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