_ 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_parse_qs,
  11     compat_str,
  12     compat_urllib_parse,
  13     compat_urllib_parse_urlparse,
  14     compat_urllib_request,
  15 )
  16 from ..utils import (
  17     ExtractorError,
  18     int_or_none,
  19     parse_duration,
  20     parse_iso8601,
  21 )
  22
  23
  24 class TwitchBaseIE(InfoExtractor):
  25     _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
  26
  27     _API_BASE = 'https://api.twitch.tv'
  28     _USHER_BASE = 'http://usher.twitch.tv'
  29     _LOGIN_URL = 'https://secure.twitch.tv/login'
  30     _LOGIN_POST_URL = 'https://passport.twitch.tv/authorize'
  31     _NETRC_MACHINE = 'twitch'
  32
  33     def _handle_error(self, response):
  34         if not isinstance(response, dict):
  35             return
  36         error = response.get('error')
  37         if error:
  38             raise ExtractorError(
  39                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  40                 expected=True)
  41
  42     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  43         headers = {
  44             'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
  45             'X-Requested-With': 'XMLHttpRequest',
  46         }
  47         for cookie in self._downloader.cookiejar:
  48             if cookie.name == 'api_token':
  49                 headers['Twitch-Api-Token'] = cookie.value
  50         request = compat_urllib_request.Request(url, headers=headers)
  51         response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
  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         login_page = self._download_webpage(
  64             self._LOGIN_URL, None, 'Downloading login page')
  65
  66         login_form = self._hidden_inputs(login_page)
  67
  68         login_form.update({
  69             'login': username.encode('utf-8'),
  70             'password': password.encode('utf-8'),
  71         })
  72
  73         request = compat_urllib_request.Request(
  74             self._LOGIN_POST_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  75         request.add_header('Referer', self._LOGIN_URL)
  76         response = self._download_webpage(
  77             request, None, 'Logging in as %s' % username)
  78
  79         error_message = self._search_regex(
  80             r'<div[^>]+class="subwindow_notice"[^>]*>([^<]+)</div>',
  81             response, 'error message', default=None)
  82         if error_message:
  83             raise ExtractorError(
  84                 'Unable to login. Twitch said: %s' % error_message, expected=True)
  85
  86         if '>Reset your password<' in response:
  87             self.report_warning('Twitch asks you to reset your password, go to https://secure.twitch.tv/reset/submit')
  88
  89     def _prefer_source(self, formats):
  90         try:
  91             source = next(f for f in formats if f['format_id'] == 'Source')
  92             source['preference'] = 10
  93         except StopIteration:
  94             pass  # No Source stream present
  95         self._sort_formats(formats)
  96
  97
  98 class TwitchItemBaseIE(TwitchBaseIE):
  99     def _download_info(self, item, item_id):
 100         return self._extract_info(self._download_json(
 101             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
 102             'Downloading %s info JSON' % self._ITEM_TYPE))
 103
 104     def _extract_media(self, item_id):
 105         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 106         response = self._download_json(
 107             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
 108             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 109         entries = []
 110         chunks = response['chunks']
 111         qualities = list(chunks.keys())
 112         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 113             formats = []
 114             for fmt_num, fragment_fmt in enumerate(fragment):
 115                 format_id = qualities[fmt_num]
 116                 fmt = {
 117                     'url': fragment_fmt['url'],
 118                     'format_id': format_id,
 119                     'quality': 1 if format_id == 'live' else 0,
 120                 }
 121                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 122                 if m:
 123                     fmt['height'] = int(m.group('height'))
 124                 formats.append(fmt)
 125             self._sort_formats(formats)
 126             entry = dict(info)
 127             entry['id'] = '%s_%d' % (entry['id'], num)
 128             entry['title'] = '%s part %d' % (entry['title'], num)
 129             entry['formats'] = formats
 130             entries.append(entry)
 131         return self.playlist_result(entries, info['id'], info['title'])
 132
 133     def _extract_info(self, info):
 134         return {
 135             'id': info['_id'],
 136             'title': info.get('title') or 'Untitled Broadcast',
 137             'description': info.get('description'),
 138             'duration': int_or_none(info.get('length')),
 139             'thumbnail': info.get('preview'),
 140             'uploader': info.get('channel', {}).get('display_name'),
 141             'uploader_id': info.get('channel', {}).get('name'),
 142             'timestamp': parse_iso8601(info.get('recorded_at')),
 143             'view_count': int_or_none(info.get('views')),
 144         }
 145
 146     def _real_extract(self, url):
 147         return self._extract_media(self._match_id(url))
 148
 149
 150 class TwitchVideoIE(TwitchItemBaseIE):
 151     IE_NAME = 'twitch:video'
 152     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 153     _ITEM_TYPE = 'video'
 154     _ITEM_SHORTCUT = 'a'
 155
 156     _TEST = {
 157         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 158         'info_dict': {
 159             'id': 'a577357806',
 160             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 161         },
 162         'playlist_mincount': 12,
 163     }
 164
 165
 166 class TwitchChapterIE(TwitchItemBaseIE):
 167     IE_NAME = 'twitch:chapter'
 168     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 169     _ITEM_TYPE = 'chapter'
 170     _ITEM_SHORTCUT = 'c'
 171
 172     _TESTS = [{
 173         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 174         'info_dict': {
 175             'id': 'c5285812',
 176             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 177         },
 178         'playlist_mincount': 3,
 179     }, {
 180         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 181         'only_matching': True,
 182     }]
 183
 184
 185 class TwitchVodIE(TwitchItemBaseIE):
 186     IE_NAME = 'twitch:vod'
 187     _VALID_URL = r'%s/[^/]+/v/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 188     _ITEM_TYPE = 'vod'
 189     _ITEM_SHORTCUT = 'v'
 190
 191     _TESTS = [{
 192         'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
 193         'info_dict': {
 194             'id': 'v6528877',
 195             'ext': 'mp4',
 196             'title': 'LCK Summer Split - Week 6 Day 1',
 197             'thumbnail': 're:^https?://.*\.jpg$',
 198             'duration': 17208,
 199             'timestamp': 1435131709,
 200             'upload_date': '20150624',
 201             'uploader': 'Riot Games',
 202             'uploader_id': 'riotgames',
 203             'view_count': int,
 204             'start_time': 310,
 205         },
 206         'params': {
 207             # m3u8 download
 208             'skip_download': True,
 209         },
 210     }, {
 211         # Untitled broadcast (title is None)
 212         'url': 'http://www.twitch.tv/belkao_o/v/11230755',
 213         'info_dict': {
 214             'id': 'v11230755',
 215             'ext': 'mp4',
 216             'title': 'Untitled Broadcast',
 217             'thumbnail': 're:^https?://.*\.jpg$',
 218             'duration': 1638,
 219             'timestamp': 1439746708,
 220             'upload_date': '20150816',
 221             'uploader': 'BelkAO_o',
 222             'uploader_id': 'belkao_o',
 223             'view_count': int,
 224         },
 225         'params': {
 226             # m3u8 download
 227             'skip_download': True,
 228         },
 229     }]
 230
 231     def _real_extract(self, url):
 232         item_id = self._match_id(url)
 233         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 234         access_token = self._download_json(
 235             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 236             'Downloading %s access token' % self._ITEM_TYPE)
 237         formats = self._extract_m3u8_formats(
 238             '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true'
 239             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 240             item_id, 'mp4')
 241         self._prefer_source(formats)
 242         info['formats'] = formats
 243
 244         parsed_url = compat_urllib_parse_urlparse(url)
 245         query = compat_parse_qs(parsed_url.query)
 246         if 't' in query:
 247             info['start_time'] = parse_duration(query['t'][0])
 248
 249         return info
 250
 251
 252 class TwitchPlaylistBaseIE(TwitchBaseIE):
 253     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 254     _PAGE_LIMIT = 100
 255
 256     def _extract_playlist(self, channel_id):
 257         info = self._download_json(
 258             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 259             channel_id, 'Downloading channel info JSON')
 260         channel_name = info.get('display_name') or info.get('name')
 261         entries = []
 262         offset = 0
 263         limit = self._PAGE_LIMIT
 264         for counter in itertools.count(1):
 265             response = self._download_json(
 266                 self._PLAYLIST_URL % (channel_id, offset, limit),
 267                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 268             page_entries = self._extract_playlist_page(response)
 269             if not page_entries:
 270                 break
 271             entries.extend(page_entries)
 272             offset += limit
 273         return self.playlist_result(
 274             [self.url_result(entry) for entry in set(entries)],
 275             channel_id, channel_name)
 276
 277     def _extract_playlist_page(self, response):
 278         videos = response.get('videos')
 279         return [video['url'] for video in videos] if videos else []
 280
 281     def _real_extract(self, url):
 282         return self._extract_playlist(self._match_id(url))
 283
 284
 285 class TwitchProfileIE(TwitchPlaylistBaseIE):
 286     IE_NAME = 'twitch:profile'
 287     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 288     _PLAYLIST_TYPE = 'profile'
 289
 290     _TEST = {
 291         'url': 'http://www.twitch.tv/vanillatv/profile',
 292         'info_dict': {
 293             'id': 'vanillatv',
 294             'title': 'VanillaTV',
 295         },
 296         'playlist_mincount': 412,
 297     }
 298
 299
 300 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 301     IE_NAME = 'twitch:past_broadcasts'
 302     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 303     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 304     _PLAYLIST_TYPE = 'past broadcasts'
 305
 306     _TEST = {
 307         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 308         'info_dict': {
 309             'id': 'spamfish',
 310             'title': 'Spamfish',
 311         },
 312         'playlist_mincount': 54,
 313     }
 314
 315
 316 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 317     IE_NAME = 'twitch:bookmarks'
 318     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 319     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 320     _PLAYLIST_TYPE = 'bookmarks'
 321
 322     _TEST = {
 323         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 324         'info_dict': {
 325             'id': 'ognos',
 326             'title': 'Ognos',
 327         },
 328         'playlist_mincount': 3,
 329     }
 330
 331     def _extract_playlist_page(self, response):
 332         entries = []
 333         for bookmark in response.get('bookmarks', []):
 334             video = bookmark.get('video')
 335             if not video:
 336                 continue
 337             entries.append(video['url'])
 338         return entries
 339
 340
 341 class TwitchStreamIE(TwitchBaseIE):
 342     IE_NAME = 'twitch:stream'
 343     _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 344
 345     _TESTS = [{
 346         'url': 'http://www.twitch.tv/shroomztv',
 347         'info_dict': {
 348             'id': '12772022048',
 349             'display_id': 'shroomztv',
 350             'ext': 'mp4',
 351             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 352             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 353             'is_live': True,
 354             'timestamp': 1421928037,
 355             'upload_date': '20150122',
 356             'uploader': 'ShroomzTV',
 357             'uploader_id': 'shroomztv',
 358             'view_count': int,
 359         },
 360         'params': {
 361             # m3u8 download
 362             'skip_download': True,
 363         },
 364     }, {
 365         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 366         'only_matching': True,
 367     }]
 368
 369     def _real_extract(self, url):
 370         channel_id = self._match_id(url)
 371
 372         stream = self._download_json(
 373             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 374             'Downloading stream JSON').get('stream')
 375
 376         # Fallback on profile extraction if stream is offline
 377         if not stream:
 378             return self.url_result(
 379                 'http://www.twitch.tv/%s/profile' % channel_id,
 380                 'TwitchProfile', channel_id)
 381
 382         # Channel name may be typed if different case than the original channel name
 383         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 384         # an invalid m3u8 URL. Working around by use of original channel name from stream
 385         # JSON and fallback to lowercase if it's not available.
 386         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 387
 388         access_token = self._download_json(
 389             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 390             'Downloading channel access token')
 391
 392         query = {
 393             'allow_source': 'true',
 394             'p': random.randint(1000000, 10000000),
 395             'player': 'twitchweb',
 396             'segment_preference': '4',
 397             'sig': access_token['sig'].encode('utf-8'),
 398             'token': access_token['token'].encode('utf-8'),
 399         }
 400         formats = self._extract_m3u8_formats(
 401             '%s/api/channel/hls/%s.m3u8?%s'
 402             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 403             channel_id, 'mp4')
 404         self._prefer_source(formats)
 405
 406         view_count = stream.get('viewers')
 407         timestamp = parse_iso8601(stream.get('created_at'))
 408
 409         channel = stream['channel']
 410         title = self._live_title(channel.get('display_name') or channel.get('name'))
 411         description = channel.get('status')
 412
 413         thumbnails = []
 414         for thumbnail_key, thumbnail_url in stream['preview'].items():
 415             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 416             if not m:
 417                 continue
 418             thumbnails.append({
 419                 'url': thumbnail_url,
 420                 'width': int(m.group('width')),
 421                 'height': int(m.group('height')),
 422             })
 423
 424         return {
 425             'id': compat_str(stream['_id']),
 426             'display_id': channel_id,
 427             'title': title,
 428             'description': description,
 429             'thumbnails': thumbnails,
 430             'uploader': channel.get('display_name'),
 431             'uploader_id': channel.get('name'),
 432             'timestamp': timestamp,
 433             'view_count': view_count,
 434             'formats': formats,
 435             'is_live': True,
 436         }