_ 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/login'
  26     _LOGIN_POST_URL = 'https://passport.twitch.tv/authorize'
  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         login_form = dict(re.findall(
  63             r'<input\s+type="hidden"\s+name="([^"]+)"\s+(?:id="[^"]+"\s+)?value="([^"]*)"',
  64             login_page))
  65
  66         login_form.update({
  67             'login': username,
  68             'password': password,
  69         })
  70
  71         request = compat_urllib_request.Request(
  72             self._LOGIN_POST_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  73         request.add_header('Referer', self._LOGIN_URL)
  74         response = self._download_webpage(
  75             request, None, 'Logging in as %s' % username)
  76
  77         error_message = self._search_regex(
  78             r'<div[^>]+class="subwindow_notice"[^>]*>([^<]+)</div>',
  79             response, 'error message', default=None)
  80         if error_message:
  81             raise ExtractorError(
  82                 'Unable to login. Twitch said: %s' % error_message, expected=True)
  83
  84     def _prefer_source(self, formats):
  85         try:
  86             source = next(f for f in formats if f['format_id'] == 'Source')
  87             source['preference'] = 10
  88         except StopIteration:
  89             pass  # No Source stream present
  90         self._sort_formats(formats)
  91
  92
  93 class TwitchItemBaseIE(TwitchBaseIE):
  94     def _download_info(self, item, item_id):
  95         return self._extract_info(self._download_json(
  96             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  97             'Downloading %s info JSON' % self._ITEM_TYPE))
  98
  99     def _extract_media(self, item_id):
 100         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 101         response = self._download_json(
 102             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
 103             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 104         entries = []
 105         chunks = response['chunks']
 106         qualities = list(chunks.keys())
 107         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 108             formats = []
 109             for fmt_num, fragment_fmt in enumerate(fragment):
 110                 format_id = qualities[fmt_num]
 111                 fmt = {
 112                     'url': fragment_fmt['url'],
 113                     'format_id': format_id,
 114                     'quality': 1 if format_id == 'live' else 0,
 115                 }
 116                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 117                 if m:
 118                     fmt['height'] = int(m.group('height'))
 119                 formats.append(fmt)
 120             self._sort_formats(formats)
 121             entry = dict(info)
 122             entry['id'] = '%s_%d' % (entry['id'], num)
 123             entry['title'] = '%s part %d' % (entry['title'], num)
 124             entry['formats'] = formats
 125             entries.append(entry)
 126         return self.playlist_result(entries, info['id'], info['title'])
 127
 128     def _extract_info(self, info):
 129         return {
 130             'id': info['_id'],
 131             'title': info['title'],
 132             'description': info['description'],
 133             'duration': info['length'],
 134             'thumbnail': info['preview'],
 135             'uploader': info['channel']['display_name'],
 136             'uploader_id': info['channel']['name'],
 137             'timestamp': parse_iso8601(info['recorded_at']),
 138             'view_count': info['views'],
 139         }
 140
 141     def _real_extract(self, url):
 142         return self._extract_media(self._match_id(url))
 143
 144
 145 class TwitchVideoIE(TwitchItemBaseIE):
 146     IE_NAME = 'twitch:video'
 147     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 148     _ITEM_TYPE = 'video'
 149     _ITEM_SHORTCUT = 'a'
 150
 151     _TEST = {
 152         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 153         'info_dict': {
 154             'id': 'a577357806',
 155             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 156         },
 157         'playlist_mincount': 12,
 158     }
 159
 160
 161 class TwitchChapterIE(TwitchItemBaseIE):
 162     IE_NAME = 'twitch:chapter'
 163     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 164     _ITEM_TYPE = 'chapter'
 165     _ITEM_SHORTCUT = 'c'
 166
 167     _TESTS = [{
 168         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 169         'info_dict': {
 170             'id': 'c5285812',
 171             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 172         },
 173         'playlist_mincount': 3,
 174     }, {
 175         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 176         'only_matching': True,
 177     }]
 178
 179
 180 class TwitchVodIE(TwitchItemBaseIE):
 181     IE_NAME = 'twitch:vod'
 182     _VALID_URL = r'%s/[^/]+/v/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 183     _ITEM_TYPE = 'vod'
 184     _ITEM_SHORTCUT = 'v'
 185
 186     _TEST = {
 187         'url': 'http://www.twitch.tv/riotgames/v/6528877',
 188         'info_dict': {
 189             'id': 'v6528877',
 190             'ext': 'mp4',
 191             'title': 'LCK Summer Split - Week 6 Day 1',
 192             'thumbnail': 're:^https?://.*\.jpg$',
 193             'duration': 17208,
 194             'timestamp': 1435131709,
 195             'upload_date': '20150624',
 196             'uploader': 'Riot Games',
 197             'uploader_id': 'riotgames',
 198             'view_count': int,
 199         },
 200         'params': {
 201             # m3u8 download
 202             'skip_download': True,
 203         },
 204     }
 205
 206     def _real_extract(self, url):
 207         item_id = self._match_id(url)
 208         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 209         access_token = self._download_json(
 210             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 211             'Downloading %s access token' % self._ITEM_TYPE)
 212         formats = self._extract_m3u8_formats(
 213             '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true'
 214             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 215             item_id, 'mp4')
 216         self._prefer_source(formats)
 217         info['formats'] = formats
 218         return info
 219
 220
 221 class TwitchPlaylistBaseIE(TwitchBaseIE):
 222     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 223     _PAGE_LIMIT = 100
 224
 225     def _extract_playlist(self, channel_id):
 226         info = self._download_json(
 227             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 228             channel_id, 'Downloading channel info JSON')
 229         channel_name = info.get('display_name') or info.get('name')
 230         entries = []
 231         offset = 0
 232         limit = self._PAGE_LIMIT
 233         for counter in itertools.count(1):
 234             response = self._download_json(
 235                 self._PLAYLIST_URL % (channel_id, offset, limit),
 236                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 237             page_entries = self._extract_playlist_page(response)
 238             if not page_entries:
 239                 break
 240             entries.extend(page_entries)
 241             offset += limit
 242         return self.playlist_result(
 243             [self.url_result(entry) for entry in set(entries)],
 244             channel_id, channel_name)
 245
 246     def _extract_playlist_page(self, response):
 247         videos = response.get('videos')
 248         return [video['url'] for video in videos] if videos else []
 249
 250     def _real_extract(self, url):
 251         return self._extract_playlist(self._match_id(url))
 252
 253
 254 class TwitchProfileIE(TwitchPlaylistBaseIE):
 255     IE_NAME = 'twitch:profile'
 256     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 257     _PLAYLIST_TYPE = 'profile'
 258
 259     _TEST = {
 260         'url': 'http://www.twitch.tv/vanillatv/profile',
 261         'info_dict': {
 262             'id': 'vanillatv',
 263             'title': 'VanillaTV',
 264         },
 265         'playlist_mincount': 412,
 266     }
 267
 268
 269 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 270     IE_NAME = 'twitch:past_broadcasts'
 271     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 272     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 273     _PLAYLIST_TYPE = 'past broadcasts'
 274
 275     _TEST = {
 276         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 277         'info_dict': {
 278             'id': 'spamfish',
 279             'title': 'Spamfish',
 280         },
 281         'playlist_mincount': 54,
 282     }
 283
 284
 285 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 286     IE_NAME = 'twitch:bookmarks'
 287     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 288     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 289     _PLAYLIST_TYPE = 'bookmarks'
 290
 291     _TEST = {
 292         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 293         'info_dict': {
 294             'id': 'ognos',
 295             'title': 'Ognos',
 296         },
 297         'playlist_mincount': 3,
 298     }
 299
 300     def _extract_playlist_page(self, response):
 301         entries = []
 302         for bookmark in response.get('bookmarks', []):
 303             video = bookmark.get('video')
 304             if not video:
 305                 continue
 306             entries.append(video['url'])
 307         return entries
 308
 309
 310 class TwitchStreamIE(TwitchBaseIE):
 311     IE_NAME = 'twitch:stream'
 312     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 313
 314     _TEST = {
 315         'url': 'http://www.twitch.tv/shroomztv',
 316         'info_dict': {
 317             'id': '12772022048',
 318             'display_id': 'shroomztv',
 319             'ext': 'mp4',
 320             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 321             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 322             'is_live': True,
 323             'timestamp': 1421928037,
 324             'upload_date': '20150122',
 325             'uploader': 'ShroomzTV',
 326             'uploader_id': 'shroomztv',
 327             'view_count': int,
 328         },
 329         'params': {
 330             # m3u8 download
 331             'skip_download': True,
 332         },
 333     }
 334
 335     def _real_extract(self, url):
 336         channel_id = self._match_id(url)
 337
 338         stream = self._download_json(
 339             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 340             'Downloading stream JSON').get('stream')
 341
 342         # Fallback on profile extraction if stream is offline
 343         if not stream:
 344             return self.url_result(
 345                 'http://www.twitch.tv/%s/profile' % channel_id,
 346                 'TwitchProfile', channel_id)
 347
 348         access_token = self._download_json(
 349             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 350             'Downloading channel access token')
 351
 352         query = {
 353             'allow_source': 'true',
 354             'p': random.randint(1000000, 10000000),
 355             'player': 'twitchweb',
 356             'segment_preference': '4',
 357             'sig': access_token['sig'].encode('utf-8'),
 358             'token': access_token['token'].encode('utf-8'),
 359         }
 360         formats = self._extract_m3u8_formats(
 361             '%s/api/channel/hls/%s.m3u8?%s'
 362             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 363             channel_id, 'mp4')
 364         self._prefer_source(formats)
 365
 366         view_count = stream.get('viewers')
 367         timestamp = parse_iso8601(stream.get('created_at'))
 368
 369         channel = stream['channel']
 370         title = self._live_title(channel.get('display_name') or channel.get('name'))
 371         description = channel.get('status')
 372
 373         thumbnails = []
 374         for thumbnail_key, thumbnail_url in stream['preview'].items():
 375             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 376             if not m:
 377                 continue
 378             thumbnails.append({
 379                 'url': thumbnail_url,
 380                 'width': int(m.group('width')),
 381                 'height': int(m.group('height')),
 382             })
 383
 384         return {
 385             'id': compat_str(stream['_id']),
 386             'display_id': channel_id,
 387             'title': title,
 388             'description': description,
 389             'thumbnails': thumbnails,
 390             'uploader': channel.get('display_name'),
 391             'uploader_id': channel.get('name'),
 392             'timestamp': timestamp,
 393             'view_count': view_count,
 394             'formats': formats,
 395             'is_live': True,
 396         }