yt_dlp/extractor/dispeak.py

   1 import re
   2
   3 from .common import InfoExtractor
   4 from ..utils import (
   5     int_or_none,
   6     parse_duration,
   7     remove_end,
   8     xpath_element,
   9     xpath_text,
  10 )
  11
  12
  13 class DigitallySpeakingIE(InfoExtractor):
  14     _VALID_URL = r'https?://(?:s?evt\.dispeak|events\.digitallyspeaking)\.com/(?:[^/]+/)+xml/(?P<id>[^.]+)\.xml'
  15
  16     _TESTS = [{
  17         # From http://gdcvault.com/play/1023460/Tenacious-Design-and-The-Interface
  18         'url': 'http://evt.dispeak.com/ubm/gdc/sf16/xml/840376_BQRC.xml',
  19         'md5': 'a8efb6c31ed06ca8739294960b2dbabd',
  20         'info_dict': {
  21             'id': '840376_BQRC',
  22             'ext': 'mp4',
  23             'title': 'Tenacious Design and The Interface of \'Destiny\'',
  24         },
  25     }, {
  26         # From http://www.gdcvault.com/play/1014631/Classic-Game-Postmortem-PAC
  27         'url': 'http://events.digitallyspeaking.com/gdc/sf11/xml/12396_1299111843500GMPX.xml',
  28         'only_matching': True,
  29     }, {
  30         # From http://www.gdcvault.com/play/1013700/Advanced-Material
  31         'url': 'http://sevt.dispeak.com/ubm/gdc/eur10/xml/11256_1282118587281VNIT.xml',
  32         'only_matching': True,
  33     }, {
  34         # From https://gdcvault.com/play/1016624, empty speakerVideo
  35         'url': 'https://sevt.dispeak.com/ubm/gdc/online12/xml/201210-822101_1349794556671DDDD.xml',
  36         'info_dict': {
  37             'id': '201210-822101_1349794556671DDDD',
  38             'ext': 'flv',
  39             'title': 'Pre-launch - Preparing to Take the Plunge',
  40         },
  41     }, {
  42         # From http://www.gdcvault.com/play/1014846/Conference-Keynote-Shigeru, empty slideVideo
  43         'url': 'http://events.digitallyspeaking.com/gdc/project25/xml/p25-miyamoto1999_1282467389849HSVB.xml',
  44         'only_matching': True,
  45     }]
  46
  47     def _parse_mp4(self, metadata):
  48         video_formats = []
  49         video_root = None
  50
  51         mp4_video = xpath_text(metadata, './mp4video', default=None)
  52         if mp4_video is not None:
  53             mobj = re.match(r'(?P<root>https?://.*?/).*', mp4_video)
  54             video_root = mobj.group('root')
  55         if video_root is None:
  56             http_host = xpath_text(metadata, 'httpHost', default=None)
  57             if http_host:
  58                 video_root = f'http://{http_host}/'
  59         if video_root is None:
  60             # Hard-coded in http://evt.dispeak.com/ubm/gdc/sf16/custom/player2.js
  61             # Works for GPUTechConf, too
  62             video_root = 'http://s3-2u.digitallyspeaking.com/'
  63
  64         formats = metadata.findall('./MBRVideos/MBRVideo')
  65         if not formats:
  66             return None
  67         for a_format in formats:
  68             stream_name = xpath_text(a_format, 'streamName', fatal=True)
  69             video_path = re.match(r'mp4\:(?P<path>.*)', stream_name).group('path')
  70             url = video_root + video_path
  71             bitrate = xpath_text(a_format, 'bitrate')
  72             tbr = int_or_none(bitrate)
  73             vbr = int_or_none(self._search_regex(
  74                 r'-(\d+)\.mp4', video_path, 'vbr', default=None))
  75             video_formats.append({
  76                 'format_id': bitrate,
  77                 'url': url,
  78                 'tbr': tbr,
  79                 'vbr': vbr,
  80             })
  81         return video_formats
  82
  83     def _parse_flv(self, metadata):
  84         formats = []
  85         akamai_url = xpath_text(metadata, './akamaiHost', fatal=True)
  86         audios = metadata.findall('./audios/audio')
  87         for audio in audios:
  88             formats.append({
  89                 'url': f'rtmp://{akamai_url}/ondemand?ovpfv=1.1',
  90                 'play_path': remove_end(audio.get('url'), '.flv'),
  91                 'ext': 'flv',
  92                 'vcodec': 'none',
  93                 'quality': 1,
  94                 'format_id': audio.get('code'),
  95             })
  96         for video_key, format_id, preference in (
  97                 ('slide', 'slides', -2), ('speaker', 'speaker', -1)):
  98             video_path = xpath_text(metadata, f'./{video_key}Video')
  99             if not video_path:
 100                 continue
 101             formats.append({
 102                 'url': f'rtmp://{akamai_url}/ondemand?ovpfv=1.1',
 103                 'play_path': remove_end(video_path, '.flv'),
 104                 'ext': 'flv',
 105                 'format_note': f'{video_key} video',
 106                 'quality': preference,
 107                 'format_id': format_id,
 108             })
 109         return formats
 110
 111     def _real_extract(self, url):
 112         video_id = self._match_id(url)
 113
 114         xml_description = self._download_xml(url, video_id)
 115         metadata = xpath_element(xml_description, 'metadata')
 116
 117         video_formats = self._parse_mp4(metadata)
 118         if video_formats is None:
 119             video_formats = self._parse_flv(metadata)
 120
 121         return {
 122             'id': video_id,
 123             'formats': video_formats,
 124             'title': xpath_text(metadata, 'title', fatal=True),
 125             'duration': parse_duration(xpath_text(metadata, 'endTime')),
 126             'creator': xpath_text(metadata, 'speaker'),
 127         }