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': '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84',
  52         'StreamMetadata': '1c719a40e481453e5c48d9bb585d971b8b372f8ebb105b17076722264dfa5b3e',
  53         'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
  54         'VideoAccessToken_Clip': '36b89d2507fce29e5ca551df756d27c1cfe079e2609642b4390aa4c35796eb11',
  55         'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
  56         'VideoMetadata': '226edb3e692509f727fd56821f5653c05740242c82b0388883e0c0e75dcbf687',
  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         self._sort_formats(formats)
 137
 138     def _download_base_gql(self, video_id, ops, note, fatal=True):
 139         headers = {
 140             'Content-Type': 'text/plain;charset=UTF-8',
 141             'Client-ID': self._CLIENT_ID,
 142         }
 143         gql_auth = self._get_cookies('https://gql.twitch.tv').get('auth-token')
 144         if gql_auth:
 145             headers['Authorization'] = 'OAuth ' + gql_auth.value
 146         return self._download_json(
 147             'https://gql.twitch.tv/gql', video_id, note,
 148             data=json.dumps(ops).encode(),
 149             headers=headers, fatal=fatal)
 150
 151     def _download_gql(self, video_id, ops, note, fatal=True):
 152         for op in ops:
 153             op['extensions'] = {
 154                 'persistedQuery': {
 155                     'version': 1,
 156                     'sha256Hash': self._OPERATION_HASHES[op['operationName']],
 157                 }
 158             }
 159         return self._download_base_gql(video_id, ops, note)
 160
 161     def _download_access_token(self, video_id, token_kind, param_name):
 162         method = '%sPlaybackAccessToken' % token_kind
 163         ops = {
 164             'query': '''{
 165               %s(
 166                 %s: "%s",
 167                 params: {
 168                   platform: "web",
 169                   playerBackend: "mediaplayer",
 170                   playerType: "site"
 171                 }
 172               )
 173               {
 174                 value
 175                 signature
 176               }
 177             }''' % (method, param_name, video_id),
 178         }
 179         return self._download_base_gql(
 180             video_id, ops,
 181             'Downloading %s access token GraphQL' % token_kind)['data'][method]
 182
 183
 184 class TwitchVodIE(TwitchBaseIE):
 185     IE_NAME = 'twitch:vod'
 186     _VALID_URL = r'''(?x)
 187                     https?://
 188                         (?:
 189                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 190                             player\.twitch\.tv/\?.*?\bvideo=v?
 191                         )
 192                         (?P<id>\d+)
 193                     '''
 194
 195     _TESTS = [{
 196         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 197         'info_dict': {
 198             'id': 'v6528877',
 199             'ext': 'mp4',
 200             'title': 'LCK Summer Split - Week 6 Day 1',
 201             'thumbnail': r're:^https?://.*\.jpg$',
 202             'duration': 17208,
 203             'timestamp': 1435131734,
 204             'upload_date': '20150624',
 205             'uploader': 'Riot Games',
 206             'uploader_id': 'riotgames',
 207             'view_count': int,
 208             'start_time': 310,
 209             'chapters': [
 210                 {
 211                     'start_time': 0,
 212                     'end_time': 17208,
 213                     'title': 'League of Legends'
 214                 }
 215             ],
 216             'live_status': 'was_live',
 217         },
 218         'params': {
 219             # m3u8 download
 220             'skip_download': True,
 221         },
 222     }, {
 223         # Untitled broadcast (title is None)
 224         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 225         'info_dict': {
 226             'id': 'v11230755',
 227             'ext': 'mp4',
 228             'title': 'Untitled Broadcast',
 229             'thumbnail': r're:^https?://.*\.jpg$',
 230             'duration': 1638,
 231             'timestamp': 1439746708,
 232             'upload_date': '20150816',
 233             'uploader': 'BelkAO_o',
 234             'uploader_id': 'belkao_o',
 235             'view_count': int,
 236         },
 237         'params': {
 238             # m3u8 download
 239             'skip_download': True,
 240         },
 241         'skip': 'HTTP Error 404: Not Found',
 242     }, {
 243         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 244         'only_matching': True,
 245     }, {
 246         'url': 'https://www.twitch.tv/videos/6528877',
 247         'only_matching': True,
 248     }, {
 249         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 250         'only_matching': True,
 251     }, {
 252         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 253         'only_matching': True,
 254     }, {
 255         'url': 'https://player.twitch.tv/?video=480452374',
 256         'only_matching': True,
 257     }, {
 258         'url': 'https://www.twitch.tv/videos/635475444',
 259         'info_dict': {
 260             'id': 'v635475444',
 261             'ext': 'mp4',
 262             'title': 'Riot Games',
 263             'duration': 11643,
 264             'uploader': 'Riot Games',
 265             'uploader_id': 'riotgames',
 266             'timestamp': 1590770569,
 267             'upload_date': '20200529',
 268             'chapters': [
 269                 {
 270                     'start_time': 0,
 271                     'end_time': 573,
 272                     'title': 'League of Legends'
 273                 },
 274                 {
 275                     'start_time': 573,
 276                     'end_time': 3922,
 277                     'title': 'Legends of Runeterra'
 278                 },
 279                 {
 280                     'start_time': 3922,
 281                     'end_time': 11643,
 282                     'title': 'Art'
 283                 }
 284             ],
 285             'live_status': 'was_live',
 286             'thumbnail': r're:^https?://.*\.jpg$',
 287             'view_count': int,
 288         },
 289         'params': {
 290             'skip_download': True
 291         },
 292     }, {
 293         'note': 'Storyboards',
 294         'url': 'https://www.twitch.tv/videos/635475444',
 295         'info_dict': {
 296             'id': 'v635475444',
 297             'format_id': 'sb0',
 298             'ext': 'mhtml',
 299             'title': 'Riot Games',
 300             'duration': 11643,
 301             'uploader': 'Riot Games',
 302             'uploader_id': 'riotgames',
 303             'timestamp': 1590770569,
 304             'upload_date': '20200529',
 305             'chapters': [
 306                 {
 307                     'start_time': 0,
 308                     'end_time': 573,
 309                     'title': 'League of Legends'
 310                 },
 311                 {
 312                     'start_time': 573,
 313                     'end_time': 3922,
 314                     'title': 'Legends of Runeterra'
 315                 },
 316                 {
 317                     'start_time': 3922,
 318                     'end_time': 11643,
 319                     'title': 'Art'
 320                 }
 321             ],
 322             'live_status': 'was_live',
 323             'thumbnail': r're:^https?://.*\.jpg$',
 324             'view_count': int,
 325             'columns': int,
 326             'rows': int,
 327         },
 328         'params': {
 329             'format': 'mhtml',
 330             'skip_download': True
 331         }
 332     }, {
 333         'note': 'VOD with single chapter',
 334         'url': 'https://www.twitch.tv/videos/1536751224',
 335         'info_dict': {
 336             'id': 'v1536751224',
 337             'ext': 'mp4',
 338             'title': 'Porter Robinson Star Guardian Stream Tour with LilyPichu',
 339             'duration': 8353,
 340             'uploader': 'Riot Games',
 341             'uploader_id': 'riotgames',
 342             'timestamp': 1658267731,
 343             'upload_date': '20220719',
 344             'chapters': [
 345                 {
 346                     'start_time': 0,
 347                     'end_time': 8353,
 348                     'title': 'League of Legends'
 349                 }
 350             ],
 351             'live_status': 'was_live',
 352             'thumbnail': r're:^https?://.*\.jpg$',
 353             'view_count': int,
 354         },
 355         'params': {
 356             'skip_download': True
 357         },
 358         'expected_warnings': ['Unable to download JSON metadata: HTTP Error 403: Forbidden']
 359     }]
 360
 361     def _download_info(self, item_id):
 362         data = self._download_gql(
 363             item_id, [{
 364                 'operationName': 'VideoMetadata',
 365                 'variables': {
 366                     'channelLogin': '',
 367                     'videoID': item_id,
 368                 },
 369             }, {
 370                 'operationName': 'VideoPlayer_ChapterSelectButtonVideo',
 371                 'variables': {
 372                     'includePrivate': False,
 373                     'videoID': item_id,
 374                 },
 375             }, {
 376                 'operationName': 'VideoPlayer_VODSeekbarPreviewVideo',
 377                 'variables': {
 378                     'includePrivate': False,
 379                     'videoID': item_id,
 380                 },
 381             }],
 382             'Downloading stream metadata GraphQL')
 383
 384         video = traverse_obj(data, (0, 'data', 'video'))
 385         video['moments'] = traverse_obj(data, (1, 'data', 'video', 'moments', 'edges', ..., 'node'))
 386         video['storyboard'] = traverse_obj(data, (2, 'data', 'video', 'seekPreviewsURL'), expected_type=url_or_none)
 387
 388         if video is None:
 389             raise ExtractorError(
 390                 'Video %s does not exist' % item_id, expected=True)
 391         return video
 392
 393     def _extract_info(self, info):
 394         status = info.get('status')
 395         if status == 'recording':
 396             is_live = True
 397         elif status == 'recorded':
 398             is_live = False
 399         else:
 400             is_live = None
 401         _QUALITIES = ('small', 'medium', 'large')
 402         quality_key = qualities(_QUALITIES)
 403         thumbnails = []
 404         preview = info.get('preview')
 405         if isinstance(preview, dict):
 406             for thumbnail_id, thumbnail_url in preview.items():
 407                 thumbnail_url = url_or_none(thumbnail_url)
 408                 if not thumbnail_url:
 409                     continue
 410                 if thumbnail_id not in _QUALITIES:
 411                     continue
 412                 thumbnails.append({
 413                     'url': thumbnail_url,
 414                     'preference': quality_key(thumbnail_id),
 415                 })
 416         return {
 417             'id': info['_id'],
 418             'title': info.get('title') or 'Untitled Broadcast',
 419             'description': info.get('description'),
 420             'duration': int_or_none(info.get('length')),
 421             'thumbnails': thumbnails,
 422             'uploader': info.get('channel', {}).get('display_name'),
 423             'uploader_id': info.get('channel', {}).get('name'),
 424             'timestamp': parse_iso8601(info.get('recorded_at')),
 425             'view_count': int_or_none(info.get('views')),
 426             'is_live': is_live,
 427             'was_live': True,
 428         }
 429
 430     def _extract_chapters(self, info, item_id):
 431         if not info.get('moments'):
 432             game = traverse_obj(info, ('game', 'displayName'))
 433             if game:
 434                 yield {'title': game}
 435             return
 436
 437         for moment in info['moments']:
 438             start_time = int_or_none(moment.get('positionMilliseconds'), 1000)
 439             duration = int_or_none(moment.get('durationMilliseconds'), 1000)
 440             name = str_or_none(moment.get('description'))
 441
 442             if start_time is None or duration is None:
 443                 self.report_warning(f'Important chapter information missing for chapter {name}', item_id)
 444                 continue
 445             yield {
 446                 'start_time': start_time,
 447                 'end_time': start_time + duration,
 448                 'title': name,
 449             }
 450
 451     def _extract_info_gql(self, info, item_id):
 452         vod_id = info.get('id') or item_id
 453         # id backward compatibility for download archives
 454         if vod_id[0] != 'v':
 455             vod_id = 'v%s' % vod_id
 456         thumbnail = url_or_none(info.get('previewThumbnailURL'))
 457         is_live = None
 458         if thumbnail:
 459             if thumbnail.endswith('/404_processing_{width}x{height}.png'):
 460                 is_live, thumbnail = True, None
 461             else:
 462                 is_live = False
 463                 for p in ('width', 'height'):
 464                     thumbnail = thumbnail.replace('{%s}' % p, '0')
 465
 466         return {
 467             'id': vod_id,
 468             'title': info.get('title') or 'Untitled Broadcast',
 469             'description': info.get('description'),
 470             'duration': int_or_none(info.get('lengthSeconds')),
 471             'thumbnail': thumbnail,
 472             'uploader': try_get(info, lambda x: x['owner']['displayName'], compat_str),
 473             'uploader_id': try_get(info, lambda x: x['owner']['login'], compat_str),
 474             'timestamp': unified_timestamp(info.get('publishedAt')),
 475             'view_count': int_or_none(info.get('viewCount')),
 476             'chapters': list(self._extract_chapters(info, item_id)),
 477             'is_live': is_live,
 478             'was_live': True,
 479         }
 480
 481     def _extract_storyboard(self, item_id, storyboard_json_url, duration):
 482         if not duration or not storyboard_json_url:
 483             return
 484         spec = self._download_json(storyboard_json_url, item_id, 'Downloading storyboard metadata JSON', fatal=False) or []
 485         # sort from highest quality to lowest
 486         # This makes sb0 the highest-quality format, sb1 - lower, etc which is consistent with youtube sb ordering
 487         spec.sort(key=lambda x: int_or_none(x.get('width')) or 0, reverse=True)
 488         base = base_url(storyboard_json_url)
 489         for i, s in enumerate(spec):
 490             count = int_or_none(s.get('count'))
 491             images = s.get('images')
 492             if not (images and count):
 493                 continue
 494             fragment_duration = duration / len(images)
 495             yield {
 496                 'format_id': f'sb{i}',
 497                 'format_note': 'storyboard',
 498                 'ext': 'mhtml',
 499                 'protocol': 'mhtml',
 500                 'acodec': 'none',
 501                 'vcodec': 'none',
 502                 'url': urljoin(base, images[0]),
 503                 'width': int_or_none(s.get('width')),
 504                 'height': int_or_none(s.get('height')),
 505                 'fps': count / duration,
 506                 'rows': int_or_none(s.get('rows')),
 507                 'columns': int_or_none(s.get('cols')),
 508                 'fragments': [{
 509                     'url': urljoin(base, path),
 510                     'duration': fragment_duration,
 511                 } for path in images],
 512             }
 513
 514     def _real_extract(self, url):
 515         vod_id = self._match_id(url)
 516
 517         video = self._download_info(vod_id)
 518         info = self._extract_info_gql(video, vod_id)
 519         access_token = self._download_access_token(vod_id, 'video', 'id')
 520
 521         formats = self._extract_m3u8_formats(
 522             '%s/vod/%s.m3u8?%s' % (
 523                 self._USHER_BASE, vod_id,
 524                 compat_urllib_parse_urlencode({
 525                     'allow_source': 'true',
 526                     'allow_audio_only': 'true',
 527                     'allow_spectre': 'true',
 528                     'player': 'twitchweb',
 529                     'playlist_include_framerate': 'true',
 530                     'nauth': access_token['value'],
 531                     'nauthsig': access_token['signature'],
 532                 })),
 533             vod_id, 'mp4', entry_protocol='m3u8_native')
 534
 535         formats.extend(self._extract_storyboard(vod_id, video.get('storyboard'), info.get('duration')))
 536
 537         self._prefer_source(formats)
 538         info['formats'] = formats
 539
 540         parsed_url = compat_urllib_parse_urlparse(url)
 541         query = compat_parse_qs(parsed_url.query)
 542         if 't' in query:
 543             info['start_time'] = parse_duration(query['t'][0])
 544
 545         if info.get('timestamp') is not None:
 546             info['subtitles'] = {
 547                 'rechat': [{
 548                     'url': update_url_query(
 549                         'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, {
 550                             'client_id': self._CLIENT_ID,
 551                         }),
 552                     'ext': 'json',
 553                 }],
 554             }
 555
 556         return info
 557
 558
 559 def _make_video_result(node):
 560     assert isinstance(node, dict)
 561     video_id = node.get('id')
 562     if not video_id:
 563         return
 564     return {
 565         '_type': 'url_transparent',
 566         'ie_key': TwitchVodIE.ie_key(),
 567         'id': 'v' + video_id,
 568         'url': 'https://www.twitch.tv/videos/%s' % video_id,
 569         'title': node.get('title'),
 570         'thumbnail': node.get('previewThumbnailURL'),
 571         'duration': float_or_none(node.get('lengthSeconds')),
 572         'view_count': int_or_none(node.get('viewCount')),
 573     }
 574
 575
 576 class TwitchCollectionIE(TwitchBaseIE):
 577     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P<id>[^/]+)'
 578
 579     _TESTS = [{
 580         'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ',
 581         'info_dict': {
 582             'id': 'wlDCoH0zEBZZbQ',
 583             'title': 'Overthrow Nook, capitalism for children',
 584         },
 585         'playlist_mincount': 13,
 586     }]
 587
 588     _OPERATION_NAME = 'CollectionSideBar'
 589
 590     def _real_extract(self, url):
 591         collection_id = self._match_id(url)
 592         collection = self._download_gql(
 593             collection_id, [{
 594                 'operationName': self._OPERATION_NAME,
 595                 'variables': {'collectionID': collection_id},
 596             }],
 597             'Downloading collection GraphQL')[0]['data']['collection']
 598         title = collection.get('title')
 599         entries = []
 600         for edge in collection['items']['edges']:
 601             if not isinstance(edge, dict):
 602                 continue
 603             node = edge.get('node')
 604             if not isinstance(node, dict):
 605                 continue
 606             video = _make_video_result(node)
 607             if video:
 608                 entries.append(video)
 609         return self.playlist_result(
 610             entries, playlist_id=collection_id, playlist_title=title)
 611
 612
 613 class TwitchPlaylistBaseIE(TwitchBaseIE):
 614     _PAGE_LIMIT = 100
 615
 616     def _entries(self, channel_name, *args):
 617         cursor = None
 618         variables_common = self._make_variables(channel_name, *args)
 619         entries_key = '%ss' % self._ENTRY_KIND
 620         for page_num in itertools.count(1):
 621             variables = variables_common.copy()
 622             variables['limit'] = self._PAGE_LIMIT
 623             if cursor:
 624                 variables['cursor'] = cursor
 625             page = self._download_gql(
 626                 channel_name, [{
 627                     'operationName': self._OPERATION_NAME,
 628                     'variables': variables,
 629                 }],
 630                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
 631                 fatal=False)
 632             if not page:
 633                 break
 634             edges = try_get(
 635                 page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
 636             if not edges:
 637                 break
 638             for edge in edges:
 639                 if not isinstance(edge, dict):
 640                     continue
 641                 if edge.get('__typename') != self._EDGE_KIND:
 642                     continue
 643                 node = edge.get('node')
 644                 if not isinstance(node, dict):
 645                     continue
 646                 if node.get('__typename') != self._NODE_KIND:
 647                     continue
 648                 entry = self._extract_entry(node)
 649                 if entry:
 650                     cursor = edge.get('cursor')
 651                     yield entry
 652             if not cursor or not isinstance(cursor, compat_str):
 653                 break
 654
 655
 656 class TwitchVideosIE(TwitchPlaylistBaseIE):
 657     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:videos|profile)'
 658
 659     _TESTS = [{
 660         # All Videos sorted by Date
 661         'url': 'https://www.twitch.tv/spamfish/videos?filter=all',
 662         'info_dict': {
 663             'id': 'spamfish',
 664             'title': 'spamfish - All Videos sorted by Date',
 665         },
 666         'playlist_mincount': 924,
 667     }, {
 668         # All Videos sorted by Popular
 669         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views',
 670         'info_dict': {
 671             'id': 'spamfish',
 672             'title': 'spamfish - All Videos sorted by Popular',
 673         },
 674         'playlist_mincount': 931,
 675     }, {
 676         # Past Broadcasts sorted by Date
 677         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives',
 678         'info_dict': {
 679             'id': 'spamfish',
 680             'title': 'spamfish - Past Broadcasts sorted by Date',
 681         },
 682         'playlist_mincount': 27,
 683     }, {
 684         # Highlights sorted by Date
 685         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights',
 686         'info_dict': {
 687             'id': 'spamfish',
 688             'title': 'spamfish - Highlights sorted by Date',
 689         },
 690         'playlist_mincount': 901,
 691     }, {
 692         # Uploads sorted by Date
 693         'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time',
 694         'info_dict': {
 695             'id': 'esl_csgo',
 696             'title': 'esl_csgo - Uploads sorted by Date',
 697         },
 698         'playlist_mincount': 5,
 699     }, {
 700         # Past Premieres sorted by Date
 701         'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres',
 702         'info_dict': {
 703             'id': 'spamfish',
 704             'title': 'spamfish - Past Premieres sorted by Date',
 705         },
 706         'playlist_mincount': 1,
 707     }, {
 708         'url': 'https://www.twitch.tv/spamfish/videos/all',
 709         'only_matching': True,
 710     }, {
 711         'url': 'https://m.twitch.tv/spamfish/videos/all',
 712         'only_matching': True,
 713     }, {
 714         'url': 'https://www.twitch.tv/spamfish/videos',
 715         'only_matching': True,
 716     }]
 717
 718     Broadcast = collections.namedtuple('Broadcast', ['type', 'label'])
 719
 720     _DEFAULT_BROADCAST = Broadcast(None, 'All Videos')
 721     _BROADCASTS = {
 722         'archives': Broadcast('ARCHIVE', 'Past Broadcasts'),
 723         'highlights': Broadcast('HIGHLIGHT', 'Highlights'),
 724         'uploads': Broadcast('UPLOAD', 'Uploads'),
 725         'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'),
 726         'all': _DEFAULT_BROADCAST,
 727     }
 728
 729     _DEFAULT_SORTED_BY = 'Date'
 730     _SORTED_BY = {
 731         'time': _DEFAULT_SORTED_BY,
 732         'views': 'Popular',
 733     }
 734
 735     _OPERATION_NAME = 'FilterableVideoTower_Videos'
 736     _ENTRY_KIND = 'video'
 737     _EDGE_KIND = 'VideoEdge'
 738     _NODE_KIND = 'Video'
 739
 740     @classmethod
 741     def suitable(cls, url):
 742         return (False
 743                 if any(ie.suitable(url) for ie in (
 744                     TwitchVideosClipsIE,
 745                     TwitchVideosCollectionsIE))
 746                 else super(TwitchVideosIE, cls).suitable(url))
 747
 748     @staticmethod
 749     def _make_variables(channel_name, broadcast_type, sort):
 750         return {
 751             'channelOwnerLogin': channel_name,
 752             'broadcastType': broadcast_type,
 753             'videoSort': sort.upper(),
 754         }
 755
 756     @staticmethod
 757     def _extract_entry(node):
 758         return _make_video_result(node)
 759
 760     def _real_extract(self, url):
 761         channel_name = self._match_id(url)
 762         qs = parse_qs(url)
 763         filter = qs.get('filter', ['all'])[0]
 764         sort = qs.get('sort', ['time'])[0]
 765         broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST)
 766         return self.playlist_result(
 767             self._entries(channel_name, broadcast.type, sort),
 768             playlist_id=channel_name,
 769             playlist_title='%s - %s sorted by %s'
 770             % (channel_name, broadcast.label,
 771                self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY)))
 772
 773
 774 class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
 775     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)'
 776
 777     _TESTS = [{
 778         # Clips
 779         'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all',
 780         'info_dict': {
 781             'id': 'vanillatv',
 782             'title': 'vanillatv - Clips Top All',
 783         },
 784         'playlist_mincount': 1,
 785     }, {
 786         'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d',
 787         'only_matching': True,
 788     }]
 789
 790     Clip = collections.namedtuple('Clip', ['filter', 'label'])
 791
 792     _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D')
 793     _RANGE = {
 794         '24hr': Clip('LAST_DAY', 'Top 24H'),
 795         '7d': _DEFAULT_CLIP,
 796         '30d': Clip('LAST_MONTH', 'Top 30D'),
 797         'all': Clip('ALL_TIME', 'Top All'),
 798     }
 799
 800     # NB: values other than 20 result in skipped videos
 801     _PAGE_LIMIT = 20
 802
 803     _OPERATION_NAME = 'ClipsCards__User'
 804     _ENTRY_KIND = 'clip'
 805     _EDGE_KIND = 'ClipEdge'
 806     _NODE_KIND = 'Clip'
 807
 808     @staticmethod
 809     def _make_variables(channel_name, filter):
 810         return {
 811             'login': channel_name,
 812             'criteria': {
 813                 'filter': filter,
 814             },
 815         }
 816
 817     @staticmethod
 818     def _extract_entry(node):
 819         assert isinstance(node, dict)
 820         clip_url = url_or_none(node.get('url'))
 821         if not clip_url:
 822             return
 823         return {
 824             '_type': 'url_transparent',
 825             'ie_key': TwitchClipsIE.ie_key(),
 826             'id': node.get('id'),
 827             'url': clip_url,
 828             'title': node.get('title'),
 829             'thumbnail': node.get('thumbnailURL'),
 830             'duration': float_or_none(node.get('durationSeconds')),
 831             'timestamp': unified_timestamp(node.get('createdAt')),
 832             'view_count': int_or_none(node.get('viewCount')),
 833             'language': node.get('language'),
 834         }
 835
 836     def _real_extract(self, url):
 837         channel_name = self._match_id(url)
 838         qs = parse_qs(url)
 839         range = qs.get('range', ['7d'])[0]
 840         clip = self._RANGE.get(range, self._DEFAULT_CLIP)
 841         return self.playlist_result(
 842             self._entries(channel_name, clip.filter),
 843             playlist_id=channel_name,
 844             playlist_title='%s - Clips %s' % (channel_name, clip.label))
 845
 846
 847 class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
 848     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/videos/*?\?.*?\bfilter=collections'
 849
 850     _TESTS = [{
 851         # Collections
 852         'url': 'https://www.twitch.tv/spamfish/videos?filter=collections',
 853         'info_dict': {
 854             'id': 'spamfish',
 855             'title': 'spamfish - Collections',
 856         },
 857         'playlist_mincount': 3,
 858     }]
 859
 860     _OPERATION_NAME = 'ChannelCollectionsContent'
 861     _ENTRY_KIND = 'collection'
 862     _EDGE_KIND = 'CollectionsItemEdge'
 863     _NODE_KIND = 'Collection'
 864
 865     @staticmethod
 866     def _make_variables(channel_name):
 867         return {
 868             'ownerLogin': channel_name,
 869         }
 870
 871     @staticmethod
 872     def _extract_entry(node):
 873         assert isinstance(node, dict)
 874         collection_id = node.get('id')
 875         if not collection_id:
 876             return
 877         return {
 878             '_type': 'url_transparent',
 879             'ie_key': TwitchCollectionIE.ie_key(),
 880             'id': collection_id,
 881             'url': 'https://www.twitch.tv/collections/%s' % collection_id,
 882             'title': node.get('title'),
 883             'thumbnail': node.get('thumbnailURL'),
 884             'duration': float_or_none(node.get('lengthSeconds')),
 885             'timestamp': unified_timestamp(node.get('updatedAt')),
 886             'view_count': int_or_none(node.get('viewCount')),
 887         }
 888
 889     def _real_extract(self, url):
 890         channel_name = self._match_id(url)
 891         return self.playlist_result(
 892             self._entries(channel_name), playlist_id=channel_name,
 893             playlist_title='%s - Collections' % channel_name)
 894
 895
 896 class TwitchStreamIE(TwitchBaseIE):
 897     IE_NAME = 'twitch:stream'
 898     _VALID_URL = r'''(?x)
 899                     https?://
 900                         (?:
 901                             (?:(?:www|go|m)\.)?twitch\.tv/|
 902                             player\.twitch\.tv/\?.*?\bchannel=
 903                         )
 904                         (?P<id>[^/#?]+)
 905                     '''
 906
 907     _TESTS = [{
 908         'url': 'http://www.twitch.tv/shroomztv',
 909         'info_dict': {
 910             'id': '12772022048',
 911             'display_id': 'shroomztv',
 912             'ext': 'mp4',
 913             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 914             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 915             'is_live': True,
 916             'timestamp': 1421928037,
 917             'upload_date': '20150122',
 918             'uploader': 'ShroomzTV',
 919             'uploader_id': 'shroomztv',
 920             'view_count': int,
 921         },
 922         'params': {
 923             # m3u8 download
 924             'skip_download': True,
 925         },
 926     }, {
 927         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 928         'only_matching': True,
 929     }, {
 930         'url': 'https://player.twitch.tv/?channel=lotsofs',
 931         'only_matching': True,
 932     }, {
 933         'url': 'https://go.twitch.tv/food',
 934         'only_matching': True,
 935     }, {
 936         'url': 'https://m.twitch.tv/food',
 937         'only_matching': True,
 938     }]
 939
 940     @classmethod
 941     def suitable(cls, url):
 942         return (False
 943                 if any(ie.suitable(url) for ie in (
 944                     TwitchVodIE,
 945                     TwitchCollectionIE,
 946                     TwitchVideosIE,
 947                     TwitchVideosClipsIE,
 948                     TwitchVideosCollectionsIE,
 949                     TwitchClipsIE))
 950                 else super(TwitchStreamIE, cls).suitable(url))
 951
 952     def _real_extract(self, url):
 953         channel_name = self._match_id(url).lower()
 954
 955         gql = self._download_gql(
 956             channel_name, [{
 957                 'operationName': 'StreamMetadata',
 958                 'variables': {'channelLogin': channel_name},
 959             }, {
 960                 'operationName': 'ComscoreStreamingQuery',
 961                 'variables': {
 962                     'channel': channel_name,
 963                     'clipSlug': '',
 964                     'isClip': False,
 965                     'isLive': True,
 966                     'isVodOrCollection': False,
 967                     'vodID': '',
 968                 },
 969             }, {
 970                 'operationName': 'VideoPreviewOverlay',
 971                 'variables': {'login': channel_name},
 972             }],
 973             'Downloading stream GraphQL')
 974
 975         user = gql[0]['data']['user']
 976
 977         if not user:
 978             raise ExtractorError(
 979                 '%s does not exist' % channel_name, expected=True)
 980
 981         stream = user['stream']
 982
 983         if not stream:
 984             raise UserNotLive(video_id=channel_name)
 985
 986         access_token = self._download_access_token(
 987             channel_name, 'stream', 'channelName')
 988         token = access_token['value']
 989
 990         stream_id = stream.get('id') or channel_name
 991         query = {
 992             'allow_source': 'true',
 993             'allow_audio_only': 'true',
 994             'allow_spectre': 'true',
 995             'p': random.randint(1000000, 10000000),
 996             'player': 'twitchweb',
 997             'playlist_include_framerate': 'true',
 998             'segment_preference': '4',
 999             'sig': access_token['signature'].encode('utf-8'),
1000             'token': token.encode('utf-8'),
1001         }
1002         formats = self._extract_m3u8_formats(
1003             '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
1004             stream_id, 'mp4', query=query)
1005         self._prefer_source(formats)
1006
1007         view_count = stream.get('viewers')
1008         timestamp = unified_timestamp(stream.get('createdAt'))
1009
1010         sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
1011         uploader = sq_user.get('displayName')
1012         description = try_get(
1013             sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
1014
1015         thumbnail = url_or_none(try_get(
1016             gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
1017             compat_str))
1018
1019         title = uploader or channel_name
1020         stream_type = stream.get('type')
1021         if stream_type in ['rerun', 'live']:
1022             title += ' (%s)' % stream_type
1023
1024         return {
1025             'id': stream_id,
1026             'display_id': channel_name,
1027             'title': title,
1028             'description': description,
1029             'thumbnail': thumbnail,
1030             'uploader': uploader,
1031             'uploader_id': channel_name,
1032             'timestamp': timestamp,
1033             'view_count': view_count,
1034             'formats': formats,
1035             'is_live': stream_type == 'live',
1036         }
1037
1038
1039 class TwitchClipsIE(TwitchBaseIE):
1040     IE_NAME = 'twitch:clips'
1041     _VALID_URL = r'''(?x)
1042                     https?://
1043                         (?:
1044                             clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)|
1045                             (?:(?:www|go|m)\.)?twitch\.tv/[^/]+/clip/
1046                         )
1047                         (?P<id>[^/?#&]+)
1048                     '''
1049
1050     _TESTS = [{
1051         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
1052         'md5': '761769e1eafce0ffebfb4089cb3847cd',
1053         'info_dict': {
1054             'id': '42850523',
1055             'display_id': 'FaintLightGullWholeWheat',
1056             'ext': 'mp4',
1057             'title': 'EA Play 2016 Live from the Novo Theatre',
1058             'thumbnail': r're:^https?://.*\.jpg',
1059             'timestamp': 1465767393,
1060             'upload_date': '20160612',
1061             'creator': 'EA',
1062             'uploader': 'stereotype_',
1063             'uploader_id': '43566419',
1064         },
1065     }, {
1066         # multiple formats
1067         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
1068         'only_matching': True,
1069     }, {
1070         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
1071         'only_matching': True,
1072     }, {
1073         'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited',
1074         'only_matching': True,
1075     }, {
1076         'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1077         'only_matching': True,
1078     }, {
1079         'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
1080         'only_matching': True,
1081     }]
1082
1083     def _real_extract(self, url):
1084         video_id = self._match_id(url)
1085
1086         clip = self._download_gql(
1087             video_id, [{
1088                 'operationName': 'VideoAccessToken_Clip',
1089                 'variables': {
1090                     'slug': video_id,
1091                 },
1092             }],
1093             'Downloading clip access token GraphQL')[0]['data']['clip']
1094
1095         if not clip:
1096             raise ExtractorError(
1097                 'This clip is no longer available', expected=True)
1098
1099         access_query = {
1100             'sig': clip['playbackAccessToken']['signature'],
1101             'token': clip['playbackAccessToken']['value'],
1102         }
1103
1104         data = self._download_base_gql(
1105             video_id, {
1106                 'query': '''{
1107   clip(slug: "%s") {
1108     broadcaster {
1109       displayName
1110     }
1111     createdAt
1112     curator {
1113       displayName
1114       id
1115     }
1116     durationSeconds
1117     id
1118     tiny: thumbnailURL(width: 86, height: 45)
1119     small: thumbnailURL(width: 260, height: 147)
1120     medium: thumbnailURL(width: 480, height: 272)
1121     title
1122     videoQualities {
1123       frameRate
1124       quality
1125       sourceURL
1126     }
1127     viewCount
1128   }
1129 }''' % video_id}, 'Downloading clip GraphQL', fatal=False)
1130
1131         if data:
1132             clip = try_get(data, lambda x: x['data']['clip'], dict) or clip
1133
1134         formats = []
1135         for option in clip.get('videoQualities', []):
1136             if not isinstance(option, dict):
1137                 continue
1138             source = url_or_none(option.get('sourceURL'))
1139             if not source:
1140                 continue
1141             formats.append({
1142                 'url': update_url_query(source, access_query),
1143                 'format_id': option.get('quality'),
1144                 'height': int_or_none(option.get('quality')),
1145                 'fps': int_or_none(option.get('frameRate')),
1146             })
1147         self._sort_formats(formats)
1148
1149         thumbnails = []
1150         for thumbnail_id in ('tiny', 'small', 'medium'):
1151             thumbnail_url = clip.get(thumbnail_id)
1152             if not thumbnail_url:
1153                 continue
1154             thumb = {
1155                 'id': thumbnail_id,
1156                 'url': thumbnail_url,
1157             }
1158             mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url)
1159             if mobj:
1160                 thumb.update({
1161                     'height': int(mobj.group(2)),
1162                     'width': int(mobj.group(1)),
1163                 })
1164             thumbnails.append(thumb)
1165
1166         old_id = self._search_regex(r'%7C(\d+)(?:-\d+)?.mp4', formats[-1]['url'], 'old id', default=None)
1167
1168         return {
1169             'id': clip.get('id') or video_id,
1170             '_old_archive_ids': [make_archive_id(self, old_id)] if old_id else None,
1171             'display_id': video_id,
1172             'title': clip.get('title'),
1173             'formats': formats,
1174             'duration': int_or_none(clip.get('durationSeconds')),
1175             'view_count': int_or_none(clip.get('viewCount')),
1176             'timestamp': unified_timestamp(clip.get('createdAt')),
1177             'thumbnails': thumbnails,
1178             'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str),
1179             'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str),
1180             'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
1181         }