youtube_dl/downloader/hls.py

   1 from __future__ import unicode_literals
   2
   3 import os.path
   4 import re
   5
   6 from .fragment import FragmentFD
   7 from .external import FFmpegFD
   8
   9 from ..compat import compat_urlparse
  10 from ..utils import (
  11     encodeFilename,
  12     sanitize_open,
  13 )
  14
  15
  16 class HlsFD(FragmentFD):
  17     """ A limited implementation that does not require ffmpeg """
  18
  19     FD_NAME = 'hlsnative'
  20
  21     @staticmethod
  22     def can_download(manifest):
  23         UNSUPPORTED_FEATURES = (
  24             r'#EXT-X-KEY:METHOD=(?!NONE)',  # encrypted streams [1]
  25             r'#EXT-X-BYTERANGE',  # playlists composed of byte ranges of media files [2]
  26             # Live streams heuristic does not always work (e.g. geo restricted to Germany
  27             # http://hls-geo.daserste.de/i/videoportal/Film/c_620000/622873/format,716451,716457,716450,716458,716459,.mp4.csmil/index_4_av.m3u8?null=0)
  28             #r'#EXT-X-MEDIA-SEQUENCE:(?!0$)',  # live streams [3]
  29             # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.4
  30             # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.2
  31             # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.2
  32         )
  33         return all(not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES)
  34
  35     def real_download(self, filename, info_dict):
  36         man_url = info_dict['url']
  37         self.to_screen('[%s] Downloading m3u8 manifest' % self.FD_NAME)
  38         manifest = self.ydl.urlopen(man_url).read()
  39
  40         s = manifest.decode('utf-8', 'ignore')
  41
  42         if not self.can_download(s):
  43             self.report_warning(
  44                 'hlsnative has detected features it does not support, '
  45                 'extraction will be delegated to ffmpeg')
  46             fd = FFmpegFD(self.ydl, self.params)
  47             for ph in self._progress_hooks:
  48                 fd.add_progress_hook(ph)
  49             return fd.real_download(filename, info_dict)
  50
  51         fragment_urls = []
  52         for line in s.splitlines():
  53             line = line.strip()
  54             if line and not line.startswith('#'):
  55                 segment_url = (
  56                     line
  57                     if re.match(r'^https?://', line)
  58                     else compat_urlparse.urljoin(man_url, line))
  59                 fragment_urls.append(segment_url)
  60                 # We only download the first fragment during the test
  61                 if self.params.get('test', False):
  62                     break
  63
  64         ctx = {
  65             'filename': filename,
  66             'total_frags': len(fragment_urls),
  67         }
  68
  69         self._prepare_and_start_frag_download(ctx)
  70
  71         frags_filenames = []
  72         for i, frag_url in enumerate(fragment_urls):
  73             frag_filename = '%s-Frag%d' % (ctx['tmpfilename'], i)
  74             success = ctx['dl'].download(frag_filename, {'url': frag_url})
  75             if not success:
  76                 return False
  77             down, frag_sanitized = sanitize_open(frag_filename, 'rb')
  78             ctx['dest_stream'].write(down.read())
  79             down.close()
  80             frags_filenames.append(frag_sanitized)
  81
  82         self._finish_frag_download(ctx)
  83
  84         for frag_file in frags_filenames:
  85             os.remove(encodeFilename(frag_file))
  86
  87         return True