_ Git - youtube-dl/blob - youtube_dl/extractor/twitch.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import re
   6 import random
   7
   8 from .common import InfoExtractor
   9 from ..compat import (
  10     compat_HTTPError,
  11     compat_parse_qs,
  12     compat_str,
  13     compat_urllib_parse_urlencode,
  14     compat_urllib_parse_urlparse,
  15 )
  16 from ..utils import (
  17     clean_html,
  18     ExtractorError,
  19     int_or_none,
  20     js_to_json,
  21     orderedSet,
  22     parse_duration,
  23     parse_iso8601,
  24     update_url_query,
  25     urlencode_postdata,
  26     urljoin,
  27 )
  28
  29
  30 class TwitchBaseIE(InfoExtractor):
  31     _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv'
  32
  33     _API_BASE = 'https://api.twitch.tv'
  34     _USHER_BASE = 'https://usher.ttvnw.net'
  35     _LOGIN_URL = 'https://www.twitch.tv/login'
  36     _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6'
  37     _NETRC_MACHINE = 'twitch'
  38
  39     def _handle_error(self, response):
  40         if not isinstance(response, dict):
  41             return
  42         error = response.get('error')
  43         if error:
  44             raise ExtractorError(
  45                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  46                 expected=True)
  47
  48     def _call_api(self, path, item_id, note):
  49         response = self._download_json(
  50             '%s/%s' % (self._API_BASE, path), item_id, note,
  51             headers={'Client-ID': self._CLIENT_ID})
  52         self._handle_error(response)
  53         return response
  54
  55     def _real_initialize(self):
  56         self._login()
  57
  58     def _login(self):
  59         (username, password) = self._get_login_info()
  60         if username is None:
  61             return
  62
  63         def fail(message):
  64             raise ExtractorError(
  65                 'Unable to login. Twitch said: %s' % message, expected=True)
  66
  67         def login_step(page, urlh, note, data):
  68             form = self._hidden_inputs(page)
  69             form.update(data)
  70
  71             page_url = urlh.geturl()
  72             post_url = self._search_regex(
  73                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  74                 'post url', default=page_url, group='url')
  75             post_url = urljoin(page_url, post_url)
  76
  77             headers = {'Referer': page_url}
  78
  79             try:
  80                 response = self._download_json(
  81                     post_url, None, note,
  82                     data=urlencode_postdata(form),
  83                     headers=headers)
  84             except ExtractorError as e:
  85                 if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400:
  86                     response = self._parse_json(
  87                         e.cause.read().decode('utf-8'), None)
  88                     fail(response.get('message') or response['errors'][0])
  89                 raise
  90
  91             if 'Authenticated successfully' in response.get('message', ''):
  92                 return None, None
  93
  94             redirect_url = urljoin(
  95                 post_url,
  96                 response.get('redirect') or response['redirect_path'])
  97             return self._download_webpage_handle(
  98                 redirect_url, None, 'Downloading login redirect page',
  99                 headers=headers)
 100
 101         login_page, handle = self._download_webpage_handle(
 102             self._LOGIN_URL, None, 'Downloading login page')
 103
 104         # Some TOR nodes and public proxies are blocked completely
 105         if 'blacklist_message' in login_page:
 106             fail(clean_html(login_page))
 107
 108         redirect_page, handle = login_step(
 109             login_page, handle, 'Logging in', {
 110                 'username': username,
 111                 'password': password,
 112             })
 113
 114         # Successful login
 115         if not redirect_page:
 116             return
 117
 118         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 119             # TODO: Add mechanism to request an SMS or phone call
 120             tfa_token = self._get_tfa_info('two-factor authentication token')
 121             login_step(redirect_page, handle, 'Submitting TFA token', {
 122                 'authy_token': tfa_token,
 123                 'remember_2fa': 'true',
 124             })
 125
 126     def _prefer_source(self, formats):
 127         try:
 128             source = next(f for f in formats if f['format_id'] == 'Source')
 129             source['preference'] = 10
 130         except StopIteration:
 131             pass  # No Source stream present
 132         self._sort_formats(formats)
 133
 134
 135 class TwitchItemBaseIE(TwitchBaseIE):
 136     def _download_info(self, item, item_id):
 137         return self._extract_info(self._call_api(
 138             'kraken/videos/%s%s' % (item, item_id), item_id,
 139             'Downloading %s info JSON' % self._ITEM_TYPE))
 140
 141     def _extract_media(self, item_id):
 142         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 143         response = self._call_api(
 144             'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id,
 145             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 146         entries = []
 147         chunks = response['chunks']
 148         qualities = list(chunks.keys())
 149         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 150             formats = []
 151             for fmt_num, fragment_fmt in enumerate(fragment):
 152                 format_id = qualities[fmt_num]
 153                 fmt = {
 154                     'url': fragment_fmt['url'],
 155                     'format_id': format_id,
 156                     'quality': 1 if format_id == 'live' else 0,
 157                 }
 158                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 159                 if m:
 160                     fmt['height'] = int(m.group('height'))
 161                 formats.append(fmt)
 162             self._sort_formats(formats)
 163             entry = dict(info)
 164             entry['id'] = '%s_%d' % (entry['id'], num)
 165             entry['title'] = '%s part %d' % (entry['title'], num)
 166             entry['formats'] = formats
 167             entries.append(entry)
 168         return self.playlist_result(entries, info['id'], info['title'])
 169
 170     def _extract_info(self, info):
 171         return {
 172             'id': info['_id'],
 173             'title': info.get('title') or 'Untitled Broadcast',
 174             'description': info.get('description'),
 175             'duration': int_or_none(info.get('length')),
 176             'thumbnail': info.get('preview'),
 177             'uploader': info.get('channel', {}).get('display_name'),
 178             'uploader_id': info.get('channel', {}).get('name'),
 179             'timestamp': parse_iso8601(info.get('recorded_at')),
 180             'view_count': int_or_none(info.get('views')),
 181         }
 182
 183     def _real_extract(self, url):
 184         return self._extract_media(self._match_id(url))
 185
 186
 187 class TwitchVideoIE(TwitchItemBaseIE):
 188     IE_NAME = 'twitch:video'
 189     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 190     _ITEM_TYPE = 'video'
 191     _ITEM_SHORTCUT = 'a'
 192
 193     _TEST = {
 194         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 195         'info_dict': {
 196             'id': 'a577357806',
 197             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 198         },
 199         'playlist_mincount': 12,
 200         'skip': 'HTTP Error 404: Not Found',
 201     }
 202
 203
 204 class TwitchChapterIE(TwitchItemBaseIE):
 205     IE_NAME = 'twitch:chapter'
 206     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 207     _ITEM_TYPE = 'chapter'
 208     _ITEM_SHORTCUT = 'c'
 209
 210     _TESTS = [{
 211         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 212         'info_dict': {
 213             'id': 'c5285812',
 214             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 215         },
 216         'playlist_mincount': 3,
 217         'skip': 'HTTP Error 404: Not Found',
 218     }, {
 219         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 220         'only_matching': True,
 221     }]
 222
 223
 224 class TwitchVodIE(TwitchItemBaseIE):
 225     IE_NAME = 'twitch:vod'
 226     _VALID_URL = r'''(?x)
 227                     https?://
 228                         (?:
 229                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v|videos)/|
 230                             player\.twitch\.tv/\?.*?\bvideo=v
 231                         )
 232                         (?P<id>\d+)
 233                     '''
 234     _ITEM_TYPE = 'vod'
 235     _ITEM_SHORTCUT = 'v'
 236
 237     _TESTS = [{
 238         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 239         'info_dict': {
 240             'id': 'v6528877',
 241             'ext': 'mp4',
 242             'title': 'LCK Summer Split - Week 6 Day 1',
 243             'thumbnail': r're:^https?://.*\.jpg$',
 244             'duration': 17208,
 245             'timestamp': 1435131709,
 246             'upload_date': '20150624',
 247             'uploader': 'Riot Games',
 248             'uploader_id': 'riotgames',
 249             'view_count': int,
 250             'start_time': 310,
 251         },
 252         'params': {
 253             # m3u8 download
 254             'skip_download': True,
 255         },
 256     }, {
 257         # Untitled broadcast (title is None)
 258         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 259         'info_dict': {
 260             'id': 'v11230755',
 261             'ext': 'mp4',
 262             'title': 'Untitled Broadcast',
 263             'thumbnail': r're:^https?://.*\.jpg$',
 264             'duration': 1638,
 265             'timestamp': 1439746708,
 266             'upload_date': '20150816',
 267             'uploader': 'BelkAO_o',
 268             'uploader_id': 'belkao_o',
 269             'view_count': int,
 270         },
 271         'params': {
 272             # m3u8 download
 273             'skip_download': True,
 274         },
 275         'skip': 'HTTP Error 404: Not Found',
 276     }, {
 277         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 278         'only_matching': True,
 279     }, {
 280         'url': 'https://www.twitch.tv/videos/6528877',
 281         'only_matching': True,
 282     }, {
 283         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 284         'only_matching': True,
 285     }]
 286
 287     def _real_extract(self, url):
 288         item_id = self._match_id(url)
 289
 290         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 291         access_token = self._call_api(
 292             'api/vods/%s/access_token' % item_id, item_id,
 293             'Downloading %s access token' % self._ITEM_TYPE)
 294
 295         formats = self._extract_m3u8_formats(
 296             '%s/vod/%s?%s' % (
 297                 self._USHER_BASE, item_id,
 298                 compat_urllib_parse_urlencode({
 299                     'allow_source': 'true',
 300                     'allow_audio_only': 'true',
 301                     'allow_spectre': 'true',
 302                     'player': 'twitchweb',
 303                     'nauth': access_token['token'],
 304                     'nauthsig': access_token['sig'],
 305                 })),
 306             item_id, 'mp4', entry_protocol='m3u8_native')
 307
 308         self._prefer_source(formats)
 309         info['formats'] = formats
 310
 311         parsed_url = compat_urllib_parse_urlparse(url)
 312         query = compat_parse_qs(parsed_url.query)
 313         if 't' in query:
 314             info['start_time'] = parse_duration(query['t'][0])
 315
 316         if info.get('timestamp') is not None:
 317             info['subtitles'] = {
 318                 'rechat': [{
 319                     'url': update_url_query(
 320                         'https://rechat.twitch.tv/rechat-messages', {
 321                             'video_id': 'v%s' % item_id,
 322                             'start': info['timestamp'],
 323                         }),
 324                     'ext': 'json',
 325                 }],
 326             }
 327
 328         return info
 329
 330
 331 class TwitchPlaylistBaseIE(TwitchBaseIE):
 332     _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
 333     _PAGE_LIMIT = 100
 334
 335     def _extract_playlist(self, channel_id):
 336         info = self._call_api(
 337             'kraken/channels/%s' % channel_id,
 338             channel_id, 'Downloading channel info JSON')
 339         channel_name = info.get('display_name') or info.get('name')
 340         entries = []
 341         offset = 0
 342         limit = self._PAGE_LIMIT
 343         broken_paging_detected = False
 344         counter_override = None
 345         for counter in itertools.count(1):
 346             response = self._call_api(
 347                 self._PLAYLIST_PATH % (channel_id, offset, limit),
 348                 channel_id,
 349                 'Downloading %s JSON page %s'
 350                 % (self._PLAYLIST_TYPE, counter_override or counter))
 351             page_entries = self._extract_playlist_page(response)
 352             if not page_entries:
 353                 break
 354             total = int_or_none(response.get('_total'))
 355             # Since the beginning of March 2016 twitch's paging mechanism
 356             # is completely broken on the twitch side. It simply ignores
 357             # a limit and returns the whole offset number of videos.
 358             # Working around by just requesting all videos at once.
 359             # Upd: pagination bug was fixed by twitch on 15.03.2016.
 360             if not broken_paging_detected and total and len(page_entries) > limit:
 361                 self.report_warning(
 362                     'Twitch pagination is broken on twitch side, requesting all videos at once',
 363                     channel_id)
 364                 broken_paging_detected = True
 365                 offset = total
 366                 counter_override = '(all at once)'
 367                 continue
 368             entries.extend(page_entries)
 369             if broken_paging_detected or total and len(page_entries) >= total:
 370                 break
 371             offset += limit
 372         return self.playlist_result(
 373             [self._make_url_result(entry) for entry in orderedSet(entries)],
 374             channel_id, channel_name)
 375
 376     def _make_url_result(self, url):
 377         try:
 378             video_id = 'v%s' % TwitchVodIE._match_id(url)
 379             return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id)
 380         except AssertionError:
 381             return self.url_result(url)
 382
 383     def _extract_playlist_page(self, response):
 384         videos = response.get('videos')
 385         return [video['url'] for video in videos] if videos else []
 386
 387     def _real_extract(self, url):
 388         return self._extract_playlist(self._match_id(url))
 389
 390
 391 class TwitchProfileIE(TwitchPlaylistBaseIE):
 392     IE_NAME = 'twitch:profile'
 393     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 394     _PLAYLIST_TYPE = 'profile'
 395
 396     _TESTS = [{
 397         'url': 'http://www.twitch.tv/vanillatv/profile',
 398         'info_dict': {
 399             'id': 'vanillatv',
 400             'title': 'VanillaTV',
 401         },
 402         'playlist_mincount': 412,
 403     }, {
 404         'url': 'http://m.twitch.tv/vanillatv/profile',
 405         'only_matching': True,
 406     }]
 407
 408
 409 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
 410     _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
 411     _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
 412
 413
 414 class TwitchAllVideosIE(TwitchVideosBaseIE):
 415     IE_NAME = 'twitch:videos:all'
 416     _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 417     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
 418     _PLAYLIST_TYPE = 'all videos'
 419
 420     _TESTS = [{
 421         'url': 'https://www.twitch.tv/spamfish/videos/all',
 422         'info_dict': {
 423             'id': 'spamfish',
 424             'title': 'Spamfish',
 425         },
 426         'playlist_mincount': 869,
 427     }, {
 428         'url': 'https://m.twitch.tv/spamfish/videos/all',
 429         'only_matching': True,
 430     }]
 431
 432
 433 class TwitchUploadsIE(TwitchVideosBaseIE):
 434     IE_NAME = 'twitch:videos:uploads'
 435     _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 436     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
 437     _PLAYLIST_TYPE = 'uploads'
 438
 439     _TESTS = [{
 440         'url': 'https://www.twitch.tv/spamfish/videos/uploads',
 441         'info_dict': {
 442             'id': 'spamfish',
 443             'title': 'Spamfish',
 444         },
 445         'playlist_mincount': 0,
 446     }, {
 447         'url': 'https://m.twitch.tv/spamfish/videos/uploads',
 448         'only_matching': True,
 449     }]
 450
 451
 452 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
 453     IE_NAME = 'twitch:videos:past-broadcasts'
 454     _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 455     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
 456     _PLAYLIST_TYPE = 'past broadcasts'
 457
 458     _TESTS = [{
 459         'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
 460         'info_dict': {
 461             'id': 'spamfish',
 462             'title': 'Spamfish',
 463         },
 464         'playlist_mincount': 0,
 465     }, {
 466         'url': 'https://m.twitch.tv/spamfish/videos/past-broadcasts',
 467         'only_matching': True,
 468     }]
 469
 470
 471 class TwitchHighlightsIE(TwitchVideosBaseIE):
 472     IE_NAME = 'twitch:videos:highlights'
 473     _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 474     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
 475     _PLAYLIST_TYPE = 'highlights'
 476
 477     _TESTS = [{
 478         'url': 'https://www.twitch.tv/spamfish/videos/highlights',
 479         'info_dict': {
 480             'id': 'spamfish',
 481             'title': 'Spamfish',
 482         },
 483         'playlist_mincount': 805,
 484     }, {
 485         'url': 'https://m.twitch.tv/spamfish/videos/highlights',
 486         'only_matching': True,
 487     }]
 488
 489
 490 class TwitchStreamIE(TwitchBaseIE):
 491     IE_NAME = 'twitch:stream'
 492     _VALID_URL = r'''(?x)
 493                     https?://
 494                         (?:
 495                             (?:(?:www|go|m)\.)?twitch\.tv/|
 496                             player\.twitch\.tv/\?.*?\bchannel=
 497                         )
 498                         (?P<id>[^/#?]+)
 499                     '''
 500
 501     _TESTS = [{
 502         'url': 'http://www.twitch.tv/shroomztv',
 503         'info_dict': {
 504             'id': '12772022048',
 505             'display_id': 'shroomztv',
 506             'ext': 'mp4',
 507             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 508             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 509             'is_live': True,
 510             'timestamp': 1421928037,
 511             'upload_date': '20150122',
 512             'uploader': 'ShroomzTV',
 513             'uploader_id': 'shroomztv',
 514             'view_count': int,
 515         },
 516         'params': {
 517             # m3u8 download
 518             'skip_download': True,
 519         },
 520     }, {
 521         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 522         'only_matching': True,
 523     }, {
 524         'url': 'https://player.twitch.tv/?channel=lotsofs',
 525         'only_matching': True,
 526     }, {
 527         'url': 'https://go.twitch.tv/food',
 528         'only_matching': True,
 529     }, {
 530         'url': 'https://m.twitch.tv/food',
 531         'only_matching': True,
 532     }]
 533
 534     @classmethod
 535     def suitable(cls, url):
 536         return (False
 537                 if any(ie.suitable(url) for ie in (
 538                     TwitchVideoIE,
 539                     TwitchChapterIE,
 540                     TwitchVodIE,
 541                     TwitchProfileIE,
 542                     TwitchAllVideosIE,
 543                     TwitchUploadsIE,
 544                     TwitchPastBroadcastsIE,
 545                     TwitchHighlightsIE))
 546                 else super(TwitchStreamIE, cls).suitable(url))
 547
 548     def _real_extract(self, url):
 549         channel_id = self._match_id(url)
 550
 551         stream = self._call_api(
 552             'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
 553             'Downloading stream JSON').get('stream')
 554
 555         if not stream:
 556             raise ExtractorError('%s is offline' % channel_id, expected=True)
 557
 558         # Channel name may be typed if different case than the original channel name
 559         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 560         # an invalid m3u8 URL. Working around by use of original channel name from stream
 561         # JSON and fallback to lowercase if it's not available.
 562         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 563
 564         access_token = self._call_api(
 565             'api/channels/%s/access_token' % channel_id, channel_id,
 566             'Downloading channel access token')
 567
 568         query = {
 569             'allow_source': 'true',
 570             'allow_audio_only': 'true',
 571             'allow_spectre': 'true',
 572             'p': random.randint(1000000, 10000000),
 573             'player': 'twitchweb',
 574             'segment_preference': '4',
 575             'sig': access_token['sig'].encode('utf-8'),
 576             'token': access_token['token'].encode('utf-8'),
 577         }
 578         formats = self._extract_m3u8_formats(
 579             '%s/api/channel/hls/%s.m3u8?%s'
 580             % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
 581             channel_id, 'mp4')
 582         self._prefer_source(formats)
 583
 584         view_count = stream.get('viewers')
 585         timestamp = parse_iso8601(stream.get('created_at'))
 586
 587         channel = stream['channel']
 588         title = self._live_title(channel.get('display_name') or channel.get('name'))
 589         description = channel.get('status')
 590
 591         thumbnails = []
 592         for thumbnail_key, thumbnail_url in stream['preview'].items():
 593             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 594             if not m:
 595                 continue
 596             thumbnails.append({
 597                 'url': thumbnail_url,
 598                 'width': int(m.group('width')),
 599                 'height': int(m.group('height')),
 600             })
 601
 602         return {
 603             'id': compat_str(stream['_id']),
 604             'display_id': channel_id,
 605             'title': title,
 606             'description': description,
 607             'thumbnails': thumbnails,
 608             'uploader': channel.get('display_name'),
 609             'uploader_id': channel.get('name'),
 610             'timestamp': timestamp,
 611             'view_count': view_count,
 612             'formats': formats,
 613             'is_live': True,
 614         }
 615
 616
 617 class TwitchClipsIE(InfoExtractor):
 618     IE_NAME = 'twitch:clips'
 619     _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
 620
 621     _TESTS = [{
 622         'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
 623         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 624         'info_dict': {
 625             'id': 'AggressiveCobraPoooound',
 626             'ext': 'mp4',
 627             'title': 'EA Play 2016 Live from the Novo Theatre',
 628             'thumbnail': r're:^https?://.*\.jpg',
 629             'creator': 'EA',
 630             'uploader': 'stereotype_',
 631             'uploader_id': 'stereotype_',
 632         },
 633     }, {
 634         # multiple formats
 635         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 636         'only_matching': True,
 637     }]
 638
 639     def _real_extract(self, url):
 640         video_id = self._match_id(url)
 641
 642         webpage = self._download_webpage(url, video_id)
 643
 644         clip = self._parse_json(
 645             self._search_regex(
 646                 r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
 647             video_id, transform_source=js_to_json)
 648
 649         title = clip.get('title') or clip.get('channel_title') or self._og_search_title(webpage)
 650
 651         formats = [{
 652             'url': option['source'],
 653             'format_id': option.get('quality'),
 654             'height': int_or_none(option.get('quality')),
 655         } for option in clip.get('quality_options', []) if option.get('source')]
 656
 657         if not formats:
 658             formats = [{
 659                 'url': clip['clip_video_url'],
 660             }]
 661
 662         self._sort_formats(formats)
 663
 664         return {
 665             'id': video_id,
 666             'title': title,
 667             'thumbnail': self._og_search_thumbnail(webpage),
 668             'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
 669             'uploader': clip.get('curator_login'),
 670             'uploader_id': clip.get('curator_display_name'),
 671             'formats': formats,
 672         }