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/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 = self._hidden_inputs(login_page)
  63
  64         login_form.update({
  65             'login': username.encode('utf-8'),
  66             'password': password.encode('utf-8'),
  67         })
  68
  69         request = compat_urllib_request.Request(
  70             self._LOGIN_POST_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  71         request.add_header('Referer', self._LOGIN_URL)
  72         response = self._download_webpage(
  73             request, None, 'Logging in as %s' % username)
  74
  75         error_message = self._search_regex(
  76             r'<div[^>]+class="subwindow_notice"[^>]*>([^<]+)</div>',
  77             response, 'error message', default=None)
  78         if error_message:
  79             raise ExtractorError(
  80                 'Unable to login. Twitch said: %s' % error_message, expected=True)
  81
  82         if '>Reset your password<' in response:
  83             self.report_warning('Twitch asks you to reset your password, go to https://secure.twitch.tv/reset/submit')
  84
  85     def _prefer_source(self, formats):
  86         try:
  87             source = next(f for f in formats if f['format_id'] == 'Source')
  88             source['preference'] = 10
  89         except StopIteration:
  90             pass  # No Source stream present
  91         self._sort_formats(formats)
  92
  93
  94 class TwitchItemBaseIE(TwitchBaseIE):
  95     def _download_info(self, item, item_id):
  96         return self._extract_info(self._download_json(
  97             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  98             'Downloading %s info JSON' % self._ITEM_TYPE))
  99
 100     def _extract_media(self, item_id):
 101         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 102         response = self._download_json(
 103             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
 104             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 105         entries = []
 106         chunks = response['chunks']
 107         qualities = list(chunks.keys())
 108         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 109             formats = []
 110             for fmt_num, fragment_fmt in enumerate(fragment):
 111                 format_id = qualities[fmt_num]
 112                 fmt = {
 113                     'url': fragment_fmt['url'],
 114                     'format_id': format_id,
 115                     'quality': 1 if format_id == 'live' else 0,
 116                 }
 117                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 118                 if m:
 119                     fmt['height'] = int(m.group('height'))
 120                 formats.append(fmt)
 121             self._sort_formats(formats)
 122             entry = dict(info)
 123             entry['id'] = '%s_%d' % (entry['id'], num)
 124             entry['title'] = '%s part %d' % (entry['title'], num)
 125             entry['formats'] = formats
 126             entries.append(entry)
 127         return self.playlist_result(entries, info['id'], info['title'])
 128
 129     def _extract_info(self, info):
 130         return {
 131             'id': info['_id'],
 132             'title': info['title'],
 133             'description': info['description'],
 134             'duration': info['length'],
 135             'thumbnail': info['preview'],
 136             'uploader': info['channel']['display_name'],
 137             'uploader_id': info['channel']['name'],
 138             'timestamp': parse_iso8601(info['recorded_at']),
 139             'view_count': info['views'],
 140         }
 141
 142     def _real_extract(self, url):
 143         return self._extract_media(self._match_id(url))
 144
 145
 146 class TwitchVideoIE(TwitchItemBaseIE):
 147     IE_NAME = 'twitch:video'
 148     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 149     _ITEM_TYPE = 'video'
 150     _ITEM_SHORTCUT = 'a'
 151
 152     _TEST = {
 153         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 154         'info_dict': {
 155             'id': 'a577357806',
 156             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 157         },
 158         'playlist_mincount': 12,
 159     }
 160
 161
 162 class TwitchChapterIE(TwitchItemBaseIE):
 163     IE_NAME = 'twitch:chapter'
 164     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 165     _ITEM_TYPE = 'chapter'
 166     _ITEM_SHORTCUT = 'c'
 167
 168     _TESTS = [{
 169         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 170         'info_dict': {
 171             'id': 'c5285812',
 172             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 173         },
 174         'playlist_mincount': 3,
 175     }, {
 176         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 177         'only_matching': True,
 178     }]
 179
 180
 181 class TwitchVodIE(TwitchItemBaseIE):
 182     IE_NAME = 'twitch:vod'
 183     _VALID_URL = r'%s/[^/]+/v/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 184     _ITEM_TYPE = 'vod'
 185     _ITEM_SHORTCUT = 'v'
 186
 187     _TEST = {
 188         'url': 'http://www.twitch.tv/riotgames/v/6528877',
 189         'info_dict': {
 190             'id': 'v6528877',
 191             'ext': 'mp4',
 192             'title': 'LCK Summer Split - Week 6 Day 1',
 193             'thumbnail': 're:^https?://.*\.jpg$',
 194             'duration': 17208,
 195             'timestamp': 1435131709,
 196             'upload_date': '20150624',
 197             'uploader': 'Riot Games',
 198             'uploader_id': 'riotgames',
 199             'view_count': int,
 200         },
 201         'params': {
 202             # m3u8 download
 203             'skip_download': True,
 204         },
 205     }
 206
 207     def _real_extract(self, url):
 208         item_id = self._match_id(url)
 209         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 210         access_token = self._download_json(
 211             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 212             'Downloading %s access token' % self._ITEM_TYPE)
 213         formats = self._extract_m3u8_formats(
 214             '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true'
 215             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 216             item_id, 'mp4')
 217         self._prefer_source(formats)
 218         info['formats'] = formats
 219         return info
 220
 221
 222 class TwitchPlaylistBaseIE(TwitchBaseIE):
 223     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 224     _PAGE_LIMIT = 100
 225
 226     def _extract_playlist(self, channel_id):
 227         info = self._download_json(
 228             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 229             channel_id, 'Downloading channel info JSON')
 230         channel_name = info.get('display_name') or info.get('name')
 231         entries = []
 232         offset = 0
 233         limit = self._PAGE_LIMIT
 234         for counter in itertools.count(1):
 235             response = self._download_json(
 236                 self._PLAYLIST_URL % (channel_id, offset, limit),
 237                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 238             page_entries = self._extract_playlist_page(response)
 239             if not page_entries:
 240                 break
 241             entries.extend(page_entries)
 242             offset += limit
 243         return self.playlist_result(
 244             [self.url_result(entry) for entry in set(entries)],
 245             channel_id, channel_name)
 246
 247     def _extract_playlist_page(self, response):
 248         videos = response.get('videos')
 249         return [video['url'] for video in videos] if videos else []
 250
 251     def _real_extract(self, url):
 252         return self._extract_playlist(self._match_id(url))
 253
 254
 255 class TwitchProfileIE(TwitchPlaylistBaseIE):
 256     IE_NAME = 'twitch:profile'
 257     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 258     _PLAYLIST_TYPE = 'profile'
 259
 260     _TEST = {
 261         'url': 'http://www.twitch.tv/vanillatv/profile',
 262         'info_dict': {
 263             'id': 'vanillatv',
 264             'title': 'VanillaTV',
 265         },
 266         'playlist_mincount': 412,
 267     }
 268
 269
 270 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 271     IE_NAME = 'twitch:past_broadcasts'
 272     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 273     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 274     _PLAYLIST_TYPE = 'past broadcasts'
 275
 276     _TEST = {
 277         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 278         'info_dict': {
 279             'id': 'spamfish',
 280             'title': 'Spamfish',
 281         },
 282         'playlist_mincount': 54,
 283     }
 284
 285
 286 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 287     IE_NAME = 'twitch:bookmarks'
 288     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 289     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 290     _PLAYLIST_TYPE = 'bookmarks'
 291
 292     _TEST = {
 293         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 294         'info_dict': {
 295             'id': 'ognos',
 296             'title': 'Ognos',
 297         },
 298         'playlist_mincount': 3,
 299     }
 300
 301     def _extract_playlist_page(self, response):
 302         entries = []
 303         for bookmark in response.get('bookmarks', []):
 304             video = bookmark.get('video')
 305             if not video:
 306                 continue
 307             entries.append(video['url'])
 308         return entries
 309
 310
 311 class TwitchStreamIE(TwitchBaseIE):
 312     IE_NAME = 'twitch:stream'
 313     _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 314
 315     _TESTS = [{
 316         'url': 'http://www.twitch.tv/shroomztv',
 317         'info_dict': {
 318             'id': '12772022048',
 319             'display_id': 'shroomztv',
 320             'ext': 'mp4',
 321             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 322             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 323             'is_live': True,
 324             'timestamp': 1421928037,
 325             'upload_date': '20150122',
 326             'uploader': 'ShroomzTV',
 327             'uploader_id': 'shroomztv',
 328             'view_count': int,
 329         },
 330         'params': {
 331             # m3u8 download
 332             'skip_download': True,
 333         },
 334     }, {
 335         'url': 'http://www.twitch.tv/miracle_doto#profile-0',
 336         'only_matching': True,
 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         # Channel name may be typed if different case than the original channel name
 353         # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
 354         # an invalid m3u8 URL. Working around by use of original channel name from stream
 355         # JSON and fallback to lowercase if it's not available.
 356         channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
 357
 358         access_token = self._download_json(
 359             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 360             'Downloading channel access token')
 361
 362         query = {
 363             'allow_source': 'true',
 364             'p': random.randint(1000000, 10000000),
 365             'player': 'twitchweb',
 366             'segment_preference': '4',
 367             'sig': access_token['sig'].encode('utf-8'),
 368             'token': access_token['token'].encode('utf-8'),
 369         }
 370         formats = self._extract_m3u8_formats(
 371             '%s/api/channel/hls/%s.m3u8?%s'
 372             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 373             channel_id, 'mp4')
 374         self._prefer_source(formats)
 375
 376         view_count = stream.get('viewers')
 377         timestamp = parse_iso8601(stream.get('created_at'))
 378
 379         channel = stream['channel']
 380         title = self._live_title(channel.get('display_name') or channel.get('name'))
 381         description = channel.get('status')
 382
 383         thumbnails = []
 384         for thumbnail_key, thumbnail_url in stream['preview'].items():
 385             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 386             if not m:
 387                 continue
 388             thumbnails.append({
 389                 'url': thumbnail_url,
 390                 'width': int(m.group('width')),
 391                 'height': int(m.group('height')),
 392             })
 393
 394         return {
 395             'id': compat_str(stream['_id']),
 396             'display_id': channel_id,
 397             'title': title,
 398             'description': description,
 399             'thumbnails': thumbnails,
 400             'uploader': channel.get('display_name'),
 401             'uploader_id': channel.get('name'),
 402             'timestamp': timestamp,
 403             'view_count': view_count,
 404             'formats': formats,
 405             'is_live': True,
 406         }