-from __future__ import unicode_literals
-
+import enum
import os.path
import re
import subprocess
import sys
import time
-try:
- from Crypto.Cipher import AES
- can_decrypt_frag = True
-except ImportError:
- can_decrypt_frag = False
-
-from .common import FileDownloader
-from ..compat import (
- compat_setenv,
- compat_str,
-)
-from ..postprocessor.ffmpeg import FFmpegPostProcessor, EXT_TO_OUT_FORMATS
+from .fragment import FragmentFD
+from ..compat import functools # isort: split
+from ..postprocessor.ffmpeg import EXT_TO_OUT_FORMATS, FFmpegPostProcessor
from ..utils import (
+ Popen,
+ _configuration_args,
+ check_executable,
+ classproperty,
+ cli_bool_option,
cli_option,
cli_valueless_option,
- cli_bool_option,
- cli_configuration_args,
- encodeFilename,
- error_to_compat_str,
+ determine_ext,
encodeArgument,
+ encodeFilename,
handle_youtubedl_headers,
- check_executable,
- is_outdated_version,
- process_communicate_or_kill,
- sanitized_Request,
- sanitize_open,
+ remove_end,
+ traverse_obj,
)
-class ExternalFD(FileDownloader):
+class Features(enum.Enum):
+ TO_STDOUT = enum.auto()
+ MULTIPLE_FORMATS = enum.auto()
+
+
+class ExternalFD(FragmentFD):
SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps')
+ SUPPORTED_FEATURES = ()
+ _CAPTURE_STDERR = True
def real_download(self, filename, info_dict):
self.report_destination(filename)
}
if filename != '-':
fsize = os.path.getsize(encodeFilename(tmpfilename))
- self.to_screen('\r[%s] Downloaded %s bytes' % (self.get_basename(), fsize))
+ self.to_screen(f'\r[{self.get_basename()}] Downloaded {fsize} bytes')
self.try_rename(tmpfilename, filename)
status.update({
'downloaded_bytes': fsize,
'total_bytes': fsize,
})
- self._hook_progress(status)
+ self._hook_progress(status, info_dict)
return True
else:
self.to_stderr('\n')
def get_basename(cls):
return cls.__name__[:-2].lower()
- @property
+ @classproperty
+ def EXE_NAME(cls):
+ return cls.get_basename()
+
+ @functools.cached_property
def exe(self):
- return self.params.get('external_downloader')
+ return self.EXE_NAME
@classmethod
def available(cls, path=None):
- return check_executable(path or cls.get_basename(), [cls.AVAILABLE_OPT])
+ path = check_executable(
+ cls.EXE_NAME if path in (None, cls.get_basename()) else path,
+ [cls.AVAILABLE_OPT])
+ if not path:
+ return False
+ cls.exe = path
+ return path
@classmethod
def supports(cls, info_dict):
- return info_dict['protocol'] in cls.SUPPORTED_PROTOCOLS
+ return all((
+ not info_dict.get('to_stdout') or Features.TO_STDOUT in cls.SUPPORTED_FEATURES,
+ '+' not in info_dict['protocol'] or Features.MULTIPLE_FORMATS in cls.SUPPORTED_FEATURES,
+ all(proto in cls.SUPPORTED_PROTOCOLS for proto in info_dict['protocol'].split('+')),
+ ))
@classmethod
def can_download(cls, info_dict, path=None):
def _valueless_option(self, command_option, param, expected_value=True):
return cli_valueless_option(self.params, command_option, param, expected_value)
- def _configuration_args(self, *args, **kwargs):
- return cli_configuration_args(
- self.params.get('external_downloader_args'),
- self.get_basename(), *args, **kwargs)
+ def _configuration_args(self, keys=None, *args, **kwargs):
+ return _configuration_args(
+ self.get_basename(), self.params.get('external_downloader_args'), self.EXE_NAME,
+ keys, *args, **kwargs)
def _call_downloader(self, tmpfilename, info_dict):
""" Either overwrite this or implement _make_cmd """
self._debug_cmd(cmd)
- p = subprocess.Popen(
- cmd, stderr=subprocess.PIPE)
- _, stderr = process_communicate_or_kill(p)
- if p.returncode != 0:
- self.to_stderr(stderr.decode('utf-8', 'replace'))
+ if 'fragments' not in info_dict:
+ _, stderr, returncode = Popen.run(
+ cmd, text=True, stderr=subprocess.PIPE if self._CAPTURE_STDERR else None)
+ if returncode and stderr:
+ self.to_stderr(stderr)
+ return returncode
+
+ fragment_retries = self.params.get('fragment_retries', 0)
+ skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True)
+
+ count = 0
+ while count <= fragment_retries:
+ _, stderr, returncode = Popen.run(cmd, text=True, stderr=subprocess.PIPE)
+ if not returncode:
+ break
- if 'fragments' in info_dict:
- file_list = []
- dest, _ = sanitize_open(tmpfilename, 'wb')
- for i, fragment in enumerate(info_dict['fragments']):
- file = '%s_%s.frag' % (tmpfilename, i)
- decrypt_info = fragment.get('decrypt_info')
- src, _ = sanitize_open(file, 'rb')
- if decrypt_info:
- if decrypt_info['METHOD'] == 'AES-128':
- iv = decrypt_info.get('IV')
- decrypt_info['KEY'] = decrypt_info.get('KEY') or self.ydl.urlopen(
- self._prepare_url(info_dict, info_dict.get('_decryption_key_url') or decrypt_info['URI'])).read()
- encrypted_data = src.read()
- decrypted_data = AES.new(
- decrypt_info['KEY'], AES.MODE_CBC, iv).decrypt(encrypted_data)
- dest.write(decrypted_data)
- else:
- fragment_data = src.read()
- dest.write(fragment_data)
- else:
- fragment_data = src.read()
- dest.write(fragment_data)
- src.close()
- file_list.append(file)
- dest.close()
+ # TODO: Decide whether to retry based on error code
+ # https://aria2.github.io/manual/en/html/aria2c.html#exit-status
+ if stderr:
+ self.to_stderr(stderr)
+ count += 1
+ if count <= fragment_retries:
+ self.to_screen(
+ '[%s] Got error. Retrying fragments (attempt %d of %s)...'
+ % (self.get_basename(), count, self.format_retries(fragment_retries)))
+ self.sleep_retry('fragment', count)
+ if count > fragment_retries:
+ if not skip_unavailable_fragments:
+ self.report_error('Giving up after %s fragment retries' % fragment_retries)
+ return -1
+
+ decrypt_fragment = self.decrypter(info_dict)
+ dest, _ = self.sanitize_open(tmpfilename, 'wb')
+ for frag_index, fragment in enumerate(info_dict['fragments']):
+ fragment_filename = '%s-Frag%d' % (tmpfilename, frag_index)
+ try:
+ src, _ = self.sanitize_open(fragment_filename, 'rb')
+ except OSError as err:
+ if skip_unavailable_fragments and frag_index > 1:
+ self.report_skip_fragment(frag_index, err)
+ continue
+ self.report_error(f'Unable to open fragment {frag_index}; {err}')
+ return -1
+ dest.write(decrypt_fragment(fragment, src.read()))
+ src.close()
if not self.params.get('keep_fragments', False):
- for file_path in file_list:
- try:
- os.remove(file_path)
- except OSError as ose:
- self.report_error("Unable to delete file %s; %s" % (file_path, error_to_compat_str(ose)))
- try:
- file_path = '%s.frag.urls' % tmpfilename
- os.remove(file_path)
- except OSError as ose:
- self.report_error("Unable to delete file %s; %s" % (file_path, error_to_compat_str(ose)))
-
- return p.returncode
-
- def _prepare_url(self, info_dict, url):
- headers = info_dict.get('http_headers')
- return sanitized_Request(url, None, headers) if headers else url
+ self.try_remove(encodeFilename(fragment_filename))
+ dest.close()
+ self.try_remove(encodeFilename('%s.frag.urls' % tmpfilename))
+ return 0
class CurlFD(ExternalFD):
AVAILABLE_OPT = '-V'
+ _CAPTURE_STDERR = False # curl writes the progress to stderr
def _make_cmd(self, tmpfilename, info_dict):
- cmd = [self.exe, '--location', '-o', tmpfilename]
+ cmd = [self.exe, '--location', '-o', tmpfilename, '--compressed']
if info_dict.get('http_headers') is not None:
for key, val in info_dict['http_headers'].items():
- cmd += ['--header', '%s: %s' % (key, val)]
+ cmd += ['--header', f'{key}: {val}']
cmd += self._bool_option('--continue-at', 'continuedl', '-', '0')
cmd += self._valueless_option('--silent', 'noprogress')
cmd += ['--', info_dict['url']]
return cmd
- def _call_downloader(self, tmpfilename, info_dict):
- cmd = [encodeArgument(a) for a in self._make_cmd(tmpfilename, info_dict)]
-
- self._debug_cmd(cmd)
-
- # curl writes the progress to stderr so don't capture it.
- p = subprocess.Popen(cmd)
- process_communicate_or_kill(p)
- return p.returncode
-
class AxelFD(ExternalFD):
AVAILABLE_OPT = '-V'
cmd = [self.exe, '-o', tmpfilename]
if info_dict.get('http_headers') is not None:
for key, val in info_dict['http_headers'].items():
- cmd += ['-H', '%s: %s' % (key, val)]
+ cmd += ['-H', f'{key}: {val}']
cmd += self._configuration_args()
cmd += ['--', info_dict['url']]
return cmd
AVAILABLE_OPT = '--version'
def _make_cmd(self, tmpfilename, info_dict):
- cmd = [self.exe, '-O', tmpfilename, '-nv', '--no-cookies']
+ cmd = [self.exe, '-O', tmpfilename, '-nv', '--no-cookies', '--compression=auto']
if info_dict.get('http_headers') is not None:
for key, val in info_dict['http_headers'].items():
- cmd += ['--header', '%s: %s' % (key, val)]
+ cmd += ['--header', f'{key}: {val}']
cmd += self._option('--limit-rate', 'ratelimit')
retry = self._option('--tries', 'retries')
if len(retry) == 2:
retry[1] = '0'
cmd += retry
cmd += self._option('--bind-address', 'source_address')
- cmd += self._option('--proxy', 'proxy')
+ proxy = self.params.get('proxy')
+ if proxy:
+ for var in ('http_proxy', 'https_proxy'):
+ cmd += ['--execute', f'{var}={proxy}']
cmd += self._valueless_option('--no-check-certificate', 'nocheckcertificate')
cmd += self._configuration_args()
cmd += ['--', info_dict['url']]
class Aria2cFD(ExternalFD):
AVAILABLE_OPT = '-v'
- SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps', 'frag_urls')
+ SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps', 'dash_frag_urls', 'm3u8_frag_urls')
@staticmethod
def supports_manifest(manifest):
return all(check_results)
def _make_cmd(self, tmpfilename, info_dict):
- cmd = [self.exe, '-c']
- dn = os.path.dirname(tmpfilename)
- if 'fragments' not in info_dict:
- cmd += ['--out', os.path.basename(tmpfilename)]
- verbose_level_args = ['--console-log-level=warn', '--summary-interval=0']
- cmd += self._configuration_args(['--file-allocation=none', '-x16', '-j16', '-s16'] + verbose_level_args)
- if dn:
- cmd += ['--dir', dn]
+ cmd = [self.exe, '-c',
+ '--console-log-level=warn', '--summary-interval=0', '--download-result=hide',
+ '--http-accept-gzip=true', '--file-allocation=none', '-x16', '-j16', '-s16']
+ if 'fragments' in info_dict:
+ cmd += ['--allow-overwrite=true', '--allow-piece-length-change=true']
+ else:
+ cmd += ['--min-split-size', '1M']
+
if info_dict.get('http_headers') is not None:
for key, val in info_dict['http_headers'].items():
- cmd += ['--header', '%s: %s' % (key, val)]
+ cmd += ['--header', f'{key}: {val}']
+ cmd += self._option('--max-overall-download-limit', 'ratelimit')
cmd += self._option('--interface', 'source_address')
cmd += self._option('--all-proxy', 'proxy')
cmd += self._bool_option('--check-certificate', 'nocheckcertificate', 'false', 'true', '=')
cmd += self._bool_option('--remote-time', 'updatetime', 'true', 'false', '=')
+ cmd += self._bool_option('--show-console-readout', 'noprogress', 'false', 'true', '=')
+ cmd += self._configuration_args()
+
+ # aria2c strips out spaces from the beginning/end of filenames and paths.
+ # We work around this issue by adding a "./" to the beginning of the
+ # filename and relative path, and adding a "/" at the end of the path.
+ # See: https://github.com/yt-dlp/yt-dlp/issues/276
+ # https://github.com/ytdl-org/youtube-dl/issues/20312
+ # https://github.com/aria2/aria2/issues/1373
+ dn = os.path.dirname(tmpfilename)
+ if dn:
+ if not os.path.isabs(dn):
+ dn = f'.{os.path.sep}{dn}'
+ cmd += ['--dir', dn + os.path.sep]
+ if 'fragments' not in info_dict:
+ cmd += ['--out', f'.{os.path.sep}{os.path.basename(tmpfilename)}']
cmd += ['--auto-file-renaming=false']
+
if 'fragments' in info_dict:
- cmd += verbose_level_args
- cmd += ['--uri-selector', 'inorder', '--download-result=hide']
+ cmd += ['--file-allocation=none', '--uri-selector=inorder']
url_list_file = '%s.frag.urls' % tmpfilename
url_list = []
- for i, fragment in enumerate(info_dict['fragments']):
- tmpsegmentname = '%s_%s.frag' % (os.path.basename(tmpfilename), i)
- url_list.append('%s\n\tout=%s' % (fragment['url'], tmpsegmentname))
- stream, _ = sanitize_open(url_list_file, 'wb')
- stream.write('\n'.join(url_list).encode('utf-8'))
+ for frag_index, fragment in enumerate(info_dict['fragments']):
+ fragment_filename = '%s-Frag%d' % (os.path.basename(tmpfilename), frag_index)
+ url_list.append('%s\n\tout=%s' % (fragment['url'], fragment_filename))
+ stream, _ = self.sanitize_open(url_list_file, 'wb')
+ stream.write('\n'.join(url_list).encode())
stream.close()
-
cmd += ['-i', url_list_file]
else:
cmd += ['--', info_dict['url']]
class HttpieFD(ExternalFD):
- @classmethod
- def available(cls, path=None):
- return check_executable(path or 'http', ['--version'])
+ AVAILABLE_OPT = '--version'
+ EXE_NAME = 'http'
def _make_cmd(self, tmpfilename, info_dict):
cmd = ['http', '--download', '--output', tmpfilename, info_dict['url']]
if info_dict.get('http_headers') is not None:
for key, val in info_dict['http_headers'].items():
- cmd += ['%s:%s' % (key, val)]
+ cmd += [f'{key}:{val}']
return cmd
class FFmpegFD(ExternalFD):
- SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps', 'm3u8', 'rtsp', 'rtmp', 'mms')
+ SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps', 'm3u8', 'm3u8_native', 'rtsp', 'rtmp', 'rtmp_ffmpeg', 'mms', 'http_dash_segments')
+ SUPPORTED_FEATURES = (Features.TO_STDOUT, Features.MULTIPLE_FORMATS)
@classmethod
- def available(cls, path=None): # path is ignored for ffmpeg
+ def available(cls, path=None):
+ # TODO: Fix path for ffmpeg
+ # Fixme: This may be wrong when --ffmpeg-location is used
return FFmpegPostProcessor().available
+ def on_process_started(self, proc, stdin):
+ """ Override this in subclasses """
+ pass
+
+ @classmethod
+ def can_merge_formats(cls, info_dict, params):
+ return (
+ info_dict.get('requested_formats')
+ and info_dict.get('protocol')
+ and not params.get('allow_unplayable_formats')
+ and 'no-direct-merge' not in params.get('compat_opts', [])
+ and cls.can_download(info_dict))
+
def _call_downloader(self, tmpfilename, info_dict):
- url = info_dict['url']
+ urls = [f['url'] for f in info_dict.get('requested_formats', [])] or [info_dict['url']]
ffpp = FFmpegPostProcessor(downloader=self)
if not ffpp.available:
self.report_error('m3u8 download detected but ffmpeg could not be found. Please install')
if self.params.get(log_level, False):
args += ['-loglevel', log_level]
break
+ if not self.params.get('verbose'):
+ args += ['-hide_banner']
+
+ args += traverse_obj(info_dict, ('downloader_options', 'ffmpeg_args'), default=[])
+ # These exists only for compatibility. Extractors should use
+ # info_dict['downloader_options']['ffmpeg_args'] instead
+ args += info_dict.get('_ffmpeg_args') or []
seekable = info_dict.get('_seekable')
if seekable is not None:
# setting -seekable prevents ffmpeg from guessing if the server
# http://trac.ffmpeg.org/ticket/6125#comment:10
args += ['-seekable', '1' if seekable else '0']
- args += self._configuration_args()
-
- # start_time = info_dict.get('start_time') or 0
- # if start_time:
- # args += ['-ss', compat_str(start_time)]
- # end_time = info_dict.get('end_time')
- # if end_time:
- # args += ['-t', compat_str(end_time - start_time)]
-
- if info_dict.get('http_headers') is not None and re.match(r'^https?://', url):
- # Trailing \r\n after each HTTP header is important to prevent warning from ffmpeg/avconv:
- # [http @ 00000000003d2fa0] No trailing CRLF found in HTTP header.
- headers = handle_youtubedl_headers(info_dict['http_headers'])
- args += [
+ http_headers = None
+ if info_dict.get('http_headers'):
+ youtubedl_headers = handle_youtubedl_headers(info_dict['http_headers'])
+ http_headers = [
+ # Trailing \r\n after each HTTP header is important to prevent warning from ffmpeg/avconv:
+ # [http @ 00000000003d2fa0] No trailing CRLF found in HTTP header.
'-headers',
- ''.join('%s: %s\r\n' % (key, val) for key, val in headers.items())]
+ ''.join(f'{key}: {val}\r\n' for key, val in youtubedl_headers.items())
+ ]
env = None
proxy = self.params.get('proxy')
# We could switch to the following code if we are able to detect version properly
# args += ['-http_proxy', proxy]
env = os.environ.copy()
- compat_setenv('HTTP_PROXY', proxy, env=env)
- compat_setenv('http_proxy', proxy, env=env)
+ env['HTTP_PROXY'] = proxy
+ env['http_proxy'] = proxy
protocol = info_dict.get('protocol')
if isinstance(conn, list):
for entry in conn:
args += ['-rtmp_conn', entry]
- elif isinstance(conn, compat_str):
+ elif isinstance(conn, str):
args += ['-rtmp_conn', conn]
- args += ['-i', url, '-c', 'copy']
+ start_time, end_time = info_dict.get('section_start') or 0, info_dict.get('section_end')
+
+ for i, url in enumerate(urls):
+ if http_headers is not None and re.match(r'^https?://', url):
+ args += http_headers
+ if start_time:
+ args += ['-ss', str(start_time)]
+ if end_time:
+ args += ['-t', str(end_time - start_time)]
+
+ args += self._configuration_args((f'_i{i + 1}', '_i')) + ['-i', url]
+
+ if not (start_time or end_time) or not self.params.get('force_keyframes_at_cuts'):
+ args += ['-c', 'copy']
+
+ if info_dict.get('requested_formats') or protocol == 'http_dash_segments':
+ for (i, fmt) in enumerate(info_dict.get('requested_formats') or [info_dict]):
+ stream_number = fmt.get('manifest_stream_number', 0)
+ args.extend(['-map', f'{i}:{stream_number}'])
if self.params.get('test', False):
- args += ['-fs', compat_str(self._TEST_FILE_SIZE)]
+ args += ['-fs', str(self._TEST_FILE_SIZE)]
+ ext = info_dict['ext']
if protocol in ('m3u8', 'm3u8_native'):
use_mpegts = (tmpfilename == '-') or self.params.get('hls_use_mpegts')
if use_mpegts is None:
args += ['-f', 'mpegts']
else:
args += ['-f', 'mp4']
- if (ffpp.basename == 'ffmpeg' and is_outdated_version(ffpp._versions['ffmpeg'], '3.2', False)) and (not info_dict.get('acodec') or info_dict['acodec'].split('.')[0] in ('aac', 'mp4a')):
+ if (ffpp.basename == 'ffmpeg' and ffpp._features.get('needs_adtstoasc')) and (not info_dict.get('acodec') or info_dict['acodec'].split('.')[0] in ('aac', 'mp4a')):
args += ['-bsf:a', 'aac_adtstoasc']
elif protocol == 'rtmp':
args += ['-f', 'flv']
+ elif ext == 'mp4' and tmpfilename == '-':
+ args += ['-f', 'mpegts']
+ elif ext == 'unknown_video':
+ ext = determine_ext(remove_end(tmpfilename, '.part'))
+ if ext == 'unknown_video':
+ self.report_warning(
+ 'The video format is unknown and cannot be downloaded by ffmpeg. '
+ 'Explicitly set the extension in the filename to attempt download in that format')
+ else:
+ self.report_warning(f'The video format is unknown. Trying to download as {ext} according to the filename')
+ args += ['-f', EXT_TO_OUT_FORMATS.get(ext, ext)]
else:
- args += ['-f', EXT_TO_OUT_FORMATS.get(info_dict['ext'], info_dict['ext'])]
+ args += ['-f', EXT_TO_OUT_FORMATS.get(ext, ext)]
+
+ args += self._configuration_args(('_o1', '_o', ''))
args = [encodeArgument(opt) for opt in args]
args.append(encodeFilename(ffpp._ffmpeg_filename_argument(tmpfilename), True))
-
self._debug_cmd(args)
- proc = subprocess.Popen(args, stdin=subprocess.PIPE, env=env)
- try:
- retval = proc.wait()
- except BaseException as e:
- # subprocces.run would send the SIGKILL signal to ffmpeg and the
- # mp4 file couldn't be played, but if we ask ffmpeg to quit it
- # produces a file that is playable (this is mostly useful for live
- # streams). Note that Windows is not affected and produces playable
- # files (see https://github.com/ytdl-org/youtube-dl/issues/8300).
- if isinstance(e, KeyboardInterrupt) and sys.platform != 'win32':
- process_communicate_or_kill(proc, b'q')
- else:
- proc.kill()
- proc.wait()
- raise
- return retval
+ with Popen(args, stdin=subprocess.PIPE, env=env) as proc:
+ if url in ('-', 'pipe:'):
+ self.on_process_started(proc, proc.stdin)
+ try:
+ retval = proc.wait()
+ except BaseException as e:
+ # subprocces.run would send the SIGKILL signal to ffmpeg and the
+ # mp4 file couldn't be played, but if we ask ffmpeg to quit it
+ # produces a file that is playable (this is mostly useful for live
+ # streams). Note that Windows is not affected and produces playable
+ # files (see https://github.com/ytdl-org/youtube-dl/issues/8300).
+ if isinstance(e, KeyboardInterrupt) and sys.platform != 'win32' and url not in ('-', 'pipe:'):
+ proc.communicate_or_kill(b'q')
+ else:
+ proc.kill(timeout=None)
+ raise
+ return retval
class AVconvFD(FFmpegFD):
pass
-_BY_NAME = dict(
- (klass.get_basename(), klass)
+_BY_NAME = {
+ klass.get_basename(): klass
for name, klass in globals().items()
- if name.endswith('FD') and name != 'ExternalFD'
-)
+ if name.endswith('FD') and name not in ('ExternalFD', 'FragmentFD')
+}
+
+_BY_EXE = {klass.EXE_NAME: klass for klass in _BY_NAME.values()}
def list_external_downloaders():
downloader . """
# Drop .exe extension on Windows
bn = os.path.splitext(os.path.basename(external_downloader))[0]
- return _BY_NAME[bn]
+ return _BY_NAME.get(bn, _BY_EXE.get(bn))