logo

youtube-dl

[mirror] Download/Watch videos from video hostersgit clone https://hacktivis.me/git/mirror/youtube-dl.git

fragment.py (11804B)

    

  1. from __future__ import division, unicode_literals
  3. import os
  4. import time
  5. import json
  7. from .common import FileDownloader
  8. from .http import HttpFD
  9. from ..utils import (
  10.     error_to_compat_str,
  11.     encodeFilename,
  12.     sanitize_open,
  13.     sanitized_Request,
  14. )
  17. class HttpQuietDownloader(HttpFD):
  18.     def to_screen(self, *args, **kargs):
  19.         pass
  22. class FragmentFD(FileDownloader):
  23.     """
  24.     A base file downloader class for fragmented media (e.g. f4m/m3u8 manifests).
  26.     Available options:
  28.     fragment_retries:   Number of times to retry a fragment for HTTP error (DASH
  29.                         and hlsnative only)
  30.     skip_unavailable_fragments:
  31.                         Skip unavailable fragments (DASH and hlsnative only)
  32.     keep_fragments:     Keep downloaded fragments on disk after downloading is
  33.                         finished
  35.     For each incomplete fragment download youtube-dl keeps on disk a special
  36.     bookkeeping file with download state and metadata (in future such files will
  37.     be used for any incomplete download handled by youtube-dl). This file is
  38.     used to properly handle resuming, check download file consistency and detect
  39.     potential errors. The file has a .ytdl extension and represents a standard
  40.     JSON file of the following format:
  42.     extractor:
  43.         Dictionary of extractor related data. TBD.
  45.     downloader:
  46.         Dictionary of downloader related data. May contain following data:
  47.             current_fragment:
  48.                 Dictionary with current (being downloaded) fragment data:
  49.                 index:  0-based index of current fragment among all fragments
  50.             fragment_count:
  51.                 Total count of fragments
  53.     This feature is experimental and file format may change in future.
  54.     """
  56.     def report_retry_fragment(self, err, frag_index, count, retries):
  57.         self.to_screen(
  58.             '[download] Got server HTTP error: %s. Retrying fragment %d (attempt %d of %s)...'
  59.             % (error_to_compat_str(err), frag_index, count, self.format_retries(retries)))
  61.     def report_skip_fragment(self, frag_index):
  62.         self.to_screen('[download] Skipping fragment %d...' % frag_index)
  64.     def _prepare_url(self, info_dict, url):
  65.         headers = info_dict.get('http_headers')
  66.         return sanitized_Request(url, None, headers) if headers else url
  68.     def _prepare_and_start_frag_download(self, ctx):
  69.         self._prepare_frag_download(ctx)
  70.         self._start_frag_download(ctx)
  72.     @staticmethod
  73.     def __do_ytdl_file(ctx):
  74.         return ctx['live'] is not True and ctx['tmpfilename'] != '-'
  76.     def _read_ytdl_file(self, ctx):
  77.         assert 'ytdl_corrupt' not in ctx
  78.         stream, _ = sanitize_open(self.ytdl_filename(ctx['filename']), 'r')
  79.         try:
  80.             ctx['fragment_index'] = json.loads(stream.read())['downloader']['current_fragment']['index']
  81.         except Exception:
  82.             ctx['ytdl_corrupt'] = True
  83.         finally:
  84.             stream.close()
  86.     def _write_ytdl_file(self, ctx):
  87.         frag_index_stream, _ = sanitize_open(self.ytdl_filename(ctx['filename']), 'w')
  88.         downloader = {
  89.             'current_fragment': {
  90.                 'index': ctx['fragment_index'],
  91.             },
  92.         }
  93.         if ctx.get('fragment_count') is not None:
  94.             downloader['fragment_count'] = ctx['fragment_count']
  95.         frag_index_stream.write(json.dumps({'downloader': downloader}))
  96.         frag_index_stream.close()
  98.     def _download_fragment(self, ctx, frag_url, info_dict, headers=None):
  99.         fragment_filename = '%s-Frag%d' % (ctx['tmpfilename'], ctx['fragment_index'])
  100.         fragment_info_dict = {
  101.             'url': frag_url,
  102.             'http_headers': headers or info_dict.get('http_headers'),
  103.         }
  104.         frag_resume_len = 0
  105.         if ctx['dl'].params.get('continuedl', True):
  106.             frag_resume_len = self.filesize_or_none(
  107.                 self.temp_name(fragment_filename))
  108.         fragment_info_dict['frag_resume_len'] = frag_resume_len
  109.         ctx['frag_resume_len'] = frag_resume_len or 0
  111.         success = ctx['dl'].download(fragment_filename, fragment_info_dict)
  112.         if not success:
  113.             return False, None
  114.         if fragment_info_dict.get('filetime'):
  115.             ctx['fragment_filetime'] = fragment_info_dict.get('filetime')
  116.         down, frag_sanitized = sanitize_open(fragment_filename, 'rb')
  117.         ctx['fragment_filename_sanitized'] = frag_sanitized
  118.         frag_content = down.read()
  119.         down.close()
  120.         return True, frag_content
  122.     def _append_fragment(self, ctx, frag_content):
  123.         try:
  124.             ctx['dest_stream'].write(frag_content)
  125.             ctx['dest_stream'].flush()
  126.         finally:
  127.             if self.__do_ytdl_file(ctx):
  128.                 self._write_ytdl_file(ctx)
  129.             if not self.params.get('keep_fragments', False):
  130.                 os.remove(encodeFilename(ctx['fragment_filename_sanitized']))
  131.             del ctx['fragment_filename_sanitized']
  133.     def _prepare_frag_download(self, ctx):
  134.         if not ctx.setdefault('live', False):
  135.             total_frags_str = '%d' % ctx['total_frags']
  136.             ad_frags = ctx.get('ad_frags', 0)
  137.             if ad_frags:
  138.                 total_frags_str += ' (not including %d ad)' % ad_frags
  139.         else:
  140.             total_frags_str = 'unknown (live)'
  141.         self.to_screen(
  142.             '[%s] Total fragments: %s' % (self.FD_NAME, total_frags_str))
  143.         self.report_destination(ctx['filename'])
  144.         continuedl = self.params.get('continuedl', True)
  145.         dl = HttpQuietDownloader(
  146.             self.ydl,
  147.             {
  148.                 'continuedl': continuedl,
  149.                 'quiet': True,
  150.                 'noprogress': True,
  151.                 'ratelimit': self.params.get('ratelimit'),
  152.                 'retries': self.params.get('retries', 0),
  153.                 'nopart': self.params.get('nopart', False),
  154.                 'test': self.params.get('test', False),
  155.             }
  156.         )
  157.         tmpfilename = self.temp_name(ctx['filename'])
  158.         open_mode = 'wb'
  160.         # Establish possible resume length
  161.         resume_len = self.filesize_or_none(tmpfilename) or 0
  162.         if resume_len > 0:
  163.             open_mode = 'ab'
  165.         # Should be initialized before ytdl file check
  166.         ctx.update({
  167.             'tmpfilename': tmpfilename,
  168.             'fragment_index': 0,
  169.         })
  171.         if self.__do_ytdl_file(ctx):
  172.             ytdl_file_exists = os.path.isfile(encodeFilename(self.ytdl_filename(ctx['filename'])))
  173.             if continuedl and ytdl_file_exists:
  174.                 self._read_ytdl_file(ctx)
  175.                 is_corrupt = ctx.get('ytdl_corrupt') is True
  176.                 is_inconsistent = ctx['fragment_index'] > 0 and resume_len == 0
  177.                 if is_corrupt or is_inconsistent:
  178.                     message = (
  179.                         '.ytdl file is corrupt' if is_corrupt else
  180.                         'Inconsistent state of incomplete fragment download')
  181.                     self.report_warning(
  182.                         '%s. Restarting from the beginning...' % message)
  183.                     ctx['fragment_index'] = resume_len = 0
  184.                     if 'ytdl_corrupt' in ctx:
  185.                         del ctx['ytdl_corrupt']
  186.                     self._write_ytdl_file(ctx)
  188.             else:
  189.                 if not continuedl:
  190.                     if ytdl_file_exists:
  191.                         self._read_ytdl_file(ctx)
  192.                     ctx['fragment_index'] = resume_len = 0
  193.                 self._write_ytdl_file(ctx)
  194.                 assert ctx['fragment_index'] == 0
  196.         dest_stream, tmpfilename = sanitize_open(tmpfilename, open_mode)
  198.         ctx.update({
  199.             'dl': dl,
  200.             'dest_stream': dest_stream,
  201.             'tmpfilename': tmpfilename,
  202.             # Total complete fragments downloaded so far in bytes
  203.             'complete_frags_downloaded_bytes': resume_len,
  204.         })
  206.     def _start_frag_download(self, ctx):
  207.         resume_len = ctx['complete_frags_downloaded_bytes']
  208.         total_frags = ctx['total_frags']
  209.         # This dict stores the download progress, it's updated by the progress
  210.         # hook
  211.         state = {
  212.             'status': 'downloading',
  213.             'downloaded_bytes': resume_len,
  214.             'fragment_index': ctx['fragment_index'],
  215.             'fragment_count': total_frags,
  216.             'filename': ctx['filename'],
  217.             'tmpfilename': ctx['tmpfilename'],
  218.         }
  220.         start = time.time()
  221.         ctx.update({
  222.             'started': start,
  223.             'fragment_started': start,
  224.             # Amount of fragment's bytes downloaded by the time of the previous
  225.             # frag progress hook invocation
  226.             'prev_frag_downloaded_bytes': 0,
  227.         })
  229.         def frag_progress_hook(s):
  230.             if s['status'] not in ('downloading', 'finished'):
  231.                 return
  233.             if not total_frags and ctx.get('fragment_count'):
  234.                 state['fragment_count'] = ctx['fragment_count']
  236.             time_now = time.time()
  237.             state['elapsed'] = time_now - start
  238.             frag_total_bytes = s.get('total_bytes') or 0
  239.             if not ctx['live']:
  240.                 estimated_size = (
  241.                     (ctx['complete_frags_downloaded_bytes'] + frag_total_bytes)
  242.                     / (state['fragment_index'] + 1) * total_frags)
  243.                 state['total_bytes_estimate'] = estimated_size
  245.             if s['status'] == 'finished':
  246.                 state['fragment_index'] += 1
  247.                 ctx['fragment_index'] = state['fragment_index']
  248.                 state['downloaded_bytes'] += frag_total_bytes - ctx['prev_frag_downloaded_bytes']
  249.                 ctx['complete_frags_downloaded_bytes'] = state['downloaded_bytes']
  250.                 ctx['speed'] = state['speed'] = self.calc_speed(
  251.                     ctx['fragment_started'], time_now, frag_total_bytes)
  252.                 ctx['fragment_started'] = time.time()
  253.                 ctx['prev_frag_downloaded_bytes'] = 0
  254.             else:
  255.                 frag_downloaded_bytes = s['downloaded_bytes']
  256.                 state['downloaded_bytes'] += frag_downloaded_bytes - ctx['prev_frag_downloaded_bytes']
  257.                 ctx['speed'] = state['speed'] = self.calc_speed(
  258.                     ctx['fragment_started'], time_now, frag_downloaded_bytes - ctx['frag_resume_len'])
  259.                 if not ctx['live']:
  260.                     state['eta'] = self.calc_eta(state['speed'], estimated_size - state['downloaded_bytes'])
  261.                 ctx['prev_frag_downloaded_bytes'] = frag_downloaded_bytes
  262.             self._hook_progress(state)
  264.         ctx['dl'].add_progress_hook(frag_progress_hook)
  266.         return start
  268.     def _finish_frag_download(self, ctx):
  269.         ctx['dest_stream'].close()
  270.         if self.__do_ytdl_file(ctx):
  271.             ytdl_filename = encodeFilename(self.ytdl_filename(ctx['filename']))
  272.             if os.path.isfile(ytdl_filename):
  273.                 os.remove(ytdl_filename)
  274.         elapsed = time.time() - ctx['started']
  276.         if ctx['tmpfilename'] == '-':
  277.             downloaded_bytes = ctx['complete_frags_downloaded_bytes']
  278.         else:
  279.             self.try_rename(ctx['tmpfilename'], ctx['filename'])
  280.             if self.params.get('updatetime', True):
  281.                 filetime = ctx.get('fragment_filetime')
  282.                 if filetime:
  283.                     try:
  284.                         os.utime(ctx['filename'], (time.time(), filetime))
  285.                     except Exception:
  286.                         pass
  287.             downloaded_bytes = self.filesize_or_none(ctx['filename']) or 0
  289.         self._hook_progress({
  290.             'downloaded_bytes': downloaded_bytes,
  291.             'total_bytes': downloaded_bytes,
  292.             'filename': ctx['filename'],
  293.             'status': 'finished',
  294.             'elapsed': elapsed,
  295.         })