_ Git - youtube-dl/blob - youtube_dl/extractor/twitch.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import re
   6 import random
   7
   8 from .common import InfoExtractor
   9 from ..compat import (
  10     compat_HTTPError,
  11     compat_parse_qs,
  12     compat_str,
  13     compat_urllib_parse_urlencode,
  14     compat_urllib_parse_urlparse,
  15     compat_urlparse,
  16 )
  17 from ..utils import (
  18     clean_html,
  19     ExtractorError,
  20     int_or_none,
  21     js_to_json,
  22     orderedSet,
  23     parse_duration,
  24     parse_iso8601,
  25     update_url_query,
  26     urlencode_postdata,
  27 )
  28
  29
  30 class TwitchBaseIE(InfoExtractor):
  31     _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
  32
  33     _API_BASE = 'https://api.twitch.tv'
  34     _USHER_BASE = 'https://usher.ttvnw.net'
  35     _LOGIN_URL = 'http://www.twitch.tv/login'
  36     _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6'
  37     _NETRC_MACHINE = 'twitch'
  38
  39     def _handle_error(self, response):
  40         if not isinstance(response, dict):
  41             return
  42         error = response.get('error')
  43         if error:
  44             raise ExtractorError(
  45                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  46                 expected=True)
  47
  48     def _call_api(self, path, item_id, note):
  49         response = self._download_json(
  50             '%s/%s' % (self._API_BASE, path), item_id, note,
  51             headers={'Client-ID': self._CLIENT_ID})
  52         self._handle_error(response)
  53         return response
  54
  55     def _real_initialize(self):
  56         self._login()
  57
  58     def _login(self):
  59         (username, password) = self._get_login_info()
  60         if username is None:
  61             return
  62
  63         def fail(message):
  64             raise ExtractorError(
  65                 'Unable to login. Twitch said: %s' % message, expected=True)
  66
  67         login_page, handle = self._download_webpage_handle(
  68             self._LOGIN_URL, None, 'Downloading login page')
  69
  70         # Some TOR nodes and public proxies are blocked completely
  71         if 'blacklist_message' in login_page:
  72             fail(clean_html(login_page))
  73
  74         login_form = self._hidden_inputs(login_page)
  75
  76         login_form.update({
  77             'username': username,
  78             'password': password,
  79         })
  80
  81         redirect_url = handle.geturl()
  82
  83         post_url = self._search_regex(
  84             r'<form[^>]+action=(["\'])(?P<url>.+?)\1', login_page,
  85             'post url', default=redirect_url, group='url')
  86
  87         if not post_url.startswith('http'):
  88             post_url = compat_urlparse.urljoin(redirect_url, post_url)
  89
  90         headers = {'Referer': redirect_url}
  91
  92         try:
  93             response = self._download_json(
  94                 post_url, None, 'Logging in as %s' % username,
  95                 data=urlencode_postdata(login_form),
  96                 headers=headers)
  97         except ExtractorError as e:
  98             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400:
  99                 response = self._parse_json(
 100                     e.cause.read().decode('utf-8'), None)
 101                 fail(response['message'])
 102             raise
 103
 104         if response.get('redirect'):
 105             self._download_webpage(
 106                 response['redirect'], None, 'Downloading login redirect page',
 107                 headers=headers)
 108
 109     def _prefer_source(self, formats):
 110         try:
 111             source = next(f for f in formats if f['format_id'] == 'Source')
 112             source['preference'] = 10
 113         except StopIteration:
 114             pass  # No Source stream present
 115         self._sort_formats(formats)
 116
 117
 118 class TwitchItemBaseIE(TwitchBaseIE):
 119     def _download_info(self, item, item_id):
 120         return self._extract_info(self._call_api(
 121             'kraken/videos/%s%s' % (item, item_id), item_id,
 122             'Downloading %s info JSON' % self._ITEM_TYPE))
 123
 124     def _extract_media(self, item_id):
 125         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 126         response = self._call_api(
 127             'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id,
 128             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 129         entries = []
 130         chunks = response['chunks']
 131         qualities = list(chunks.keys())
 132         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 133             formats = []
 134             for fmt_num, fragment_fmt in enumerate(fragment):
 135                 format_id = qualities[fmt_num]
 136                 fmt = {
 137                     'url': fragment_fmt['url'],
 138                     'format_id': format_id,
 139                     'quality': 1 if format_id == 'live' else 0,
 140                 }
 141                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 142                 if m:
 143                     fmt['height'] = int(m.group('height'))
 144                 formats.append(fmt)
 145             self._sort_formats(formats)
 146             entry = dict(info)
 147             entry['id'] = '%s_%d' % (entry['id'], num)
 148             entry['title'] = '%s part %d' % (entry['title'], num)
 149             entry['formats'] = formats
 150             entries.append(entry)
 151         return self.playlist_result(entries, info['id'], info['title'])
 152
 153     def _extract_info(self, info):
 154         return {
 155             'id': info['_id'],
 156             'title': info.get('title') or 'Untitled Broadcast',
 157             'description': info.get('description'),
 158             'duration': int_or_none(info.get('length')),
 159             'thumbnail': info.get('preview'),
 160             'uploader': info.get('channel', {}).get('display_name'),
 161             'uploader_id': info.get('channel', {}).get('name'),
 162             'timestamp': parse_iso8601(info.get('recorded_at')),
 163             'view_count': int_or_none(info.get('views')),
 164         }
 165
 166     def _real_extract(self, url):
 167         return self._extract_media(self._match_id(url))
 168
 169
 170 class TwitchVideoIE(TwitchItemBaseIE):
 171     IE_NAME = 'twitch:video'
 172     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 173     _ITEM_TYPE = 'video'
 174     _ITEM_SHORTCUT = 'a'
 175
 176     _TEST = {
 177         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 178         'info_dict': {
 179             'id': 'a577357806',
 180             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 181         },
 182         'playlist_mincount': 12,
 183         'skip': 'HTTP Error 404: Not Found',
 184     }
 185
 186
 187 class TwitchChapterIE(TwitchItemBaseIE):
 188     IE_NAME = 'twitch:chapter'
 189     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 190     _ITEM_TYPE = 'chapter'
 191     _ITEM_SHORTCUT = 'c'
 192
 193     _TESTS = [{
 194         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 195         'info_dict': {
 196             'id': 'c5285812',
 197             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 198         },
 199         'playlist_mincount': 3,
 200         'skip': 'HTTP Error 404: Not Found',
 201     }, {
 202         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 203         'only_matching': True,
 204     }]
 205
 206
 207 class TwitchVodIE(TwitchItemBaseIE):
 208     IE_NAME = 'twitch:vod'
 209     _VALID_URL_BASE = r'https?://(?:www\.|player\.)?twitch\.tv'
 210     _VALID_URL = r'%s/(?:[^/]+/v/|\?video=v)(?P<id>\d+)' % _VALID_URL_BASE
 211     _ITEM_TYPE = 'vod'
 212     _ITEM_SHORTCUT = 'v'
 213
 214     _TESTS = [{
 215         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 216         'info_dict': {
 217             'id': 'v6528877',
 218             'ext': 'mp4',
 219             'title': 'LCK Summer Split - Week 6 Day 1',
 220             'thumbnail': 're:^https?://.*\.jpg$',
 221             'duration': 17208,
 222             'timestamp': 1435131709,
 223             'upload_date': '20150624',
 224             'uploader': 'Riot Games',
 225             'uploader_id': 'riotgames',
 226             'view_count': int,
 227             'start_time': 310,
 228         },
 229         'params': {
 230             # m3u8 download
 231             'skip_download': True,
 232         },
 233     }, {
 234         # player.twitch.tv URL
 235         'url': 'http://player.twitch.tv/?video=v6528877&t=5m10s',
 236         'info_dict': {
 237             'id': 'v6528877',
 238             'ext': 'mp4',
 239             'title': 'LCK Summer Split - Week 6 Day 1',
 240             'thumbnail': 're:^https?://.*\.jpg$',
 241             'duration': 17208,
 242             'timestamp': 1435131709,
 243             'upload_date': '20150624',
 244             'uploader': 'Riot Games',
 245             'uploader_id': 'riotgames',
 246             'view_count': int,
 247             'start_time': 310,
 248         },
 249         'params': {
 250             # m3u8 download
 251             'skip_download': True,
 252         },
 253     }, {
 254         # Untitled broadcast (title is None)
 255         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 256         'info_dict': {
 257             'id': 'v11230755',
 258             'ext': 'mp4',
 259             'title': 'Untitled Broadcast',
 260             'thumbnail': 're:^https?://.*\.jpg$',
 261             'duration': 1638,
 262             'timestamp': 1439746708,
 263             'upload_date': '20150816',
 264             'uploader': 'BelkAO_o',
 265             'uploader_id': 'belkao_o',
 266             'view_count': int,
 267         },
 268         'params': {
 269             # m3u8 download
 270             'skip_download': True,
 271         },
 272         'skip': 'HTTP Error 404: Not Found',
 273     }]
 274
 275     def _real_extract(self, url):
 276         item_id = self._match_id(url)
 277
 278         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 279         access_token = self._call_api(
 280             'api/vods/%s/access_token' % item_id, item_id,
 281             'Downloading %s access token' % self._ITEM_TYPE)
 282
 283         formats = self._extract_m3u8_formats(
 284             '%s/vod/%s?%s' % (
 285                 self._USHER_BASE, item_id,
 286                 compat_urllib_parse_urlencode({
 287                     'allow_source': 'true',
 288                     'allow_audio_only': 'true',
 289                     'allow_spectre': 'true',
 290                     'player': 'twitchweb',
 291                     'nauth': access_token['token'],
 292                     'nauthsig': access_token['sig'],
 293                 })),
 294             item_id, 'mp4', entry_protocol='m3u8_native')
 295
 296         self._prefer_source(formats)
 297         info['formats'] = formats
 298
 299         parsed_url = compat_urllib_parse_urlparse(url)
 300         query = compat_parse_qs(parsed_url.query)
 301         if 't' in query:
 302             info['start_time'] = parse_duration(query['t'][0])
 303
 304         if info.get('timestamp') is not None:
 305             info['subtitles'] = {
 306                 'rechat': [{
 307                     'url': update_url_query(
 308                         'https://rechat.twitch.tv/rechat-messages', {
 309                             'video_id': 'v%s' % item_id,
 310                             'start': info['timestamp'],
 311                         }),
 312                     'ext': 'json',
 313                 }],
 314             }
 315
 316         return info
 317
 318
 319 class TwitchPlaylistBaseIE(TwitchBaseIE):
 320     _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
 321     _PAGE_LIMIT = 100
 322
 323     def _extract_playlist(self, channel_id):
 324         info = self._call_api(
 325             'kraken/channels/%s' % channel_id,
 326             channel_id, 'Downloading channel info JSON')
 327         channel_name = info.get('display_name') or info.get('name')
 328         entries = []
 329         offset = 0
 330         limit = self._PAGE_LIMIT
 331         broken_paging_detected = False
 332         counter_override = None
 333         for counter in itertools.count(1):
 334             response = self._call_api(
 335                 self._PLAYLIST_PATH % (channel_id, offset, limit),
 336                 channel_id,
 337                 'Downloading %s JSON page %s'
 338                 % (self._PLAYLIST_TYPE, counter_override or counter))
 339             page_entries = self._extract_playlist_page(response)
 340             if not page_entries:
 341                 break
 342             total = int_or_none(response.get('_total'))
 343             # Since the beginning of March 2016 twitch's paging mechanism
 344             # is completely broken on the twitch side. It simply ignores
 345             # a limit and returns the whole offset number of videos.
 346             # Working around by just requesting all videos at once.
 347             # Upd: pagination bug was fixed by twitch on 15.03.2016.
 348             if not broken_paging_detected and total and len(page_entries) > limit:
 349                 self.report_warning(
 350                     'Twitch pagination is broken on twitch side, requesting all videos at once',
 351                     channel_id)
 352                 broken_paging_detected = True
 353                 offset = total
 354                 counter_override = '(all at once)'
 355                 continue
 356             entries.extend(page_entries)
 357             if broken_paging_detected or total and len(page_entries) >= total:
 358                 break
 359             offset += limit
 360         return self.playlist_result(
 361             [self.url_result(entry) for entry in orderedSet(entries)],
 362             channel_id, channel_name)
 363
 364     def _extract_playlist_page(self, response):
 365         videos = response.get('videos')
 366         return [video['url'] for video in videos] if videos else []
 367
 368     def _real_extract(self, url):
 369         return self._extract_playlist(self._match_id(url))
 370
 371
 372 class TwitchProfileIE(TwitchPlaylistBaseIE):
 373     IE_NAME = 'twitch:profile'
 374     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 375     _PLAYLIST_TYPE = 'profile'
 376
 377     _TEST = {
 378         'url': 'http://www.twitch.tv/vanillatv/profile',
 379         'info_dict': {
 380             'id': 'vanillatv',
 381             'title': 'VanillaTV',
 382         },
 383         'playlist_mincount': 412,
 384     }
 385
 386
 387 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
 388     _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
 389     _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
 390
 391
 392 class TwitchAllVideosIE(TwitchVideosBaseIE):
 393     IE_NAME = 'twitch:videos:all'
 394     _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 395     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
 396     _PLAYLIST_TYPE = 'all videos'
 397
 398     _TEST = {
 399         'url': 'https://www.twitch.tv/spamfish/videos/all',
 400         'info_dict': {
 401             'id': 'spamfish',
 402             'title': 'Spamfish',
 403         },
 404         'playlist_mincount': 869,
 405     }
 406
 407
 408 class TwitchUploadsIE(TwitchVideosBaseIE):
 409     IE_NAME = 'twitch:videos:uploads'
 410     _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 411     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
 412     _PLAYLIST_TYPE = 'uploads'
 413
 414     _TEST = {
 415         'url': 'https://www.twitch.tv/spamfish/videos/uploads',
 416         'info_dict': {
 417             'id': 'spamfish',
 418             'title': 'Spamfish',
 419         },
 420         'playlist_mincount': 0,
 421     }
 422
 423
 424 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
 425     IE_NAME = 'twitch:videos:past-broadcasts'
 426     _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 427     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
 428     _PLAYLIST_TYPE = 'past broadcasts'
 429
 430     _TEST = {
 431         'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
 432         'info_dict': {
 433             'id': 'spamfish',
 434             'title': 'Spamfish',
 435         },
 436         'playlist_mincount': 0,
 437     }
 438
 439
 440 class TwitchHighlightsIE(TwitchVideosBaseIE):
 441     IE_NAME = 'twitch:videos:highlights'
 442     _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 443     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
 444     _PLAYLIST_TYPE = 'highlights'
 445
 446     _TEST = {
 447         'url': 'https://www.twitch.tv/spamfish/videos/highlights',
 448         'info_dict': {
 449             'id': 'spamfish',
 450             'title': 'Spamfish',
 451         },
 452         'playlist_mincount': 805,
 453     }
 454
 455
 456 class TwitchStreamIE(TwitchBaseIE):
 457     IE_NAME = 'twitch:stream'
 458     _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 459
 460     _TESTS = [{
 461         'url': 'http://www.twitch.tv/shroomztv',
 462         'info_dict': {
 463             'id': '12772022048',
 464             'display_id': 'shroomztv',
 465             'ext': 'mp4',
 466             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 467             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 468             'is_live': True,
 469             'timestamp': 1421928037,
 470             'upload_date': '20150122',
 471             'uploader': 'ShroomzTV',
 472             'uploader_id': 'shroomztv',
 473             'view_count': int,
 474         },
 475         'params': {
 476             # m3u8 download
 477             'skip_download': True,
 478         },
 479     }, {
 480         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 481         'only_matching': True,
 482     }]
 483
 484     def _real_extract(self, url):
 485         channel_id = self._match_id(url)
 486
 487         stream = self._call_api(
 488             'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
 489             'Downloading stream JSON').get('stream')
 490
 491         if not stream:
 492             raise ExtractorError('%s is offline' % channel_id, expected=True)
 493
 494         # Channel name may be typed if different case than the original channel name
 495         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 496         # an invalid m3u8 URL. Working around by use of original channel name from stream
 497         # JSON and fallback to lowercase if it's not available.
 498         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 499
 500         access_token = self._call_api(
 501             'api/channels/%s/access_token' % channel_id, channel_id,
 502             'Downloading channel access token')
 503
 504         query = {
 505             'allow_source': 'true',
 506             'allow_audio_only': 'true',
 507             'allow_spectre': 'true',
 508             'p': random.randint(1000000, 10000000),
 509             'player': 'twitchweb',
 510             'segment_preference': '4',
 511             'sig': access_token['sig'].encode('utf-8'),
 512             'token': access_token['token'].encode('utf-8'),
 513         }
 514         formats = self._extract_m3u8_formats(
 515             '%s/api/channel/hls/%s.m3u8?%s'
 516             % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
 517             channel_id, 'mp4')
 518         self._prefer_source(formats)
 519
 520         view_count = stream.get('viewers')
 521         timestamp = parse_iso8601(stream.get('created_at'))
 522
 523         channel = stream['channel']
 524         title = self._live_title(channel.get('display_name') or channel.get('name'))
 525         description = channel.get('status')
 526
 527         thumbnails = []
 528         for thumbnail_key, thumbnail_url in stream['preview'].items():
 529             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 530             if not m:
 531                 continue
 532             thumbnails.append({
 533                 'url': thumbnail_url,
 534                 'width': int(m.group('width')),
 535                 'height': int(m.group('height')),
 536             })
 537
 538         return {
 539             'id': compat_str(stream['_id']),
 540             'display_id': channel_id,
 541             'title': title,
 542             'description': description,
 543             'thumbnails': thumbnails,
 544             'uploader': channel.get('display_name'),
 545             'uploader_id': channel.get('name'),
 546             'timestamp': timestamp,
 547             'view_count': view_count,
 548             'formats': formats,
 549             'is_live': True,
 550         }
 551
 552
 553 class TwitchClipsIE(InfoExtractor):
 554     IE_NAME = 'twitch:clips'
 555     _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
 556
 557     _TESTS = [{
 558         'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
 559         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 560         'info_dict': {
 561             'id': 'AggressiveCobraPoooound',
 562             'ext': 'mp4',
 563             'title': 'EA Play 2016 Live from the Novo Theatre',
 564             'thumbnail': 're:^https?://.*\.jpg',
 565             'creator': 'EA',
 566             'uploader': 'stereotype_',
 567             'uploader_id': 'stereotype_',
 568         },
 569     }, {
 570         # multiple formats
 571         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 572         'only_matching': True,
 573     }]
 574
 575     def _real_extract(self, url):
 576         video_id = self._match_id(url)
 577
 578         webpage = self._download_webpage(url, video_id)
 579
 580         clip = self._parse_json(
 581             self._search_regex(
 582                 r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
 583             video_id, transform_source=js_to_json)
 584
 585         title = clip.get('channel_title') or self._og_search_title(webpage)
 586
 587         formats = [{
 588             'url': option['source'],
 589             'format_id': option.get('quality'),
 590             'height': int_or_none(option.get('quality')),
 591         } for option in clip.get('quality_options', []) if option.get('source')]
 592
 593         if not formats:
 594             formats = [{
 595                 'url': clip['clip_video_url'],
 596             }]
 597
 598         self._sort_formats(formats)
 599
 600         return {
 601             'id': video_id,
 602             'title': title,
 603             'thumbnail': self._og_search_thumbnail(webpage),
 604             'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
 605             'uploader': clip.get('curator_login'),
 606             'uploader_id': clip.get('curator_display_name'),
 607             'formats': formats,
 608         }