7 from .fragment
import FragmentFD
8 from ..compat
import functools
9 from ..compat
import compat_setenv
, compat_str
10 from ..postprocessor
.ffmpeg
import EXT_TO_OUT_FORMATS
, FFmpegPostProcessor
22 handle_youtubedl_headers
,
28 class ExternalFD(FragmentFD
):
29 SUPPORTED_PROTOCOLS
= ('http', 'https', 'ftp', 'ftps')
30 can_download_to_stdout
= False
32 def real_download(self
, filename
, info_dict
):
33 self
.report_destination(filename
)
34 tmpfilename
= self
.temp_name(filename
)
38 retval
= self
._call
_downloader
(tmpfilename
, info_dict
)
39 except KeyboardInterrupt:
40 if not info_dict
.get('is_live'):
42 # Live stream downloading cancellation should be considered as
43 # correct and expected termination thus all postprocessing
46 self
.to_screen('[%s] Interrupted by user' % self
.get_basename())
52 'elapsed': time
.time() - started
,
55 fsize
= os
.path
.getsize(encodeFilename(tmpfilename
))
56 self
.to_screen(f
'\r[{self.get_basename()}] Downloaded {fsize} bytes')
57 self
.try_rename(tmpfilename
, filename
)
59 'downloaded_bytes': fsize
,
62 self
._hook
_progress
(status
, info_dict
)
66 self
.report_error('%s exited with code %d' % (
67 self
.get_basename(), retval
))
71 def get_basename(cls
):
72 return cls
.__name
__[:-2].lower()
76 return cls
.get_basename()
78 @functools.cached_property
83 def available(cls
, path
=None):
84 path
= check_executable(
85 cls
.EXE_NAME
if path
in (None, cls
.get_basename()) else path
,
93 def supports(cls
, info_dict
):
95 (cls
.can_download_to_stdout
or not info_dict
.get('to_stdout'))
96 and info_dict
['protocol'] in cls
.SUPPORTED_PROTOCOLS
)
99 def can_download(cls
, info_dict
, path
=None):
100 return cls
.available(path
) and cls
.supports(info_dict
)
102 def _option(self
, command_option
, param
):
103 return cli_option(self
.params
, command_option
, param
)
105 def _bool_option(self
, command_option
, param
, true_value
='true', false_value
='false', separator
=None):
106 return cli_bool_option(self
.params
, command_option
, param
, true_value
, false_value
, separator
)
108 def _valueless_option(self
, command_option
, param
, expected_value
=True):
109 return cli_valueless_option(self
.params
, command_option
, param
, expected_value
)
111 def _configuration_args(self
, keys
=None, *args
, **kwargs
):
112 return _configuration_args(
113 self
.get_basename(), self
.params
.get('external_downloader_args'), self
.EXE_NAME
,
114 keys
, *args
, **kwargs
)
116 def _call_downloader(self
, tmpfilename
, info_dict
):
117 """ Either overwrite this or implement _make_cmd """
118 cmd
= [encodeArgument(a
) for a
in self
._make
_cmd
(tmpfilename
, info_dict
)]
122 if 'fragments' not in info_dict
:
123 p
= Popen(cmd
, stderr
=subprocess
.PIPE
)
124 _
, stderr
= p
.communicate_or_kill()
125 if p
.returncode
!= 0:
126 self
.to_stderr(stderr
.decode('utf-8', 'replace'))
129 fragment_retries
= self
.params
.get('fragment_retries', 0)
130 skip_unavailable_fragments
= self
.params
.get('skip_unavailable_fragments', True)
133 while count
<= fragment_retries
:
134 p
= Popen(cmd
, stderr
=subprocess
.PIPE
)
135 _
, stderr
= p
.communicate_or_kill()
136 if p
.returncode
== 0:
138 # TODO: Decide whether to retry based on error code
139 # https://aria2.github.io/manual/en/html/aria2c.html#exit-status
140 self
.to_stderr(stderr
.decode('utf-8', 'replace'))
142 if count
<= fragment_retries
:
144 '[%s] Got error. Retrying fragments (attempt %d of %s)...'
145 % (self
.get_basename(), count
, self
.format_retries(fragment_retries
)))
146 self
.sleep_retry('fragment', count
)
147 if count
> fragment_retries
:
148 if not skip_unavailable_fragments
:
149 self
.report_error('Giving up after %s fragment retries' % fragment_retries
)
152 decrypt_fragment
= self
.decrypter(info_dict
)
153 dest
, _
= self
.sanitize_open(tmpfilename
, 'wb')
154 for frag_index
, fragment
in enumerate(info_dict
['fragments']):
155 fragment_filename
= '%s-Frag%d' % (tmpfilename
, frag_index
)
157 src
, _
= self
.sanitize_open(fragment_filename
, 'rb')
158 except OSError as err
:
159 if skip_unavailable_fragments
and frag_index
> 1:
160 self
.report_skip_fragment(frag_index
, err
)
162 self
.report_error(f
'Unable to open fragment {frag_index}; {err}')
164 dest
.write(decrypt_fragment(fragment
, src
.read()))
166 if not self
.params
.get('keep_fragments', False):
167 self
.try_remove(encodeFilename(fragment_filename
))
169 self
.try_remove(encodeFilename('%s.frag.urls' % tmpfilename
))
173 class CurlFD(ExternalFD
):
176 def _make_cmd(self
, tmpfilename
, info_dict
):
177 cmd
= [self
.exe
, '--location', '-o', tmpfilename
, '--compressed']
178 if info_dict
.get('http_headers') is not None:
179 for key
, val
in info_dict
['http_headers'].items():
180 cmd
+= ['--header', f
'{key}: {val}']
182 cmd
+= self
._bool
_option
('--continue-at', 'continuedl', '-', '0')
183 cmd
+= self
._valueless
_option
('--silent', 'noprogress')
184 cmd
+= self
._valueless
_option
('--verbose', 'verbose')
185 cmd
+= self
._option
('--limit-rate', 'ratelimit')
186 retry
= self
._option
('--retry', 'retries')
188 if retry
[1] in ('inf', 'infinite'):
189 retry
[1] = '2147483647'
191 cmd
+= self
._option
('--max-filesize', 'max_filesize')
192 cmd
+= self
._option
('--interface', 'source_address')
193 cmd
+= self
._option
('--proxy', 'proxy')
194 cmd
+= self
._valueless
_option
('--insecure', 'nocheckcertificate')
195 cmd
+= self
._configuration
_args
()
196 cmd
+= ['--', info_dict
['url']]
199 def _call_downloader(self
, tmpfilename
, info_dict
):
200 cmd
= [encodeArgument(a
) for a
in self
._make
_cmd
(tmpfilename
, info_dict
)]
204 # curl writes the progress to stderr so don't capture it.
206 p
.communicate_or_kill()
210 class AxelFD(ExternalFD
):
213 def _make_cmd(self
, tmpfilename
, info_dict
):
214 cmd
= [self
.exe
, '-o', tmpfilename
]
215 if info_dict
.get('http_headers') is not None:
216 for key
, val
in info_dict
['http_headers'].items():
217 cmd
+= ['-H', f
'{key}: {val}']
218 cmd
+= self
._configuration
_args
()
219 cmd
+= ['--', info_dict
['url']]
223 class WgetFD(ExternalFD
):
224 AVAILABLE_OPT
= '--version'
226 def _make_cmd(self
, tmpfilename
, info_dict
):
227 cmd
= [self
.exe
, '-O', tmpfilename
, '-nv', '--no-cookies', '--compression=auto']
228 if info_dict
.get('http_headers') is not None:
229 for key
, val
in info_dict
['http_headers'].items():
230 cmd
+= ['--header', f
'{key}: {val}']
231 cmd
+= self
._option
('--limit-rate', 'ratelimit')
232 retry
= self
._option
('--tries', 'retries')
234 if retry
[1] in ('inf', 'infinite'):
237 cmd
+= self
._option
('--bind-address', 'source_address')
238 proxy
= self
.params
.get('proxy')
240 for var
in ('http_proxy', 'https_proxy'):
241 cmd
+= ['--execute', f
'{var}={proxy}']
242 cmd
+= self
._valueless
_option
('--no-check-certificate', 'nocheckcertificate')
243 cmd
+= self
._configuration
_args
()
244 cmd
+= ['--', info_dict
['url']]
248 class Aria2cFD(ExternalFD
):
250 SUPPORTED_PROTOCOLS
= ('http', 'https', 'ftp', 'ftps', 'dash_frag_urls', 'm3u8_frag_urls')
253 def supports_manifest(manifest
):
254 UNSUPPORTED_FEATURES
= [
255 r
'#EXT-X-BYTERANGE', # playlists composed of byte ranges of media files [1]
256 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.2
258 check_results
= (not re
.search(feature
, manifest
) for feature
in UNSUPPORTED_FEATURES
)
259 return all(check_results
)
261 def _make_cmd(self
, tmpfilename
, info_dict
):
262 cmd
= [self
.exe
, '-c',
263 '--console-log-level=warn', '--summary-interval=0', '--download-result=hide',
264 '--http-accept-gzip=true', '--file-allocation=none', '-x16', '-j16', '-s16']
265 if 'fragments' in info_dict
:
266 cmd
+= ['--allow-overwrite=true', '--allow-piece-length-change=true']
268 cmd
+= ['--min-split-size', '1M']
270 if info_dict
.get('http_headers') is not None:
271 for key
, val
in info_dict
['http_headers'].items():
272 cmd
+= ['--header', f
'{key}: {val}']
273 cmd
+= self
._option
('--max-overall-download-limit', 'ratelimit')
274 cmd
+= self
._option
('--interface', 'source_address')
275 cmd
+= self
._option
('--all-proxy', 'proxy')
276 cmd
+= self
._bool
_option
('--check-certificate', 'nocheckcertificate', 'false', 'true', '=')
277 cmd
+= self
._bool
_option
('--remote-time', 'updatetime', 'true', 'false', '=')
278 cmd
+= self
._bool
_option
('--show-console-readout', 'noprogress', 'false', 'true', '=')
279 cmd
+= self
._configuration
_args
()
281 # aria2c strips out spaces from the beginning/end of filenames and paths.
282 # We work around this issue by adding a "./" to the beginning of the
283 # filename and relative path, and adding a "/" at the end of the path.
284 # See: https://github.com/yt-dlp/yt-dlp/issues/276
285 # https://github.com/ytdl-org/youtube-dl/issues/20312
286 # https://github.com/aria2/aria2/issues/1373
287 dn
= os
.path
.dirname(tmpfilename
)
289 if not os
.path
.isabs(dn
):
290 dn
= f
'.{os.path.sep}{dn}'
291 cmd
+= ['--dir', dn
+ os
.path
.sep
]
292 if 'fragments' not in info_dict
:
293 cmd
+= ['--out', f
'.{os.path.sep}{os.path.basename(tmpfilename)}']
294 cmd
+= ['--auto-file-renaming=false']
296 if 'fragments' in info_dict
:
297 cmd
+= ['--file-allocation=none', '--uri-selector=inorder']
298 url_list_file
= '%s.frag.urls' % tmpfilename
300 for frag_index
, fragment
in enumerate(info_dict
['fragments']):
301 fragment_filename
= '%s-Frag%d' % (os
.path
.basename(tmpfilename
), frag_index
)
302 url_list
.append('%s\n\tout=%s' % (fragment
['url'], fragment_filename
))
303 stream
, _
= self
.sanitize_open(url_list_file
, 'wb')
304 stream
.write('\n'.join(url_list
).encode())
306 cmd
+= ['-i', url_list_file
]
308 cmd
+= ['--', info_dict
['url']]
312 class HttpieFD(ExternalFD
):
313 AVAILABLE_OPT
= '--version'
316 def _make_cmd(self
, tmpfilename
, info_dict
):
317 cmd
= ['http', '--download', '--output', tmpfilename
, info_dict
['url']]
319 if info_dict
.get('http_headers') is not None:
320 for key
, val
in info_dict
['http_headers'].items():
321 cmd
+= [f
'{key}:{val}']
325 class FFmpegFD(ExternalFD
):
326 SUPPORTED_PROTOCOLS
= ('http', 'https', 'ftp', 'ftps', 'm3u8', 'm3u8_native', 'rtsp', 'rtmp', 'rtmp_ffmpeg', 'mms', 'http_dash_segments')
327 can_download_to_stdout
= True
330 def available(cls
, path
=None):
331 # TODO: Fix path for ffmpeg
332 # Fixme: This may be wrong when --ffmpeg-location is used
333 return FFmpegPostProcessor().available
336 def supports(cls
, info_dict
):
337 return all(proto
in cls
.SUPPORTED_PROTOCOLS
for proto
in info_dict
['protocol'].split('+'))
339 def on_process_started(self
, proc
, stdin
):
340 """ Override this in subclasses """
344 def can_merge_formats(cls
, info_dict
, params
):
346 info_dict
.get('requested_formats')
347 and info_dict
.get('protocol')
348 and not params
.get('allow_unplayable_formats')
349 and 'no-direct-merge' not in params
.get('compat_opts', [])
350 and cls
.can_download(info_dict
))
352 def _call_downloader(self
, tmpfilename
, info_dict
):
353 urls
= [f
['url'] for f
in info_dict
.get('requested_formats', [])] or [info_dict
['url']]
354 ffpp
= FFmpegPostProcessor(downloader
=self
)
355 if not ffpp
.available
:
356 self
.report_error('m3u8 download detected but ffmpeg could not be found. Please install')
360 args
= [ffpp
.executable
, '-y']
362 for log_level
in ('quiet', 'verbose'):
363 if self
.params
.get(log_level
, False):
364 args
+= ['-loglevel', log_level
]
366 if not self
.params
.get('verbose'):
367 args
+= ['-hide_banner']
369 args
+= traverse_obj(info_dict
, ('downloader_options', 'ffmpeg_args'), default
=[])
371 # These exists only for compatibility. Extractors should use
372 # info_dict['downloader_options']['ffmpeg_args'] instead
373 args
+= info_dict
.get('_ffmpeg_args') or []
374 seekable
= info_dict
.get('_seekable')
375 if seekable
is not None:
376 # setting -seekable prevents ffmpeg from guessing if the server
377 # supports seeking(by adding the header `Range: bytes=0-`), which
378 # can cause problems in some cases
379 # https://github.com/ytdl-org/youtube-dl/issues/11800#issuecomment-275037127
380 # http://trac.ffmpeg.org/ticket/6125#comment:10
381 args
+= ['-seekable', '1' if seekable
else '0']
383 # start_time = info_dict.get('start_time') or 0
385 # args += ['-ss', compat_str(start_time)]
386 # end_time = info_dict.get('end_time')
388 # args += ['-t', compat_str(end_time - start_time)]
391 if info_dict
.get('http_headers'):
392 youtubedl_headers
= handle_youtubedl_headers(info_dict
['http_headers'])
394 # Trailing \r\n after each HTTP header is important to prevent warning from ffmpeg/avconv:
395 # [http @ 00000000003d2fa0] No trailing CRLF found in HTTP header.
397 ''.join(f
'{key}: {val}\r\n' for key
, val
in youtubedl_headers
.items())
401 proxy
= self
.params
.get('proxy')
403 if not re
.match(r
'^[\da-zA-Z]+://', proxy
):
404 proxy
= 'http://%s' % proxy
406 if proxy
.startswith('socks'):
408 '%s does not support SOCKS proxies. Downloading is likely to fail. '
409 'Consider adding --hls-prefer-native to your command.' % self
.get_basename())
411 # Since December 2015 ffmpeg supports -http_proxy option (see
412 # http://git.videolan.org/?p=ffmpeg.git;a=commit;h=b4eb1f29ebddd60c41a2eb39f5af701e38e0d3fd)
413 # We could switch to the following code if we are able to detect version properly
414 # args += ['-http_proxy', proxy]
415 env
= os
.environ
.copy()
416 compat_setenv('HTTP_PROXY', proxy
, env
=env
)
417 compat_setenv('http_proxy', proxy
, env
=env
)
419 protocol
= info_dict
.get('protocol')
421 if protocol
== 'rtmp':
422 player_url
= info_dict
.get('player_url')
423 page_url
= info_dict
.get('page_url')
424 app
= info_dict
.get('app')
425 play_path
= info_dict
.get('play_path')
426 tc_url
= info_dict
.get('tc_url')
427 flash_version
= info_dict
.get('flash_version')
428 live
= info_dict
.get('rtmp_live', False)
429 conn
= info_dict
.get('rtmp_conn')
430 if player_url
is not None:
431 args
+= ['-rtmp_swfverify', player_url
]
432 if page_url
is not None:
433 args
+= ['-rtmp_pageurl', page_url
]
435 args
+= ['-rtmp_app', app
]
436 if play_path
is not None:
437 args
+= ['-rtmp_playpath', play_path
]
438 if tc_url
is not None:
439 args
+= ['-rtmp_tcurl', tc_url
]
440 if flash_version
is not None:
441 args
+= ['-rtmp_flashver', flash_version
]
443 args
+= ['-rtmp_live', 'live']
444 if isinstance(conn
, list):
446 args
+= ['-rtmp_conn', entry
]
447 elif isinstance(conn
, compat_str
):
448 args
+= ['-rtmp_conn', conn
]
450 for i
, url
in enumerate(urls
):
451 # We need to specify headers for each http input stream
452 # otherwise, it will only be applied to the first.
453 # https://github.com/yt-dlp/yt-dlp/issues/2696
454 if http_headers
is not None and re
.match(r
'^https?://', url
):
456 args
+= self
._configuration
_args
((f
'_i{i + 1}', '_i')) + ['-i', url
]
458 args
+= ['-c', 'copy']
459 if info_dict
.get('requested_formats') or protocol
== 'http_dash_segments':
460 for (i
, fmt
) in enumerate(info_dict
.get('requested_formats') or [info_dict
]):
461 stream_number
= fmt
.get('manifest_stream_number', 0)
462 args
.extend(['-map', f
'{i}:{stream_number}'])
464 if self
.params
.get('test', False):
465 args
+= ['-fs', compat_str(self
._TEST
_FILE
_SIZE
)]
467 ext
= info_dict
['ext']
468 if protocol
in ('m3u8', 'm3u8_native'):
469 use_mpegts
= (tmpfilename
== '-') or self
.params
.get('hls_use_mpegts')
470 if use_mpegts
is None:
471 use_mpegts
= info_dict
.get('is_live')
473 args
+= ['-f', 'mpegts']
475 args
+= ['-f', 'mp4']
476 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')):
477 args
+= ['-bsf:a', 'aac_adtstoasc']
478 elif protocol
== 'rtmp':
479 args
+= ['-f', 'flv']
480 elif ext
== 'mp4' and tmpfilename
== '-':
481 args
+= ['-f', 'mpegts']
482 elif ext
== 'unknown_video':
483 ext
= determine_ext(remove_end(tmpfilename
, '.part'))
484 if ext
== 'unknown_video':
486 'The video format is unknown and cannot be downloaded by ffmpeg. '
487 'Explicitly set the extension in the filename to attempt download in that format')
489 self
.report_warning(f
'The video format is unknown. Trying to download as {ext} according to the filename')
490 args
+= ['-f', EXT_TO_OUT_FORMATS
.get(ext
, ext
)]
492 args
+= ['-f', EXT_TO_OUT_FORMATS
.get(ext
, ext
)]
494 args
+= self
._configuration
_args
(('_o1', '_o', ''))
496 args
= [encodeArgument(opt
) for opt
in args
]
497 args
.append(encodeFilename(ffpp
._ffmpeg
_filename
_argument
(tmpfilename
), True))
498 self
._debug
_cmd
(args
)
500 proc
= Popen(args
, stdin
=subprocess
.PIPE
, env
=env
)
501 if url
in ('-', 'pipe:'):
502 self
.on_process_started(proc
, proc
.stdin
)
505 except BaseException
as e
:
506 # subprocces.run would send the SIGKILL signal to ffmpeg and the
507 # mp4 file couldn't be played, but if we ask ffmpeg to quit it
508 # produces a file that is playable (this is mostly useful for live
509 # streams). Note that Windows is not affected and produces playable
510 # files (see https://github.com/ytdl-org/youtube-dl/issues/8300).
511 if isinstance(e
, KeyboardInterrupt) and sys
.platform
!= 'win32' and url
not in ('-', 'pipe:'):
512 proc
.communicate_or_kill(b
'q')
520 class AVconvFD(FFmpegFD
):
525 klass
.get_basename(): klass
526 for name
, klass
in globals().items()
527 if name
.endswith('FD') and name
not in ('ExternalFD', 'FragmentFD')
530 _BY_EXE
= {klass.EXE_NAME: klass for klass in _BY_NAME.values()}
533 def list_external_downloaders():
534 return sorted(_BY_NAME
.keys())
537 def get_external_downloader(external_downloader
):
538 """ Given the name of the executable, see whether we support the given
540 # Drop .exe extension on Windows
541 bn
= os
.path
.splitext(os
.path
.basename(external_downloader
))[0]
542 return _BY_NAME
.get(bn
, _BY_EXE
.get(bn
))