youtube_dl/extractor/twitch.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_str,
  10     compat_urllib_parse,
  11     compat_urllib_request,
  12 )
  13 from ..utils import (
  14     ExtractorError,
  15     parse_iso8601,
  16 )
  17
  18
  19 class TwitchBaseIE(InfoExtractor):
  20     _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
  21
  22     _API_BASE = 'https://api.twitch.tv'
  23     _USHER_BASE = 'http://usher.twitch.tv'
  24     _LOGIN_URL = 'https://secure.twitch.tv/user/login'
  25
  26     def _handle_error(self, response):
  27         if not isinstance(response, dict):
  28             return
  29         error = response.get('error')
  30         if error:
  31             raise ExtractorError(
  32                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  33                 expected=True)
  34
  35     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  36         response = super(TwitchBaseIE, self)._download_json(url, video_id, note)
  37         self._handle_error(response)
  38         return response
  39
  40     def _real_initialize(self):
  41         self._login()
  42
  43     def _login(self):
  44         (username, password) = self._get_login_info()
  45         if username is None:
  46             return
  47
  48         login_page = self._download_webpage(
  49             self._LOGIN_URL, None, 'Downloading login page')
  50
  51         authenticity_token = self._search_regex(
  52             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  53             login_page, 'authenticity token')
  54
  55         login_form = {
  56             'utf8': '✓'.encode('utf-8'),
  57             'authenticity_token': authenticity_token,
  58             'redirect_on_login': '',
  59             'embed_form': 'false',
  60             'mp_source_action': '',
  61             'follow': '',
  62             'user[login]': username,
  63             'user[password]': password,
  64         }
  65
  66         request = compat_urllib_request.Request(
  67             self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  68         request.add_header('Referer', self._LOGIN_URL)
  69         response = self._download_webpage(
  70             request, None, 'Logging in as %s' % username)
  71
  72         m = re.search(
  73             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  74         if m:
  75             raise ExtractorError(
  76                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  77
  78
  79 class TwitchItemBaseIE(TwitchBaseIE):
  80     def _download_info(self, item, item_id):
  81         return self._extract_info(self._download_json(
  82             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  83             'Downloading %s info JSON' % self._ITEM_TYPE))
  84
  85     def _extract_media(self, item_id):
  86         info = self._download_info(self._ITEM_SHORTCUT, item_id)
  87         response = self._download_json(
  88             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
  89             'Downloading %s playlist JSON' % self._ITEM_TYPE)
  90         entries = []
  91         chunks = response['chunks']
  92         qualities = list(chunks.keys())
  93         for num, fragment in enumerate(zip(*chunks.values()), start=1):
  94             formats = []
  95             for fmt_num, fragment_fmt in enumerate(fragment):
  96                 format_id = qualities[fmt_num]
  97                 fmt = {
  98                     'url': fragment_fmt['url'],
  99                     'format_id': format_id,
 100                     'quality': 1 if format_id == 'live' else 0,
 101                 }
 102                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 103                 if m:
 104                     fmt['height'] = int(m.group('height'))
 105                 formats.append(fmt)
 106             self._sort_formats(formats)
 107             entry = dict(info)
 108             entry['id'] = '%s_%d' % (entry['id'], num)
 109             entry['title'] = '%s part %d' % (entry['title'], num)
 110             entry['formats'] = formats
 111             entries.append(entry)
 112         return self.playlist_result(entries, info['id'], info['title'])
 113
 114     def _extract_info(self, info):
 115         return {
 116             'id': info['_id'],
 117             'title': info['title'],
 118             'description': info['description'],
 119             'duration': info['length'],
 120             'thumbnail': info['preview'],
 121             'uploader': info['channel']['display_name'],
 122             'uploader_id': info['channel']['name'],
 123             'timestamp': parse_iso8601(info['recorded_at']),
 124             'view_count': info['views'],
 125         }
 126
 127     def _real_extract(self, url):
 128         return self._extract_media(self._match_id(url))
 129
 130
 131 class TwitchVideoIE(TwitchItemBaseIE):
 132     IE_NAME = 'twitch:video'
 133     _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 134     _ITEM_TYPE = 'video'
 135     _ITEM_SHORTCUT = 'a'
 136
 137     _TEST = {
 138         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 139         'info_dict': {
 140             'id': 'a577357806',
 141             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 142         },
 143         'playlist_mincount': 12,
 144     }
 145
 146
 147 class TwitchChapterIE(TwitchItemBaseIE):
 148     IE_NAME = 'twitch:chapter'
 149     _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 150     _ITEM_TYPE = 'chapter'
 151     _ITEM_SHORTCUT = 'c'
 152
 153     _TESTS = [{
 154         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 155         'info_dict': {
 156             'id': 'c5285812',
 157             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 158         },
 159         'playlist_mincount': 3,
 160     }, {
 161         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 162         'only_matching': True,
 163     }]
 164
 165
 166 class TwitchVodIE(TwitchItemBaseIE):
 167     IE_NAME = 'twitch:vod'
 168     _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 169     _ITEM_TYPE = 'vod'
 170     _ITEM_SHORTCUT = 'v'
 171
 172     _TEST = {
 173         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 174         'info_dict': {
 175             'id': 'v3622000',
 176             'ext': 'mp4',
 177             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 178             'thumbnail': 're:^https?://.*\.jpg$',
 179             'duration': 6951,
 180             'timestamp': 1419028564,
 181             'upload_date': '20141219',
 182             'uploader': 'KSPTV',
 183             'uploader_id': 'ksptv',
 184             'view_count': int,
 185         },
 186         'params': {
 187             # m3u8 download
 188             'skip_download': True,
 189         },
 190     }
 191
 192     def _real_extract(self, url):
 193         item_id = self._match_id(url)
 194         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 195         access_token = self._download_json(
 196             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 197             'Downloading %s access token' % self._ITEM_TYPE)
 198         formats = self._extract_m3u8_formats(
 199             '%s/vod/%s?nauth=%s&nauthsig=%s'
 200             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 201             item_id, 'mp4')
 202         info['formats'] = formats
 203         return info
 204
 205
 206 class TwitchPlaylistBaseIE(TwitchBaseIE):
 207     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 208     _PAGE_LIMIT = 100
 209
 210     def _extract_playlist(self, channel_id):
 211         info = self._download_json(
 212             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 213             channel_id, 'Downloading channel info JSON')
 214         channel_name = info.get('display_name') or info.get('name')
 215         entries = []
 216         offset = 0
 217         limit = self._PAGE_LIMIT
 218         for counter in itertools.count(1):
 219             response = self._download_json(
 220                 self._PLAYLIST_URL % (channel_id, offset, limit),
 221                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 222             videos = response['videos']
 223             if not videos:
 224                 break
 225             entries.extend([self.url_result(video['url']) for video in videos])
 226             offset += limit
 227         return self.playlist_result(entries, channel_id, channel_name)
 228
 229     def _real_extract(self, url):
 230         return self._extract_playlist(self._match_id(url))
 231
 232
 233 class TwitchProfileIE(TwitchPlaylistBaseIE):
 234     IE_NAME = 'twitch:profile'
 235     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 236     _PLAYLIST_TYPE = 'profile'
 237
 238     _TEST = {
 239         'url': 'http://www.twitch.tv/vanillatv/profile',
 240         'info_dict': {
 241             'id': 'vanillatv',
 242             'title': 'VanillaTV',
 243         },
 244         'playlist_mincount': 412,
 245     }
 246
 247
 248 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 249     IE_NAME = 'twitch:past_broadcasts'
 250     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 251     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 252     _PLAYLIST_TYPE = 'past broadcasts'
 253
 254     _TEST = {
 255         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 256         'info_dict': {
 257             'id': 'spamfish',
 258             'title': 'Spamfish',
 259         },
 260         'playlist_mincount': 54,
 261     }
 262
 263
 264 class TwitchStreamIE(TwitchBaseIE):
 265     IE_NAME = 'twitch:stream'
 266     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 267
 268     _TEST = {
 269         'url': 'http://www.twitch.tv/shroomztv',
 270         'info_dict': {
 271             'id': '12772022048',
 272             'display_id': 'shroomztv',
 273             'ext': 'mp4',
 274             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 275             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 276             'is_live': True,
 277             'timestamp': 1421928037,
 278             'upload_date': '20150122',
 279             'uploader': 'ShroomzTV',
 280             'uploader_id': 'shroomztv',
 281             'view_count': int,
 282         },
 283         'params': {
 284             # m3u8 download
 285             'skip_download': True,
 286         },
 287     }
 288
 289     def _real_extract(self, url):
 290         channel_id = self._match_id(url)
 291
 292         stream = self._download_json(
 293             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 294             'Downloading stream JSON').get('stream')
 295
 296         # Fallback on profile extraction if stream is offline
 297         if not stream:
 298             return self.url_result(
 299                 'http://www.twitch.tv/%s/profile' % channel_id,
 300                 'TwitchProfile', channel_id)
 301
 302         access_token = self._download_json(
 303             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 304             'Downloading channel access token')
 305
 306         query = {
 307             'allow_source': 'true',
 308             'p': '9386337',
 309             'player': 'twitchweb',
 310             'segment_preference': '4',
 311             'sig': access_token['sig'],
 312             'token': access_token['token'],
 313         }
 314
 315         formats = self._extract_m3u8_formats(
 316             '%s/api/channel/hls/%s.m3u8?%s'
 317             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query).encode('utf-8')),
 318             channel_id, 'mp4')
 319
 320         view_count = stream.get('viewers')
 321         timestamp = parse_iso8601(stream.get('created_at'))
 322
 323         channel = stream['channel']
 324         title = self._live_title(channel.get('display_name') or channel.get('name'))
 325         description = channel.get('status')
 326
 327         thumbnails = []
 328         for thumbnail_key, thumbnail_url in stream['preview'].items():
 329             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 330             if not m:
 331                 continue
 332             thumbnails.append({
 333                 'url': thumbnail_url,
 334                 'width': int(m.group('width')),
 335                 'height': int(m.group('height')),
 336             })
 337
 338         return {
 339             'id': compat_str(stream['_id']),
 340             'display_id': channel_id,
 341             'title': title,
 342             'description': description,
 343             'thumbnails': thumbnails,
 344             'uploader': channel.get('display_name'),
 345             'uploader_id': channel.get('name'),
 346             'timestamp': timestamp,
 347             'view_count': view_count,
 348             'formats': formats,
 349             'is_live': True,
 350         }