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