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         if '>Reset your password<' in response:
  85             self.report_warning('Twitch asks you to reset your password, go to https://secure.twitch.tv/reset/submit')
  86
  87     def _prefer_source(self, formats):
  88         try:
  89             source = next(f for f in formats if f['format_id'] == 'Source')
  90             source['preference'] = 10
  91         except StopIteration:
  92             pass  # No Source stream present
  93         self._sort_formats(formats)
  94
  95
  96 class TwitchItemBaseIE(TwitchBaseIE):
  97     def _download_info(self, item, item_id):
  98         return self._extract_info(self._download_json(
  99             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
 100             'Downloading %s info JSON' % self._ITEM_TYPE))
 101
 102     def _extract_media(self, item_id):
 103         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 104         response = self._download_json(
 105             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
 106             'Downloading %s playlist JSON' % self._ITEM_TYPE)
 107         entries = []
 108         chunks = response['chunks']
 109         qualities = list(chunks.keys())
 110         for num, fragment in enumerate(zip(*chunks.values()), start=1):
 111             formats = []
 112             for fmt_num, fragment_fmt in enumerate(fragment):
 113                 format_id = qualities[fmt_num]
 114                 fmt = {
 115                     'url': fragment_fmt['url'],
 116                     'format_id': format_id,
 117                     'quality': 1 if format_id == 'live' else 0,
 118                 }
 119                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 120                 if m:
 121                     fmt['height'] = int(m.group('height'))
 122                 formats.append(fmt)
 123             self._sort_formats(formats)
 124             entry = dict(info)
 125             entry['id'] = '%s_%d' % (entry['id'], num)
 126             entry['title'] = '%s part %d' % (entry['title'], num)
 127             entry['formats'] = formats
 128             entries.append(entry)
 129         return self.playlist_result(entries, info['id'], info['title'])
 130
 131     def _extract_info(self, info):
 132         return {
 133             'id': info['_id'],
 134             'title': info['title'],
 135             'description': info['description'],
 136             'duration': info['length'],
 137             'thumbnail': info['preview'],
 138             'uploader': info['channel']['display_name'],
 139             'uploader_id': info['channel']['name'],
 140             'timestamp': parse_iso8601(info['recorded_at']),
 141             'view_count': info['views'],
 142         }
 143
 144     def _real_extract(self, url):
 145         return self._extract_media(self._match_id(url))
 146
 147
 148 class TwitchVideoIE(TwitchItemBaseIE):
 149     IE_NAME = 'twitch:video'
 150     _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 151     _ITEM_TYPE = 'video'
 152     _ITEM_SHORTCUT = 'a'
 153
 154     _TEST = {
 155         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 156         'info_dict': {
 157             'id': 'a577357806',
 158             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 159         },
 160         'playlist_mincount': 12,
 161     }
 162
 163
 164 class TwitchChapterIE(TwitchItemBaseIE):
 165     IE_NAME = 'twitch:chapter'
 166     _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 167     _ITEM_TYPE = 'chapter'
 168     _ITEM_SHORTCUT = 'c'
 169
 170     _TESTS = [{
 171         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 172         'info_dict': {
 173             'id': 'c5285812',
 174             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 175         },
 176         'playlist_mincount': 3,
 177     }, {
 178         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 179         'only_matching': True,
 180     }]
 181
 182
 183 class TwitchVodIE(TwitchItemBaseIE):
 184     IE_NAME = 'twitch:vod'
 185     _VALID_URL = r'%s/[^/]+/v/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
 186     _ITEM_TYPE = 'vod'
 187     _ITEM_SHORTCUT = 'v'
 188
 189     _TEST = {
 190         'url': 'http://www.twitch.tv/riotgames/v/6528877',
 191         'info_dict': {
 192             'id': 'v6528877',
 193             'ext': 'mp4',
 194             'title': 'LCK Summer Split - Week 6 Day 1',
 195             'thumbnail': 're:^https?://.*\.jpg$',
 196             'duration': 17208,
 197             'timestamp': 1435131709,
 198             'upload_date': '20150624',
 199             'uploader': 'Riot Games',
 200             'uploader_id': 'riotgames',
 201             'view_count': int,
 202         },
 203         'params': {
 204             # m3u8 download
 205             'skip_download': True,
 206         },
 207     }
 208
 209     def _real_extract(self, url):
 210         item_id = self._match_id(url)
 211         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 212         access_token = self._download_json(
 213             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 214             'Downloading %s access token' % self._ITEM_TYPE)
 215         formats = self._extract_m3u8_formats(
 216             '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true'
 217             % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
 218             item_id, 'mp4')
 219         self._prefer_source(formats)
 220         info['formats'] = formats
 221         return info
 222
 223
 224 class TwitchPlaylistBaseIE(TwitchBaseIE):
 225     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 226     _PAGE_LIMIT = 100
 227
 228     def _extract_playlist(self, channel_id):
 229         info = self._download_json(
 230             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 231             channel_id, 'Downloading channel info JSON')
 232         channel_name = info.get('display_name') or info.get('name')
 233         entries = []
 234         offset = 0
 235         limit = self._PAGE_LIMIT
 236         for counter in itertools.count(1):
 237             response = self._download_json(
 238                 self._PLAYLIST_URL % (channel_id, offset, limit),
 239                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 240             page_entries = self._extract_playlist_page(response)
 241             if not page_entries:
 242                 break
 243             entries.extend(page_entries)
 244             offset += limit
 245         return self.playlist_result(
 246             [self.url_result(entry) for entry in set(entries)],
 247             channel_id, channel_name)
 248
 249     def _extract_playlist_page(self, response):
 250         videos = response.get('videos')
 251         return [video['url'] for video in videos] if videos else []
 252
 253     def _real_extract(self, url):
 254         return self._extract_playlist(self._match_id(url))
 255
 256
 257 class TwitchProfileIE(TwitchPlaylistBaseIE):
 258     IE_NAME = 'twitch:profile'
 259     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 260     _PLAYLIST_TYPE = 'profile'
 261
 262     _TEST = {
 263         'url': 'http://www.twitch.tv/vanillatv/profile',
 264         'info_dict': {
 265             'id': 'vanillatv',
 266             'title': 'VanillaTV',
 267         },
 268         'playlist_mincount': 412,
 269     }
 270
 271
 272 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 273     IE_NAME = 'twitch:past_broadcasts'
 274     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 275     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 276     _PLAYLIST_TYPE = 'past broadcasts'
 277
 278     _TEST = {
 279         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 280         'info_dict': {
 281             'id': 'spamfish',
 282             'title': 'Spamfish',
 283         },
 284         'playlist_mincount': 54,
 285     }
 286
 287
 288 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 289     IE_NAME = 'twitch:bookmarks'
 290     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 291     _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 292     _PLAYLIST_TYPE = 'bookmarks'
 293
 294     _TEST = {
 295         'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
 296         'info_dict': {
 297             'id': 'ognos',
 298             'title': 'Ognos',
 299         },
 300         'playlist_mincount': 3,
 301     }
 302
 303     def _extract_playlist_page(self, response):
 304         entries = []
 305         for bookmark in response.get('bookmarks', []):
 306             video = bookmark.get('video')
 307             if not video:
 308                 continue
 309             entries.append(video['url'])
 310         return entries
 311
 312
 313 class TwitchStreamIE(TwitchBaseIE):
 314     IE_NAME = 'twitch:stream'
 315     _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 316
 317     _TEST = {
 318         'url': 'http://www.twitch.tv/shroomztv',
 319         'info_dict': {
 320             'id': '12772022048',
 321             'display_id': 'shroomztv',
 322             'ext': 'mp4',
 323             'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
 324             'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
 325             'is_live': True,
 326             'timestamp': 1421928037,
 327             'upload_date': '20150122',
 328             'uploader': 'ShroomzTV',
 329             'uploader_id': 'shroomztv',
 330             'view_count': int,
 331         },
 332         'params': {
 333             # m3u8 download
 334             'skip_download': True,
 335         },
 336     }
 337
 338     def _real_extract(self, url):
 339         channel_id = self._match_id(url)
 340
 341         stream = self._download_json(
 342             '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
 343             'Downloading stream JSON').get('stream')
 344
 345         # Fallback on profile extraction if stream is offline
 346         if not stream:
 347             return self.url_result(
 348                 'http://www.twitch.tv/%s/profile' % channel_id,
 349                 'TwitchProfile', channel_id)
 350
 351         access_token = self._download_json(
 352             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
 353             'Downloading channel access token')
 354
 355         query = {
 356             'allow_source': 'true',
 357             'p': random.randint(1000000, 10000000),
 358             'player': 'twitchweb',
 359             'segment_preference': '4',
 360             'sig': access_token['sig'].encode('utf-8'),
 361             'token': access_token['token'].encode('utf-8'),
 362         }
 363         formats = self._extract_m3u8_formats(
 364             '%s/api/channel/hls/%s.m3u8?%s'
 365             % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)),
 366             channel_id, 'mp4')
 367         self._prefer_source(formats)
 368
 369         view_count = stream.get('viewers')
 370         timestamp = parse_iso8601(stream.get('created_at'))
 371
 372         channel = stream['channel']
 373         title = self._live_title(channel.get('display_name') or channel.get('name'))
 374         description = channel.get('status')
 375
 376         thumbnails = []
 377         for thumbnail_key, thumbnail_url in stream['preview'].items():
 378             m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
 379             if not m:
 380                 continue
 381             thumbnails.append({
 382                 'url': thumbnail_url,
 383                 'width': int(m.group('width')),
 384                 'height': int(m.group('height')),
 385             })
 386
 387         return {
 388             'id': compat_str(stream['_id']),
 389             'display_id': channel_id,
 390             'title': title,
 391             'description': description,
 392             'thumbnails': thumbnails,
 393             'uploader': channel.get('display_name'),
 394             'uploader_id': channel.get('name'),
 395             'timestamp': timestamp,
 396             'view_count': view_count,
 397             'formats': formats,
 398             'is_live': True,
 399         }