2015-06-03 11:10:18 -04:00
|
|
|
from __future__ import unicode_literals
|
|
|
|
|
2021-02-08 11:46:01 -05:00
|
|
|
from ..downloader import _get_real_downloader
|
2016-02-09 11:25:02 -05:00
|
|
|
from .fragment import FragmentFD
|
2021-02-08 11:46:01 -05:00
|
|
|
|
2016-03-19 10:42:23 -04:00
|
|
|
from ..compat import compat_urllib_error
|
2018-07-07 21:22:56 -04:00
|
|
|
from ..utils import (
|
|
|
|
DownloadError,
|
|
|
|
urljoin,
|
|
|
|
)
|
2015-06-04 10:12:05 -04:00
|
|
|
|
2015-06-03 11:10:18 -04:00
|
|
|
|
2016-02-09 11:25:02 -05:00
|
|
|
class DashSegmentsFD(FragmentFD):
|
2015-06-03 11:10:18 -04:00
|
|
|
"""
|
2021-03-10 10:26:24 -05:00
|
|
|
Download segments in a DASH manifest. External downloaders can take over
|
|
|
|
the fragment downloads by supporting the 'frag_urls' protocol
|
2015-06-03 11:10:18 -04:00
|
|
|
"""
|
|
|
|
|
2016-02-09 11:25:02 -05:00
|
|
|
FD_NAME = 'dashsegments'
|
2015-06-10 02:45:54 -04:00
|
|
|
|
2016-02-09 11:25:02 -05:00
|
|
|
def real_download(self, filename, info_dict):
|
2017-08-04 19:57:19 -04:00
|
|
|
fragment_base_url = info_dict.get('fragment_base_url')
|
|
|
|
fragments = info_dict['fragments'][:1] if self.params.get(
|
2016-09-17 09:35:22 -04:00
|
|
|
'test', False) else info_dict['fragments']
|
2015-06-10 02:45:54 -04:00
|
|
|
|
2021-02-08 11:46:01 -05:00
|
|
|
real_downloader = _get_real_downloader(info_dict, 'frag_urls', self.params, None)
|
|
|
|
|
2016-02-09 11:25:02 -05:00
|
|
|
ctx = {
|
|
|
|
'filename': filename,
|
2017-08-04 19:57:19 -04:00
|
|
|
'total_frags': len(fragments),
|
2016-02-09 11:25:02 -05:00
|
|
|
}
|
2015-06-10 02:45:54 -04:00
|
|
|
|
2021-02-08 11:46:01 -05:00
|
|
|
if real_downloader:
|
|
|
|
self._prepare_external_frag_download(ctx)
|
|
|
|
else:
|
|
|
|
self._prepare_and_start_frag_download(ctx)
|
2015-06-03 11:10:18 -04:00
|
|
|
|
2016-03-19 10:42:23 -04:00
|
|
|
fragment_retries = self.params.get('fragment_retries', 0)
|
2016-08-26 17:55:55 -04:00
|
|
|
skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True)
|
2016-03-19 10:42:23 -04:00
|
|
|
|
2021-03-10 10:26:24 -05:00
|
|
|
fragments_to_download = []
|
2016-06-28 13:07:50 -04:00
|
|
|
frag_index = 0
|
2017-08-04 19:57:19 -04:00
|
|
|
for i, fragment in enumerate(fragments):
|
2016-06-28 13:07:50 -04:00
|
|
|
frag_index += 1
|
2017-04-22 11:42:24 -04:00
|
|
|
if frag_index <= ctx['fragment_index']:
|
2016-06-28 13:07:50 -04:00
|
|
|
continue
|
2021-02-08 11:46:01 -05:00
|
|
|
fragment_url = fragment.get('url')
|
|
|
|
if not fragment_url:
|
|
|
|
assert fragment_base_url
|
|
|
|
fragment_url = urljoin(fragment_base_url, fragment['path'])
|
|
|
|
|
|
|
|
if real_downloader:
|
2021-03-10 10:26:24 -05:00
|
|
|
fragments_to_download.append({
|
2021-03-10 09:39:40 -05:00
|
|
|
'url': fragment_url,
|
|
|
|
})
|
2021-02-08 11:46:01 -05:00
|
|
|
continue
|
|
|
|
|
2016-09-17 09:35:22 -04:00
|
|
|
# In DASH, the first segment contains necessary headers to
|
|
|
|
# generate a valid MP4 file, so always abort for the first segment
|
2016-06-28 13:07:50 -04:00
|
|
|
fatal = i == 0 or not skip_unavailable_fragments
|
2016-03-19 10:42:23 -04:00
|
|
|
count = 0
|
|
|
|
while count <= fragment_retries:
|
|
|
|
try:
|
2017-08-04 19:57:19 -04:00
|
|
|
success, frag_content = self._download_fragment(ctx, fragment_url, info_dict)
|
2016-03-19 10:42:23 -04:00
|
|
|
if not success:
|
|
|
|
return False
|
2016-06-28 13:07:50 -04:00
|
|
|
self._append_fragment(ctx, frag_content)
|
2016-03-19 10:42:23 -04:00
|
|
|
break
|
2016-08-26 17:57:59 -04:00
|
|
|
except compat_urllib_error.HTTPError as err:
|
2016-03-19 10:42:23 -04:00
|
|
|
# YouTube may often return 404 HTTP error for a fragment causing the
|
|
|
|
# whole download to fail. However if the same fragment is immediately
|
2019-07-26 11:30:18 -04:00
|
|
|
# retried with the same request data this usually succeeds (1-2 attempts
|
2016-03-19 10:42:23 -04:00
|
|
|
# is usually enough) thus allowing to download the whole file successfully.
|
2016-08-26 17:55:55 -04:00
|
|
|
# To be future-proof we will retry all fragments that fail with any
|
|
|
|
# HTTP error.
|
2016-03-19 10:42:23 -04:00
|
|
|
count += 1
|
|
|
|
if count <= fragment_retries:
|
2016-06-28 13:07:50 -04:00
|
|
|
self.report_retry_fragment(err, frag_index, count, fragment_retries)
|
2018-07-07 21:22:56 -04:00
|
|
|
except DownloadError:
|
|
|
|
# Don't retry fragment if error occurred during HTTP downloading
|
|
|
|
# itself since it has own retry settings
|
|
|
|
if not fatal:
|
|
|
|
self.report_skip_fragment(frag_index)
|
|
|
|
break
|
|
|
|
raise
|
|
|
|
|
2016-03-19 10:42:23 -04:00
|
|
|
if count > fragment_retries:
|
2016-09-03 11:00:52 -04:00
|
|
|
if not fatal:
|
2016-06-28 13:07:50 -04:00
|
|
|
self.report_skip_fragment(frag_index)
|
|
|
|
continue
|
2016-03-19 10:42:23 -04:00
|
|
|
self.report_error('giving up after %s fragment retries' % fragment_retries)
|
2016-02-09 11:25:02 -05:00
|
|
|
return False
|
|
|
|
|
2021-02-08 11:46:01 -05:00
|
|
|
if real_downloader:
|
|
|
|
info_copy = info_dict.copy()
|
2021-03-10 10:26:24 -05:00
|
|
|
info_copy['fragments'] = fragments_to_download
|
2021-02-08 11:46:01 -05:00
|
|
|
fd = real_downloader(self.ydl, self.params)
|
|
|
|
# TODO: Make progress updates work without hooking twice
|
|
|
|
# for ph in self._progress_hooks:
|
|
|
|
# fd.add_progress_hook(ph)
|
|
|
|
success = fd.real_download(filename, info_copy)
|
|
|
|
if not success:
|
|
|
|
return False
|
|
|
|
else:
|
|
|
|
self._finish_frag_download(ctx)
|
2015-06-03 11:10:18 -04:00
|
|
|
return True
|