mirror of
https://github.com/ytdl-org/youtube-dl.git
synced 2024-01-07 17:16:08 +00:00
[downloader/dash] Implement dashsegments fd in terms of fragment fd
This commit is contained in:
parent
d413095f7e
commit
c43fe0268c
|
@ -1,67 +1,58 @@
|
|||
from __future__ import unicode_literals
|
||||
|
||||
import os
|
||||
import re
|
||||
|
||||
from .common import FileDownloader
|
||||
from ..utils import sanitized_Request
|
||||
from .fragment import FragmentFD
|
||||
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')
|
||||
|
||||
is_test = self.params.get('test', False)
|
||||
remaining_bytes = self._TEST_FILE_SIZE if is_test else None
|
||||
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, remaining_bytes=None):
|
||||
self.to_screen('[DashSegments] %s: Downloading %s' % (info_dict['id'], target_name))
|
||||
req = sanitized_Request(target_url)
|
||||
if remaining_bytes is not None:
|
||||
req.add_header('Range', 'bytes=0-%d' % (remaining_bytes - 1))
|
||||
|
||||
data = self.ydl.urlopen(req).read()
|
||||
|
||||
if remaining_bytes is not None:
|
||||
data = data[:remaining_bytes]
|
||||
|
||||
outf.write(data)
|
||||
return 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%s' % (base_url, '' if base_url.endswith('/') else '/', target_url)
|
||||
|
||||
with open(tmpfilename, 'wb') as outf:
|
||||
if info_dict.get('initialization_url'):
|
||||
append_url_to_file(
|
||||
outf, combine_url(base_url, info_dict['initialization_url']),
|
||||
'initialization segment')
|
||||
for i, segment_url in enumerate(segment_urls):
|
||||
segment_len = append_url_to_file(
|
||||
outf, combine_url(base_url, segment_url),
|
||||
'segment %d / %d' % (i + 1, len(segment_urls)),
|
||||
remaining_bytes)
|
||||
byte_counter += segment_len
|
||||
if remaining_bytes is not None:
|
||||
remaining_bytes -= segment_len
|
||||
if remaining_bytes <= 0:
|
||||
break
|
||||
segments_filenames = []
|
||||
def append_url_to_file(target_url, target_filename):
|
||||
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)
|
||||
|
||||
self.try_rename(tmpfilename, filename)
|
||||
if initialization_url:
|
||||
append_url_to_file(initialization_url, ctx['tmpfilename'] + '-Init')
|
||||
for i, segment_url in enumerate(segment_urls):
|
||||
segment_filename = '%s-Seg%d' % (ctx['tmpfilename'], i)
|
||||
append_url_to_file(segment_url, segment_filename)
|
||||
|
||||
self._hook_progress({
|
||||
'downloaded_bytes': byte_counter,
|
||||
'total_bytes': byte_counter,
|
||||
'filename': filename,
|
||||
'status': 'finished',
|
||||
})
|
||||
self._finish_frag_download(ctx)
|
||||
|
||||
for segment_file in segments_filenames:
|
||||
os.remove(encodeFilename(segment_file))
|
||||
|
||||
return True
|
||||
|
|
Loading…
Reference in a new issue