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     _NETRC_MACHINE = 'twitch'
  45
  46     _OPERATION_HASHES = {
  47         'CollectionSideBar': '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14',
  48         'FilterableVideoTower_Videos': 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb',
  49         'ClipsCards__User': 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777',
  50         'ChannelCollectionsContent': '447aec6a0cc1e8d0a8d7732d47eb0762c336a2294fdb009e9c9d854e49d484b9',
  51         'StreamMetadata': 'a647c2a13599e5991e175155f798ca7f1ecddde73f7f341f39009c14dbf59962',
  52         'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
  53         'VideoAccessToken_Clip': '36b89d2507fce29e5ca551df756d27c1cfe079e2609642b4390aa4c35796eb11',
  54         'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
  55         'VideoMetadata': '49b5b8f268cdeb259d75b58dcb0c1a748e3b575003448a2333dc5cdafd49adad',
  56         'VideoPlayer_ChapterSelectButtonVideo': '8d2793384aac3773beab5e59bd5d6f585aedb923d292800119e03d40cd0f9b41',
  57         'VideoPlayer_VODSeekbarPreviewVideo': '07e99e4d56c5a7c67117a154777b0baf85a5ffefa393b213f4bc712ccaf85dd6',
  58     }
  59
  60     @property
  61     def _CLIENT_ID(self):
  62         return self._configuration_arg(
  63             'client_id', ['ue6666qo983tsx6so1t0vnawi233wa'], ie_key=TwitchStreamIE, casesense=True)[0]
  64
  65     def _perform_login(self, username, password):
  66         def fail(message):
  67             raise ExtractorError(
  68                 'Unable to login. Twitch said: %s' % message, expected=True)
  69
  70         def login_step(page, urlh, note, data):
  71             form = self._hidden_inputs(page)
  72             form.update(data)
  73
  74             page_url = urlh.geturl()
  75             post_url = self._search_regex(
  76                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  77                 'post url', default=self._LOGIN_POST_URL, group='url')
  78             post_url = urljoin(page_url, post_url)
  79
  80             headers = {
  81                 'Referer': page_url,
  82                 'Origin': 'https://www.twitch.tv',
  83                 'Content-Type': 'text/plain;charset=UTF-8',
  84             }
  85
  86             response = self._download_json(
  87                 post_url, None, note, data=json.dumps(form).encode(),
  88                 headers=headers, expected_status=400)
  89             error = dict_get(response, ('error', 'error_description', 'error_code'))
  90             if error:
  91                 fail(error)
  92
  93             if 'Authenticated successfully' in response.get('message', ''):
  94                 return None, None
  95
  96             redirect_url = urljoin(
  97                 post_url,
  98                 response.get('redirect') or response['redirect_path'])
  99             return self._download_webpage_handle(
 100                 redirect_url, None, 'Downloading login redirect page',
 101                 headers=headers)
 102
 103         login_page, handle = self._download_webpage_handle(
 104             self._LOGIN_FORM_URL, None, 'Downloading login page')
 105
 106         # Some TOR nodes and public proxies are blocked completely
 107         if 'blacklist_message' in login_page:
 108             fail(clean_html(login_page))
 109
 110         redirect_page, handle = login_step(
 111             login_page, handle, 'Logging in', {
 112                 'username': username,
 113                 'password': password,
 114                 'client_id': self._CLIENT_ID,
 115             })
 116
 117         # Successful login
 118         if not redirect_page:
 119             return
 120
 121         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 122             # TODO: Add mechanism to request an SMS or phone call
 123             tfa_token = self._get_tfa_info('two-factor authentication token')
 124             login_step(redirect_page, handle, 'Submitting TFA token', {
 125                 'authy_token': tfa_token,
 126                 'remember_2fa': 'true',
 127             })
 128
 129     def _prefer_source(self, formats):
 130         try:
 131             source = next(f for f in formats if f['format_id'] == 'Source')
 132             source['quality'] = 10
 133         except StopIteration:
 134             for f in formats:
 135                 if '/chunked/' in f['url']:
 136                     f.update({
 137                         'quality': 10,
 138                         'format_note': 'Source',
 139                     })
 140
 141     def _download_base_gql(self, video_id, ops, note, fatal=True):
 142         headers = {
 143             'Content-Type': 'text/plain;charset=UTF-8',
 144             'Client-ID': self._CLIENT_ID,
 145         }
 146         gql_auth = self._get_cookies('https://gql.twitch.tv').get('auth-token')
 147         if gql_auth:
 148             headers['Authorization'] = 'OAuth ' + gql_auth.value
 149         return self._download_json(
 150             'https://gql.twitch.tv/gql', video_id, note,
 151             data=json.dumps(ops).encode(),
 152             headers=headers, fatal=fatal)
 153
 154     def _download_gql(self, video_id, ops, note, fatal=True):
 155         for op in ops:
 156             op['extensions'] = {
 157                 'persistedQuery': {
 158                     'version': 1,
 159                     'sha256Hash': self._OPERATION_HASHES[op['operationName']],
 160                 }
 161             }
 162         return self._download_base_gql(video_id, ops, note)
 163
 164     def _download_access_token(self, video_id, token_kind, param_name):
 165         method = '%sPlaybackAccessToken' % token_kind
 166         ops = {
 167             'query': '''{
 168               %s(
 169                 %s: "%s",
 170                 params: {
 171                   platform: "web",
 172                   playerBackend: "mediaplayer",
 173                   playerType: "site"
 174                 }
 175               )
 176               {
 177                 value
 178                 signature
 179               }
 180             }''' % (method, param_name, video_id),
 181         }
 182         return self._download_base_gql(
 183             video_id, ops,
 184             'Downloading %s access token GraphQL' % token_kind)['data'][method]
 185
 186     def _get_thumbnails(self, thumbnail):
 187         return [{
 188             'url': re.sub(r'\d+x\d+(\.\w+)($|(?=[?#]))', r'0x0\g<1>', thumbnail),
 189             'preference': 1,
 190         }, {
 191             'url': thumbnail,
 192         }] if thumbnail else None
 193
 194
 195 class TwitchVodIE(TwitchBaseIE):
 196     IE_NAME = 'twitch:vod'
 197     _VALID_URL = r'''(?x)
 198                     https?://
 199                         (?:
 200                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 201                             player\.twitch\.tv/\?.*?\bvideo=v?|
 202                             www\.twitch\.tv/[^/]+/schedule\?vodID=
 203                         )
 204                         (?P<id>\d+)
 205                     '''
 206
 207     _TESTS = [{
 208         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 209         'info_dict': {
 210             'id': 'v6528877',
 211             'ext': 'mp4',
 212             'title': 'LCK Summer Split - Week 6 Day 1',
 213             'thumbnail': r're:^https?://.*\.jpg$',
 214             'duration': 17208,
 215             'timestamp': 1435131734,
 216             'upload_date': '20150624',
 217             'uploader': 'Riot Games',
 218             'uploader_id': 'riotgames',
 219             'view_count': int,
 220             'start_time': 310,
 221             'chapters': [
 222                 {
 223                     'start_time': 0,
 224                     'end_time': 17208,
 225                     'title': 'League of Legends'
 226                 }
 227             ],
 228             'live_status': 'was_live',
 229         },
 230         'params': {
 231             # m3u8 download
 232             'skip_download': True,
 233         },
 234     }, {
 235         # Untitled broadcast (title is None)
 236         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 237         'info_dict': {
 238             'id': 'v11230755',
 239             'ext': 'mp4',
 240             'title': 'Untitled Broadcast',
 241             'thumbnail': r're:^https?://.*\.jpg$',
 242             'duration': 1638,
 243             'timestamp': 1439746708,
 244             'upload_date': '20150816',
 245             'uploader': 'BelkAO_o',
 246             'uploader_id': 'belkao_o',
 247             'view_count': int,
 248         },
 249         'params': {
 250             # m3u8 download
 251             'skip_download': True,
 252         },
 253         'skip': 'HTTP Error 404: Not Found',
 254     }, {
 255         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 256         'only_matching': True,
 257     }, {
 258         'url': 'https://www.twitch.tv/videos/6528877',
 259         'only_matching': True,
 260     }, {
 261         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 262         'only_matching': True,
 263     }, {
 264         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 265         'only_matching': True,
 266     }, {
 267         'url': 'https://player.twitch.tv/?video=480452374',
 268         'only_matching': True,
 269     }, {
 270         'url': 'https://www.twitch.tv/videos/635475444',
 271         'info_dict': {
 272             'id': 'v635475444',
 273             'ext': 'mp4',
 274             'title': 'Riot Games',
 275             'duration': 11643,
 276             'uploader': 'Riot Games',
 277             'uploader_id': 'riotgames',
 278             'timestamp': 1590770569,
 279             'upload_date': '20200529',
 280             'chapters': [
 281                 {
 282                     'start_time': 0,
 283                     'end_time': 573,
 284                     'title': 'League of Legends'
 285                 },
 286                 {
 287                     'start_time': 573,
 288                     'end_time': 3922,
 289                     'title': 'Legends of Runeterra'
 290                 },
 291                 {
 292                     'start_time': 3922,
 293                     'end_time': 11643,
 294                     'title': 'Art'
 295                 }
 296             ],
 297             'live_status': 'was_live',
 298             'thumbnail': r're:^https?://.*\.jpg$',
 299             'view_count': int,
 300         },
 301         'params': {
 302             'skip_download': True
 303         },
 304     }, {
 305         'note': 'Storyboards',
 306         'url': 'https://www.twitch.tv/videos/635475444',
 307         'info_dict': {
 308             'id': 'v635475444',
 309             'format_id': 'sb0',
 310             'ext': 'mhtml',
 311             'title': 'Riot Games',
 312             'duration': 11643,
 313             'uploader': 'Riot Games',
 314             'uploader_id': 'riotgames',
 315             'timestamp': 1590770569,
 316             'upload_date': '20200529',
 317             'chapters': [
 318                 {
 319                     'start_time': 0,
 320                     'end_time': 573,
 321                     'title': 'League of Legends'
 322                 },
 323                 {
 324                     'start_time': 573,
 325                     'end_time': 3922,
 326                     'title': 'Legends of Runeterra'
 327                 },
 328                 {
 329                     'start_time': 3922,
 330                     'end_time': 11643,
 331                     'title': 'Art'
 332                 }
 333             ],
 334             'live_status': 'was_live',
 335             'thumbnail': r're:^https?://.*\.jpg$',
 336             'view_count': int,
 337             'columns': int,
 338             'rows': int,
 339         },
 340         'params': {
 341             'format': 'mhtml',
 342             'skip_download': True
 343         }
 344     }, {
 345         'note': 'VOD with single chapter',
 346         'url': 'https://www.twitch.tv/videos/1536751224',
 347         'info_dict': {
 348             'id': 'v1536751224',
 349             'ext': 'mp4',
 350             'title': 'Porter Robinson Star Guardian Stream Tour with LilyPichu',
 351             'duration': 8353,
 352             'uploader': 'Riot Games',
 353             'uploader_id': 'riotgames',
 354             'timestamp': 1658267731,
 355             'upload_date': '20220719',
 356             'chapters': [
 357                 {
 358                     'start_time': 0,
 359                     'end_time': 8353,
 360                     'title': 'League of Legends'
 361                 }
 362             ],
 363             'live_status': 'was_live',
 364             'thumbnail': r're:^https?://.*\.jpg$',
 365             'view_count': int,
 366         },
 367         'params': {
 368             'skip_download': True
 369         },
 370         'expected_warnings': ['Unable to download JSON metadata: HTTP Error 403: Forbidden']
 371     }, {
 372         'url': 'https://www.twitch.tv/tangotek/schedule?vodID=1822395420',
 373         'only_matching': True,
 374     }]
 375
 376     def _download_info(self, item_id):
 377         data = self._download_gql(
 378             item_id, [{
 379                 'operationName': 'VideoMetadata',
 380                 'variables': {
 381                     'channelLogin': '',
 382                     'videoID': item_id,
 383                 },
 384             }, {
 385                 'operationName': 'VideoPlayer_ChapterSelectButtonVideo',
 386                 'variables': {
 387                     'includePrivate': False,
 388                     'videoID': item_id,
 389                 },
 390             }, {
 391                 'operationName': 'VideoPlayer_VODSeekbarPreviewVideo',
 392                 'variables': {
 393                     'includePrivate': False,
 394                     'videoID': item_id,
 395                 },
 396             }],
 397             'Downloading stream metadata GraphQL')
 398
 399         video = traverse_obj(data, (..., 'data', 'video'), get_all=False)
 400         if video is None:
 401             raise ExtractorError(f'Video {item_id} does not exist', expected=True)
 402
 403         video['moments'] = traverse_obj(data, (..., 'data', 'video', 'moments', 'edges', ..., 'node'))
 404         video['storyboard'] = traverse_obj(
 405             data, (..., 'data', 'video', 'seekPreviewsURL', {url_or_none}), get_all=False)
 406
 407         return video
 408
 409     def _extract_info(self, info):
 410         status = info.get('status')
 411         if status == 'recording':
 412             is_live = True
 413         elif status == 'recorded':
 414             is_live = False
 415         else:
 416             is_live = None
 417         _QUALITIES = ('small', 'medium', 'large')
 418         quality_key = qualities(_QUALITIES)
 419         thumbnails = []
 420         preview = info.get('preview')
 421         if isinstance(preview, dict):
 422             for thumbnail_id, thumbnail_url in preview.items():
 423                 thumbnail_url = url_or_none(thumbnail_url)
 424                 if not thumbnail_url:
 425                     continue
 426                 if thumbnail_id not in _QUALITIES:
 427                     continue
 428                 thumbnails.append({
 429                     'url': thumbnail_url,
 430                     'preference': quality_key(thumbnail_id),
 431                 })
 432         return {
 433             'id': info['_id'],
 434             'title': info.get('title') or 'Untitled Broadcast',
 435             'description': info.get('description'),
 436             'duration': int_or_none(info.get('length')),
 437             'thumbnails': thumbnails,
 438             'uploader': info.get('channel', {}).get('display_name'),
 439             'uploader_id': info.get('channel', {}).get('name'),
 440             'timestamp': parse_iso8601(info.get('recorded_at')),
 441             'view_count': int_or_none(info.get('views')),
 442             'is_live': is_live,
 443             'was_live': True,
 444         }
 445
 446     def _extract_chapters(self, info, item_id):
 447         if not info.get('moments'):
 448             game = traverse_obj(info, ('game', 'displayName'))
 449             if game:
 450                 yield {'title': game}
 451             return
 452
 453         for moment in info['moments']:
 454             start_time = int_or_none(moment.get('positionMilliseconds'), 1000)
 455             duration = int_or_none(moment.get('durationMilliseconds'), 1000)
 456             name = str_or_none(moment.get('description'))
 457
 458             if start_time is None or duration is None:
 459                 self.report_warning(f'Important chapter information missing for chapter {name}', item_id)
 460                 continue
 461             yield {
 462                 'start_time': start_time,
 463                 'end_time': start_time + duration,
 464                 'title': name,
 465             }
 466
 467     def _extract_info_gql(self, info, item_id):
 468         vod_id = info.get('id') or item_id
 469         # id backward compatibility for download archives
 470         if vod_id[0] != 'v':
 471             vod_id = 'v%s' % vod_id
 472         thumbnail = url_or_none(info.get('previewThumbnailURL'))
 473         is_live = None
 474         if thumbnail:
 475             if re.findall(r'/404_processing_[^.?#]+\.png', thumbnail):
 476                 is_live, thumbnail = True, None
 477             else:
 478                 is_live = False
 479
 480         return {
 481             'id': vod_id,
 482             'title': info.get('title') or 'Untitled Broadcast',
 483             'description': info.get('description'),
 484             'duration': int_or_none(info.get('lengthSeconds')),
 485             'thumbnails': self._get_thumbnails(thumbnail),
 486             'uploader': try_get(info, lambda x: x['owner']['displayName'], compat_str),
 487             'uploader_id': try_get(info, lambda x: x['owner']['login'], compat_str),
 488             'timestamp': unified_timestamp(info.get('publishedAt')),
 489             'view_count': int_or_none(info.get('viewCount')),
 490             'chapters': list(self._extract_chapters(info, item_id)),
 491             'is_live': is_live,
 492             'was_live': True,
 493         }
 494
 495     def _extract_storyboard(self, item_id, storyboard_json_url, duration):
 496         if not duration or not storyboard_json_url:
 497             return
 498         spec = self._download_json(storyboard_json_url, item_id, 'Downloading storyboard metadata JSON', fatal=False) or []
 499         # sort from highest quality to lowest
 500         # This makes sb0 the highest-quality format, sb1 - lower, etc which is consistent with youtube sb ordering
 501         spec.sort(key=lambda x: int_or_none(x.get('width')) or 0, reverse=True)
 502         base = base_url(storyboard_json_url)
 503         for i, s in enumerate(spec):
 504             count = int_or_none(s.get('count'))
 505             images = s.get('images')
 506             if not (images and count):
 507                 continue
 508             fragment_duration = duration / len(images)
 509             yield {
 510                 'format_id': f'sb{i}',
 511                 'format_note': 'storyboard',
 512                 'ext': 'mhtml',
 513                 'protocol': 'mhtml',
 514                 'acodec': 'none',
 515                 'vcodec': 'none',
 516                 'url': urljoin(base, images[0]),
 517                 'width': int_or_none(s.get('width')),
 518                 'height': int_or_none(s.get('height')),
 519                 'fps': count / duration,
 520                 'rows': int_or_none(s.get('rows')),
 521                 'columns': int_or_none(s.get('cols')),
 522                 'fragments': [{
 523                     'url': urljoin(base, path),
 524                     'duration': fragment_duration,
 525                 } for path in images],
 526             }
 527
 528     def _real_extract(self, url):
 529         vod_id = self._match_id(url)
 530
 531         video = self._download_info(vod_id)
 532         info = self._extract_info_gql(video, vod_id)
 533         access_token = self._download_access_token(vod_id, 'video', 'id')
 534
 535         formats = self._extract_m3u8_formats(
 536             '%s/vod/%s.m3u8?%s' % (
 537                 self._USHER_BASE, vod_id,
 538                 compat_urllib_parse_urlencode({
 539                     'allow_source': 'true',
 540                     'allow_audio_only': 'true',
 541                     'allow_spectre': 'true',
 542                     'player': 'twitchweb',
 543                     'playlist_include_framerate': 'true',
 544                     'nauth': access_token['value'],
 545                     'nauthsig': access_token['signature'],
 546                 })),
 547             vod_id, 'mp4', entry_protocol='m3u8_native')
 548
 549         formats.extend(self._extract_storyboard(vod_id, video.get('storyboard'), info.get('duration')))
 550
 551         self._prefer_source(formats)
 552         info['formats'] = formats
 553
 554         parsed_url = compat_urllib_parse_urlparse(url)
 555         query = compat_parse_qs(parsed_url.query)
 556         if 't' in query:
 557             info['start_time'] = parse_duration(query['t'][0])
 558
 559         if info.get('timestamp') is not None:
 560             info['subtitles'] = {
 561                 'rechat': [{
 562                     'url': update_url_query(
 563                         'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, {
 564                             'client_id': self._CLIENT_ID,
 565                         }),
 566                     'ext': 'json',
 567                 }],
 568             }
 569
 570         return info
 571
 572
 573 def _make_video_result(node):
 574     assert isinstance(node, dict)
 575     video_id = node.get('id')
 576     if not video_id:
 577         return
 578     return {
 579         '_type': 'url_transparent',
 580         'ie_key': TwitchVodIE.ie_key(),
 581         'id': 'v' + video_id,
 582         'url': 'https://www.twitch.tv/videos/%s' % video_id,
 583         'title': node.get('title'),
 584         'thumbnail': node.get('previewThumbnailURL'),
 585         'duration': float_or_none(node.get('lengthSeconds')),
 586         'view_count': int_or_none(node.get('viewCount')),
 587     }
 588
 589
 590 class TwitchCollectionIE(TwitchBaseIE):
 591     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P<id>[^/]+)'
 592
 593     _TESTS = [{
 594         'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ',
 595         'info_dict': {
 596             'id': 'wlDCoH0zEBZZbQ',
 597             'title': 'Overthrow Nook, capitalism for children',
 598         },
 599         'playlist_mincount': 13,
 600     }]
 601
 602     _OPERATION_NAME = 'CollectionSideBar'
 603
 604     def _real_extract(self, url):
 605         collection_id = self._match_id(url)
 606         collection = self._download_gql(
 607             collection_id, [{
 608                 'operationName': self._OPERATION_NAME,
 609                 'variables': {'collectionID': collection_id},
 610             }],
 611             'Downloading collection GraphQL')[0]['data']['collection']
 612         title = collection.get('title')
 613         entries = []
 614         for edge in collection['items']['edges']:
 615             if not isinstance(edge, dict):
 616                 continue
 617             node = edge.get('node')
 618             if not isinstance(node, dict):
 619                 continue
 620             video = _make_video_result(node)
 621             if video:
 622                 entries.append(video)
 623         return self.playlist_result(
 624             entries, playlist_id=collection_id, playlist_title=title)
 625
 626
 627 class TwitchPlaylistBaseIE(TwitchBaseIE):
 628     _PAGE_LIMIT = 100
 629
 630     def _entries(self, channel_name, *args):
 631         cursor = None
 632         variables_common = self._make_variables(channel_name, *args)
 633         entries_key = '%ss' % self._ENTRY_KIND
 634         for page_num in itertools.count(1):
 635             variables = variables_common.copy()
 636             variables['limit'] = self._PAGE_LIMIT
 637             if cursor:
 638                 variables['cursor'] = cursor
 639             page = self._download_gql(
 640                 channel_name, [{
 641                     'operationName': self._OPERATION_NAME,
 642                     'variables': variables,
 643                 }],
 644                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
 645                 fatal=False)
 646             if not page:
 647                 break
 648             edges = try_get(
 649                 page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
 650             if not edges:
 651                 break
 652             for edge in edges:
 653                 if not isinstance(edge, dict):
 654                     continue
 655                 if edge.get('__typename') != self._EDGE_KIND:
 656                     continue
 657                 node = edge.get('node')
 658                 if not isinstance(node, dict):
 659                     continue
 660                 if node.get('__typename') != self._NODE_KIND:
 661                     continue
 662                 entry = self._extract_entry(node)
 663                 if entry:
 664                     cursor = edge.get('cursor')
 665                     yield entry
 666             if not cursor or not isinstance(cursor, compat_str):
 667                 break
 668
 669
 670 class TwitchVideosIE(TwitchPlaylistBaseIE):
 671     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:videos|profile)'
 672
 673     _TESTS = [{
 674         # All Videos sorted by Date
 675         'url': 'https://www.twitch.tv/spamfish/videos?filter=all',
 676         'info_dict': {
 677             'id': 'spamfish',
 678             'title': 'spamfish - All Videos sorted by Date',
 679         },
 680         'playlist_mincount': 924,
 681     }, {
 682         # All Videos sorted by Popular
 683         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views',
 684         'info_dict': {
 685             'id': 'spamfish',
 686             'title': 'spamfish - All Videos sorted by Popular',
 687         },
 688         'playlist_mincount': 931,
 689     }, {
 690         # Past Broadcasts sorted by Date
 691         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives',
 692         'info_dict': {
 693             'id': 'spamfish',
 694             'title': 'spamfish - Past Broadcasts sorted by Date',
 695         },
 696         'playlist_mincount': 27,
 697     }, {
 698         # Highlights sorted by Date
 699         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights',
 700         'info_dict': {
 701             'id': 'spamfish',
 702             'title': 'spamfish - Highlights sorted by Date',
 703         },
 704         'playlist_mincount': 901,
 705     }, {
 706         # Uploads sorted by Date
 707         'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time',
 708         'info_dict': {
 709             'id': 'esl_csgo',
 710             'title': 'esl_csgo - Uploads sorted by Date',
 711         },
 712         'playlist_mincount': 5,
 713     }, {
 714         # Past Premieres sorted by Date
 715         'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres',
 716         'info_dict': {
 717             'id': 'spamfish',
 718             'title': 'spamfish - Past Premieres sorted by Date',
 719         },
 720         'playlist_mincount': 1,
 721     }, {
 722         'url': 'https://www.twitch.tv/spamfish/videos/all',
 723         'only_matching': True,
 724     }, {
 725         'url': 'https://m.twitch.tv/spamfish/videos/all',
 726         'only_matching': True,
 727     }, {
 728         'url': 'https://www.twitch.tv/spamfish/videos',
 729         'only_matching': True,
 730     }]
 731
 732     Broadcast = collections.namedtuple('Broadcast', ['type', 'label'])
 733
 734     _DEFAULT_BROADCAST = Broadcast(None, 'All Videos')
 735     _BROADCASTS = {
 736         'archives': Broadcast('ARCHIVE', 'Past Broadcasts'),
 737         'highlights': Broadcast('HIGHLIGHT', 'Highlights'),
 738         'uploads': Broadcast('UPLOAD', 'Uploads'),
 739         'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'),
 740         'all': _DEFAULT_BROADCAST,
 741     }
 742
 743     _DEFAULT_SORTED_BY = 'Date'
 744     _SORTED_BY = {
 745         'time': _DEFAULT_SORTED_BY,
 746         'views': 'Popular',
 747     }
 748
 749     _OPERATION_NAME = 'FilterableVideoTower_Videos'
 750     _ENTRY_KIND = 'video'
 751     _EDGE_KIND = 'VideoEdge'
 752     _NODE_KIND = 'Video'
 753
 754     @classmethod
 755     def suitable(cls, url):
 756         return (False
 757                 if any(ie.suitable(url) for ie in (
 758                     TwitchVideosClipsIE,
 759                     TwitchVideosCollectionsIE))
 760                 else super(TwitchVideosIE, cls).suitable(url))
 761
 762     @staticmethod
 763     def _make_variables(channel_name, broadcast_type, sort):
 764         return {
 765             'channelOwnerLogin': channel_name,
 766             'broadcastType': broadcast_type,
 767             'videoSort': sort.upper(),
 768         }
 769
 770     @staticmethod
 771     def _extract_entry(node):
 772         return _make_video_result(node)
 773
 774     def _real_extract(self, url):
 775         channel_name = self._match_id(url)
 776         qs = parse_qs(url)
 777         filter = qs.get('filter', ['all'])[0]
 778         sort = qs.get('sort', ['time'])[0]
 779         broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST)
 780         return self.playlist_result(
 781             self._entries(channel_name, broadcast.type, sort),
 782             playlist_id=channel_name,
 783             playlist_title='%s - %s sorted by %s'
 784             % (channel_name, broadcast.label,
 785                self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY)))
 786
 787
 788 class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
 789     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)'
 790
 791     _TESTS = [{
 792         # Clips
 793         'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all',
 794         'info_dict': {
 795             'id': 'vanillatv',
 796             'title': 'vanillatv - Clips Top All',
 797         },
 798         'playlist_mincount': 1,
 799     }, {
 800         'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d',
 801         'only_matching': True,
 802     }]
 803
 804     Clip = collections.namedtuple('Clip', ['filter', 'label'])
 805
 806     _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D')
 807     _RANGE = {
 808         '24hr': Clip('LAST_DAY', 'Top 24H'),
 809         '7d': _DEFAULT_CLIP,
 810         '30d': Clip('LAST_MONTH', 'Top 30D'),
 811         'all': Clip('ALL_TIME', 'Top All'),
 812     }
 813
 814     # NB: values other than 20 result in skipped videos
 815     _PAGE_LIMIT = 20
 816
 817     _OPERATION_NAME = 'ClipsCards__User'
 818     _ENTRY_KIND = 'clip'
 819     _EDGE_KIND = 'ClipEdge'
 820     _NODE_KIND = 'Clip'
 821
 822     @staticmethod
 823     def _make_variables(channel_name, filter):
 824         return {
 825             'login': channel_name,
 826             'criteria': {
 827                 'filter': filter,
 828             },
 829         }
 830
 831     @staticmethod
 832     def _extract_entry(node):
 833         assert isinstance(node, dict)
 834         clip_url = url_or_none(node.get('url'))
 835         if not clip_url:
 836             return
 837         return {
 838             '_type': 'url_transparent',
 839             'ie_key': TwitchClipsIE.ie_key(),
 840             'id': node.get('id'),
 841             'url': clip_url,
 842             'title': node.get('title'),
 843             'thumbnail': node.get('thumbnailURL'),
 844             'duration': float_or_none(node.get('durationSeconds')),
 845             'timestamp': unified_timestamp(node.get('createdAt')),
 846             'view_count': int_or_none(node.get('viewCount')),
 847             'language': node.get('language'),
 848         }
 849
 850     def _real_extract(self, url):
 851         channel_name = self._match_id(url)
 852         qs = parse_qs(url)
 853         range = qs.get('range', ['7d'])[0]
 854         clip = self._RANGE.get(range, self._DEFAULT_CLIP)
 855         return self.playlist_result(
 856             self._entries(channel_name, clip.filter),
 857             playlist_id=channel_name,
 858             playlist_title='%s - Clips %s' % (channel_name, clip.label))
 859
 860
 861 class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
 862     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/videos/*?\?.*?\bfilter=collections'
 863
 864     _TESTS = [{
 865         # Collections
 866         'url': 'https://www.twitch.tv/spamfish/videos?filter=collections',
 867         'info_dict': {
 868             'id': 'spamfish',
 869             'title': 'spamfish - Collections',
 870         },
 871         'playlist_mincount': 3,
 872     }, {
 873         'url': 'https://www.twitch.tv/monstercat/videos?filter=collections',
 874         'info_dict': {
 875             'id': 'monstercat',
 876             'title': 'monstercat - Collections',
 877         },
 878         'playlist_mincount': 13,
 879     }]
 880
 881     _OPERATION_NAME = 'ChannelCollectionsContent'
 882     _ENTRY_KIND = 'collection'
 883     _EDGE_KIND = 'CollectionsItemEdge'
 884     _NODE_KIND = 'Collection'
 885
 886     @staticmethod
 887     def _make_variables(channel_name):
 888         return {
 889             'ownerLogin': channel_name,
 890         }
 891
 892     @staticmethod
 893     def _extract_entry(node):
 894         assert isinstance(node, dict)
 895         collection_id = node.get('id')
 896         if not collection_id:
 897             return
 898         return {
 899             '_type': 'url_transparent',
 900             'ie_key': TwitchCollectionIE.ie_key(),
 901             'id': collection_id,
 902             'url': 'https://www.twitch.tv/collections/%s' % collection_id,
 903             'title': node.get('title'),
 904             'thumbnail': node.get('thumbnailURL'),
 905             'duration': float_or_none(node.get('lengthSeconds')),
 906             'timestamp': unified_timestamp(node.get('updatedAt')),
 907             'view_count': int_or_none(node.get('viewCount')),
 908         }
 909
 910     def _real_extract(self, url):
 911         channel_name = self._match_id(url)
 912         return self.playlist_result(
 913             self._entries(channel_name), playlist_id=channel_name,
 914             playlist_title='%s - Collections' % channel_name)
 915
 916
 917 class TwitchStreamIE(TwitchBaseIE):
 918     IE_NAME = 'twitch:stream'
 919     _VALID_URL = r'''(?x)
 920                     https?://
 921                         (?:
 922                             (?:(?:www|go|m)\.)?twitch\.tv/|
 923                             player\.twitch\.tv/\?.*?\bchannel=
 924                         )
 925                         (?P<id>[^/#?]+)
 926                     '''
 927
 928     _TESTS = [{
 929         'url': 'http://www.twitch.tv/shroomztv',
 930         'info_dict': {
 931             'id': '12772022048',
 932             'display_id': 'shroomztv',
 933             'ext': 'mp4',
 934             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 935             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 936             'is_live': True,
 937             'timestamp': 1421928037,
 938             'upload_date': '20150122',
 939             'uploader': 'ShroomzTV',
 940             'uploader_id': 'shroomztv',
 941             'view_count': int,
 942         },
 943         'params': {
 944             # m3u8 download
 945             'skip_download': True,
 946         },
 947         'skip': 'User does not exist',
 948     }, {
 949         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 950         'only_matching': True,
 951     }, {
 952         'url': 'https://player.twitch.tv/?channel=lotsofs',
 953         'only_matching': True,
 954     }, {
 955         'url': 'https://go.twitch.tv/food',
 956         'only_matching': True,
 957     }, {
 958         'url': 'https://m.twitch.tv/food',
 959         'only_matching': True,
 960     }, {
 961         'url': 'https://www.twitch.tv/monstercat',
 962         'info_dict': {
 963             'id': '40500071752',
 964             'display_id': 'monstercat',
 965             'title': 're:Monstercat',
 966             'description': 'md5:0945ad625e615bc8f0469396537d87d9',
 967             'is_live': True,
 968             'timestamp': 1677107190,
 969             'upload_date': '20230222',
 970             'uploader': 'Monstercat',
 971             'uploader_id': 'monstercat',
 972             'live_status': 'is_live',
 973             'thumbnail': 're:https://.*.jpg',
 974             'ext': 'mp4',
 975         },
 976         'params': {
 977             'skip_download': 'Livestream',
 978         },
 979     }]
 980
 981     @classmethod
 982     def suitable(cls, url):
 983         return (False
 984                 if any(ie.suitable(url) for ie in (
 985                     TwitchVodIE,
 986                     TwitchCollectionIE,
 987                     TwitchVideosIE,
 988                     TwitchVideosClipsIE,
 989                     TwitchVideosCollectionsIE,
 990                     TwitchClipsIE))
 991                 else super(TwitchStreamIE, cls).suitable(url))
 992
 993     def _real_extract(self, url):
 994         channel_name = self._match_id(url).lower()
 995
 996         gql = self._download_gql(
 997             channel_name, [{
 998                 'operationName': 'StreamMetadata',
 999                 'variables': {'channelLogin': channel_name},
1000             }, {
1001                 'operationName': 'ComscoreStreamingQuery',
1002                 'variables': {
1003                     'channel': channel_name,
1004                     'clipSlug': '',
1005                     'isClip': False,
1006                     'isLive': True,
1007                     'isVodOrCollection': False,
1008                     'vodID': '',
1009                 },
1010             }, {
1011                 'operationName': 'VideoPreviewOverlay',
1012                 'variables': {'login': channel_name},
1013             }],
1014             'Downloading stream GraphQL')
1015
1016         user = gql[0]['data']['user']
1017
1018         if not user:
1019             raise ExtractorError(
1020                 '%s does not exist' % channel_name, expected=True)
1021
1022         stream = user['stream']
1023
1024         if not stream:
1025             raise UserNotLive(video_id=channel_name)
1026
1027         access_token = self._download_access_token(
1028             channel_name, 'stream', 'channelName')
1029         token = access_token['value']
1030
1031         stream_id = stream.get('id') or channel_name
1032         query = {
1033             'allow_source': 'true',
1034             'allow_audio_only': 'true',
1035             'allow_spectre': 'true',
1036             'p': random.randint(1000000, 10000000),
1037             'player': 'twitchweb',
1038             'playlist_include_framerate': 'true',
1039             'segment_preference': '4',
1040             'sig': access_token['signature'].encode('utf-8'),
1041             'token': token.encode('utf-8'),
1042         }
1043         formats = self._extract_m3u8_formats(
1044             '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
1045             stream_id, 'mp4', query=query)
1046         self._prefer_source(formats)
1047
1048         view_count = stream.get('viewers')
1049         timestamp = unified_timestamp(stream.get('createdAt'))
1050
1051         sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
1052         uploader = sq_user.get('displayName')
1053         description = try_get(
1054             sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
1055
1056         thumbnail = url_or_none(try_get(
1057             gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
1058             compat_str))
1059
1060         title = uploader or channel_name
1061         stream_type = stream.get('type')
1062         if stream_type in ['rerun', 'live']:
1063             title += ' (%s)' % stream_type
1064
1065         return {
1066             'id': stream_id,
1067             'display_id': channel_name,
1068             'title': title,
1069             'description': description,
1070             'thumbnails': self._get_thumbnails(thumbnail),
1071             'uploader': uploader,
1072             'uploader_id': channel_name,
1073             'timestamp': timestamp,
1074             'view_count': view_count,
1075             'formats': formats,
1076             'is_live': stream_type == 'live',
1077         }
1078
1079
1080 class TwitchClipsIE(TwitchBaseIE):
1081     IE_NAME = 'twitch:clips'
1082     _VALID_URL = r'''(?x)
1083                     https?://
1084                         (?:
1085                             clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)|
1086                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/)?clip/
1087                         )
1088                         (?P<id>[^/?#&]+)
1089                     '''
1090
1091     _TESTS = [{
1092         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
1093         'md5': '761769e1eafce0ffebfb4089cb3847cd',
1094         'info_dict': {
1095             'id': '42850523',
1096             'display_id': 'FaintLightGullWholeWheat',
1097             'ext': 'mp4',
1098             'title': 'EA Play 2016 Live from the Novo Theatre',
1099             'thumbnail': r're:^https?://.*\.jpg',
1100             'timestamp': 1465767393,
1101             'upload_date': '20160612',
1102             'creator': 'EA',
1103             'uploader': 'stereotype_',
1104             'uploader_id': '43566419',
1105         },
1106     }, {
1107         # multiple formats
1108         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
1109         'only_matching': True,
1110     }, {
1111         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
1112         'only_matching': True,
1113     }, {
1114         'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited',
1115         'only_matching': True,
1116     }, {
1117         'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1118         'only_matching': True,
1119     }, {
1120         'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1121         'only_matching': True,
1122     }, {
1123         'url': 'https://m.twitch.tv/clip/FaintLightGullWholeWheat',
1124         'only_matching': True,
1125     }]
1126
1127     def _real_extract(self, url):
1128         video_id = self._match_id(url)
1129
1130         clip = self._download_gql(
1131             video_id, [{
1132                 'operationName': 'VideoAccessToken_Clip',
1133                 'variables': {
1134                     'slug': video_id,
1135                 },
1136             }],
1137             'Downloading clip access token GraphQL')[0]['data']['clip']
1138
1139         if not clip:
1140             raise ExtractorError(
1141                 'This clip is no longer available', expected=True)
1142
1143         access_query = {
1144             'sig': clip['playbackAccessToken']['signature'],
1145             'token': clip['playbackAccessToken']['value'],
1146         }
1147
1148         data = self._download_base_gql(
1149             video_id, {
1150                 'query': '''{
1151   clip(slug: "%s") {
1152     broadcaster {
1153       displayName
1154     }
1155     createdAt
1156     curator {
1157       displayName
1158       id
1159     }
1160     durationSeconds
1161     id
1162     tiny: thumbnailURL(width: 86, height: 45)
1163     small: thumbnailURL(width: 260, height: 147)
1164     medium: thumbnailURL(width: 480, height: 272)
1165     title
1166     videoQualities {
1167       frameRate
1168       quality
1169       sourceURL
1170     }
1171     viewCount
1172   }
1173 }''' % video_id}, 'Downloading clip GraphQL', fatal=False)
1174
1175         if data:
1176             clip = try_get(data, lambda x: x['data']['clip'], dict) or clip
1177
1178         formats = []
1179         for option in clip.get('videoQualities', []):
1180             if not isinstance(option, dict):
1181                 continue
1182             source = url_or_none(option.get('sourceURL'))
1183             if not source:
1184                 continue
1185             formats.append({
1186                 'url': update_url_query(source, access_query),
1187                 'format_id': option.get('quality'),
1188                 'height': int_or_none(option.get('quality')),
1189                 'fps': int_or_none(option.get('frameRate')),
1190             })
1191
1192         thumbnails = []
1193         for thumbnail_id in ('tiny', 'small', 'medium'):
1194             thumbnail_url = clip.get(thumbnail_id)
1195             if not thumbnail_url:
1196                 continue
1197             thumb = {
1198                 'id': thumbnail_id,
1199                 'url': thumbnail_url,
1200             }
1201             mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url)
1202             if mobj:
1203                 thumb.update({
1204                     'height': int(mobj.group(2)),
1205                     'width': int(mobj.group(1)),
1206                 })
1207             thumbnails.append(thumb)
1208
1209         old_id = self._search_regex(r'%7C(\d+)(?:-\d+)?.mp4', formats[-1]['url'], 'old id', default=None)
1210
1211         return {
1212             'id': clip.get('id') or video_id,
1213             '_old_archive_ids': [make_archive_id(self, old_id)] if old_id else None,
1214             'display_id': video_id,
1215             'title': clip.get('title'),
1216             'formats': formats,
1217             'duration': int_or_none(clip.get('durationSeconds')),
1218             'view_count': int_or_none(clip.get('viewCount')),
1219             'timestamp': unified_timestamp(clip.get('createdAt')),
1220             'thumbnails': thumbnails,
1221             'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str),
1222             'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str),
1223             'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
1224         }