2015-06-03 15:10:18 +00:00
|
|
|
from __future__ import unicode_literals
|
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
import os
|
2015-06-03 15:10:18 +00:00
|
|
|
import re
|
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
from .fragment import FragmentFD
|
2016-03-19 14:42:23 +00:00
|
|
|
from ..compat import compat_urllib_error
|
2016-02-09 16:25:02 +00:00
|
|
|
from ..utils import (
|
|
|
|
sanitize_open,
|
|
|
|
encodeFilename,
|
|
|
|
)
|
2015-06-04 14:12:05 +00:00
|
|
|
|
2015-06-03 15:10:18 +00:00
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
class DashSegmentsFD(FragmentFD):
|
2015-06-03 15:10:18 +00:00
|
|
|
"""
|
|
|
|
Download segments in a DASH manifest
|
|
|
|
"""
|
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
FD_NAME = 'dashsegments'
|
2015-06-10 06:45:54 +00:00
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
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 06:45:54 +00:00
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
ctx = {
|
|
|
|
'filename': filename,
|
|
|
|
'total_frags': len(segment_urls) + (1 if initialization_url else 0),
|
|
|
|
}
|
2015-06-10 06:45:54 +00:00
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
self._prepare_and_start_frag_download(ctx)
|
2015-06-03 15:10:18 +00:00
|
|
|
|
|
|
|
def combine_url(base_url, target_url):
|
|
|
|
if re.match(r'^https?://', target_url):
|
|
|
|
return target_url
|
2015-07-23 10:09:30 +00:00
|
|
|
return '%s%s%s' % (base_url, '' if base_url.endswith('/') else '/', target_url)
|
2015-06-03 15:10:18 +00:00
|
|
|
|
2016-02-09 16:25:02 +00:00
|
|
|
segments_filenames = []
|
2016-02-14 08:13:09 +00:00
|
|
|
|
2016-03-19 14:42:23 +00:00
|
|
|
fragment_retries = self.params.get('fragment_retries', 0)
|
2016-08-26 21:55:55 +00:00
|
|
|
skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True)
|
2016-03-19 14:42:23 +00:00
|
|
|
|
2016-09-03 15:00:52 +00:00
|
|
|
def process_segment(segment, tmp_filename, fatal):
|
|
|
|
target_url, segment_name = segment
|
2016-03-19 14:42:23 +00:00
|
|
|
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
|
2016-08-26 21:57:59 +00:00
|
|
|
except compat_urllib_error.HTTPError as err:
|
2016-03-19 14:42:23 +00:00
|
|
|
# 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.
|
2016-08-26 21:55:55 +00:00
|
|
|
# To be future-proof we will retry all fragments that fail with any
|
|
|
|
# HTTP error.
|
2016-03-19 14:42:23 +00:00
|
|
|
count += 1
|
|
|
|
if count <= fragment_retries:
|
2016-08-26 21:57:59 +00:00
|
|
|
self.report_retry_fragment(err, segment_name, count, fragment_retries)
|
2016-03-19 14:42:23 +00:00
|
|
|
if count > fragment_retries:
|
2016-09-03 15:00:52 +00:00
|
|
|
if not fatal:
|
2016-08-26 21:55:55 +00:00
|
|
|
self.report_skip_fragment(segment_name)
|
2016-08-30 15:28:14 +00:00
|
|
|
return True
|
2016-03-19 14:42:23 +00:00
|
|
|
self.report_error('giving up after %s fragment retries' % fragment_retries)
|
2016-02-09 16:25:02 +00:00
|
|
|
return False
|
2016-08-30 15:28:14 +00:00
|
|
|
return True
|
2016-02-09 16:25:02 +00:00
|
|
|
|
2016-09-03 15:00:52 +00:00
|
|
|
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):
|
2016-08-30 15:28:14 +00:00
|
|
|
return False
|
2016-02-09 16:25:02 +00:00
|
|
|
|
|
|
|
self._finish_frag_download(ctx)
|
|
|
|
|
|
|
|
for segment_file in segments_filenames:
|
|
|
|
os.remove(encodeFilename(segment_file))
|
2015-06-03 15:10:18 +00:00
|
|
|
|
|
|
|
return True
|