_ 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     _NETRC_MACHINE = 'twitch'
  27
  28     def _handle_error(self, response):
  29         if not isinstance(response, dict):
  30             return
  31         error = response.get('error')
  32         if error:
  33             raise ExtractorError(
  34                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  35                 expected=True)
  36
  37     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  38         headers = {
  39             'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
  40             'X-Requested-With': 'XMLHttpRequest',
  41         }
  42         for cookie in self._downloader.cookiejar:
  43             if cookie.name == 'api_token':
  44                 headers['Twitch-Api-Token'] = cookie.value
  45         request = compat_urllib_request.Request(url, headers=headers)
  46         response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
  47         self._handle_error(response)
  48         return response
  49
  50     def _real_initialize(self):
  51         self._login()
  52
  53     def _login(self):
  54         (username, password) = self._get_login_info()
  55         if username is None:
  56             return
  57
  58         login_page = self._download_webpage(
  59             self._LOGIN_URL, None, 'Downloading login page')
  60
  61         authenticity_token = self._search_regex(
  62             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  63             login_page, 'authenticity token')
  64
  65         login_form = {
  66             'utf8': '✓'.encode('utf-8'),
  67             'authenticity_token': authenticity_token,
  68             'redirect_on_login': '',
  69             'embed_form': 'false',
  70             'mp_source_action': '',
  71             'follow': '',
  72             'user[login]': username,
  73             'user[password]': password,
  74         }
  75
  76         request = compat_urllib_request.Request(
  77             self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  78         request.add_header('Referer', self._LOGIN_URL)
  79         response = self._download_webpage(
  80             request, None, 'Logging in as %s' % username)
  81
  82         m = re.search(
  83             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  84         if m:
  85             raise ExtractorError(
  86                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  87
  88     def _prefer_source(self, formats):
  89         try:
  90             source = next(f for f in formats if f['format_id'] == 'Source')
  91             source['preference'] = 10
  92         except StopIteration:
  93             pass  # No Source stream present
  94         self._sort_formats(formats)
  95
  96
  97 class TwitchItemBaseIE(TwitchBaseIE):
  98     def _download_info(self, item, item_id):
  99         return self._extract_info(self._download_json(
 100             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
 101             'Downloading %s info JSON' % self._ITEM_TYPE))
 102
 103     def _extract_media(self, item_id):
 104         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 105         response = self._download_json(
 106             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
 107             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 108         entries = []
 109         chunks = response['chunks']
 110         qualities = list(chunks.keys())
 111         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 112             formats = []
 113             for fmt_num, fragment_fmt in enumerate(fragment):
 114                 format_id = qualities[fmt_num]
 115                 fmt = {
 116                     'url': fragment_fmt['url'],
 117                     'format_id': format_id,
 118                     'quality': 1 if format_id == 'live' else 0,
 119                 }
 120                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 121                 if m:
 122                     fmt['height'] = int(m.group('height'))
 123                 formats.append(fmt)
 124             self._sort_formats(formats)
 125             entry = dict(info)
 126             entry['id'] = '%s_%d' % (entry['id'], num)
 127             entry['title'] = '%s part %d' % (entry['title'], num)
 128             entry['formats'] = formats
 129             entries.append(entry)
 130         return self.playlist_result(entries, info['id'], info['title'])
 131
 132     def _extract_info(self, info):
 133         return {
 134             'id': info['_id'],
 135             'title': info['title'],
 136             'description': info['description'],
 137             'duration': info['length'],
 138             'thumbnail': info['preview'],
 139             'uploader': info['channel']['display_name'],
 140             'uploader_id': info['channel']['name'],
 141             'timestamp': parse_iso8601(info['recorded_at']),
 142             'view_count': info['views'],
 143         }
 144
 145     def _real_extract(self, url):
 146         return self._extract_media(self._match_id(url))
 147
 148
 149 class TwitchVideoIE(TwitchItemBaseIE):
 150     IE_NAME = 'twitch:video'
 151     _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 152     _ITEM_TYPE = 'video'
 153     _ITEM_SHORTCUT = 'a'
 154
 155     _TEST = {
 156         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 157         'info_dict': {
 158             'id': 'a577357806',
 159             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 160         },
 161         'playlist_mincount': 12,
 162     }
 163
 164
 165 class TwitchChapterIE(TwitchItemBaseIE):
 166     IE_NAME = 'twitch:chapter'
 167     _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 168     _ITEM_TYPE = 'chapter'
 169     _ITEM_SHORTCUT = 'c'
 170
 171     _TESTS = [{
 172         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 173         'info_dict': {
 174             'id': 'c5285812',
 175             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 176         },
 177         'playlist_mincount': 3,
 178     }, {
 179         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 180         'only_matching': True,
 181     }]
 182
 183
 184 class TwitchVodIE(TwitchItemBaseIE):
 185     IE_NAME = 'twitch:vod'
 186     _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 187     _ITEM_TYPE = 'vod'
 188     _ITEM_SHORTCUT = 'v'
 189
 190     _TEST = {
 191         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 192         'info_dict': {
 193             'id': 'v3622000',
 194             'ext': 'mp4',
 195             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 196             'thumbnail': 're:^https?://.*\.jpg$',
 197             'duration': 6951,
 198             'timestamp': 1419028564,
 199             'upload_date': '20141219',
 200             'uploader': 'KSPTV',
 201             'uploader_id': 'ksptv',
 202             'view_count': int,
 203         },
 204         'params': {
 205             # m3u8 download
 206             'skip_download': True,
 207         },
 208     }
 209
 210     def _real_extract(self, url):
 211         item_id = self._match_id(url)
 212         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 213         access_token = self._download_json(
 214             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 215             'Downloading %s access token' % self._ITEM_TYPE)
 216         formats = self._extract_m3u8_formats(
 217             '%s/vod/%s?nauth=%s&nauthsig=%s'
 218             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 219             item_id, 'mp4')
 220         self._prefer_source(formats)
 221         info['formats'] = formats
 222         return info
 223
 224
 225 class TwitchPlaylistBaseIE(TwitchBaseIE):
 226     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 227     _PAGE_LIMIT = 100
 228
 229     def _extract_playlist(self, channel_id):
 230         info = self._download_json(
 231             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 232             channel_id, 'Downloading channel info JSON')
 233         channel_name = info.get('display_name') or info.get('name')
 234         entries = []
 235         offset = 0
 236         limit = self._PAGE_LIMIT
 237         for counter in itertools.count(1):
 238             response = self._download_json(
 239                 self._PLAYLIST_URL % (channel_id, offset, limit),
 240                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 241             page_entries = self._extract_playlist_page(response)
 242             if not page_entries:
 243                 break
 244             entries.extend(page_entries)
 245             offset += limit
 246         return self.playlist_result(
 247             [self.url_result(entry) for entry in set(entries)],
 248             channel_id, channel_name)
 249
 250     def _extract_playlist_page(self, response):
 251         videos = response.get('videos')
 252         return [video['url'] for video in videos] if videos else []
 253
 254     def _real_extract(self, url):
 255         return self._extract_playlist(self._match_id(url))
 256
 257
 258 class TwitchProfileIE(TwitchPlaylistBaseIE):
 259     IE_NAME = 'twitch:profile'
 260     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 261     _PLAYLIST_TYPE = 'profile'
 262
 263     _TEST = {
 264         'url': 'http://www.twitch.tv/vanillatv/profile',
 265         'info_dict': {
 266             'id': 'vanillatv',
 267             'title': 'VanillaTV',
 268         },
 269         'playlist_mincount': 412,
 270     }
 271
 272
 273 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 274     IE_NAME = 'twitch:past_broadcasts'
 275     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 276     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 277     _PLAYLIST_TYPE = 'past broadcasts'
 278
 279     _TEST = {
 280         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 281         'info_dict': {
 282             'id': 'spamfish',
 283             'title': 'Spamfish',
 284         },
 285         'playlist_mincount': 54,
 286     }
 287
 288
 289 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 290     IE_NAME = 'twitch:bookmarks'
 291     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 292     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 293     _PLAYLIST_TYPE = 'bookmarks'
 294
 295     _TEST = {
 296         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 297         'info_dict': {
 298             'id': 'ognos',
 299             'title': 'Ognos',
 300         },
 301         'playlist_mincount': 3,
 302     }
 303
 304     def _extract_playlist_page(self, response):
 305         entries = []
 306         for bookmark in response.get('bookmarks', []):
 307             video = bookmark.get('video')
 308             if not video:
 309                 continue
 310             entries.append(video['url'])
 311         return entries
 312
 313
 314 class TwitchStreamIE(TwitchBaseIE):
 315     IE_NAME = 'twitch:stream'
 316     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 317
 318     _TEST = {
 319         'url': 'http://www.twitch.tv/shroomztv',
 320         'info_dict': {
 321             'id': '12772022048',
 322             'display_id': 'shroomztv',
 323             'ext': 'mp4',
 324             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 325             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 326             'is_live': True,
 327             'timestamp': 1421928037,
 328             'upload_date': '20150122',
 329             'uploader': 'ShroomzTV',
 330             'uploader_id': 'shroomztv',
 331             'view_count': int,
 332         },
 333         'params': {
 334             # m3u8 download
 335             'skip_download': True,
 336         },
 337     }
 338
 339     def _real_extract(self, url):
 340         channel_id = self._match_id(url)
 341
 342         stream = self._download_json(
 343             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 344             'Downloading stream JSON').get('stream')
 345
 346         # Fallback on profile extraction if stream is offline
 347         if not stream:
 348             return self.url_result(
 349                 'http://www.twitch.tv/%s/profile' % channel_id,
 350                 'TwitchProfile', channel_id)
 351
 352         access_token = self._download_json(
 353             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 354             'Downloading channel access token')
 355
 356         query = {
 357             'allow_source': 'true',
 358             'p': random.randint(1000000, 10000000),
 359             'player': 'twitchweb',
 360             'segment_preference': '4',
 361             'sig': access_token['sig'].encode('utf-8'),
 362             'token': access_token['token'].encode('utf-8'),
 363         }
 364         formats = self._extract_m3u8_formats(
 365             '%s/api/channel/hls/%s.m3u8?%s'
 366             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 367             channel_id, 'mp4')
 368         self._prefer_source(formats)
 369
 370         view_count = stream.get('viewers')
 371         timestamp = parse_iso8601(stream.get('created_at'))
 372
 373         channel = stream['channel']
 374         title = self._live_title(channel.get('display_name') or channel.get('name'))
 375         description = channel.get('status')
 376
 377         thumbnails = []
 378         for thumbnail_key, thumbnail_url in stream['preview'].items():
 379             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 380             if not m:
 381                 continue
 382             thumbnails.append({
 383                 'url': thumbnail_url,
 384                 'width': int(m.group('width')),
 385                 'height': int(m.group('height')),
 386             })
 387
 388         return {
 389             'id': compat_str(stream['_id']),
 390             'display_id': channel_id,
 391             'title': title,
 392             'description': description,
 393             'thumbnails': thumbnails,
 394             'uploader': channel.get('display_name'),
 395             'uploader_id': channel.get('name'),
 396             'timestamp': timestamp,
 397             'view_count': view_count,
 398             'formats': formats,
 399             'is_live': True,
 400         }