_ Git - youtube-dl/blob - youtube_dl/downloader/fragment.py

   1 from __future__ import division, unicode_literals
   2
   3 import os
   4 import time
   5 import io
   6 import json
   7
   8 from .common import FileDownloader
   9 from .http import HttpFD
  10 from ..utils import (
  11     error_to_compat_str,
  12     encodeFilename,
  13     sanitize_open,
  14     sanitized_Request,
  15     compat_str,
  16 )
  17
  18
  19 class HttpQuietDownloader(HttpFD):
  20     def to_screen(self, *args, **kargs):
  21         pass
  22
  23
  24 class FragmentFD(FileDownloader):
  25     """
  26     A base file downloader class for fragmented media (e.g. f4m/m3u8 manifests).
  27
  28     Available options:
  29
  30     fragment_retries:   Number of times to retry a fragment for HTTP error (DASH
  31                         and hlsnative only)
  32     skip_unavailable_fragments:
  33                         Skip unavailable fragments (DASH and hlsnative only)
  34     """
  35
  36     def report_retry_fragment(self, err, frag_index, count, retries):
  37         self.to_screen(
  38             '[download] Got server HTTP error: %s. Retrying fragment %d (attempt %d of %s)...'
  39             % (error_to_compat_str(err), frag_index, count, self.format_retries(retries)))
  40
  41     def report_skip_fragment(self, frag_index):
  42         self.to_screen('[download] Skipping fragment %d...' % frag_index)
  43
  44     def _prepare_url(self, info_dict, url):
  45         headers = info_dict.get('http_headers')
  46         return sanitized_Request(url, None, headers) if headers else url
  47
  48     def _prepare_and_start_frag_download(self, ctx):
  49         self._prepare_frag_download(ctx)
  50         self._start_frag_download(ctx)
  51
  52     def _download_fragment(self, ctx, frag_url, info_dict, headers=None):
  53         down = io.BytesIO()
  54         success = ctx['dl'].download(down, {
  55             'url': frag_url,
  56             'http_headers': headers or info_dict.get('http_headers'),
  57         })
  58         if not success:
  59             return False, None
  60         frag_content = down.getvalue()
  61         down.close()
  62         return True, frag_content
  63
  64     def _append_fragment(self, ctx, frag_content):
  65         ctx['dest_stream'].write(frag_content)
  66         if not (ctx.get('live') or ctx['tmpfilename'] == '-'):
  67             frag_index_stream, _ = sanitize_open(self.ytdl_filename(ctx['filename']), 'w')
  68             frag_index_stream.write(json.dumps({
  69                 'frag_index': ctx['frag_index']
  70             }))
  71             frag_index_stream.close()
  72
  73     def _prepare_frag_download(self, ctx):
  74         if 'live' not in ctx:
  75             ctx['live'] = False
  76         self.to_screen(
  77             '[%s] Total fragments: %s'
  78             % (self.FD_NAME, ctx['total_frags'] if not ctx['live'] else 'unknown (live)'))
  79         self.report_destination(ctx['filename'])
  80         dl = HttpQuietDownloader(
  81             self.ydl,
  82             {
  83                 'continuedl': True,
  84                 'quiet': True,
  85                 'noprogress': True,
  86                 'ratelimit': self.params.get('ratelimit'),
  87                 'retries': self.params.get('retries', 0),
  88                 'nopart': self.params.get('nopart', False),
  89                 'test': self.params.get('test', False),
  90             }
  91         )
  92         tmpfilename = self.temp_name(ctx['filename'])
  93         open_mode = 'wb'
  94         resume_len = 0
  95         frag_index = 0
  96         # Establish possible resume length
  97         if os.path.isfile(encodeFilename(tmpfilename)):
  98             open_mode = 'ab'
  99             resume_len = os.path.getsize(encodeFilename(tmpfilename))
 100             ytdl_filename = encodeFilename(self.ytdl_filename(ctx['filename']))
 101             if os.path.isfile(ytdl_filename):
 102                 frag_index_stream, _ = sanitize_open(ytdl_filename, 'r')
 103                 frag_index = json.loads(frag_index_stream.read())['frag_index']
 104                 frag_index_stream.close()
 105         dest_stream, tmpfilename = sanitize_open(tmpfilename, open_mode)
 106
 107         ctx.update({
 108             'dl': dl,
 109             'dest_stream': dest_stream,
 110             'tmpfilename': tmpfilename,
 111             'frag_index': frag_index,
 112             # Total complete fragments downloaded so far in bytes
 113             'complete_frags_downloaded_bytes': resume_len,
 114         })
 115
 116     def _start_frag_download(self, ctx):
 117         total_frags = ctx['total_frags']
 118         # This dict stores the download progress, it's updated by the progress
 119         # hook
 120         state = {
 121             'status': 'downloading',
 122             'downloaded_bytes': ctx['complete_frags_downloaded_bytes'],
 123             'frag_index': ctx['frag_index'],
 124             'frag_count': total_frags,
 125             'filename': ctx['filename'],
 126             'tmpfilename': ctx['tmpfilename'],
 127         }
 128
 129         start = time.time()
 130         ctx.update({
 131             'started': start,
 132             # Amount of fragment's bytes downloaded by the time of the previous
 133             # frag progress hook invocation
 134             'prev_frag_downloaded_bytes': 0,
 135         })
 136
 137         def frag_progress_hook(s):
 138             if s['status'] not in ('downloading', 'finished'):
 139                 return
 140
 141             time_now = time.time()
 142             state['elapsed'] = time_now - start
 143             frag_total_bytes = s.get('total_bytes') or 0
 144             if not ctx['live']:
 145                 estimated_size = (
 146                     (ctx['complete_frags_downloaded_bytes'] + frag_total_bytes) /
 147                     (state['frag_index'] + 1) * total_frags)
 148                 state['total_bytes_estimate'] = estimated_size
 149
 150             if s['status'] == 'finished':
 151                 state['frag_index'] += 1
 152                 ctx['frag_index'] = state['frag_index']
 153                 state['downloaded_bytes'] += frag_total_bytes - ctx['prev_frag_downloaded_bytes']
 154                 ctx['complete_frags_downloaded_bytes'] = state['downloaded_bytes']
 155                 ctx['prev_frag_downloaded_bytes'] = 0
 156             else:
 157                 frag_downloaded_bytes = s['downloaded_bytes']
 158                 state['downloaded_bytes'] += frag_downloaded_bytes - ctx['prev_frag_downloaded_bytes']
 159                 if not ctx['live']:
 160                     state['eta'] = self.calc_eta(
 161                         start, time_now, estimated_size,
 162                         state['downloaded_bytes'])
 163                 state['speed'] = s.get('speed') or ctx.get('speed')
 164                 ctx['speed'] = state['speed']
 165                 ctx['prev_frag_downloaded_bytes'] = frag_downloaded_bytes
 166             self._hook_progress(state)
 167
 168         ctx['dl'].add_progress_hook(frag_progress_hook)
 169
 170         return start
 171
 172     def _finish_frag_download(self, ctx):
 173         ctx['dest_stream'].close()
 174         ytdl_filename = encodeFilename(self.ytdl_filename(ctx['filename']))
 175         if os.path.isfile(ytdl_filename):
 176             os.remove(ytdl_filename)
 177         elapsed = time.time() - ctx['started']
 178         self.try_rename(ctx['tmpfilename'], ctx['filename'])
 179         fsize = os.path.getsize(encodeFilename(ctx['filename']))
 180
 181         self._hook_progress({
 182             'downloaded_bytes': fsize,
 183             'total_bytes': fsize,
 184             'filename': ctx['filename'],
 185             'status': 'finished',
 186             'elapsed': elapsed,
 187         })