yt_dlp/extractor/radiko.py

   1 import base64
   2 import random
   3 import re
   4 import urllib.parse
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     ExtractorError,
   9     clean_html,
  10     time_seconds,
  11     try_call,
  12     unified_timestamp,
  13     update_url_query,
  14 )
  15 from ..utils.traversal import traverse_obj
  16
  17
  18 class RadikoBaseIE(InfoExtractor):
  19     _GEO_BYPASS = False
  20     _FULL_KEY = None
  21     _HOSTS_FOR_TIME_FREE_FFMPEG_UNSUPPORTED = (
  22         'https://c-rpaa.smartstream.ne.jp',
  23         'https://si-c-radiko.smartstream.ne.jp',
  24         'https://tf-f-rpaa-radiko.smartstream.ne.jp',
  25         'https://tf-c-rpaa-radiko.smartstream.ne.jp',
  26         'https://si-f-radiko.smartstream.ne.jp',
  27         'https://rpaa.smartstream.ne.jp',
  28     )
  29     _HOSTS_FOR_TIME_FREE_FFMPEG_SUPPORTED = (
  30         'https://rd-wowza-radiko.radiko-cf.com',
  31         'https://radiko.jp',
  32         'https://f-radiko.smartstream.ne.jp',
  33     )
  34     # Following URL forcibly connects not Time Free but Live
  35     _HOSTS_FOR_LIVE = (
  36         'https://c-radiko.smartstream.ne.jp',
  37     )
  38
  39     def _negotiate_token(self):
  40         _, auth1_handle = self._download_webpage_handle(
  41             'https://radiko.jp/v2/api/auth1', None, 'Downloading authentication page',
  42             headers={
  43                 'x-radiko-app': 'pc_html5',
  44                 'x-radiko-app-version': '0.0.1',
  45                 'x-radiko-device': 'pc',
  46                 'x-radiko-user': 'dummy_user',
  47             })
  48         auth1_header = auth1_handle.headers
  49
  50         auth_token = auth1_header['X-Radiko-AuthToken']
  51         kl = int(auth1_header['X-Radiko-KeyLength'])
  52         ko = int(auth1_header['X-Radiko-KeyOffset'])
  53         raw_partial_key = self._extract_full_key()[ko:ko + kl]
  54         partial_key = base64.b64encode(raw_partial_key).decode()
  55
  56         area_id = self._download_webpage(
  57             'https://radiko.jp/v2/api/auth2', None, 'Authenticating',
  58             headers={
  59                 'x-radiko-device': 'pc',
  60                 'x-radiko-user': 'dummy_user',
  61                 'x-radiko-authtoken': auth_token,
  62                 'x-radiko-partialkey': partial_key,
  63             }).split(',')[0]
  64
  65         if area_id == 'OUT':
  66             self.raise_geo_restricted(countries=['JP'])
  67
  68         auth_data = (auth_token, area_id)
  69         self.cache.store('radiko', 'auth_data', auth_data)
  70         return auth_data
  71
  72     def _auth_client(self):
  73         cachedata = self.cache.load('radiko', 'auth_data')
  74         if cachedata is not None:
  75             response = self._download_webpage(
  76                 'https://radiko.jp/v2/api/auth_check', None, 'Checking cached token', expected_status=401,
  77                 headers={'X-Radiko-AuthToken': cachedata[0], 'X-Radiko-AreaId': cachedata[1]})
  78             if response == 'OK':
  79                 return cachedata
  80         return self._negotiate_token()
  81
  82     def _extract_full_key(self):
  83         if self._FULL_KEY:
  84             return self._FULL_KEY
  85
  86         jscode = self._download_webpage(
  87             'https://radiko.jp/apps/js/playerCommon.js', None,
  88             note='Downloading player js code')
  89         full_key = self._search_regex(
  90             (r"RadikoJSPlayer\([^,]*,\s*(['\"])pc_html5\1,\s*(['\"])(?P<fullkey>[0-9a-f]+)\2,\s*{"),
  91             jscode, 'full key', fatal=False, group='fullkey')
  92
  93         if full_key:
  94             full_key = full_key.encode()
  95         else:  # use only full key ever known
  96             full_key = b'bcd151073c03b352e1ef2fd66c32209da9ca0afa'
  97
  98         self._FULL_KEY = full_key
  99         return full_key
 100
 101     def _find_program(self, video_id, station, cursor):
 102         station_program = self._download_xml(
 103             'https://radiko.jp/v3/program/station/weekly/%s.xml' % station, video_id,
 104             note='Downloading radio program for %s station' % station)
 105
 106         prog = None
 107         for p in station_program.findall('.//prog'):
 108             ft_str, to_str = p.attrib['ft'], p.attrib['to']
 109             ft = unified_timestamp(ft_str, False)
 110             to = unified_timestamp(to_str, False)
 111             if ft <= cursor and cursor < to:
 112                 prog = p
 113                 break
 114         if not prog:
 115             raise ExtractorError('Cannot identify radio program to download!')
 116         assert ft, to
 117         return prog, station_program, ft, ft_str, to_str
 118
 119     def _extract_formats(self, video_id, station, is_onair, ft, cursor, auth_token, area_id, query):
 120         m3u8_playlist_data = self._download_xml(
 121             f'https://radiko.jp/v3/station/stream/pc_html5/{station}.xml', video_id,
 122             note='Downloading stream information')
 123
 124         formats = []
 125         found = set()
 126
 127         timefree_int = 0 if is_onair else 1
 128
 129         for element in m3u8_playlist_data.findall(f'.//url[@timefree="{timefree_int}"]/playlist_create_url'):
 130             pcu = element.text
 131             if pcu in found:
 132                 continue
 133             found.add(pcu)
 134             playlist_url = update_url_query(pcu, {
 135                 'station_id': station,
 136                 **query,
 137                 'l': '15',
 138                 'lsid': ''.join(random.choices('0123456789abcdef', k=32)),
 139                 'type': 'b',
 140             })
 141
 142             time_to_skip = None if is_onair else cursor - ft
 143
 144             domain = urllib.parse.urlparse(playlist_url).netloc
 145             subformats = self._extract_m3u8_formats(
 146                 playlist_url, video_id, ext='m4a',
 147                 live=True, fatal=False, m3u8_id=domain,
 148                 note=f'Downloading m3u8 information from {domain}',
 149                 headers={
 150                     'X-Radiko-AreaId': area_id,
 151                     'X-Radiko-AuthToken': auth_token,
 152                 })
 153             for sf in subformats:
 154                 if (is_onair ^ pcu.startswith(self._HOSTS_FOR_LIVE)) or (
 155                         not is_onair and pcu.startswith(self._HOSTS_FOR_TIME_FREE_FFMPEG_UNSUPPORTED)):
 156                     sf['preference'] = -100
 157                     sf['format_note'] = 'not preferred'
 158                 if not is_onair and timefree_int == 1 and time_to_skip:
 159                     sf['downloader_options'] = {'ffmpeg_args': ['-ss', str(time_to_skip)]}
 160             formats.extend(subformats)
 161
 162         return formats
 163
 164     def _extract_performers(self, prog):
 165         return traverse_obj(prog, (
 166             'pfm/text()', ..., {lambda x: re.split(r'[/／、　,，]', x)}, ..., {str.strip})) or None
 167
 168
 169 class RadikoIE(RadikoBaseIE):
 170     _VALID_URL = r'https?://(?:www\.)?radiko\.jp/#!/ts/(?P<station>[A-Z0-9-]+)/(?P<id>\d+)'
 171
 172     _TESTS = [{
 173         # QRR (文化放送) station provides <desc>
 174         'url': 'https://radiko.jp/#!/ts/QRR/20210425101300',
 175         'only_matching': True,
 176     }, {
 177         # FMT (TOKYO FM) station does not provide <desc>
 178         'url': 'https://radiko.jp/#!/ts/FMT/20210810150000',
 179         'only_matching': True,
 180     }, {
 181         'url': 'https://radiko.jp/#!/ts/JOAK-FM/20210509090000',
 182         'only_matching': True,
 183     }]
 184
 185     def _real_extract(self, url):
 186         station, video_id = self._match_valid_url(url).groups()
 187         vid_int = unified_timestamp(video_id, False)
 188         prog, station_program, ft, radio_begin, radio_end = self._find_program(video_id, station, vid_int)
 189
 190         auth_token, area_id = self._auth_client()
 191
 192         return {
 193             'id': video_id,
 194             'title': try_call(lambda: prog.find('title').text),
 195             'cast': self._extract_performers(prog),
 196             'description': clean_html(try_call(lambda: prog.find('info').text)),
 197             'uploader': try_call(lambda: station_program.find('.//name').text),
 198             'uploader_id': station,
 199             'timestamp': vid_int,
 200             'duration': try_call(lambda: unified_timestamp(radio_end, False) - unified_timestamp(radio_begin, False)),
 201             'is_live': True,
 202             'formats': self._extract_formats(
 203                 video_id=video_id, station=station, is_onair=False,
 204                 ft=ft, cursor=vid_int, auth_token=auth_token, area_id=area_id,
 205                 query={
 206                     'start_at': radio_begin,
 207                     'ft': radio_begin,
 208                     'end_at': radio_end,
 209                     'to': radio_end,
 210                     'seek': video_id
 211                 }
 212             ),
 213         }
 214
 215
 216 class RadikoRadioIE(RadikoBaseIE):
 217     _VALID_URL = r'https?://(?:www\.)?radiko\.jp/#!/live/(?P<id>[A-Z0-9-]+)'
 218
 219     _TESTS = [{
 220         # QRR (文化放送) station provides <desc>
 221         'url': 'https://radiko.jp/#!/live/QRR',
 222         'only_matching': True,
 223     }, {
 224         # FMT (TOKYO FM) station does not provide <desc>
 225         'url': 'https://radiko.jp/#!/live/FMT',
 226         'only_matching': True,
 227     }, {
 228         'url': 'https://radiko.jp/#!/live/JOAK-FM',
 229         'only_matching': True,
 230     }]
 231
 232     def _real_extract(self, url):
 233         station = self._match_id(url)
 234         self.report_warning('Downloader will not stop at the end of the program! Press Ctrl+C to stop')
 235
 236         auth_token, area_id = self._auth_client()
 237         # get current time in JST (GMT+9:00 w/o DST)
 238         vid_now = time_seconds(hours=9)
 239
 240         prog, station_program, ft, _, _ = self._find_program(station, station, vid_now)
 241
 242         title = prog.find('title').text
 243         description = clean_html(prog.find('info').text)
 244         station_name = station_program.find('.//name').text
 245
 246         formats = self._extract_formats(
 247             video_id=station, station=station, is_onair=True,
 248             ft=ft, cursor=vid_now, auth_token=auth_token, area_id=area_id,
 249             query={})
 250
 251         return {
 252             'id': station,
 253             'title': title,
 254             'cast': self._extract_performers(prog),
 255             'description': description,
 256             'uploader': station_name,
 257             'uploader_id': station,
 258             'timestamp': ft,
 259             'formats': formats,
 260             'is_live': True,
 261         }