yt_dlp/extractor/turner.py

   1 import re
   2
   3 from .adobepass import AdobePassIE
   4 from ..utils import (
   5     ExtractorError,
   6     determine_ext,
   7     fix_xml_ampersands,
   8     float_or_none,
   9     int_or_none,
  10     parse_duration,
  11     strip_or_none,
  12     update_url_query,
  13     url_or_none,
  14     xpath_attr,
  15     xpath_text,
  16 )
  17
  18
  19 class TurnerBaseIE(AdobePassIE):
  20     _AKAMAI_SPE_TOKEN_CACHE = {}
  21
  22     def _extract_timestamp(self, video_data):
  23         return int_or_none(xpath_attr(video_data, 'dateCreated', 'uts'))
  24
  25     def _add_akamai_spe_token(self, tokenizer_src, video_url, content_id, ap_data, custom_tokenizer_query=None):
  26         secure_path = self._search_regex(r'https?://[^/]+(.+/)', video_url, 'secure path') + '*'
  27         token = self._AKAMAI_SPE_TOKEN_CACHE.get(secure_path)
  28         if not token:
  29             query = {
  30                 'path': secure_path,
  31             }
  32             if custom_tokenizer_query:
  33                 query.update(custom_tokenizer_query)
  34             else:
  35                 query['videoId'] = content_id
  36             if ap_data.get('auth_required'):
  37                 query['accessToken'] = self._extract_mvpd_auth(ap_data['url'], content_id, ap_data['site_name'], ap_data['site_name'])
  38             auth = self._download_xml(
  39                 tokenizer_src, content_id, query=query)
  40             error_msg = xpath_text(auth, 'error/msg')
  41             if error_msg:
  42                 raise ExtractorError(error_msg, expected=True)
  43             token = xpath_text(auth, 'token')
  44             if not token:
  45                 return video_url
  46             self._AKAMAI_SPE_TOKEN_CACHE[secure_path] = token
  47         return video_url + '?hdnea=' + token
  48
  49     def _extract_cvp_info(self, data_src, video_id, path_data={}, ap_data={}, fatal=False):
  50         video_data = self._download_xml(
  51             data_src, video_id,
  52             transform_source=lambda s: fix_xml_ampersands(s).strip(),
  53             fatal=fatal)
  54         if not video_data:
  55             return {}
  56         video_id = video_data.attrib['id']
  57         title = xpath_text(video_data, 'headline', fatal=True)
  58         content_id = xpath_text(video_data, 'contentId') or video_id
  59         # rtmp_src = xpath_text(video_data, 'akamai/src')
  60         # if rtmp_src:
  61         #     split_rtmp_src = rtmp_src.split(',')
  62         #     if len(split_rtmp_src) == 2:
  63         #         rtmp_src = split_rtmp_src[1]
  64         # aifp = xpath_text(video_data, 'akamai/aifp', default='')
  65
  66         urls = []
  67         formats = []
  68         thumbnails = []
  69         subtitles = {}
  70         rex = re.compile(
  71             r'(?P<width>[0-9]+)x(?P<height>[0-9]+)(?:_(?P<bitrate>[0-9]+))?')
  72         # Possible formats locations: files/file, files/groupFiles/files
  73         # and maybe others
  74         for video_file in video_data.findall('.//file'):
  75             video_url = url_or_none(video_file.text.strip())
  76             if not video_url:
  77                 continue
  78             ext = determine_ext(video_url)
  79             if video_url.startswith('/mp4:protected/'):
  80                 continue
  81                 # TODO: Correct extraction for these files
  82                 # protected_path_data = path_data.get('protected')
  83                 # if not protected_path_data or not rtmp_src:
  84                 #     continue
  85                 # protected_path = self._search_regex(
  86                 #     r'/mp4:(.+)\.[a-z0-9]', video_url, 'secure path')
  87                 # auth = self._download_webpage(
  88                 #     protected_path_data['tokenizer_src'], query={
  89                 #         'path': protected_path,
  90                 #         'videoId': content_id,
  91                 #         'aifp': aifp,
  92                 #     })
  93                 # token = xpath_text(auth, 'token')
  94                 # if not token:
  95                 #     continue
  96                 # video_url = rtmp_src + video_url + '?' + token
  97             elif video_url.startswith('/secure/'):
  98                 secure_path_data = path_data.get('secure')
  99                 if not secure_path_data:
 100                     continue
 101                 video_url = self._add_akamai_spe_token(
 102                     secure_path_data['tokenizer_src'],
 103                     secure_path_data['media_src'] + video_url,
 104                     content_id, ap_data)
 105             elif not re.match('https?://', video_url):
 106                 base_path_data = path_data.get(ext, path_data.get('default', {}))
 107                 media_src = base_path_data.get('media_src')
 108                 if not media_src:
 109                     continue
 110                 video_url = media_src + video_url
 111             if video_url in urls:
 112                 continue
 113             urls.append(video_url)
 114             format_id = video_file.get('bitrate')
 115             if ext in ('scc', 'srt', 'vtt'):
 116                 subtitles.setdefault('en', []).append({
 117                     'ext': ext,
 118                     'url': video_url,
 119                 })
 120             elif ext == 'png':
 121                 thumbnails.append({
 122                     'id': format_id,
 123                     'url': video_url,
 124                 })
 125             elif ext == 'smil':
 126                 formats.extend(self._extract_smil_formats(
 127                     video_url, video_id, fatal=False))
 128             elif re.match(r'https?://[^/]+\.akamaihd\.net/[iz]/', video_url):
 129                 formats.extend(self._extract_akamai_formats(
 130                     video_url, video_id, {
 131                         'hds': path_data.get('f4m', {}).get('host'),
 132                         # nba.cdn.turner.com, ht.cdn.turner.com, ht2.cdn.turner.com
 133                         # ht3.cdn.turner.com, i.cdn.turner.com, s.cdn.turner.com
 134                         # ssl.cdn.turner.com
 135                         'http': 'pmd.cdn.turner.com',
 136                     }))
 137             elif ext == 'm3u8':
 138                 m3u8_formats = self._extract_m3u8_formats(
 139                     video_url, video_id, 'mp4',
 140                     m3u8_id=format_id or 'hls', fatal=False)
 141                 if '/secure/' in video_url and '?hdnea=' in video_url:
 142                     for f in m3u8_formats:
 143                         f['downloader_options'] = {'ffmpeg_args': ['-seekable', '0']}
 144                 formats.extend(m3u8_formats)
 145             elif ext == 'f4m':
 146                 formats.extend(self._extract_f4m_formats(
 147                     update_url_query(video_url, {'hdcore': '3.7.0'}),
 148                     video_id, f4m_id=format_id or 'hds', fatal=False))
 149             else:
 150                 f = {
 151                     'format_id': format_id,
 152                     'url': video_url,
 153                     'ext': ext,
 154                 }
 155                 mobj = rex.search(video_url)
 156                 if mobj:
 157                     f.update({
 158                         'width': int(mobj.group('width')),
 159                         'height': int(mobj.group('height')),
 160                         'tbr': int_or_none(mobj.group('bitrate')),
 161                     })
 162                 elif isinstance(format_id, str):
 163                     if format_id.isdigit():
 164                         f['tbr'] = int(format_id)
 165                     else:
 166                         mobj = re.match(r'ios_(audio|[0-9]+)$', format_id)
 167                         if mobj:
 168                             if mobj.group(1) == 'audio':
 169                                 f.update({
 170                                     'vcodec': 'none',
 171                                     'ext': 'm4a',
 172                                 })
 173                             else:
 174                                 f['tbr'] = int(mobj.group(1))
 175                 formats.append(f)
 176
 177         for source in video_data.findall('closedCaptions/source'):
 178             for track in source.findall('track'):
 179                 track_url = url_or_none(track.get('url'))
 180                 if not track_url or track_url.endswith('/big'):
 181                     continue
 182                 lang = track.get('lang') or track.get('label') or 'en'
 183                 subtitles.setdefault(lang, []).append({
 184                     'url': track_url,
 185                     'ext': {
 186                         'scc': 'scc',
 187                         'webvtt': 'vtt',
 188                         'smptett': 'tt',
 189                     }.get(source.get('format')),
 190                 })
 191
 192         thumbnails.extend({
 193             'id': image.get('cut') or image.get('name'),
 194             'url': image.text,
 195             'width': int_or_none(image.get('width')),
 196             'height': int_or_none(image.get('height')),
 197         } for image in video_data.findall('images/image'))
 198
 199         is_live = xpath_text(video_data, 'isLive') == 'true'
 200
 201         return {
 202             'id': video_id,
 203             'title': title,
 204             'formats': formats,
 205             'subtitles': subtitles,
 206             'thumbnails': thumbnails,
 207             'thumbnail': xpath_text(video_data, 'poster'),
 208             'description': strip_or_none(xpath_text(video_data, 'description')),
 209             'duration': parse_duration(xpath_text(video_data, 'length') or xpath_text(video_data, 'trt')),
 210             'timestamp': self._extract_timestamp(video_data),
 211             'upload_date': xpath_attr(video_data, 'metas', 'version'),
 212             'series': xpath_text(video_data, 'showTitle'),
 213             'season_number': int_or_none(xpath_text(video_data, 'seasonNumber')),
 214             'episode_number': int_or_none(xpath_text(video_data, 'episodeNumber')),
 215             'is_live': is_live,
 216         }
 217
 218     def _extract_ngtv_info(self, media_id, tokenizer_query, ap_data=None):
 219         is_live = ap_data.get('is_live')
 220         streams_data = self._download_json(
 221             f'http://medium.ngtv.io/media/{media_id}/tv',
 222             media_id)['media']['tv']
 223         duration = None
 224         chapters = []
 225         formats = []
 226         for supported_type in ('unprotected', 'bulkaes'):
 227             stream_data = streams_data.get(supported_type, {})
 228             m3u8_url = stream_data.get('secureUrl') or stream_data.get('url')
 229             if not m3u8_url:
 230                 continue
 231             if stream_data.get('playlistProtection') == 'spe':
 232                 m3u8_url = self._add_akamai_spe_token(
 233                     'http://token.ngtv.io/token/token_spe',
 234                     m3u8_url, media_id, ap_data or {}, tokenizer_query)
 235             formats.extend(self._extract_m3u8_formats(
 236                 m3u8_url, media_id, 'mp4', m3u8_id='hls', live=is_live, fatal=False))
 237
 238             duration = float_or_none(stream_data.get('totalRuntime'))
 239
 240             if not chapters and not is_live:
 241                 for chapter in stream_data.get('contentSegments', []):
 242                     start_time = float_or_none(chapter.get('start'))
 243                     chapter_duration = float_or_none(chapter.get('duration'))
 244                     if start_time is None or chapter_duration is None:
 245                         continue
 246                     chapters.append({
 247                         'start_time': start_time,
 248                         'end_time': start_time + chapter_duration,
 249                     })
 250
 251         return {
 252             'formats': formats,
 253             'chapters': chapters,
 254             'duration': duration,
 255         }