]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/downloader/fragment.py
debian/control: Update list of supported sites.
[youtubedl] / youtube_dl / downloader / fragment.py
index 5a64b29eefd9bf7a35f6220efc25bcbd999ce33c..02f35459e82ddb0e39dad3b7790f2db7a90d85fa 100644 (file)
@@ -2,12 +2,15 @@ from __future__ import division, unicode_literals
 
 import os
 import time
+import json
 
 from .common import FileDownloader
 from .http import HttpFD
 from ..utils import (
+    error_to_compat_str,
     encodeFilename,
     sanitize_open,
+    sanitized_Request,
 )
 
 
@@ -19,14 +22,116 @@ class HttpQuietDownloader(HttpFD):
 class FragmentFD(FileDownloader):
     """
     A base file downloader class for fragmented media (e.g. f4m/m3u8 manifests).
+
+    Available options:
+
+    fragment_retries:   Number of times to retry a fragment for HTTP error (DASH
+                        and hlsnative only)
+    skip_unavailable_fragments:
+                        Skip unavailable fragments (DASH and hlsnative only)
+    keep_fragments:     Keep downloaded fragments on disk after downloading is
+                        finished
+
+    For each incomplete fragment download youtube-dl keeps on disk a special
+    bookkeeping file with download state and metadata (in future such files will
+    be used for any incomplete download handled by youtube-dl). This file is
+    used to properly handle resuming, check download file consistency and detect
+    potential errors. The file has a .ytdl extension and represents a standard
+    JSON file of the following format:
+
+    extractor:
+        Dictionary of extractor related data. TBD.
+
+    downloader:
+        Dictionary of downloader related data. May contain following data:
+            current_fragment:
+                Dictionary with current (being downloaded) fragment data:
+                index:  0-based index of current fragment among all fragments
+            fragment_count:
+                Total count of fragments
+
+    This feature is experimental and file format may change in future.
     """
 
+    def report_retry_fragment(self, err, frag_index, count, retries):
+        self.to_screen(
+            '[download] Got server HTTP error: %s. Retrying fragment %d (attempt %d of %s)...'
+            % (error_to_compat_str(err), frag_index, count, self.format_retries(retries)))
+
+    def report_skip_fragment(self, frag_index):
+        self.to_screen('[download] Skipping fragment %d...' % frag_index)
+
+    def _prepare_url(self, info_dict, url):
+        headers = info_dict.get('http_headers')
+        return sanitized_Request(url, None, headers) if headers else url
+
     def _prepare_and_start_frag_download(self, ctx):
         self._prepare_frag_download(ctx)
         self._start_frag_download(ctx)
 
+    @staticmethod
+    def __do_ytdl_file(ctx):
+        return not ctx['live'] and not ctx['tmpfilename'] == '-'
+
+    def _read_ytdl_file(self, ctx):
+        assert 'ytdl_corrupt' not in ctx
+        stream, _ = sanitize_open(self.ytdl_filename(ctx['filename']), 'r')
+        try:
+            ctx['fragment_index'] = json.loads(stream.read())['downloader']['current_fragment']['index']
+        except Exception:
+            ctx['ytdl_corrupt'] = True
+        finally:
+            stream.close()
+
+    def _write_ytdl_file(self, ctx):
+        frag_index_stream, _ = sanitize_open(self.ytdl_filename(ctx['filename']), 'w')
+        downloader = {
+            'current_fragment': {
+                'index': ctx['fragment_index'],
+            },
+        }
+        if ctx.get('fragment_count') is not None:
+            downloader['fragment_count'] = ctx['fragment_count']
+        frag_index_stream.write(json.dumps({'downloader': downloader}))
+        frag_index_stream.close()
+
+    def _download_fragment(self, ctx, frag_url, info_dict, headers=None):
+        fragment_filename = '%s-Frag%d' % (ctx['tmpfilename'], ctx['fragment_index'])
+        success = ctx['dl'].download(fragment_filename, {
+            'url': frag_url,
+            'http_headers': headers or info_dict.get('http_headers'),
+        })
+        if not success:
+            return False, None
+        down, frag_sanitized = sanitize_open(fragment_filename, 'rb')
+        ctx['fragment_filename_sanitized'] = frag_sanitized
+        frag_content = down.read()
+        down.close()
+        return True, frag_content
+
+    def _append_fragment(self, ctx, frag_content):
+        try:
+            ctx['dest_stream'].write(frag_content)
+            ctx['dest_stream'].flush()
+        finally:
+            if self.__do_ytdl_file(ctx):
+                self._write_ytdl_file(ctx)
+            if not self.params.get('keep_fragments', False):
+                os.remove(encodeFilename(ctx['fragment_filename_sanitized']))
+            del ctx['fragment_filename_sanitized']
+
     def _prepare_frag_download(self, ctx):
-        self.to_screen('[%s] Total fragments: %d' % (self.FD_NAME, ctx['total_frags']))
+        if 'live' not in ctx:
+            ctx['live'] = False
+        if not ctx['live']:
+            total_frags_str = '%d' % ctx['total_frags']
+            ad_frags = ctx.get('ad_frags', 0)
+            if ad_frags:
+                total_frags_str += ' (not including %d ad)' % ad_frags
+        else:
+            total_frags_str = 'unknown (live)'
+        self.to_screen(
+            '[%s] Total fragments: %s' % (self.FD_NAME, total_frags_str))
         self.report_destination(ctx['filename'])
         dl = HttpQuietDownloader(
             self.ydl,
@@ -34,62 +139,107 @@ class FragmentFD(FileDownloader):
                 'continuedl': True,
                 'quiet': True,
                 'noprogress': True,
-                'ratelimit': self.params.get('ratelimit', None),
+                'ratelimit': self.params.get('ratelimit'),
                 'retries': self.params.get('retries', 0),
+                'nopart': self.params.get('nopart', False),
                 'test': self.params.get('test', False),
             }
         )
         tmpfilename = self.temp_name(ctx['filename'])
-        dest_stream, tmpfilename = sanitize_open(tmpfilename, 'wb')
+        open_mode = 'wb'
+        resume_len = 0
+
+        # Establish possible resume length
+        if os.path.isfile(encodeFilename(tmpfilename)):
+            open_mode = 'ab'
+            resume_len = os.path.getsize(encodeFilename(tmpfilename))
+
+        # Should be initialized before ytdl file check
+        ctx.update({
+            'tmpfilename': tmpfilename,
+            'fragment_index': 0,
+        })
+
+        if self.__do_ytdl_file(ctx):
+            if os.path.isfile(encodeFilename(self.ytdl_filename(ctx['filename']))):
+                self._read_ytdl_file(ctx)
+                is_corrupt = ctx.get('ytdl_corrupt') is True
+                is_inconsistent = ctx['fragment_index'] > 0 and resume_len == 0
+                if is_corrupt or is_inconsistent:
+                    message = (
+                        '.ytdl file is corrupt' if is_corrupt else
+                        'Inconsistent state of incomplete fragment download')
+                    self.report_warning(
+                        '%s. Restarting from the beginning...' % message)
+                    ctx['fragment_index'] = resume_len = 0
+                    if 'ytdl_corrupt' in ctx:
+                        del ctx['ytdl_corrupt']
+                    self._write_ytdl_file(ctx)
+            else:
+                self._write_ytdl_file(ctx)
+                assert ctx['fragment_index'] == 0
+
+        dest_stream, tmpfilename = sanitize_open(tmpfilename, open_mode)
+
         ctx.update({
             'dl': dl,
             'dest_stream': dest_stream,
             'tmpfilename': tmpfilename,
+            # Total complete fragments downloaded so far in bytes
+            'complete_frags_downloaded_bytes': resume_len,
         })
 
     def _start_frag_download(self, ctx):
+        resume_len = ctx['complete_frags_downloaded_bytes']
         total_frags = ctx['total_frags']
         # This dict stores the download progress, it's updated by the progress
         # hook
         state = {
             'status': 'downloading',
-            'downloaded_bytes': 0,
-            'frag_index': 0,
-            'frag_count': total_frags,
+            'downloaded_bytes': resume_len,
+            'fragment_index': ctx['fragment_index'],
+            'fragment_count': total_frags,
             'filename': ctx['filename'],
             'tmpfilename': ctx['tmpfilename'],
         }
+
         start = time.time()
-        ctx['started'] = start
+        ctx.update({
+            'started': start,
+            # Amount of fragment's bytes downloaded by the time of the previous
+            # frag progress hook invocation
+            'prev_frag_downloaded_bytes': 0,
+        })
 
         def frag_progress_hook(s):
             if s['status'] not in ('downloading', 'finished'):
                 return
 
-            frag_total_bytes = s.get('total_bytes', 0)
-            if s['status'] == 'finished':
-                state['downloaded_bytes'] += frag_total_bytes
-                state['frag_index'] += 1
-
-            estimated_size = (
-                (state['downloaded_bytes'] + frag_total_bytes) /
-                (state['frag_index'] + 1) * total_frags)
             time_now = time.time()
-            state['total_bytes_estimate'] = estimated_size
             state['elapsed'] = time_now - start
+            frag_total_bytes = s.get('total_bytes') or 0
+            if not ctx['live']:
+                estimated_size = (
+                    (ctx['complete_frags_downloaded_bytes'] + frag_total_bytes)
+                    / (state['fragment_index'] + 1) * total_frags)
+                state['total_bytes_estimate'] = estimated_size
 
             if s['status'] == 'finished':
-                progress = self.calc_percent(state['frag_index'], total_frags)
+                state['fragment_index'] += 1
+                ctx['fragment_index'] = state['fragment_index']
+                state['downloaded_bytes'] += frag_total_bytes - ctx['prev_frag_downloaded_bytes']
+                ctx['complete_frags_downloaded_bytes'] = state['downloaded_bytes']
+                ctx['prev_frag_downloaded_bytes'] = 0
             else:
                 frag_downloaded_bytes = s['downloaded_bytes']
-                frag_progress = self.calc_percent(frag_downloaded_bytes,
-                                                  frag_total_bytes)
-                progress = self.calc_percent(state['frag_index'], total_frags)
-                progress += frag_progress / float(total_frags)
-
-                state['eta'] = self.calc_eta(
-                    start, time_now, estimated_size, state['downloaded_bytes'] + frag_downloaded_bytes)
-                state['speed'] = s.get('speed')
+                state['downloaded_bytes'] += frag_downloaded_bytes - ctx['prev_frag_downloaded_bytes']
+                if not ctx['live']:
+                    state['eta'] = self.calc_eta(
+                        start, time_now, estimated_size - resume_len,
+                        state['downloaded_bytes'] - resume_len)
+                state['speed'] = s.get('speed') or ctx.get('speed')
+                ctx['speed'] = state['speed']
+                ctx['prev_frag_downloaded_bytes'] = frag_downloaded_bytes
             self._hook_progress(state)
 
         ctx['dl'].add_progress_hook(frag_progress_hook)
@@ -98,13 +248,21 @@ class FragmentFD(FileDownloader):
 
     def _finish_frag_download(self, ctx):
         ctx['dest_stream'].close()
+        if self.__do_ytdl_file(ctx):
+            ytdl_filename = encodeFilename(self.ytdl_filename(ctx['filename']))
+            if os.path.isfile(ytdl_filename):
+                os.remove(ytdl_filename)
         elapsed = time.time() - ctx['started']
-        self.try_rename(ctx['tmpfilename'], ctx['filename'])
-        fsize = os.path.getsize(encodeFilename(ctx['filename']))
+
+        if ctx['tmpfilename'] == '-':
+            downloaded_bytes = ctx['complete_frags_downloaded_bytes']
+        else:
+            self.try_rename(ctx['tmpfilename'], ctx['filename'])
+            downloaded_bytes = os.path.getsize(encodeFilename(ctx['filename']))
 
         self._hook_progress({
-            'downloaded_bytes': fsize,
-            'total_bytes': fsize,
+            'downloaded_bytes': downloaded_bytes,
+            'total_bytes': downloaded_bytes,
             'filename': ctx['filename'],
             'status': 'finished',
             'elapsed': elapsed,