X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fdownloader%2Fdash.py;h=cbcee324df957c36cdcd00ef7e49a4c30bf0d949;hb=25afc2a7830e281e849609202b4f70728664bdb7;hp=18eca2c0426d22193155a1408f34249904881e02;hpb=6800d3372f35e08dcc4d34d06601815bf0cb0a3d;p=youtube-dl diff --git a/youtube_dl/downloader/dash.py b/youtube_dl/downloader/dash.py index 18eca2c04..cbcee324d 100644 --- a/youtube_dl/downloader/dash.py +++ b/youtube_dl/downloader/dash.py @@ -1,50 +1,83 @@ from __future__ import unicode_literals -from .common import FileDownloader -from ..compat import compat_urllib_request +import os import re +from .fragment import FragmentFD +from ..compat import compat_urllib_error +from ..utils import ( + sanitize_open, + encodeFilename, +) -class DashSegmentsFD(FileDownloader): + +class DashSegmentsFD(FragmentFD): """ Download segments in a DASH manifest """ + + FD_NAME = 'dashsegments' + def real_download(self, filename, info_dict): - self.report_destination(filename) - tmpfilename = self.temp_name(filename) base_url = info_dict['url'] - segment_urls = info_dict['segment_urls'] + segment_urls = [info_dict['segment_urls'][0]] if self.params.get('test', False) else info_dict['segment_urls'] + initialization_url = info_dict.get('initialization_url') - self.byte_counter = 0 + ctx = { + 'filename': filename, + 'total_frags': len(segment_urls) + (1 if initialization_url else 0), + } - def append_url_to_file(outf, target_url, target_name): - self.to_screen('[DashSegments] %s: Downloading %s' % (info_dict['id'], target_name)) - req = compat_urllib_request.Request(target_url) - data = self.ydl.urlopen(req).read() - outf.write(data) - self.byte_counter += len(data) + self._prepare_and_start_frag_download(ctx) def combine_url(base_url, target_url): if re.match(r'^https?://', target_url): return target_url - return '%s/%s' % (base_url, target_url) - - with open(tmpfilename, 'wb') as outf: - append_url_to_file( - outf, combine_url(base_url, info_dict['initialization_url']), - 'initialization segment') - for i, segment_url in enumerate(segment_urls): - append_url_to_file( - outf, combine_url(base_url, segment_url), - 'segment %d / %d' % (i + 1, len(segment_urls))) - - self.try_rename(tmpfilename, filename) - - self._hook_progress({ - 'downloaded_bytes': self.byte_counter, - 'total_bytes': self.byte_counter, - 'filename': filename, - 'status': 'finished', - }) + return '%s%s%s' % (base_url, '' if base_url.endswith('/') else '/', target_url) + + segments_filenames = [] + + fragment_retries = self.params.get('fragment_retries', 0) + skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True) + + def append_url_to_file(target_url, tmp_filename, segment_name): + target_filename = '%s-%s' % (tmp_filename, segment_name) + count = 0 + while count <= fragment_retries: + try: + success = ctx['dl'].download(target_filename, {'url': combine_url(base_url, target_url)}) + if not success: + return False + down, target_sanitized = sanitize_open(target_filename, 'rb') + ctx['dest_stream'].write(down.read()) + down.close() + segments_filenames.append(target_sanitized) + break + except compat_urllib_error.HTTPError: + # YouTube may often return 404 HTTP error for a fragment causing the + # whole download to fail. However if the same fragment is immediately + # retried with the same request data this usually succeeds (1-2 attemps + # is usually enough) thus allowing to download the whole file successfully. + # To be future-proof we will retry all fragments that fail with any + # HTTP error. + count += 1 + if count <= fragment_retries: + self.report_retry_fragment(segment_name, count, fragment_retries) + if count > fragment_retries: + if skip_unavailable_fragments: + self.report_skip_fragment(segment_name) + return + self.report_error('giving up after %s fragment retries' % fragment_retries) + return False + + if initialization_url: + append_url_to_file(initialization_url, ctx['tmpfilename'], 'Init') + for i, segment_url in enumerate(segment_urls): + append_url_to_file(segment_url, ctx['tmpfilename'], 'Seg%d' % i) + + self._finish_frag_download(ctx) + + for segment_file in segments_filenames: + os.remove(encodeFilename(segment_file)) return True