_ 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_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
  27     def _handle_error(self, response):
  28         if not isinstance(response, dict):
  29             return
  30         error = response.get('error')
  31         if error:
  32             raise ExtractorError(
  33                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  34                 expected=True)
  35
  36     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  37         response = super(TwitchBaseIE, self)._download_json(url, video_id, note)
  38         self._handle_error(response)
  39         return response
  40
  41     def _real_initialize(self):
  42         self._login()
  43
  44     def _login(self):
  45         (username, password) = self._get_login_info()
  46         if username is None:
  47             return
  48
  49         login_page = self._download_webpage(
  50             self._LOGIN_URL, None, 'Downloading login page')
  51
  52         authenticity_token = self._search_regex(
  53             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  54             login_page, 'authenticity token')
  55
  56         login_form = {
  57             'utf8': '✓'.encode('utf-8'),
  58             'authenticity_token': authenticity_token,
  59             'redirect_on_login': '',
  60             'embed_form': 'false',
  61             'mp_source_action': '',
  62             'follow': '',
  63             'user[login]': username,
  64             'user[password]': password,
  65         }
  66
  67         request = compat_urllib_request.Request(
  68             self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  69         request.add_header('Referer', self._LOGIN_URL)
  70         response = self._download_webpage(
  71             request, None, 'Logging in as %s' % username)
  72
  73         m = re.search(
  74             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  75         if m:
  76             raise ExtractorError(
  77                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  78
  79
  80 class TwitchItemBaseIE(TwitchBaseIE):
  81     def _download_info(self, item, item_id):
  82         return self._extract_info(self._download_json(
  83             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  84             'Downloading %s info JSON' % self._ITEM_TYPE))
  85
  86     def _extract_media(self, item_id):
  87         info = self._download_info(self._ITEM_SHORTCUT, item_id)
  88         response = self._download_json(
  89             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
  90             'Downloading %s playlist JSON' % self._ITEM_TYPE)
  91         entries = []
  92         chunks = response['chunks']
  93         qualities = list(chunks.keys())
  94         for num, fragment in enumerate(zip(*chunks.values()), start=1):
  95             formats = []
  96             for fmt_num, fragment_fmt in enumerate(fragment):
  97                 format_id = qualities[fmt_num]
  98                 fmt = {
  99                     'url': fragment_fmt['url'],
 100                     'format_id': format_id,
 101                     'quality': 1 if format_id == 'live' else 0,
 102                 }
 103                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 104                 if m:
 105                     fmt['height'] = int(m.group('height'))
 106                 formats.append(fmt)
 107             self._sort_formats(formats)
 108             entry = dict(info)
 109             entry['id'] = '%s_%d' % (entry['id'], num)
 110             entry['title'] = '%s part %d' % (entry['title'], num)
 111             entry['formats'] = formats
 112             entries.append(entry)
 113         return self.playlist_result(entries, info['id'], info['title'])
 114
 115     def _extract_info(self, info):
 116         return {
 117             'id': info['_id'],
 118             'title': info['title'],
 119             'description': info['description'],
 120             'duration': info['length'],
 121             'thumbnail': info['preview'],
 122             'uploader': info['channel']['display_name'],
 123             'uploader_id': info['channel']['name'],
 124             'timestamp': parse_iso8601(info['recorded_at']),
 125             'view_count': info['views'],
 126         }
 127
 128     def _real_extract(self, url):
 129         return self._extract_media(self._match_id(url))
 130
 131
 132 class TwitchVideoIE(TwitchItemBaseIE):
 133     IE_NAME = 'twitch:video'
 134     _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 135     _ITEM_TYPE = 'video'
 136     _ITEM_SHORTCUT = 'a'
 137
 138     _TEST = {
 139         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 140         'info_dict': {
 141             'id': 'a577357806',
 142             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 143         },
 144         'playlist_mincount': 12,
 145     }
 146
 147
 148 class TwitchChapterIE(TwitchItemBaseIE):
 149     IE_NAME = 'twitch:chapter'
 150     _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 151     _ITEM_TYPE = 'chapter'
 152     _ITEM_SHORTCUT = 'c'
 153
 154     _TESTS = [{
 155         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 156         'info_dict': {
 157             'id': 'c5285812',
 158             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 159         },
 160         'playlist_mincount': 3,
 161     }, {
 162         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 163         'only_matching': True,
 164     }]
 165
 166
 167 class TwitchVodIE(TwitchItemBaseIE):
 168     IE_NAME = 'twitch:vod'
 169     _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 170     _ITEM_TYPE = 'vod'
 171     _ITEM_SHORTCUT = 'v'
 172
 173     _TEST = {
 174         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 175         'info_dict': {
 176             'id': 'v3622000',
 177             'ext': 'mp4',
 178             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 179             'thumbnail': 're:^https?://.*\.jpg$',
 180             'duration': 6951,
 181             'timestamp': 1419028564,
 182             'upload_date': '20141219',
 183             'uploader': 'KSPTV',
 184             'uploader_id': 'ksptv',
 185             'view_count': int,
 186         },
 187         'params': {
 188             # m3u8 download
 189             'skip_download': True,
 190         },
 191     }
 192
 193     def _real_extract(self, url):
 194         item_id = self._match_id(url)
 195         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 196         access_token = self._download_json(
 197             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 198             'Downloading %s access token' % self._ITEM_TYPE)
 199         formats = self._extract_m3u8_formats(
 200             '%s/vod/%s?nauth=%s&nauthsig=%s'
 201             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 202             item_id, 'mp4')
 203         info['formats'] = formats
 204         return info
 205
 206
 207 class TwitchPlaylistBaseIE(TwitchBaseIE):
 208     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 209     _PAGE_LIMIT = 100
 210
 211     def _extract_playlist(self, channel_id):
 212         info = self._download_json(
 213             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 214             channel_id, 'Downloading channel info JSON')
 215         channel_name = info.get('display_name') or info.get('name')
 216         entries = []
 217         offset = 0
 218         limit = self._PAGE_LIMIT
 219         for counter in itertools.count(1):
 220             response = self._download_json(
 221                 self._PLAYLIST_URL % (channel_id, offset, limit),
 222                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 223             videos = response['videos']
 224             if not videos:
 225                 break
 226             entries.extend([self.url_result(video['url']) for video in videos])
 227             offset += limit
 228         return self.playlist_result(entries, channel_id, channel_name)
 229
 230     def _real_extract(self, url):
 231         return self._extract_playlist(self._match_id(url))
 232
 233
 234 class TwitchProfileIE(TwitchPlaylistBaseIE):
 235     IE_NAME = 'twitch:profile'
 236     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 237     _PLAYLIST_TYPE = 'profile'
 238
 239     _TEST = {
 240         'url': 'http://www.twitch.tv/vanillatv/profile',
 241         'info_dict': {
 242             'id': 'vanillatv',
 243             'title': 'VanillaTV',
 244         },
 245         'playlist_mincount': 412,
 246     }
 247
 248
 249 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 250     IE_NAME = 'twitch:past_broadcasts'
 251     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 252     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 253     _PLAYLIST_TYPE = 'past broadcasts'
 254
 255     _TEST = {
 256         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 257         'info_dict': {
 258             'id': 'spamfish',
 259             'title': 'Spamfish',
 260         },
 261         'playlist_mincount': 54,
 262     }
 263
 264
 265 class TwitchStreamIE(TwitchBaseIE):
 266     IE_NAME = 'twitch:stream'
 267     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 268
 269     _TEST = {
 270         'url': 'http://www.twitch.tv/shroomztv',
 271         'info_dict': {
 272             'id': '12772022048',
 273             'display_id': 'shroomztv',
 274             'ext': 'mp4',
 275             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 276             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 277             'is_live': True,
 278             'timestamp': 1421928037,
 279             'upload_date': '20150122',
 280             'uploader': 'ShroomzTV',
 281             'uploader_id': 'shroomztv',
 282             'view_count': int,
 283         },
 284         'params': {
 285             # m3u8 download
 286             'skip_download': True,
 287         },
 288     }
 289
 290     def _real_extract(self, url):
 291         channel_id = self._match_id(url)
 292
 293         stream = self._download_json(
 294             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 295             'Downloading stream JSON').get('stream')
 296
 297         # Fallback on profile extraction if stream is offline
 298         if not stream:
 299             return self.url_result(
 300                 'http://www.twitch.tv/%s/profile' % channel_id,
 301                 'TwitchProfile', channel_id)
 302
 303         access_token = self._download_json(
 304             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 305             'Downloading channel access token')
 306
 307         query = {
 308             'allow_source': 'true',
 309             'p': random.randint(1000000, 10000000),
 310             'player': 'twitchweb',
 311             'segment_preference': '4',
 312             'sig': access_token['sig'],
 313             'token': access_token['token'],
 314         }
 315
 316         formats = self._extract_m3u8_formats(
 317             '%s/api/channel/hls/%s.m3u8?%s'
 318             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query).encode('utf-8')),
 319             channel_id, 'mp4')
 320
 321         view_count = stream.get('viewers')
 322         timestamp = parse_iso8601(stream.get('created_at'))
 323
 324         channel = stream['channel']
 325         title = self._live_title(channel.get('display_name') or channel.get('name'))
 326         description = channel.get('status')
 327
 328         thumbnails = []
 329         for thumbnail_key, thumbnail_url in stream['preview'].items():
 330             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 331             if not m:
 332                 continue
 333             thumbnails.append({
 334                 'url': thumbnail_url,
 335                 'width': int(m.group('width')),
 336                 'height': int(m.group('height')),
 337             })
 338
 339         return {
 340             'id': compat_str(stream['_id']),
 341             'display_id': channel_id,
 342             'title': title,
 343             'description': description,
 344             'thumbnails': thumbnails,
 345             'uploader': channel.get('display_name'),
 346             'uploader_id': channel.get('name'),
 347             'timestamp': timestamp,
 348             'view_count': view_count,
 349             'formats': formats,
 350             'is_live': True,
 351         }