
 966             'uploader': 'Monstercat',
 967             'uploader_id': 'monstercat',
 968             'live_status': 'is_live',
 969             'thumbnail': 're:https://.*.jpg',
 970             'ext': 'mp4',
 971         },
 972         'params': {
 973             'skip_download': 'Livestream',
 974         },
 975     }]
 976
 977     @classmethod
 978     def suitable(cls, url):
 979         return (False
 980                 if any(ie.suitable(url) for ie in (
 981                     TwitchVodIE,
 982                     TwitchCollectionIE,
 983                     TwitchVideosIE,
 984                     TwitchVideosClipsIE,
 985                     TwitchVideosCollectionsIE,
 986                     TwitchClipsIE))
 987                 else super(TwitchStreamIE, cls).suitable(url))
 988
 989     def _real_extract(self, url):
 990         channel_name = self._match_id(url).lower()
 991
 992         gql = self._download_gql(
 993             channel_name, [{
 994                 'operationName': 'StreamMetadata',
 995                 'variables': {'channelLogin': channel_name},
 996             }, {
 997                 'operationName': 'ComscoreStreamingQuery',
 998                 'variables': {
 999                     'channel': channel_name,
1000                     'clipSlug': '',
1001                     'isClip': False,
1002                     'isLive': True,
1003                     'isVodOrCollection': False,
1004                     'vodID': '',
1005                 },
1006             }, {
1007                 'operationName': 'VideoPreviewOverlay',
1008                 'variables': {'login': channel_name},
1009             }],
1010             'Downloading stream GraphQL')
1011
1012         user = gql[0]['data']['user']
1013
1014         if not user:
1015             raise ExtractorError(
1016                 '%s does not exist' % channel_name, expected=True)
1017
1018         stream = user['stream']
1019
1020         if not stream:
1021             raise UserNotLive(video_id=channel_name)
1022
1023         access_token = self._download_access_token(
1024             channel_name, 'stream', 'channelName')
1025         token = access_token['value']
1026
1027         stream_id = stream.get('id') or channel_name
1028         query = {
1029             'allow_source': 'true',
1030             'allow_audio_only': 'true',
1031             'allow_spectre': 'true',
1032             'p': random.randint(1000000, 10000000),
1033             'player': 'twitchweb',
1034             'playlist_include_framerate': 'true',
1035             'segment_preference': '4',
1036             'sig': access_token['signature'].encode('utf-8'),
1037             'token': token.encode('utf-8'),
1038         }
1039         formats = self._extract_m3u8_formats(
1040             '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
1041             stream_id, 'mp4', query=query)
1042         self._prefer_source(formats)
1043
1044         view_count = stream.get('viewers')
1045         timestamp = unified_timestamp(stream.get('createdAt'))
1046
1047         sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
1048         uploader = sq_user.get('displayName')
1049         description = try_get(
1050             sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
1051
1052         thumbnail = url_or_none(try_get(
1053             gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
1054             compat_str))
1055
1056         title = uploader or channel_name
1057         stream_type = stream.get('type')
1058         if stream_type in ['rerun', 'live']:
1059             title += ' (%s)' % stream_type
1060
1061         return {
1062             'id': stream_id,
1063             'display_id': channel_name,
1064             'title': title,
1065             'description': description,
1066             'thumbnails': self._get_thumbnails(thumbnail),
1067             'uploader': uploader,
1068             'uploader_id': channel_name,
1069             'timestamp': timestamp,
1070             'view_count': view_count,
1071             'formats': formats,
1072             'is_live': stream_type == 'live',
1073         }
1074
1075
1076 class TwitchClipsIE(TwitchBaseIE):
1077     IE_NAME = 'twitch:clips'
1078     _VALID_URL = r'''(?x)
1079                     https?://
1080                         (?:
1081                             clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)|
1082                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/)?clip/
1083                         )
1084                         (?P<id>[^/?#&]+)
1085                     '''
1086
1087     _TESTS = [{
1088         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
1089         'md5': '761769e1eafce0ffebfb4089cb3847cd',
1090         'info_dict': {
1091             'id': '42850523',
1092             'display_id': 'FaintLightGullWholeWheat',
1093             'ext': 'mp4',
1094             'title': 'EA Play 2016 Live from the Novo Theatre',
1095             'thumbnail': r're:^https?://.*\.jpg',
1096             'timestamp': 1465767393,
1097             'upload_date': '20160612',
1098             'creator': 'EA',
1099             'uploader': 'stereotype_',
1100             'uploader_id': '43566419',
1101         },
1102     }, {
1103         # multiple formats
1104         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
1105         'only_matching': True,
1106     }, {
1107         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
1108         'only_matching': True,
1109     }, {
1110         'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited',
1111         'only_matching': True,
1112     }, {
1113         'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1114         'only_matching': True,
1115     }, {
1116         'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1117         'only_matching': True,
1118     }, {
1119         'url': 'https://m.twitch.tv/clip/FaintLightGullWholeWheat',
1120         'only_matching': True,
1121     }]
1122
1123     def _real_extract(self, url):
1124         video_id = self._match_id(url)
1125
1126         clip = self._download_gql(
1127             video_id, [{
1128                 'operationName': 'VideoAccessToken_Clip',
1129                 'variables': {
1130                     'slug': video_id,
1131                 },
1132             }],
1133             'Downloading clip access token GraphQL')[0]['data']['clip']
1134
1135         if not clip:
1136             raise ExtractorError(
1137                 'This clip is no longer available', expected=True)
1138
1139         access_query = {
1140             'sig': clip['playbackAccessToken']['signature'],
1141             'token': clip['playbackAccessToken']['value'],
1142         }
1143
1144         data = self._download_base_gql(
1145             video_id, {
1146                 'query': '''{
1147   clip(slug: "%s") {
1148     broadcaster {
1149       displayName
1150     }
1151     createdAt
1152     curator {
1153       displayName
1154       id
1155     }
1156     durationSeconds
1157     id
1158     tiny: thumbnailURL(width: 86, height: 45)
1159     small: thumbnailURL(width: 260, height: 147)
1160     medium: thumbnailURL(width: 480, height: 272)
1161     title
1162     videoQualities {
1163       frameRate
1164       quality
1165       sourceURL
1166     }
1167     viewCount
1168   }
1169 }''' % video_id}, 'Downloading clip GraphQL', fatal=False)
1170
1171         if data:
1172             clip = try_get(data, lambda x: x['data']['clip'], dict) or clip
1173
1174         formats = []
1175         for option in clip.get('videoQualities', []):
1176             if not isinstance(option, dict):
1177                 continue
1178             source = url_or_none(option.get('sourceURL'))
1179             if not source:
1180                 continue
1181             formats.append({
1182                 'url': update_url_query(source, access_query),
1183                 'format_id': option.get('quality'),
1184                 'height': int_or_none(option.get('quality')),
1185                 'fps': int_or_none(option.get('frameRate')),
1186             })
1187
1188         thumbnails = []
1189         for thumbnail_id in ('tiny', 'small', 'medium'):
1190             thumbnail_url = clip.get(thumbnail_id)
1191             if not thumbnail_url:
1192                 continue
1193             thumb = {
1194                 'id': thumbnail_id,
1195                 'url': thumbnail_url,
1196             }
1197             mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url)
1198             if mobj:
1199                 thumb.update({
1200                     'height': int(mobj.group(2)),
1201                     'width': int(mobj.group(1)),
1202                 })
1203             thumbnails.append(thumb)
1204
1205         old_id = self._search_regex(r'%7C(\d+)(?:-\d+)?.mp4', formats[-1]['url'], 'old id', default=None)
1206
1207         return {
1208             'id': clip.get('id') or video_id,
1209             '_old_archive_ids': [make_archive_id(self, old_id)] if old_id else None,
1210             'display_id': video_id,
1211             'title': clip.get('title'),
1212             'formats': formats,
1213             'duration': int_or_none(clip.get('durationSeconds')),
1214             'view_count': int_or_none(clip.get('viewCount')),
1215             'timestamp': unified_timestamp(clip.get('createdAt')),
1216             'thumbnails': thumbnails,
1217             'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str),
1218             'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str),
1219             'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
1220         }