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
  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         headers = {
  38             'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
  39             'X-Requested-With': 'XMLHttpRequest',
  40         }
  41         for cookie in self._downloader.cookiejar:
  42             if cookie.name == 'api_token':
  43                 headers['Twitch-Api-Token'] = cookie.value
  44         request = compat_urllib_request.Request(url, headers=headers)
  45         response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
  46         self._handle_error(response)
  47         return response
  48
  49     def _real_initialize(self):
  50         self._login()
  51
  52     def _login(self):
  53         (username, password) = self._get_login_info()
  54         if username is None:
  55             return
  56
  57         login_page = self._download_webpage(
  58             self._LOGIN_URL, None, 'Downloading login page')
  59
  60         authenticity_token = self._search_regex(
  61             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  62             login_page, 'authenticity token')
  63
  64         login_form = {
  65             'utf8': '✓'.encode('utf-8'),
  66             'authenticity_token': authenticity_token,
  67             'redirect_on_login': '',
  68             'embed_form': 'false',
  69             'mp_source_action': '',
  70             'follow': '',
  71             'user[login]': username,
  72             'user[password]': password,
  73         }
  74
  75         request = compat_urllib_request.Request(
  76             self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  77         request.add_header('Referer', self._LOGIN_URL)
  78         response = self._download_webpage(
  79             request, None, 'Logging in as %s' % username)
  80
  81         m = re.search(
  82             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  83         if m:
  84             raise ExtractorError(
  85                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  86
  87
  88 class TwitchItemBaseIE(TwitchBaseIE):
  89     def _download_info(self, item, item_id):
  90         return self._extract_info(self._download_json(
  91             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  92             'Downloading %s info JSON' % self._ITEM_TYPE))
  93
  94     def _extract_media(self, item_id):
  95         info = self._download_info(self._ITEM_SHORTCUT, item_id)
  96         response = self._download_json(
  97             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
  98             'Downloading %s playlist JSON' % self._ITEM_TYPE)
  99         entries = []
 100         chunks = response['chunks']
 101         qualities = list(chunks.keys())
 102         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 103             formats = []
 104             for fmt_num, fragment_fmt in enumerate(fragment):
 105                 format_id = qualities[fmt_num]
 106                 fmt = {
 107                     'url': fragment_fmt['url'],
 108                     'format_id': format_id,
 109                     'quality': 1 if format_id == 'live' else 0,
 110                 }
 111                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 112                 if m:
 113                     fmt['height'] = int(m.group('height'))
 114                 formats.append(fmt)
 115             self._sort_formats(formats)
 116             entry = dict(info)
 117             entry['id'] = '%s_%d' % (entry['id'], num)
 118             entry['title'] = '%s part %d' % (entry['title'], num)
 119             entry['formats'] = formats
 120             entries.append(entry)
 121         return self.playlist_result(entries, info['id'], info['title'])
 122
 123     def _extract_info(self, info):
 124         return {
 125             'id': info['_id'],
 126             'title': info['title'],
 127             'description': info['description'],
 128             'duration': info['length'],
 129             'thumbnail': info['preview'],
 130             'uploader': info['channel']['display_name'],
 131             'uploader_id': info['channel']['name'],
 132             'timestamp': parse_iso8601(info['recorded_at']),
 133             'view_count': info['views'],
 134         }
 135
 136     def _real_extract(self, url):
 137         return self._extract_media(self._match_id(url))
 138
 139
 140 class TwitchVideoIE(TwitchItemBaseIE):
 141     IE_NAME = 'twitch:video'
 142     _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 143     _ITEM_TYPE = 'video'
 144     _ITEM_SHORTCUT = 'a'
 145
 146     _TEST = {
 147         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 148         'info_dict': {
 149             'id': 'a577357806',
 150             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 151         },
 152         'playlist_mincount': 12,
 153     }
 154
 155
 156 class TwitchChapterIE(TwitchItemBaseIE):
 157     IE_NAME = 'twitch:chapter'
 158     _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 159     _ITEM_TYPE = 'chapter'
 160     _ITEM_SHORTCUT = 'c'
 161
 162     _TESTS = [{
 163         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 164         'info_dict': {
 165             'id': 'c5285812',
 166             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 167         },
 168         'playlist_mincount': 3,
 169     }, {
 170         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 171         'only_matching': True,
 172     }]
 173
 174
 175 class TwitchVodIE(TwitchItemBaseIE):
 176     IE_NAME = 'twitch:vod'
 177     _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 178     _ITEM_TYPE = 'vod'
 179     _ITEM_SHORTCUT = 'v'
 180
 181     _TEST = {
 182         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 183         'info_dict': {
 184             'id': 'v3622000',
 185             'ext': 'mp4',
 186             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 187             'thumbnail': 're:^https?://.*\.jpg$',
 188             'duration': 6951,
 189             'timestamp': 1419028564,
 190             'upload_date': '20141219',
 191             'uploader': 'KSPTV',
 192             'uploader_id': 'ksptv',
 193             'view_count': int,
 194         },
 195         'params': {
 196             # m3u8 download
 197             'skip_download': True,
 198         },
 199     }
 200
 201     def _real_extract(self, url):
 202         item_id = self._match_id(url)
 203         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 204         access_token = self._download_json(
 205             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 206             'Downloading %s access token' % self._ITEM_TYPE)
 207         formats = self._extract_m3u8_formats(
 208             '%s/vod/%s?nauth=%s&nauthsig=%s'
 209             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 210             item_id, 'mp4')
 211         info['formats'] = formats
 212         return info
 213
 214
 215 class TwitchPlaylistBaseIE(TwitchBaseIE):
 216     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 217     _PAGE_LIMIT = 100
 218
 219     def _extract_playlist(self, channel_id):
 220         info = self._download_json(
 221             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 222             channel_id, 'Downloading channel info JSON')
 223         channel_name = info.get('display_name') or info.get('name')
 224         entries = []
 225         offset = 0
 226         limit = self._PAGE_LIMIT
 227         for counter in itertools.count(1):
 228             response = self._download_json(
 229                 self._PLAYLIST_URL % (channel_id, offset, limit),
 230                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 231             page_entries = self._extract_playlist_page(response)
 232             if not page_entries:
 233                 break
 234             entries.extend(page_entries)
 235             offset += limit
 236         return self.playlist_result(
 237             [self.url_result(entry) for entry in set(entries)],
 238             channel_id, channel_name)
 239
 240     def _extract_playlist_page(self, response):
 241         videos = response.get('videos')
 242         return [video['url'] for video in videos] if videos else []
 243
 244     def _real_extract(self, url):
 245         return self._extract_playlist(self._match_id(url))
 246
 247
 248 class TwitchProfileIE(TwitchPlaylistBaseIE):
 249     IE_NAME = 'twitch:profile'
 250     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 251     _PLAYLIST_TYPE = 'profile'
 252
 253     _TEST = {
 254         'url': 'http://www.twitch.tv/vanillatv/profile',
 255         'info_dict': {
 256             'id': 'vanillatv',
 257             'title': 'VanillaTV',
 258         },
 259         'playlist_mincount': 412,
 260     }
 261
 262
 263 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 264     IE_NAME = 'twitch:past_broadcasts'
 265     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 266     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 267     _PLAYLIST_TYPE = 'past broadcasts'
 268
 269     _TEST = {
 270         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 271         'info_dict': {
 272             'id': 'spamfish',
 273             'title': 'Spamfish',
 274         },
 275         'playlist_mincount': 54,
 276     }
 277
 278
 279 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 280     IE_NAME = 'twitch:bookmarks'
 281     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 282     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 283     _PLAYLIST_TYPE = 'bookmarks'
 284
 285     _TEST = {
 286         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 287         'info_dict': {
 288             'id': 'ognos',
 289             'title': 'Ognos',
 290         },
 291         'playlist_mincount': 3,
 292     }
 293
 294     def _extract_playlist_page(self, response):
 295         entries = []
 296         for bookmark in response.get('bookmarks', []):
 297             video = bookmark.get('video')
 298             if not video:
 299                 continue
 300             entries.append(video['url'])
 301         return entries
 302
 303
 304 class TwitchStreamIE(TwitchBaseIE):
 305     IE_NAME = 'twitch:stream'
 306     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 307
 308     _TEST = {
 309         'url': 'http://www.twitch.tv/shroomztv',
 310         'info_dict': {
 311             'id': '12772022048',
 312             'display_id': 'shroomztv',
 313             'ext': 'mp4',
 314             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 315             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 316             'is_live': True,
 317             'timestamp': 1421928037,
 318             'upload_date': '20150122',
 319             'uploader': 'ShroomzTV',
 320             'uploader_id': 'shroomztv',
 321             'view_count': int,
 322         },
 323         'params': {
 324             # m3u8 download
 325             'skip_download': True,
 326         },
 327     }
 328
 329     def _real_extract(self, url):
 330         channel_id = self._match_id(url)
 331
 332         stream = self._download_json(
 333             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 334             'Downloading stream JSON').get('stream')
 335
 336         # Fallback on profile extraction if stream is offline
 337         if not stream:
 338             return self.url_result(
 339                 'http://www.twitch.tv/%s/profile' % channel_id,
 340                 'TwitchProfile', channel_id)
 341
 342         access_token = self._download_json(
 343             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 344             'Downloading channel access token')
 345
 346         query = {
 347             'allow_source': 'true',
 348             'p': random.randint(1000000, 10000000),
 349             'player': 'twitchweb',
 350             'segment_preference': '4',
 351             'sig': access_token['sig'],
 352             'token': access_token['token'],
 353         }
 354
 355         formats = self._extract_m3u8_formats(
 356             '%s/api/channel/hls/%s.m3u8?%s'
 357             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query).encode('utf-8')),
 358             channel_id, 'mp4')
 359
 360         # prefer the 'source' stream, the others are limited to 30 fps
 361         def _sort_source(f):
 362             if f.get('m3u8_media') is not None and f['m3u8_media'].get('NAME') == 'Source':
 363                 return 1
 364             return 0
 365         formats = sorted(formats, key=_sort_source)
 366
 367         view_count = stream.get('viewers')
 368         timestamp = parse_iso8601(stream.get('created_at'))
 369
 370         channel = stream['channel']
 371         title = self._live_title(channel.get('display_name') or channel.get('name'))
 372         description = channel.get('status')
 373
 374         thumbnails = []
 375         for thumbnail_key, thumbnail_url in stream['preview'].items():
 376             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 377             if not m:
 378                 continue
 379             thumbnails.append({
 380                 'url': thumbnail_url,
 381                 'width': int(m.group('width')),
 382                 'height': int(m.group('height')),
 383             })
 384
 385         return {
 386             'id': compat_str(stream['_id']),
 387             'display_id': channel_id,
 388             'title': title,
 389             'description': description,
 390             'thumbnails': thumbnails,
 391             'uploader': channel.get('display_name'),
 392             'uploader_id': channel.get('name'),
 393             'timestamp': timestamp,
 394             'view_count': view_count,
 395             'formats': formats,
 396             'is_live': True,
 397         }