Raphaël G. Git Repositories - youtubedl/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_str,
  11     compat_urllib_parse,
  12     compat_urllib_request,
  13 )
  14 from ..utils import (
  15     ExtractorError,
  16     parse_iso8601,
  17 )
  18
  19
  20 class TwitchBaseIE(InfoExtractor):
  21     _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
  22
  23     _API_BASE = 'https://api.twitch.tv'
  24     _USHER_BASE = 'http://usher.twitch.tv'
  25     _LOGIN_URL = 'https://secure.twitch.tv/user/login'
  26     _LOGIN_POST_URL = 'https://secure-login.twitch.tv/login'
  27     _NETRC_MACHINE = 'twitch'
  28
  29     def _handle_error(self, response):
  30         if not isinstance(response, dict):
  31             return
  32         error = response.get('error')
  33         if error:
  34             raise ExtractorError(
  35                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  36                 expected=True)
  37
  38     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  39         headers = {
  40             'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
  41             'X-Requested-With': 'XMLHttpRequest',
  42         }
  43         for cookie in self._downloader.cookiejar:
  44             if cookie.name == 'api_token':
  45                 headers['Twitch-Api-Token'] = cookie.value
  46         request = compat_urllib_request.Request(url, headers=headers)
  47         response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
  48         self._handle_error(response)
  49         return response
  50
  51     def _real_initialize(self):
  52         self._login()
  53
  54     def _login(self):
  55         (username, password) = self._get_login_info()
  56         if username is None:
  57             return
  58
  59         login_page = self._download_webpage(
  60             self._LOGIN_URL, None, 'Downloading login page')
  61
  62         authenticity_token = self._search_regex(
  63             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  64             login_page, 'authenticity token')
  65
  66         login_form = {
  67             'utf8': '✓'.encode('utf-8'),
  68             'authenticity_token': authenticity_token,
  69             'redirect_on_login': '',
  70             'embed_form': 'false',
  71             'mp_source_action': 'login-button',
  72             'follow': '',
  73             'login': username,
  74             'password': password,
  75         }
  76
  77         request = compat_urllib_request.Request(
  78             self._LOGIN_POST_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  79         request.add_header('Referer', self._LOGIN_URL)
  80         response = self._download_webpage(
  81             request, None, 'Logging in as %s' % username)
  82
  83         m = re.search(
  84             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  85         if m:
  86             raise ExtractorError(
  87                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  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['title'],
 137             'description': info['description'],
 138             'duration': info['length'],
 139             'thumbnail': info['preview'],
 140             'uploader': info['channel']['display_name'],
 141             'uploader_id': info['channel']['name'],
 142             'timestamp': parse_iso8601(info['recorded_at']),
 143             'view_count': info['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     _TEST = {
 192         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 193         'info_dict': {
 194             'id': 'v3622000',
 195             'ext': 'mp4',
 196             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 197             'thumbnail': 're:^https?://.*\.jpg$',
 198             'duration': 6951,
 199             'timestamp': 1419028564,
 200             'upload_date': '20141219',
 201             'uploader': 'KSPTV',
 202             'uploader_id': 'ksptv',
 203             'view_count': int,
 204         },
 205         'params': {
 206             # m3u8 download
 207             'skip_download': True,
 208         },
 209     }
 210
 211     def _real_extract(self, url):
 212         item_id = self._match_id(url)
 213         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 214         access_token = self._download_json(
 215             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 216             'Downloading %s access token' % self._ITEM_TYPE)
 217         formats = self._extract_m3u8_formats(
 218             '%s/vod/%s?nauth=%s&nauthsig=%s'
 219             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 220             item_id, 'mp4')
 221         self._prefer_source(formats)
 222         info['formats'] = formats
 223         return info
 224
 225
 226 class TwitchPlaylistBaseIE(TwitchBaseIE):
 227     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 228     _PAGE_LIMIT = 100
 229
 230     def _extract_playlist(self, channel_id):
 231         info = self._download_json(
 232             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 233             channel_id, 'Downloading channel info JSON')
 234         channel_name = info.get('display_name') or info.get('name')
 235         entries = []
 236         offset = 0
 237         limit = self._PAGE_LIMIT
 238         for counter in itertools.count(1):
 239             response = self._download_json(
 240                 self._PLAYLIST_URL % (channel_id, offset, limit),
 241                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 242             page_entries = self._extract_playlist_page(response)
 243             if not page_entries:
 244                 break
 245             entries.extend(page_entries)
 246             offset += limit
 247         return self.playlist_result(
 248             [self.url_result(entry) for entry in set(entries)],
 249             channel_id, channel_name)
 250
 251     def _extract_playlist_page(self, response):
 252         videos = response.get('videos')
 253         return [video['url'] for video in videos] if videos else []
 254
 255     def _real_extract(self, url):
 256         return self._extract_playlist(self._match_id(url))
 257
 258
 259 class TwitchProfileIE(TwitchPlaylistBaseIE):
 260     IE_NAME = 'twitch:profile'
 261     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 262     _PLAYLIST_TYPE = 'profile'
 263
 264     _TEST = {
 265         'url': 'http://www.twitch.tv/vanillatv/profile',
 266         'info_dict': {
 267             'id': 'vanillatv',
 268             'title': 'VanillaTV',
 269         },
 270         'playlist_mincount': 412,
 271     }
 272
 273
 274 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 275     IE_NAME = 'twitch:past_broadcasts'
 276     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 277     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 278     _PLAYLIST_TYPE = 'past broadcasts'
 279
 280     _TEST = {
 281         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 282         'info_dict': {
 283             'id': 'spamfish',
 284             'title': 'Spamfish',
 285         },
 286         'playlist_mincount': 54,
 287     }
 288
 289
 290 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 291     IE_NAME = 'twitch:bookmarks'
 292     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 293     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 294     _PLAYLIST_TYPE = 'bookmarks'
 295
 296     _TEST = {
 297         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 298         'info_dict': {
 299             'id': 'ognos',
 300             'title': 'Ognos',
 301         },
 302         'playlist_mincount': 3,
 303     }
 304
 305     def _extract_playlist_page(self, response):
 306         entries = []
 307         for bookmark in response.get('bookmarks', []):
 308             video = bookmark.get('video')
 309             if not video:
 310                 continue
 311             entries.append(video['url'])
 312         return entries
 313
 314
 315 class TwitchStreamIE(TwitchBaseIE):
 316     IE_NAME = 'twitch:stream'
 317     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 318
 319     _TEST = {
 320         'url': 'http://www.twitch.tv/shroomztv',
 321         'info_dict': {
 322             'id': '12772022048',
 323             'display_id': 'shroomztv',
 324             'ext': 'mp4',
 325             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 326             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 327             'is_live': True,
 328             'timestamp': 1421928037,
 329             'upload_date': '20150122',
 330             'uploader': 'ShroomzTV',
 331             'uploader_id': 'shroomztv',
 332             'view_count': int,
 333         },
 334         'params': {
 335             # m3u8 download
 336             'skip_download': True,
 337         },
 338     }
 339
 340     def _real_extract(self, url):
 341         channel_id = self._match_id(url)
 342
 343         stream = self._download_json(
 344             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 345             'Downloading stream JSON').get('stream')
 346
 347         # Fallback on profile extraction if stream is offline
 348         if not stream:
 349             return self.url_result(
 350                 'http://www.twitch.tv/%s/profile' % channel_id,
 351                 'TwitchProfile', channel_id)
 352
 353         access_token = self._download_json(
 354             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 355             'Downloading channel access token')
 356
 357         query = {
 358             'allow_source': 'true',
 359             'p': random.randint(1000000, 10000000),
 360             'player': 'twitchweb',
 361             'segment_preference': '4',
 362             'sig': access_token['sig'].encode('utf-8'),
 363             'token': access_token['token'].encode('utf-8'),
 364         }
 365         formats = self._extract_m3u8_formats(
 366             '%s/api/channel/hls/%s.m3u8?%s'
 367             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 368             channel_id, 'mp4')
 369         self._prefer_source(formats)
 370
 371         view_count = stream.get('viewers')
 372         timestamp = parse_iso8601(stream.get('created_at'))
 373
 374         channel = stream['channel']
 375         title = self._live_title(channel.get('display_name') or channel.get('name'))
 376         description = channel.get('status')
 377
 378         thumbnails = []
 379         for thumbnail_key, thumbnail_url in stream['preview'].items():
 380             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 381             if not m:
 382                 continue
 383             thumbnails.append({
 384                 'url': thumbnail_url,
 385                 'width': int(m.group('width')),
 386                 'height': int(m.group('height')),
 387             })
 388
 389         return {
 390             'id': compat_str(stream['_id']),
 391             'display_id': channel_id,
 392             'title': title,
 393             'description': description,
 394             'thumbnails': thumbnails,
 395             'uploader': channel.get('display_name'),
 396             'uploader_id': channel.get('name'),
 397             'timestamp': timestamp,
 398             'view_count': view_count,
 399             'formats': formats,
 400             'is_live': True,
 401         }