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