yt_dlp/extractor/turner.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .adobepass import AdobePassIE
   7 from ..compat import compat_str
   8 from ..utils import (
   9     fix_xml_ampersands,
  10     xpath_text,
  11     int_or_none,
  12     determine_ext,
  13     float_or_none,
  14     parse_duration,
  15     xpath_attr,
  16     update_url_query,
  17     ExtractorError,
  18     strip_or_none,
  19     url_or_none,
  20 )
  21
  22
  23 class TurnerBaseIE(AdobePassIE):
  24     _AKAMAI_SPE_TOKEN_CACHE = {}
  25
  26     def _extract_timestamp(self, video_data):
  27         return int_or_none(xpath_attr(video_data, 'dateCreated', 'uts'))
  28
  29     def _add_akamai_spe_token(self, tokenizer_src, video_url, content_id, ap_data, custom_tokenizer_query=None):
  30         secure_path = self._search_regex(r'https?://[^/]+(.+/)', video_url, 'secure path') + '*'
  31         token = self._AKAMAI_SPE_TOKEN_CACHE.get(secure_path)
  32         if not token:
  33             query = {
  34                 'path': secure_path,
  35             }
  36             if custom_tokenizer_query:
  37                 query.update(custom_tokenizer_query)
  38             else:
  39                 query['videoId'] = content_id
  40             if ap_data.get('auth_required'):
  41                 query['accessToken'] = self._extract_mvpd_auth(ap_data['url'], content_id, ap_data['site_name'], ap_data['site_name'])
  42             auth = self._download_xml(
  43                 tokenizer_src, content_id, query=query)
  44             error_msg = xpath_text(auth, 'error/msg')
  45             if error_msg:
  46                 raise ExtractorError(error_msg, expected=True)
  47             token = xpath_text(auth, 'token')
  48             if not token:
  49                 return video_url
  50             self._AKAMAI_SPE_TOKEN_CACHE[secure_path] = token
  51         return video_url + '?hdnea=' + token
  52
  53     def _extract_cvp_info(self, data_src, video_id, path_data={}, ap_data={}, fatal=False):
  54         video_data = self._download_xml(
  55             data_src, video_id,
  56             transform_source=lambda s: fix_xml_ampersands(s).strip(),
  57             fatal=fatal)
  58         if not video_data:
  59             return {}
  60         video_id = video_data.attrib['id']
  61         title = xpath_text(video_data, 'headline', fatal=True)
  62         content_id = xpath_text(video_data, 'contentId') or video_id
  63         # rtmp_src = xpath_text(video_data, 'akamai/src')
  64         # if rtmp_src:
  65         #     split_rtmp_src = rtmp_src.split(',')
  66         #     if len(split_rtmp_src) == 2:
  67         #         rtmp_src = split_rtmp_src[1]
  68         # aifp = xpath_text(video_data, 'akamai/aifp', default='')
  69
  70         urls = []
  71         formats = []
  72         thumbnails = []
  73         subtitles = {}
  74         rex = re.compile(
  75             r'(?P<width>[0-9]+)x(?P<height>[0-9]+)(?:_(?P<bitrate>[0-9]+))?')
  76         # Possible formats locations: files/file, files/groupFiles/files
  77         # and maybe others
  78         for video_file in video_data.findall('.//file'):
  79             video_url = url_or_none(video_file.text.strip())
  80             if not video_url:
  81                 continue
  82             ext = determine_ext(video_url)
  83             if video_url.startswith('/mp4:protected/'):
  84                 continue
  85                 # TODO Correct extraction for these files
  86                 # protected_path_data = path_data.get('protected')
  87                 # if not protected_path_data or not rtmp_src:
  88                 #     continue
  89                 # protected_path = self._search_regex(
  90                 #     r'/mp4:(.+)\.[a-z0-9]', video_url, 'secure path')
  91                 # auth = self._download_webpage(
  92                 #     protected_path_data['tokenizer_src'], query={
  93                 #         'path': protected_path,
  94                 #         'videoId': content_id,
  95                 #         'aifp': aifp,
  96                 #     })
  97                 # token = xpath_text(auth, 'token')
  98                 # if not token:
  99                 #     continue
 100                 # video_url = rtmp_src + video_url + '?' + token
 101             elif video_url.startswith('/secure/'):
 102                 secure_path_data = path_data.get('secure')
 103                 if not secure_path_data:
 104                     continue
 105                 video_url = self._add_akamai_spe_token(
 106                     secure_path_data['tokenizer_src'],
 107                     secure_path_data['media_src'] + video_url,
 108                     content_id, ap_data)
 109             elif not re.match('https?://', video_url):
 110                 base_path_data = path_data.get(ext, path_data.get('default', {}))
 111                 media_src = base_path_data.get('media_src')
 112                 if not media_src:
 113                     continue
 114                 video_url = media_src + video_url
 115             if video_url in urls:
 116                 continue
 117             urls.append(video_url)
 118             format_id = video_file.get('bitrate')
 119             if ext in ('scc', 'srt', 'vtt'):
 120                 subtitles.setdefault('en', []).append({
 121                     'ext': ext,
 122                     'url': video_url,
 123                 })
 124             elif ext == 'png':
 125                 thumbnails.append({
 126                     'id': format_id,
 127                     'url': video_url,
 128                 })
 129             elif ext == 'smil':
 130                 formats.extend(self._extract_smil_formats(
 131                     video_url, video_id, fatal=False))
 132             elif re.match(r'https?://[^/]+\.akamaihd\.net/[iz]/', video_url):
 133                 formats.extend(self._extract_akamai_formats(
 134                     video_url, video_id, {
 135                         'hds': path_data.get('f4m', {}).get('host'),
 136                         # nba.cdn.turner.com, ht.cdn.turner.com, ht2.cdn.turner.com
 137                         # ht3.cdn.turner.com, i.cdn.turner.com, s.cdn.turner.com
 138                         # ssl.cdn.turner.com
 139                         'http': 'pmd.cdn.turner.com',
 140                     }))
 141             elif ext == 'm3u8':
 142                 m3u8_formats = self._extract_m3u8_formats(
 143                     video_url, video_id, 'mp4',
 144                     m3u8_id=format_id or 'hls', fatal=False)
 145                 if '/secure/' in video_url and '?hdnea=' in video_url:
 146                     for f in m3u8_formats:
 147                         f['_seekable'] = False
 148                 formats.extend(m3u8_formats)
 149             elif ext == 'f4m':
 150                 formats.extend(self._extract_f4m_formats(
 151                     update_url_query(video_url, {'hdcore': '3.7.0'}),
 152                     video_id, f4m_id=format_id or 'hds', fatal=False))
 153             else:
 154                 f = {
 155                     'format_id': format_id,
 156                     'url': video_url,
 157                     'ext': ext,
 158                 }
 159                 mobj = rex.search(video_url)
 160                 if mobj:
 161                     f.update({
 162                         'width': int(mobj.group('width')),
 163                         'height': int(mobj.group('height')),
 164                         'tbr': int_or_none(mobj.group('bitrate')),
 165                     })
 166                 elif isinstance(format_id, compat_str):
 167                     if format_id.isdigit():
 168                         f['tbr'] = int(format_id)
 169                     else:
 170                         mobj = re.match(r'ios_(audio|[0-9]+)$', format_id)
 171                         if mobj:
 172                             if mobj.group(1) == 'audio':
 173                                 f.update({
 174                                     'vcodec': 'none',
 175                                     'ext': 'm4a',
 176                                 })
 177                             else:
 178                                 f['tbr'] = int(mobj.group(1))
 179                 formats.append(f)
 180         self._sort_formats(formats)
 181
 182         for source in video_data.findall('closedCaptions/source'):
 183             for track in source.findall('track'):
 184                 track_url = url_or_none(track.get('url'))
 185                 if not track_url or track_url.endswith('/big'):
 186                     continue
 187                 lang = track.get('lang') or track.get('label') or 'en'
 188                 subtitles.setdefault(lang, []).append({
 189                     'url': track_url,
 190                     'ext': {
 191                         'scc': 'scc',
 192                         'webvtt': 'vtt',
 193                         'smptett': 'tt',
 194                     }.get(source.get('format'))
 195                 })
 196
 197         thumbnails.extend({
 198             'id': image.get('cut') or image.get('name'),
 199             'url': image.text,
 200             'width': int_or_none(image.get('width')),
 201             'height': int_or_none(image.get('height')),
 202         } for image in video_data.findall('images/image'))
 203
 204         is_live = xpath_text(video_data, 'isLive') == 'true'
 205
 206         return {
 207             'id': video_id,
 208             'title': self._live_title(title) if is_live else title,
 209             'formats': formats,
 210             'subtitles': subtitles,
 211             'thumbnails': thumbnails,
 212             'thumbnail': xpath_text(video_data, 'poster'),
 213             'description': strip_or_none(xpath_text(video_data, 'description')),
 214             'duration': parse_duration(xpath_text(video_data, 'length') or xpath_text(video_data, 'trt')),
 215             'timestamp': self._extract_timestamp(video_data),
 216             'upload_date': xpath_attr(video_data, 'metas', 'version'),
 217             'series': xpath_text(video_data, 'showTitle'),
 218             'season_number': int_or_none(xpath_text(video_data, 'seasonNumber')),
 219             'episode_number': int_or_none(xpath_text(video_data, 'episodeNumber')),
 220             'is_live': is_live,
 221         }
 222
 223     def _extract_ngtv_info(self, media_id, tokenizer_query, ap_data=None):
 224         streams_data = self._download_json(
 225             'http://medium.ngtv.io/media/%s/tv' % media_id,
 226             media_id)['media']['tv']
 227         duration = None
 228         chapters = []
 229         formats = []
 230         for supported_type in ('unprotected', 'bulkaes'):
 231             stream_data = streams_data.get(supported_type, {})
 232             m3u8_url = stream_data.get('secureUrl') or stream_data.get('url')
 233             if not m3u8_url:
 234                 continue
 235             if stream_data.get('playlistProtection') == 'spe':
 236                 m3u8_url = self._add_akamai_spe_token(
 237                     'http://token.ngtv.io/token/token_spe',
 238                     m3u8_url, media_id, ap_data or {}, tokenizer_query)
 239             formats.extend(self._extract_m3u8_formats(
 240                 m3u8_url, media_id, 'mp4', m3u8_id='hls', fatal=False))
 241
 242             duration = float_or_none(stream_data.get('totalRuntime'))
 243
 244             if not chapters:
 245                 for chapter in stream_data.get('contentSegments', []):
 246                     start_time = float_or_none(chapter.get('start'))
 247                     chapter_duration = float_or_none(chapter.get('duration'))
 248                     if start_time is None or chapter_duration is None:
 249                         continue
 250                     chapters.append({
 251                         'start_time': start_time,
 252                         'end_time': start_time + chapter_duration,
 253                     })
 254         self._sort_formats(formats)
 255
 256         return {
 257             'formats': formats,
 258             'chapters': chapters,
 259             'duration': duration,
 260         }