_ 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_kwargs,
  12     compat_parse_qs,
  13     compat_str,
  14     compat_urllib_parse_urlencode,
  15     compat_urllib_parse_urlparse,
  16 )
  17 from ..utils import (
  18     clean_html,
  19     ExtractorError,
  20     float_or_none,
  21     int_or_none,
  22     orderedSet,
  23     parse_duration,
  24     parse_iso8601,
  25     qualities,
  26     try_get,
  27     unified_timestamp,
  28     update_url_query,
  29     urlencode_postdata,
  30     url_or_none,
  31     urljoin,
  32 )
  33
  34
  35 class TwitchBaseIE(InfoExtractor):
  36     _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv'
  37
  38     _API_BASE = 'https://api.twitch.tv'
  39     _USHER_BASE = 'https://usher.ttvnw.net'
  40     _LOGIN_URL = 'https://www.twitch.tv/login'
  41     _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6'
  42     _NETRC_MACHINE = 'twitch'
  43
  44     def _handle_error(self, response):
  45         if not isinstance(response, dict):
  46             return
  47         error = response.get('error')
  48         if error:
  49             raise ExtractorError(
  50                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  51                 expected=True)
  52
  53     def _call_api(self, path, item_id, *args, **kwargs):
  54         kwargs.setdefault('headers', {})['Client-ID'] = self._CLIENT_ID
  55         response = self._download_json(
  56             '%s/%s' % (self._API_BASE, path), item_id,
  57             *args, **compat_kwargs(kwargs))
  58         self._handle_error(response)
  59         return response
  60
  61     def _real_initialize(self):
  62         self._login()
  63
  64     def _login(self):
  65         username, password = self._get_login_info()
  66         if username is None:
  67             return
  68
  69         def fail(message):
  70             raise ExtractorError(
  71                 'Unable to login. Twitch said: %s' % message, expected=True)
  72
  73         def login_step(page, urlh, note, data):
  74             form = self._hidden_inputs(page)
  75             form.update(data)
  76
  77             page_url = urlh.geturl()
  78             post_url = self._search_regex(
  79                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  80                 'post url', default=page_url, group='url')
  81             post_url = urljoin(page_url, post_url)
  82
  83             headers = {'Referer': page_url}
  84
  85             try:
  86                 response = self._download_json(
  87                     post_url, None, note,
  88                     data=urlencode_postdata(form),
  89                     headers=headers)
  90             except ExtractorError as e:
  91                 if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400:
  92                     response = self._parse_json(
  93                         e.cause.read().decode('utf-8'), None)
  94                     fail(response.get('message') or response['errors'][0])
  95                 raise
  96
  97             if 'Authenticated successfully' in response.get('message', ''):
  98                 return None, None
  99
 100             redirect_url = urljoin(
 101                 post_url,
 102                 response.get('redirect') or response['redirect_path'])
 103             return self._download_webpage_handle(
 104                 redirect_url, None, 'Downloading login redirect page',
 105                 headers=headers)
 106
 107         login_page, handle = self._download_webpage_handle(
 108             self._LOGIN_URL, None, 'Downloading login page')
 109
 110         # Some TOR nodes and public proxies are blocked completely
 111         if 'blacklist_message' in login_page:
 112             fail(clean_html(login_page))
 113
 114         redirect_page, handle = login_step(
 115             login_page, handle, 'Logging in', {
 116                 'username': username,
 117                 'password': password,
 118             })
 119
 120         # Successful login
 121         if not redirect_page:
 122             return
 123
 124         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 125             # TODO: Add mechanism to request an SMS or phone call
 126             tfa_token = self._get_tfa_info('two-factor authentication token')
 127             login_step(redirect_page, handle, 'Submitting TFA token', {
 128                 'authy_token': tfa_token,
 129                 'remember_2fa': 'true',
 130             })
 131
 132     def _prefer_source(self, formats):
 133         try:
 134             source = next(f for f in formats if f['format_id'] == 'Source')
 135             source['preference'] = 10
 136         except StopIteration:
 137             pass  # No Source stream present
 138         self._sort_formats(formats)
 139
 140
 141 class TwitchItemBaseIE(TwitchBaseIE):
 142     def _download_info(self, item, item_id):
 143         return self._extract_info(self._call_api(
 144             'kraken/videos/%s%s' % (item, item_id), item_id,
 145             'Downloading %s info JSON' % self._ITEM_TYPE))
 146
 147     def _extract_media(self, item_id):
 148         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 149         response = self._call_api(
 150             'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id,
 151             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 152         entries = []
 153         chunks = response['chunks']
 154         qualities = list(chunks.keys())
 155         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 156             formats = []
 157             for fmt_num, fragment_fmt in enumerate(fragment):
 158                 format_id = qualities[fmt_num]
 159                 fmt = {
 160                     'url': fragment_fmt['url'],
 161                     'format_id': format_id,
 162                     'quality': 1 if format_id == 'live' else 0,
 163                 }
 164                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 165                 if m:
 166                     fmt['height'] = int(m.group('height'))
 167                 formats.append(fmt)
 168             self._sort_formats(formats)
 169             entry = dict(info)
 170             entry['id'] = '%s_%d' % (entry['id'], num)
 171             entry['title'] = '%s part %d' % (entry['title'], num)
 172             entry['formats'] = formats
 173             entries.append(entry)
 174         return self.playlist_result(entries, info['id'], info['title'])
 175
 176     def _extract_info(self, info):
 177         status = info.get('status')
 178         if status == 'recording':
 179             is_live = True
 180         elif status == 'recorded':
 181             is_live = False
 182         else:
 183             is_live = None
 184         return {
 185             'id': info['_id'],
 186             'title': info.get('title') or 'Untitled Broadcast',
 187             'description': info.get('description'),
 188             'duration': int_or_none(info.get('length')),
 189             'thumbnail': info.get('preview'),
 190             'uploader': info.get('channel', {}).get('display_name'),
 191             'uploader_id': info.get('channel', {}).get('name'),
 192             'timestamp': parse_iso8601(info.get('recorded_at')),
 193             'view_count': int_or_none(info.get('views')),
 194             'is_live': is_live,
 195         }
 196
 197     def _real_extract(self, url):
 198         return self._extract_media(self._match_id(url))
 199
 200
 201 class TwitchVideoIE(TwitchItemBaseIE):
 202     IE_NAME = 'twitch:video'
 203     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 204     _ITEM_TYPE = 'video'
 205     _ITEM_SHORTCUT = 'a'
 206
 207     _TEST = {
 208         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 209         'info_dict': {
 210             'id': 'a577357806',
 211             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 212         },
 213         'playlist_mincount': 12,
 214         'skip': 'HTTP Error 404: Not Found',
 215     }
 216
 217
 218 class TwitchChapterIE(TwitchItemBaseIE):
 219     IE_NAME = 'twitch:chapter'
 220     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 221     _ITEM_TYPE = 'chapter'
 222     _ITEM_SHORTCUT = 'c'
 223
 224     _TESTS = [{
 225         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 226         'info_dict': {
 227             'id': 'c5285812',
 228             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 229         },
 230         'playlist_mincount': 3,
 231         'skip': 'HTTP Error 404: Not Found',
 232     }, {
 233         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 234         'only_matching': True,
 235     }]
 236
 237
 238 class TwitchVodIE(TwitchItemBaseIE):
 239     IE_NAME = 'twitch:vod'
 240     _VALID_URL = r'''(?x)
 241                     https?://
 242                         (?:
 243                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v|videos)/|
 244                             player\.twitch\.tv/\?.*?\bvideo=v
 245                         )
 246                         (?P<id>\d+)
 247                     '''
 248     _ITEM_TYPE = 'vod'
 249     _ITEM_SHORTCUT = 'v'
 250
 251     _TESTS = [{
 252         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 253         'info_dict': {
 254             'id': 'v6528877',
 255             'ext': 'mp4',
 256             'title': 'LCK Summer Split - Week 6 Day 1',
 257             'thumbnail': r're:^https?://.*\.jpg$',
 258             'duration': 17208,
 259             'timestamp': 1435131709,
 260             'upload_date': '20150624',
 261             'uploader': 'Riot Games',
 262             'uploader_id': 'riotgames',
 263             'view_count': int,
 264             'start_time': 310,
 265         },
 266         'params': {
 267             # m3u8 download
 268             'skip_download': True,
 269         },
 270     }, {
 271         # Untitled broadcast (title is None)
 272         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 273         'info_dict': {
 274             'id': 'v11230755',
 275             'ext': 'mp4',
 276             'title': 'Untitled Broadcast',
 277             'thumbnail': r're:^https?://.*\.jpg$',
 278             'duration': 1638,
 279             'timestamp': 1439746708,
 280             'upload_date': '20150816',
 281             'uploader': 'BelkAO_o',
 282             'uploader_id': 'belkao_o',
 283             'view_count': int,
 284         },
 285         'params': {
 286             # m3u8 download
 287             'skip_download': True,
 288         },
 289         'skip': 'HTTP Error 404: Not Found',
 290     }, {
 291         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 292         'only_matching': True,
 293     }, {
 294         'url': 'https://www.twitch.tv/videos/6528877',
 295         'only_matching': True,
 296     }, {
 297         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 298         'only_matching': True,
 299     }]
 300
 301     def _real_extract(self, url):
 302         item_id = self._match_id(url)
 303
 304         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 305         access_token = self._call_api(
 306             'api/vods/%s/access_token' % item_id, item_id,
 307             'Downloading %s access token' % self._ITEM_TYPE)
 308
 309         formats = self._extract_m3u8_formats(
 310             '%s/vod/%s?%s' % (
 311                 self._USHER_BASE, item_id,
 312                 compat_urllib_parse_urlencode({
 313                     'allow_source': 'true',
 314                     'allow_audio_only': 'true',
 315                     'allow_spectre': 'true',
 316                     'player': 'twitchweb',
 317                     'nauth': access_token['token'],
 318                     'nauthsig': access_token['sig'],
 319                 })),
 320             item_id, 'mp4', entry_protocol='m3u8_native')
 321
 322         self._prefer_source(formats)
 323         info['formats'] = formats
 324
 325         parsed_url = compat_urllib_parse_urlparse(url)
 326         query = compat_parse_qs(parsed_url.query)
 327         if 't' in query:
 328             info['start_time'] = parse_duration(query['t'][0])
 329
 330         if info.get('timestamp') is not None:
 331             info['subtitles'] = {
 332                 'rechat': [{
 333                     'url': update_url_query(
 334                         'https://rechat.twitch.tv/rechat-messages', {
 335                             'video_id': 'v%s' % item_id,
 336                             'start': info['timestamp'],
 337                         }),
 338                     'ext': 'json',
 339                 }],
 340             }
 341
 342         return info
 343
 344
 345 class TwitchPlaylistBaseIE(TwitchBaseIE):
 346     _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
 347     _PAGE_LIMIT = 100
 348
 349     def _extract_playlist(self, channel_id):
 350         info = self._call_api(
 351             'kraken/channels/%s' % channel_id,
 352             channel_id, 'Downloading channel info JSON')
 353         channel_name = info.get('display_name') or info.get('name')
 354         entries = []
 355         offset = 0
 356         limit = self._PAGE_LIMIT
 357         broken_paging_detected = False
 358         counter_override = None
 359         for counter in itertools.count(1):
 360             response = self._call_api(
 361                 self._PLAYLIST_PATH % (channel_id, offset, limit),
 362                 channel_id,
 363                 'Downloading %s JSON page %s'
 364                 % (self._PLAYLIST_TYPE, counter_override or counter))
 365             page_entries = self._extract_playlist_page(response)
 366             if not page_entries:
 367                 break
 368             total = int_or_none(response.get('_total'))
 369             # Since the beginning of March 2016 twitch's paging mechanism
 370             # is completely broken on the twitch side. It simply ignores
 371             # a limit and returns the whole offset number of videos.
 372             # Working around by just requesting all videos at once.
 373             # Upd: pagination bug was fixed by twitch on 15.03.2016.
 374             if not broken_paging_detected and total and len(page_entries) > limit:
 375                 self.report_warning(
 376                     'Twitch pagination is broken on twitch side, requesting all videos at once',
 377                     channel_id)
 378                 broken_paging_detected = True
 379                 offset = total
 380                 counter_override = '(all at once)'
 381                 continue
 382             entries.extend(page_entries)
 383             if broken_paging_detected or total and len(page_entries) >= total:
 384                 break
 385             offset += limit
 386         return self.playlist_result(
 387             [self._make_url_result(entry) for entry in orderedSet(entries)],
 388             channel_id, channel_name)
 389
 390     def _make_url_result(self, url):
 391         try:
 392             video_id = 'v%s' % TwitchVodIE._match_id(url)
 393             return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id)
 394         except AssertionError:
 395             return self.url_result(url)
 396
 397     def _extract_playlist_page(self, response):
 398         videos = response.get('videos')
 399         return [video['url'] for video in videos] if videos else []
 400
 401     def _real_extract(self, url):
 402         return self._extract_playlist(self._match_id(url))
 403
 404
 405 class TwitchProfileIE(TwitchPlaylistBaseIE):
 406     IE_NAME = 'twitch:profile'
 407     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 408     _PLAYLIST_TYPE = 'profile'
 409
 410     _TESTS = [{
 411         'url': 'http://www.twitch.tv/vanillatv/profile',
 412         'info_dict': {
 413             'id': 'vanillatv',
 414             'title': 'VanillaTV',
 415         },
 416         'playlist_mincount': 412,
 417     }, {
 418         'url': 'http://m.twitch.tv/vanillatv/profile',
 419         'only_matching': True,
 420     }]
 421
 422
 423 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
 424     _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
 425     _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
 426
 427
 428 class TwitchAllVideosIE(TwitchVideosBaseIE):
 429     IE_NAME = 'twitch:videos:all'
 430     _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 431     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
 432     _PLAYLIST_TYPE = 'all videos'
 433
 434     _TESTS = [{
 435         'url': 'https://www.twitch.tv/spamfish/videos/all',
 436         'info_dict': {
 437             'id': 'spamfish',
 438             'title': 'Spamfish',
 439         },
 440         'playlist_mincount': 869,
 441     }, {
 442         'url': 'https://m.twitch.tv/spamfish/videos/all',
 443         'only_matching': True,
 444     }]
 445
 446
 447 class TwitchUploadsIE(TwitchVideosBaseIE):
 448     IE_NAME = 'twitch:videos:uploads'
 449     _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 450     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
 451     _PLAYLIST_TYPE = 'uploads'
 452
 453     _TESTS = [{
 454         'url': 'https://www.twitch.tv/spamfish/videos/uploads',
 455         'info_dict': {
 456             'id': 'spamfish',
 457             'title': 'Spamfish',
 458         },
 459         'playlist_mincount': 0,
 460     }, {
 461         'url': 'https://m.twitch.tv/spamfish/videos/uploads',
 462         'only_matching': True,
 463     }]
 464
 465
 466 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
 467     IE_NAME = 'twitch:videos:past-broadcasts'
 468     _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 469     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
 470     _PLAYLIST_TYPE = 'past broadcasts'
 471
 472     _TESTS = [{
 473         'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
 474         'info_dict': {
 475             'id': 'spamfish',
 476             'title': 'Spamfish',
 477         },
 478         'playlist_mincount': 0,
 479     }, {
 480         'url': 'https://m.twitch.tv/spamfish/videos/past-broadcasts',
 481         'only_matching': True,
 482     }]
 483
 484
 485 class TwitchHighlightsIE(TwitchVideosBaseIE):
 486     IE_NAME = 'twitch:videos:highlights'
 487     _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 488     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
 489     _PLAYLIST_TYPE = 'highlights'
 490
 491     _TESTS = [{
 492         'url': 'https://www.twitch.tv/spamfish/videos/highlights',
 493         'info_dict': {
 494             'id': 'spamfish',
 495             'title': 'Spamfish',
 496         },
 497         'playlist_mincount': 805,
 498     }, {
 499         'url': 'https://m.twitch.tv/spamfish/videos/highlights',
 500         'only_matching': True,
 501     }]
 502
 503
 504 class TwitchStreamIE(TwitchBaseIE):
 505     IE_NAME = 'twitch:stream'
 506     _VALID_URL = r'''(?x)
 507                     https?://
 508                         (?:
 509                             (?:(?:www|go|m)\.)?twitch\.tv/|
 510                             player\.twitch\.tv/\?.*?\bchannel=
 511                         )
 512                         (?P<id>[^/#?]+)
 513                     '''
 514
 515     _TESTS = [{
 516         'url': 'http://www.twitch.tv/shroomztv',
 517         'info_dict': {
 518             'id': '12772022048',
 519             'display_id': 'shroomztv',
 520             'ext': 'mp4',
 521             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 522             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 523             'is_live': True,
 524             'timestamp': 1421928037,
 525             'upload_date': '20150122',
 526             'uploader': 'ShroomzTV',
 527             'uploader_id': 'shroomztv',
 528             'view_count': int,
 529         },
 530         'params': {
 531             # m3u8 download
 532             'skip_download': True,
 533         },
 534     }, {
 535         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 536         'only_matching': True,
 537     }, {
 538         'url': 'https://player.twitch.tv/?channel=lotsofs',
 539         'only_matching': True,
 540     }, {
 541         'url': 'https://go.twitch.tv/food',
 542         'only_matching': True,
 543     }, {
 544         'url': 'https://m.twitch.tv/food',
 545         'only_matching': True,
 546     }]
 547
 548     @classmethod
 549     def suitable(cls, url):
 550         return (False
 551                 if any(ie.suitable(url) for ie in (
 552                     TwitchVideoIE,
 553                     TwitchChapterIE,
 554                     TwitchVodIE,
 555                     TwitchProfileIE,
 556                     TwitchAllVideosIE,
 557                     TwitchUploadsIE,
 558                     TwitchPastBroadcastsIE,
 559                     TwitchHighlightsIE))
 560                 else super(TwitchStreamIE, cls).suitable(url))
 561
 562     def _real_extract(self, url):
 563         channel_id = self._match_id(url)
 564
 565         stream = self._call_api(
 566             'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
 567             'Downloading stream JSON').get('stream')
 568
 569         if not stream:
 570             raise ExtractorError('%s is offline' % channel_id, expected=True)
 571
 572         # Channel name may be typed if different case than the original channel name
 573         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 574         # an invalid m3u8 URL. Working around by use of original channel name from stream
 575         # JSON and fallback to lowercase if it's not available.
 576         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 577
 578         access_token = self._call_api(
 579             'api/channels/%s/access_token' % channel_id, channel_id,
 580             'Downloading channel access token')
 581
 582         query = {
 583             'allow_source': 'true',
 584             'allow_audio_only': 'true',
 585             'allow_spectre': 'true',
 586             'p': random.randint(1000000, 10000000),
 587             'player': 'twitchweb',
 588             'segment_preference': '4',
 589             'sig': access_token['sig'].encode('utf-8'),
 590             'token': access_token['token'].encode('utf-8'),
 591         }
 592         formats = self._extract_m3u8_formats(
 593             '%s/api/channel/hls/%s.m3u8?%s'
 594             % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
 595             channel_id, 'mp4')
 596         self._prefer_source(formats)
 597
 598         view_count = stream.get('viewers')
 599         timestamp = parse_iso8601(stream.get('created_at'))
 600
 601         channel = stream['channel']
 602         title = self._live_title(channel.get('display_name') or channel.get('name'))
 603         description = channel.get('status')
 604
 605         thumbnails = []
 606         for thumbnail_key, thumbnail_url in stream['preview'].items():
 607             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 608             if not m:
 609                 continue
 610             thumbnails.append({
 611                 'url': thumbnail_url,
 612                 'width': int(m.group('width')),
 613                 'height': int(m.group('height')),
 614             })
 615
 616         return {
 617             'id': compat_str(stream['_id']),
 618             'display_id': channel_id,
 619             'title': title,
 620             'description': description,
 621             'thumbnails': thumbnails,
 622             'uploader': channel.get('display_name'),
 623             'uploader_id': channel.get('name'),
 624             'timestamp': timestamp,
 625             'view_count': view_count,
 626             'formats': formats,
 627             'is_live': True,
 628         }
 629
 630
 631 class TwitchClipsIE(TwitchBaseIE):
 632     IE_NAME = 'twitch:clips'
 633     _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
 634
 635     _TESTS = [{
 636         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
 637         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 638         'info_dict': {
 639             'id': '42850523',
 640             'ext': 'mp4',
 641             'title': 'EA Play 2016 Live from the Novo Theatre',
 642             'thumbnail': r're:^https?://.*\.jpg',
 643             'timestamp': 1465767393,
 644             'upload_date': '20160612',
 645             'creator': 'EA',
 646             'uploader': 'stereotype_',
 647             'uploader_id': '43566419',
 648         },
 649     }, {
 650         # multiple formats
 651         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 652         'only_matching': True,
 653     }]
 654
 655     def _real_extract(self, url):
 656         video_id = self._match_id(url)
 657
 658         status = self._download_json(
 659             'https://clips.twitch.tv/api/v2/clips/%s/status' % video_id,
 660             video_id)
 661
 662         formats = []
 663
 664         for option in status['quality_options']:
 665             if not isinstance(option, dict):
 666                 continue
 667             source = url_or_none(option.get('source'))
 668             if not source:
 669                 continue
 670             formats.append({
 671                 'url': source,
 672                 'format_id': option.get('quality'),
 673                 'height': int_or_none(option.get('quality')),
 674                 'fps': int_or_none(option.get('frame_rate')),
 675             })
 676
 677         self._sort_formats(formats)
 678
 679         info = {
 680             'formats': formats,
 681         }
 682
 683         clip = self._call_api(
 684             'kraken/clips/%s' % video_id, video_id, fatal=False, headers={
 685                 'Accept': 'application/vnd.twitchtv.v5+json',
 686             })
 687
 688         if clip:
 689             quality_key = qualities(('tiny', 'small', 'medium'))
 690             thumbnails = []
 691             thumbnails_dict = clip.get('thumbnails')
 692             if isinstance(thumbnails_dict, dict):
 693                 for thumbnail_id, thumbnail_url in thumbnails_dict.items():
 694                     thumbnails.append({
 695                         'id': thumbnail_id,
 696                         'url': thumbnail_url,
 697                         'preference': quality_key(thumbnail_id),
 698                     })
 699
 700             info.update({
 701                 'id': clip.get('tracking_id') or video_id,
 702                 'title': clip.get('title') or video_id,
 703                 'duration': float_or_none(clip.get('duration')),
 704                 'views': int_or_none(clip.get('views')),
 705                 'timestamp': unified_timestamp(clip.get('created_at')),
 706                 'thumbnails': thumbnails,
 707                 'creator': try_get(clip, lambda x: x['broadcaster']['display_name'], compat_str),
 708                 'uploader': try_get(clip, lambda x: x['curator']['display_name'], compat_str),
 709                 'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
 710             })
 711         else:
 712             info.update({
 713                 'title': video_id,
 714                 'id': video_id,
 715             })
 716
 717         return info