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