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 import json
   8
   9 from .common import InfoExtractor
  10 from ..compat import (
  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     url_or_none,
  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_FORM_URL = 'https://www.twitch.tv/login'
  40     _LOGIN_POST_URL = 'https://passport.twitch.tv/login'
  41     _CLIENT_ID = 'kimne78kx3ncx6brgo4mv6wki5h1ko'
  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         headers = kwargs.get('headers', {}).copy()
  55         headers['Client-ID'] = self._CLIENT_ID
  56         kwargs['headers'] = headers
  57         response = self._download_json(
  58             '%s/%s' % (self._API_BASE, path), item_id,
  59             *args, **compat_kwargs(kwargs))
  60         self._handle_error(response)
  61         return response
  62
  63     def _real_initialize(self):
  64         self._login()
  65
  66     def _login(self):
  67         username, password = self._get_login_info()
  68         if username is None:
  69             return
  70
  71         def fail(message):
  72             raise ExtractorError(
  73                 'Unable to login. Twitch said: %s' % message, expected=True)
  74
  75         def login_step(page, urlh, note, data):
  76             form = self._hidden_inputs(page)
  77             form.update(data)
  78
  79             page_url = urlh.geturl()
  80             post_url = self._search_regex(
  81                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  82                 'post url', default=self._LOGIN_POST_URL, group='url')
  83             post_url = urljoin(page_url, post_url)
  84
  85             headers = {
  86                 'Referer': page_url,
  87                 'Origin': page_url,
  88                 'Content-Type': 'text/plain;charset=UTF-8',
  89             }
  90
  91             response = self._download_json(
  92                 post_url, None, note, data=json.dumps(form).encode(),
  93                 headers=headers, expected_status=400)
  94             error = response.get('error_description') or response.get('error_code')
  95             if error:
  96                 fail(error)
  97
  98             if 'Authenticated successfully' in response.get('message', ''):
  99                 return None, None
 100
 101             redirect_url = urljoin(
 102                 post_url,
 103                 response.get('redirect') or response['redirect_path'])
 104             return self._download_webpage_handle(
 105                 redirect_url, None, 'Downloading login redirect page',
 106                 headers=headers)
 107
 108         login_page, handle = self._download_webpage_handle(
 109             self._LOGIN_FORM_URL, None, 'Downloading login page')
 110
 111         # Some TOR nodes and public proxies are blocked completely
 112         if 'blacklist_message' in login_page:
 113             fail(clean_html(login_page))
 114
 115         redirect_page, handle = login_step(
 116             login_page, handle, 'Logging in', {
 117                 'username': username,
 118                 'password': password,
 119                 'client_id': self._CLIENT_ID,
 120             })
 121
 122         # Successful login
 123         if not redirect_page:
 124             return
 125
 126         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 127             # TODO: Add mechanism to request an SMS or phone call
 128             tfa_token = self._get_tfa_info('two-factor authentication token')
 129             login_step(redirect_page, handle, 'Submitting TFA token', {
 130                 'authy_token': tfa_token,
 131                 'remember_2fa': 'true',
 132             })
 133
 134     def _prefer_source(self, formats):
 135         try:
 136             source = next(f for f in formats if f['format_id'] == 'Source')
 137             source['quality'] = 10
 138         except StopIteration:
 139             for f in formats:
 140                 if '/chunked/' in f['url']:
 141                     f.update({
 142                         'quality': 10,
 143                         'format_note': 'Source',
 144                     })
 145         self._sort_formats(formats)
 146
 147
 148 class TwitchItemBaseIE(TwitchBaseIE):
 149     def _download_info(self, item, item_id):
 150         return self._extract_info(self._call_api(
 151             'kraken/videos/%s%s' % (item, item_id), item_id,
 152             'Downloading %s info JSON' % self._ITEM_TYPE))
 153
 154     def _extract_media(self, item_id):
 155         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 156         response = self._call_api(
 157             'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id,
 158             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 159         entries = []
 160         chunks = response['chunks']
 161         qualities = list(chunks.keys())
 162         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 163             formats = []
 164             for fmt_num, fragment_fmt in enumerate(fragment):
 165                 format_id = qualities[fmt_num]
 166                 fmt = {
 167                     'url': fragment_fmt['url'],
 168                     'format_id': format_id,
 169                     'quality': 1 if format_id == 'live' else 0,
 170                 }
 171                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 172                 if m:
 173                     fmt['height'] = int(m.group('height'))
 174                 formats.append(fmt)
 175             self._sort_formats(formats)
 176             entry = dict(info)
 177             entry['id'] = '%s_%d' % (entry['id'], num)
 178             entry['title'] = '%s part %d' % (entry['title'], num)
 179             entry['formats'] = formats
 180             entries.append(entry)
 181         return self.playlist_result(entries, info['id'], info['title'])
 182
 183     def _extract_info(self, info):
 184         status = info.get('status')
 185         if status == 'recording':
 186             is_live = True
 187         elif status == 'recorded':
 188             is_live = False
 189         else:
 190             is_live = None
 191         return {
 192             'id': info['_id'],
 193             'title': info.get('title') or 'Untitled Broadcast',
 194             'description': info.get('description'),
 195             'duration': int_or_none(info.get('length')),
 196             'thumbnail': info.get('preview'),
 197             'uploader': info.get('channel', {}).get('display_name'),
 198             'uploader_id': info.get('channel', {}).get('name'),
 199             'timestamp': parse_iso8601(info.get('recorded_at')),
 200             'view_count': int_or_none(info.get('views')),
 201             'is_live': is_live,
 202         }
 203
 204     def _real_extract(self, url):
 205         return self._extract_media(self._match_id(url))
 206
 207
 208 class TwitchVideoIE(TwitchItemBaseIE):
 209     IE_NAME = 'twitch:video'
 210     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 211     _ITEM_TYPE = 'video'
 212     _ITEM_SHORTCUT = 'a'
 213
 214     _TEST = {
 215         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 216         'info_dict': {
 217             'id': 'a577357806',
 218             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 219         },
 220         'playlist_mincount': 12,
 221         'skip': 'HTTP Error 404: Not Found',
 222     }
 223
 224
 225 class TwitchChapterIE(TwitchItemBaseIE):
 226     IE_NAME = 'twitch:chapter'
 227     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 228     _ITEM_TYPE = 'chapter'
 229     _ITEM_SHORTCUT = 'c'
 230
 231     _TESTS = [{
 232         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 233         'info_dict': {
 234             'id': 'c5285812',
 235             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 236         },
 237         'playlist_mincount': 3,
 238         'skip': 'HTTP Error 404: Not Found',
 239     }, {
 240         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 241         'only_matching': True,
 242     }]
 243
 244
 245 class TwitchVodIE(TwitchItemBaseIE):
 246     IE_NAME = 'twitch:vod'
 247     _VALID_URL = r'''(?x)
 248                     https?://
 249                         (?:
 250                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 251                             player\.twitch\.tv/\?.*?\bvideo=v
 252                         )
 253                         (?P<id>\d+)
 254                     '''
 255     _ITEM_TYPE = 'vod'
 256     _ITEM_SHORTCUT = 'v'
 257
 258     _TESTS = [{
 259         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 260         'info_dict': {
 261             'id': 'v6528877',
 262             'ext': 'mp4',
 263             'title': 'LCK Summer Split - Week 6 Day 1',
 264             'thumbnail': r're:^https?://.*\.jpg$',
 265             'duration': 17208,
 266             'timestamp': 1435131709,
 267             'upload_date': '20150624',
 268             'uploader': 'Riot Games',
 269             'uploader_id': 'riotgames',
 270             'view_count': int,
 271             'start_time': 310,
 272         },
 273         'params': {
 274             # m3u8 download
 275             'skip_download': True,
 276         },
 277     }, {
 278         # Untitled broadcast (title is None)
 279         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 280         'info_dict': {
 281             'id': 'v11230755',
 282             'ext': 'mp4',
 283             'title': 'Untitled Broadcast',
 284             'thumbnail': r're:^https?://.*\.jpg$',
 285             'duration': 1638,
 286             'timestamp': 1439746708,
 287             'upload_date': '20150816',
 288             'uploader': 'BelkAO_o',
 289             'uploader_id': 'belkao_o',
 290             'view_count': int,
 291         },
 292         'params': {
 293             # m3u8 download
 294             'skip_download': True,
 295         },
 296         'skip': 'HTTP Error 404: Not Found',
 297     }, {
 298         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 299         'only_matching': True,
 300     }, {
 301         'url': 'https://www.twitch.tv/videos/6528877',
 302         'only_matching': True,
 303     }, {
 304         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 305         'only_matching': True,
 306     }, {
 307         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 308         'only_matching': True,
 309     }]
 310
 311     def _real_extract(self, url):
 312         item_id = self._match_id(url)
 313
 314         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 315         access_token = self._call_api(
 316             'api/vods/%s/access_token' % item_id, item_id,
 317             'Downloading %s access token' % self._ITEM_TYPE)
 318
 319         formats = self._extract_m3u8_formats(
 320             '%s/vod/%s.m3u8?%s' % (
 321                 self._USHER_BASE, item_id,
 322                 compat_urllib_parse_urlencode({
 323                     'allow_source': 'true',
 324                     'allow_audio_only': 'true',
 325                     'allow_spectre': 'true',
 326                     'player': 'twitchweb',
 327                     'nauth': access_token['token'],
 328                     'nauthsig': access_token['sig'],
 329                 })),
 330             item_id, 'mp4', entry_protocol='m3u8_native')
 331
 332         self._prefer_source(formats)
 333         info['formats'] = formats
 334
 335         parsed_url = compat_urllib_parse_urlparse(url)
 336         query = compat_parse_qs(parsed_url.query)
 337         if 't' in query:
 338             info['start_time'] = parse_duration(query['t'][0])
 339
 340         if info.get('timestamp') is not None:
 341             info['subtitles'] = {
 342                 'rechat': [{
 343                     'url': update_url_query(
 344                         'https://rechat.twitch.tv/rechat-messages', {
 345                             'video_id': 'v%s' % item_id,
 346                             'start': info['timestamp'],
 347                         }),
 348                     'ext': 'json',
 349                 }],
 350             }
 351
 352         return info
 353
 354
 355 class TwitchPlaylistBaseIE(TwitchBaseIE):
 356     _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
 357     _PAGE_LIMIT = 100
 358
 359     def _extract_playlist(self, channel_id):
 360         info = self._call_api(
 361             'kraken/channels/%s' % channel_id,
 362             channel_id, 'Downloading channel info JSON')
 363         channel_name = info.get('display_name') or info.get('name')
 364         entries = []
 365         offset = 0
 366         limit = self._PAGE_LIMIT
 367         broken_paging_detected = False
 368         counter_override = None
 369         for counter in itertools.count(1):
 370             response = self._call_api(
 371                 self._PLAYLIST_PATH % (channel_id, offset, limit),
 372                 channel_id,
 373                 'Downloading %s JSON page %s'
 374                 % (self._PLAYLIST_TYPE, counter_override or counter))
 375             page_entries = self._extract_playlist_page(response)
 376             if not page_entries:
 377                 break
 378             total = int_or_none(response.get('_total'))
 379             # Since the beginning of March 2016 twitch's paging mechanism
 380             # is completely broken on the twitch side. It simply ignores
 381             # a limit and returns the whole offset number of videos.
 382             # Working around by just requesting all videos at once.
 383             # Upd: pagination bug was fixed by twitch on 15.03.2016.
 384             if not broken_paging_detected and total and len(page_entries) > limit:
 385                 self.report_warning(
 386                     'Twitch pagination is broken on twitch side, requesting all videos at once',
 387                     channel_id)
 388                 broken_paging_detected = True
 389                 offset = total
 390                 counter_override = '(all at once)'
 391                 continue
 392             entries.extend(page_entries)
 393             if broken_paging_detected or total and len(page_entries) >= total:
 394                 break
 395             offset += limit
 396         return self.playlist_result(
 397             [self._make_url_result(entry) for entry in orderedSet(entries)],
 398             channel_id, channel_name)
 399
 400     def _make_url_result(self, url):
 401         try:
 402             video_id = 'v%s' % TwitchVodIE._match_id(url)
 403             return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id)
 404         except AssertionError:
 405             return self.url_result(url)
 406
 407     def _extract_playlist_page(self, response):
 408         videos = response.get('videos')
 409         return [video['url'] for video in videos] if videos else []
 410
 411     def _real_extract(self, url):
 412         return self._extract_playlist(self._match_id(url))
 413
 414
 415 class TwitchProfileIE(TwitchPlaylistBaseIE):
 416     IE_NAME = 'twitch:profile'
 417     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 418     _PLAYLIST_TYPE = 'profile'
 419
 420     _TESTS = [{
 421         'url': 'http://www.twitch.tv/vanillatv/profile',
 422         'info_dict': {
 423             'id': 'vanillatv',
 424             'title': 'VanillaTV',
 425         },
 426         'playlist_mincount': 412,
 427     }, {
 428         'url': 'http://m.twitch.tv/vanillatv/profile',
 429         'only_matching': True,
 430     }]
 431
 432
 433 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
 434     _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
 435     _VALID_URL_VIDEOS_FILTERS = r'\?(?:.*?[&;])??filter=%s'
 436     _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
 437
 438
 439 class TwitchAllVideosIE(TwitchVideosBaseIE):
 440     IE_NAME = 'twitch:videos:all'
 441     _VALID_URL = '%s/?(?:(?:%s)|$)' % (
 442         TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE,
 443         TwitchVideosBaseIE._VALID_URL_VIDEOS_FILTERS % 'all'
 444     )
 445     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
 446     _PLAYLIST_TYPE = 'all videos'
 447
 448     _TESTS = [{
 449         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=time',
 450         'info_dict': {
 451             'id': 'spamfish',
 452             'title': 'Spamfish',
 453         },
 454         'playlist_mincount': 869,
 455     }, {
 456         'url': 'https://m.twitch.tv/spamfish/videos/',
 457         'only_matching': True,
 458     }]
 459
 460
 461 class TwitchUploadsIE(TwitchVideosBaseIE):
 462     IE_NAME = 'twitch:videos:uploads'
 463     _VALID_URL = '%s/?(?:%s)' % (
 464         TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE,
 465         TwitchVideosBaseIE._VALID_URL_VIDEOS_FILTERS % 'uploads'
 466     )
 467     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
 468     _PLAYLIST_TYPE = 'uploads'
 469
 470     _TESTS = [{
 471         'url': 'https://www.twitch.tv/spamfish/videos?filter=uploads&sort=time',
 472         'info_dict': {
 473             'id': 'spamfish',
 474             'title': 'Spamfish',
 475         },
 476         'playlist_mincount': 0,
 477     }]
 478
 479
 480 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
 481     IE_NAME = 'twitch:videos:past-broadcasts'
 482     _VALID_URL = '%s/?(?:%s)' % (
 483         TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE,
 484         TwitchVideosBaseIE._VALID_URL_VIDEOS_FILTERS % 'archives'
 485     )
 486     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
 487     _PLAYLIST_TYPE = 'past broadcasts'
 488
 489     _TESTS = [{
 490         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives&sort=time',
 491         'info_dict': {
 492             'id': 'spamfish',
 493             'title': 'Spamfish',
 494         },
 495         'playlist_mincount': 0,
 496     }]
 497
 498
 499 class TwitchHighlightsIE(TwitchVideosBaseIE):
 500     IE_NAME = 'twitch:videos:highlights'
 501     _VALID_URL = '%s/?(?:%s)' % (
 502         TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE,
 503         TwitchVideosBaseIE._VALID_URL_VIDEOS_FILTERS % 'highlights'
 504     )
 505     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
 506     _PLAYLIST_TYPE = 'highlights'
 507
 508     _TESTS = [{
 509         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights&sort=views',
 510         'info_dict': {
 511             'id': 'spamfish',
 512             'title': 'Spamfish',
 513         },
 514         'playlist_mincount': 805,
 515     }]
 516
 517
 518 class TwitchStreamIE(TwitchBaseIE):
 519     IE_NAME = 'twitch:stream'
 520     _VALID_URL = r'''(?x)
 521                     https?://
 522                         (?:
 523                             (?:(?:www|go|m)\.)?twitch\.tv/|
 524                             player\.twitch\.tv/\?.*?\bchannel=
 525                         )
 526                         (?P<id>[^/#?]+)
 527                     '''
 528
 529     _TESTS = [{
 530         'url': 'http://www.twitch.tv/shroomztv',
 531         'info_dict': {
 532             'id': '12772022048',
 533             'display_id': 'shroomztv',
 534             'ext': 'mp4',
 535             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 536             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 537             'is_live': True,
 538             'timestamp': 1421928037,
 539             'upload_date': '20150122',
 540             'uploader': 'ShroomzTV',
 541             'uploader_id': 'shroomztv',
 542             'view_count': int,
 543         },
 544         'params': {
 545             # m3u8 download
 546             'skip_download': True,
 547         },
 548     }, {
 549         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 550         'only_matching': True,
 551     }, {
 552         'url': 'https://player.twitch.tv/?channel=lotsofs',
 553         'only_matching': True,
 554     }, {
 555         'url': 'https://go.twitch.tv/food',
 556         'only_matching': True,
 557     }, {
 558         'url': 'https://m.twitch.tv/food',
 559         'only_matching': True,
 560     }]
 561
 562     @classmethod
 563     def suitable(cls, url):
 564         return (False
 565                 if any(ie.suitable(url) for ie in (
 566                     TwitchVideoIE,
 567                     TwitchChapterIE,
 568                     TwitchVodIE,
 569                     TwitchProfileIE,
 570                     TwitchAllVideosIE,
 571                     TwitchUploadsIE,
 572                     TwitchPastBroadcastsIE,
 573                     TwitchHighlightsIE,
 574                     TwitchClipsIE))
 575                 else super(TwitchStreamIE, cls).suitable(url))
 576
 577     def _real_extract(self, url):
 578         channel_id = self._match_id(url)
 579
 580         stream = self._call_api(
 581             'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
 582             'Downloading stream JSON').get('stream')
 583
 584         if not stream:
 585             raise ExtractorError('%s is offline' % channel_id, expected=True)
 586
 587         # Channel name may be typed if different case than the original channel name
 588         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 589         # an invalid m3u8 URL. Working around by use of original channel name from stream
 590         # JSON and fallback to lowercase if it's not available.
 591         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 592
 593         access_token = self._call_api(
 594             'api/channels/%s/access_token' % channel_id, channel_id,
 595             'Downloading channel access token')
 596
 597         query = {
 598             'allow_source': 'true',
 599             'allow_audio_only': 'true',
 600             'allow_spectre': 'true',
 601             'p': random.randint(1000000, 10000000),
 602             'player': 'twitchweb',
 603             'segment_preference': '4',
 604             'sig': access_token['sig'].encode('utf-8'),
 605             'token': access_token['token'].encode('utf-8'),
 606         }
 607         formats = self._extract_m3u8_formats(
 608             '%s/api/channel/hls/%s.m3u8?%s'
 609             % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
 610             channel_id, 'mp4')
 611         self._prefer_source(formats)
 612
 613         view_count = stream.get('viewers')
 614         timestamp = parse_iso8601(stream.get('created_at'))
 615
 616         channel = stream['channel']
 617         title = self._live_title(channel.get('display_name') or channel.get('name'))
 618         description = channel.get('status')
 619
 620         thumbnails = []
 621         for thumbnail_key, thumbnail_url in stream['preview'].items():
 622             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 623             if not m:
 624                 continue
 625             thumbnails.append({
 626                 'url': thumbnail_url,
 627                 'width': int(m.group('width')),
 628                 'height': int(m.group('height')),
 629             })
 630
 631         return {
 632             'id': compat_str(stream['_id']),
 633             'display_id': channel_id,
 634             'title': title,
 635             'description': description,
 636             'thumbnails': thumbnails,
 637             'uploader': channel.get('display_name'),
 638             'uploader_id': channel.get('name'),
 639             'timestamp': timestamp,
 640             'view_count': view_count,
 641             'formats': formats,
 642             'is_live': True,
 643         }
 644
 645
 646 class TwitchClipsIE(TwitchBaseIE):
 647     IE_NAME = 'twitch:clips'
 648     _VALID_URL = r'https?://(?:clips\.twitch\.tv/(?:[^/]+/)*|(?:www\.)?twitch\.tv/[^/]+/clip/)(?P<id>[^/?#&]+)'
 649
 650     _TESTS = [{
 651         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
 652         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 653         'info_dict': {
 654             'id': '42850523',
 655             'ext': 'mp4',
 656             'title': 'EA Play 2016 Live from the Novo Theatre',
 657             'thumbnail': r're:^https?://.*\.jpg',
 658             'timestamp': 1465767393,
 659             'upload_date': '20160612',
 660             'creator': 'EA',
 661             'uploader': 'stereotype_',
 662             'uploader_id': '43566419',
 663         },
 664     }, {
 665         # multiple formats
 666         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 667         'only_matching': True,
 668     }, {
 669         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
 670         'only_matching': True,
 671     }]
 672
 673     def _real_extract(self, url):
 674         video_id = self._match_id(url)
 675
 676         status = self._download_json(
 677             'https://clips.twitch.tv/api/v2/clips/%s/status' % video_id,
 678             video_id)
 679
 680         formats = []
 681
 682         for option in status['quality_options']:
 683             if not isinstance(option, dict):
 684                 continue
 685             source = url_or_none(option.get('source'))
 686             if not source:
 687                 continue
 688             formats.append({
 689                 'url': source,
 690                 'format_id': option.get('quality'),
 691                 'height': int_or_none(option.get('quality')),
 692                 'fps': int_or_none(option.get('frame_rate')),
 693             })
 694
 695         self._sort_formats(formats)
 696
 697         info = {
 698             'formats': formats,
 699         }
 700
 701         clip = self._call_api(
 702             'kraken/clips/%s' % video_id, video_id, fatal=False, headers={
 703                 'Accept': 'application/vnd.twitchtv.v5+json',
 704             })
 705
 706         if clip:
 707             quality_key = qualities(('tiny', 'small', 'medium'))
 708             thumbnails = []
 709             thumbnails_dict = clip.get('thumbnails')
 710             if isinstance(thumbnails_dict, dict):
 711                 for thumbnail_id, thumbnail_url in thumbnails_dict.items():
 712                     thumbnails.append({
 713                         'id': thumbnail_id,
 714                         'url': thumbnail_url,
 715                         'preference': quality_key(thumbnail_id),
 716                     })
 717
 718             info.update({
 719                 'id': clip.get('tracking_id') or video_id,
 720                 'title': clip.get('title') or video_id,
 721                 'duration': float_or_none(clip.get('duration')),
 722                 'views': int_or_none(clip.get('views')),
 723                 'timestamp': unified_timestamp(clip.get('created_at')),
 724                 'thumbnails': thumbnails,
 725                 'creator': try_get(clip, lambda x: x['broadcaster']['display_name'], compat_str),
 726                 'uploader': try_get(clip, lambda x: x['curator']['display_name'], compat_str),
 727                 'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
 728             })
 729         else:
 730             info.update({
 731                 'title': video_id,
 732                 'id': video_id,
 733             })
 734
 735         return info