46 import xml
.etree
.ElementTree
49 from .compat
import functools
# isort: split
51 compat_etree_fromstring
,
53 compat_HTMLParseError
,
57 from .dependencies
import brotli
, certifi
, websockets
, xattr
58 from .socks
import ProxyType
, sockssocket
61 def register_socks_protocols():
62 # "Register" SOCKS protocols
63 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
64 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
65 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
66 if scheme
not in urllib
.parse
.uses_netloc
:
67 urllib
.parse
.uses_netloc
.append(scheme
)
70 # This is not clearly defined otherwise
71 compiled_regex_type
= type(re
.compile(''))
74 def random_user_agent():
75 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
116 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
119 SUPPORTED_ENCODINGS
= [
123 SUPPORTED_ENCODINGS
.append('br')
126 'User-Agent': random_user_agent(),
127 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
128 'Accept-Language': 'en-us,en;q=0.5',
129 'Sec-Fetch-Mode': 'navigate',
134 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
138 NO_DEFAULT
= object()
139 IDENTITY
= lambda x
: x
141 ENGLISH_MONTH_NAMES
= [
142 'January', 'February', 'March', 'April', 'May', 'June',
143 'July', 'August', 'September', 'October', 'November', 'December']
146 'en': ENGLISH_MONTH_NAMES
,
148 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
149 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
152 # needed for sanitizing filenames in restricted mode
153 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
154 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
155 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
185 '%Y-%m-%d %H:%M:%S.%f',
186 '%Y-%m-%d %H:%M:%S:%f',
189 '%Y-%m-%dT%H:%M:%SZ',
190 '%Y-%m-%dT%H:%M:%S.%fZ',
191 '%Y-%m-%dT%H:%M:%S.%f0Z',
193 '%Y-%m-%dT%H:%M:%S.%f',
196 '%b %d %Y at %H:%M:%S',
198 '%B %d %Y at %H:%M:%S',
202 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
203 DATE_FORMATS_DAY_FIRST
.extend([
213 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
214 DATE_FORMATS_MONTH_FIRST
.extend([
222 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
223 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}
)\s
*</script
>'
225 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
229 def preferredencoding():
230 """Get preferred encoding.
232 Returns the best encoding scheme for the system, based on
233 locale.getpreferredencoding() and some further tweaks.
236 pref = locale.getpreferredencoding()
244 def write_json_file(obj, fn):
245 """ Encode obj as JSON and write it to fn, atomically if possible """
247 tf = tempfile.NamedTemporaryFile(
248 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
249 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
253 json.dump(obj, tf, ensure_ascii=False)
254 if sys.platform == 'win32
':
255 # Need to remove existing file on Windows, else os.rename raises
256 # WindowsError or FileExistsError.
257 with contextlib.suppress(OSError):
259 with contextlib.suppress(OSError):
262 os.chmod(tf.name, 0o666 & ~mask)
263 os.rename(tf.name, fn)
265 with contextlib.suppress(OSError):
270 def find_xpath_attr(node, xpath, key, val=None):
271 """ Find the xpath xpath[@key=val] """
272 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
273 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
274 return node.find(expr)
276 # On python2.6 the xml.etree.ElementTree.Element methods don't support
277 # the namespace parameter
280 def xpath_with_ns(path
, ns_map
):
281 components
= [c
.split(':') for c
in path
.split('/')]
285 replaced
.append(c
[0])
288 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
289 return '/'.join(replaced
)
292 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
293 def _find_xpath(xpath
):
294 return node
.find(xpath
)
296 if isinstance(xpath
, str):
297 n
= _find_xpath(xpath
)
305 if default
is not NO_DEFAULT
:
308 name
= xpath
if name
is None else name
309 raise ExtractorError('Could not find XML element %s' % name
)
315 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
316 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
317 if n
is None or n
== default
:
320 if default
is not NO_DEFAULT
:
323 name
= xpath
if name
is None else name
324 raise ExtractorError('Could not find XML element\'s text %s' % name
)
330 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
331 n
= find_xpath_attr(node
, xpath
, key
)
333 if default
is not NO_DEFAULT
:
336 name
= f
'{xpath}[@{key}]' if name
is None else name
337 raise ExtractorError('Could not find XML attribute %s' % name
)
343 def get_element_by_id(id, html
, **kwargs
):
344 """Return the content of the tag with the specified ID in the passed HTML document"""
345 return get_element_by_attribute('id', id, html
, **kwargs
)
348 def get_element_html_by_id(id, html
, **kwargs
):
349 """Return the html of the tag with the specified ID in the passed HTML document"""
350 return get_element_html_by_attribute('id', id, html
, **kwargs
)
353 def get_element_by_class(class_name
, html
):
354 """Return the content of the first tag with the specified class in the passed HTML document"""
355 retval
= get_elements_by_class(class_name
, html
)
356 return retval
[0] if retval
else None
359 def get_element_html_by_class(class_name
, html
):
360 """Return the html of the first tag with the specified class in the passed HTML document"""
361 retval
= get_elements_html_by_class(class_name
, html
)
362 return retval
[0] if retval
else None
365 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
366 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
367 return retval
[0] if retval
else None
370 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
371 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
372 return retval
[0] if retval
else None
375 def get_elements_by_class(class_name
, html
, **kargs
):
376 """Return the content of all tags with the specified class in the passed HTML document as a list"""
377 return get_elements_by_attribute(
378 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
379 html, escape_value=False)
382 def get_elements_html_by_class(class_name, html):
383 """Return the html of all tags with the specified class in the passed HTML document as a list"""
384 return get_elements_html_by_attribute(
385 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
386 html, escape_value=False)
389 def get_elements_by_attribute(*args, **kwargs):
390 """Return the content of the tag with the specified attribute in the passed HTML document"""
391 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
394 def get_elements_html_by_attribute(*args, **kwargs):
395 """Return the html of the tag with the specified attribute in the passed HTML document"""
396 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
399 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
401 Return the text (content) and the html (whole) of the tag with the specified
402 attribute in the passed HTML document
405 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
407 value = re.escape(value) if escape_value else value
409 partial_element_re = rf'''(?x
)
410 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
411 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
412 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
415 for m in re.finditer(partial_element_re, html):
416 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
419 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
424 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
426 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
427 closing tag for the first opening tag it has encountered, and can be used
431 class HTMLBreakOnClosingTagException(Exception):
435 self.tagstack = collections.deque()
436 html.parser.HTMLParser.__init__(self)
441 def __exit__(self, *_):
445 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
446 # so data remains buffered; we no longer have any interest in it, thus
447 # override this method to discard it
450 def handle_starttag(self, tag, _):
451 self.tagstack.append(tag)
453 def handle_endtag(self, tag):
454 if not self.tagstack:
455 raise compat_HTMLParseError('no tags
in the stack
')
457 inner_tag = self.tagstack.pop()
461 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
462 if not self.tagstack:
463 raise self.HTMLBreakOnClosingTagException()
466 def get_element_text_and_html_by_tag(tag, html):
468 For the first element with the specified tag in the passed HTML document
469 return its' content (text
) and the whole
element (html
)
471 def find_or_raise(haystack, needle, exc):
473 return haystack.index(needle)
476 closing_tag = f'</{tag}>'
477 whole_start = find_or_raise(
478 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
479 content_start = find_or_raise(
480 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
481 content_start += whole_start + 1
482 with HTMLBreakOnClosingTagParser() as parser:
483 parser.feed(html[whole_start:content_start])
484 if not parser.tagstack or parser.tagstack[0] != tag:
485 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
486 offset = content_start
487 while offset < len(html):
488 next_closing_tag_start = find_or_raise(
489 html[offset:], closing_tag,
490 compat_HTMLParseError(f'closing {tag} tag not found'))
491 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
493 parser.feed(html[offset:offset + next_closing_tag_end])
494 offset += next_closing_tag_end
495 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
496 return html[content_start:offset + next_closing_tag_start], \
497 html[whole_start:offset + next_closing_tag_end]
498 raise compat_HTMLParseError('unexpected end of html')
501 class HTMLAttributeParser(html.parser.HTMLParser):
502 """Trivial HTML parser to gather the attributes
for a single element
"""
506 html.parser.HTMLParser.__init__(self)
508 def handle_starttag(self, tag, attrs):
509 self.attrs = dict(attrs)
512 class HTMLListAttrsParser(html.parser.HTMLParser):
513 """HTML parser to gather the attributes
for the elements of a
list"""
516 html.parser.HTMLParser.__init__(self)
520 def handle_starttag(self, tag, attrs):
521 if tag == 'li' and self._level == 0:
522 self.items.append(dict(attrs))
525 def handle_endtag(self, tag):
529 def extract_attributes(html_element):
530 """Given a string
for an HTML element such
as
532 a
="foo" B
="bar" c
="&98;az" d
=boz
533 empty
= noval entity
="&"
536 Decode
and return a dictionary of attributes
.
538 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
539 'empty': '', 'noval': None, 'entity': '&',
540 'sq': '"', 'dq': '\''
543 parser = HTMLAttributeParser()
544 with contextlib.suppress(compat_HTMLParseError):
545 parser.feed(html_element)
550 def parse_list(webpage):
551 """Given a string
for an series of HTML
<li
> elements
,
552 return a dictionary of their attributes
"""
553 parser = HTMLListAttrsParser()
559 def clean_html(html):
560 """Clean an HTML snippet into a readable string
"""
562 if html is None: # Convenience for sanitizing descriptions etc.
565 html = re.sub(r'\s+', ' ', html)
566 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
567 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
569 html = re.sub('<.*?>', '', html)
570 # Replace html entities
571 html = unescapeHTML(html)
575 class LenientJSONDecoder(json.JSONDecoder):
576 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
577 self.transform_source, self.ignore_extra = transform_source, ignore_extra
578 super().__init__(*args, **kwargs)
581 if self.transform_source:
582 s = self.transform_source(s)
583 if self.ignore_extra:
584 return self.raw_decode(s.lstrip())[0]
585 return super().decode(s)
588 def sanitize_open(filename, open_mode):
589 """Try to
open the given filename
, and slightly tweak it
if this fails
.
591 Attempts to
open the given filename
. If this fails
, it tries to change
592 the filename slightly
, step by step
, until it
's either able to open it
593 or it fails and raises a final exception, like the standard open()
596 It returns the tuple (stream, definitive_file_name).
599 if sys.platform == 'win32
':
602 # stdout may be any IO stream. Eg, when using contextlib.redirect_stdout
603 with contextlib.suppress(io.UnsupportedOperation):
604 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
605 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
607 for attempt in range(2):
610 if sys.platform == 'win32
':
611 # FIXME: An exclusive lock also locks the file from being read.
612 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
613 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
614 raise LockingUnsupportedError()
615 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
617 stream
= open(filename
, open_mode
)
618 return stream
, filename
619 except OSError as err
:
620 if attempt
or err
.errno
in (errno
.EACCES
,):
622 old_filename
, filename
= filename
, sanitize_path(filename
)
623 if old_filename
== filename
:
627 def timeconvert(timestr
):
628 """Convert RFC 2822 defined time string into system timestamp"""
630 timetuple
= email
.utils
.parsedate_tz(timestr
)
631 if timetuple
is not None:
632 timestamp
= email
.utils
.mktime_tz(timetuple
)
636 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
637 """Sanitizes a string so it could be used as part of a filename.
638 @param restricted Use a stricter subset of allowed characters
639 @param is_id Whether this is an ID that should be kept unchanged if possible.
640 If unset, yt-dlp's new sanitization rules are in effect
645 def replace_insane(char
):
646 if restricted
and char
in ACCENT_CHARS
:
647 return ACCENT_CHARS
[char
]
648 elif not restricted
and char
== '\n':
650 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
653 return '' if restricted
else '\''
655 return '\0_\0-' if restricted
else '\0 \0-'
656 elif char
in '\\/|*<>':
658 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
662 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
663 result
= ''.join(map(replace_insane
, s
))
664 if is_id
is NO_DEFAULT
:
665 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
666 STRIP_RE
= r
'(?:\0.|[ _-])*'
667 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
668 result
= result
.replace('\0', '') or '_'
671 while '__' in result
:
672 result
= result
.replace('__', '_')
673 result
= result
.strip('_')
674 # Common case of "Foreign band name - English song title"
675 if restricted
and result
.startswith('-_'):
677 if result
.startswith('-'):
678 result
= '_' + result
[len('-'):]
679 result
= result
.lstrip('.')
685 def sanitize_path(s
, force
=False):
686 """Sanitizes and normalizes path on Windows"""
687 if sys
.platform
== 'win32':
689 drive_or_unc
, _
= os
.path
.splitdrive(s
)
695 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
699 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
700 for path_part
in norm_path
]
702 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
703 elif force
and s
and s
[0] == os
.path
.sep
:
704 sanitized_path
.insert(0, os
.path
.sep
)
705 return os
.path
.join(*sanitized_path
)
708 def sanitize_url(url
, *, scheme
='http'):
709 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
710 # the number of unwanted failures due to missing protocol
713 elif url
.startswith('//'):
714 return f
'{scheme}:{url}'
715 # Fix some common typos seen so far
717 # https://github.com/ytdl-org/youtube-dl/issues/15649
718 (r
'^httpss://', r
'https://'),
719 # https://bx1.be/lives/direct-tv/
720 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
722 for mistake
, fixup
in COMMON_TYPOS
:
723 if re
.match(mistake
, url
):
724 return re
.sub(mistake
, fixup
, url
)
728 def extract_basic_auth(url
):
729 parts
= urllib
.parse
.urlsplit(url
)
730 if parts
.username
is None:
732 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
733 parts
.hostname
if parts
.port
is None
734 else '%s:%d' % (parts
.hostname
, parts
.port
))))
735 auth_payload
= base64
.b64encode(
736 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
737 return url
, f
'Basic {auth_payload.decode()}'
740 def sanitized_Request(url
, *args
, **kwargs
):
741 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
742 if auth_header
is not None:
743 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
744 headers
['Authorization'] = auth_header
745 return urllib
.request
.Request(url
, *args
, **kwargs
)
749 """Expand shell variables and ~"""
750 return os
.path
.expandvars(compat_expanduser(s
))
753 def orderedSet(iterable
, *, lazy
=False):
754 """Remove all duplicates from the input iterable"""
756 seen
= [] # Do not use set since the items can be unhashable
762 return _iter() if lazy
else list(_iter())
765 def _htmlentity_transform(entity_with_semicolon
):
766 """Transforms an HTML entity to a character."""
767 entity
= entity_with_semicolon
[:-1]
769 # Known non-numeric HTML entity
770 if entity
in html
.entities
.name2codepoint
:
771 return chr(html
.entities
.name2codepoint
[entity
])
773 # TODO: HTML5 allows entities without a semicolon. For example,
774 # 'Éric' should be decoded as 'Éric'.
775 if entity_with_semicolon
in html
.entities
.html5
:
776 return html
.entities
.html5
[entity_with_semicolon
]
778 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
780 numstr
= mobj
.group(1)
781 if numstr
.startswith('x'):
783 numstr
= '0%s' % numstr
786 # See https://github.com/ytdl-org/youtube-dl/issues/7518
787 with contextlib
.suppress(ValueError):
788 return chr(int(numstr
, base
))
790 # Unknown entity in name, return its literal representation
791 return '&%s;' % entity
797 assert isinstance(s
, str)
800 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
803 def escapeHTML(text
):
806 .replace('&', '&')
807 .replace('<', '<')
808 .replace('>', '>')
809 .replace('"', '"')
810 .replace("'", ''')
814 def process_communicate_or_kill(p
, *args
, **kwargs
):
815 write_string('DeprecationWarning: yt_dlp.utils.process_communicate_or_kill is deprecated '
816 'and may be removed in a future version. Use yt_dlp.utils.Popen.communicate_or_kill instead')
817 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
820 class Popen(subprocess
.Popen
):
821 if sys
.platform
== 'win32':
822 _startupinfo
= subprocess
.STARTUPINFO()
823 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
827 def __init__(self
, *args
, text
=False, **kwargs
):
829 kwargs
['universal_newlines'] = True # For 3.6 compatibility
830 kwargs
.setdefault('encoding', 'utf-8')
831 kwargs
.setdefault('errors', 'replace')
832 super().__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
834 def communicate_or_kill(self
, *args
, **kwargs
):
836 return self
.communicate(*args
, **kwargs
)
837 except BaseException
: # Including KeyboardInterrupt
838 self
.kill(timeout
=None)
841 def kill(self
, *, timeout
=0):
844 self
.wait(timeout
=timeout
)
847 def run(cls
, *args
, **kwargs
):
848 with cls(*args
, **kwargs
) as proc
:
849 stdout
, stderr
= proc
.communicate_or_kill()
850 return stdout
or '', stderr
or '', proc
.returncode
853 def get_subprocess_encoding():
854 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
855 # For subprocess calls, encode with locale encoding
856 # Refer to http://stackoverflow.com/a/9951851/35070
857 encoding
= preferredencoding()
859 encoding
= sys
.getfilesystemencoding()
865 def encodeFilename(s
, for_subprocess
=False):
866 assert isinstance(s
, str)
870 def decodeFilename(b
, for_subprocess
=False):
874 def encodeArgument(s
):
875 # Legacy code that uses byte strings
876 # Uncomment the following line after fixing all post processors
877 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
878 return s
if isinstance(s
, str) else s
.decode('ascii')
881 def decodeArgument(b
):
885 def decodeOption(optval
):
888 if isinstance(optval
, bytes):
889 optval
= optval
.decode(preferredencoding())
891 assert isinstance(optval
, str)
895 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
898 def timetuple_from_msec(msec
):
899 secs
, msec
= divmod(msec
, 1000)
900 mins
, secs
= divmod(secs
, 60)
901 hrs
, mins
= divmod(mins
, 60)
902 return _timetuple(hrs
, mins
, secs
, msec
)
905 def formatSeconds(secs
, delim
=':', msec
=False):
906 time
= timetuple_from_msec(secs
* 1000)
908 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
910 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
912 ret
= '%d' % time
.seconds
913 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
916 def _ssl_load_windows_store_certs(ssl_context
, storename
):
917 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
919 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
920 if encoding
== 'x509_asn' and (
921 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
922 except PermissionError
:
925 with contextlib
.suppress(ssl
.SSLError
):
926 ssl_context
.load_verify_locations(cadata
=cert
)
929 def make_HTTPS_handler(params
, **kwargs
):
930 opts_check_certificate
= not params
.get('nocheckcertificate')
931 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
932 context
.check_hostname
= opts_check_certificate
933 if params
.get('legacyserverconnect'):
934 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
935 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
936 context
.set_ciphers('DEFAULT')
938 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
939 if opts_check_certificate
:
940 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
941 context
.load_verify_locations(cafile
=certifi
.where())
944 context
.load_default_certs()
945 # Work around the issue in load_default_certs when there are bad certificates. See:
946 # https://github.com/yt-dlp/yt-dlp/issues/1060,
947 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
949 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
950 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
951 for storename
in ('CA', 'ROOT'):
952 _ssl_load_windows_store_certs(context
, storename
)
953 context
.set_default_verify_paths()
955 client_certfile
= params
.get('client_certificate')
958 context
.load_cert_chain(
959 client_certfile
, keyfile
=params
.get('client_certificate_key'),
960 password
=params
.get('client_certificate_password'))
962 raise YoutubeDLError('Unable to load client certificate')
964 # Some servers may reject requests if ALPN extension is not sent. See:
965 # https://github.com/python/cpython/issues/85140
966 # https://github.com/yt-dlp/yt-dlp/issues/3878
967 with contextlib
.suppress(NotImplementedError):
968 context
.set_alpn_protocols(['http/1.1'])
970 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
973 def bug_reports_message(before
=';'):
974 from .update
import REPOSITORY
976 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
977 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
979 before
= before
.rstrip()
980 if not before
or before
.endswith(('.', '!', '?')):
981 msg
= msg
[0].title() + msg
[1:]
983 return (before
+ ' ' if before
else '') + msg
986 class YoutubeDLError(Exception):
987 """Base exception for YoutubeDL errors."""
990 def __init__(self
, msg
=None):
993 elif self
.msg
is None:
994 self
.msg
= type(self
).__name
__
995 super().__init
__(self
.msg
)
998 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
999 if hasattr(ssl
, 'CertificateError'):
1000 network_exceptions
.append(ssl
.CertificateError
)
1001 network_exceptions
= tuple(network_exceptions
)
1004 class ExtractorError(YoutubeDLError
):
1005 """Error during info extraction."""
1007 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1008 """ tb, if given, is the original traceback (so that it can be printed out).
1009 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1011 if sys
.exc_info()[0] in network_exceptions
:
1014 self
.orig_msg
= str(msg
)
1016 self
.expected
= expected
1018 self
.video_id
= video_id
1020 self
.exc_info
= sys
.exc_info() # preserve original exception
1021 if isinstance(self
.exc_info
[1], ExtractorError
):
1022 self
.exc_info
= self
.exc_info
[1].exc_info
1024 super().__init
__(''.join((
1025 format_field(ie
, None, '[%s] '),
1026 format_field(video_id
, None, '%s: '),
1028 format_field(cause
, None, ' (caused by %r)'),
1029 '' if expected
else bug_reports_message())))
1031 def format_traceback(self
):
1032 return join_nonempty(
1033 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1034 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1038 class UnsupportedError(ExtractorError
):
1039 def __init__(self
, url
):
1041 'Unsupported URL: %s' % url
, expected
=True)
1045 class RegexNotFoundError(ExtractorError
):
1046 """Error when a regex didn't match"""
1050 class GeoRestrictedError(ExtractorError
):
1051 """Geographic restriction Error exception.
1053 This exception may be thrown when a video is not available from your
1054 geographic location due to geographic restrictions imposed by a website.
1057 def __init__(self
, msg
, countries
=None, **kwargs
):
1058 kwargs
['expected'] = True
1059 super().__init
__(msg
, **kwargs
)
1060 self
.countries
= countries
1063 class UserNotLive(ExtractorError
):
1064 """Error when a channel/user is not live"""
1066 def __init__(self
, msg
=None, **kwargs
):
1067 kwargs
['expected'] = True
1068 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1071 class DownloadError(YoutubeDLError
):
1072 """Download Error exception.
1074 This exception may be thrown by FileDownloader objects if they are not
1075 configured to continue on errors. They will contain the appropriate
1079 def __init__(self
, msg
, exc_info
=None):
1080 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1081 super().__init
__(msg
)
1082 self
.exc_info
= exc_info
1085 class EntryNotInPlaylist(YoutubeDLError
):
1086 """Entry not in playlist exception.
1088 This exception will be thrown by YoutubeDL when a requested entry
1089 is not found in the playlist info_dict
1091 msg
= 'Entry not found in info'
1094 class SameFileError(YoutubeDLError
):
1095 """Same File exception.
1097 This exception will be thrown by FileDownloader objects if they detect
1098 multiple files would have to be downloaded to the same file on disk.
1100 msg
= 'Fixed output name but more than one file to download'
1102 def __init__(self
, filename
=None):
1103 if filename
is not None:
1104 self
.msg
+= f
': {filename}'
1105 super().__init
__(self
.msg
)
1108 class PostProcessingError(YoutubeDLError
):
1109 """Post Processing exception.
1111 This exception may be raised by PostProcessor's .run() method to
1112 indicate an error in the postprocessing task.
1116 class DownloadCancelled(YoutubeDLError
):
1117 """ Exception raised when the download queue should be interrupted """
1118 msg
= 'The download was cancelled'
1121 class ExistingVideoReached(DownloadCancelled
):
1122 """ --break-on-existing triggered """
1123 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1126 class RejectedVideoReached(DownloadCancelled
):
1127 """ --break-on-reject triggered """
1128 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1131 class MaxDownloadsReached(DownloadCancelled
):
1132 """ --max-downloads limit has been reached. """
1133 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1136 class ReExtractInfo(YoutubeDLError
):
1137 """ Video info needs to be re-extracted. """
1139 def __init__(self
, msg
, expected
=False):
1140 super().__init
__(msg
)
1141 self
.expected
= expected
1144 class ThrottledDownload(ReExtractInfo
):
1145 """ Download speed below --throttled-rate. """
1146 msg
= 'The download speed is below throttle limit'
1149 super().__init
__(self
.msg
, expected
=False)
1152 class UnavailableVideoError(YoutubeDLError
):
1153 """Unavailable Format exception.
1155 This exception will be thrown when a video is requested
1156 in a format that is not available for that video.
1158 msg
= 'Unable to download video'
1160 def __init__(self
, err
=None):
1162 self
.msg
+= f
': {err}'
1163 super().__init
__(self
.msg
)
1166 class ContentTooShortError(YoutubeDLError
):
1167 """Content Too Short exception.
1169 This exception may be raised by FileDownloader objects when a file they
1170 download is too small for what the server announced first, indicating
1171 the connection was probably interrupted.
1174 def __init__(self
, downloaded
, expected
):
1175 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1177 self
.downloaded
= downloaded
1178 self
.expected
= expected
1181 class XAttrMetadataError(YoutubeDLError
):
1182 def __init__(self
, code
=None, msg
='Unknown error'):
1183 super().__init
__(msg
)
1187 # Parsing code and msg
1188 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1189 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1190 self
.reason
= 'NO_SPACE'
1191 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1192 self
.reason
= 'VALUE_TOO_LONG'
1194 self
.reason
= 'NOT_SUPPORTED'
1197 class XAttrUnavailableError(YoutubeDLError
):
1201 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1202 hc
= http_class(*args
, **kwargs
)
1203 source_address
= ydl_handler
._params
.get('source_address')
1205 if source_address
is not None:
1206 # This is to workaround _create_connection() from socket where it will try all
1207 # address data from getaddrinfo() including IPv6. This filters the result from
1208 # getaddrinfo() based on the source_address value.
1209 # This is based on the cpython socket.create_connection() function.
1210 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1211 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1212 host
, port
= address
1214 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1215 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1216 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1217 if addrs
and not ip_addrs
:
1218 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1220 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1221 % (ip_version
, source_address
[0]))
1222 for res
in ip_addrs
:
1223 af
, socktype
, proto
, canonname
, sa
= res
1226 sock
= socket
.socket(af
, socktype
, proto
)
1227 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1228 sock
.settimeout(timeout
)
1229 sock
.bind(source_address
)
1231 err
= None # Explicitly break reference cycle
1233 except OSError as _
:
1235 if sock
is not None:
1240 raise OSError('getaddrinfo returns an empty list')
1241 if hasattr(hc
, '_create_connection'):
1242 hc
._create
_connection
= _create_connection
1243 hc
.source_address
= (source_address
, 0)
1248 def handle_youtubedl_headers(headers
):
1249 filtered_headers
= headers
1251 if 'Youtubedl-no-compression' in filtered_headers
:
1252 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1253 del filtered_headers
['Youtubedl-no-compression']
1255 return filtered_headers
1258 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1259 """Handler for HTTP requests and responses.
1261 This class, when installed with an OpenerDirector, automatically adds
1262 the standard headers to every HTTP request and handles gzipped and
1263 deflated responses from web servers. If compression is to be avoided in
1264 a particular request, the original request in the program code only has
1265 to include the HTTP header "Youtubedl-no-compression", which will be
1266 removed before making the real request.
1268 Part of this code was copied from:
1270 http://techknack.net/python-urllib2-handlers/
1272 Andrew Rowls, the author of that code, agreed to release it to the
1276 def __init__(self
, params
, *args
, **kwargs
):
1277 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1278 self
._params
= params
1280 def http_open(self
, req
):
1281 conn_class
= http
.client
.HTTPConnection
1283 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1285 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1286 del req
.headers
['Ytdl-socks-proxy']
1288 return self
.do_open(functools
.partial(
1289 _create_http_connection
, self
, conn_class
, False),
1297 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1299 return zlib
.decompress(data
)
1305 return brotli
.decompress(data
)
1307 def http_request(self
, req
):
1308 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1309 # always respected by websites, some tend to give out URLs with non percent-encoded
1310 # non-ASCII characters (see telemb.py, ard.py [#3412])
1311 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1312 # To work around aforementioned issue we will replace request's original URL with
1313 # percent-encoded one
1314 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1315 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1316 url
= req
.get_full_url()
1317 url_escaped
= escape_url(url
)
1319 # Substitute URL if any change after escaping
1320 if url
!= url_escaped
:
1321 req
= update_Request(req
, url
=url_escaped
)
1323 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1324 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1325 # The dict keys are capitalized because of this bug by urllib
1326 if h
.capitalize() not in req
.headers
:
1327 req
.add_header(h
, v
)
1329 if 'Accept-encoding' not in req
.headers
:
1330 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1332 req
.headers
= handle_youtubedl_headers(req
.headers
)
1334 return super().do_request_(req
)
1336 def http_response(self
, req
, resp
):
1339 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1340 content
= resp
.read()
1341 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1343 uncompressed
= io
.BytesIO(gz
.read())
1344 except OSError as original_ioerror
:
1345 # There may be junk add the end of the file
1346 # See http://stackoverflow.com/q/4928560/35070 for details
1347 for i
in range(1, 1024):
1349 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1350 uncompressed
= io
.BytesIO(gz
.read())
1355 raise original_ioerror
1356 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1357 resp
.msg
= old_resp
.msg
1358 del resp
.headers
['Content-encoding']
1360 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1361 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1362 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1363 resp
.msg
= old_resp
.msg
1364 del resp
.headers
['Content-encoding']
1366 if resp
.headers
.get('Content-encoding', '') == 'br':
1367 resp
= urllib
.request
.addinfourl(
1368 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1369 resp
.msg
= old_resp
.msg
1370 del resp
.headers
['Content-encoding']
1371 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1372 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1373 if 300 <= resp
.code
< 400:
1374 location
= resp
.headers
.get('Location')
1376 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1377 location
= location
.encode('iso-8859-1').decode()
1378 location_escaped
= escape_url(location
)
1379 if location
!= location_escaped
:
1380 del resp
.headers
['Location']
1381 resp
.headers
['Location'] = location_escaped
1384 https_request
= http_request
1385 https_response
= http_response
1388 def make_socks_conn_class(base_class
, socks_proxy
):
1389 assert issubclass(base_class
, (
1390 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1392 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1393 if url_components
.scheme
.lower() == 'socks5':
1394 socks_type
= ProxyType
.SOCKS5
1395 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1396 socks_type
= ProxyType
.SOCKS4
1397 elif url_components
.scheme
.lower() == 'socks4a':
1398 socks_type
= ProxyType
.SOCKS4A
1400 def unquote_if_non_empty(s
):
1403 return urllib
.parse
.unquote_plus(s
)
1407 url_components
.hostname
, url_components
.port
or 1080,
1409 unquote_if_non_empty(url_components
.username
),
1410 unquote_if_non_empty(url_components
.password
),
1413 class SocksConnection(base_class
):
1415 self
.sock
= sockssocket()
1416 self
.sock
.setproxy(*proxy_args
)
1417 if isinstance(self
.timeout
, (int, float)):
1418 self
.sock
.settimeout(self
.timeout
)
1419 self
.sock
.connect((self
.host
, self
.port
))
1421 if isinstance(self
, http
.client
.HTTPSConnection
):
1422 if hasattr(self
, '_context'): # Python > 2.6
1423 self
.sock
= self
._context
.wrap_socket(
1424 self
.sock
, server_hostname
=self
.host
)
1426 self
.sock
= ssl
.wrap_socket(self
.sock
)
1428 return SocksConnection
1431 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1432 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1433 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1434 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1435 self
._params
= params
1437 def https_open(self
, req
):
1439 conn_class
= self
._https
_conn
_class
1441 if hasattr(self
, '_context'): # python > 2.6
1442 kwargs
['context'] = self
._context
1443 if hasattr(self
, '_check_hostname'): # python 3.x
1444 kwargs
['check_hostname'] = self
._check
_hostname
1446 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1448 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1449 del req
.headers
['Ytdl-socks-proxy']
1452 return self
.do_open(
1453 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1454 except urllib
.error
.URLError
as e
:
1455 if (isinstance(e
.reason
, ssl
.SSLError
)
1456 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1457 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1461 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1463 See [1] for cookie file format.
1465 1. https://curl.haxx.se/docs/http-cookies.html
1467 _HTTPONLY_PREFIX
= '#HttpOnly_'
1469 _HEADER
= '''# Netscape HTTP Cookie File
1470 # This file is generated by yt-dlp. Do not edit.
1473 _CookieFileEntry
= collections
.namedtuple(
1475 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1477 def __init__(self
, filename
=None, *args
, **kwargs
):
1478 super().__init
__(None, *args
, **kwargs
)
1479 if self
.is_path(filename
):
1480 filename
= os
.fspath(filename
)
1481 self
.filename
= filename
1484 def _true_or_false(cndn
):
1485 return 'TRUE' if cndn
else 'FALSE'
1489 return isinstance(file, (str, bytes, os
.PathLike
))
1491 @contextlib.contextmanager
1492 def open(self
, file, *, write
=False):
1493 if self
.is_path(file):
1494 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1501 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1504 if (not ignore_discard
and cookie
.discard
1505 or not ignore_expires
and cookie
.is_expired(now
)):
1507 name
, value
= cookie
.name
, cookie
.value
1509 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1510 # with no name, whereas http.cookiejar regards it as a
1511 # cookie with no value.
1512 name
, value
= '', name
1513 f
.write('%s\n' % '\t'.join((
1515 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1517 self
._true
_or
_false
(cookie
.secure
),
1518 str_or_none(cookie
.expires
, default
=''),
1522 def save(self
, filename
=None, *args
, **kwargs
):
1524 Save cookies to a file.
1525 Code is taken from CPython 3.6
1526 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1528 if filename
is None:
1529 if self
.filename
is not None:
1530 filename
= self
.filename
1532 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1534 # Store session cookies with `expires` set to 0 instead of an empty string
1536 if cookie
.expires
is None:
1539 with self
.open(filename
, write
=True) as f
:
1540 f
.write(self
._HEADER
)
1541 self
._really
_save
(f
, *args
, **kwargs
)
1543 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1544 """Load cookies from a file."""
1545 if filename
is None:
1546 if self
.filename
is not None:
1547 filename
= self
.filename
1549 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1551 def prepare_line(line
):
1552 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1553 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1554 # comments and empty lines are fine
1555 if line
.startswith('#') or not line
.strip():
1557 cookie_list
= line
.split('\t')
1558 if len(cookie_list
) != self
._ENTRY
_LEN
:
1559 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1560 cookie
= self
._CookieFileEntry
(*cookie_list
)
1561 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1562 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1566 with self
.open(filename
) as f
:
1569 cf
.write(prepare_line(line
))
1570 except http
.cookiejar
.LoadError
as e
:
1571 if f
'{line.strip()} '[0] in '[{"':
1572 raise http
.cookiejar
.LoadError(
1573 'Cookies file must be Netscape formatted, not JSON. See '
1574 'https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl')
1575 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1578 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1579 # Session cookies are denoted by either `expires` field set to
1580 # an empty string or 0. MozillaCookieJar only recognizes the former
1581 # (see [1]). So we need force the latter to be recognized as session
1582 # cookies on our own.
1583 # Session cookies may be important for cookies-based authentication,
1584 # e.g. usually, when user does not check 'Remember me' check box while
1585 # logging in on a site, some important cookies are stored as session
1586 # cookies so that not recognizing them will result in failed login.
1587 # 1. https://bugs.python.org/issue17164
1589 # Treat `expires=0` cookies as session cookies
1590 if cookie
.expires
== 0:
1591 cookie
.expires
= None
1592 cookie
.discard
= True
1595 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1596 def __init__(self
, cookiejar
=None):
1597 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1599 def http_response(self
, request
, response
):
1600 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1602 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1603 https_response
= http_response
1606 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1607 """YoutubeDL redirect handler
1609 The code is based on HTTPRedirectHandler implementation from CPython [1].
1611 This redirect handler solves two issues:
1612 - ensures redirect URL is always unicode under python 2
1613 - introduces support for experimental HTTP response status code
1614 308 Permanent Redirect [2] used by some sites [3]
1616 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1617 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1618 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1621 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1623 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1624 """Return a Request or None in response to a redirect.
1626 This is called by the http_error_30x methods when a
1627 redirection response is received. If a redirection should
1628 take place, return a new Request to allow http_error_30x to
1629 perform the redirect. Otherwise, raise HTTPError if no-one
1630 else should try to handle this url. Return None if you can't
1631 but another Handler might.
1633 m
= req
.get_method()
1634 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1635 or code
in (301, 302, 303) and m
== "POST")):
1636 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1637 # Strictly (according to RFC 2616), 301 or 302 in response to
1638 # a POST MUST NOT cause a redirection without confirmation
1639 # from the user (of urllib.request, in this case). In practice,
1640 # essentially all clients do redirect in this case, so we do
1643 # Be conciliant with URIs containing a space. This is mainly
1644 # redundant with the more complete encoding done in http_error_302(),
1645 # but it is kept for compatibility with other callers.
1646 newurl
= newurl
.replace(' ', '%20')
1648 CONTENT_HEADERS
= ("content-length", "content-type")
1649 # NB: don't use dict comprehension for python 2.6 compatibility
1650 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1652 # A 303 must either use GET or HEAD for subsequent request
1653 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1654 if code
== 303 and m
!= 'HEAD':
1656 # 301 and 302 redirects are commonly turned into a GET from a POST
1657 # for subsequent requests by browsers, so we'll do the same.
1658 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1659 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1660 if code
in (301, 302) and m
== 'POST':
1663 return urllib
.request
.Request(
1664 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1665 unverifiable
=True, method
=m
)
1668 def extract_timezone(date_str
):
1671 ^.{8,}? # >=8 char non-TZ prefix, if present
1672 (?P<tz>Z| # just the UTC Z, or
1673 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1674 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1675 [ ]? # optional space
1676 (?P<sign>\+|-) # +/-
1677 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1681 timezone
= datetime
.timedelta()
1683 date_str
= date_str
[:-len(m
.group('tz'))]
1684 if not m
.group('sign'):
1685 timezone
= datetime
.timedelta()
1687 sign
= 1 if m
.group('sign') == '+' else -1
1688 timezone
= datetime
.timedelta(
1689 hours
=sign
* int(m
.group('hours')),
1690 minutes
=sign
* int(m
.group('minutes')))
1691 return timezone
, date_str
1694 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1695 """ Return a UNIX timestamp from the given date """
1697 if date_str
is None:
1700 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1702 if timezone
is None:
1703 timezone
, date_str
= extract_timezone(date_str
)
1705 with contextlib
.suppress(ValueError):
1706 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1707 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1708 return calendar
.timegm(dt
.timetuple())
1711 def date_formats(day_first
=True):
1712 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1715 def unified_strdate(date_str
, day_first
=True):
1716 """Return a string with the date in the format YYYYMMDD"""
1718 if date_str
is None:
1722 date_str
= date_str
.replace(',', ' ')
1723 # Remove AM/PM + timezone
1724 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1725 _
, date_str
= extract_timezone(date_str
)
1727 for expression
in date_formats(day_first
):
1728 with contextlib
.suppress(ValueError):
1729 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1730 if upload_date
is None:
1731 timetuple
= email
.utils
.parsedate_tz(date_str
)
1733 with contextlib
.suppress(ValueError):
1734 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1735 if upload_date
is not None:
1736 return str(upload_date
)
1739 def unified_timestamp(date_str
, day_first
=True):
1740 if date_str
is None:
1743 date_str
= re
.sub(r
'[,|]', '', date_str
)
1745 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1746 timezone
, date_str
= extract_timezone(date_str
)
1748 # Remove AM/PM + timezone
1749 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1751 # Remove unrecognized timezones from ISO 8601 alike timestamps
1752 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1754 date_str
= date_str
[:-len(m
.group('tz'))]
1756 # Python only supports microseconds, so remove nanoseconds
1757 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1759 date_str
= m
.group(1)
1761 for expression
in date_formats(day_first
):
1762 with contextlib
.suppress(ValueError):
1763 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1764 return calendar
.timegm(dt
.timetuple())
1765 timetuple
= email
.utils
.parsedate_tz(date_str
)
1767 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1770 def determine_ext(url
, default_ext
='unknown_video'):
1771 if url
is None or '.' not in url
:
1773 guess
= url
.partition('?')[0].rpartition('.')[2]
1774 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1776 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1777 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1778 return guess
.rstrip('/')
1783 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1784 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1787 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1789 Return a datetime object from a string.
1791 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1793 @param format strftime format of DATE
1794 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1795 auto: round to the unit provided in date_str (if applicable).
1797 auto_precision
= False
1798 if precision
== 'auto':
1799 auto_precision
= True
1800 precision
= 'microsecond'
1801 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1802 if date_str
in ('now', 'today'):
1804 if date_str
== 'yesterday':
1805 return today
- datetime
.timedelta(days
=1)
1807 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1809 if match
is not None:
1810 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1811 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1812 unit
= match
.group('unit')
1813 if unit
== 'month' or unit
== 'year':
1814 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1820 delta
= datetime
.timedelta(**{unit + 's': time}
)
1821 new_date
= start_time
+ delta
1823 return datetime_round(new_date
, unit
)
1826 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1829 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1831 Return a date object from a string using datetime_from_str
1833 @param strict Restrict allowed patterns to "YYYYMMDD" and
1834 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1836 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1837 raise ValueError(f
'Invalid date format "{date_str}"')
1838 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1841 def datetime_add_months(dt
, months
):
1842 """Increment/Decrement a datetime object by months."""
1843 month
= dt
.month
+ months
- 1
1844 year
= dt
.year
+ month
// 12
1845 month
= month
% 12 + 1
1846 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1847 return dt
.replace(year
, month
, day
)
1850 def datetime_round(dt
, precision
='day'):
1852 Round a datetime object's time to a specific precision
1854 if precision
== 'microsecond':
1863 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1864 timestamp
= calendar
.timegm(dt
.timetuple())
1865 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1868 def hyphenate_date(date_str
):
1870 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1871 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1872 if match
is not None:
1873 return '-'.join(match
.groups())
1879 """Represents a time interval between two dates"""
1881 def __init__(self
, start
=None, end
=None):
1882 """start and end must be strings in the format accepted by date"""
1883 if start
is not None:
1884 self
.start
= date_from_str(start
, strict
=True)
1886 self
.start
= datetime
.datetime
.min.date()
1888 self
.end
= date_from_str(end
, strict
=True)
1890 self
.end
= datetime
.datetime
.max.date()
1891 if self
.start
> self
.end
:
1892 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1896 """Returns a range that only contains the given day"""
1897 return cls(day
, day
)
1899 def __contains__(self
, date
):
1900 """Check if the date is in the range"""
1901 if not isinstance(date
, datetime
.date
):
1902 date
= date_from_str(date
)
1903 return self
.start
<= date
<= self
.end
1906 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1908 def __eq__(self
, other
):
1909 return (isinstance(other
, DateRange
)
1910 and self
.start
== other
.start
and self
.end
== other
.end
)
1913 def platform_name():
1914 """ Returns the platform name as a str """
1915 write_string('DeprecationWarning: yt_dlp.utils.platform_name is deprecated, use platform.platform instead')
1916 return platform
.platform()
1920 def system_identifier():
1921 python_implementation
= platform
.python_implementation()
1922 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
1923 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
1925 return 'Python %s (%s %s) - %s %s' % (
1926 platform
.python_version(),
1927 python_implementation
,
1928 platform
.architecture()[0],
1929 platform
.platform(),
1930 format_field(join_nonempty(*platform
.libc_ver(), delim
=' '), None, '(%s)'),
1935 def get_windows_version():
1936 ''' Get Windows version. returns () if it's not running on Windows '''
1937 if compat_os_name
== 'nt':
1938 return version_tuple(platform
.win32_ver()[1])
1943 def write_string(s
, out
=None, encoding
=None):
1944 assert isinstance(s
, str)
1945 out
= out
or sys
.stderr
1947 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
1948 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1950 enc
, buffer = None, out
1951 if 'b' in getattr(out
, 'mode', ''):
1952 enc
= encoding
or preferredencoding()
1953 elif hasattr(out
, 'buffer'):
1955 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1957 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
1961 def bytes_to_intlist(bs
):
1964 if isinstance(bs
[0], int): # Python 3
1967 return [ord(c
) for c
in bs
]
1970 def intlist_to_bytes(xs
):
1973 return struct
.pack('%dB' % len(xs
), *xs
)
1976 class LockingUnsupportedError(OSError):
1977 msg
= 'File locking is not supported'
1980 super().__init
__(self
.msg
)
1983 # Cross-platform file locking
1984 if sys
.platform
== 'win32':
1986 import ctypes
.wintypes
1989 class OVERLAPPED(ctypes
.Structure
):
1991 ('Internal', ctypes
.wintypes
.LPVOID
),
1992 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1993 ('Offset', ctypes
.wintypes
.DWORD
),
1994 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1995 ('hEvent', ctypes
.wintypes
.HANDLE
),
1998 kernel32
= ctypes
.windll
.kernel32
1999 LockFileEx
= kernel32
.LockFileEx
2000 LockFileEx
.argtypes
= [
2001 ctypes
.wintypes
.HANDLE
, # hFile
2002 ctypes
.wintypes
.DWORD
, # dwFlags
2003 ctypes
.wintypes
.DWORD
, # dwReserved
2004 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2005 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2006 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2008 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2009 UnlockFileEx
= kernel32
.UnlockFileEx
2010 UnlockFileEx
.argtypes
= [
2011 ctypes
.wintypes
.HANDLE
, # hFile
2012 ctypes
.wintypes
.DWORD
, # dwReserved
2013 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2014 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2015 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2017 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2018 whole_low
= 0xffffffff
2019 whole_high
= 0x7fffffff
2021 def _lock_file(f
, exclusive
, block
):
2022 overlapped
= OVERLAPPED()
2023 overlapped
.Offset
= 0
2024 overlapped
.OffsetHigh
= 0
2025 overlapped
.hEvent
= 0
2026 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2028 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2029 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2030 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2031 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2032 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2034 def _unlock_file(f
):
2035 assert f
._lock
_file
_overlapped
_p
2036 handle
= msvcrt
.get_osfhandle(f
.fileno())
2037 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2038 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2044 def _lock_file(f
, exclusive
, block
):
2045 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2047 flags |
= fcntl
.LOCK_NB
2049 fcntl
.flock(f
, flags
)
2050 except BlockingIOError
:
2052 except OSError: # AOSP does not have flock()
2053 fcntl
.lockf(f
, flags
)
2055 def _unlock_file(f
):
2057 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2059 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2063 def _lock_file(f
, exclusive
, block
):
2064 raise LockingUnsupportedError()
2066 def _unlock_file(f
):
2067 raise LockingUnsupportedError()
2073 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2074 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2075 raise NotImplementedError(mode
)
2076 self
.mode
, self
.block
= mode
, block
2078 writable
= any(f
in mode
for f
in 'wax+')
2079 readable
= any(f
in mode
for f
in 'r+')
2080 flags
= functools
.reduce(operator
.ior
, (
2081 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2082 getattr(os
, 'O_BINARY', 0), # Windows only
2083 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2084 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2085 os
.O_APPEND
if 'a' in mode
else 0,
2086 os
.O_EXCL
if 'x' in mode
else 0,
2087 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2090 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2092 def __enter__(self
):
2093 exclusive
= 'r' not in self
.mode
2095 _lock_file(self
.f
, exclusive
, self
.block
)
2100 if 'w' in self
.mode
:
2103 except OSError as e
:
2105 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2106 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2115 _unlock_file(self
.f
)
2119 def __exit__(self
, *_
):
2128 def __getattr__(self
, attr
):
2129 return getattr(self
.f
, attr
)
2136 def get_filesystem_encoding():
2137 encoding
= sys
.getfilesystemencoding()
2138 return encoding
if encoding
is not None else 'utf-8'
2141 def shell_quote(args
):
2143 encoding
= get_filesystem_encoding()
2145 if isinstance(a
, bytes):
2146 # We may get a filename encoded with 'encodeFilename'
2147 a
= a
.decode(encoding
)
2148 quoted_args
.append(compat_shlex_quote(a
))
2149 return ' '.join(quoted_args
)
2152 def smuggle_url(url
, data
):
2153 """ Pass additional data in a URL for internal use. """
2155 url
, idata
= unsmuggle_url(url
, {})
2157 sdata
= urllib
.parse
.urlencode(
2158 {'__youtubedl_smuggle': json.dumps(data)}
)
2159 return url
+ '#' + sdata
2162 def unsmuggle_url(smug_url
, default
=None):
2163 if '#__youtubedl_smuggle' not in smug_url
:
2164 return smug_url
, default
2165 url
, _
, sdata
= smug_url
.rpartition('#')
2166 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2167 data
= json
.loads(jsond
)
2171 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2172 """ Formats numbers with decimal sufixes like K, M, etc """
2173 num
, factor
= float_or_none(num
), float(factor
)
2174 if num
is None or num
< 0:
2176 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2177 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2178 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2180 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2181 converted
= num
/ (factor
** exponent
)
2182 return fmt
% (converted
, suffix
)
2185 def format_bytes(bytes):
2186 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2189 def lookup_unit_table(unit_table
, s
):
2190 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2192 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2195 num_str
= m
.group('num').replace(',', '.')
2196 mult
= unit_table
[m
.group('unit')]
2197 return int(float(num_str
) * mult
)
2200 def parse_filesize(s
):
2204 # The lower-case forms are of course incorrect and unofficial,
2205 # but we support those too
2222 'megabytes': 1000 ** 2,
2223 'mebibytes': 1024 ** 2,
2229 'gigabytes': 1000 ** 3,
2230 'gibibytes': 1024 ** 3,
2236 'terabytes': 1000 ** 4,
2237 'tebibytes': 1024 ** 4,
2243 'petabytes': 1000 ** 5,
2244 'pebibytes': 1024 ** 5,
2250 'exabytes': 1000 ** 6,
2251 'exbibytes': 1024 ** 6,
2257 'zettabytes': 1000 ** 7,
2258 'zebibytes': 1024 ** 7,
2264 'yottabytes': 1000 ** 8,
2265 'yobibytes': 1024 ** 8,
2268 return lookup_unit_table(_UNIT_TABLE
, s
)
2275 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2277 if re
.match(r
'^[\d,.]+$', s
):
2278 return str_to_int(s
)
2291 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2295 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2297 return str_to_int(mobj
.group(1))
2300 def parse_resolution(s
, *, lenient
=False):
2305 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2307 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2310 'width': int(mobj
.group('w')),
2311 'height': int(mobj
.group('h')),
2314 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2316 return {'height': int(mobj.group(1))}
2318 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2320 return {'height': int(mobj.group(1)) * 540}
2325 def parse_bitrate(s
):
2326 if not isinstance(s
, str):
2328 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2330 return int(mobj
.group(1))
2333 def month_by_name(name
, lang
='en'):
2334 """ Return the number of a month by (locale-independently) English name """
2336 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2339 return month_names
.index(name
) + 1
2344 def month_by_abbreviation(abbrev
):
2345 """ Return the number of a month by (locale-independently) English
2349 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2354 def fix_xml_ampersands(xml_str
):
2355 """Replace all the '&' by '&' in XML"""
2357 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2362 def setproctitle(title
):
2363 assert isinstance(title
, str)
2365 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2372 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2376 # LoadLibrary in Windows Python 2.7.13 only expects
2377 # a bytestring, but since unicode_literals turns
2378 # every string into a unicode string, it fails.
2380 title_bytes
= title
.encode()
2381 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2382 buf
.value
= title_bytes
2384 libc
.prctl(15, buf
, 0, 0, 0)
2385 except AttributeError:
2386 return # Strange libc, just skip this
2389 def remove_start(s
, start
):
2390 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2393 def remove_end(s
, end
):
2394 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2397 def remove_quotes(s
):
2398 if s
is None or len(s
) < 2:
2400 for quote
in ('"', "'", ):
2401 if s
[0] == quote
and s
[-1] == quote
:
2406 def get_domain(url
):
2408 This implementation is inconsistent, but is kept for compatibility.
2409 Use this only for "webpage_url_domain"
2411 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2414 def url_basename(url
):
2415 path
= urllib
.parse
.urlparse(url
).path
2416 return path
.strip('/').split('/')[-1]
2420 return re
.match(r
'https?://[^?#&]+/', url
).group()
2423 def urljoin(base
, path
):
2424 if isinstance(path
, bytes):
2425 path
= path
.decode()
2426 if not isinstance(path
, str) or not path
:
2428 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2430 if isinstance(base
, bytes):
2431 base
= base
.decode()
2432 if not isinstance(base
, str) or not re
.match(
2433 r
'^(?:https?:)?//', base
):
2435 return urllib
.parse
.urljoin(base
, path
)
2438 class HEADRequest(urllib
.request
.Request
):
2439 def get_method(self
):
2443 class PUTRequest(urllib
.request
.Request
):
2444 def get_method(self
):
2448 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2449 if get_attr
and v
is not None:
2450 v
= getattr(v
, get_attr
, None)
2452 return int(v
) * invscale
// scale
2453 except (ValueError, TypeError, OverflowError):
2457 def str_or_none(v
, default
=None):
2458 return default
if v
is None else str(v
)
2461 def str_to_int(int_str
):
2462 """ A more relaxed version of int_or_none """
2463 if isinstance(int_str
, int):
2465 elif isinstance(int_str
, str):
2466 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2467 return int_or_none(int_str
)
2470 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2474 return float(v
) * invscale
/ scale
2475 except (ValueError, TypeError):
2479 def bool_or_none(v
, default
=None):
2480 return v
if isinstance(v
, bool) else default
2483 def strip_or_none(v
, default
=None):
2484 return v
.strip() if isinstance(v
, str) else default
2487 def url_or_none(url
):
2488 if not url
or not isinstance(url
, str):
2491 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2494 def request_to_url(req
):
2495 if isinstance(req
, urllib
.request
.Request
):
2496 return req
.get_full_url()
2501 def strftime_or_none(timestamp
, date_format
, default
=None):
2502 datetime_object
= None
2504 if isinstance(timestamp
, (int, float)): # unix timestamp
2505 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2506 elif isinstance(timestamp
, str): # assume YYYYMMDD
2507 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2508 return datetime_object
.strftime(date_format
)
2509 except (ValueError, TypeError, AttributeError):
2513 def parse_duration(s
):
2514 if not isinstance(s
, str):
2520 days
, hours
, mins
, secs
, ms
= [None] * 5
2521 m
= re
.match(r
'''(?x)
2523 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2524 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2525 (?P<ms>[.:][0-9]+)?Z?$
2528 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2533 [0-9]+\s*y(?:ears?)?,?\s*
2536 [0-9]+\s*m(?:onths?)?,?\s*
2539 [0-9]+\s*w(?:eeks?)?,?\s*
2542 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2546 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2549 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2552 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2555 days
, hours
, mins
, secs
, ms
= m
.groups()
2557 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2559 hours
, mins
= m
.groups()
2564 ms
= ms
.replace(':', '.')
2565 return sum(float(part
or 0) * mult
for part
, mult
in (
2566 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2569 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2570 name
, real_ext
= os
.path
.splitext(filename
)
2572 f
'{name}.{ext}{real_ext}'
2573 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2574 else f
'{filename}.{ext}')
2577 def replace_extension(filename
, ext
, expected_real_ext
=None):
2578 name
, real_ext
= os
.path
.splitext(filename
)
2579 return '{}.{}'.format(
2580 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2584 def check_executable(exe
, args
=[]):
2585 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2586 args can be a list of arguments for a short output (like -version) """
2588 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2594 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2596 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2598 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2599 # SIGTTOU if yt-dlp is run in the background.
2600 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2601 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2602 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2608 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2609 assert isinstance(output
, str)
2610 if version_re
is None:
2611 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2612 m
= re
.search(version_re
, output
)
2619 def get_exe_version(exe
, args
=['--version'],
2620 version_re
=None, unrecognized
='present'):
2621 """ Returns the version of the specified executable,
2622 or False if the executable is not present """
2623 out
= _get_exe_version_output(exe
, args
)
2624 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2627 def frange(start
=0, stop
=None, step
=1):
2630 start
, stop
= 0, start
2631 sign
= [-1, 1][step
> 0] if step
else 0
2632 while sign
* start
< sign
* stop
:
2637 class LazyList(collections
.abc
.Sequence
):
2638 """Lazy immutable list from an iterable
2639 Note that slices of a LazyList are lists and not LazyList"""
2641 class IndexError(IndexError):
2644 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2645 self
._iterable
= iter(iterable
)
2646 self
._cache
= [] if _cache
is None else _cache
2647 self
._reversed
= reverse
2651 # We need to consume the entire iterable to iterate in reverse
2652 yield from self
.exhaust()
2654 yield from self
._cache
2655 for item
in self
._iterable
:
2656 self
._cache
.append(item
)
2660 self
._cache
.extend(self
._iterable
)
2661 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2665 """Evaluate the entire iterable"""
2666 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2669 def _reverse_index(x
):
2670 return None if x
is None else ~x
2672 def __getitem__(self
, idx
):
2673 if isinstance(idx
, slice):
2675 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2676 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2677 elif isinstance(idx
, int):
2679 idx
= self
._reverse
_index
(idx
)
2680 start
, stop
, step
= idx
, idx
, 0
2682 raise TypeError('indices must be integers or slices')
2683 if ((start
or 0) < 0 or (stop
or 0) < 0
2684 or (start
is None and step
< 0)
2685 or (stop
is None and step
> 0)):
2686 # We need to consume the entire iterable to be able to slice from the end
2687 # Obviously, never use this with infinite iterables
2690 return self
._cache
[idx
]
2691 except IndexError as e
:
2692 raise self
.IndexError(e
) from e
2693 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2695 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2697 return self
._cache
[idx
]
2698 except IndexError as e
:
2699 raise self
.IndexError(e
) from e
2703 self
[-1] if self
._reversed
else self
[0]
2704 except self
.IndexError:
2710 return len(self
._cache
)
2712 def __reversed__(self
):
2713 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2716 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2719 # repr and str should mimic a list. So we exhaust the iterable
2720 return repr(self
.exhaust())
2723 return repr(self
.exhaust())
2728 class IndexError(IndexError):
2732 # This is only useful for tests
2733 return len(self
.getslice())
2735 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2736 self
._pagefunc
= pagefunc
2737 self
._pagesize
= pagesize
2738 self
._pagecount
= float('inf')
2739 self
._use
_cache
= use_cache
2742 def getpage(self
, pagenum
):
2743 page_results
= self
._cache
.get(pagenum
)
2744 if page_results
is None:
2745 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2747 self
._cache
[pagenum
] = page_results
2750 def getslice(self
, start
=0, end
=None):
2751 return list(self
._getslice
(start
, end
))
2753 def _getslice(self
, start
, end
):
2754 raise NotImplementedError('This method must be implemented by subclasses')
2756 def __getitem__(self
, idx
):
2757 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2758 if not isinstance(idx
, int) or idx
< 0:
2759 raise TypeError('indices must be non-negative integers')
2760 entries
= self
.getslice(idx
, idx
+ 1)
2762 raise self
.IndexError()
2766 class OnDemandPagedList(PagedList
):
2767 """Download pages until a page with less than maximum results"""
2769 def _getslice(self
, start
, end
):
2770 for pagenum
in itertools
.count(start
// self
._pagesize
):
2771 firstid
= pagenum
* self
._pagesize
2772 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2773 if start
>= nextfirstid
:
2777 start
% self
._pagesize
2778 if firstid
<= start
< nextfirstid
2781 ((end
- 1) % self
._pagesize
) + 1
2782 if (end
is not None and firstid
<= end
<= nextfirstid
)
2786 page_results
= self
.getpage(pagenum
)
2788 self
._pagecount
= pagenum
- 1
2790 if startv
!= 0 or endv
is not None:
2791 page_results
= page_results
[startv
:endv
]
2792 yield from page_results
2794 # A little optimization - if current page is not "full", ie. does
2795 # not contain page_size videos then we can assume that this page
2796 # is the last one - there are no more ids on further pages -
2797 # i.e. no need to query again.
2798 if len(page_results
) + startv
< self
._pagesize
:
2801 # If we got the whole page, but the next page is not interesting,
2802 # break out early as well
2803 if end
== nextfirstid
:
2807 class InAdvancePagedList(PagedList
):
2808 """PagedList with total number of pages known in advance"""
2810 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2811 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2812 self
._pagecount
= pagecount
2814 def _getslice(self
, start
, end
):
2815 start_page
= start
// self
._pagesize
2816 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2817 skip_elems
= start
- start_page
* self
._pagesize
2818 only_more
= None if end
is None else end
- start
2819 for pagenum
in range(start_page
, end_page
):
2820 page_results
= self
.getpage(pagenum
)
2822 page_results
= page_results
[skip_elems
:]
2824 if only_more
is not None:
2825 if len(page_results
) < only_more
:
2826 only_more
-= len(page_results
)
2828 yield from page_results
[:only_more
]
2830 yield from page_results
2833 class PlaylistEntries
:
2834 MissingEntry
= object()
2835 is_exhausted
= False
2837 def __init__(self
, ydl
, info_dict
):
2840 # _entries must be assigned now since infodict can change during iteration
2841 entries
= info_dict
.get('entries')
2843 raise EntryNotInPlaylist('There are no entries')
2844 elif isinstance(entries
, list):
2845 self
.is_exhausted
= True
2847 requested_entries
= info_dict
.get('requested_entries')
2848 self
.is_incomplete
= bool(requested_entries
)
2849 if self
.is_incomplete
:
2850 assert self
.is_exhausted
2851 self
._entries
= [self
.MissingEntry
] * max(requested_entries
)
2852 for i
, entry
in zip(requested_entries
, entries
):
2853 self
._entries
[i
- 1] = entry
2854 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2855 self
._entries
= entries
2857 self
._entries
= LazyList(entries
)
2859 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2860 (?P<start>[+-]?\d+)?
2862 (?P<end>[+-]?\d+|inf(?:inite)?)?
2863 (?::(?P<step>[+-]?\d+))?
2867 def parse_playlist_items(cls
, string
):
2868 for segment
in string
.split(','):
2870 raise ValueError('There is two or more consecutive commas')
2871 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2873 raise ValueError(f
'{segment!r} is not a valid specification')
2874 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2875 if int_or_none(step
) == 0:
2876 raise ValueError(f
'Step in {segment!r} cannot be zero')
2877 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2879 def get_requested_items(self
):
2880 playlist_items
= self
.ydl
.params
.get('playlist_items')
2881 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2882 playlist_end
= self
.ydl
.params
.get('playlistend')
2883 # For backwards compatibility, interpret -1 as whole list
2884 if playlist_end
in (-1, None):
2886 if not playlist_items
:
2887 playlist_items
= f
'{playlist_start}:{playlist_end}'
2888 elif playlist_start
!= 1 or playlist_end
:
2889 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2891 for index
in self
.parse_playlist_items(playlist_items
):
2892 for i
, entry
in self
[index
]:
2897 # TODO: Add auto-generated fields
2898 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
2899 except (ExistingVideoReached
, RejectedVideoReached
):
2902 def get_full_count(self
):
2903 if self
.is_exhausted
and not self
.is_incomplete
:
2905 elif isinstance(self
._entries
, InAdvancePagedList
):
2906 if self
._entries
._pagesize
== 1:
2907 return self
._entries
._pagecount
2909 @functools.cached_property
2911 if isinstance(self
._entries
, list):
2914 entry
= self
._entries
[i
]
2916 entry
= self
.MissingEntry
2917 if not self
.is_incomplete
:
2918 raise self
.IndexError()
2919 if entry
is self
.MissingEntry
:
2920 raise EntryNotInPlaylist(f
'Entry {i} cannot be found')
2925 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
2926 except (LazyList
.IndexError, PagedList
.IndexError):
2927 raise self
.IndexError()
2930 def __getitem__(self
, idx
):
2931 if isinstance(idx
, int):
2932 idx
= slice(idx
, idx
)
2934 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
2935 step
= 1 if idx
.step
is None else idx
.step
2936 if idx
.start
is None:
2937 start
= 0 if step
> 0 else len(self
) - 1
2939 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
2941 # NB: Do not call len(self) when idx == [:]
2942 if idx
.stop
is None:
2943 stop
= 0 if step
< 0 else float('inf')
2945 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
2946 stop
+= [-1, 1][step
> 0]
2948 for i
in frange(start
, stop
, step
):
2952 entry
= self
._getter
(i
)
2953 except self
.IndexError:
2954 self
.is_exhausted
= True
2961 return len(tuple(self
[:]))
2963 class IndexError(IndexError):
2967 def uppercase_escape(s
):
2968 unicode_escape
= codecs
.getdecoder('unicode_escape')
2970 r
'\\U[0-9a-fA-F]{8}',
2971 lambda m
: unicode_escape(m
.group(0))[0],
2975 def lowercase_escape(s
):
2976 unicode_escape
= codecs
.getdecoder('unicode_escape')
2978 r
'\\u[0-9a-fA-F]{4}',
2979 lambda m
: unicode_escape(m
.group(0))[0],
2983 def escape_rfc3986(s
):
2984 """Escape non-ASCII characters as suggested by RFC 3986"""
2985 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2988 def escape_url(url
):
2989 """Escape URL as suggested by RFC 3986"""
2990 url_parsed
= urllib
.parse
.urlparse(url
)
2991 return url_parsed
._replace
(
2992 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2993 path
=escape_rfc3986(url_parsed
.path
),
2994 params
=escape_rfc3986(url_parsed
.params
),
2995 query
=escape_rfc3986(url_parsed
.query
),
2996 fragment
=escape_rfc3986(url_parsed
.fragment
)
3001 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
)
3004 def read_batch_urls(batch_fd
):
3006 if not isinstance(url
, str):
3007 url
= url
.decode('utf-8', 'replace')
3008 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3009 for bom
in BOM_UTF8
:
3010 if url
.startswith(bom
):
3011 url
= url
[len(bom
):]
3013 if not url
or url
.startswith(('#', ';', ']')):
3015 # "#" cannot be stripped out since it is part of the URI
3016 # However, it can be safely stripped out if following a whitespace
3017 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3019 with contextlib
.closing(batch_fd
) as fd
:
3020 return [url
for url
in map(fixup
, fd
) if url
]
3023 def urlencode_postdata(*args
, **kargs
):
3024 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3027 def update_url_query(url
, query
):
3030 parsed_url
= urllib
.parse
.urlparse(url
)
3031 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3033 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3034 query
=urllib
.parse
.urlencode(qs
, True)))
3037 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3038 req_headers
= req
.headers
.copy()
3039 req_headers
.update(headers
or {})
3040 req_data
= data
or req
.data
3041 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3042 req_get_method
= req
.get_method()
3043 if req_get_method
== 'HEAD':
3044 req_type
= HEADRequest
3045 elif req_get_method
== 'PUT':
3046 req_type
= PUTRequest
3048 req_type
= urllib
.request
.Request
3050 req_url
, data
=req_data
, headers
=req_headers
,
3051 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3052 if hasattr(req
, 'timeout'):
3053 new_req
.timeout
= req
.timeout
3057 def _multipart_encode_impl(data
, boundary
):
3058 content_type
= 'multipart/form-data; boundary=%s' % boundary
3061 for k
, v
in data
.items():
3062 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3063 if isinstance(k
, str):
3065 if isinstance(v
, str):
3067 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3068 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3069 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3070 if boundary
.encode('ascii') in content
:
3071 raise ValueError('Boundary overlaps with data')
3074 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3076 return out
, content_type
3079 def multipart_encode(data
, boundary
=None):
3081 Encode a dict to RFC 7578-compliant form-data
3084 A dict where keys and values can be either Unicode or bytes-like
3087 If specified a Unicode object, it's used as the boundary. Otherwise
3088 a random boundary is generated.
3090 Reference: https://tools.ietf.org/html/rfc7578
3092 has_specified_boundary
= boundary
is not None
3095 if boundary
is None:
3096 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3099 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3102 if has_specified_boundary
:
3106 return out
, content_type
3109 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3110 for val
in map(d
.get
, variadic(key_or_keys
)):
3111 if val
is not None and (val
or not skip_false_values
):
3116 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3119 val
= f(*args
, **kwargs
)
3120 except (AttributeError, KeyError, TypeError, IndexError, ZeroDivisionError):
3123 if expected_type
is None or isinstance(val
, expected_type
):
3127 def try_get(src
, getter
, expected_type
=None):
3128 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3131 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3132 return {k: v for k, v in dct.items() if cndn(k, v)}
3135 def merge_dicts(*dicts
):
3137 for a_dict
in dicts
:
3138 for k
, v
in a_dict
.items():
3139 if (v
is not None and k
not in merged
3140 or isinstance(v
, str) and merged
[k
] == ''):
3145 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3146 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3158 TV_PARENTAL_GUIDELINES
= {
3168 def parse_age_limit(s
):
3169 # isinstance(False, int) is True. So type() must be used instead
3170 if type(s
) is int: # noqa: E721
3171 return s
if 0 <= s
<= 21 else None
3172 elif not isinstance(s
, str):
3174 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3176 return int(m
.group('age'))
3179 return US_RATINGS
[s
]
3180 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3182 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3186 def strip_jsonp(code
):
3189 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3190 (?:\s*&&\s*(?P=func_name))?
3191 \s*\(\s*(?P<callback_data>.*)\);?
3192 \s*?(?://[^\n]*)*$''',
3193 r
'\g<callback_data>', code
)
3196 def js_to_json(code
, vars={}):
3197 # vars is a dict of var, val pairs to substitute
3198 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3199 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3201 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3202 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3207 if v
in ('true', 'false', 'null'):
3209 elif v
in ('undefined', 'void 0'):
3211 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3214 if v
[0] in ("'", '"'):
3215 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3220 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3222 for regex
, base
in INTEGER_TABLE
:
3223 im
= re
.match(regex
, v
)
3225 i
= int(im
.group(1), base
)
3226 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3233 def create_map(mobj
):
3234 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3236 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3237 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3239 return re
.sub(r
'''(?sx)
3240 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3241 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3242 {comment}|,(?={skip}[\]}}])|
3243 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3244 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3247 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3250 def qualities(quality_ids
):
3251 """ Get a numeric quality value out of a list of possible values """
3254 return quality_ids
.index(qid
)
3260 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3264 'default': '%(title)s [%(id)s].%(ext)s',
3265 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3271 'description': 'description',
3272 'annotation': 'annotations.xml',
3273 'infojson': 'info.json',
3276 'pl_thumbnail': None,
3277 'pl_description': 'description',
3278 'pl_infojson': 'info.json',
3281 # As of [1] format syntax is:
3282 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3283 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3284 STR_FORMAT_RE_TMPL
= r
'''(?x)
3285 (?<!%)(?P<prefix>(?:%%)*)
3287 (?P<has_key>\((?P<key>{0})\))?
3289 (?P<conversion>[#0\-+ ]+)?
3291 (?P<precision>\.\d+)?
3292 (?P<len_mod>[hlL])? # unused in python
3293 {1} # conversion type
3298 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3301 def limit_length(s
, length
):
3302 """ Add ellipses to overly long strings """
3307 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3311 def version_tuple(v
):
3312 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3315 def is_outdated_version(version
, limit
, assume_new
=True):
3317 return not assume_new
3319 return version_tuple(version
) < version_tuple(limit
)
3321 return not assume_new
3324 def ytdl_is_updateable():
3325 """ Returns if yt-dlp can be updated with -U """
3327 from .update
import is_non_updateable
3329 return not is_non_updateable()
3332 def args_to_str(args
):
3333 # Get a short string representation for a subprocess command
3334 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3337 def error_to_compat_str(err
):
3341 def error_to_str(err
):
3342 return f
'{type(err).__name__}: {err}'
3345 def mimetype2ext(mt
):
3349 mt
, _
, params
= mt
.partition(';')
3354 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3355 # it's the most popular one
3356 'audio/mpeg': 'mp3',
3357 'audio/x-wav': 'wav',
3359 'audio/wave': 'wav',
3362 ext
= FULL_MAP
.get(mt
)
3368 'smptett+xml': 'tt',
3372 'x-mp4-fragmented': 'mp4',
3373 'x-ms-sami': 'sami',
3376 'x-mpegurl': 'm3u8',
3377 'vnd.apple.mpegurl': 'm3u8',
3381 'vnd.ms-sstr+xml': 'ism',
3385 'filmstrip+json': 'fs',
3389 _
, _
, subtype
= mt
.rpartition('/')
3390 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3401 _
, _
, suffix
= subtype
.partition('+')
3402 ext
= SUFFIX_MAP
.get(suffix
)
3406 return subtype
.replace('+', '.')
3409 def ext2mimetype(ext_or_url
):
3412 if '.' not in ext_or_url
:
3413 ext_or_url
= f
'file.{ext_or_url}'
3414 return mimetypes
.guess_type(ext_or_url
)[0]
3417 def parse_codecs(codecs_str
):
3418 # http://tools.ietf.org/html/rfc6381
3421 split_codecs
= list(filter(None, map(
3422 str.strip
, codecs_str
.strip().strip(',').split(','))))
3423 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3424 for full_codec
in split_codecs
:
3425 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3426 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3427 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3431 if parts
[0] in ('dvh1', 'dvhe'):
3433 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3435 elif parts
[:2] == ['vp9', '2']:
3437 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3438 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3439 acodec
= acodec
or full_codec
3440 elif parts
[0] in ('stpp', 'wvtt'):
3441 scodec
= scodec
or full_codec
3443 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3444 if vcodec
or acodec
or scodec
:
3446 'vcodec': vcodec
or 'none',
3447 'acodec': acodec
or 'none',
3448 'dynamic_range': hdr
,
3449 **({'scodec': scodec}
if scodec
is not None else {}),
3451 elif len(split_codecs
) == 2:
3453 'vcodec': split_codecs
[0],
3454 'acodec': split_codecs
[1],
3459 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3460 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3462 allow_mkv
= not preferences
or 'mkv' in preferences
3464 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3465 return 'mkv' # TODO: any other format allows this?
3467 # TODO: All codecs supported by parse_codecs isn't handled here
3468 COMPATIBLE_CODECS
= {
3470 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3471 'h264', 'aacl', # Set in ISM
3474 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3475 'vp9x', 'vp8x', # in the webm spec
3479 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
.split('.')[0].replace('0', ''))
3480 vcodec
, acodec
= sanitize_codec(vcodecs
[0]), sanitize_codec(acodecs
[0])
3482 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3483 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3484 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3488 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3491 for ext
in preferences
or vexts
:
3492 current_exts
= {ext, *vexts, *aexts}
3493 if ext
== 'mkv' or current_exts
== {ext}
or any(
3494 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3496 return 'mkv' if allow_mkv
else preferences
[-1]
3499 def urlhandle_detect_ext(url_handle
):
3500 getheader
= url_handle
.headers
.get
3502 cd
= getheader('Content-Disposition')
3504 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3506 e
= determine_ext(m
.group('filename'), default_ext
=None)
3510 return mimetype2ext(getheader('Content-Type'))
3513 def encode_data_uri(data
, mime_type
):
3514 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3517 def age_restricted(content_limit
, age_limit
):
3518 """ Returns True iff the content should be blocked """
3520 if age_limit
is None: # No limit set
3522 if content_limit
is None:
3523 return False # Content available for everyone
3524 return age_limit
< content_limit
3527 # List of known byte-order-marks (BOM)
3529 (b
'\xef\xbb\xbf', 'utf-8'),
3530 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3531 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3532 (b
'\xff\xfe', 'utf-16-le'),
3533 (b
'\xfe\xff', 'utf-16-be'),
3537 def is_html(first_bytes
):
3538 """ Detect whether a file contains HTML by examining its first bytes. """
3541 for bom
, enc
in BOMS
:
3542 while first_bytes
.startswith(bom
):
3543 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3545 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3548 def determine_protocol(info_dict
):
3549 protocol
= info_dict
.get('protocol')
3550 if protocol
is not None:
3553 url
= sanitize_url(info_dict
['url'])
3554 if url
.startswith('rtmp'):
3556 elif url
.startswith('mms'):
3558 elif url
.startswith('rtsp'):
3561 ext
= determine_ext(url
)
3567 return urllib
.parse
.urlparse(url
).scheme
3570 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3571 """ Render a list of rows, each as a list of values.
3572 Text after a \t will be right aligned """
3574 return len(remove_terminal_sequences(string
).replace('\t', ''))
3576 def get_max_lens(table
):
3577 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3579 def filter_using_list(row
, filterArray
):
3580 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3582 max_lens
= get_max_lens(data
) if hide_empty
else []
3583 header_row
= filter_using_list(header_row
, max_lens
)
3584 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3586 table
= [header_row
] + data
3587 max_lens
= get_max_lens(table
)
3590 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3591 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3593 for pos
, text
in enumerate(map(str, row
)):
3595 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3597 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3598 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3602 def _match_one(filter_part
, dct
, incomplete
):
3603 # TODO: Generalize code with YoutubeDL._build_format_filter
3604 STRING_OPERATORS
= {
3605 '*=': operator
.contains
,
3606 '^=': lambda attr
, value
: attr
.startswith(value
),
3607 '$=': lambda attr
, value
: attr
.endswith(value
),
3608 '~=': lambda attr
, value
: re
.search(value
, attr
),
3610 COMPARISON_OPERATORS
= {
3612 '<=': operator
.le
, # "<=" must be defined above "<"
3619 if isinstance(incomplete
, bool):
3620 is_incomplete
= lambda _
: incomplete
3622 is_incomplete
= lambda k
: k
in incomplete
3624 operator_rex
= re
.compile(r
'''(?x)
3626 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3628 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3631 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3632 m = operator_rex.fullmatch(filter_part.strip())
3635 unnegated_op = COMPARISON_OPERATORS[m['op']]
3637 op = lambda attr, value: not unnegated_op(attr, value)
3640 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3642 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3643 actual_value = dct.get(m['key'])
3644 numeric_comparison = None
3645 if isinstance(actual_value, (int, float)):
3646 # If the original field is a string and matching comparisonvalue is
3647 # a number we should respect the origin of the original field
3648 # and process comparison value as a string (see
3649 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3651 numeric_comparison = int(comparison_value)
3653 numeric_comparison = parse_filesize(comparison_value)
3654 if numeric_comparison is None:
3655 numeric_comparison = parse_filesize(f'{comparison_value}B')
3656 if numeric_comparison is None:
3657 numeric_comparison = parse_duration(comparison_value)
3658 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3659 raise ValueError('Operator %s only supports string values!' % m['op'])
3660 if actual_value is None:
3661 return is_incomplete(m['key']) or m['none_inclusive']
3662 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3665 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3666 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3668 operator_rex = re.compile(r'''(?x
)
3669 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3670 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3671 m = operator_rex.fullmatch(filter_part.strip())
3673 op = UNARY_OPERATORS[m.group('op')]
3674 actual_value = dct.get(m.group('key'))
3675 if is_incomplete(m.group('key')) and actual_value is None:
3677 return op(actual_value)
3679 raise ValueError('Invalid filter part %r' % filter_part)
3682 def match_str(filter_str, dct, incomplete=False):
3683 """ Filter a dictionary with a simple string syntax.
3684 @returns Whether the filter passes
3685 @param incomplete Set of keys that is expected to be missing from dct.
3686 Can be True/False to indicate all/none of the keys may be missing.
3687 All conditions on incomplete keys pass if the key is missing
3690 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3691 for filter_part in re.split(r'(?<!\\)&', filter_str))
3694 def match_filter_func(filters):
3697 filters = set(variadic(filters))
3699 interactive = '-' in filters
3703 def _match_func(info_dict, incomplete=False):
3704 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3705 return NO_DEFAULT if interactive and not incomplete else None
3707 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3708 filter_str = ') | ('.join(map(str.strip, filters))
3709 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3713 class download_range_func:
3714 def __init__(self, chapters, ranges):
3715 self.chapters, self.ranges = chapters, ranges
3717 def __call__(self, info_dict, ydl):
3718 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3719 else 'Cannot match chapters since chapter information is unavailable')
3720 for regex in self.chapters or []:
3721 for i, chapter in enumerate(info_dict.get('chapters') or []):
3722 if re.search(regex, chapter['title']):
3724 yield {**chapter, 'index': i}
3725 if self.chapters and warning:
3726 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3728 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3730 def __eq__(self, other):
3731 return (isinstance(other, download_range_func)
3732 and self.chapters == other.chapters and self.ranges == other.ranges)
3735 def parse_dfxp_time_expr(time_expr):
3739 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3741 return float(mobj.group('time_offset'))
3743 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3745 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3748 def srt_subtitles_timecode(seconds):
3749 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3752 def ass_subtitles_timecode(seconds):
3753 time = timetuple_from_msec(seconds * 1000)
3754 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3757 def dfxp2srt(dfxp_data):
3759 @param dfxp_data A
bytes-like
object containing DFXP data
3760 @returns A
unicode object containing converted SRT data
3762 LEGACY_NAMESPACES = (
3763 (b'http://www.w3.org/ns/ttml', [
3764 b'http://www.w3.org/2004/11/ttaf1',
3765 b'http://www.w3.org/2006/04/ttaf1',
3766 b'http://www.w3.org/2006/10/ttaf1',
3768 (b'http://www.w3.org/ns/ttml#styling', [
3769 b'http://www.w3.org/ns/ttml#style',
3773 SUPPORTED_STYLING = [
3782 _x = functools.partial(xpath_with_ns, ns_map={
3783 'xml': 'http://www.w3.org/XML/1998/namespace',
3784 'ttml': 'http://www.w3.org/ns/ttml',
3785 'tts': 'http://www.w3.org/ns/ttml#styling',
3791 class TTMLPElementParser:
3793 _unclosed_elements = []
3794 _applied_styles = []
3796 def start(self, tag, attrib):
3797 if tag in (_x('ttml:br'), 'br'):
3800 unclosed_elements = []
3802 element_style_id = attrib.get('style')
3804 style.update(default_style)
3805 if element_style_id:
3806 style.update(styles.get(element_style_id, {}))
3807 for prop in SUPPORTED_STYLING:
3808 prop_val = attrib.get(_x('tts:' + prop))
3810 style[prop] = prop_val
3813 for k, v in sorted(style.items()):
3814 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3817 font += ' color="%s"' % v
3818 elif k == 'fontSize':
3819 font += ' size="%s"' % v
3820 elif k == 'fontFamily':
3821 font += ' face="%s"' % v
3822 elif k == 'fontWeight' and v == 'bold':
3824 unclosed_elements.append('b')
3825 elif k == 'fontStyle' and v == 'italic':
3827 unclosed_elements.append('i')
3828 elif k == 'textDecoration' and v == 'underline':
3830 unclosed_elements.append('u')
3832 self._out += '<font' + font + '>'
3833 unclosed_elements.append('font')
3835 if self._applied_styles:
3836 applied_style.update(self._applied_styles[-1])
3837 applied_style.update(style)
3838 self._applied_styles.append(applied_style)
3839 self._unclosed_elements.append(unclosed_elements)
3842 if tag not in (_x('ttml:br'), 'br'):
3843 unclosed_elements = self._unclosed_elements.pop()
3844 for element in reversed(unclosed_elements):
3845 self._out += '</%s>' % element
3846 if unclosed_elements and self._applied_styles:
3847 self._applied_styles.pop()
3849 def data(self, data):
3853 return self._out.strip()
3855 def parse_node(node):
3856 target = TTMLPElementParser()
3857 parser = xml.etree.ElementTree.XMLParser(target=target)
3858 parser.feed(xml.etree.ElementTree.tostring(node))
3859 return parser.close()
3861 for k, v in LEGACY_NAMESPACES:
3863 dfxp_data = dfxp_data.replace(ns, k)
3865 dfxp = compat_etree_fromstring(dfxp_data)
3867 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3870 raise ValueError('Invalid dfxp/TTML subtitle')
3874 for style in dfxp.findall(_x('.//ttml:style')):
3875 style_id = style.get('id') or style.get(_x('xml:id'))
3878 parent_style_id = style.get('style')
3880 if parent_style_id not in styles:
3883 styles[style_id] = styles[parent_style_id].copy()
3884 for prop in SUPPORTED_STYLING:
3885 prop_val = style.get(_x('tts:' + prop))
3887 styles.setdefault(style_id, {})[prop] = prop_val
3893 for p in ('body', 'div'):
3894 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3897 style = styles.get(ele.get('style'))
3900 default_style.update(style)
3902 for para, index in zip(paras, itertools.count(1)):
3903 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3904 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3905 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3906 if begin_time is None:
3911 end_time = begin_time + dur
3912 out.append('%d\n%s --> %s\n%s\n\n' % (
3914 srt_subtitles_timecode(begin_time),
3915 srt_subtitles_timecode(end_time),
3921 def cli_option(params, command_option, param, separator=None):
3922 param = params.get(param)
3923 return ([] if param is None
3924 else [command_option, str(param)] if separator is None
3925 else [f'{command_option}{separator}{param}'])
3928 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3929 param = params.get(param)
3930 assert param in (True, False, None)
3931 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
3934 def cli_valueless_option(params, command_option, param, expected_value=True):
3935 return [command_option] if params.get(param) == expected_value else []
3938 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3939 if isinstance(argdict, (list, tuple)): # for backward compatibility
3946 assert isinstance(argdict, dict)
3948 assert isinstance(keys, (list, tuple))
3949 for key_list in keys:
3950 arg_list = list(filter(
3951 lambda x: x is not None,
3952 [argdict.get(key.lower()) for key in variadic(key_list)]))
3954 return [arg for args in arg_list for arg in args]
3958 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3959 main_key, exe = main_key.lower(), exe.lower()
3960 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3961 keys = [f'{root_key}{k}' for k in (keys or [''])]
3962 if root_key in keys:
3964 keys.append((main_key, exe))
3965 keys.append('default')
3968 return cli_configuration_args(argdict, keys, default, use_compat)
3972 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4031 'iw': 'heb', # Replaced by he in 1989 revision
4041 'in': 'ind', # Replaced by id in 1989 revision
4156 'ji': 'yid', # Replaced by yi in 1989 revision
4164 def short2long(cls, code):
4165 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4166 return cls._lang_map.get(code[:2])
4169 def long2short(cls, code):
4170 """Convert language code from ISO 639-2/T to ISO 639-1"""
4171 for short_name, long_name in cls._lang_map.items():
4172 if long_name == code:
4177 # From http://data.okfn.org/data/core/country-list
4179 'AF': 'Afghanistan',
4180 'AX': 'Åland Islands',
4183 'AS': 'American Samoa',
4188 'AG': 'Antigua and Barbuda',
4205 'BO': 'Bolivia, Plurinational State of',
4206 'BQ': 'Bonaire, Sint Eustatius and Saba',
4207 'BA': 'Bosnia and Herzegovina',
4209 'BV': 'Bouvet Island',
4211 'IO': 'British Indian Ocean Territory',
4212 'BN': 'Brunei Darussalam',
4214 'BF': 'Burkina Faso',
4220 'KY': 'Cayman Islands',
4221 'CF': 'Central African Republic',
4225 'CX': 'Christmas Island',
4226 'CC': 'Cocos (Keeling) Islands',
4230 'CD': 'Congo, the Democratic Republic of the',
4231 'CK': 'Cook Islands',
4233 'CI': 'Côte d\'Ivoire',
4238 'CZ': 'Czech Republic',
4242 'DO': 'Dominican Republic',
4245 'SV': 'El Salvador',
4246 'GQ': 'Equatorial Guinea',
4250 'FK': 'Falkland Islands (Malvinas)',
4251 'FO': 'Faroe Islands',
4255 'GF': 'French Guiana',
4256 'PF': 'French Polynesia',
4257 'TF': 'French Southern Territories',
4272 'GW': 'Guinea-Bissau',
4275 'HM': 'Heard Island and McDonald Islands',
4276 'VA': 'Holy See (Vatican City State)',
4283 'IR': 'Iran, Islamic Republic of',
4286 'IM': 'Isle of Man',
4296 'KP': 'Korea, Democratic People\'s Republic of',
4297 'KR': 'Korea, Republic of',
4300 'LA': 'Lao People\'s Democratic Republic',
4306 'LI': 'Liechtenstein',
4310 'MK': 'Macedonia, the Former Yugoslav Republic of',
4317 'MH': 'Marshall Islands',
4323 'FM': 'Micronesia, Federated States of',
4324 'MD': 'Moldova, Republic of',
4335 'NL': 'Netherlands',
4336 'NC': 'New Caledonia',
4337 'NZ': 'New Zealand',
4342 'NF': 'Norfolk Island',
4343 'MP': 'Northern Mariana Islands',
4348 'PS': 'Palestine, State of',
4350 'PG': 'Papua New Guinea',
4353 'PH': 'Philippines',
4357 'PR': 'Puerto Rico',
4361 'RU': 'Russian Federation',
4363 'BL': 'Saint Barthélemy',
4364 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4365 'KN': 'Saint Kitts and Nevis',
4366 'LC': 'Saint Lucia',
4367 'MF': 'Saint Martin (French part)',
4368 'PM': 'Saint Pierre and Miquelon',
4369 'VC': 'Saint Vincent and the Grenadines',
4372 'ST': 'Sao Tome and Principe',
4373 'SA': 'Saudi Arabia',
4377 'SL': 'Sierra Leone',
4379 'SX': 'Sint Maarten (Dutch part)',
4382 'SB': 'Solomon Islands',
4384 'ZA': 'South Africa',
4385 'GS': 'South Georgia and the South Sandwich Islands',
4386 'SS': 'South Sudan',
4391 'SJ': 'Svalbard and Jan Mayen',
4394 'CH': 'Switzerland',
4395 'SY': 'Syrian Arab Republic',
4396 'TW': 'Taiwan, Province of China',
4398 'TZ': 'Tanzania, United Republic of',
4400 'TL': 'Timor-Leste',
4404 'TT': 'Trinidad and Tobago',
4407 'TM': 'Turkmenistan',
4408 'TC': 'Turks and Caicos Islands',
4412 'AE': 'United Arab Emirates',
4413 'GB': 'United Kingdom',
4414 'US': 'United States',
4415 'UM': 'United States Minor Outlying Islands',
4419 'VE': 'Venezuela, Bolivarian Republic of',
4421 'VG': 'Virgin Islands, British',
4422 'VI': 'Virgin Islands, U.S.',
4423 'WF': 'Wallis and Futuna',
4424 'EH': 'Western Sahara',
4428 # Not ISO 3166 codes, but used for IP blocks
4429 'AP': 'Asia/Pacific Region',
4434 def short2full(cls, code):
4435 """Convert an ISO 3166-2 country code to the corresponding full name"""
4436 return cls._country_map.get(code.upper())
4440 # Major IPv4 address blocks per country
4442 'AD': '46.172.224.0/19',
4443 'AE': '94.200.0.0/13',
4444 'AF': '149.54.0.0/17',
4445 'AG': '209.59.64.0/18',
4446 'AI': '204.14.248.0/21',
4447 'AL': '46.99.0.0/16',
4448 'AM': '46.70.0.0/15',
4449 'AO': '105.168.0.0/13',
4450 'AP': '182.50.184.0/21',
4451 'AQ': '23.154.160.0/24',
4452 'AR': '181.0.0.0/12',
4453 'AS': '202.70.112.0/20',
4454 'AT': '77.116.0.0/14',
4455 'AU': '1.128.0.0/11',
4456 'AW': '181.41.0.0/18',
4457 'AX': '185.217.4.0/22',
4458 'AZ': '5.197.0.0/16',
4459 'BA': '31.176.128.0/17',
4460 'BB': '65.48.128.0/17',
4461 'BD': '114.130.0.0/16',
4463 'BF': '102.178.0.0/15',
4464 'BG': '95.42.0.0/15',
4465 'BH': '37.131.0.0/17',
4466 'BI': '154.117.192.0/18',
4467 'BJ': '137.255.0.0/16',
4468 'BL': '185.212.72.0/23',
4469 'BM': '196.12.64.0/18',
4470 'BN': '156.31.0.0/16',
4471 'BO': '161.56.0.0/16',
4472 'BQ': '161.0.80.0/20',
4473 'BR': '191.128.0.0/12',
4474 'BS': '24.51.64.0/18',
4475 'BT': '119.2.96.0/19',
4476 'BW': '168.167.0.0/16',
4477 'BY': '178.120.0.0/13',
4478 'BZ': '179.42.192.0/18',
4479 'CA': '99.224.0.0/11',
4480 'CD': '41.243.0.0/16',
4481 'CF': '197.242.176.0/21',
4482 'CG': '160.113.0.0/16',
4483 'CH': '85.0.0.0/13',
4484 'CI': '102.136.0.0/14',
4485 'CK': '202.65.32.0/19',
4486 'CL': '152.172.0.0/14',
4487 'CM': '102.244.0.0/14',
4488 'CN': '36.128.0.0/10',
4489 'CO': '181.240.0.0/12',
4490 'CR': '201.192.0.0/12',
4491 'CU': '152.206.0.0/15',
4492 'CV': '165.90.96.0/19',
4493 'CW': '190.88.128.0/17',
4494 'CY': '31.153.0.0/16',
4495 'CZ': '88.100.0.0/14',
4497 'DJ': '197.241.0.0/17',
4498 'DK': '87.48.0.0/12',
4499 'DM': '192.243.48.0/20',
4500 'DO': '152.166.0.0/15',
4501 'DZ': '41.96.0.0/12',
4502 'EC': '186.68.0.0/15',
4503 'EE': '90.190.0.0/15',
4504 'EG': '156.160.0.0/11',
4505 'ER': '196.200.96.0/20',
4506 'ES': '88.0.0.0/11',
4507 'ET': '196.188.0.0/14',
4508 'EU': '2.16.0.0/13',
4509 'FI': '91.152.0.0/13',
4510 'FJ': '144.120.0.0/16',
4511 'FK': '80.73.208.0/21',
4512 'FM': '119.252.112.0/20',
4513 'FO': '88.85.32.0/19',
4515 'GA': '41.158.0.0/15',
4517 'GD': '74.122.88.0/21',
4518 'GE': '31.146.0.0/16',
4519 'GF': '161.22.64.0/18',
4520 'GG': '62.68.160.0/19',
4521 'GH': '154.160.0.0/12',
4522 'GI': '95.164.0.0/16',
4523 'GL': '88.83.0.0/19',
4524 'GM': '160.182.0.0/15',
4525 'GN': '197.149.192.0/18',
4526 'GP': '104.250.0.0/19',
4527 'GQ': '105.235.224.0/20',
4528 'GR': '94.64.0.0/13',
4529 'GT': '168.234.0.0/16',
4530 'GU': '168.123.0.0/16',
4531 'GW': '197.214.80.0/20',
4532 'GY': '181.41.64.0/18',
4533 'HK': '113.252.0.0/14',
4534 'HN': '181.210.0.0/16',
4535 'HR': '93.136.0.0/13',
4536 'HT': '148.102.128.0/17',
4537 'HU': '84.0.0.0/14',
4538 'ID': '39.192.0.0/10',
4539 'IE': '87.32.0.0/12',
4540 'IL': '79.176.0.0/13',
4541 'IM': '5.62.80.0/20',
4542 'IN': '117.192.0.0/10',
4543 'IO': '203.83.48.0/21',
4544 'IQ': '37.236.0.0/14',
4545 'IR': '2.176.0.0/12',
4546 'IS': '82.221.0.0/16',
4547 'IT': '79.0.0.0/10',
4548 'JE': '87.244.64.0/18',
4549 'JM': '72.27.0.0/17',
4550 'JO': '176.29.0.0/16',
4551 'JP': '133.0.0.0/8',
4552 'KE': '105.48.0.0/12',
4553 'KG': '158.181.128.0/17',
4554 'KH': '36.37.128.0/17',
4555 'KI': '103.25.140.0/22',
4556 'KM': '197.255.224.0/20',
4557 'KN': '198.167.192.0/19',
4558 'KP': '175.45.176.0/22',
4559 'KR': '175.192.0.0/10',
4560 'KW': '37.36.0.0/14',
4561 'KY': '64.96.0.0/15',
4562 'KZ': '2.72.0.0/13',
4563 'LA': '115.84.64.0/18',
4564 'LB': '178.135.0.0/16',
4565 'LC': '24.92.144.0/20',
4566 'LI': '82.117.0.0/19',
4567 'LK': '112.134.0.0/15',
4568 'LR': '102.183.0.0/16',
4569 'LS': '129.232.0.0/17',
4570 'LT': '78.56.0.0/13',
4571 'LU': '188.42.0.0/16',
4572 'LV': '46.109.0.0/16',
4573 'LY': '41.252.0.0/14',
4574 'MA': '105.128.0.0/11',
4575 'MC': '88.209.64.0/18',
4576 'MD': '37.246.0.0/16',
4577 'ME': '178.175.0.0/17',
4578 'MF': '74.112.232.0/21',
4579 'MG': '154.126.0.0/17',
4580 'MH': '117.103.88.0/21',
4581 'MK': '77.28.0.0/15',
4582 'ML': '154.118.128.0/18',
4583 'MM': '37.111.0.0/17',
4584 'MN': '49.0.128.0/17',
4585 'MO': '60.246.0.0/16',
4586 'MP': '202.88.64.0/20',
4587 'MQ': '109.203.224.0/19',
4588 'MR': '41.188.64.0/18',
4589 'MS': '208.90.112.0/22',
4590 'MT': '46.11.0.0/16',
4591 'MU': '105.16.0.0/12',
4592 'MV': '27.114.128.0/18',
4593 'MW': '102.70.0.0/15',
4594 'MX': '187.192.0.0/11',
4595 'MY': '175.136.0.0/13',
4596 'MZ': '197.218.0.0/15',
4597 'NA': '41.182.0.0/16',
4598 'NC': '101.101.0.0/18',
4599 'NE': '197.214.0.0/18',
4600 'NF': '203.17.240.0/22',
4601 'NG': '105.112.0.0/12',
4602 'NI': '186.76.0.0/15',
4603 'NL': '145.96.0.0/11',
4604 'NO': '84.208.0.0/13',
4605 'NP': '36.252.0.0/15',
4606 'NR': '203.98.224.0/19',
4607 'NU': '49.156.48.0/22',
4608 'NZ': '49.224.0.0/14',
4609 'OM': '5.36.0.0/15',
4610 'PA': '186.72.0.0/15',
4611 'PE': '186.160.0.0/14',
4612 'PF': '123.50.64.0/18',
4613 'PG': '124.240.192.0/19',
4614 'PH': '49.144.0.0/13',
4615 'PK': '39.32.0.0/11',
4616 'PL': '83.0.0.0/11',
4617 'PM': '70.36.0.0/20',
4618 'PR': '66.50.0.0/16',
4619 'PS': '188.161.0.0/16',
4620 'PT': '85.240.0.0/13',
4621 'PW': '202.124.224.0/20',
4622 'PY': '181.120.0.0/14',
4623 'QA': '37.210.0.0/15',
4624 'RE': '102.35.0.0/16',
4625 'RO': '79.112.0.0/13',
4626 'RS': '93.86.0.0/15',
4627 'RU': '5.136.0.0/13',
4628 'RW': '41.186.0.0/16',
4629 'SA': '188.48.0.0/13',
4630 'SB': '202.1.160.0/19',
4631 'SC': '154.192.0.0/11',
4632 'SD': '102.120.0.0/13',
4633 'SE': '78.64.0.0/12',
4634 'SG': '8.128.0.0/10',
4635 'SI': '188.196.0.0/14',
4636 'SK': '78.98.0.0/15',
4637 'SL': '102.143.0.0/17',
4638 'SM': '89.186.32.0/19',
4639 'SN': '41.82.0.0/15',
4640 'SO': '154.115.192.0/18',
4641 'SR': '186.179.128.0/17',
4642 'SS': '105.235.208.0/21',
4643 'ST': '197.159.160.0/19',
4644 'SV': '168.243.0.0/16',
4645 'SX': '190.102.0.0/20',
4647 'SZ': '41.84.224.0/19',
4648 'TC': '65.255.48.0/20',
4649 'TD': '154.68.128.0/19',
4650 'TG': '196.168.0.0/14',
4651 'TH': '171.96.0.0/13',
4652 'TJ': '85.9.128.0/18',
4653 'TK': '27.96.24.0/21',
4654 'TL': '180.189.160.0/20',
4655 'TM': '95.85.96.0/19',
4656 'TN': '197.0.0.0/11',
4657 'TO': '175.176.144.0/21',
4658 'TR': '78.160.0.0/11',
4659 'TT': '186.44.0.0/15',
4660 'TV': '202.2.96.0/19',
4661 'TW': '120.96.0.0/11',
4662 'TZ': '156.156.0.0/14',
4663 'UA': '37.52.0.0/14',
4664 'UG': '102.80.0.0/13',
4666 'UY': '167.56.0.0/13',
4667 'UZ': '84.54.64.0/18',
4668 'VA': '212.77.0.0/19',
4669 'VC': '207.191.240.0/21',
4670 'VE': '186.88.0.0/13',
4671 'VG': '66.81.192.0/20',
4672 'VI': '146.226.0.0/16',
4673 'VN': '14.160.0.0/11',
4674 'VU': '202.80.32.0/20',
4675 'WF': '117.20.32.0/21',
4676 'WS': '202.4.32.0/19',
4677 'YE': '134.35.0.0/16',
4678 'YT': '41.242.116.0/22',
4679 'ZA': '41.0.0.0/11',
4680 'ZM': '102.144.0.0/13',
4681 'ZW': '102.177.192.0/18',
4685 def random_ipv4(cls, code_or_block):
4686 if len(code_or_block) == 2:
4687 block = cls._country_ip_map.get(code_or_block.upper())
4691 block = code_or_block
4692 addr, preflen = block.split('/')
4693 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4694 addr_max = addr_min | (0xffffffff >> int(preflen))
4695 return str(socket.inet_ntoa(
4696 struct.pack('!L', random.randint(addr_min, addr_max))))
4699 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4700 def __init__(self, proxies=None):
4701 # Set default handlers
4702 for type in ('http', 'https'):
4703 setattr(self, '%s_open' % type,
4704 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4705 meth(r, proxy, type))
4706 urllib.request.ProxyHandler.__init__(self, proxies)
4708 def proxy_open(self, req, proxy, type):
4709 req_proxy = req.headers.get('Ytdl-request-proxy')
4710 if req_proxy is not None:
4712 del req.headers['Ytdl-request-proxy']
4714 if proxy == '__noproxy__':
4715 return None # No Proxy
4716 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4717 req.add_header('Ytdl-socks-proxy', proxy)
4718 # yt-dlp's http/https handlers do wrapping the socket with socks
4720 return urllib.request.ProxyHandler.proxy_open(
4721 self, req, proxy, type)
4724 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4725 # released into Public Domain
4726 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4728 def long_to_bytes(n, blocksize=0):
4729 """long_to_bytes(n:long, blocksize:int) : string
4730 Convert a long integer to a byte string.
4732 If optional blocksize is given and greater than zero, pad the front of the
4733 byte string with binary zeros so that the length is a multiple of
4736 # after much testing, this algorithm was deemed to be the fastest
4740 s = struct.pack('>I', n & 0xffffffff) + s
4742 # strip off leading zeros
4743 for i in range(len(s)):
4744 if s[i] != b'\000'[0]:
4747 # only happens when n == 0
4751 # add back some pad bytes. this could be done more efficiently w.r.t. the
4752 # de-padding being done above, but sigh...
4753 if blocksize > 0 and len(s) % blocksize:
4754 s = (blocksize - len(s) % blocksize) * b'\000' + s
4758 def bytes_to_long(s):
4759 """bytes_to_long(string) : long
4760 Convert a byte string to a long integer.
4762 This is (essentially) the inverse of long_to_bytes().
4767 extra = (4 - length % 4)
4768 s = b'\000' * extra + s
4769 length = length + extra
4770 for i in range(0, length, 4):
4771 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4775 def ohdave_rsa_encrypt(data, exponent, modulus):
4777 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4780 data: data to encrypt, bytes-like object
4781 exponent, modulus: parameter e and N of RSA algorithm, both integer
4782 Output: hex string of encrypted data
4784 Limitation: supports one block encryption only
4787 payload = int(binascii.hexlify(data[::-1]), 16)
4788 encrypted = pow(payload, exponent, modulus)
4789 return '%x' % encrypted
4792 def pkcs1pad(data, length):
4794 Padding input data with PKCS#1 scheme
4796 @param {int[]} data input data
4797 @param {int} length target length
4798 @returns {int[]} padded data
4800 if len(data) > length - 11:
4801 raise ValueError('Input data too
long for PKCS
#1 padding')
4803 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4804 return [0, 2] + pseudo_random
+ [0] + data
4807 def _base_n_table(n
, table
):
4808 if not table
and not n
:
4809 raise ValueError('Either table or n must be specified')
4810 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4812 if n
and n
!= len(table
):
4813 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4817 def encode_base_n(num
, n
=None, table
=None):
4818 """Convert given int to a base-n string"""
4819 table
= _base_n_table(n
, table
)
4823 result
, base
= '', len(table
)
4825 result
= table
[num
% base
] + result
4830 def decode_base_n(string
, n
=None, table
=None):
4831 """Convert given base-n string to int"""
4832 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4833 result
, base
= 0, len(table
)
4835 result
= result
* base
+ table
[char
]
4839 def decode_base(value
, digits
):
4840 write_string('DeprecationWarning: yt_dlp.utils.decode_base is deprecated '
4841 'and may be removed in a future version. Use yt_dlp.decode_base_n instead')
4842 return decode_base_n(value
, table
=digits
)
4845 def decode_packed_codes(code
):
4846 mobj
= re
.search(PACKED_CODES_RE
, code
)
4847 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4850 symbols
= symbols
.split('|')
4855 base_n_count
= encode_base_n(count
, base
)
4856 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4859 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4863 def caesar(s
, alphabet
, shift
):
4868 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4873 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4876 def parse_m3u8_attributes(attrib
):
4878 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4879 if val
.startswith('"'):
4885 def urshift(val
, n
):
4886 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4889 # Based on png2str() written by @gdkchan and improved by @yokrysty
4890 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4891 def decode_png(png_data
):
4892 # Reference: https://www.w3.org/TR/PNG/
4893 header
= png_data
[8:]
4895 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4896 raise OSError('Not a valid PNG file.')
4898 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4899 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
4904 length
= unpack_integer(header
[:4])
4907 chunk_type
= header
[:4]
4910 chunk_data
= header
[:length
]
4911 header
= header
[length
:]
4913 header
= header
[4:] # Skip CRC
4921 ihdr
= chunks
[0]['data']
4923 width
= unpack_integer(ihdr
[:4])
4924 height
= unpack_integer(ihdr
[4:8])
4928 for chunk
in chunks
:
4929 if chunk
['type'] == b
'IDAT':
4930 idat
+= chunk
['data']
4933 raise OSError('Unable to read PNG data.')
4935 decompressed_data
= bytearray(zlib
.decompress(idat
))
4940 def _get_pixel(idx
):
4945 for y
in range(height
):
4946 basePos
= y
* (1 + stride
)
4947 filter_type
= decompressed_data
[basePos
]
4951 pixels
.append(current_row
)
4953 for x
in range(stride
):
4954 color
= decompressed_data
[1 + basePos
+ x
]
4955 basex
= y
* stride
+ x
4960 left
= _get_pixel(basex
- 3)
4962 up
= _get_pixel(basex
- stride
)
4964 if filter_type
== 1: # Sub
4965 color
= (color
+ left
) & 0xff
4966 elif filter_type
== 2: # Up
4967 color
= (color
+ up
) & 0xff
4968 elif filter_type
== 3: # Average
4969 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4970 elif filter_type
== 4: # Paeth
4976 c
= _get_pixel(basex
- stride
- 3)
4984 if pa
<= pb
and pa
<= pc
:
4985 color
= (color
+ a
) & 0xff
4987 color
= (color
+ b
) & 0xff
4989 color
= (color
+ c
) & 0xff
4991 current_row
.append(color
)
4993 return width
, height
, pixels
4996 def write_xattr(path
, key
, value
):
4997 # Windows: Write xattrs to NTFS Alternate Data Streams:
4998 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4999 if compat_os_name
== 'nt':
5000 assert ':' not in key
5001 assert os
.path
.exists(path
)
5004 with open(f
'{path}:{key}', 'wb') as f
:
5006 except OSError as e
:
5007 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5010 # UNIX Method 1. Use xattrs/pyxattrs modules
5013 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5014 # Unicode arguments are not supported in pyxattr until version 0.5.0
5015 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5016 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5017 setxattr
= xattr
.set
5019 setxattr
= xattr
.setxattr
5023 setxattr(path
, key
, value
)
5024 except OSError as e
:
5025 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5028 # UNIX Method 2. Use setfattr/xattr executables
5029 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5030 else 'xattr' if check_executable('xattr', ['-h']) else None)
5032 raise XAttrUnavailableError(
5033 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5034 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5036 value
= value
.decode()
5038 _
, stderr
, returncode
= Popen
.run(
5039 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5040 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5041 except OSError as e
:
5042 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5044 raise XAttrMetadataError(returncode
, stderr
)
5047 def random_birthday(year_field
, month_field
, day_field
):
5048 start_date
= datetime
.date(1950, 1, 1)
5049 end_date
= datetime
.date(1995, 12, 31)
5050 offset
= random
.randint(0, (end_date
- start_date
).days
)
5051 random_date
= start_date
+ datetime
.timedelta(offset
)
5053 year_field
: str(random_date
.year
),
5054 month_field
: str(random_date
.month
),
5055 day_field
: str(random_date
.day
),
5059 # Templates for internet shortcut files, which are plain text files.
5060 DOT_URL_LINK_TEMPLATE
= '''\
5065 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5066 <?xml version="1.0" encoding="UTF-8"?>
5067 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5068 <plist version="1.0">
5071 \t<string>%(url)s</string>
5076 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5086 'url': DOT_URL_LINK_TEMPLATE
,
5087 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5088 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5092 def iri_to_uri(iri
):
5094 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5096 The function doesn't add an additional layer of escaping; e.g., it doesn't escape `%3C` as `%253C`. Instead, it percent-escapes characters with an underlying UTF-8 encoding *besides* those already escaped, leaving the URI intact.
5099 iri_parts
= urllib
.parse
.urlparse(iri
)
5101 if '[' in iri_parts
.netloc
:
5102 raise ValueError('IPv6 URIs are not, yet, supported.')
5103 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5105 # The `safe` argument values, that the following code uses, contain the characters that should not be percent-encoded. Everything else but letters, digits and '_.-' will be percent-encoded with an underlying UTF-8 encoding. Everything already percent-encoded will be left as is.
5108 if iri_parts
.username
:
5109 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5110 if iri_parts
.password
is not None:
5111 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5114 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5115 # The 'idna' encoding produces ASCII text.
5116 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5117 net_location
+= ':' + str(iri_parts
.port
)
5119 return urllib
.parse
.urlunparse(
5123 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5125 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5126 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5128 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5129 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5131 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5133 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5136 def to_high_limit_path(path
):
5137 if sys
.platform
in ['win32', 'cygwin']:
5138 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5139 return '\\\\?\\' + os
.path
.abspath(path
)
5144 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5145 val
= traverse_obj(obj
, *variadic(field
))
5146 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5148 return template
% func(val
)
5151 def clean_podcast_url(url
):
5152 return re
.sub(r
'''(?x)
5156 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5159 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5162 cn\.co| # https://podcorn.com/analytics-prefix/
5163 st\.fm # https://podsights.com/docs/
5168 _HEX_TABLE
= '0123456789abcdef'
5171 def random_uuidv4():
5172 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5175 def make_dir(path
, to_screen
=None):
5177 dn
= os
.path
.dirname(path
)
5178 if dn
and not os
.path
.exists(dn
):
5181 except OSError as err
:
5182 if callable(to_screen
) is not None:
5183 to_screen('unable to create directory ' + error_to_compat_str(err
))
5187 def get_executable_path():
5188 from .update
import _get_variant_and_executable_path
5190 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5193 def load_plugins(name
, suffix
, namespace
):
5195 with contextlib
.suppress(FileNotFoundError
):
5196 plugins_spec
= importlib
.util
.spec_from_file_location(
5197 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5198 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5199 sys
.modules
[plugins_spec
.name
] = plugins
5200 plugins_spec
.loader
.exec_module(plugins
)
5201 for name
in dir(plugins
):
5202 if name
in namespace
:
5204 if not name
.endswith(suffix
):
5206 klass
= getattr(plugins
, name
)
5207 classes
[name
] = namespace
[name
] = klass
5212 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
5213 casesense
=True, is_user_input
=False, traverse_string
=False):
5214 ''' Traverse nested list/dict/tuple
5215 @param path_list A list of paths which are checked one by one.
5216 Each path is a list of keys where each key is a:
5218 - string: A dictionary key
5219 - int: An index into a list
5220 - tuple: A list of keys all of which will be traversed
5221 - Ellipsis: Fetch all values in the object
5222 - Function: Takes the key and value as arguments
5223 and returns whether the key matches or not
5224 @param default Default value to return
5225 @param expected_type Only accept final value of this type (Can also be any callable)
5226 @param get_all Return all the values obtained from a path or only the first one
5227 @param casesense Whether to consider dictionary keys as case sensitive
5228 @param is_user_input Whether the keys are generated from user input. If True,
5229 strings are converted to int/slice if necessary
5230 @param traverse_string Whether to traverse inside strings. If True, any
5231 non-compatible object will also be converted into a string
5235 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
5236 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
5238 def _traverse_obj(obj
, path
, _current_depth
=0):
5240 path
= tuple(variadic(path
))
5241 for i
, key
in enumerate(path
):
5242 if None in (key
, obj
):
5244 if isinstance(key
, (list, tuple)):
5245 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
5248 obj
= (obj
.values() if isinstance(obj
, dict)
5249 else obj
if isinstance(obj
, (list, tuple, LazyList
))
5250 else str(obj
) if traverse_string
else [])
5252 depth
= max(depth
, _current_depth
)
5253 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
5255 if isinstance(obj
, (list, tuple, LazyList
)):
5256 obj
= enumerate(obj
)
5257 elif isinstance(obj
, dict):
5260 if not traverse_string
:
5264 depth
= max(depth
, _current_depth
)
5265 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if try_call(key
, args
=(k
, v
))]
5266 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
5267 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
5268 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
5271 key
= (int_or_none(key
) if ':' not in key
5272 else slice(*map(int_or_none
, key
.split(':'))))
5273 if key
== slice(None):
5274 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
5275 if not isinstance(key
, (int, slice)):
5277 if not isinstance(obj
, (list, tuple, LazyList
)):
5278 if not traverse_string
:
5287 if isinstance(expected_type
, type):
5288 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5290 type_test
= expected_type
or IDENTITY
5292 for path
in path_list
:
5294 val
= _traverse_obj(obj
, path
)
5297 for _
in range(depth
- 1):
5298 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5299 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5301 return val
if get_all
else val
[0]
5303 val
= type_test(val
)
5309 def traverse_dict(dictn
, keys
, casesense
=True):
5310 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5311 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5312 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5315 def get_first(obj
, keys
, **kwargs
):
5316 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5319 def variadic(x
, allowed_types
=(str, bytes, dict)):
5320 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5323 def time_seconds(**kwargs
):
5324 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5325 return t
.timestamp()
5328 # create a JSON Web Signature (jws) with HS256 algorithm
5329 # the resulting format is in JWS Compact Serialization
5330 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5331 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5332 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5338 header_data
.update(headers
)
5339 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5340 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5341 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5342 signature_b64
= base64
.b64encode(h
.digest())
5343 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5347 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5348 def jwt_decode_hs256(jwt
):
5349 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5350 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5354 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5358 def supports_terminal_sequences(stream
):
5359 if compat_os_name
== 'nt':
5360 if not WINDOWS_VT_MODE
:
5362 elif not os
.getenv('TERM'):
5365 return stream
.isatty()
5366 except BaseException
:
5370 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5371 if get_windows_version() < (10, 0, 10586):
5373 global WINDOWS_VT_MODE
5375 Popen
.run('', shell
=True)
5379 WINDOWS_VT_MODE
= True
5380 supports_terminal_sequences
.cache_clear()
5383 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5386 def remove_terminal_sequences(string
):
5387 return _terminal_sequences_re
.sub('', string
)
5390 def number_of_digits(number
):
5391 return len('%d' % number
)
5394 def join_nonempty(*values
, delim
='-', from_dict
=None):
5395 if from_dict
is not None:
5396 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5397 return delim
.join(map(str, filter(None, values
)))
5400 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5402 Find the largest format dimensions in terms of video width and, for each thumbnail:
5403 * Modify the URL: Match the width with the provided regex and replace with the former width
5406 This function is useful with video services that scale the provided thumbnails on demand
5408 _keys
= ('width', 'height')
5409 max_dimensions
= max(
5410 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5412 if not max_dimensions
[0]:
5416 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5417 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5418 for thumbnail
in thumbnails
5422 def parse_http_range(range):
5423 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5425 return None, None, None
5426 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5428 return None, None, None
5429 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5432 def read_stdin(what
):
5433 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5434 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5438 def determine_file_encoding(data
):
5440 Detect the text encoding used
5441 @returns (encoding, bytes to skip)
5444 # BOM marks are given priority over declarations
5445 for bom
, enc
in BOMS
:
5446 if data
.startswith(bom
):
5447 return enc
, len(bom
)
5449 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5450 # We ignore the endianness to get a good enough match
5451 data
= data
.replace(b
'\0', b
'')
5452 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5453 return mobj
.group(1).decode() if mobj
else None, 0
5460 __initialized
= False
5462 def __init__(self
, parser
, label
=None):
5463 self
.parser
, self
.label
= parser
, label
5464 self
._loaded
_paths
, self
.configs
= set(), []
5466 def init(self
, args
=None, filename
=None):
5467 assert not self
.__initialized
5468 self
.own_args
, self
.filename
= args
, filename
5469 return self
.load_configs()
5471 def load_configs(self
):
5474 location
= os
.path
.realpath(self
.filename
)
5475 directory
= os
.path
.dirname(location
)
5476 if location
in self
._loaded
_paths
:
5478 self
._loaded
_paths
.add(location
)
5480 self
.__initialized
= True
5481 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5482 self
.parsed_args
= self
.own_args
5483 for location
in opts
.config_locations
or []:
5485 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5487 location
= os
.path
.join(directory
, expand_path(location
))
5488 if os
.path
.isdir(location
):
5489 location
= os
.path
.join(location
, 'yt-dlp.conf')
5490 if not os
.path
.exists(location
):
5491 self
.parser
.error(f
'config location {location} does not exist')
5492 self
.append_config(self
.read_file(location
), location
)
5496 label
= join_nonempty(
5497 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5499 return join_nonempty(
5500 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5501 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5505 def read_file(filename
, default
=[]):
5507 optionf
= open(filename
, 'rb')
5509 return default
# silently skip if file is not present
5511 enc
, skip
= determine_file_encoding(optionf
.read(512))
5512 optionf
.seek(skip
, io
.SEEK_SET
)
5514 enc
= None # silently skip read errors
5516 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5517 contents
= optionf
.read().decode(enc
or preferredencoding())
5518 res
= shlex
.split(contents
, comments
=True)
5519 except Exception as err
:
5520 raise ValueError(f
'Unable to parse "{filename}": {err}')
5526 def hide_login_info(opts
):
5527 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5528 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5533 return m
.group('key') + '=PRIVATE'
5537 opts
= list(map(_scrub_eq
, opts
))
5538 for idx
, opt
in enumerate(opts
):
5539 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5540 opts
[idx
+ 1] = 'PRIVATE'
5543 def append_config(self
, *args
, label
=None):
5544 config
= type(self
)(self
.parser
, label
)
5545 config
._loaded
_paths
= self
._loaded
_paths
5546 if config
.init(*args
):
5547 self
.configs
.append(config
)
5551 for config
in reversed(self
.configs
):
5552 yield from config
.all_args
5553 yield from self
.parsed_args
or []
5555 def parse_known_args(self
, **kwargs
):
5556 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5558 def parse_args(self
):
5559 return self
.parser
.parse_args(self
.all_args
)
5562 class WebSocketsWrapper():
5563 """Wraps websockets module to use in non-async scopes"""
5566 def __init__(self
, url
, headers
=None, connect
=True):
5567 self
.loop
= asyncio
.new_event_loop()
5568 # XXX: "loop" is deprecated
5569 self
.conn
= websockets
.connect(
5570 url
, extra_headers
=headers
, ping_interval
=None,
5571 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5574 atexit
.register(self
.__exit
__, None, None, None)
5576 def __enter__(self
):
5578 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5581 def send(self
, *args
):
5582 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5584 def recv(self
, *args
):
5585 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5587 def __exit__(self
, type, value
, traceback
):
5589 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5592 self
._cancel
_all
_tasks
(self
.loop
)
5594 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5595 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5597 def run_with_loop(main
, loop
):
5598 if not asyncio
.iscoroutine(main
):
5599 raise ValueError(f
'a coroutine was expected, got {main!r}')
5602 return loop
.run_until_complete(main
)
5604 loop
.run_until_complete(loop
.shutdown_asyncgens())
5605 if hasattr(loop
, 'shutdown_default_executor'):
5606 loop
.run_until_complete(loop
.shutdown_default_executor())
5609 def _cancel_all_tasks(loop
):
5610 to_cancel
= asyncio
.all_tasks(loop
)
5615 for task
in to_cancel
:
5618 # XXX: "loop" is removed in python 3.10+
5619 loop
.run_until_complete(
5620 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5622 for task
in to_cancel
:
5623 if task
.cancelled():
5625 if task
.exception() is not None:
5626 loop
.call_exception_handler({
5627 'message': 'unhandled exception during asyncio.run() shutdown',
5628 'exception': task
.exception(),
5633 def merge_headers(*dicts
):
5634 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5635 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5638 def cached_method(f
):
5639 """Cache a method"""
5640 signature
= inspect
.signature(f
)
5643 def wrapper(self
, *args
, **kwargs
):
5644 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5645 bound_args
.apply_defaults()
5646 key
= tuple(bound_args
.arguments
.values())
5648 if not hasattr(self
, '__cached_method__cache'):
5649 self
.__cached
_method
__cache
= {}
5650 cache
= self
.__cached
_method
__cache
.setdefault(f
.__name
__, {})
5651 if key
not in cache
:
5652 cache
[key
] = f(self
, *args
, **kwargs
)
5657 class classproperty
:
5658 """property access for class methods"""
5660 def __init__(self
, func
):
5661 functools
.update_wrapper(self
, func
)
5664 def __get__(self
, _
, cls
):
5665 return self
.func(cls
)
5668 class Namespace(types
.SimpleNamespace
):
5669 """Immutable namespace"""
5672 return iter(self
.__dict
__.values())
5676 return self
.__dict
__.items()
5679 MEDIA_EXTENSIONS
= Namespace(
5680 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5681 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5682 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5683 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5684 thumbnails
=('jpg', 'png', 'webp'),
5685 storyboards
=('mhtml', ),
5686 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5687 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5689 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5690 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5692 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5697 for retry in RetryManager(...):
5700 except SomeException as err:
5704 attempt
, _error
= 0, None
5706 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5707 self
.retries
= _retries
or 0
5708 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5710 def _should_retry(self
):
5711 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5715 if self
._error
is NO_DEFAULT
:
5720 def error(self
, value
):
5724 while self
._should
_retry
():
5725 self
.error
= NO_DEFAULT
5729 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5732 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5733 """Utility function for reporting retries"""
5736 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5741 elif isinstance(e
, ExtractorError
):
5742 e
= remove_end(e
.cause
or e
.orig_msg
, '.')
5743 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5745 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5747 info(f
'Sleeping {delay:.2f} seconds ...')
5751 def make_archive_id(ie
, video_id
):
5752 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5753 return f
'{ie_key.lower()} {video_id}'
5757 has_certifi
= bool(certifi
)
5758 has_websockets
= bool(websockets
)