yt_dlp/extractor/twitch.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import collections
   5 import itertools
   6 import json
   7 import random
   8 import re
   9
  10 from .common import InfoExtractor
  11 from ..compat import (
  12     compat_parse_qs,
  13     compat_str,
  14     compat_urlparse,
  15     compat_urllib_parse_urlencode,
  16     compat_urllib_parse_urlparse,
  17 )
  18 from ..utils import (
  19     clean_html,
  20     dict_get,
  21     ExtractorError,
  22     float_or_none,
  23     int_or_none,
  24     parse_duration,
  25     parse_iso8601,
  26     qualities,
  27     try_get,
  28     unified_timestamp,
  29     update_url_query,
  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_FORM_URL = 'https://www.twitch.tv/login'
  41     _LOGIN_POST_URL = 'https://passport.twitch.tv/login'
  42     _CLIENT_ID = 'kimne78kx3ncx6brgo4mv6wki5h1ko'
  43     _NETRC_MACHINE = 'twitch'
  44
  45     _OPERATION_HASHES = {
  46         'CollectionSideBar': '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14',
  47         'FilterableVideoTower_Videos': 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb',
  48         'ClipsCards__User': 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777',
  49         'ChannelCollectionsContent': '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84',
  50         'StreamMetadata': '1c719a40e481453e5c48d9bb585d971b8b372f8ebb105b17076722264dfa5b3e',
  51         'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
  52         'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
  53         'VideoMetadata': '226edb3e692509f727fd56821f5653c05740242c82b0388883e0c0e75dcbf687',
  54     }
  55
  56     def _real_initialize(self):
  57         self._login()
  58
  59     def _login(self):
  60         username, password = self._get_login_info()
  61         if username is None:
  62             return
  63
  64         def fail(message):
  65             raise ExtractorError(
  66                 'Unable to login. Twitch said: %s' % message, expected=True)
  67
  68         def login_step(page, urlh, note, data):
  69             form = self._hidden_inputs(page)
  70             form.update(data)
  71
  72             page_url = urlh.geturl()
  73             post_url = self._search_regex(
  74                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  75                 'post url', default=self._LOGIN_POST_URL, group='url')
  76             post_url = urljoin(page_url, post_url)
  77
  78             headers = {
  79                 'Referer': page_url,
  80                 'Origin': 'https://www.twitch.tv',
  81                 'Content-Type': 'text/plain;charset=UTF-8',
  82             }
  83
  84             response = self._download_json(
  85                 post_url, None, note, data=json.dumps(form).encode(),
  86                 headers=headers, expected_status=400)
  87             error = dict_get(response, ('error', 'error_description', 'error_code'))
  88             if error:
  89                 fail(error)
  90
  91             if 'Authenticated successfully' in response.get('message', ''):
  92                 return None, None
  93
  94             redirect_url = urljoin(
  95                 post_url,
  96                 response.get('redirect') or response['redirect_path'])
  97             return self._download_webpage_handle(
  98                 redirect_url, None, 'Downloading login redirect page',
  99                 headers=headers)
 100
 101         login_page, handle = self._download_webpage_handle(
 102             self._LOGIN_FORM_URL, None, 'Downloading login page')
 103
 104         # Some TOR nodes and public proxies are blocked completely
 105         if 'blacklist_message' in login_page:
 106             fail(clean_html(login_page))
 107
 108         redirect_page, handle = login_step(
 109             login_page, handle, 'Logging in', {
 110                 'username': username,
 111                 'password': password,
 112                 'client_id': self._CLIENT_ID,
 113             })
 114
 115         # Successful login
 116         if not redirect_page:
 117             return
 118
 119         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 120             # TODO: Add mechanism to request an SMS or phone call
 121             tfa_token = self._get_tfa_info('two-factor authentication token')
 122             login_step(redirect_page, handle, 'Submitting TFA token', {
 123                 'authy_token': tfa_token,
 124                 'remember_2fa': 'true',
 125             })
 126
 127     def _prefer_source(self, formats):
 128         try:
 129             source = next(f for f in formats if f['format_id'] == 'Source')
 130             source['quality'] = 10
 131         except StopIteration:
 132             for f in formats:
 133                 if '/chunked/' in f['url']:
 134                     f.update({
 135                         'quality': 10,
 136                         'format_note': 'Source',
 137                     })
 138         self._sort_formats(formats)
 139
 140     def _download_base_gql(self, video_id, ops, note, fatal=True):
 141         headers = {
 142             'Content-Type': 'text/plain;charset=UTF-8',
 143             'Client-ID': self._CLIENT_ID,
 144         }
 145         gql_auth = self._get_cookies('https://gql.twitch.tv').get('auth-token')
 146         if gql_auth:
 147             headers['Authorization'] = 'OAuth ' + gql_auth.value
 148         return self._download_json(
 149             'https://gql.twitch.tv/gql', video_id, note,
 150             data=json.dumps(ops).encode(),
 151             headers=headers, fatal=fatal)
 152
 153     def _download_gql(self, video_id, ops, note, fatal=True):
 154         for op in ops:
 155             op['extensions'] = {
 156                 'persistedQuery': {
 157                     'version': 1,
 158                     'sha256Hash': self._OPERATION_HASHES[op['operationName']],
 159                 }
 160             }
 161         return self._download_base_gql(video_id, ops, note)
 162
 163     def _download_access_token(self, video_id, token_kind, param_name):
 164         method = '%sPlaybackAccessToken' % token_kind
 165         ops = {
 166             'query': '''{
 167               %s(
 168                 %s: "%s",
 169                 params: {
 170                   platform: "web",
 171                   playerBackend: "mediaplayer",
 172                   playerType: "site"
 173                 }
 174               )
 175               {
 176                 value
 177                 signature
 178               }
 179             }''' % (method, param_name, video_id),
 180         }
 181         return self._download_base_gql(
 182             video_id, ops,
 183             'Downloading %s access token GraphQL' % token_kind)['data'][method]
 184
 185
 186 class TwitchVodIE(TwitchBaseIE):
 187     IE_NAME = 'twitch:vod'
 188     _VALID_URL = r'''(?x)
 189                     https?://
 190                         (?:
 191                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 192                             player\.twitch\.tv/\?.*?\bvideo=v?
 193                         )
 194                         (?P<id>\d+)
 195                     '''
 196
 197     _TESTS = [{
 198         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 199         'info_dict': {
 200             'id': 'v6528877',
 201             'ext': 'mp4',
 202             'title': 'LCK Summer Split - Week 6 Day 1',
 203             'thumbnail': r're:^https?://.*\.jpg$',
 204             'duration': 17208,
 205             'timestamp': 1435131734,
 206             'upload_date': '20150624',
 207             'uploader': 'Riot Games',
 208             'uploader_id': 'riotgames',
 209             'view_count': int,
 210             'start_time': 310,
 211         },
 212         'params': {
 213             # m3u8 download
 214             'skip_download': True,
 215         },
 216     }, {
 217         # Untitled broadcast (title is None)
 218         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 219         'info_dict': {
 220             'id': 'v11230755',
 221             'ext': 'mp4',
 222             'title': 'Untitled Broadcast',
 223             'thumbnail': r're:^https?://.*\.jpg$',
 224             'duration': 1638,
 225             'timestamp': 1439746708,
 226             'upload_date': '20150816',
 227             'uploader': 'BelkAO_o',
 228             'uploader_id': 'belkao_o',
 229             'view_count': int,
 230         },
 231         'params': {
 232             # m3u8 download
 233             'skip_download': True,
 234         },
 235         'skip': 'HTTP Error 404: Not Found',
 236     }, {
 237         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 238         'only_matching': True,
 239     }, {
 240         'url': 'https://www.twitch.tv/videos/6528877',
 241         'only_matching': True,
 242     }, {
 243         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 244         'only_matching': True,
 245     }, {
 246         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 247         'only_matching': True,
 248     }, {
 249         'url': 'https://player.twitch.tv/?video=480452374',
 250         'only_matching': True,
 251     }]
 252
 253     def _download_info(self, item_id):
 254         data = self._download_gql(
 255             item_id, [{
 256                 'operationName': 'VideoMetadata',
 257                 'variables': {
 258                     'channelLogin': '',
 259                     'videoID': item_id,
 260                 },
 261             }],
 262             'Downloading stream metadata GraphQL')[0]['data']
 263         video = data.get('video')
 264         if video is None:
 265             raise ExtractorError(
 266                 'Video %s does not exist' % item_id, expected=True)
 267         return self._extract_info_gql(video, item_id)
 268
 269     @staticmethod
 270     def _extract_info(info):
 271         status = info.get('status')
 272         if status == 'recording':
 273             is_live = True
 274         elif status == 'recorded':
 275             is_live = False
 276         else:
 277             is_live = None
 278         _QUALITIES = ('small', 'medium', 'large')
 279         quality_key = qualities(_QUALITIES)
 280         thumbnails = []
 281         preview = info.get('preview')
 282         if isinstance(preview, dict):
 283             for thumbnail_id, thumbnail_url in preview.items():
 284                 thumbnail_url = url_or_none(thumbnail_url)
 285                 if not thumbnail_url:
 286                     continue
 287                 if thumbnail_id not in _QUALITIES:
 288                     continue
 289                 thumbnails.append({
 290                     'url': thumbnail_url,
 291                     'preference': quality_key(thumbnail_id),
 292                 })
 293         return {
 294             'id': info['_id'],
 295             'title': info.get('title') or 'Untitled Broadcast',
 296             'description': info.get('description'),
 297             'duration': int_or_none(info.get('length')),
 298             'thumbnails': thumbnails,
 299             'uploader': info.get('channel', {}).get('display_name'),
 300             'uploader_id': info.get('channel', {}).get('name'),
 301             'timestamp': parse_iso8601(info.get('recorded_at')),
 302             'view_count': int_or_none(info.get('views')),
 303             'is_live': is_live,
 304         }
 305
 306     @staticmethod
 307     def _extract_info_gql(info, item_id):
 308         vod_id = info.get('id') or item_id
 309         # id backward compatibility for download archives
 310         if vod_id[0] != 'v':
 311             vod_id = 'v%s' % vod_id
 312         thumbnail = url_or_none(info.get('previewThumbnailURL'))
 313         if thumbnail:
 314             for p in ('width', 'height'):
 315                 thumbnail = thumbnail.replace('{%s}' % p, '0')
 316         return {
 317             'id': vod_id,
 318             'title': info.get('title') or 'Untitled Broadcast',
 319             'description': info.get('description'),
 320             'duration': int_or_none(info.get('lengthSeconds')),
 321             'thumbnail': thumbnail,
 322             'uploader': try_get(info, lambda x: x['owner']['displayName'], compat_str),
 323             'uploader_id': try_get(info, lambda x: x['owner']['login'], compat_str),
 324             'timestamp': unified_timestamp(info.get('publishedAt')),
 325             'view_count': int_or_none(info.get('viewCount')),
 326         }
 327
 328     def _real_extract(self, url):
 329         vod_id = self._match_id(url)
 330
 331         info = self._download_info(vod_id)
 332         access_token = self._download_access_token(vod_id, 'video', 'id')
 333
 334         formats = self._extract_m3u8_formats(
 335             '%s/vod/%s.m3u8?%s' % (
 336                 self._USHER_BASE, vod_id,
 337                 compat_urllib_parse_urlencode({
 338                     'allow_source': 'true',
 339                     'allow_audio_only': 'true',
 340                     'allow_spectre': 'true',
 341                     'player': 'twitchweb',
 342                     'playlist_include_framerate': 'true',
 343                     'nauth': access_token['value'],
 344                     'nauthsig': access_token['signature'],
 345                 })),
 346             vod_id, 'mp4', entry_protocol='m3u8_native')
 347
 348         self._prefer_source(formats)
 349         info['formats'] = formats
 350
 351         parsed_url = compat_urllib_parse_urlparse(url)
 352         query = compat_parse_qs(parsed_url.query)
 353         if 't' in query:
 354             info['start_time'] = parse_duration(query['t'][0])
 355
 356         if info.get('timestamp') is not None:
 357             info['subtitles'] = {
 358                 'rechat': [{
 359                     'url': update_url_query(
 360                         'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, {
 361                             'client_id': self._CLIENT_ID,
 362                         }),
 363                     'ext': 'json',
 364                 }],
 365             }
 366
 367         return info
 368
 369
 370 def _make_video_result(node):
 371     assert isinstance(node, dict)
 372     video_id = node.get('id')
 373     if not video_id:
 374         return
 375     return {
 376         '_type': 'url_transparent',
 377         'ie_key': TwitchVodIE.ie_key(),
 378         'id': 'v' + video_id,
 379         'url': 'https://www.twitch.tv/videos/%s' % video_id,
 380         'title': node.get('title'),
 381         'thumbnail': node.get('previewThumbnailURL'),
 382         'duration': float_or_none(node.get('lengthSeconds')),
 383         'view_count': int_or_none(node.get('viewCount')),
 384     }
 385
 386
 387 class TwitchCollectionIE(TwitchBaseIE):
 388     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P<id>[^/]+)'
 389
 390     _TESTS = [{
 391         'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ',
 392         'info_dict': {
 393             'id': 'wlDCoH0zEBZZbQ',
 394             'title': 'Overthrow Nook, capitalism for children',
 395         },
 396         'playlist_mincount': 13,
 397     }]
 398
 399     _OPERATION_NAME = 'CollectionSideBar'
 400
 401     def _real_extract(self, url):
 402         collection_id = self._match_id(url)
 403         collection = self._download_gql(
 404             collection_id, [{
 405                 'operationName': self._OPERATION_NAME,
 406                 'variables': {'collectionID': collection_id},
 407             }],
 408             'Downloading collection GraphQL')[0]['data']['collection']
 409         title = collection.get('title')
 410         entries = []
 411         for edge in collection['items']['edges']:
 412             if not isinstance(edge, dict):
 413                 continue
 414             node = edge.get('node')
 415             if not isinstance(node, dict):
 416                 continue
 417             video = _make_video_result(node)
 418             if video:
 419                 entries.append(video)
 420         return self.playlist_result(
 421             entries, playlist_id=collection_id, playlist_title=title)
 422
 423
 424 class TwitchPlaylistBaseIE(TwitchBaseIE):
 425     _PAGE_LIMIT = 100
 426
 427     def _entries(self, channel_name, *args):
 428         cursor = None
 429         variables_common = self._make_variables(channel_name, *args)
 430         entries_key = '%ss' % self._ENTRY_KIND
 431         for page_num in itertools.count(1):
 432             variables = variables_common.copy()
 433             variables['limit'] = self._PAGE_LIMIT
 434             if cursor:
 435                 variables['cursor'] = cursor
 436             page = self._download_gql(
 437                 channel_name, [{
 438                     'operationName': self._OPERATION_NAME,
 439                     'variables': variables,
 440                 }],
 441                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
 442                 fatal=False)
 443             if not page:
 444                 break
 445             edges = try_get(
 446                 page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
 447             if not edges:
 448                 break
 449             for edge in edges:
 450                 if not isinstance(edge, dict):
 451                     continue
 452                 if edge.get('__typename') != self._EDGE_KIND:
 453                     continue
 454                 node = edge.get('node')
 455                 if not isinstance(node, dict):
 456                     continue
 457                 if node.get('__typename') != self._NODE_KIND:
 458                     continue
 459                 entry = self._extract_entry(node)
 460                 if entry:
 461                     cursor = edge.get('cursor')
 462                     yield entry
 463             if not cursor or not isinstance(cursor, compat_str):
 464                 break
 465
 466
 467 class TwitchVideosIE(TwitchPlaylistBaseIE):
 468     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:videos|profile)'
 469
 470     _TESTS = [{
 471         # All Videos sorted by Date
 472         'url': 'https://www.twitch.tv/spamfish/videos?filter=all',
 473         'info_dict': {
 474             'id': 'spamfish',
 475             'title': 'spamfish - All Videos sorted by Date',
 476         },
 477         'playlist_mincount': 924,
 478     }, {
 479         # All Videos sorted by Popular
 480         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views',
 481         'info_dict': {
 482             'id': 'spamfish',
 483             'title': 'spamfish - All Videos sorted by Popular',
 484         },
 485         'playlist_mincount': 931,
 486     }, {
 487         # Past Broadcasts sorted by Date
 488         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives',
 489         'info_dict': {
 490             'id': 'spamfish',
 491             'title': 'spamfish - Past Broadcasts sorted by Date',
 492         },
 493         'playlist_mincount': 27,
 494     }, {
 495         # Highlights sorted by Date
 496         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights',
 497         'info_dict': {
 498             'id': 'spamfish',
 499             'title': 'spamfish - Highlights sorted by Date',
 500         },
 501         'playlist_mincount': 901,
 502     }, {
 503         # Uploads sorted by Date
 504         'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time',
 505         'info_dict': {
 506             'id': 'esl_csgo',
 507             'title': 'esl_csgo - Uploads sorted by Date',
 508         },
 509         'playlist_mincount': 5,
 510     }, {
 511         # Past Premieres sorted by Date
 512         'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres',
 513         'info_dict': {
 514             'id': 'spamfish',
 515             'title': 'spamfish - Past Premieres sorted by Date',
 516         },
 517         'playlist_mincount': 1,
 518     }, {
 519         'url': 'https://www.twitch.tv/spamfish/videos/all',
 520         'only_matching': True,
 521     }, {
 522         'url': 'https://m.twitch.tv/spamfish/videos/all',
 523         'only_matching': True,
 524     }, {
 525         'url': 'https://www.twitch.tv/spamfish/videos',
 526         'only_matching': True,
 527     }]
 528
 529     Broadcast = collections.namedtuple('Broadcast', ['type', 'label'])
 530
 531     _DEFAULT_BROADCAST = Broadcast(None, 'All Videos')
 532     _BROADCASTS = {
 533         'archives': Broadcast('ARCHIVE', 'Past Broadcasts'),
 534         'highlights': Broadcast('HIGHLIGHT', 'Highlights'),
 535         'uploads': Broadcast('UPLOAD', 'Uploads'),
 536         'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'),
 537         'all': _DEFAULT_BROADCAST,
 538     }
 539
 540     _DEFAULT_SORTED_BY = 'Date'
 541     _SORTED_BY = {
 542         'time': _DEFAULT_SORTED_BY,
 543         'views': 'Popular',
 544     }
 545
 546     _OPERATION_NAME = 'FilterableVideoTower_Videos'
 547     _ENTRY_KIND = 'video'
 548     _EDGE_KIND = 'VideoEdge'
 549     _NODE_KIND = 'Video'
 550
 551     @classmethod
 552     def suitable(cls, url):
 553         return (False
 554                 if any(ie.suitable(url) for ie in (
 555                     TwitchVideosClipsIE,
 556                     TwitchVideosCollectionsIE))
 557                 else super(TwitchVideosIE, cls).suitable(url))
 558
 559     @staticmethod
 560     def _make_variables(channel_name, broadcast_type, sort):
 561         return {
 562             'channelOwnerLogin': channel_name,
 563             'broadcastType': broadcast_type,
 564             'videoSort': sort.upper(),
 565         }
 566
 567     @staticmethod
 568     def _extract_entry(node):
 569         return _make_video_result(node)
 570
 571     def _real_extract(self, url):
 572         channel_name = self._match_id(url)
 573         qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 574         filter = qs.get('filter', ['all'])[0]
 575         sort = qs.get('sort', ['time'])[0]
 576         broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST)
 577         return self.playlist_result(
 578             self._entries(channel_name, broadcast.type, sort),
 579             playlist_id=channel_name,
 580             playlist_title='%s - %s sorted by %s'
 581             % (channel_name, broadcast.label,
 582                self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY)))
 583
 584
 585 class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
 586     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)'
 587
 588     _TESTS = [{
 589         # Clips
 590         'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all',
 591         'info_dict': {
 592             'id': 'vanillatv',
 593             'title': 'vanillatv - Clips Top All',
 594         },
 595         'playlist_mincount': 1,
 596     }, {
 597         'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d',
 598         'only_matching': True,
 599     }]
 600
 601     Clip = collections.namedtuple('Clip', ['filter', 'label'])
 602
 603     _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D')
 604     _RANGE = {
 605         '24hr': Clip('LAST_DAY', 'Top 24H'),
 606         '7d': _DEFAULT_CLIP,
 607         '30d': Clip('LAST_MONTH', 'Top 30D'),
 608         'all': Clip('ALL_TIME', 'Top All'),
 609     }
 610
 611     # NB: values other than 20 result in skipped videos
 612     _PAGE_LIMIT = 20
 613
 614     _OPERATION_NAME = 'ClipsCards__User'
 615     _ENTRY_KIND = 'clip'
 616     _EDGE_KIND = 'ClipEdge'
 617     _NODE_KIND = 'Clip'
 618
 619     @staticmethod
 620     def _make_variables(channel_name, filter):
 621         return {
 622             'login': channel_name,
 623             'criteria': {
 624                 'filter': filter,
 625             },
 626         }
 627
 628     @staticmethod
 629     def _extract_entry(node):
 630         assert isinstance(node, dict)
 631         clip_url = url_or_none(node.get('url'))
 632         if not clip_url:
 633             return
 634         return {
 635             '_type': 'url_transparent',
 636             'ie_key': TwitchClipsIE.ie_key(),
 637             'id': node.get('id'),
 638             'url': clip_url,
 639             'title': node.get('title'),
 640             'thumbnail': node.get('thumbnailURL'),
 641             'duration': float_or_none(node.get('durationSeconds')),
 642             'timestamp': unified_timestamp(node.get('createdAt')),
 643             'view_count': int_or_none(node.get('viewCount')),
 644             'language': node.get('language'),
 645         }
 646
 647     def _real_extract(self, url):
 648         channel_name = self._match_id(url)
 649         qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 650         range = qs.get('range', ['7d'])[0]
 651         clip = self._RANGE.get(range, self._DEFAULT_CLIP)
 652         return self.playlist_result(
 653             self._entries(channel_name, clip.filter),
 654             playlist_id=channel_name,
 655             playlist_title='%s - Clips %s' % (channel_name, clip.label))
 656
 657
 658 class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
 659     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/videos/*?\?.*?\bfilter=collections'
 660
 661     _TESTS = [{
 662         # Collections
 663         'url': 'https://www.twitch.tv/spamfish/videos?filter=collections',
 664         'info_dict': {
 665             'id': 'spamfish',
 666             'title': 'spamfish - Collections',
 667         },
 668         'playlist_mincount': 3,
 669     }]
 670
 671     _OPERATION_NAME = 'ChannelCollectionsContent'
 672     _ENTRY_KIND = 'collection'
 673     _EDGE_KIND = 'CollectionsItemEdge'
 674     _NODE_KIND = 'Collection'
 675
 676     @staticmethod
 677     def _make_variables(channel_name):
 678         return {
 679             'ownerLogin': channel_name,
 680         }
 681
 682     @staticmethod
 683     def _extract_entry(node):
 684         assert isinstance(node, dict)
 685         collection_id = node.get('id')
 686         if not collection_id:
 687             return
 688         return {
 689             '_type': 'url_transparent',
 690             'ie_key': TwitchCollectionIE.ie_key(),
 691             'id': collection_id,
 692             'url': 'https://www.twitch.tv/collections/%s' % collection_id,
 693             'title': node.get('title'),
 694             'thumbnail': node.get('thumbnailURL'),
 695             'duration': float_or_none(node.get('lengthSeconds')),
 696             'timestamp': unified_timestamp(node.get('updatedAt')),
 697             'view_count': int_or_none(node.get('viewCount')),
 698         }
 699
 700     def _real_extract(self, url):
 701         channel_name = self._match_id(url)
 702         return self.playlist_result(
 703             self._entries(channel_name), playlist_id=channel_name,
 704             playlist_title='%s - Collections' % channel_name)
 705
 706
 707 class TwitchStreamIE(TwitchBaseIE):
 708     IE_NAME = 'twitch:stream'
 709     _VALID_URL = r'''(?x)
 710                     https?://
 711                         (?:
 712                             (?:(?:www|go|m)\.)?twitch\.tv/|
 713                             player\.twitch\.tv/\?.*?\bchannel=
 714                         )
 715                         (?P<id>[^/#?]+)
 716                     '''
 717
 718     _TESTS = [{
 719         'url': 'http://www.twitch.tv/shroomztv',
 720         'info_dict': {
 721             'id': '12772022048',
 722             'display_id': 'shroomztv',
 723             'ext': 'mp4',
 724             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 725             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 726             'is_live': True,
 727             'timestamp': 1421928037,
 728             'upload_date': '20150122',
 729             'uploader': 'ShroomzTV',
 730             'uploader_id': 'shroomztv',
 731             'view_count': int,
 732         },
 733         'params': {
 734             # m3u8 download
 735             'skip_download': True,
 736         },
 737     }, {
 738         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 739         'only_matching': True,
 740     }, {
 741         'url': 'https://player.twitch.tv/?channel=lotsofs',
 742         'only_matching': True,
 743     }, {
 744         'url': 'https://go.twitch.tv/food',
 745         'only_matching': True,
 746     }, {
 747         'url': 'https://m.twitch.tv/food',
 748         'only_matching': True,
 749     }]
 750
 751     @classmethod
 752     def suitable(cls, url):
 753         return (False
 754                 if any(ie.suitable(url) for ie in (
 755                     TwitchVodIE,
 756                     TwitchCollectionIE,
 757                     TwitchVideosIE,
 758                     TwitchVideosClipsIE,
 759                     TwitchVideosCollectionsIE,
 760                     TwitchClipsIE))
 761                 else super(TwitchStreamIE, cls).suitable(url))
 762
 763     def _real_extract(self, url):
 764         channel_name = self._match_id(url).lower()
 765
 766         gql = self._download_gql(
 767             channel_name, [{
 768                 'operationName': 'StreamMetadata',
 769                 'variables': {'channelLogin': channel_name},
 770             }, {
 771                 'operationName': 'ComscoreStreamingQuery',
 772                 'variables': {
 773                     'channel': channel_name,
 774                     'clipSlug': '',
 775                     'isClip': False,
 776                     'isLive': True,
 777                     'isVodOrCollection': False,
 778                     'vodID': '',
 779                 },
 780             }, {
 781                 'operationName': 'VideoPreviewOverlay',
 782                 'variables': {'login': channel_name},
 783             }],
 784             'Downloading stream GraphQL')
 785
 786         user = gql[0]['data']['user']
 787
 788         if not user:
 789             raise ExtractorError(
 790                 '%s does not exist' % channel_name, expected=True)
 791
 792         stream = user['stream']
 793
 794         if not stream:
 795             raise ExtractorError('%s is offline' % channel_name, expected=True)
 796
 797         access_token = self._download_access_token(
 798             channel_name, 'stream', 'channelName')
 799         token = access_token['value']
 800
 801         stream_id = stream.get('id') or channel_name
 802         query = {
 803             'allow_source': 'true',
 804             'allow_audio_only': 'true',
 805             'allow_spectre': 'true',
 806             'p': random.randint(1000000, 10000000),
 807             'player': 'twitchweb',
 808             'playlist_include_framerate': 'true',
 809             'segment_preference': '4',
 810             'sig': access_token['signature'].encode('utf-8'),
 811             'token': token.encode('utf-8'),
 812         }
 813         formats = self._extract_m3u8_formats(
 814             '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
 815             stream_id, 'mp4', query=query)
 816         self._prefer_source(formats)
 817
 818         view_count = stream.get('viewers')
 819         timestamp = unified_timestamp(stream.get('createdAt'))
 820
 821         sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
 822         uploader = sq_user.get('displayName')
 823         description = try_get(
 824             sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
 825
 826         thumbnail = url_or_none(try_get(
 827             gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
 828             compat_str))
 829
 830         title = uploader or channel_name
 831         stream_type = stream.get('type')
 832         if stream_type in ['rerun', 'live']:
 833             title += ' (%s)' % stream_type
 834
 835         return {
 836             'id': stream_id,
 837             'display_id': channel_name,
 838             'title': self._live_title(title),
 839             'description': description,
 840             'thumbnail': thumbnail,
 841             'uploader': uploader,
 842             'uploader_id': channel_name,
 843             'timestamp': timestamp,
 844             'view_count': view_count,
 845             'formats': formats,
 846             'is_live': stream_type == 'live',
 847         }
 848
 849
 850 class TwitchClipsIE(TwitchBaseIE):
 851     IE_NAME = 'twitch:clips'
 852     _VALID_URL = r'''(?x)
 853                     https?://
 854                         (?:
 855                             clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)|
 856                             (?:(?:www|go|m)\.)?twitch\.tv/[^/]+/clip/
 857                         )
 858                         (?P<id>[^/?#&]+)
 859                     '''
 860
 861     _TESTS = [{
 862         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
 863         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 864         'info_dict': {
 865             'id': '42850523',
 866             'ext': 'mp4',
 867             'title': 'EA Play 2016 Live from the Novo Theatre',
 868             'thumbnail': r're:^https?://.*\.jpg',
 869             'timestamp': 1465767393,
 870             'upload_date': '20160612',
 871             'creator': 'EA',
 872             'uploader': 'stereotype_',
 873             'uploader_id': '43566419',
 874         },
 875     }, {
 876         # multiple formats
 877         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 878         'only_matching': True,
 879     }, {
 880         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
 881         'only_matching': True,
 882     }, {
 883         'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited',
 884         'only_matching': True,
 885     }, {
 886         'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
 887         'only_matching': True,
 888     }, {
 889         'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
 890         'only_matching': True,
 891     }]
 892
 893     def _real_extract(self, url):
 894         video_id = self._match_id(url)
 895
 896         clip = self._download_base_gql(
 897             video_id, {
 898                 'query': '''{
 899   clip(slug: "%s") {
 900     broadcaster {
 901       displayName
 902     }
 903     createdAt
 904     curator {
 905       displayName
 906       id
 907     }
 908     durationSeconds
 909     id
 910     tiny: thumbnailURL(width: 86, height: 45)
 911     small: thumbnailURL(width: 260, height: 147)
 912     medium: thumbnailURL(width: 480, height: 272)
 913     title
 914     videoQualities {
 915       frameRate
 916       quality
 917       sourceURL
 918     }
 919     viewCount
 920   }
 921 }''' % video_id}, 'Downloading clip GraphQL')['data']['clip']
 922
 923         if not clip:
 924             raise ExtractorError(
 925                 'This clip is no longer available', expected=True)
 926
 927         formats = []
 928         for option in clip.get('videoQualities', []):
 929             if not isinstance(option, dict):
 930                 continue
 931             source = url_or_none(option.get('sourceURL'))
 932             if not source:
 933                 continue
 934             formats.append({
 935                 'url': source,
 936                 'format_id': option.get('quality'),
 937                 'height': int_or_none(option.get('quality')),
 938                 'fps': int_or_none(option.get('frameRate')),
 939             })
 940         self._sort_formats(formats)
 941
 942         thumbnails = []
 943         for thumbnail_id in ('tiny', 'small', 'medium'):
 944             thumbnail_url = clip.get(thumbnail_id)
 945             if not thumbnail_url:
 946                 continue
 947             thumb = {
 948                 'id': thumbnail_id,
 949                 'url': thumbnail_url,
 950             }
 951             mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url)
 952             if mobj:
 953                 thumb.update({
 954                     'height': int(mobj.group(2)),
 955                     'width': int(mobj.group(1)),
 956                 })
 957             thumbnails.append(thumb)
 958
 959         return {
 960             'id': clip.get('id') or video_id,
 961             'title': clip.get('title') or video_id,
 962             'formats': formats,
 963             'duration': int_or_none(clip.get('durationSeconds')),
 964             'views': int_or_none(clip.get('viewCount')),
 965             'timestamp': unified_timestamp(clip.get('createdAt')),
 966             'thumbnails': thumbnails,
 967             'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str),
 968             'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str),
 969             'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
 970         }