youtube_dlc/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_kwargs,
  13     compat_parse_qs,
  14     compat_str,
  15     compat_urlparse,
  16     compat_urllib_parse_urlencode,
  17     compat_urllib_parse_urlparse,
  18 )
  19 from ..utils import (
  20     clean_html,
  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     def _handle_error(self, response):
  46         if not isinstance(response, dict):
  47             return
  48         error = response.get('error')
  49         if error:
  50             raise ExtractorError(
  51                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  52                 expected=True)
  53
  54     def _call_api(self, path, item_id, *args, **kwargs):
  55         headers = kwargs.get('headers', {}).copy()
  56         headers.update({
  57             'Accept': 'application/vnd.twitchtv.v5+json; charset=UTF-8',
  58             'Client-ID': self._CLIENT_ID,
  59         })
  60         kwargs.update({
  61             'headers': headers,
  62             'expected_status': (400, 410),
  63         })
  64         response = self._download_json(
  65             '%s/%s' % (self._API_BASE, path), item_id,
  66             *args, **compat_kwargs(kwargs))
  67         self._handle_error(response)
  68         return response
  69
  70     def _real_initialize(self):
  71         self._login()
  72
  73     def _login(self):
  74         username, password = self._get_login_info()
  75         if username is None:
  76             return
  77
  78         def fail(message):
  79             raise ExtractorError(
  80                 'Unable to login. Twitch said: %s' % message, expected=True)
  81
  82         def login_step(page, urlh, note, data):
  83             form = self._hidden_inputs(page)
  84             form.update(data)
  85
  86             page_url = urlh.geturl()
  87             post_url = self._search_regex(
  88                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
  89                 'post url', default=self._LOGIN_POST_URL, group='url')
  90             post_url = urljoin(page_url, post_url)
  91
  92             headers = {
  93                 'Referer': page_url,
  94                 'Origin': page_url,
  95                 'Content-Type': 'text/plain;charset=UTF-8',
  96             }
  97
  98             response = self._download_json(
  99                 post_url, None, note, data=json.dumps(form).encode(),
 100                 headers=headers, expected_status=400)
 101             error = response.get('error_description') or response.get('error_code')
 102             if error:
 103                 fail(error)
 104
 105             if 'Authenticated successfully' in response.get('message', ''):
 106                 return None, None
 107
 108             redirect_url = urljoin(
 109                 post_url,
 110                 response.get('redirect') or response['redirect_path'])
 111             return self._download_webpage_handle(
 112                 redirect_url, None, 'Downloading login redirect page',
 113                 headers=headers)
 114
 115         login_page, handle = self._download_webpage_handle(
 116             self._LOGIN_FORM_URL, None, 'Downloading login page')
 117
 118         # Some TOR nodes and public proxies are blocked completely
 119         if 'blacklist_message' in login_page:
 120             fail(clean_html(login_page))
 121
 122         redirect_page, handle = login_step(
 123             login_page, handle, 'Logging in', {
 124                 'username': username,
 125                 'password': password,
 126                 'client_id': self._CLIENT_ID,
 127             })
 128
 129         # Successful login
 130         if not redirect_page:
 131             return
 132
 133         if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
 134             # TODO: Add mechanism to request an SMS or phone call
 135             tfa_token = self._get_tfa_info('two-factor authentication token')
 136             login_step(redirect_page, handle, 'Submitting TFA token', {
 137                 'authy_token': tfa_token,
 138                 'remember_2fa': 'true',
 139             })
 140
 141     def _prefer_source(self, formats):
 142         try:
 143             source = next(f for f in formats if f['format_id'] == 'Source')
 144             source['quality'] = 10
 145         except StopIteration:
 146             for f in formats:
 147                 if '/chunked/' in f['url']:
 148                     f.update({
 149                         'quality': 10,
 150                         'format_note': 'Source',
 151                     })
 152         self._sort_formats(formats)
 153
 154     def _download_access_token(self, channel_name):
 155         return self._call_api(
 156             'api/channels/%s/access_token' % channel_name, channel_name,
 157             'Downloading access token JSON')
 158
 159     def _extract_channel_id(self, token, channel_name):
 160         return compat_str(self._parse_json(token, channel_name)['channel_id'])
 161
 162
 163 class TwitchVodIE(TwitchBaseIE):
 164     IE_NAME = 'twitch:vod'
 165     _VALID_URL = r'''(?x)
 166                     https?://
 167                         (?:
 168                             (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
 169                             player\.twitch\.tv/\?.*?\bvideo=v?
 170                         )
 171                         (?P<id>\d+)
 172                     '''
 173     _ITEM_TYPE = 'vod'
 174     _ITEM_SHORTCUT = 'v'
 175
 176     _TESTS = [{
 177         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 178         'info_dict': {
 179             'id': 'v6528877',
 180             'ext': 'mp4',
 181             'title': 'LCK Summer Split - Week 6 Day 1',
 182             'thumbnail': r're:^https?://.*\.jpg$',
 183             'duration': 17208,
 184             'timestamp': 1435131709,
 185             'upload_date': '20150624',
 186             'uploader': 'Riot Games',
 187             'uploader_id': 'riotgames',
 188             'view_count': int,
 189             'start_time': 310,
 190         },
 191         'params': {
 192             # m3u8 download
 193             'skip_download': True,
 194         },
 195     }, {
 196         # Untitled broadcast (title is None)
 197         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 198         'info_dict': {
 199             'id': 'v11230755',
 200             'ext': 'mp4',
 201             'title': 'Untitled Broadcast',
 202             'thumbnail': r're:^https?://.*\.jpg$',
 203             'duration': 1638,
 204             'timestamp': 1439746708,
 205             'upload_date': '20150816',
 206             'uploader': 'BelkAO_o',
 207             'uploader_id': 'belkao_o',
 208             'view_count': int,
 209         },
 210         'params': {
 211             # m3u8 download
 212             'skip_download': True,
 213         },
 214         'skip': 'HTTP Error 404: Not Found',
 215     }, {
 216         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 217         'only_matching': True,
 218     }, {
 219         'url': 'https://www.twitch.tv/videos/6528877',
 220         'only_matching': True,
 221     }, {
 222         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
 223         'only_matching': True,
 224     }, {
 225         'url': 'https://www.twitch.tv/northernlion/video/291940395',
 226         'only_matching': True,
 227     }, {
 228         'url': 'https://player.twitch.tv/?video=480452374',
 229         'only_matching': True,
 230     }]
 231
 232     def _download_info(self, item_id):
 233         return self._extract_info(
 234             self._call_api(
 235                 'kraken/videos/%s' % item_id, item_id,
 236                 'Downloading video info JSON'))
 237
 238     @staticmethod
 239     def _extract_info(info):
 240         status = info.get('status')
 241         if status == 'recording':
 242             is_live = True
 243         elif status == 'recorded':
 244             is_live = False
 245         else:
 246             is_live = None
 247         _QUALITIES = ('small', 'medium', 'large')
 248         quality_key = qualities(_QUALITIES)
 249         thumbnails = []
 250         preview = info.get('preview')
 251         if isinstance(preview, dict):
 252             for thumbnail_id, thumbnail_url in preview.items():
 253                 thumbnail_url = url_or_none(thumbnail_url)
 254                 if not thumbnail_url:
 255                     continue
 256                 if thumbnail_id not in _QUALITIES:
 257                     continue
 258                 thumbnails.append({
 259                     'url': thumbnail_url,
 260                     'preference': quality_key(thumbnail_id),
 261                 })
 262         return {
 263             'id': info['_id'],
 264             'title': info.get('title') or 'Untitled Broadcast',
 265             'description': info.get('description'),
 266             'duration': int_or_none(info.get('length')),
 267             'thumbnails': thumbnails,
 268             'uploader': info.get('channel', {}).get('display_name'),
 269             'uploader_id': info.get('channel', {}).get('name'),
 270             'timestamp': parse_iso8601(info.get('recorded_at')),
 271             'view_count': int_or_none(info.get('views')),
 272             'is_live': is_live,
 273         }
 274
 275     def _real_extract(self, url):
 276         vod_id = self._match_id(url)
 277
 278         info = self._download_info(vod_id)
 279         access_token = self._call_api(
 280             'api/vods/%s/access_token' % vod_id, vod_id,
 281             'Downloading %s access token' % self._ITEM_TYPE)
 282
 283         formats = self._extract_m3u8_formats(
 284             '%s/vod/%s.m3u8?%s' % (
 285                 self._USHER_BASE, vod_id,
 286                 compat_urllib_parse_urlencode({
 287                     'allow_source': 'true',
 288                     'allow_audio_only': 'true',
 289                     'allow_spectre': 'true',
 290                     'player': 'twitchweb',
 291                     'playlist_include_framerate': 'true',
 292                     'nauth': access_token['token'],
 293                     'nauthsig': access_token['sig'],
 294                 })),
 295             vod_id, 'mp4', entry_protocol='m3u8_native')
 296
 297         self._prefer_source(formats)
 298         info['formats'] = formats
 299
 300         parsed_url = compat_urllib_parse_urlparse(url)
 301         query = compat_parse_qs(parsed_url.query)
 302         if 't' in query:
 303             info['start_time'] = parse_duration(query['t'][0])
 304
 305         if info.get('timestamp') is not None:
 306             info['subtitles'] = {
 307                 'rechat': [{
 308                     'url': update_url_query(
 309                         'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, {
 310                             'client_id': self._CLIENT_ID,
 311                         }),
 312                     'ext': 'json',
 313                 }],
 314             }
 315
 316         return info
 317
 318
 319 def _make_video_result(node):
 320     assert isinstance(node, dict)
 321     video_id = node.get('id')
 322     if not video_id:
 323         return
 324     return {
 325         '_type': 'url_transparent',
 326         'ie_key': TwitchVodIE.ie_key(),
 327         'id': video_id,
 328         'url': 'https://www.twitch.tv/videos/%s' % video_id,
 329         'title': node.get('title'),
 330         'thumbnail': node.get('previewThumbnailURL'),
 331         'duration': float_or_none(node.get('lengthSeconds')),
 332         'view_count': int_or_none(node.get('viewCount')),
 333     }
 334
 335
 336 class TwitchGraphQLBaseIE(TwitchBaseIE):
 337     _PAGE_LIMIT = 100
 338
 339     _OPERATION_HASHES = {
 340         'CollectionSideBar': '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14',
 341         'FilterableVideoTower_Videos': 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb',
 342         'ClipsCards__User': 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777',
 343         'ChannelCollectionsContent': '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84',
 344         'StreamMetadata': '1c719a40e481453e5c48d9bb585d971b8b372f8ebb105b17076722264dfa5b3e',
 345         'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
 346         'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
 347     }
 348
 349     def _download_gql(self, video_id, ops, note, fatal=True):
 350         for op in ops:
 351             op['extensions'] = {
 352                 'persistedQuery': {
 353                     'version': 1,
 354                     'sha256Hash': self._OPERATION_HASHES[op['operationName']],
 355                 }
 356             }
 357         return self._download_json(
 358             'https://gql.twitch.tv/gql', video_id, note,
 359             data=json.dumps(ops).encode(),
 360             headers={
 361                 'Content-Type': 'text/plain;charset=UTF-8',
 362                 'Client-ID': self._CLIENT_ID,
 363             }, fatal=fatal)
 364
 365
 366 class TwitchCollectionIE(TwitchGraphQLBaseIE):
 367     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P<id>[^/]+)'
 368
 369     _TESTS = [{
 370         'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ',
 371         'info_dict': {
 372             'id': 'wlDCoH0zEBZZbQ',
 373             'title': 'Overthrow Nook, capitalism for children',
 374         },
 375         'playlist_mincount': 13,
 376     }]
 377
 378     _OPERATION_NAME = 'CollectionSideBar'
 379
 380     def _real_extract(self, url):
 381         collection_id = self._match_id(url)
 382         collection = self._download_gql(
 383             collection_id, [{
 384                 'operationName': self._OPERATION_NAME,
 385                 'variables': {'collectionID': collection_id},
 386             }],
 387             'Downloading collection GraphQL')[0]['data']['collection']
 388         title = collection.get('title')
 389         entries = []
 390         for edge in collection['items']['edges']:
 391             if not isinstance(edge, dict):
 392                 continue
 393             node = edge.get('node')
 394             if not isinstance(node, dict):
 395                 continue
 396             video = _make_video_result(node)
 397             if video:
 398                 entries.append(video)
 399         return self.playlist_result(
 400             entries, playlist_id=collection_id, playlist_title=title)
 401
 402
 403 class TwitchPlaylistBaseIE(TwitchGraphQLBaseIE):
 404     def _entries(self, channel_name, *args):
 405         cursor = None
 406         variables_common = self._make_variables(channel_name, *args)
 407         entries_key = '%ss' % self._ENTRY_KIND
 408         for page_num in itertools.count(1):
 409             variables = variables_common.copy()
 410             variables['limit'] = self._PAGE_LIMIT
 411             if cursor:
 412                 variables['cursor'] = cursor
 413             page = self._download_gql(
 414                 channel_name, [{
 415                     'operationName': self._OPERATION_NAME,
 416                     'variables': variables,
 417                 }],
 418                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
 419                 fatal=False)
 420             if not page:
 421                 break
 422             edges = try_get(
 423                 page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
 424             if not edges:
 425                 break
 426             for edge in edges:
 427                 if not isinstance(edge, dict):
 428                     continue
 429                 if edge.get('__typename') != self._EDGE_KIND:
 430                     continue
 431                 node = edge.get('node')
 432                 if not isinstance(node, dict):
 433                     continue
 434                 if node.get('__typename') != self._NODE_KIND:
 435                     continue
 436                 entry = self._extract_entry(node)
 437                 if entry:
 438                     cursor = edge.get('cursor')
 439                     yield entry
 440             if not cursor or not isinstance(cursor, compat_str):
 441                 break
 442
 443     # Deprecated kraken v5 API
 444     def _entries_kraken(self, channel_name, broadcast_type, sort):
 445         access_token = self._download_access_token(channel_name)
 446         channel_id = self._extract_channel_id(access_token['token'], channel_name)
 447         offset = 0
 448         counter_override = None
 449         for counter in itertools.count(1):
 450             response = self._call_api(
 451                 'kraken/channels/%s/videos/' % channel_id,
 452                 channel_id,
 453                 'Downloading video JSON page %s' % (counter_override or counter),
 454                 query={
 455                     'offset': offset,
 456                     'limit': self._PAGE_LIMIT,
 457                     'broadcast_type': broadcast_type,
 458                     'sort': sort,
 459                 })
 460             videos = response.get('videos')
 461             if not isinstance(videos, list):
 462                 break
 463             for video in videos:
 464                 if not isinstance(video, dict):
 465                     continue
 466                 video_url = url_or_none(video.get('url'))
 467                 if not video_url:
 468                     continue
 469                 yield {
 470                     '_type': 'url_transparent',
 471                     'ie_key': TwitchVodIE.ie_key(),
 472                     'id': video.get('_id'),
 473                     'url': video_url,
 474                     'title': video.get('title'),
 475                     'description': video.get('description'),
 476                     'timestamp': unified_timestamp(video.get('published_at')),
 477                     'duration': float_or_none(video.get('length')),
 478                     'view_count': int_or_none(video.get('views')),
 479                     'language': video.get('language'),
 480                 }
 481             offset += self._PAGE_LIMIT
 482             total = int_or_none(response.get('_total'))
 483             if total and offset >= total:
 484                 break
 485
 486
 487 class TwitchVideosIE(TwitchPlaylistBaseIE):
 488     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:videos|profile)'
 489
 490     _TESTS = [{
 491         # All Videos sorted by Date
 492         'url': 'https://www.twitch.tv/spamfish/videos?filter=all',
 493         'info_dict': {
 494             'id': 'spamfish',
 495             'title': 'spamfish - All Videos sorted by Date',
 496         },
 497         'playlist_mincount': 924,
 498     }, {
 499         # All Videos sorted by Popular
 500         'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views',
 501         'info_dict': {
 502             'id': 'spamfish',
 503             'title': 'spamfish - All Videos sorted by Popular',
 504         },
 505         'playlist_mincount': 931,
 506     }, {
 507         # Past Broadcasts sorted by Date
 508         'url': 'https://www.twitch.tv/spamfish/videos?filter=archives',
 509         'info_dict': {
 510             'id': 'spamfish',
 511             'title': 'spamfish - Past Broadcasts sorted by Date',
 512         },
 513         'playlist_mincount': 27,
 514     }, {
 515         # Highlights sorted by Date
 516         'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights',
 517         'info_dict': {
 518             'id': 'spamfish',
 519             'title': 'spamfish - Highlights sorted by Date',
 520         },
 521         'playlist_mincount': 901,
 522     }, {
 523         # Uploads sorted by Date
 524         'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time',
 525         'info_dict': {
 526             'id': 'esl_csgo',
 527             'title': 'esl_csgo - Uploads sorted by Date',
 528         },
 529         'playlist_mincount': 5,
 530     }, {
 531         # Past Premieres sorted by Date
 532         'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres',
 533         'info_dict': {
 534             'id': 'spamfish',
 535             'title': 'spamfish - Past Premieres sorted by Date',
 536         },
 537         'playlist_mincount': 1,
 538     }, {
 539         'url': 'https://www.twitch.tv/spamfish/videos/all',
 540         'only_matching': True,
 541     }, {
 542         'url': 'https://m.twitch.tv/spamfish/videos/all',
 543         'only_matching': True,
 544     }, {
 545         'url': 'https://www.twitch.tv/spamfish/videos',
 546         'only_matching': True,
 547     }]
 548
 549     Broadcast = collections.namedtuple('Broadcast', ['type', 'label'])
 550
 551     _DEFAULT_BROADCAST = Broadcast(None, 'All Videos')
 552     _BROADCASTS = {
 553         'archives': Broadcast('ARCHIVE', 'Past Broadcasts'),
 554         'highlights': Broadcast('HIGHLIGHT', 'Highlights'),
 555         'uploads': Broadcast('UPLOAD', 'Uploads'),
 556         'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'),
 557         'all': _DEFAULT_BROADCAST,
 558     }
 559
 560     _DEFAULT_SORTED_BY = 'Date'
 561     _SORTED_BY = {
 562         'time': _DEFAULT_SORTED_BY,
 563         'views': 'Popular',
 564     }
 565
 566     _OPERATION_NAME = 'FilterableVideoTower_Videos'
 567     _ENTRY_KIND = 'video'
 568     _EDGE_KIND = 'VideoEdge'
 569     _NODE_KIND = 'Video'
 570
 571     @classmethod
 572     def suitable(cls, url):
 573         return (False
 574                 if any(ie.suitable(url) for ie in (
 575                     TwitchVideosClipsIE,
 576                     TwitchVideosCollectionsIE))
 577                 else super(TwitchVideosIE, cls).suitable(url))
 578
 579     @staticmethod
 580     def _make_variables(channel_name, broadcast_type, sort):
 581         return {
 582             'channelOwnerLogin': channel_name,
 583             'broadcastType': broadcast_type,
 584             'videoSort': sort.upper(),
 585         }
 586
 587     @staticmethod
 588     def _extract_entry(node):
 589         return _make_video_result(node)
 590
 591     def _real_extract(self, url):
 592         channel_name = self._match_id(url)
 593         qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 594         filter = qs.get('filter', ['all'])[0]
 595         sort = qs.get('sort', ['time'])[0]
 596         broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST)
 597         return self.playlist_result(
 598             self._entries(channel_name, broadcast.type, sort),
 599             playlist_id=channel_name,
 600             playlist_title='%s - %s sorted by %s'
 601             % (channel_name, broadcast.label,
 602                self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY)))
 603
 604
 605 class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
 606     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)'
 607
 608     _TESTS = [{
 609         # Clips
 610         'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all',
 611         'info_dict': {
 612             'id': 'vanillatv',
 613             'title': 'vanillatv - Clips Top All',
 614         },
 615         'playlist_mincount': 1,
 616     }, {
 617         'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d',
 618         'only_matching': True,
 619     }]
 620
 621     Clip = collections.namedtuple('Clip', ['filter', 'label'])
 622
 623     _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D')
 624     _RANGE = {
 625         '24hr': Clip('LAST_DAY', 'Top 24H'),
 626         '7d': _DEFAULT_CLIP,
 627         '30d': Clip('LAST_MONTH', 'Top 30D'),
 628         'all': Clip('ALL_TIME', 'Top All'),
 629     }
 630
 631     # NB: values other than 20 result in skipped videos
 632     _PAGE_LIMIT = 20
 633
 634     _OPERATION_NAME = 'ClipsCards__User'
 635     _ENTRY_KIND = 'clip'
 636     _EDGE_KIND = 'ClipEdge'
 637     _NODE_KIND = 'Clip'
 638
 639     @staticmethod
 640     def _make_variables(channel_name, filter):
 641         return {
 642             'login': channel_name,
 643             'criteria': {
 644                 'filter': filter,
 645             },
 646         }
 647
 648     @staticmethod
 649     def _extract_entry(node):
 650         assert isinstance(node, dict)
 651         clip_url = url_or_none(node.get('url'))
 652         if not clip_url:
 653             return
 654         return {
 655             '_type': 'url_transparent',
 656             'ie_key': TwitchClipsIE.ie_key(),
 657             'id': node.get('id'),
 658             'url': clip_url,
 659             'title': node.get('title'),
 660             'thumbnail': node.get('thumbnailURL'),
 661             'duration': float_or_none(node.get('durationSeconds')),
 662             'timestamp': unified_timestamp(node.get('createdAt')),
 663             'view_count': int_or_none(node.get('viewCount')),
 664             'language': node.get('language'),
 665         }
 666
 667     def _real_extract(self, url):
 668         channel_name = self._match_id(url)
 669         qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 670         range = qs.get('range', ['7d'])[0]
 671         clip = self._RANGE.get(range, self._DEFAULT_CLIP)
 672         return self.playlist_result(
 673             self._entries(channel_name, clip.filter),
 674             playlist_id=channel_name,
 675             playlist_title='%s - Clips %s' % (channel_name, clip.label))
 676
 677
 678 class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
 679     _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P<id>[^/]+)/videos/*?\?.*?\bfilter=collections'
 680
 681     _TESTS = [{
 682         # Collections
 683         'url': 'https://www.twitch.tv/spamfish/videos?filter=collections',
 684         'info_dict': {
 685             'id': 'spamfish',
 686             'title': 'spamfish - Collections',
 687         },
 688         'playlist_mincount': 3,
 689     }]
 690
 691     _OPERATION_NAME = 'ChannelCollectionsContent'
 692     _ENTRY_KIND = 'collection'
 693     _EDGE_KIND = 'CollectionsItemEdge'
 694     _NODE_KIND = 'Collection'
 695
 696     @staticmethod
 697     def _make_variables(channel_name):
 698         return {
 699             'ownerLogin': channel_name,
 700         }
 701
 702     @staticmethod
 703     def _extract_entry(node):
 704         assert isinstance(node, dict)
 705         collection_id = node.get('id')
 706         if not collection_id:
 707             return
 708         return {
 709             '_type': 'url_transparent',
 710             'ie_key': TwitchCollectionIE.ie_key(),
 711             'id': collection_id,
 712             'url': 'https://www.twitch.tv/collections/%s' % collection_id,
 713             'title': node.get('title'),
 714             'thumbnail': node.get('thumbnailURL'),
 715             'duration': float_or_none(node.get('lengthSeconds')),
 716             'timestamp': unified_timestamp(node.get('updatedAt')),
 717             'view_count': int_or_none(node.get('viewCount')),
 718         }
 719
 720     def _real_extract(self, url):
 721         channel_name = self._match_id(url)
 722         return self.playlist_result(
 723             self._entries(channel_name), playlist_id=channel_name,
 724             playlist_title='%s - Collections' % channel_name)
 725
 726
 727 class TwitchStreamIE(TwitchGraphQLBaseIE):
 728     IE_NAME = 'twitch:stream'
 729     _VALID_URL = r'''(?x)
 730                     https?://
 731                         (?:
 732                             (?:(?:www|go|m)\.)?twitch\.tv/|
 733                             player\.twitch\.tv/\?.*?\bchannel=
 734                         )
 735                         (?P<id>[^/#?]+)
 736                     '''
 737
 738     _TESTS = [{
 739         'url': 'http://www.twitch.tv/shroomztv',
 740         'info_dict': {
 741             'id': '12772022048',
 742             'display_id': 'shroomztv',
 743             'ext': 'mp4',
 744             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 745             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 746             'is_live': True,
 747             'timestamp': 1421928037,
 748             'upload_date': '20150122',
 749             'uploader': 'ShroomzTV',
 750             'uploader_id': 'shroomztv',
 751             'view_count': int,
 752         },
 753         'params': {
 754             # m3u8 download
 755             'skip_download': True,
 756         },
 757     }, {
 758         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 759         'only_matching': True,
 760     }, {
 761         'url': 'https://player.twitch.tv/?channel=lotsofs',
 762         'only_matching': True,
 763     }, {
 764         'url': 'https://go.twitch.tv/food',
 765         'only_matching': True,
 766     }, {
 767         'url': 'https://m.twitch.tv/food',
 768         'only_matching': True,
 769     }]
 770
 771     @classmethod
 772     def suitable(cls, url):
 773         return (False
 774                 if any(ie.suitable(url) for ie in (
 775                     TwitchVodIE,
 776                     TwitchCollectionIE,
 777                     TwitchVideosIE,
 778                     TwitchVideosClipsIE,
 779                     TwitchVideosCollectionsIE,
 780                     TwitchClipsIE))
 781                 else super(TwitchStreamIE, cls).suitable(url))
 782
 783     def _real_extract(self, url):
 784         channel_name = self._match_id(url).lower()
 785
 786         gql = self._download_gql(
 787             channel_name, [{
 788                 'operationName': 'StreamMetadata',
 789                 'variables': {'channelLogin': channel_name},
 790             }, {
 791                 'operationName': 'ComscoreStreamingQuery',
 792                 'variables': {
 793                     'channel': channel_name,
 794                     'clipSlug': '',
 795                     'isClip': False,
 796                     'isLive': True,
 797                     'isVodOrCollection': False,
 798                     'vodID': '',
 799                 },
 800             }, {
 801                 'operationName': 'VideoPreviewOverlay',
 802                 'variables': {'login': channel_name},
 803             }],
 804             'Downloading stream GraphQL')
 805
 806         user = gql[0]['data']['user']
 807
 808         if not user:
 809             raise ExtractorError(
 810                 '%s does not exist' % channel_name, expected=True)
 811
 812         stream = user['stream']
 813
 814         if not stream:
 815             raise ExtractorError('%s is offline' % channel_name, expected=True)
 816
 817         access_token = self._download_access_token(channel_name)
 818         token = access_token['token']
 819
 820         stream_id = stream.get('id') or channel_name
 821         query = {
 822             'allow_source': 'true',
 823             'allow_audio_only': 'true',
 824             'allow_spectre': 'true',
 825             'p': random.randint(1000000, 10000000),
 826             'player': 'twitchweb',
 827             'playlist_include_framerate': 'true',
 828             'segment_preference': '4',
 829             'sig': access_token['sig'].encode('utf-8'),
 830             'token': token.encode('utf-8'),
 831         }
 832         formats = self._extract_m3u8_formats(
 833             '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
 834             stream_id, 'mp4', query=query)
 835         self._prefer_source(formats)
 836
 837         view_count = stream.get('viewers')
 838         timestamp = unified_timestamp(stream.get('createdAt'))
 839
 840         sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
 841         uploader = sq_user.get('displayName')
 842         description = try_get(
 843             sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
 844
 845         thumbnail = url_or_none(try_get(
 846             gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
 847             compat_str))
 848
 849         title = uploader or channel_name
 850         stream_type = stream.get('type')
 851         if stream_type in ['rerun', 'live']:
 852             title += ' (%s)' % stream_type
 853
 854         return {
 855             'id': stream_id,
 856             'display_id': channel_name,
 857             'title': self._live_title(title),
 858             'description': description,
 859             'thumbnail': thumbnail,
 860             'uploader': uploader,
 861             'uploader_id': channel_name,
 862             'timestamp': timestamp,
 863             'view_count': view_count,
 864             'formats': formats,
 865             'is_live': stream_type == 'live',
 866         }
 867
 868
 869 class TwitchClipsIE(TwitchBaseIE):
 870     IE_NAME = 'twitch:clips'
 871     _VALID_URL = r'''(?x)
 872                     https?://
 873                         (?:
 874                             clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)|
 875                             (?:(?:www|go|m)\.)?twitch\.tv/[^/]+/clip/
 876                         )
 877                         (?P<id>[^/?#&]+)
 878                     '''
 879
 880     _TESTS = [{
 881         'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat',
 882         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 883         'info_dict': {
 884             'id': '42850523',
 885             'ext': 'mp4',
 886             'title': 'EA Play 2016 Live from the Novo Theatre',
 887             'thumbnail': r're:^https?://.*\.jpg',
 888             'timestamp': 1465767393,
 889             'upload_date': '20160612',
 890             'creator': 'EA',
 891             'uploader': 'stereotype_',
 892             'uploader_id': '43566419',
 893         },
 894     }, {
 895         # multiple formats
 896         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 897         'only_matching': True,
 898     }, {
 899         'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan',
 900         'only_matching': True,
 901     }, {
 902         'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited',
 903         'only_matching': True,
 904     }, {
 905         'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
 906         'only_matching': True,
 907     }, {
 908         'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank',
 909         'only_matching': True,
 910     }]
 911
 912     def _real_extract(self, url):
 913         video_id = self._match_id(url)
 914
 915         clip = self._download_json(
 916             'https://gql.twitch.tv/gql', video_id, data=json.dumps({
 917                 'query': '''{
 918   clip(slug: "%s") {
 919     broadcaster {
 920       displayName
 921     }
 922     createdAt
 923     curator {
 924       displayName
 925       id
 926     }
 927     durationSeconds
 928     id
 929     tiny: thumbnailURL(width: 86, height: 45)
 930     small: thumbnailURL(width: 260, height: 147)
 931     medium: thumbnailURL(width: 480, height: 272)
 932     title
 933     videoQualities {
 934       frameRate
 935       quality
 936       sourceURL
 937     }
 938     viewCount
 939   }
 940 }''' % video_id,
 941             }).encode(), headers={
 942                 'Client-ID': self._CLIENT_ID,
 943             })['data']['clip']
 944
 945         if not clip:
 946             raise ExtractorError(
 947                 'This clip is no longer available', expected=True)
 948
 949         formats = []
 950         for option in clip.get('videoQualities', []):
 951             if not isinstance(option, dict):
 952                 continue
 953             source = url_or_none(option.get('sourceURL'))
 954             if not source:
 955                 continue
 956             formats.append({
 957                 'url': source,
 958                 'format_id': option.get('quality'),
 959                 'height': int_or_none(option.get('quality')),
 960                 'fps': int_or_none(option.get('frameRate')),
 961             })
 962         self._sort_formats(formats)
 963
 964         thumbnails = []
 965         for thumbnail_id in ('tiny', 'small', 'medium'):
 966             thumbnail_url = clip.get(thumbnail_id)
 967             if not thumbnail_url:
 968                 continue
 969             thumb = {
 970                 'id': thumbnail_id,
 971                 'url': thumbnail_url,
 972             }
 973             mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url)
 974             if mobj:
 975                 thumb.update({
 976                     'height': int(mobj.group(2)),
 977                     'width': int(mobj.group(1)),
 978                 })
 979             thumbnails.append(thumb)
 980
 981         return {
 982             'id': clip.get('id') or video_id,
 983             'title': clip.get('title') or video_id,
 984             'formats': formats,
 985             'duration': int_or_none(clip.get('durationSeconds')),
 986             'views': int_or_none(clip.get('viewCount')),
 987             'timestamp': unified_timestamp(clip.get('createdAt')),
 988             'thumbnails': thumbnails,
 989             'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str),
 990             'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str),
 991             'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str),
 992         }