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 = r'''(?x)
 210                     https?://
 211                         (?:
 212                             (?:www\.)?twitch\.tv/[^/]+/v/|
 213                             player\.twitch\.tv/\?.*?\bvideo=v
 214                         )
 215                         (?P<id>\d+)
 216                     '''
 217     _ITEM_TYPE = 'vod'
 218     _ITEM_SHORTCUT = 'v'
 219
 220     _TESTS = [{
 221         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 222         'info_dict': {
 223             'id': 'v6528877',
 224             'ext': 'mp4',
 225             'title': 'LCK Summer Split - Week 6 Day 1',
 226             'thumbnail': r're:^https?://.*\.jpg$',
 227             'duration': 17208,
 228             'timestamp': 1435131709,
 229             'upload_date': '20150624',
 230             'uploader': 'Riot Games',
 231             'uploader_id': 'riotgames',
 232             'view_count': int,
 233             'start_time': 310,
 234         },
 235         'params': {
 236             # m3u8 download
 237             'skip_download': True,
 238         },
 239     }, {
 240         # Untitled broadcast (title is None)
 241         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 242         'info_dict': {
 243             'id': 'v11230755',
 244             'ext': 'mp4',
 245             'title': 'Untitled Broadcast',
 246             'thumbnail': r're:^https?://.*\.jpg$',
 247             'duration': 1638,
 248             'timestamp': 1439746708,
 249             'upload_date': '20150816',
 250             'uploader': 'BelkAO_o',
 251             'uploader_id': 'belkao_o',
 252             'view_count': int,
 253         },
 254         'params': {
 255             # m3u8 download
 256             'skip_download': True,
 257         },
 258         'skip': 'HTTP Error 404: Not Found',
 259     }, {
 260         'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
 261         'only_matching': True,
 262     }]
 263
 264     def _real_extract(self, url):
 265         item_id = self._match_id(url)
 266
 267         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 268         access_token = self._call_api(
 269             'api/vods/%s/access_token' % item_id, item_id,
 270             'Downloading %s access token' % self._ITEM_TYPE)
 271
 272         formats = self._extract_m3u8_formats(
 273             '%s/vod/%s?%s' % (
 274                 self._USHER_BASE, item_id,
 275                 compat_urllib_parse_urlencode({
 276                     'allow_source': 'true',
 277                     'allow_audio_only': 'true',
 278                     'allow_spectre': 'true',
 279                     'player': 'twitchweb',
 280                     'nauth': access_token['token'],
 281                     'nauthsig': access_token['sig'],
 282                 })),
 283             item_id, 'mp4', entry_protocol='m3u8_native')
 284
 285         self._prefer_source(formats)
 286         info['formats'] = formats
 287
 288         parsed_url = compat_urllib_parse_urlparse(url)
 289         query = compat_parse_qs(parsed_url.query)
 290         if 't' in query:
 291             info['start_time'] = parse_duration(query['t'][0])
 292
 293         if info.get('timestamp') is not None:
 294             info['subtitles'] = {
 295                 'rechat': [{
 296                     'url': update_url_query(
 297                         'https://rechat.twitch.tv/rechat-messages', {
 298                             'video_id': 'v%s' % item_id,
 299                             'start': info['timestamp'],
 300                         }),
 301                     'ext': 'json',
 302                 }],
 303             }
 304
 305         return info
 306
 307
 308 class TwitchPlaylistBaseIE(TwitchBaseIE):
 309     _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
 310     _PAGE_LIMIT = 100
 311
 312     def _extract_playlist(self, channel_id):
 313         info = self._call_api(
 314             'kraken/channels/%s' % channel_id,
 315             channel_id, 'Downloading channel info JSON')
 316         channel_name = info.get('display_name') or info.get('name')
 317         entries = []
 318         offset = 0
 319         limit = self._PAGE_LIMIT
 320         broken_paging_detected = False
 321         counter_override = None
 322         for counter in itertools.count(1):
 323             response = self._call_api(
 324                 self._PLAYLIST_PATH % (channel_id, offset, limit),
 325                 channel_id,
 326                 'Downloading %s JSON page %s'
 327                 % (self._PLAYLIST_TYPE, counter_override or counter))
 328             page_entries = self._extract_playlist_page(response)
 329             if not page_entries:
 330                 break
 331             total = int_or_none(response.get('_total'))
 332             # Since the beginning of March 2016 twitch's paging mechanism
 333             # is completely broken on the twitch side. It simply ignores
 334             # a limit and returns the whole offset number of videos.
 335             # Working around by just requesting all videos at once.
 336             # Upd: pagination bug was fixed by twitch on 15.03.2016.
 337             if not broken_paging_detected and total and len(page_entries) > limit:
 338                 self.report_warning(
 339                     'Twitch pagination is broken on twitch side, requesting all videos at once',
 340                     channel_id)
 341                 broken_paging_detected = True
 342                 offset = total
 343                 counter_override = '(all at once)'
 344                 continue
 345             entries.extend(page_entries)
 346             if broken_paging_detected or total and len(page_entries) >= total:
 347                 break
 348             offset += limit
 349         return self.playlist_result(
 350             [self.url_result(entry) for entry in orderedSet(entries)],
 351             channel_id, channel_name)
 352
 353     def _extract_playlist_page(self, response):
 354         videos = response.get('videos')
 355         return [video['url'] for video in videos] if videos else []
 356
 357     def _real_extract(self, url):
 358         return self._extract_playlist(self._match_id(url))
 359
 360
 361 class TwitchProfileIE(TwitchPlaylistBaseIE):
 362     IE_NAME = 'twitch:profile'
 363     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 364     _PLAYLIST_TYPE = 'profile'
 365
 366     _TEST = {
 367         'url': 'http://www.twitch.tv/vanillatv/profile',
 368         'info_dict': {
 369             'id': 'vanillatv',
 370             'title': 'VanillaTV',
 371         },
 372         'playlist_mincount': 412,
 373     }
 374
 375
 376 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
 377     _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
 378     _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
 379
 380
 381 class TwitchAllVideosIE(TwitchVideosBaseIE):
 382     IE_NAME = 'twitch:videos:all'
 383     _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 384     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
 385     _PLAYLIST_TYPE = 'all videos'
 386
 387     _TEST = {
 388         'url': 'https://www.twitch.tv/spamfish/videos/all',
 389         'info_dict': {
 390             'id': 'spamfish',
 391             'title': 'Spamfish',
 392         },
 393         'playlist_mincount': 869,
 394     }
 395
 396
 397 class TwitchUploadsIE(TwitchVideosBaseIE):
 398     IE_NAME = 'twitch:videos:uploads'
 399     _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 400     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
 401     _PLAYLIST_TYPE = 'uploads'
 402
 403     _TEST = {
 404         'url': 'https://www.twitch.tv/spamfish/videos/uploads',
 405         'info_dict': {
 406             'id': 'spamfish',
 407             'title': 'Spamfish',
 408         },
 409         'playlist_mincount': 0,
 410     }
 411
 412
 413 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
 414     IE_NAME = 'twitch:videos:past-broadcasts'
 415     _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 416     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
 417     _PLAYLIST_TYPE = 'past broadcasts'
 418
 419     _TEST = {
 420         'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
 421         'info_dict': {
 422             'id': 'spamfish',
 423             'title': 'Spamfish',
 424         },
 425         'playlist_mincount': 0,
 426     }
 427
 428
 429 class TwitchHighlightsIE(TwitchVideosBaseIE):
 430     IE_NAME = 'twitch:videos:highlights'
 431     _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
 432     _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
 433     _PLAYLIST_TYPE = 'highlights'
 434
 435     _TEST = {
 436         'url': 'https://www.twitch.tv/spamfish/videos/highlights',
 437         'info_dict': {
 438             'id': 'spamfish',
 439             'title': 'Spamfish',
 440         },
 441         'playlist_mincount': 805,
 442     }
 443
 444
 445 class TwitchStreamIE(TwitchBaseIE):
 446     IE_NAME = 'twitch:stream'
 447     _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 448
 449     _TESTS = [{
 450         'url': 'http://www.twitch.tv/shroomztv',
 451         'info_dict': {
 452             'id': '12772022048',
 453             'display_id': 'shroomztv',
 454             'ext': 'mp4',
 455             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 456             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 457             'is_live': True,
 458             'timestamp': 1421928037,
 459             'upload_date': '20150122',
 460             'uploader': 'ShroomzTV',
 461             'uploader_id': 'shroomztv',
 462             'view_count': int,
 463         },
 464         'params': {
 465             # m3u8 download
 466             'skip_download': True,
 467         },
 468     }, {
 469         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 470         'only_matching': True,
 471     }]
 472
 473     def _real_extract(self, url):
 474         channel_id = self._match_id(url)
 475
 476         stream = self._call_api(
 477             'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
 478             'Downloading stream JSON').get('stream')
 479
 480         if not stream:
 481             raise ExtractorError('%s is offline' % channel_id, expected=True)
 482
 483         # Channel name may be typed if different case than the original channel name
 484         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 485         # an invalid m3u8 URL. Working around by use of original channel name from stream
 486         # JSON and fallback to lowercase if it's not available.
 487         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 488
 489         access_token = self._call_api(
 490             'api/channels/%s/access_token' % channel_id, channel_id,
 491             'Downloading channel access token')
 492
 493         query = {
 494             'allow_source': 'true',
 495             'allow_audio_only': 'true',
 496             'allow_spectre': 'true',
 497             'p': random.randint(1000000, 10000000),
 498             'player': 'twitchweb',
 499             'segment_preference': '4',
 500             'sig': access_token['sig'].encode('utf-8'),
 501             'token': access_token['token'].encode('utf-8'),
 502         }
 503         formats = self._extract_m3u8_formats(
 504             '%s/api/channel/hls/%s.m3u8?%s'
 505             % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
 506             channel_id, 'mp4')
 507         self._prefer_source(formats)
 508
 509         view_count = stream.get('viewers')
 510         timestamp = parse_iso8601(stream.get('created_at'))
 511
 512         channel = stream['channel']
 513         title = self._live_title(channel.get('display_name') or channel.get('name'))
 514         description = channel.get('status')
 515
 516         thumbnails = []
 517         for thumbnail_key, thumbnail_url in stream['preview'].items():
 518             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 519             if not m:
 520                 continue
 521             thumbnails.append({
 522                 'url': thumbnail_url,
 523                 'width': int(m.group('width')),
 524                 'height': int(m.group('height')),
 525             })
 526
 527         return {
 528             'id': compat_str(stream['_id']),
 529             'display_id': channel_id,
 530             'title': title,
 531             'description': description,
 532             'thumbnails': thumbnails,
 533             'uploader': channel.get('display_name'),
 534             'uploader_id': channel.get('name'),
 535             'timestamp': timestamp,
 536             'view_count': view_count,
 537             'formats': formats,
 538             'is_live': True,
 539         }
 540
 541
 542 class TwitchClipsIE(InfoExtractor):
 543     IE_NAME = 'twitch:clips'
 544     _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
 545
 546     _TESTS = [{
 547         'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
 548         'md5': '761769e1eafce0ffebfb4089cb3847cd',
 549         'info_dict': {
 550             'id': 'AggressiveCobraPoooound',
 551             'ext': 'mp4',
 552             'title': 'EA Play 2016 Live from the Novo Theatre',
 553             'thumbnail': r're:^https?://.*\.jpg',
 554             'creator': 'EA',
 555             'uploader': 'stereotype_',
 556             'uploader_id': 'stereotype_',
 557         },
 558     }, {
 559         # multiple formats
 560         'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
 561         'only_matching': True,
 562     }]
 563
 564     def _real_extract(self, url):
 565         video_id = self._match_id(url)
 566
 567         webpage = self._download_webpage(url, video_id)
 568
 569         clip = self._parse_json(
 570             self._search_regex(
 571                 r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
 572             video_id, transform_source=js_to_json)
 573
 574         title = clip.get('channel_title') or self._og_search_title(webpage)
 575
 576         formats = [{
 577             'url': option['source'],
 578             'format_id': option.get('quality'),
 579             'height': int_or_none(option.get('quality')),
 580         } for option in clip.get('quality_options', []) if option.get('source')]
 581
 582         if not formats:
 583             formats = [{
 584                 'url': clip['clip_video_url'],
 585             }]
 586
 587         self._sort_formats(formats)
 588
 589         return {
 590             'id': video_id,
 591             'title': title,
 592             'thumbnail': self._og_search_thumbnail(webpage),
 593             'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
 594             'uploader': clip.get('curator_login'),
 595             'uploader_id': clip.get('curator_display_name'),
 596             'formats': formats,
 597         }