yt_dlp/extractor/turner.py

   1 import re
   2
   3 from .adobepass import AdobePassIE
   4 from ..compat import compat_str
   5 from ..utils import (
   6     ExtractorError,
   7     determine_ext,
   8     fix_xml_ampersands,
   9     float_or_none,
  10     int_or_none,
  11     parse_duration,
  12     strip_or_none,
  13     update_url_query,
  14     url_or_none,
  15     xpath_attr,
  16     xpath_text,
  17 )
  18
  19
  20 class TurnerBaseIE(AdobePassIE):
  21     _AKAMAI_SPE_TOKEN_CACHE = {}
  22
  23     def _extract_timestamp(self, video_data):
  24         return int_or_none(xpath_attr(video_data, 'dateCreated', 'uts'))
  25
  26     def _add_akamai_spe_token(self, tokenizer_src, video_url, content_id, ap_data, custom_tokenizer_query=None):
  27         secure_path = self._search_regex(r'https?://[^/]+(.+/)', video_url, 'secure path') + '*'
  28         token = self._AKAMAI_SPE_TOKEN_CACHE.get(secure_path)
  29         if not token:
  30             query = {
  31                 'path': secure_path,
  32             }
  33             if custom_tokenizer_query:
  34                 query.update(custom_tokenizer_query)
  35             else:
  36                 query['videoId'] = content_id
  37             if ap_data.get('auth_required'):
  38                 query['accessToken'] = self._extract_mvpd_auth(ap_data['url'], content_id, ap_data['site_name'], ap_data['site_name'])
  39             auth = self._download_xml(
  40                 tokenizer_src, content_id, query=query)
  41             error_msg = xpath_text(auth, 'error/msg')
  42             if error_msg:
  43                 raise ExtractorError(error_msg, expected=True)
  44             token = xpath_text(auth, 'token')
  45             if not token:
  46                 return video_url
  47             self._AKAMAI_SPE_TOKEN_CACHE[secure_path] = token
  48         return video_url + '?hdnea=' + token
  49
  50     def _extract_cvp_info(self, data_src, video_id, path_data={}, ap_data={}, fatal=False):
  51         video_data = self._download_xml(
  52             data_src, video_id,
  53             transform_source=lambda s: fix_xml_ampersands(s).strip(),
  54             fatal=fatal)
  55         if not video_data:
  56             return {}
  57         video_id = video_data.attrib['id']
  58         title = xpath_text(video_data, 'headline', fatal=True)
  59         content_id = xpath_text(video_data, 'contentId') or video_id
  60         # rtmp_src = xpath_text(video_data, 'akamai/src')
  61         # if rtmp_src:
  62         #     split_rtmp_src = rtmp_src.split(',')
  63         #     if len(split_rtmp_src) == 2:
  64         #         rtmp_src = split_rtmp_src[1]
  65         # aifp = xpath_text(video_data, 'akamai/aifp', default='')
  66
  67         urls = []
  68         formats = []
  69         thumbnails = []
  70         subtitles = {}
  71         rex = re.compile(
  72             r'(?P<width>[0-9]+)x(?P<height>[0-9]+)(?:_(?P<bitrate>[0-9]+))?')
  73         # Possible formats locations: files/file, files/groupFiles/files
  74         # and maybe others
  75         for video_file in video_data.findall('.//file'):
  76             video_url = url_or_none(video_file.text.strip())
  77             if not video_url:
  78                 continue
  79             ext = determine_ext(video_url)
  80             if video_url.startswith('/mp4:protected/'):
  81                 continue
  82                 # TODO Correct extraction for these files
  83                 # protected_path_data = path_data.get('protected')
  84                 # if not protected_path_data or not rtmp_src:
  85                 #     continue
  86                 # protected_path = self._search_regex(
  87                 #     r'/mp4:(.+)\.[a-z0-9]', video_url, 'secure path')
  88                 # auth = self._download_webpage(
  89                 #     protected_path_data['tokenizer_src'], query={
  90                 #         'path': protected_path,
  91                 #         'videoId': content_id,
  92                 #         'aifp': aifp,
  93                 #     })
  94                 # token = xpath_text(auth, 'token')
  95                 # if not token:
  96                 #     continue
  97                 # video_url = rtmp_src + video_url + '?' + token
  98             elif video_url.startswith('/secure/'):
  99                 secure_path_data = path_data.get('secure')
 100                 if not secure_path_data:
 101                     continue
 102                 video_url = self._add_akamai_spe_token(
 103                     secure_path_data['tokenizer_src'],
 104                     secure_path_data['media_src'] + video_url,
 105                     content_id, ap_data)
 106             elif not re.match('https?://', video_url):
 107                 base_path_data = path_data.get(ext, path_data.get('default', {}))
 108                 media_src = base_path_data.get('media_src')
 109                 if not media_src:
 110                     continue
 111                 video_url = media_src + video_url
 112             if video_url in urls:
 113                 continue
 114             urls.append(video_url)
 115             format_id = video_file.get('bitrate')
 116             if ext in ('scc', 'srt', 'vtt'):
 117                 subtitles.setdefault('en', []).append({
 118                     'ext': ext,
 119                     'url': video_url,
 120                 })
 121             elif ext == 'png':
 122                 thumbnails.append({
 123                     'id': format_id,
 124                     'url': video_url,
 125                 })
 126             elif ext == 'smil':
 127                 formats.extend(self._extract_smil_formats(
 128                     video_url, video_id, fatal=False))
 129             elif re.match(r'https?://[^/]+\.akamaihd\.net/[iz]/', video_url):
 130                 formats.extend(self._extract_akamai_formats(
 131                     video_url, video_id, {
 132                         'hds': path_data.get('f4m', {}).get('host'),
 133                         # nba.cdn.turner.com, ht.cdn.turner.com, ht2.cdn.turner.com
 134                         # ht3.cdn.turner.com, i.cdn.turner.com, s.cdn.turner.com
 135                         # ssl.cdn.turner.com
 136                         'http': 'pmd.cdn.turner.com',
 137                     }))
 138             elif ext == 'm3u8':
 139                 m3u8_formats = self._extract_m3u8_formats(
 140                     video_url, video_id, 'mp4',
 141                     m3u8_id=format_id or 'hls', fatal=False)
 142                 if '/secure/' in video_url and '?hdnea=' in video_url:
 143                     for f in m3u8_formats:
 144                         f['downloader_options'] = {'ffmpeg_args': ['-seekable', '0']}
 145                 formats.extend(m3u8_formats)
 146             elif ext == 'f4m':
 147                 formats.extend(self._extract_f4m_formats(
 148                     update_url_query(video_url, {'hdcore': '3.7.0'}),
 149                     video_id, f4m_id=format_id or 'hds', fatal=False))
 150             else:
 151                 f = {
 152                     'format_id': format_id,
 153                     'url': video_url,
 154                     'ext': ext,
 155                 }
 156                 mobj = rex.search(video_url)
 157                 if mobj:
 158                     f.update({
 159                         'width': int(mobj.group('width')),
 160                         'height': int(mobj.group('height')),
 161                         'tbr': int_or_none(mobj.group('bitrate')),
 162                     })
 163                 elif isinstance(format_id, compat_str):
 164                     if format_id.isdigit():
 165                         f['tbr'] = int(format_id)
 166                     else:
 167                         mobj = re.match(r'ios_(audio|[0-9]+)$', format_id)
 168                         if mobj:
 169                             if mobj.group(1) == 'audio':
 170                                 f.update({
 171                                     'vcodec': 'none',
 172                                     'ext': 'm4a',
 173                                 })
 174                             else:
 175                                 f['tbr'] = int(mobj.group(1))
 176                 formats.append(f)
 177
 178         for source in video_data.findall('closedCaptions/source'):
 179             for track in source.findall('track'):
 180                 track_url = url_or_none(track.get('url'))
 181                 if not track_url or track_url.endswith('/big'):
 182                     continue
 183                 lang = track.get('lang') or track.get('label') or 'en'
 184                 subtitles.setdefault(lang, []).append({
 185                     'url': track_url,
 186                     'ext': {
 187                         'scc': 'scc',
 188                         'webvtt': 'vtt',
 189                         'smptett': 'tt',
 190                     }.get(source.get('format'))
 191                 })
 192
 193         thumbnails.extend({
 194             'id': image.get('cut') or image.get('name'),
 195             'url': image.text,
 196             'width': int_or_none(image.get('width')),
 197             'height': int_or_none(image.get('height')),
 198         } for image in video_data.findall('images/image'))
 199
 200         is_live = xpath_text(video_data, 'isLive') == 'true'
 201
 202         return {
 203             'id': video_id,
 204             'title': title,
 205             'formats': formats,
 206             'subtitles': subtitles,
 207             'thumbnails': thumbnails,
 208             'thumbnail': xpath_text(video_data, 'poster'),
 209             'description': strip_or_none(xpath_text(video_data, 'description')),
 210             'duration': parse_duration(xpath_text(video_data, 'length') or xpath_text(video_data, 'trt')),
 211             'timestamp': self._extract_timestamp(video_data),
 212             'upload_date': xpath_attr(video_data, 'metas', 'version'),
 213             'series': xpath_text(video_data, 'showTitle'),
 214             'season_number': int_or_none(xpath_text(video_data, 'seasonNumber')),
 215             'episode_number': int_or_none(xpath_text(video_data, 'episodeNumber')),
 216             'is_live': is_live,
 217         }
 218
 219     def _extract_ngtv_info(self, media_id, tokenizer_query, ap_data=None):
 220         is_live = ap_data.get('is_live')
 221         streams_data = self._download_json(
 222             'http://medium.ngtv.io/media/%s/tv' % media_id,
 223             media_id)['media']['tv']
 224         duration = None
 225         chapters = []
 226         formats = []
 227         for supported_type in ('unprotected', 'bulkaes'):
 228             stream_data = streams_data.get(supported_type, {})
 229             m3u8_url = stream_data.get('secureUrl') or stream_data.get('url')
 230             if not m3u8_url:
 231                 continue
 232             if stream_data.get('playlistProtection') == 'spe':
 233                 m3u8_url = self._add_akamai_spe_token(
 234                     'http://token.ngtv.io/token/token_spe',
 235                     m3u8_url, media_id, ap_data or {}, tokenizer_query)
 236             formats.extend(self._extract_m3u8_formats(
 237                 m3u8_url, media_id, 'mp4', m3u8_id='hls', live=is_live, fatal=False))
 238
 239             duration = float_or_none(stream_data.get('totalRuntime'))
 240
 241             if not chapters and not is_live:
 242                 for chapter in stream_data.get('contentSegments', []):
 243                     start_time = float_or_none(chapter.get('start'))
 244                     chapter_duration = float_or_none(chapter.get('duration'))
 245                     if start_time is None or chapter_duration is None:
 246                         continue
 247                     chapters.append({
 248                         'start_time': start_time,
 249                         'end_time': start_time + chapter_duration,
 250                     })
 251
 252         return {
 253             'formats': formats,
 254             'chapters': chapters,
 255             'duration': duration,
 256         }