youtube-dl/youtube_dl/downloader/dash.py

93 lines
3.6 KiB
Python
Raw Normal View History

from __future__ import unicode_literals
import os
import re
from .fragment import FragmentFD
from ..compat import compat_urllib_error
from ..utils import (
sanitize_open,
encodeFilename,
)
2015-06-04 16:12:05 +02:00
class DashSegmentsFD(FragmentFD):
"""
Download segments in a DASH manifest
"""
FD_NAME = 'dashsegments'
2015-06-10 08:45:54 +02:00
def real_download(self, filename, info_dict):
base_url = info_dict['url']
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')
2015-06-10 08:45:54 +02:00
ctx = {
'filename': filename,
'total_frags': len(segment_urls) + (1 if initialization_url else 0),
}
2015-06-10 08:45:54 +02:00
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%s' % (base_url, '' if base_url.endswith('/') else '/', target_url)
segments_filenames = []
2016-02-14 09:13:09 +01:00
fragment_retries = self.params.get('fragment_retries', 0)
skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True)
def process_segment(segment, tmp_filename, fatal):
target_url, segment_name = segment
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 as err:
# 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(err, segment_name, count, fragment_retries)
if count > fragment_retries:
if not fatal:
self.report_skip_fragment(segment_name)
return True
self.report_error('giving up after %s fragment retries' % fragment_retries)
return False
return True
segments_to_download = [(initialization_url, 'Init')] if initialization_url else []
segments_to_download.extend([
(segment_url, 'Seg%d' % i)
for i, segment_url in enumerate(segment_urls)])
for i, segment in enumerate(segments_to_download):
# In DASH, the first segment contains necessary headers to
# generate a valid MP4 file, so always abort for the first segment
fatal = i == 0 or not skip_unavailable_fragments
if not process_segment(segment, ctx['tmpfilename'], fatal):
return False
self._finish_frag_download(ctx)
for segment_file in segments_filenames:
os.remove(encodeFilename(segment_file))
return True