From 25afc2a7830e281e849609202b4f70728664bdb7 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Sergey=20M=E2=80=A4?= Date: Sat, 27 Aug 2016 04:55:55 +0700 Subject: [PATCH] [downloader/dash:hls] Respect --fragment-retries and --skip-unavailable-fragments (Closes #10165, closes #10448) --- youtube_dl/downloader/dash.py | 12 +++++----- youtube_dl/downloader/hls.py | 41 +++++++++++++++++++++++++++++------ 2 files changed, 41 insertions(+), 12 deletions(-) diff --git a/youtube_dl/downloader/dash.py b/youtube_dl/downloader/dash.py index 8bbab9dbc..cbcee324d 100644 --- a/youtube_dl/downloader/dash.py +++ b/youtube_dl/downloader/dash.py @@ -38,6 +38,7 @@ class DashSegmentsFD(FragmentFD): 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) @@ -52,19 +53,20 @@ class DashSegmentsFD(FragmentFD): down.close() segments_filenames.append(target_sanitized) break - except (compat_urllib_error.HTTPError, ) as err: + 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. - # So, we will retry all fragments that fail with 404 HTTP error for now. - if err.code != 404: - raise - # Retry fragment + # 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 diff --git a/youtube_dl/downloader/hls.py b/youtube_dl/downloader/hls.py index baaff44d5..7412620a5 100644 --- a/youtube_dl/downloader/hls.py +++ b/youtube_dl/downloader/hls.py @@ -13,6 +13,7 @@ from .fragment import FragmentFD from .external import FFmpegFD from ..compat import ( + compat_urllib_error, compat_urlparse, compat_struct_pack, ) @@ -83,6 +84,10 @@ class HlsFD(FragmentFD): self._prepare_and_start_frag_download(ctx) + fragment_retries = self.params.get('fragment_retries', 0) + skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True) + test = self.params.get('test', False) + extra_query = None extra_param_to_segment_url = info_dict.get('extra_param_to_segment_url') if extra_param_to_segment_url: @@ -99,15 +104,37 @@ class HlsFD(FragmentFD): line if re.match(r'^https?://', line) else compat_urlparse.urljoin(man_url, line)) - frag_filename = '%s-Frag%d' % (ctx['tmpfilename'], i) + frag_name = 'Frag%d' % i + frag_filename = '%s-%s' % (ctx['tmpfilename'], frag_name) if extra_query: frag_url = update_url_query(frag_url, extra_query) - success = ctx['dl'].download(frag_filename, {'url': frag_url}) - if not success: + count = 0 + while count <= fragment_retries: + try: + success = ctx['dl'].download(frag_filename, {'url': frag_url}) + if not success: + return False + down, frag_sanitized = sanitize_open(frag_filename, 'rb') + frag_content = down.read() + down.close() + break + except compat_urllib_error.HTTPError: + # Unavailable (possibly temporary) fragments may be served. + # First we try to retry then either skip or abort. + # See https://github.com/rg3/youtube-dl/issues/10165, + # https://github.com/rg3/youtube-dl/issues/10448). + count += 1 + if count <= fragment_retries: + self.report_retry_fragment(frag_name, count, fragment_retries) + if count > fragment_retries: + if skip_unavailable_fragments: + i += 1 + media_sequence += 1 + self.report_skip_fragment(frag_name) + continue + self.report_error( + 'giving up after %s fragment retries' % fragment_retries) return False - down, frag_sanitized = sanitize_open(frag_filename, 'rb') - frag_content = down.read() - down.close() if decrypt_info['METHOD'] == 'AES-128': iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', media_sequence) frag_content = AES.new( @@ -115,7 +142,7 @@ class HlsFD(FragmentFD): ctx['dest_stream'].write(frag_content) frags_filenames.append(frag_sanitized) # We only download the first fragment during the test - if self.params.get('test', False): + if test: break i += 1 media_sequence += 1 -- 2.39.5