youtube_dl/extractor/twitch.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_urllib_parse,
  10     compat_urllib_request,
  11 )
  12 from ..utils import (
  13     ExtractorError,
  14     parse_iso8601,
  15 )
  16
  17
  18 class TwitchBaseIE(InfoExtractor):
  19     _VALID_URL_BASE = r'http://(?:www\.)?twitch\.tv'
  20
  21     _API_BASE = 'https://api.twitch.tv'
  22     _LOGIN_URL = 'https://secure.twitch.tv/user/login'
  23
  24     def _handle_error(self, response):
  25         if not isinstance(response, dict):
  26             return
  27         error = response.get('error')
  28         if error:
  29             raise ExtractorError(
  30                 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
  31                 expected=True)
  32
  33     def _download_json(self, url, video_id, note='Downloading JSON metadata'):
  34         response = super(TwitchBaseIE, self)._download_json(url, video_id, note)
  35         self._handle_error(response)
  36         return response
  37
  38     def _real_initialize(self):
  39         self._login()
  40
  41     def _login(self):
  42         (username, password) = self._get_login_info()
  43         if username is None:
  44             return
  45
  46         login_page = self._download_webpage(
  47             self._LOGIN_URL, None, 'Downloading login page')
  48
  49         authenticity_token = self._search_regex(
  50             r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
  51             login_page, 'authenticity token')
  52
  53         login_form = {
  54             'utf8': '✓'.encode('utf-8'),
  55             'authenticity_token': authenticity_token,
  56             'redirect_on_login': '',
  57             'embed_form': 'false',
  58             'mp_source_action': '',
  59             'follow': '',
  60             'user[login]': username,
  61             'user[password]': password,
  62         }
  63
  64         request = compat_urllib_request.Request(
  65             self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
  66         request.add_header('Referer', self._LOGIN_URL)
  67         response = self._download_webpage(
  68             request, None, 'Logging in as %s' % username)
  69
  70         m = re.search(
  71             r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
  72         if m:
  73             raise ExtractorError(
  74                 'Unable to login: %s' % m.group('msg').strip(), expected=True)
  75
  76
  77 class TwitchItemBaseIE(TwitchBaseIE):
  78     def _download_info(self, item, item_id):
  79         return self._extract_info(self._download_json(
  80             '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
  81             'Downloading %s info JSON' % self._ITEM_TYPE))
  82
  83     def _extract_media(self, item_id):
  84         info = self._download_info(self._ITEM_SHORTCUT, item_id)
  85         response = self._download_json(
  86             '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
  87             'Downloading %s playlist JSON' % self._ITEM_TYPE)
  88         entries = []
  89         chunks = response['chunks']
  90         qualities = list(chunks.keys())
  91         for num, fragment in enumerate(zip(*chunks.values()), start=1):
  92             formats = []
  93             for fmt_num, fragment_fmt in enumerate(fragment):
  94                 format_id = qualities[fmt_num]
  95                 fmt = {
  96                     'url': fragment_fmt['url'],
  97                     'format_id': format_id,
  98                     'quality': 1 if format_id == 'live' else 0,
  99                 }
 100                 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
 101                 if m:
 102                     fmt['height'] = int(m.group('height'))
 103                 formats.append(fmt)
 104             self._sort_formats(formats)
 105             entry = dict(info)
 106             entry['id'] = '%s_%d' % (entry['id'], num)
 107             entry['title'] = '%s part %d' % (entry['title'], num)
 108             entry['formats'] = formats
 109             entries.append(entry)
 110         return self.playlist_result(entries, info['id'], info['title'])
 111
 112     def _extract_info(self, info):
 113         return {
 114             'id': info['_id'],
 115             'title': info['title'],
 116             'description': info['description'],
 117             'duration': info['length'],
 118             'thumbnail': info['preview'],
 119             'uploader': info['channel']['display_name'],
 120             'uploader_id': info['channel']['name'],
 121             'timestamp': parse_iso8601(info['recorded_at']),
 122             'view_count': info['views'],
 123         }
 124
 125     def _real_extract(self, url):
 126         return self._extract_media(self._match_id(url))
 127
 128
 129 class TwitchVideoIE(TwitchItemBaseIE):
 130     IE_NAME = 'twitch:video'
 131     _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 132     _ITEM_TYPE = 'video'
 133     _ITEM_SHORTCUT = 'a'
 134
 135     _TEST = {
 136         'url': 'http://www.twitch.tv/riotgames/b/577357806',
 137         'info_dict': {
 138             'id': 'a577357806',
 139             'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
 140         },
 141         'playlist_mincount': 12,
 142     }
 143
 144
 145 class TwitchChapterIE(TwitchItemBaseIE):
 146     IE_NAME = 'twitch:chapter'
 147     _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 148     _ITEM_TYPE = 'chapter'
 149     _ITEM_SHORTCUT = 'c'
 150
 151     _TESTS = [{
 152         'url': 'http://www.twitch.tv/acracingleague/c/5285812',
 153         'info_dict': {
 154             'id': 'c5285812',
 155             'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
 156         },
 157         'playlist_mincount': 3,
 158     }, {
 159         'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
 160         'only_matching': True,
 161     }]
 162
 163
 164 class TwitchVodIE(TwitchItemBaseIE):
 165     IE_NAME = 'twitch:vod'
 166     _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
 167     _ITEM_TYPE = 'vod'
 168     _ITEM_SHORTCUT = 'v'
 169
 170     _TEST = {
 171         'url': 'http://www.twitch.tv/ksptv/v/3622000',
 172         'info_dict': {
 173             'id': 'v3622000',
 174             'ext': 'mp4',
 175             'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
 176             'thumbnail': 're:^https?://.*\.jpg$',
 177             'duration': 6951,
 178             'timestamp': 1419028564,
 179             'upload_date': '20141219',
 180             'uploader': 'KSPTV',
 181             'uploader_id': 'ksptv',
 182             'view_count': int,
 183         },
 184         'params': {
 185             # m3u8 download
 186             'skip_download': True,
 187         },
 188     }
 189
 190     def _real_extract(self, url):
 191         item_id = self._match_id(url)
 192         info = self._download_info(self._ITEM_SHORTCUT, item_id)
 193         access_token = self._download_json(
 194             '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
 195             'Downloading %s access token' % self._ITEM_TYPE)
 196         formats = self._extract_m3u8_formats(
 197             'http://usher.twitch.tv/vod/%s?nauth=%s&nauthsig=%s'
 198             % (item_id, access_token['token'], access_token['sig']),
 199             item_id, 'mp4')
 200         info['formats'] = formats
 201         return info
 202
 203
 204 class TwitchPlaylistBaseIE(TwitchBaseIE):
 205     _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
 206     _PAGE_LIMIT = 100
 207
 208     def _extract_playlist(self, channel_id):
 209         info = self._download_json(
 210             '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
 211             channel_id, 'Downloading channel info JSON')
 212         channel_name = info.get('display_name') or info.get('name')
 213         entries = []
 214         offset = 0
 215         limit = self._PAGE_LIMIT
 216         for counter in itertools.count(1):
 217             response = self._download_json(
 218                 self._PLAYLIST_URL % (channel_id, offset, limit),
 219                 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
 220             videos = response['videos']
 221             if not videos:
 222                 break
 223             entries.extend([self.url_result(video['url']) for video in videos])
 224             offset += limit
 225         return self.playlist_result(entries, channel_id, channel_name)
 226
 227     def _real_extract(self, url):
 228         return self._extract_playlist(self._match_id(url))
 229
 230
 231 class TwitchProfileIE(TwitchPlaylistBaseIE):
 232     IE_NAME = 'twitch:profile'
 233     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 234     _PLAYLIST_TYPE = 'profile'
 235
 236     _TEST = {
 237         'url': 'http://www.twitch.tv/vanillatv/profile',
 238         'info_dict': {
 239             'id': 'vanillatv',
 240             'title': 'VanillaTV',
 241         },
 242         'playlist_mincount': 412,
 243     }
 244
 245
 246 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
 247     IE_NAME = 'twitch:past_broadcasts'
 248     _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 249     _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
 250     _PLAYLIST_TYPE = 'past broadcasts'
 251
 252     _TEST = {
 253         'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
 254         'info_dict': {
 255             'id': 'spamfish',
 256             'title': 'Spamfish',
 257         },
 258         'playlist_mincount': 54,
 259     }