47 import xml
.etree
.ElementTree
50 from .compat
import functools
# isort: split
52 compat_etree_fromstring
,
54 compat_HTMLParseError
,
58 from .dependencies
import brotli
, certifi
, websockets
, xattr
59 from .socks
import ProxyType
, sockssocket
62 def register_socks_protocols():
63 # "Register" SOCKS protocols
64 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
65 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
66 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
67 if scheme
not in urllib
.parse
.uses_netloc
:
68 urllib
.parse
.uses_netloc
.append(scheme
)
71 # This is not clearly defined otherwise
72 compiled_regex_type
= type(re
.compile(''))
75 def random_user_agent():
76 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
117 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
120 SUPPORTED_ENCODINGS
= [
124 SUPPORTED_ENCODINGS
.append('br')
127 'User-Agent': random_user_agent(),
128 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
129 'Accept-Language': 'en-us,en;q=0.5',
130 'Sec-Fetch-Mode': 'navigate',
135 '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',
139 NO_DEFAULT
= object()
140 IDENTITY
= lambda x
: x
142 ENGLISH_MONTH_NAMES
= [
143 'January', 'February', 'March', 'April', 'May', 'June',
144 'July', 'August', 'September', 'October', 'November', 'December']
147 'en': ENGLISH_MONTH_NAMES
,
149 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
150 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
151 # these follow the genitive grammatical case (dopełniacz)
152 # some websites might be using nominative, which will require another month list
153 # https://en.wikibooks.org/wiki/Polish/Noun_cases
154 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
155 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
158 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
160 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
161 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
162 'EST': -5, 'EDT': -4, # Eastern
163 'CST': -6, 'CDT': -5, # Central
164 'MST': -7, 'MDT': -6, # Mountain
165 'PST': -8, 'PDT': -7 # Pacific
168 # needed for sanitizing filenames in restricted mode
169 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
170 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
171 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
201 '%Y-%m-%d %H:%M:%S.%f',
202 '%Y-%m-%d %H:%M:%S:%f',
205 '%Y-%m-%dT%H:%M:%SZ',
206 '%Y-%m-%dT%H:%M:%S.%fZ',
207 '%Y-%m-%dT%H:%M:%S.%f0Z',
209 '%Y-%m-%dT%H:%M:%S.%f',
212 '%b %d %Y at %H:%M:%S',
214 '%B %d %Y at %H:%M:%S',
218 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
219 DATE_FORMATS_DAY_FIRST
.extend([
229 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
230 DATE_FORMATS_MONTH_FIRST
.extend([
238 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
239 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
241 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
245 def preferredencoding():
246 """Get preferred encoding.
248 Returns the best encoding scheme for the system, based on
249 locale.getpreferredencoding() and some further tweaks.
252 pref = locale.getpreferredencoding()
260 def write_json_file(obj, fn):
261 """ Encode obj as JSON and write it to fn, atomically if possible """
263 tf = tempfile.NamedTemporaryFile(
264 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
265 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
269 json.dump(obj, tf, ensure_ascii=False)
270 if sys.platform == 'win32
':
271 # Need to remove existing file on Windows, else os.rename raises
272 # WindowsError or FileExistsError.
273 with contextlib.suppress(OSError):
275 with contextlib.suppress(OSError):
278 os.chmod(tf.name, 0o666 & ~mask)
279 os.rename(tf.name, fn)
281 with contextlib.suppress(OSError):
286 def find_xpath_attr(node, xpath, key, val=None):
287 """ Find the xpath xpath[@key=val] """
288 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
289 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
290 return node.find(expr)
292 # On python2.6 the xml.etree.ElementTree.Element methods don't support
293 # the namespace parameter
296 def xpath_with_ns(path
, ns_map
):
297 components
= [c
.split(':') for c
in path
.split('/')]
301 replaced
.append(c
[0])
304 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
305 return '/'.join(replaced
)
308 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
309 def _find_xpath(xpath
):
310 return node
.find(xpath
)
312 if isinstance(xpath
, str):
313 n
= _find_xpath(xpath
)
321 if default
is not NO_DEFAULT
:
324 name
= xpath
if name
is None else name
325 raise ExtractorError('Could not find XML element %s' % name
)
331 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
332 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
333 if n
is None or n
== default
:
336 if default
is not NO_DEFAULT
:
339 name
= xpath
if name
is None else name
340 raise ExtractorError('Could not find XML element\'s text %s' % name
)
346 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
347 n
= find_xpath_attr(node
, xpath
, key
)
349 if default
is not NO_DEFAULT
:
352 name
= f
'{xpath}[@{key}]' if name
is None else name
353 raise ExtractorError('Could not find XML attribute %s' % name
)
359 def get_element_by_id(id, html
, **kwargs
):
360 """Return the content of the tag with the specified ID in the passed HTML document"""
361 return get_element_by_attribute('id', id, html
, **kwargs
)
364 def get_element_html_by_id(id, html
, **kwargs
):
365 """Return the html of the tag with the specified ID in the passed HTML document"""
366 return get_element_html_by_attribute('id', id, html
, **kwargs
)
369 def get_element_by_class(class_name
, html
):
370 """Return the content of the first tag with the specified class in the passed HTML document"""
371 retval
= get_elements_by_class(class_name
, html
)
372 return retval
[0] if retval
else None
375 def get_element_html_by_class(class_name
, html
):
376 """Return the html of the first tag with the specified class in the passed HTML document"""
377 retval
= get_elements_html_by_class(class_name
, html
)
378 return retval
[0] if retval
else None
381 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
382 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
383 return retval
[0] if retval
else None
386 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
387 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
388 return retval
[0] if retval
else None
391 def get_elements_by_class(class_name
, html
, **kargs
):
392 """Return the content of all tags with the specified class in the passed HTML document as a list"""
393 return get_elements_by_attribute(
394 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
395 html, escape_value=False)
398 def get_elements_html_by_class(class_name, html):
399 """Return the html of all tags with the specified class in the passed HTML document as a list"""
400 return get_elements_html_by_attribute(
401 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
402 html, escape_value=False)
405 def get_elements_by_attribute(*args, **kwargs):
406 """Return the content of the tag with the specified attribute in the passed HTML document"""
407 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
410 def get_elements_html_by_attribute(*args, **kwargs):
411 """Return the html of the tag with the specified attribute in the passed HTML document"""
412 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
415 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
417 Return the text (content) and the html (whole) of the tag with the specified
418 attribute in the passed HTML document
423 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
425 value = re.escape(value) if escape_value else value
427 partial_element_re = rf'''(?x
)
429 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
430 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
433 for m in re.finditer(partial_element_re, html):
434 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
437 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
442 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
444 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
445 closing tag for the first opening tag it has encountered, and can be used
449 class HTMLBreakOnClosingTagException(Exception):
453 self.tagstack = collections.deque()
454 html.parser.HTMLParser.__init__(self)
459 def __exit__(self, *_):
463 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
464 # so data remains buffered; we no longer have any interest in it, thus
465 # override this method to discard it
468 def handle_starttag(self, tag, _):
469 self.tagstack.append(tag)
471 def handle_endtag(self, tag):
472 if not self.tagstack:
473 raise compat_HTMLParseError('no tags
in the stack
')
475 inner_tag = self.tagstack.pop()
479 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
480 if not self.tagstack:
481 raise self.HTMLBreakOnClosingTagException()
484 # XXX: This should be far less strict
485 def get_element_text_and_html_by_tag(tag, html):
487 For the first element with the specified tag in the passed HTML document
488 return its' content (text
) and the whole
element (html
)
490 def find_or_raise(haystack, needle, exc):
492 return haystack.index(needle)
495 closing_tag = f'</{tag}>'
496 whole_start = find_or_raise(
497 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
498 content_start = find_or_raise(
499 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
500 content_start += whole_start + 1
501 with HTMLBreakOnClosingTagParser() as parser:
502 parser.feed(html[whole_start:content_start])
503 if not parser.tagstack or parser.tagstack[0] != tag:
504 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
505 offset = content_start
506 while offset < len(html):
507 next_closing_tag_start = find_or_raise(
508 html[offset:], closing_tag,
509 compat_HTMLParseError(f'closing {tag} tag not found'))
510 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
512 parser.feed(html[offset:offset + next_closing_tag_end])
513 offset += next_closing_tag_end
514 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
515 return html[content_start:offset + next_closing_tag_start], \
516 html[whole_start:offset + next_closing_tag_end]
517 raise compat_HTMLParseError('unexpected end of html')
520 class HTMLAttributeParser(html.parser.HTMLParser):
521 """Trivial HTML parser to gather the attributes
for a single element
"""
525 html.parser.HTMLParser.__init__(self)
527 def handle_starttag(self, tag, attrs):
528 self.attrs = dict(attrs)
529 raise compat_HTMLParseError('done')
532 class HTMLListAttrsParser(html.parser.HTMLParser):
533 """HTML parser to gather the attributes
for the elements of a
list"""
536 html.parser.HTMLParser.__init__(self)
540 def handle_starttag(self, tag, attrs):
541 if tag == 'li' and self._level == 0:
542 self.items.append(dict(attrs))
545 def handle_endtag(self, tag):
549 def extract_attributes(html_element):
550 """Given a string
for an HTML element such
as
552 a
="foo" B
="bar" c
="&98;az" d
=boz
553 empty
= noval entity
="&"
556 Decode
and return a dictionary of attributes
.
558 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
559 'empty': '', 'noval': None, 'entity': '&',
560 'sq': '"', 'dq': '\''
563 parser = HTMLAttributeParser()
564 with contextlib.suppress(compat_HTMLParseError):
565 parser.feed(html_element)
570 def parse_list(webpage):
571 """Given a string
for an series of HTML
<li
> elements
,
572 return a dictionary of their attributes
"""
573 parser = HTMLListAttrsParser()
579 def clean_html(html):
580 """Clean an HTML snippet into a readable string
"""
582 if html is None: # Convenience for sanitizing descriptions etc.
585 html = re.sub(r'\s+', ' ', html)
586 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
587 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
589 html = re.sub('<.*?>', '', html)
590 # Replace html entities
591 html = unescapeHTML(html)
595 class LenientJSONDecoder(json.JSONDecoder):
597 def __init__(self, *args, transform_source=None, ignore_extra=False, close_objects=0, **kwargs):
598 self.transform_source, self.ignore_extra = transform_source, ignore_extra
599 self._close_attempts = 2 * close_objects
600 super().__init__(*args, **kwargs)
603 def _close_object(err):
604 doc = err.doc[:err.pos]
605 # We need to add comma first to get the correct error message
606 if err.msg.startswith('Expecting \',\''):
608 elif not doc.endswith(','):
611 if err.msg.startswith('Expecting property name'):
612 return doc[:-1] + '}'
613 elif err.msg.startswith('Expecting value'):
614 return doc[:-1] + ']'
617 if self.transform_source:
618 s = self.transform_source(s)
619 for attempt in range(self._close_attempts + 1):
621 if self.ignore_extra:
622 return self.raw_decode(s.lstrip())[0]
623 return super().decode(s)
624 except json.JSONDecodeError as e:
627 elif attempt < self._close_attempts:
628 s = self._close_object(e)
631 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
632 assert False, 'Too many attempts to decode JSON'
635 def sanitize_open(filename, open_mode):
636 """Try to
open the given filename
, and slightly tweak it
if this fails
.
638 Attempts to
open the given filename
. If this fails
, it tries to change
639 the filename slightly
, step by step
, until it
's either able to open it
640 or it fails and raises a final exception, like the standard open()
643 It returns the tuple (stream, definitive_file_name).
646 if sys.platform == 'win32
':
649 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
650 with contextlib.suppress(io.UnsupportedOperation):
651 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
652 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
654 for attempt in range(2):
657 if sys.platform == 'win32
':
658 # FIXME: An exclusive lock also locks the file from being read.
659 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
660 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
661 raise LockingUnsupportedError()
662 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
664 stream
= open(filename
, open_mode
)
665 return stream
, filename
666 except OSError as err
:
667 if attempt
or err
.errno
in (errno
.EACCES
,):
669 old_filename
, filename
= filename
, sanitize_path(filename
)
670 if old_filename
== filename
:
674 def timeconvert(timestr
):
675 """Convert RFC 2822 defined time string into system timestamp"""
677 timetuple
= email
.utils
.parsedate_tz(timestr
)
678 if timetuple
is not None:
679 timestamp
= email
.utils
.mktime_tz(timetuple
)
683 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
684 """Sanitizes a string so it could be used as part of a filename.
685 @param restricted Use a stricter subset of allowed characters
686 @param is_id Whether this is an ID that should be kept unchanged if possible.
687 If unset, yt-dlp's new sanitization rules are in effect
692 def replace_insane(char
):
693 if restricted
and char
in ACCENT_CHARS
:
694 return ACCENT_CHARS
[char
]
695 elif not restricted
and char
== '\n':
697 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
698 # Replace with their full-width unicode counterparts
699 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
700 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
703 return '' if restricted
else '\''
705 return '\0_\0-' if restricted
else '\0 \0-'
706 elif char
in '\\/|*<>':
708 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
712 # Replace look-alike Unicode glyphs
713 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
714 s
= unicodedata
.normalize('NFKC', s
)
715 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
716 result
= ''.join(map(replace_insane
, s
))
717 if is_id
is NO_DEFAULT
:
718 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
719 STRIP_RE
= r
'(?:\0.|[ _-])*'
720 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
721 result
= result
.replace('\0', '') or '_'
724 while '__' in result
:
725 result
= result
.replace('__', '_')
726 result
= result
.strip('_')
727 # Common case of "Foreign band name - English song title"
728 if restricted
and result
.startswith('-_'):
730 if result
.startswith('-'):
731 result
= '_' + result
[len('-'):]
732 result
= result
.lstrip('.')
738 def sanitize_path(s
, force
=False):
739 """Sanitizes and normalizes path on Windows"""
740 if sys
.platform
== 'win32':
742 drive_or_unc
, _
= os
.path
.splitdrive(s
)
748 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
752 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
753 for path_part
in norm_path
]
755 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
756 elif force
and s
and s
[0] == os
.path
.sep
:
757 sanitized_path
.insert(0, os
.path
.sep
)
758 return os
.path
.join(*sanitized_path
)
761 def sanitize_url(url
, *, scheme
='http'):
762 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
763 # the number of unwanted failures due to missing protocol
766 elif url
.startswith('//'):
767 return f
'{scheme}:{url}'
768 # Fix some common typos seen so far
770 # https://github.com/ytdl-org/youtube-dl/issues/15649
771 (r
'^httpss://', r
'https://'),
772 # https://bx1.be/lives/direct-tv/
773 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
775 for mistake
, fixup
in COMMON_TYPOS
:
776 if re
.match(mistake
, url
):
777 return re
.sub(mistake
, fixup
, url
)
781 def extract_basic_auth(url
):
782 parts
= urllib
.parse
.urlsplit(url
)
783 if parts
.username
is None:
785 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
786 parts
.hostname
if parts
.port
is None
787 else '%s:%d' % (parts
.hostname
, parts
.port
))))
788 auth_payload
= base64
.b64encode(
789 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
790 return url
, f
'Basic {auth_payload.decode()}'
793 def sanitized_Request(url
, *args
, **kwargs
):
794 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
795 if auth_header
is not None:
796 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
797 headers
['Authorization'] = auth_header
798 return urllib
.request
.Request(url
, *args
, **kwargs
)
802 """Expand shell variables and ~"""
803 return os
.path
.expandvars(compat_expanduser(s
))
806 def orderedSet(iterable
, *, lazy
=False):
807 """Remove all duplicates from the input iterable"""
809 seen
= [] # Do not use set since the items can be unhashable
815 return _iter() if lazy
else list(_iter())
818 def _htmlentity_transform(entity_with_semicolon
):
819 """Transforms an HTML entity to a character."""
820 entity
= entity_with_semicolon
[:-1]
822 # Known non-numeric HTML entity
823 if entity
in html
.entities
.name2codepoint
:
824 return chr(html
.entities
.name2codepoint
[entity
])
826 # TODO: HTML5 allows entities without a semicolon.
827 # E.g. 'Éric' should be decoded as 'Éric'.
828 if entity_with_semicolon
in html
.entities
.html5
:
829 return html
.entities
.html5
[entity_with_semicolon
]
831 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
833 numstr
= mobj
.group(1)
834 if numstr
.startswith('x'):
836 numstr
= '0%s' % numstr
839 # See https://github.com/ytdl-org/youtube-dl/issues/7518
840 with contextlib
.suppress(ValueError):
841 return chr(int(numstr
, base
))
843 # Unknown entity in name, return its literal representation
844 return '&%s;' % entity
850 assert isinstance(s
, str)
853 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
856 def escapeHTML(text
):
859 .replace('&', '&')
860 .replace('<', '<')
861 .replace('>', '>')
862 .replace('"', '"')
863 .replace("'", ''')
867 def process_communicate_or_kill(p
, *args
, **kwargs
):
868 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
869 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
870 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
873 class Popen(subprocess
.Popen
):
874 if sys
.platform
== 'win32':
875 _startupinfo
= subprocess
.STARTUPINFO()
876 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
881 def _fix_pyinstaller_ld_path(env
):
882 """Restore LD_LIBRARY_PATH when using PyInstaller
883 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
884 https://github.com/yt-dlp/yt-dlp/issues/4573
886 if not hasattr(sys
, '_MEIPASS'):
890 orig
= env
.get(f
'{key}_ORIG')
896 _fix('LD_LIBRARY_PATH') # Linux
897 _fix('DYLD_LIBRARY_PATH') # macOS
899 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
901 env
= os
.environ
.copy()
902 self
._fix
_pyinstaller
_ld
_path
(env
)
904 self
.__text
_mode
= kwargs
.get('encoding') or kwargs
.get('errors') or text
or kwargs
.get('universal_newlines')
906 kwargs
['universal_newlines'] = True # For 3.6 compatibility
907 kwargs
.setdefault('encoding', 'utf-8')
908 kwargs
.setdefault('errors', 'replace')
909 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
911 def communicate_or_kill(self
, *args
, **kwargs
):
913 return self
.communicate(*args
, **kwargs
)
914 except BaseException
: # Including KeyboardInterrupt
915 self
.kill(timeout
=None)
918 def kill(self
, *, timeout
=0):
921 self
.wait(timeout
=timeout
)
924 def run(cls
, *args
, timeout
=None, **kwargs
):
925 with cls(*args
, **kwargs
) as proc
:
926 default
= '' if proc
.__text
_mode
else b
''
927 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
928 return stdout
or default
, stderr
or default
, proc
.returncode
931 def get_subprocess_encoding():
932 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
933 # For subprocess calls, encode with locale encoding
934 # Refer to http://stackoverflow.com/a/9951851/35070
935 encoding
= preferredencoding()
937 encoding
= sys
.getfilesystemencoding()
943 def encodeFilename(s
, for_subprocess
=False):
944 assert isinstance(s
, str)
948 def decodeFilename(b
, for_subprocess
=False):
952 def encodeArgument(s
):
953 # Legacy code that uses byte strings
954 # Uncomment the following line after fixing all post processors
955 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
956 return s
if isinstance(s
, str) else s
.decode('ascii')
959 def decodeArgument(b
):
963 def decodeOption(optval
):
966 if isinstance(optval
, bytes):
967 optval
= optval
.decode(preferredencoding())
969 assert isinstance(optval
, str)
973 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
976 def timetuple_from_msec(msec
):
977 secs
, msec
= divmod(msec
, 1000)
978 mins
, secs
= divmod(secs
, 60)
979 hrs
, mins
= divmod(mins
, 60)
980 return _timetuple(hrs
, mins
, secs
, msec
)
983 def formatSeconds(secs
, delim
=':', msec
=False):
984 time
= timetuple_from_msec(secs
* 1000)
986 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
988 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
990 ret
= '%d' % time
.seconds
991 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
994 def _ssl_load_windows_store_certs(ssl_context
, storename
):
995 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
997 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
998 if encoding
== 'x509_asn' and (
999 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
1000 except PermissionError
:
1003 with contextlib
.suppress(ssl
.SSLError
):
1004 ssl_context
.load_verify_locations(cadata
=cert
)
1007 def make_HTTPS_handler(params
, **kwargs
):
1008 opts_check_certificate
= not params
.get('nocheckcertificate')
1009 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
1010 context
.check_hostname
= opts_check_certificate
1011 if params
.get('legacyserverconnect'):
1012 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
1013 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
1014 context
.set_ciphers('DEFAULT')
1016 sys
.version_info
< (3, 10)
1017 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
1018 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
1020 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
1021 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
1022 # in some situations [2][3].
1023 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
1024 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1025 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1026 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1027 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1028 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1029 # 4. https://peps.python.org/pep-0644/
1030 # 5. https://peps.python.org/pep-0644/#libressl-support
1031 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1032 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1033 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1035 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1036 if opts_check_certificate
:
1037 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1038 context
.load_verify_locations(cafile
=certifi
.where())
1041 context
.load_default_certs()
1042 # Work around the issue in load_default_certs when there are bad certificates. See:
1043 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1044 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1045 except ssl
.SSLError
:
1046 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1047 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1048 for storename
in ('CA', 'ROOT'):
1049 _ssl_load_windows_store_certs(context
, storename
)
1050 context
.set_default_verify_paths()
1052 client_certfile
= params
.get('client_certificate')
1055 context
.load_cert_chain(
1056 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1057 password
=params
.get('client_certificate_password'))
1058 except ssl
.SSLError
:
1059 raise YoutubeDLError('Unable to load client certificate')
1061 # Some servers may reject requests if ALPN extension is not sent. See:
1062 # https://github.com/python/cpython/issues/85140
1063 # https://github.com/yt-dlp/yt-dlp/issues/3878
1064 with contextlib
.suppress(NotImplementedError):
1065 context
.set_alpn_protocols(['http/1.1'])
1067 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1070 def bug_reports_message(before
=';'):
1071 from .update
import REPOSITORY
1073 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1074 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1076 before
= before
.rstrip()
1077 if not before
or before
.endswith(('.', '!', '?')):
1078 msg
= msg
[0].title() + msg
[1:]
1080 return (before
+ ' ' if before
else '') + msg
1083 class YoutubeDLError(Exception):
1084 """Base exception for YoutubeDL errors."""
1087 def __init__(self
, msg
=None):
1090 elif self
.msg
is None:
1091 self
.msg
= type(self
).__name
__
1092 super().__init
__(self
.msg
)
1095 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1096 if hasattr(ssl
, 'CertificateError'):
1097 network_exceptions
.append(ssl
.CertificateError
)
1098 network_exceptions
= tuple(network_exceptions
)
1101 class ExtractorError(YoutubeDLError
):
1102 """Error during info extraction."""
1104 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1105 """ tb, if given, is the original traceback (so that it can be printed out).
1106 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1108 if sys
.exc_info()[0] in network_exceptions
:
1111 self
.orig_msg
= str(msg
)
1113 self
.expected
= expected
1115 self
.video_id
= video_id
1117 self
.exc_info
= sys
.exc_info() # preserve original exception
1118 if isinstance(self
.exc_info
[1], ExtractorError
):
1119 self
.exc_info
= self
.exc_info
[1].exc_info
1120 super().__init
__(self
.__msg
)
1125 format_field(self
.ie
, None, '[%s] '),
1126 format_field(self
.video_id
, None, '%s: '),
1128 format_field(self
.cause
, None, ' (caused by %r)'),
1129 '' if self
.expected
else bug_reports_message()))
1131 def format_traceback(self
):
1132 return join_nonempty(
1133 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1134 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1137 def __setattr__(self
, name
, value
):
1138 super().__setattr
__(name
, value
)
1139 if getattr(self
, 'msg', None) and name
not in ('msg', 'args'):
1140 self
.msg
= self
.__msg
or type(self
).__name
__
1141 self
.args
= (self
.msg
, ) # Cannot be property
1144 class UnsupportedError(ExtractorError
):
1145 def __init__(self
, url
):
1147 'Unsupported URL: %s' % url
, expected
=True)
1151 class RegexNotFoundError(ExtractorError
):
1152 """Error when a regex didn't match"""
1156 class GeoRestrictedError(ExtractorError
):
1157 """Geographic restriction Error exception.
1159 This exception may be thrown when a video is not available from your
1160 geographic location due to geographic restrictions imposed by a website.
1163 def __init__(self
, msg
, countries
=None, **kwargs
):
1164 kwargs
['expected'] = True
1165 super().__init
__(msg
, **kwargs
)
1166 self
.countries
= countries
1169 class UserNotLive(ExtractorError
):
1170 """Error when a channel/user is not live"""
1172 def __init__(self
, msg
=None, **kwargs
):
1173 kwargs
['expected'] = True
1174 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1177 class DownloadError(YoutubeDLError
):
1178 """Download Error exception.
1180 This exception may be thrown by FileDownloader objects if they are not
1181 configured to continue on errors. They will contain the appropriate
1185 def __init__(self
, msg
, exc_info
=None):
1186 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1187 super().__init
__(msg
)
1188 self
.exc_info
= exc_info
1191 class EntryNotInPlaylist(YoutubeDLError
):
1192 """Entry not in playlist exception.
1194 This exception will be thrown by YoutubeDL when a requested entry
1195 is not found in the playlist info_dict
1197 msg
= 'Entry not found in info'
1200 class SameFileError(YoutubeDLError
):
1201 """Same File exception.
1203 This exception will be thrown by FileDownloader objects if they detect
1204 multiple files would have to be downloaded to the same file on disk.
1206 msg
= 'Fixed output name but more than one file to download'
1208 def __init__(self
, filename
=None):
1209 if filename
is not None:
1210 self
.msg
+= f
': {filename}'
1211 super().__init
__(self
.msg
)
1214 class PostProcessingError(YoutubeDLError
):
1215 """Post Processing exception.
1217 This exception may be raised by PostProcessor's .run() method to
1218 indicate an error in the postprocessing task.
1222 class DownloadCancelled(YoutubeDLError
):
1223 """ Exception raised when the download queue should be interrupted """
1224 msg
= 'The download was cancelled'
1227 class ExistingVideoReached(DownloadCancelled
):
1228 """ --break-on-existing triggered """
1229 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1232 class RejectedVideoReached(DownloadCancelled
):
1233 """ --break-on-reject triggered """
1234 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1237 class MaxDownloadsReached(DownloadCancelled
):
1238 """ --max-downloads limit has been reached. """
1239 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1242 class ReExtractInfo(YoutubeDLError
):
1243 """ Video info needs to be re-extracted. """
1245 def __init__(self
, msg
, expected
=False):
1246 super().__init
__(msg
)
1247 self
.expected
= expected
1250 class ThrottledDownload(ReExtractInfo
):
1251 """ Download speed below --throttled-rate. """
1252 msg
= 'The download speed is below throttle limit'
1255 super().__init
__(self
.msg
, expected
=False)
1258 class UnavailableVideoError(YoutubeDLError
):
1259 """Unavailable Format exception.
1261 This exception will be thrown when a video is requested
1262 in a format that is not available for that video.
1264 msg
= 'Unable to download video'
1266 def __init__(self
, err
=None):
1268 self
.msg
+= f
': {err}'
1269 super().__init
__(self
.msg
)
1272 class ContentTooShortError(YoutubeDLError
):
1273 """Content Too Short exception.
1275 This exception may be raised by FileDownloader objects when a file they
1276 download is too small for what the server announced first, indicating
1277 the connection was probably interrupted.
1280 def __init__(self
, downloaded
, expected
):
1281 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1283 self
.downloaded
= downloaded
1284 self
.expected
= expected
1287 class XAttrMetadataError(YoutubeDLError
):
1288 def __init__(self
, code
=None, msg
='Unknown error'):
1289 super().__init
__(msg
)
1293 # Parsing code and msg
1294 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1295 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1296 self
.reason
= 'NO_SPACE'
1297 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1298 self
.reason
= 'VALUE_TOO_LONG'
1300 self
.reason
= 'NOT_SUPPORTED'
1303 class XAttrUnavailableError(YoutubeDLError
):
1307 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1308 hc
= http_class(*args
, **kwargs
)
1309 source_address
= ydl_handler
._params
.get('source_address')
1311 if source_address
is not None:
1312 # This is to workaround _create_connection() from socket where it will try all
1313 # address data from getaddrinfo() including IPv6. This filters the result from
1314 # getaddrinfo() based on the source_address value.
1315 # This is based on the cpython socket.create_connection() function.
1316 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1317 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1318 host
, port
= address
1320 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1321 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1322 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1323 if addrs
and not ip_addrs
:
1324 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1326 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1327 % (ip_version
, source_address
[0]))
1328 for res
in ip_addrs
:
1329 af
, socktype
, proto
, canonname
, sa
= res
1332 sock
= socket
.socket(af
, socktype
, proto
)
1333 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1334 sock
.settimeout(timeout
)
1335 sock
.bind(source_address
)
1337 err
= None # Explicitly break reference cycle
1339 except OSError as _
:
1341 if sock
is not None:
1346 raise OSError('getaddrinfo returns an empty list')
1347 if hasattr(hc
, '_create_connection'):
1348 hc
._create
_connection
= _create_connection
1349 hc
.source_address
= (source_address
, 0)
1354 def handle_youtubedl_headers(headers
):
1355 filtered_headers
= headers
1357 if 'Youtubedl-no-compression' in filtered_headers
:
1358 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1359 del filtered_headers
['Youtubedl-no-compression']
1361 return filtered_headers
1364 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1365 """Handler for HTTP requests and responses.
1367 This class, when installed with an OpenerDirector, automatically adds
1368 the standard headers to every HTTP request and handles gzipped and
1369 deflated responses from web servers. If compression is to be avoided in
1370 a particular request, the original request in the program code only has
1371 to include the HTTP header "Youtubedl-no-compression", which will be
1372 removed before making the real request.
1374 Part of this code was copied from:
1376 http://techknack.net/python-urllib2-handlers/
1378 Andrew Rowls, the author of that code, agreed to release it to the
1382 def __init__(self
, params
, *args
, **kwargs
):
1383 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1384 self
._params
= params
1386 def http_open(self
, req
):
1387 conn_class
= http
.client
.HTTPConnection
1389 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1391 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1392 del req
.headers
['Ytdl-socks-proxy']
1394 return self
.do_open(functools
.partial(
1395 _create_http_connection
, self
, conn_class
, False),
1403 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1405 return zlib
.decompress(data
)
1411 return brotli
.decompress(data
)
1413 def http_request(self
, req
):
1414 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1415 # always respected by websites, some tend to give out URLs with non percent-encoded
1416 # non-ASCII characters (see telemb.py, ard.py [#3412])
1417 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1418 # To work around aforementioned issue we will replace request's original URL with
1419 # percent-encoded one
1420 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1421 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1422 url
= req
.get_full_url()
1423 url_escaped
= escape_url(url
)
1425 # Substitute URL if any change after escaping
1426 if url
!= url_escaped
:
1427 req
= update_Request(req
, url
=url_escaped
)
1429 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1430 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1431 # The dict keys are capitalized because of this bug by urllib
1432 if h
.capitalize() not in req
.headers
:
1433 req
.add_header(h
, v
)
1435 if 'Accept-encoding' not in req
.headers
:
1436 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1438 req
.headers
= handle_youtubedl_headers(req
.headers
)
1440 return super().do_request_(req
)
1442 def http_response(self
, req
, resp
):
1445 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1446 content
= resp
.read()
1447 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1449 uncompressed
= io
.BytesIO(gz
.read())
1450 except OSError as original_ioerror
:
1451 # There may be junk add the end of the file
1452 # See http://stackoverflow.com/q/4928560/35070 for details
1453 for i
in range(1, 1024):
1455 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1456 uncompressed
= io
.BytesIO(gz
.read())
1461 raise original_ioerror
1462 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1463 resp
.msg
= old_resp
.msg
1465 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1466 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1467 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1468 resp
.msg
= old_resp
.msg
1470 if resp
.headers
.get('Content-encoding', '') == 'br':
1471 resp
= urllib
.request
.addinfourl(
1472 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1473 resp
.msg
= old_resp
.msg
1474 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1475 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1476 if 300 <= resp
.code
< 400:
1477 location
= resp
.headers
.get('Location')
1479 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1480 location
= location
.encode('iso-8859-1').decode()
1481 location_escaped
= escape_url(location
)
1482 if location
!= location_escaped
:
1483 del resp
.headers
['Location']
1484 resp
.headers
['Location'] = location_escaped
1487 https_request
= http_request
1488 https_response
= http_response
1491 def make_socks_conn_class(base_class
, socks_proxy
):
1492 assert issubclass(base_class
, (
1493 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1495 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1496 if url_components
.scheme
.lower() == 'socks5':
1497 socks_type
= ProxyType
.SOCKS5
1498 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1499 socks_type
= ProxyType
.SOCKS4
1500 elif url_components
.scheme
.lower() == 'socks4a':
1501 socks_type
= ProxyType
.SOCKS4A
1503 def unquote_if_non_empty(s
):
1506 return urllib
.parse
.unquote_plus(s
)
1510 url_components
.hostname
, url_components
.port
or 1080,
1512 unquote_if_non_empty(url_components
.username
),
1513 unquote_if_non_empty(url_components
.password
),
1516 class SocksConnection(base_class
):
1518 self
.sock
= sockssocket()
1519 self
.sock
.setproxy(*proxy_args
)
1520 if isinstance(self
.timeout
, (int, float)):
1521 self
.sock
.settimeout(self
.timeout
)
1522 self
.sock
.connect((self
.host
, self
.port
))
1524 if isinstance(self
, http
.client
.HTTPSConnection
):
1525 if hasattr(self
, '_context'): # Python > 2.6
1526 self
.sock
= self
._context
.wrap_socket(
1527 self
.sock
, server_hostname
=self
.host
)
1529 self
.sock
= ssl
.wrap_socket(self
.sock
)
1531 return SocksConnection
1534 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1535 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1536 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1537 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1538 self
._params
= params
1540 def https_open(self
, req
):
1542 conn_class
= self
._https
_conn
_class
1544 if hasattr(self
, '_context'): # python > 2.6
1545 kwargs
['context'] = self
._context
1546 if hasattr(self
, '_check_hostname'): # python 3.x
1547 kwargs
['check_hostname'] = self
._check
_hostname
1549 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1551 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1552 del req
.headers
['Ytdl-socks-proxy']
1555 return self
.do_open(
1556 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1557 except urllib
.error
.URLError
as e
:
1558 if (isinstance(e
.reason
, ssl
.SSLError
)
1559 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1560 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1564 def is_path_like(f
):
1565 return isinstance(f
, (str, bytes, os
.PathLike
))
1568 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1570 See [1] for cookie file format.
1572 1. https://curl.haxx.se/docs/http-cookies.html
1574 _HTTPONLY_PREFIX
= '#HttpOnly_'
1576 _HEADER
= '''# Netscape HTTP Cookie File
1577 # This file is generated by yt-dlp. Do not edit.
1580 _CookieFileEntry
= collections
.namedtuple(
1582 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1584 def __init__(self
, filename
=None, *args
, **kwargs
):
1585 super().__init
__(None, *args
, **kwargs
)
1586 if is_path_like(filename
):
1587 filename
= os
.fspath(filename
)
1588 self
.filename
= filename
1591 def _true_or_false(cndn
):
1592 return 'TRUE' if cndn
else 'FALSE'
1594 @contextlib.contextmanager
1595 def open(self
, file, *, write
=False):
1596 if is_path_like(file):
1597 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1604 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1607 if (not ignore_discard
and cookie
.discard
1608 or not ignore_expires
and cookie
.is_expired(now
)):
1610 name
, value
= cookie
.name
, cookie
.value
1612 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1613 # with no name, whereas http.cookiejar regards it as a
1614 # cookie with no value.
1615 name
, value
= '', name
1616 f
.write('%s\n' % '\t'.join((
1618 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1620 self
._true
_or
_false
(cookie
.secure
),
1621 str_or_none(cookie
.expires
, default
=''),
1625 def save(self
, filename
=None, *args
, **kwargs
):
1627 Save cookies to a file.
1628 Code is taken from CPython 3.6
1629 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1631 if filename
is None:
1632 if self
.filename
is not None:
1633 filename
= self
.filename
1635 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1637 # Store session cookies with `expires` set to 0 instead of an empty string
1639 if cookie
.expires
is None:
1642 with self
.open(filename
, write
=True) as f
:
1643 f
.write(self
._HEADER
)
1644 self
._really
_save
(f
, *args
, **kwargs
)
1646 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1647 """Load cookies from a file."""
1648 if filename
is None:
1649 if self
.filename
is not None:
1650 filename
= self
.filename
1652 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1654 def prepare_line(line
):
1655 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1656 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1657 # comments and empty lines are fine
1658 if line
.startswith('#') or not line
.strip():
1660 cookie_list
= line
.split('\t')
1661 if len(cookie_list
) != self
._ENTRY
_LEN
:
1662 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1663 cookie
= self
._CookieFileEntry
(*cookie_list
)
1664 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1665 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1669 with self
.open(filename
) as f
:
1672 cf
.write(prepare_line(line
))
1673 except http
.cookiejar
.LoadError
as e
:
1674 if f
'{line.strip()} '[0] in '[{"':
1675 raise http
.cookiejar
.LoadError(
1676 'Cookies file must be Netscape formatted, not JSON. See '
1677 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1678 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1681 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1682 # Session cookies are denoted by either `expires` field set to
1683 # an empty string or 0. MozillaCookieJar only recognizes the former
1684 # (see [1]). So we need force the latter to be recognized as session
1685 # cookies on our own.
1686 # Session cookies may be important for cookies-based authentication,
1687 # e.g. usually, when user does not check 'Remember me' check box while
1688 # logging in on a site, some important cookies are stored as session
1689 # cookies so that not recognizing them will result in failed login.
1690 # 1. https://bugs.python.org/issue17164
1692 # Treat `expires=0` cookies as session cookies
1693 if cookie
.expires
== 0:
1694 cookie
.expires
= None
1695 cookie
.discard
= True
1698 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1699 def __init__(self
, cookiejar
=None):
1700 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1702 def http_response(self
, request
, response
):
1703 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1705 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1706 https_response
= http_response
1709 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1710 """YoutubeDL redirect handler
1712 The code is based on HTTPRedirectHandler implementation from CPython [1].
1714 This redirect handler solves two issues:
1715 - ensures redirect URL is always unicode under python 2
1716 - introduces support for experimental HTTP response status code
1717 308 Permanent Redirect [2] used by some sites [3]
1719 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1720 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1721 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1724 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1726 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1727 """Return a Request or None in response to a redirect.
1729 This is called by the http_error_30x methods when a
1730 redirection response is received. If a redirection should
1731 take place, return a new Request to allow http_error_30x to
1732 perform the redirect. Otherwise, raise HTTPError if no-one
1733 else should try to handle this url. Return None if you can't
1734 but another Handler might.
1736 m
= req
.get_method()
1737 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1738 or code
in (301, 302, 303) and m
== "POST")):
1739 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1740 # Strictly (according to RFC 2616), 301 or 302 in response to
1741 # a POST MUST NOT cause a redirection without confirmation
1742 # from the user (of urllib.request, in this case). In practice,
1743 # essentially all clients do redirect in this case, so we do
1746 # Be conciliant with URIs containing a space. This is mainly
1747 # redundant with the more complete encoding done in http_error_302(),
1748 # but it is kept for compatibility with other callers.
1749 newurl
= newurl
.replace(' ', '%20')
1751 CONTENT_HEADERS
= ("content-length", "content-type")
1752 # NB: don't use dict comprehension for python 2.6 compatibility
1753 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1755 # A 303 must either use GET or HEAD for subsequent request
1756 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1757 if code
== 303 and m
!= 'HEAD':
1759 # 301 and 302 redirects are commonly turned into a GET from a POST
1760 # for subsequent requests by browsers, so we'll do the same.
1761 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1762 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1763 if code
in (301, 302) and m
== 'POST':
1766 return urllib
.request
.Request(
1767 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1768 unverifiable
=True, method
=m
)
1771 def extract_timezone(date_str
):
1774 ^.{8,}? # >=8 char non-TZ prefix, if present
1775 (?P<tz>Z| # just the UTC Z, or
1776 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1777 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1778 [ ]? # optional space
1779 (?P<sign>\+|-) # +/-
1780 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1784 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1785 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1786 if timezone
is not None:
1787 date_str
= date_str
[:-len(m
.group('tz'))]
1788 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1790 date_str
= date_str
[:-len(m
.group('tz'))]
1791 if not m
.group('sign'):
1792 timezone
= datetime
.timedelta()
1794 sign
= 1 if m
.group('sign') == '+' else -1
1795 timezone
= datetime
.timedelta(
1796 hours
=sign
* int(m
.group('hours')),
1797 minutes
=sign
* int(m
.group('minutes')))
1798 return timezone
, date_str
1801 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1802 """ Return a UNIX timestamp from the given date """
1804 if date_str
is None:
1807 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1809 if timezone
is None:
1810 timezone
, date_str
= extract_timezone(date_str
)
1812 with contextlib
.suppress(ValueError):
1813 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1814 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1815 return calendar
.timegm(dt
.timetuple())
1818 def date_formats(day_first
=True):
1819 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1822 def unified_strdate(date_str
, day_first
=True):
1823 """Return a string with the date in the format YYYYMMDD"""
1825 if date_str
is None:
1829 date_str
= date_str
.replace(',', ' ')
1830 # Remove AM/PM + timezone
1831 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1832 _
, date_str
= extract_timezone(date_str
)
1834 for expression
in date_formats(day_first
):
1835 with contextlib
.suppress(ValueError):
1836 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1837 if upload_date
is None:
1838 timetuple
= email
.utils
.parsedate_tz(date_str
)
1840 with contextlib
.suppress(ValueError):
1841 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1842 if upload_date
is not None:
1843 return str(upload_date
)
1846 def unified_timestamp(date_str
, day_first
=True):
1847 if date_str
is None:
1850 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1851 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1853 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1854 timezone
, date_str
= extract_timezone(date_str
)
1856 # Remove AM/PM + timezone
1857 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1859 # Remove unrecognized timezones from ISO 8601 alike timestamps
1860 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1862 date_str
= date_str
[:-len(m
.group('tz'))]
1864 # Python only supports microseconds, so remove nanoseconds
1865 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1867 date_str
= m
.group(1)
1869 for expression
in date_formats(day_first
):
1870 with contextlib
.suppress(ValueError):
1871 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1872 return calendar
.timegm(dt
.timetuple())
1874 timetuple
= email
.utils
.parsedate_tz(date_str
)
1876 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1879 def determine_ext(url
, default_ext
='unknown_video'):
1880 if url
is None or '.' not in url
:
1882 guess
= url
.partition('?')[0].rpartition('.')[2]
1883 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1885 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1886 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1887 return guess
.rstrip('/')
1892 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1893 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1896 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1898 Return a datetime object from a string.
1900 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1902 @param format strftime format of DATE
1903 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1904 auto: round to the unit provided in date_str (if applicable).
1906 auto_precision
= False
1907 if precision
== 'auto':
1908 auto_precision
= True
1909 precision
= 'microsecond'
1910 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1911 if date_str
in ('now', 'today'):
1913 if date_str
== 'yesterday':
1914 return today
- datetime
.timedelta(days
=1)
1916 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1918 if match
is not None:
1919 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1920 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1921 unit
= match
.group('unit')
1922 if unit
== 'month' or unit
== 'year':
1923 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1929 delta
= datetime
.timedelta(**{unit + 's': time}
)
1930 new_date
= start_time
+ delta
1932 return datetime_round(new_date
, unit
)
1935 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1938 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1940 Return a date object from a string using datetime_from_str
1942 @param strict Restrict allowed patterns to "YYYYMMDD" and
1943 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1945 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1946 raise ValueError(f
'Invalid date format "{date_str}"')
1947 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1950 def datetime_add_months(dt
, months
):
1951 """Increment/Decrement a datetime object by months."""
1952 month
= dt
.month
+ months
- 1
1953 year
= dt
.year
+ month
// 12
1954 month
= month
% 12 + 1
1955 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1956 return dt
.replace(year
, month
, day
)
1959 def datetime_round(dt
, precision
='day'):
1961 Round a datetime object's time to a specific precision
1963 if precision
== 'microsecond':
1972 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1973 timestamp
= calendar
.timegm(dt
.timetuple())
1974 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1977 def hyphenate_date(date_str
):
1979 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1980 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1981 if match
is not None:
1982 return '-'.join(match
.groups())
1988 """Represents a time interval between two dates"""
1990 def __init__(self
, start
=None, end
=None):
1991 """start and end must be strings in the format accepted by date"""
1992 if start
is not None:
1993 self
.start
= date_from_str(start
, strict
=True)
1995 self
.start
= datetime
.datetime
.min.date()
1997 self
.end
= date_from_str(end
, strict
=True)
1999 self
.end
= datetime
.datetime
.max.date()
2000 if self
.start
> self
.end
:
2001 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
2005 """Returns a range that only contains the given day"""
2006 return cls(day
, day
)
2008 def __contains__(self
, date
):
2009 """Check if the date is in the range"""
2010 if not isinstance(date
, datetime
.date
):
2011 date
= date_from_str(date
)
2012 return self
.start
<= date
<= self
.end
2015 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
2017 def __eq__(self
, other
):
2018 return (isinstance(other
, DateRange
)
2019 and self
.start
== other
.start
and self
.end
== other
.end
)
2022 def platform_name():
2023 """ Returns the platform name as a str """
2024 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
2025 return platform
.platform()
2029 def system_identifier():
2030 python_implementation
= platform
.python_implementation()
2031 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2032 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2034 with contextlib
.suppress(OSError): # We may not have access to the executable
2035 libc_ver
= platform
.libc_ver()
2037 return 'Python %s (%s %s %s) - %s (%s%s)' % (
2038 platform
.python_version(),
2039 python_implementation
,
2041 platform
.architecture()[0],
2042 platform
.platform(),
2043 ssl
.OPENSSL_VERSION
,
2044 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2049 def get_windows_version():
2050 ''' Get Windows version. returns () if it's not running on Windows '''
2051 if compat_os_name
== 'nt':
2052 return version_tuple(platform
.win32_ver()[1])
2057 def write_string(s
, out
=None, encoding
=None):
2058 assert isinstance(s
, str)
2059 out
= out
or sys
.stderr
2061 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2062 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2064 enc
, buffer = None, out
2065 if 'b' in getattr(out
, 'mode', ''):
2066 enc
= encoding
or preferredencoding()
2067 elif hasattr(out
, 'buffer'):
2069 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2071 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2075 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2076 from . import _IN_CLI
2078 if msg
in deprecation_warning
._cache
:
2080 deprecation_warning
._cache
.add(msg
)
2082 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2083 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2086 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2089 deprecation_warning
._cache
= set()
2092 def bytes_to_intlist(bs
):
2095 if isinstance(bs
[0], int): # Python 3
2098 return [ord(c
) for c
in bs
]
2101 def intlist_to_bytes(xs
):
2104 return struct
.pack('%dB' % len(xs
), *xs
)
2107 class LockingUnsupportedError(OSError):
2108 msg
= 'File locking is not supported'
2111 super().__init
__(self
.msg
)
2114 # Cross-platform file locking
2115 if sys
.platform
== 'win32':
2117 import ctypes
.wintypes
2120 class OVERLAPPED(ctypes
.Structure
):
2122 ('Internal', ctypes
.wintypes
.LPVOID
),
2123 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2124 ('Offset', ctypes
.wintypes
.DWORD
),
2125 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2126 ('hEvent', ctypes
.wintypes
.HANDLE
),
2129 kernel32
= ctypes
.WinDLL('kernel32')
2130 LockFileEx
= kernel32
.LockFileEx
2131 LockFileEx
.argtypes
= [
2132 ctypes
.wintypes
.HANDLE
, # hFile
2133 ctypes
.wintypes
.DWORD
, # dwFlags
2134 ctypes
.wintypes
.DWORD
, # dwReserved
2135 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2136 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2137 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2139 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2140 UnlockFileEx
= kernel32
.UnlockFileEx
2141 UnlockFileEx
.argtypes
= [
2142 ctypes
.wintypes
.HANDLE
, # hFile
2143 ctypes
.wintypes
.DWORD
, # dwReserved
2144 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2145 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2146 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2148 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2149 whole_low
= 0xffffffff
2150 whole_high
= 0x7fffffff
2152 def _lock_file(f
, exclusive
, block
):
2153 overlapped
= OVERLAPPED()
2154 overlapped
.Offset
= 0
2155 overlapped
.OffsetHigh
= 0
2156 overlapped
.hEvent
= 0
2157 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2159 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2160 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2161 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2162 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2163 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2165 def _unlock_file(f
):
2166 assert f
._lock
_file
_overlapped
_p
2167 handle
= msvcrt
.get_osfhandle(f
.fileno())
2168 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2169 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2175 def _lock_file(f
, exclusive
, block
):
2176 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2178 flags |
= fcntl
.LOCK_NB
2180 fcntl
.flock(f
, flags
)
2181 except BlockingIOError
:
2183 except OSError: # AOSP does not have flock()
2184 fcntl
.lockf(f
, flags
)
2186 def _unlock_file(f
):
2188 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2190 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2194 def _lock_file(f
, exclusive
, block
):
2195 raise LockingUnsupportedError()
2197 def _unlock_file(f
):
2198 raise LockingUnsupportedError()
2204 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2205 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2206 raise NotImplementedError(mode
)
2207 self
.mode
, self
.block
= mode
, block
2209 writable
= any(f
in mode
for f
in 'wax+')
2210 readable
= any(f
in mode
for f
in 'r+')
2211 flags
= functools
.reduce(operator
.ior
, (
2212 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2213 getattr(os
, 'O_BINARY', 0), # Windows only
2214 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2215 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2216 os
.O_APPEND
if 'a' in mode
else 0,
2217 os
.O_EXCL
if 'x' in mode
else 0,
2218 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2221 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2223 def __enter__(self
):
2224 exclusive
= 'r' not in self
.mode
2226 _lock_file(self
.f
, exclusive
, self
.block
)
2231 if 'w' in self
.mode
:
2234 except OSError as e
:
2236 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2237 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2246 _unlock_file(self
.f
)
2250 def __exit__(self
, *_
):
2259 def __getattr__(self
, attr
):
2260 return getattr(self
.f
, attr
)
2267 def get_filesystem_encoding():
2268 encoding
= sys
.getfilesystemencoding()
2269 return encoding
if encoding
is not None else 'utf-8'
2272 def shell_quote(args
):
2274 encoding
= get_filesystem_encoding()
2276 if isinstance(a
, bytes):
2277 # We may get a filename encoded with 'encodeFilename'
2278 a
= a
.decode(encoding
)
2279 quoted_args
.append(compat_shlex_quote(a
))
2280 return ' '.join(quoted_args
)
2283 def smuggle_url(url
, data
):
2284 """ Pass additional data in a URL for internal use. """
2286 url
, idata
= unsmuggle_url(url
, {})
2288 sdata
= urllib
.parse
.urlencode(
2289 {'__youtubedl_smuggle': json.dumps(data)}
)
2290 return url
+ '#' + sdata
2293 def unsmuggle_url(smug_url
, default
=None):
2294 if '#__youtubedl_smuggle' not in smug_url
:
2295 return smug_url
, default
2296 url
, _
, sdata
= smug_url
.rpartition('#')
2297 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2298 data
= json
.loads(jsond
)
2302 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2303 """ Formats numbers with decimal sufixes like K, M, etc """
2304 num
, factor
= float_or_none(num
), float(factor
)
2305 if num
is None or num
< 0:
2307 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2308 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2309 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2311 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2312 converted
= num
/ (factor
** exponent
)
2313 return fmt
% (converted
, suffix
)
2316 def format_bytes(bytes):
2317 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2320 def lookup_unit_table(unit_table
, s
, strict
=False):
2321 num_re
= NUMBER_RE
if strict
else NUMBER_RE
.replace(R
'\.', '[,.]')
2322 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2323 m
= (re
.fullmatch
if strict
else re
.match
)(
2324 rf
'(?P<num>{num_re})\s*(?P<unit>{units_re})\b', s
)
2328 num
= float(m
.group('num').replace(',', '.'))
2329 mult
= unit_table
[m
.group('unit')]
2330 return round(num
* mult
)
2334 """Parse a string indicating a byte quantity into an integer"""
2335 return lookup_unit_table(
2336 {u: 1024**i for i, u in enumerate(['', *'KMGTPEZY'])}
,
2337 s
.upper(), strict
=True)
2340 def parse_filesize(s
):
2344 # The lower-case forms are of course incorrect and unofficial,
2345 # but we support those too
2362 'megabytes': 1000 ** 2,
2363 'mebibytes': 1024 ** 2,
2369 'gigabytes': 1000 ** 3,
2370 'gibibytes': 1024 ** 3,
2376 'terabytes': 1000 ** 4,
2377 'tebibytes': 1024 ** 4,
2383 'petabytes': 1000 ** 5,
2384 'pebibytes': 1024 ** 5,
2390 'exabytes': 1000 ** 6,
2391 'exbibytes': 1024 ** 6,
2397 'zettabytes': 1000 ** 7,
2398 'zebibytes': 1024 ** 7,
2404 'yottabytes': 1000 ** 8,
2405 'yobibytes': 1024 ** 8,
2408 return lookup_unit_table(_UNIT_TABLE
, s
)
2415 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2417 if re
.match(r
'^[\d,.]+$', s
):
2418 return str_to_int(s
)
2431 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2435 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2437 return str_to_int(mobj
.group(1))
2440 def parse_resolution(s
, *, lenient
=False):
2445 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2447 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2450 'width': int(mobj
.group('w')),
2451 'height': int(mobj
.group('h')),
2454 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2456 return {'height': int(mobj.group(1))}
2458 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2460 return {'height': int(mobj.group(1)) * 540}
2465 def parse_bitrate(s
):
2466 if not isinstance(s
, str):
2468 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2470 return int(mobj
.group(1))
2473 def month_by_name(name
, lang
='en'):
2474 """ Return the number of a month by (locale-independently) English name """
2476 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2479 return month_names
.index(name
) + 1
2484 def month_by_abbreviation(abbrev
):
2485 """ Return the number of a month by (locale-independently) English
2489 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2494 def fix_xml_ampersands(xml_str
):
2495 """Replace all the '&' by '&' in XML"""
2497 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2502 def setproctitle(title
):
2503 assert isinstance(title
, str)
2505 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2512 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2516 # LoadLibrary in Windows Python 2.7.13 only expects
2517 # a bytestring, but since unicode_literals turns
2518 # every string into a unicode string, it fails.
2520 title_bytes
= title
.encode()
2521 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2522 buf
.value
= title_bytes
2524 libc
.prctl(15, buf
, 0, 0, 0)
2525 except AttributeError:
2526 return # Strange libc, just skip this
2529 def remove_start(s
, start
):
2530 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2533 def remove_end(s
, end
):
2534 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2537 def remove_quotes(s
):
2538 if s
is None or len(s
) < 2:
2540 for quote
in ('"', "'", ):
2541 if s
[0] == quote
and s
[-1] == quote
:
2546 def get_domain(url
):
2548 This implementation is inconsistent, but is kept for compatibility.
2549 Use this only for "webpage_url_domain"
2551 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2554 def url_basename(url
):
2555 path
= urllib
.parse
.urlparse(url
).path
2556 return path
.strip('/').split('/')[-1]
2560 return re
.match(r
'https?://[^?#]+/', url
).group()
2563 def urljoin(base
, path
):
2564 if isinstance(path
, bytes):
2565 path
= path
.decode()
2566 if not isinstance(path
, str) or not path
:
2568 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2570 if isinstance(base
, bytes):
2571 base
= base
.decode()
2572 if not isinstance(base
, str) or not re
.match(
2573 r
'^(?:https?:)?//', base
):
2575 return urllib
.parse
.urljoin(base
, path
)
2578 class HEADRequest(urllib
.request
.Request
):
2579 def get_method(self
):
2583 class PUTRequest(urllib
.request
.Request
):
2584 def get_method(self
):
2588 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2589 if get_attr
and v
is not None:
2590 v
= getattr(v
, get_attr
, None)
2592 return int(v
) * invscale
// scale
2593 except (ValueError, TypeError, OverflowError):
2597 def str_or_none(v
, default
=None):
2598 return default
if v
is None else str(v
)
2601 def str_to_int(int_str
):
2602 """ A more relaxed version of int_or_none """
2603 if isinstance(int_str
, int):
2605 elif isinstance(int_str
, str):
2606 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2607 return int_or_none(int_str
)
2610 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2614 return float(v
) * invscale
/ scale
2615 except (ValueError, TypeError):
2619 def bool_or_none(v
, default
=None):
2620 return v
if isinstance(v
, bool) else default
2623 def strip_or_none(v
, default
=None):
2624 return v
.strip() if isinstance(v
, str) else default
2627 def url_or_none(url
):
2628 if not url
or not isinstance(url
, str):
2631 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2634 def request_to_url(req
):
2635 if isinstance(req
, urllib
.request
.Request
):
2636 return req
.get_full_url()
2641 def strftime_or_none(timestamp
, date_format
, default
=None):
2642 datetime_object
= None
2644 if isinstance(timestamp
, (int, float)): # unix timestamp
2645 # Using naive datetime here can break timestamp() in Windows
2646 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2647 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2648 elif isinstance(timestamp
, str): # assume YYYYMMDD
2649 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2650 date_format
= re
.sub( # Support %s on windows
2651 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2652 return datetime_object
.strftime(date_format
)
2653 except (ValueError, TypeError, AttributeError):
2657 def parse_duration(s
):
2658 if not isinstance(s
, str):
2664 days
, hours
, mins
, secs
, ms
= [None] * 5
2665 m
= re
.match(r
'''(?x)
2667 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2668 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2669 (?P<ms>[.:][0-9]+)?Z?$
2672 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2677 [0-9]+\s*y(?:ears?)?,?\s*
2680 [0-9]+\s*m(?:onths?)?,?\s*
2683 [0-9]+\s*w(?:eeks?)?,?\s*
2686 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2690 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2693 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2696 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2699 days
, hours
, mins
, secs
, ms
= m
.groups()
2701 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2703 hours
, mins
= m
.groups()
2708 ms
= ms
.replace(':', '.')
2709 return sum(float(part
or 0) * mult
for part
, mult
in (
2710 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2713 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2714 name
, real_ext
= os
.path
.splitext(filename
)
2716 f
'{name}.{ext}{real_ext}'
2717 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2718 else f
'{filename}.{ext}')
2721 def replace_extension(filename
, ext
, expected_real_ext
=None):
2722 name
, real_ext
= os
.path
.splitext(filename
)
2723 return '{}.{}'.format(
2724 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2728 def check_executable(exe
, args
=[]):
2729 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2730 args can be a list of arguments for a short output (like -version) """
2732 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2738 def _get_exe_version_output(exe
, args
):
2740 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2741 # SIGTTOU if yt-dlp is run in the background.
2742 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2743 stdout
, _
, ret
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2744 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2752 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2753 assert isinstance(output
, str)
2754 if version_re
is None:
2755 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2756 m
= re
.search(version_re
, output
)
2763 def get_exe_version(exe
, args
=['--version'],
2764 version_re
=None, unrecognized
=('present', 'broken')):
2765 """ Returns the version of the specified executable,
2766 or False if the executable is not present """
2767 unrecognized
= variadic(unrecognized
)
2768 assert len(unrecognized
) in (1, 2)
2769 out
= _get_exe_version_output(exe
, args
)
2771 return unrecognized
[-1]
2772 return out
and detect_exe_version(out
, version_re
, unrecognized
[0])
2775 def frange(start
=0, stop
=None, step
=1):
2778 start
, stop
= 0, start
2779 sign
= [-1, 1][step
> 0] if step
else 0
2780 while sign
* start
< sign
* stop
:
2785 class LazyList(collections
.abc
.Sequence
):
2786 """Lazy immutable list from an iterable
2787 Note that slices of a LazyList are lists and not LazyList"""
2789 class IndexError(IndexError):
2792 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2793 self
._iterable
= iter(iterable
)
2794 self
._cache
= [] if _cache
is None else _cache
2795 self
._reversed
= reverse
2799 # We need to consume the entire iterable to iterate in reverse
2800 yield from self
.exhaust()
2802 yield from self
._cache
2803 for item
in self
._iterable
:
2804 self
._cache
.append(item
)
2808 self
._cache
.extend(self
._iterable
)
2809 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2813 """Evaluate the entire iterable"""
2814 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2817 def _reverse_index(x
):
2818 return None if x
is None else ~x
2820 def __getitem__(self
, idx
):
2821 if isinstance(idx
, slice):
2823 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2824 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2825 elif isinstance(idx
, int):
2827 idx
= self
._reverse
_index
(idx
)
2828 start
, stop
, step
= idx
, idx
, 0
2830 raise TypeError('indices must be integers or slices')
2831 if ((start
or 0) < 0 or (stop
or 0) < 0
2832 or (start
is None and step
< 0)
2833 or (stop
is None and step
> 0)):
2834 # We need to consume the entire iterable to be able to slice from the end
2835 # Obviously, never use this with infinite iterables
2838 return self
._cache
[idx
]
2839 except IndexError as e
:
2840 raise self
.IndexError(e
) from e
2841 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2843 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2845 return self
._cache
[idx
]
2846 except IndexError as e
:
2847 raise self
.IndexError(e
) from e
2851 self
[-1] if self
._reversed
else self
[0]
2852 except self
.IndexError:
2858 return len(self
._cache
)
2860 def __reversed__(self
):
2861 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2864 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2867 # repr and str should mimic a list. So we exhaust the iterable
2868 return repr(self
.exhaust())
2871 return repr(self
.exhaust())
2876 class IndexError(IndexError):
2880 # This is only useful for tests
2881 return len(self
.getslice())
2883 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2884 self
._pagefunc
= pagefunc
2885 self
._pagesize
= pagesize
2886 self
._pagecount
= float('inf')
2887 self
._use
_cache
= use_cache
2890 def getpage(self
, pagenum
):
2891 page_results
= self
._cache
.get(pagenum
)
2892 if page_results
is None:
2893 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2895 self
._cache
[pagenum
] = page_results
2898 def getslice(self
, start
=0, end
=None):
2899 return list(self
._getslice
(start
, end
))
2901 def _getslice(self
, start
, end
):
2902 raise NotImplementedError('This method must be implemented by subclasses')
2904 def __getitem__(self
, idx
):
2905 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2906 if not isinstance(idx
, int) or idx
< 0:
2907 raise TypeError('indices must be non-negative integers')
2908 entries
= self
.getslice(idx
, idx
+ 1)
2910 raise self
.IndexError()
2914 class OnDemandPagedList(PagedList
):
2915 """Download pages until a page with less than maximum results"""
2917 def _getslice(self
, start
, end
):
2918 for pagenum
in itertools
.count(start
// self
._pagesize
):
2919 firstid
= pagenum
* self
._pagesize
2920 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2921 if start
>= nextfirstid
:
2925 start
% self
._pagesize
2926 if firstid
<= start
< nextfirstid
2929 ((end
- 1) % self
._pagesize
) + 1
2930 if (end
is not None and firstid
<= end
<= nextfirstid
)
2934 page_results
= self
.getpage(pagenum
)
2936 self
._pagecount
= pagenum
- 1
2938 if startv
!= 0 or endv
is not None:
2939 page_results
= page_results
[startv
:endv
]
2940 yield from page_results
2942 # A little optimization - if current page is not "full", ie. does
2943 # not contain page_size videos then we can assume that this page
2944 # is the last one - there are no more ids on further pages -
2945 # i.e. no need to query again.
2946 if len(page_results
) + startv
< self
._pagesize
:
2949 # If we got the whole page, but the next page is not interesting,
2950 # break out early as well
2951 if end
== nextfirstid
:
2955 class InAdvancePagedList(PagedList
):
2956 """PagedList with total number of pages known in advance"""
2958 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2959 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2960 self
._pagecount
= pagecount
2962 def _getslice(self
, start
, end
):
2963 start_page
= start
// self
._pagesize
2964 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2965 skip_elems
= start
- start_page
* self
._pagesize
2966 only_more
= None if end
is None else end
- start
2967 for pagenum
in range(start_page
, end_page
):
2968 page_results
= self
.getpage(pagenum
)
2970 page_results
= page_results
[skip_elems
:]
2972 if only_more
is not None:
2973 if len(page_results
) < only_more
:
2974 only_more
-= len(page_results
)
2976 yield from page_results
[:only_more
]
2978 yield from page_results
2981 class PlaylistEntries
:
2982 MissingEntry
= object()
2983 is_exhausted
= False
2985 def __init__(self
, ydl
, info_dict
):
2988 # _entries must be assigned now since infodict can change during iteration
2989 entries
= info_dict
.get('entries')
2991 raise EntryNotInPlaylist('There are no entries')
2992 elif isinstance(entries
, list):
2993 self
.is_exhausted
= True
2995 requested_entries
= info_dict
.get('requested_entries')
2996 self
.is_incomplete
= requested_entries
is not None
2997 if self
.is_incomplete
:
2998 assert self
.is_exhausted
2999 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
3000 for i
, entry
in zip(requested_entries
, entries
):
3001 self
._entries
[i
- 1] = entry
3002 elif isinstance(entries
, (list, PagedList
, LazyList
)):
3003 self
._entries
= entries
3005 self
._entries
= LazyList(entries
)
3007 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
3008 (?P<start>[+-]?\d+)?
3010 (?P<end>[+-]?\d+|inf(?:inite)?)?
3011 (?::(?P<step>[+-]?\d+))?
3015 def parse_playlist_items(cls
, string
):
3016 for segment
in string
.split(','):
3018 raise ValueError('There is two or more consecutive commas')
3019 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
3021 raise ValueError(f
'{segment!r} is not a valid specification')
3022 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
3023 if int_or_none(step
) == 0:
3024 raise ValueError(f
'Step in {segment!r} cannot be zero')
3025 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
3027 def get_requested_items(self
):
3028 playlist_items
= self
.ydl
.params
.get('playlist_items')
3029 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
3030 playlist_end
= self
.ydl
.params
.get('playlistend')
3031 # For backwards compatibility, interpret -1 as whole list
3032 if playlist_end
in (-1, None):
3034 if not playlist_items
:
3035 playlist_items
= f
'{playlist_start}:{playlist_end}'
3036 elif playlist_start
!= 1 or playlist_end
:
3037 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
3039 for index
in self
.parse_playlist_items(playlist_items
):
3040 for i
, entry
in self
[index
]:
3045 # TODO: Add auto-generated fields
3046 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3047 except (ExistingVideoReached
, RejectedVideoReached
):
3050 def get_full_count(self
):
3051 if self
.is_exhausted
and not self
.is_incomplete
:
3053 elif isinstance(self
._entries
, InAdvancePagedList
):
3054 if self
._entries
._pagesize
== 1:
3055 return self
._entries
._pagecount
3057 @functools.cached_property
3059 if isinstance(self
._entries
, list):
3062 entry
= self
._entries
[i
]
3064 entry
= self
.MissingEntry
3065 if not self
.is_incomplete
:
3066 raise self
.IndexError()
3067 if entry
is self
.MissingEntry
:
3068 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
3073 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3074 except (LazyList
.IndexError, PagedList
.IndexError):
3075 raise self
.IndexError()
3078 def __getitem__(self
, idx
):
3079 if isinstance(idx
, int):
3080 idx
= slice(idx
, idx
)
3082 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3083 step
= 1 if idx
.step
is None else idx
.step
3084 if idx
.start
is None:
3085 start
= 0 if step
> 0 else len(self
) - 1
3087 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3089 # NB: Do not call len(self) when idx == [:]
3090 if idx
.stop
is None:
3091 stop
= 0 if step
< 0 else float('inf')
3093 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3094 stop
+= [-1, 1][step
> 0]
3096 for i
in frange(start
, stop
, step
):
3100 entry
= self
._getter
(i
)
3101 except self
.IndexError:
3102 self
.is_exhausted
= True
3109 return len(tuple(self
[:]))
3111 class IndexError(IndexError):
3115 def uppercase_escape(s
):
3116 unicode_escape
= codecs
.getdecoder('unicode_escape')
3118 r
'\\U[0-9a-fA-F]{8}',
3119 lambda m
: unicode_escape(m
.group(0))[0],
3123 def lowercase_escape(s
):
3124 unicode_escape
= codecs
.getdecoder('unicode_escape')
3126 r
'\\u[0-9a-fA-F]{4}',
3127 lambda m
: unicode_escape(m
.group(0))[0],
3131 def escape_rfc3986(s
):
3132 """Escape non-ASCII characters as suggested by RFC 3986"""
3133 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3136 def escape_url(url
):
3137 """Escape URL as suggested by RFC 3986"""
3138 url_parsed
= urllib
.parse
.urlparse(url
)
3139 return url_parsed
._replace
(
3140 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3141 path
=escape_rfc3986(url_parsed
.path
),
3142 params
=escape_rfc3986(url_parsed
.params
),
3143 query
=escape_rfc3986(url_parsed
.query
),
3144 fragment
=escape_rfc3986(url_parsed
.fragment
)
3148 def parse_qs(url
, **kwargs
):
3149 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3152 def read_batch_urls(batch_fd
):
3154 if not isinstance(url
, str):
3155 url
= url
.decode('utf-8', 'replace')
3156 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3157 for bom
in BOM_UTF8
:
3158 if url
.startswith(bom
):
3159 url
= url
[len(bom
):]
3161 if not url
or url
.startswith(('#', ';', ']')):
3163 # "#" cannot be stripped out since it is part of the URI
3164 # However, it can be safely stripped out if following a whitespace
3165 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3167 with contextlib
.closing(batch_fd
) as fd
:
3168 return [url
for url
in map(fixup
, fd
) if url
]
3171 def urlencode_postdata(*args
, **kargs
):
3172 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3175 def update_url(url
, *, query_update
=None, **kwargs
):
3176 """Replace URL components specified by kwargs
3177 @param url str or parse url tuple
3178 @param query_update update query
3181 if isinstance(url
, str):
3182 if not kwargs
and not query_update
:
3185 url
= urllib
.parse
.urlparse(url
)
3187 assert 'query' not in kwargs
, 'query_update and query cannot be specified at the same time'
3188 kwargs
['query'] = urllib
.parse
.urlencode({
3189 **urllib
.parse
.parse_qs(url
.query
),
3192 return urllib
.parse
.urlunparse(url
._replace
(**kwargs
))
3195 def update_url_query(url
, query
):
3196 return update_url(url
, query_update
=query
)
3199 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3200 req_headers
= req
.headers
.copy()
3201 req_headers
.update(headers
or {})
3202 req_data
= data
or req
.data
3203 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3204 req_get_method
= req
.get_method()
3205 if req_get_method
== 'HEAD':
3206 req_type
= HEADRequest
3207 elif req_get_method
== 'PUT':
3208 req_type
= PUTRequest
3210 req_type
= urllib
.request
.Request
3212 req_url
, data
=req_data
, headers
=req_headers
,
3213 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3214 if hasattr(req
, 'timeout'):
3215 new_req
.timeout
= req
.timeout
3219 def _multipart_encode_impl(data
, boundary
):
3220 content_type
= 'multipart/form-data; boundary=%s' % boundary
3223 for k
, v
in data
.items():
3224 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3225 if isinstance(k
, str):
3227 if isinstance(v
, str):
3229 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3230 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3231 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3232 if boundary
.encode('ascii') in content
:
3233 raise ValueError('Boundary overlaps with data')
3236 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3238 return out
, content_type
3241 def multipart_encode(data
, boundary
=None):
3243 Encode a dict to RFC 7578-compliant form-data
3246 A dict where keys and values can be either Unicode or bytes-like
3249 If specified a Unicode object, it's used as the boundary. Otherwise
3250 a random boundary is generated.
3252 Reference: https://tools.ietf.org/html/rfc7578
3254 has_specified_boundary
= boundary
is not None
3257 if boundary
is None:
3258 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3261 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3264 if has_specified_boundary
:
3268 return out
, content_type
3271 def variadic(x
, allowed_types
=(str, bytes, dict)):
3272 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3275 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3276 for val
in map(d
.get
, variadic(key_or_keys
)):
3277 if val
is not None and (val
or not skip_false_values
):
3282 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3285 val
= f(*args
, **kwargs
)
3286 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3289 if expected_type
is None or isinstance(val
, expected_type
):
3293 def try_get(src
, getter
, expected_type
=None):
3294 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3297 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3298 return {k: v for k, v in dct.items() if cndn(k, v)}
3301 def merge_dicts(*dicts
):
3303 for a_dict
in dicts
:
3304 for k
, v
in a_dict
.items():
3305 if (v
is not None and k
not in merged
3306 or isinstance(v
, str) and merged
[k
] == ''):
3311 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3312 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3324 TV_PARENTAL_GUIDELINES
= {
3334 def parse_age_limit(s
):
3335 # isinstance(False, int) is True. So type() must be used instead
3336 if type(s
) is int: # noqa: E721
3337 return s
if 0 <= s
<= 21 else None
3338 elif not isinstance(s
, str):
3340 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3342 return int(m
.group('age'))
3345 return US_RATINGS
[s
]
3346 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3348 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3352 def strip_jsonp(code
):
3355 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3356 (?:\s*&&\s*(?P=func_name))?
3357 \s*\(\s*(?P<callback_data>.*)\);?
3358 \s*?(?://[^\n]*)*$''',
3359 r
'\g<callback_data>', code
)
3362 def js_to_json(code
, vars={}, *, strict
=False):
3363 # vars is a dict of var, val pairs to substitute
3364 STRING_QUOTES
= '\'"'
3365 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3366 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3367 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3369 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3370 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3373 def process_escape(match
):
3374 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3375 escape
= match
.group(1) or match
.group(2)
3377 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3378 else R
'\u00' if escape
== 'x'
3379 else '' if escape
== '\n'
3384 if v
in ('true', 'false', 'null'):
3386 elif v
in ('undefined', 'void 0'):
3388 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3391 if v
[0] in STRING_QUOTES
:
3392 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3393 return f
'"{escaped}"'
3395 for regex
, base
in INTEGER_TABLE
:
3396 im
= re
.match(regex
, v
)
3398 i
= int(im
.group(1), base
)
3399 return f
'"{i}":' if v
.endswith(':') else str(i
)
3405 except json
.JSONDecodeError
:
3406 return json
.dumps(vars[v
])
3413 raise ValueError(f
'Unknown value: {v}')
3415 def create_map(mobj
):
3416 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3418 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3420 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3421 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3422 code
= re
.sub(r
'parseInt\([^\d]+(\d+)[^\d]+\)', r
'\1', code
)
3423 code
= re
.sub(r
'\(function\([^)]*\)\s*\{[^}]*\}\s*\)\s*\(\s*(["\'][^
)]*["\'])\s*\)', r'\1', code)
3425 return re.sub(rf'''(?sx)
3427 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3428 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3429 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3430 [0-9]+(?={SKIP_RE}:)|
3435 def qualities(quality_ids):
3436 """ Get a numeric quality value out of a list of possible values """
3439 return quality_ids.index(qid)
3445 POSTPROCESS_WHEN = ('pre_process', 'after_filter', 'video', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3449 'default': '%(title)s [%(id)s].%(ext)s',
3450 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3456 'description': 'description',
3457 'annotation': 'annotations.xml',
3458 'infojson': 'info.json',
3461 'pl_thumbnail': None,
3462 'pl_description': 'description',
3463 'pl_infojson': 'info.json',
3466 # As of [1] format syntax is:
3467 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3468 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3469 STR_FORMAT_RE_TMPL = r'''(?x)
3470 (?<!%)(?P<prefix>(?:%%)*)
3472 (?P<has_key>\((?P<key>{0})\))?
3474 (?P<conversion>[#0\-+ ]+)?
3476 (?P<precision>\.\d+)?
3477 (?P<len_mod>[hlL])? # unused in python
3478 {1} # conversion type
3483 STR_FORMAT_TYPES = 'diouxXeEfFgGcrs'
3486 def limit_length(s, length):
3487 """ Add ellipses to overly long strings """
3492 return s[:length - len(ELLIPSES)] + ELLIPSES
3496 def version_tuple(v):
3497 return tuple(int(e) for e in re.split(r'[-.]', v))
3500 def is_outdated_version(version, limit, assume_new=True):
3502 return not assume_new
3504 return version_tuple(version) < version_tuple(limit)
3506 return not assume_new
3509 def ytdl_is_updateable():
3510 """ Returns if yt-dlp can be updated with -U """
3512 from .update import is_non_updateable
3514 return not is_non_updateable()
3517 def args_to_str(args):
3518 # Get a short string representation for a subprocess command
3519 return ' '.join(compat_shlex_quote(a) for a in args)
3522 def error_to_compat_str(err):
3526 def error_to_str(err):
3527 return f'{type(err).__name__}: {err}'
3530 def mimetype2ext(mt, default=NO_DEFAULT):
3531 if not isinstance(mt, str):
3532 if default is not NO_DEFAULT:
3548 'x-matroska': 'mkv',
3550 'x-mp4-fragmented': 'mp4',
3555 # application (streaming playlists)
3559 'vnd.apple.mpegurl': 'm3u8',
3560 'vnd.ms-sstr+xml': 'ism',
3561 'x-mpegurl': 'm3u8',
3565 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3.
3566 # Using .mp3 as it's the most popular one
3567 'audio/mpeg': 'mp3',
3568 'audio/webm': 'webm',
3569 'audio/x-matroska': 'mka',
3570 'audio/x-mpegurl': 'm3u',
3578 'x-realaudio': 'ra',
3589 'vnd.wap.wbmp': 'wbmp',
3596 'filmstrip+json': 'fs',
3597 'smptett+xml': 'tt',
3600 'x-ms-sami': 'sami',
3609 mimetype = mt.partition(';')[0].strip().lower()
3610 _, _, subtype = mimetype.rpartition('/')
3612 ext = traverse_obj(MAP, mimetype, subtype, subtype.rsplit('+')[-1])
3615 elif default is not NO_DEFAULT:
3617 return subtype.replace('+', '.')
3620 def ext2mimetype(ext_or_url):
3623 if '.' not in ext_or_url:
3624 ext_or_url = f'file.{ext_or_url}'
3625 return mimetypes.guess_type(ext_or_url)[0]
3628 def parse_codecs(codecs_str):
3629 # http://tools.ietf.org/html/rfc6381
3632 split_codecs = list(filter(None, map(
3633 str.strip, codecs_str.strip().strip(',').split(','))))
3634 vcodec, acodec, scodec, hdr = None, None, None, None
3635 for full_codec in split_codecs:
3636 parts = re.sub(r'0+(?=\d)', '', full_codec).split('.')
3637 if parts[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3638 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3642 if parts[0] in ('dvh1', 'dvhe'):
3644 elif parts[0] == 'av1' and traverse_obj(parts, 3) == '10':
3646 elif parts[:2] == ['vp9', '2']:
3648 elif parts[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-4',
3649 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3650 acodec = acodec or full_codec
3651 elif parts[0] in ('stpp', 'wvtt'):
3652 scodec = scodec or full_codec
3654 write_string(f'WARNING: Unknown codec {full_codec}\n')
3655 if vcodec or acodec or scodec:
3657 'vcodec': vcodec or 'none',
3658 'acodec': acodec or 'none',
3659 'dynamic_range': hdr,
3660 **({'scodec': scodec} if scodec is not None else {}),
3662 elif len(split_codecs) == 2:
3664 'vcodec': split_codecs[0],
3665 'acodec': split_codecs[1],
3670 def get_compatible_ext(*, vcodecs, acodecs, vexts, aexts, preferences=None):
3671 assert len(vcodecs) == len(vexts) and len(acodecs) == len(aexts)
3673 allow_mkv = not preferences or 'mkv' in preferences
3675 if allow_mkv and max(len(acodecs), len(vcodecs)) > 1:
3676 return 'mkv' # TODO: any other format allows this?
3678 # TODO: All codecs supported by parse_codecs isn't handled here
3679 COMPATIBLE_CODECS = {
3681 'av1', 'hevc', 'avc1', 'mp4a', 'ac-4', # fourcc (m3u8, mpd)
3682 'h264', 'aacl', 'ec-3', # Set in ISM
3685 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3686 'vp9x', 'vp8x', # in the webm spec
3690 sanitize_codec = functools.partial(
3691 try_get, getter=lambda x: x[0].split('.')[0].replace('0', '').lower())
3692 vcodec, acodec = sanitize_codec(vcodecs), sanitize_codec(acodecs)
3694 for ext in preferences or COMPATIBLE_CODECS.keys():
3695 codec_set = COMPATIBLE_CODECS.get(ext, set())
3696 if ext == 'mkv' or codec_set.issuperset((vcodec, acodec)):
3700 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'},
3703 for ext in preferences or vexts:
3704 current_exts = {ext, *vexts, *aexts}
3705 if ext == 'mkv' or current_exts == {ext} or any(
3706 ext_sets.issuperset(current_exts) for ext_sets in COMPATIBLE_EXTS):
3708 return 'mkv' if allow_mkv else preferences[-1]
3711 def urlhandle_detect_ext(url_handle, default=NO_DEFAULT):
3712 getheader = url_handle.headers.get
3714 cd = getheader('Content-Disposition')
3716 m = re.match(r'attachment;\s*filename="(?P
<filename
>[^
"]+)"', cd)
3718 e = determine_ext(m.group('filename
'), default_ext=None)
3722 meta_ext = getheader('x
-amz
-meta
-name
')
3724 e = meta_ext.rpartition('.')[2]
3728 return mimetype2ext(getheader('Content
-Type
'), default=default)
3731 def encode_data_uri(data, mime_type):
3732 return 'data
:%s;base64
,%s' % (mime_type, base64.b64encode(data).decode('ascii
'))
3735 def age_restricted(content_limit, age_limit):
3736 """ Returns True iff the content should be blocked """
3738 if age_limit is None: # No limit set
3740 if content_limit is None:
3741 return False # Content available for everyone
3742 return age_limit < content_limit
3745 # List of known byte-order-marks (BOM)
3747 (b'\xef\xbb\xbf', 'utf
-8'),
3748 (b'\x00\x00\xfe\xff', 'utf
-32-be
'),
3749 (b'\xff\xfe\x00\x00', 'utf
-32-le
'),
3750 (b'\xff\xfe', 'utf
-16-le
'),
3751 (b'\xfe\xff', 'utf
-16-be
'),
3755 def is_html(first_bytes):
3756 """ Detect whether a file contains HTML by examining its first bytes. """
3759 for bom, enc in BOMS:
3760 while first_bytes.startswith(bom):
3761 encoding, first_bytes = enc, first_bytes[len(bom):]
3763 return re.match(r'^\s
*<', first_bytes.decode(encoding, 'replace
'))
3766 def determine_protocol(info_dict):
3767 protocol = info_dict.get('protocol
')
3768 if protocol is not None:
3771 url = sanitize_url(info_dict['url
'])
3772 if url.startswith('rtmp
'):
3774 elif url.startswith('mms
'):
3776 elif url.startswith('rtsp
'):
3779 ext = determine_ext(url)
3781 return 'm3u8
' if info_dict.get('is_live
') else 'm3u8_native
'
3785 return urllib.parse.urlparse(url).scheme
3788 def render_table(header_row, data, delim=False, extra_gap=0, hide_empty=False):
3789 """ Render a list of rows, each as a list of values.
3790 Text after a \t will be right aligned """
3792 return len(remove_terminal_sequences(string).replace('\t', ''))
3794 def get_max_lens(table):
3795 return [max(width(str(v)) for v in col) for col in zip(*table)]
3797 def filter_using_list(row, filterArray):
3798 return [col for take, col in itertools.zip_longest(filterArray, row, fillvalue=True) if take]
3800 max_lens = get_max_lens(data) if hide_empty else []
3801 header_row = filter_using_list(header_row, max_lens)
3802 data = [filter_using_list(row, max_lens) for row in data]
3804 table = [header_row] + data
3805 max_lens = get_max_lens(table)
3808 table = [header_row, [delim * (ml + extra_gap) for ml in max_lens]] + data
3809 table[1][-1] = table[1][-1][:-extra_gap * len(delim)] # Remove extra_gap from end of delimiter
3811 for pos, text in enumerate(map(str, row)):
3813 row[pos] = text.replace('\t', ' ' * (max_lens[pos] - width(text))) + ' ' * extra_gap
3815 row[pos] = text + ' ' * (max_lens[pos] - width(text) + extra_gap)
3816 ret = '\n'.join(''.join(row).rstrip() for row in table)
3820 def _match_one(filter_part, dct, incomplete):
3821 # TODO: Generalize code with YoutubeDL._build_format_filter
3822 STRING_OPERATORS = {
3823 '*=': operator.contains,
3824 '^
=': lambda attr, value: attr.startswith(value),
3825 '$
=': lambda attr, value: attr.endswith(value),
3826 '~
=': lambda attr, value: re.search(value, attr),
3828 COMPARISON_OPERATORS = {
3830 '<=': operator.le, # "<=" must be defined above "<"
3837 if isinstance(incomplete, bool):
3838 is_incomplete = lambda _: incomplete
3840 is_incomplete = lambda k: k in incomplete
3842 operator_rex = re.compile(r'''(?x)
3844 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3846 (?P<quote>["\'])(?P<quotedstrval>.+?)(?P=quote)|
3849 ''' % '|
'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3850 m = operator_rex.fullmatch(filter_part.strip())
3853 unnegated_op = COMPARISON_OPERATORS[m['op
']]
3855 op = lambda attr, value: not unnegated_op(attr, value)
3858 comparison_value = m['quotedstrval
'] or m['strval
'] or m['intval
']
3860 comparison_value = comparison_value.replace(r'\
%s' % m['quote
'], m['quote
'])
3861 actual_value = dct.get(m['key
'])
3862 numeric_comparison = None
3863 if isinstance(actual_value, (int, float)):
3864 # If the original field is a string and matching comparisonvalue is
3865 # a number we should respect the origin of the original field
3866 # and process comparison value as a string (see
3867 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3869 numeric_comparison = int(comparison_value)
3871 numeric_comparison = parse_filesize(comparison_value)
3872 if numeric_comparison is None:
3873 numeric_comparison = parse_filesize(f'{comparison_value}B
')
3874 if numeric_comparison is None:
3875 numeric_comparison = parse_duration(comparison_value)
3876 if numeric_comparison is not None and m['op
'] in STRING_OPERATORS:
3877 raise ValueError('Operator
%s only supports string values
!' % m['op
'])
3878 if actual_value is None:
3879 return is_incomplete(m['key
']) or m['none_inclusive
']
3880 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3883 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3884 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3886 operator_rex = re.compile(r'''(?x)
3887 (?P<op>%s)\s*(?P<key>[a-z_]+)
3888 ''' % '|
'.join(map(re.escape, UNARY_OPERATORS.keys())))
3889 m = operator_rex.fullmatch(filter_part.strip())
3891 op = UNARY_OPERATORS[m.group('op
')]
3892 actual_value = dct.get(m.group('key
'))
3893 if is_incomplete(m.group('key
')) and actual_value is None:
3895 return op(actual_value)
3897 raise ValueError('Invalid
filter part
%r' % filter_part)
3900 def match_str(filter_str, dct, incomplete=False):
3901 """ Filter a dictionary with a simple string syntax.
3902 @returns Whether the filter passes
3903 @param incomplete Set of keys that is expected to be missing from dct.
3904 Can be True/False to indicate all/none of the keys may be missing.
3905 All conditions on incomplete keys pass if the key is missing
3908 _match_one(filter_part.replace(r'\
&', '&'), dct, incomplete)
3909 for filter_part in re.split(r'(?
<!\\)&', filter_str))
3912 def match_filter_func(filters):
3915 filters = set(variadic(filters))
3917 interactive = '-' in filters
3921 def _match_func(info_dict, incomplete=False):
3922 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3923 return NO_DEFAULT if interactive and not incomplete else None
3925 video_title = info_dict.get('title
') or info_dict.get('id') or 'entry
'
3926 filter_str = ') |
('.join(map(str.strip, filters))
3927 return f'{video_title} does
not pass filter ({filter_str}
), skipping
..'
3931 class download_range_func:
3932 def __init__(self, chapters, ranges):
3933 self.chapters, self.ranges = chapters, ranges
3935 def __call__(self, info_dict, ydl):
3936 if not self.ranges and not self.chapters:
3939 warning = ('There are no chapters matching the regex
' if info_dict.get('chapters
')
3940 else 'Cannot match chapters since chapter information
is unavailable
')
3941 for regex in self.chapters or []:
3942 for i, chapter in enumerate(info_dict.get('chapters
') or []):
3943 if re.search(regex, chapter['title
']):
3945 yield {**chapter, 'index': i}
3946 if self.chapters and warning:
3947 ydl.to_screen(f'[info
] {info_dict["id"]}
: {warning}
')
3949 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3951 def __eq__(self, other):
3952 return (isinstance(other, download_range_func)
3953 and self.chapters == other.chapters and self.ranges == other.ranges)
3956 return f'{__name__}
.{type(self).__name__}
({self.chapters}
, {self.ranges}
)'
3959 def parse_dfxp_time_expr(time_expr):
3963 mobj = re.match(rf'^
(?P
<time_offset
>{NUMBER_RE}
)s?$
', time_expr)
3965 return float(mobj.group('time_offset
'))
3967 mobj = re.match(r'^
(\d
+):(\d\d
):(\d\
d(?
:(?
:\
.|
:)\d
+)?
)$
', time_expr)
3969 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3972 def srt_subtitles_timecode(seconds):
3973 return '%02d
:%02d
:%02d
,%03d
' % timetuple_from_msec(seconds * 1000)
3976 def ass_subtitles_timecode(seconds):
3977 time = timetuple_from_msec(seconds * 1000)
3978 return '%01d
:%02d
:%02d
.%02d
' % (*time[:-1], time.milliseconds / 10)
3981 def dfxp2srt(dfxp_data):
3983 @param dfxp_data A bytes-like object containing DFXP data
3984 @returns A unicode object containing converted SRT data
3986 LEGACY_NAMESPACES = (
3987 (b'http
://www
.w3
.org
/ns
/ttml
', [
3988 b'http
://www
.w3
.org
/2004/11/ttaf1
',
3989 b'http
://www
.w3
.org
/2006/04/ttaf1
',
3990 b'http
://www
.w3
.org
/2006/10/ttaf1
',
3992 (b'http
://www
.w3
.org
/ns
/ttml
#styling', [
3993 b
'http://www.w3.org/ns/ttml#style',
3997 SUPPORTED_STYLING
= [
4006 _x
= functools
.partial(xpath_with_ns
, ns_map
={
4007 'xml': 'http://www.w3.org/XML/1998/namespace',
4008 'ttml': 'http://www.w3.org/ns/ttml',
4009 'tts': 'http://www.w3.org/ns/ttml#styling',
4015 class TTMLPElementParser
:
4017 _unclosed_elements
= []
4018 _applied_styles
= []
4020 def start(self
, tag
, attrib
):
4021 if tag
in (_x('ttml:br'), 'br'):
4024 unclosed_elements
= []
4026 element_style_id
= attrib
.get('style')
4028 style
.update(default_style
)
4029 if element_style_id
:
4030 style
.update(styles
.get(element_style_id
, {}))
4031 for prop
in SUPPORTED_STYLING
:
4032 prop_val
= attrib
.get(_x('tts:' + prop
))
4034 style
[prop
] = prop_val
4037 for k
, v
in sorted(style
.items()):
4038 if self
._applied
_styles
and self
._applied
_styles
[-1].get(k
) == v
:
4041 font
+= ' color="%s"' % v
4042 elif k
== 'fontSize':
4043 font
+= ' size="%s"' % v
4044 elif k
== 'fontFamily':
4045 font
+= ' face="%s"' % v
4046 elif k
== 'fontWeight' and v
== 'bold':
4048 unclosed_elements
.append('b')
4049 elif k
== 'fontStyle' and v
== 'italic':
4051 unclosed_elements
.append('i')
4052 elif k
== 'textDecoration' and v
== 'underline':
4054 unclosed_elements
.append('u')
4056 self
._out
+= '<font' + font
+ '>'
4057 unclosed_elements
.append('font')
4059 if self
._applied
_styles
:
4060 applied_style
.update(self
._applied
_styles
[-1])
4061 applied_style
.update(style
)
4062 self
._applied
_styles
.append(applied_style
)
4063 self
._unclosed
_elements
.append(unclosed_elements
)
4066 if tag
not in (_x('ttml:br'), 'br'):
4067 unclosed_elements
= self
._unclosed
_elements
.pop()
4068 for element
in reversed(unclosed_elements
):
4069 self
._out
+= '</%s>' % element
4070 if unclosed_elements
and self
._applied
_styles
:
4071 self
._applied
_styles
.pop()
4073 def data(self
, data
):
4077 return self
._out
.strip()
4079 def parse_node(node
):
4080 target
= TTMLPElementParser()
4081 parser
= xml
.etree
.ElementTree
.XMLParser(target
=target
)
4082 parser
.feed(xml
.etree
.ElementTree
.tostring(node
))
4083 return parser
.close()
4085 for k
, v
in LEGACY_NAMESPACES
:
4087 dfxp_data
= dfxp_data
.replace(ns
, k
)
4089 dfxp
= compat_etree_fromstring(dfxp_data
)
4091 paras
= dfxp
.findall(_x('.//ttml:p')) or dfxp
.findall('.//p')
4094 raise ValueError('Invalid dfxp/TTML subtitle')
4098 for style
in dfxp
.findall(_x('.//ttml:style')):
4099 style_id
= style
.get('id') or style
.get(_x('xml:id'))
4102 parent_style_id
= style
.get('style')
4104 if parent_style_id
not in styles
:
4107 styles
[style_id
] = styles
[parent_style_id
].copy()
4108 for prop
in SUPPORTED_STYLING
:
4109 prop_val
= style
.get(_x('tts:' + prop
))
4111 styles
.setdefault(style_id
, {})[prop
] = prop_val
4117 for p
in ('body', 'div'):
4118 ele
= xpath_element(dfxp
, [_x('.//ttml:' + p
), './/' + p
])
4121 style
= styles
.get(ele
.get('style'))
4124 default_style
.update(style
)
4126 for para
, index
in zip(paras
, itertools
.count(1)):
4127 begin_time
= parse_dfxp_time_expr(para
.attrib
.get('begin'))
4128 end_time
= parse_dfxp_time_expr(para
.attrib
.get('end'))
4129 dur
= parse_dfxp_time_expr(para
.attrib
.get('dur'))
4130 if begin_time
is None:
4135 end_time
= begin_time
+ dur
4136 out
.append('%d\n%s --> %s\n%s\n\n' % (
4138 srt_subtitles_timecode(begin_time
),
4139 srt_subtitles_timecode(end_time
),
4145 def cli_option(params
, command_option
, param
, separator
=None):
4146 param
= params
.get(param
)
4147 return ([] if param
is None
4148 else [command_option
, str(param
)] if separator
is None
4149 else [f
'{command_option}{separator}{param}'])
4152 def cli_bool_option(params
, command_option
, param
, true_value
='true', false_value
='false', separator
=None):
4153 param
= params
.get(param
)
4154 assert param
in (True, False, None)
4155 return cli_option({True: true_value, False: false_value}
, command_option
, param
, separator
)
4158 def cli_valueless_option(params
, command_option
, param
, expected_value
=True):
4159 return [command_option
] if params
.get(param
) == expected_value
else []
4162 def cli_configuration_args(argdict
, keys
, default
=[], use_compat
=True):
4163 if isinstance(argdict
, (list, tuple)): # for backward compatibility
4170 assert isinstance(argdict
, dict)
4172 assert isinstance(keys
, (list, tuple))
4173 for key_list
in keys
:
4174 arg_list
= list(filter(
4175 lambda x
: x
is not None,
4176 [argdict
.get(key
.lower()) for key
in variadic(key_list
)]))
4178 return [arg
for args
in arg_list
for arg
in args
]
4182 def _configuration_args(main_key
, argdict
, exe
, keys
=None, default
=[], use_compat
=True):
4183 main_key
, exe
= main_key
.lower(), exe
.lower()
4184 root_key
= exe
if main_key
== exe
else f
'{main_key}+{exe}'
4185 keys
= [f
'{root_key}{k}' for k
in (keys
or [''])]
4186 if root_key
in keys
:
4188 keys
.append((main_key
, exe
))
4189 keys
.append('default')
4192 return cli_configuration_args(argdict
, keys
, default
, use_compat
)
4196 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4255 'iw': 'heb', # Replaced by he in 1989 revision
4265 'in': 'ind', # Replaced by id in 1989 revision
4380 'ji': 'yid', # Replaced by yi in 1989 revision
4388 def short2long(cls
, code
):
4389 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4390 return cls
._lang
_map
.get(code
[:2])
4393 def long2short(cls
, code
):
4394 """Convert language code from ISO 639-2/T to ISO 639-1"""
4395 for short_name
, long_name
in cls
._lang
_map
.items():
4396 if long_name
== code
:
4401 # From http://data.okfn.org/data/core/country-list
4403 'AF': 'Afghanistan',
4404 'AX': 'Åland Islands',
4407 'AS': 'American Samoa',
4412 'AG': 'Antigua and Barbuda',
4429 'BO': 'Bolivia, Plurinational State of',
4430 'BQ': 'Bonaire, Sint Eustatius and Saba',
4431 'BA': 'Bosnia and Herzegovina',
4433 'BV': 'Bouvet Island',
4435 'IO': 'British Indian Ocean Territory',
4436 'BN': 'Brunei Darussalam',
4438 'BF': 'Burkina Faso',
4444 'KY': 'Cayman Islands',
4445 'CF': 'Central African Republic',
4449 'CX': 'Christmas Island',
4450 'CC': 'Cocos (Keeling) Islands',
4454 'CD': 'Congo, the Democratic Republic of the',
4455 'CK': 'Cook Islands',
4457 'CI': 'Côte d\'Ivoire',
4462 'CZ': 'Czech Republic',
4466 'DO': 'Dominican Republic',
4469 'SV': 'El Salvador',
4470 'GQ': 'Equatorial Guinea',
4474 'FK': 'Falkland Islands (Malvinas)',
4475 'FO': 'Faroe Islands',
4479 'GF': 'French Guiana',
4480 'PF': 'French Polynesia',
4481 'TF': 'French Southern Territories',
4496 'GW': 'Guinea-Bissau',
4499 'HM': 'Heard Island and McDonald Islands',
4500 'VA': 'Holy See (Vatican City State)',
4507 'IR': 'Iran, Islamic Republic of',
4510 'IM': 'Isle of Man',
4520 'KP': 'Korea, Democratic People\'s Republic of',
4521 'KR': 'Korea, Republic of',
4524 'LA': 'Lao People\'s Democratic Republic',
4530 'LI': 'Liechtenstein',
4534 'MK': 'Macedonia, the Former Yugoslav Republic of',
4541 'MH': 'Marshall Islands',
4547 'FM': 'Micronesia, Federated States of',
4548 'MD': 'Moldova, Republic of',
4559 'NL': 'Netherlands',
4560 'NC': 'New Caledonia',
4561 'NZ': 'New Zealand',
4566 'NF': 'Norfolk Island',
4567 'MP': 'Northern Mariana Islands',
4572 'PS': 'Palestine, State of',
4574 'PG': 'Papua New Guinea',
4577 'PH': 'Philippines',
4581 'PR': 'Puerto Rico',
4585 'RU': 'Russian Federation',
4587 'BL': 'Saint Barthélemy',
4588 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4589 'KN': 'Saint Kitts and Nevis',
4590 'LC': 'Saint Lucia',
4591 'MF': 'Saint Martin (French part)',
4592 'PM': 'Saint Pierre and Miquelon',
4593 'VC': 'Saint Vincent and the Grenadines',
4596 'ST': 'Sao Tome and Principe',
4597 'SA': 'Saudi Arabia',
4601 'SL': 'Sierra Leone',
4603 'SX': 'Sint Maarten (Dutch part)',
4606 'SB': 'Solomon Islands',
4608 'ZA': 'South Africa',
4609 'GS': 'South Georgia and the South Sandwich Islands',
4610 'SS': 'South Sudan',
4615 'SJ': 'Svalbard and Jan Mayen',
4618 'CH': 'Switzerland',
4619 'SY': 'Syrian Arab Republic',
4620 'TW': 'Taiwan, Province of China',
4622 'TZ': 'Tanzania, United Republic of',
4624 'TL': 'Timor-Leste',
4628 'TT': 'Trinidad and Tobago',
4631 'TM': 'Turkmenistan',
4632 'TC': 'Turks and Caicos Islands',
4636 'AE': 'United Arab Emirates',
4637 'GB': 'United Kingdom',
4638 'US': 'United States',
4639 'UM': 'United States Minor Outlying Islands',
4643 'VE': 'Venezuela, Bolivarian Republic of',
4645 'VG': 'Virgin Islands, British',
4646 'VI': 'Virgin Islands, U.S.',
4647 'WF': 'Wallis and Futuna',
4648 'EH': 'Western Sahara',
4652 # Not ISO 3166 codes, but used for IP blocks
4653 'AP': 'Asia/Pacific Region',
4658 def short2full(cls
, code
):
4659 """Convert an ISO 3166-2 country code to the corresponding full name"""
4660 return cls
._country
_map
.get(code
.upper())
4664 # Major IPv4 address blocks per country
4666 'AD': '46.172.224.0/19',
4667 'AE': '94.200.0.0/13',
4668 'AF': '149.54.0.0/17',
4669 'AG': '209.59.64.0/18',
4670 'AI': '204.14.248.0/21',
4671 'AL': '46.99.0.0/16',
4672 'AM': '46.70.0.0/15',
4673 'AO': '105.168.0.0/13',
4674 'AP': '182.50.184.0/21',
4675 'AQ': '23.154.160.0/24',
4676 'AR': '181.0.0.0/12',
4677 'AS': '202.70.112.0/20',
4678 'AT': '77.116.0.0/14',
4679 'AU': '1.128.0.0/11',
4680 'AW': '181.41.0.0/18',
4681 'AX': '185.217.4.0/22',
4682 'AZ': '5.197.0.0/16',
4683 'BA': '31.176.128.0/17',
4684 'BB': '65.48.128.0/17',
4685 'BD': '114.130.0.0/16',
4687 'BF': '102.178.0.0/15',
4688 'BG': '95.42.0.0/15',
4689 'BH': '37.131.0.0/17',
4690 'BI': '154.117.192.0/18',
4691 'BJ': '137.255.0.0/16',
4692 'BL': '185.212.72.0/23',
4693 'BM': '196.12.64.0/18',
4694 'BN': '156.31.0.0/16',
4695 'BO': '161.56.0.0/16',
4696 'BQ': '161.0.80.0/20',
4697 'BR': '191.128.0.0/12',
4698 'BS': '24.51.64.0/18',
4699 'BT': '119.2.96.0/19',
4700 'BW': '168.167.0.0/16',
4701 'BY': '178.120.0.0/13',
4702 'BZ': '179.42.192.0/18',
4703 'CA': '99.224.0.0/11',
4704 'CD': '41.243.0.0/16',
4705 'CF': '197.242.176.0/21',
4706 'CG': '160.113.0.0/16',
4707 'CH': '85.0.0.0/13',
4708 'CI': '102.136.0.0/14',
4709 'CK': '202.65.32.0/19',
4710 'CL': '152.172.0.0/14',
4711 'CM': '102.244.0.0/14',
4712 'CN': '36.128.0.0/10',
4713 'CO': '181.240.0.0/12',
4714 'CR': '201.192.0.0/12',
4715 'CU': '152.206.0.0/15',
4716 'CV': '165.90.96.0/19',
4717 'CW': '190.88.128.0/17',
4718 'CY': '31.153.0.0/16',
4719 'CZ': '88.100.0.0/14',
4721 'DJ': '197.241.0.0/17',
4722 'DK': '87.48.0.0/12',
4723 'DM': '192.243.48.0/20',
4724 'DO': '152.166.0.0/15',
4725 'DZ': '41.96.0.0/12',
4726 'EC': '186.68.0.0/15',
4727 'EE': '90.190.0.0/15',
4728 'EG': '156.160.0.0/11',
4729 'ER': '196.200.96.0/20',
4730 'ES': '88.0.0.0/11',
4731 'ET': '196.188.0.0/14',
4732 'EU': '2.16.0.0/13',
4733 'FI': '91.152.0.0/13',
4734 'FJ': '144.120.0.0/16',
4735 'FK': '80.73.208.0/21',
4736 'FM': '119.252.112.0/20',
4737 'FO': '88.85.32.0/19',
4739 'GA': '41.158.0.0/15',
4741 'GD': '74.122.88.0/21',
4742 'GE': '31.146.0.0/16',
4743 'GF': '161.22.64.0/18',
4744 'GG': '62.68.160.0/19',
4745 'GH': '154.160.0.0/12',
4746 'GI': '95.164.0.0/16',
4747 'GL': '88.83.0.0/19',
4748 'GM': '160.182.0.0/15',
4749 'GN': '197.149.192.0/18',
4750 'GP': '104.250.0.0/19',
4751 'GQ': '105.235.224.0/20',
4752 'GR': '94.64.0.0/13',
4753 'GT': '168.234.0.0/16',
4754 'GU': '168.123.0.0/16',
4755 'GW': '197.214.80.0/20',
4756 'GY': '181.41.64.0/18',
4757 'HK': '113.252.0.0/14',
4758 'HN': '181.210.0.0/16',
4759 'HR': '93.136.0.0/13',
4760 'HT': '148.102.128.0/17',
4761 'HU': '84.0.0.0/14',
4762 'ID': '39.192.0.0/10',
4763 'IE': '87.32.0.0/12',
4764 'IL': '79.176.0.0/13',
4765 'IM': '5.62.80.0/20',
4766 'IN': '117.192.0.0/10',
4767 'IO': '203.83.48.0/21',
4768 'IQ': '37.236.0.0/14',
4769 'IR': '2.176.0.0/12',
4770 'IS': '82.221.0.0/16',
4771 'IT': '79.0.0.0/10',
4772 'JE': '87.244.64.0/18',
4773 'JM': '72.27.0.0/17',
4774 'JO': '176.29.0.0/16',
4775 'JP': '133.0.0.0/8',
4776 'KE': '105.48.0.0/12',
4777 'KG': '158.181.128.0/17',
4778 'KH': '36.37.128.0/17',
4779 'KI': '103.25.140.0/22',
4780 'KM': '197.255.224.0/20',
4781 'KN': '198.167.192.0/19',
4782 'KP': '175.45.176.0/22',
4783 'KR': '175.192.0.0/10',
4784 'KW': '37.36.0.0/14',
4785 'KY': '64.96.0.0/15',
4786 'KZ': '2.72.0.0/13',
4787 'LA': '115.84.64.0/18',
4788 'LB': '178.135.0.0/16',
4789 'LC': '24.92.144.0/20',
4790 'LI': '82.117.0.0/19',
4791 'LK': '112.134.0.0/15',
4792 'LR': '102.183.0.0/16',
4793 'LS': '129.232.0.0/17',
4794 'LT': '78.56.0.0/13',
4795 'LU': '188.42.0.0/16',
4796 'LV': '46.109.0.0/16',
4797 'LY': '41.252.0.0/14',
4798 'MA': '105.128.0.0/11',
4799 'MC': '88.209.64.0/18',
4800 'MD': '37.246.0.0/16',
4801 'ME': '178.175.0.0/17',
4802 'MF': '74.112.232.0/21',
4803 'MG': '154.126.0.0/17',
4804 'MH': '117.103.88.0/21',
4805 'MK': '77.28.0.0/15',
4806 'ML': '154.118.128.0/18',
4807 'MM': '37.111.0.0/17',
4808 'MN': '49.0.128.0/17',
4809 'MO': '60.246.0.0/16',
4810 'MP': '202.88.64.0/20',
4811 'MQ': '109.203.224.0/19',
4812 'MR': '41.188.64.0/18',
4813 'MS': '208.90.112.0/22',
4814 'MT': '46.11.0.0/16',
4815 'MU': '105.16.0.0/12',
4816 'MV': '27.114.128.0/18',
4817 'MW': '102.70.0.0/15',
4818 'MX': '187.192.0.0/11',
4819 'MY': '175.136.0.0/13',
4820 'MZ': '197.218.0.0/15',
4821 'NA': '41.182.0.0/16',
4822 'NC': '101.101.0.0/18',
4823 'NE': '197.214.0.0/18',
4824 'NF': '203.17.240.0/22',
4825 'NG': '105.112.0.0/12',
4826 'NI': '186.76.0.0/15',
4827 'NL': '145.96.0.0/11',
4828 'NO': '84.208.0.0/13',
4829 'NP': '36.252.0.0/15',
4830 'NR': '203.98.224.0/19',
4831 'NU': '49.156.48.0/22',
4832 'NZ': '49.224.0.0/14',
4833 'OM': '5.36.0.0/15',
4834 'PA': '186.72.0.0/15',
4835 'PE': '186.160.0.0/14',
4836 'PF': '123.50.64.0/18',
4837 'PG': '124.240.192.0/19',
4838 'PH': '49.144.0.0/13',
4839 'PK': '39.32.0.0/11',
4840 'PL': '83.0.0.0/11',
4841 'PM': '70.36.0.0/20',
4842 'PR': '66.50.0.0/16',
4843 'PS': '188.161.0.0/16',
4844 'PT': '85.240.0.0/13',
4845 'PW': '202.124.224.0/20',
4846 'PY': '181.120.0.0/14',
4847 'QA': '37.210.0.0/15',
4848 'RE': '102.35.0.0/16',
4849 'RO': '79.112.0.0/13',
4850 'RS': '93.86.0.0/15',
4851 'RU': '5.136.0.0/13',
4852 'RW': '41.186.0.0/16',
4853 'SA': '188.48.0.0/13',
4854 'SB': '202.1.160.0/19',
4855 'SC': '154.192.0.0/11',
4856 'SD': '102.120.0.0/13',
4857 'SE': '78.64.0.0/12',
4858 'SG': '8.128.0.0/10',
4859 'SI': '188.196.0.0/14',
4860 'SK': '78.98.0.0/15',
4861 'SL': '102.143.0.0/17',
4862 'SM': '89.186.32.0/19',
4863 'SN': '41.82.0.0/15',
4864 'SO': '154.115.192.0/18',
4865 'SR': '186.179.128.0/17',
4866 'SS': '105.235.208.0/21',
4867 'ST': '197.159.160.0/19',
4868 'SV': '168.243.0.0/16',
4869 'SX': '190.102.0.0/20',
4871 'SZ': '41.84.224.0/19',
4872 'TC': '65.255.48.0/20',
4873 'TD': '154.68.128.0/19',
4874 'TG': '196.168.0.0/14',
4875 'TH': '171.96.0.0/13',
4876 'TJ': '85.9.128.0/18',
4877 'TK': '27.96.24.0/21',
4878 'TL': '180.189.160.0/20',
4879 'TM': '95.85.96.0/19',
4880 'TN': '197.0.0.0/11',
4881 'TO': '175.176.144.0/21',
4882 'TR': '78.160.0.0/11',
4883 'TT': '186.44.0.0/15',
4884 'TV': '202.2.96.0/19',
4885 'TW': '120.96.0.0/11',
4886 'TZ': '156.156.0.0/14',
4887 'UA': '37.52.0.0/14',
4888 'UG': '102.80.0.0/13',
4890 'UY': '167.56.0.0/13',
4891 'UZ': '84.54.64.0/18',
4892 'VA': '212.77.0.0/19',
4893 'VC': '207.191.240.0/21',
4894 'VE': '186.88.0.0/13',
4895 'VG': '66.81.192.0/20',
4896 'VI': '146.226.0.0/16',
4897 'VN': '14.160.0.0/11',
4898 'VU': '202.80.32.0/20',
4899 'WF': '117.20.32.0/21',
4900 'WS': '202.4.32.0/19',
4901 'YE': '134.35.0.0/16',
4902 'YT': '41.242.116.0/22',
4903 'ZA': '41.0.0.0/11',
4904 'ZM': '102.144.0.0/13',
4905 'ZW': '102.177.192.0/18',
4909 def random_ipv4(cls
, code_or_block
):
4910 if len(code_or_block
) == 2:
4911 block
= cls
._country
_ip
_map
.get(code_or_block
.upper())
4915 block
= code_or_block
4916 addr
, preflen
= block
.split('/')
4917 addr_min
= struct
.unpack('!L', socket
.inet_aton(addr
))[0]
4918 addr_max
= addr_min |
(0xffffffff >> int(preflen
))
4919 return str(socket
.inet_ntoa(
4920 struct
.pack('!L', random
.randint(addr_min
, addr_max
))))
4923 class PerRequestProxyHandler(urllib
.request
.ProxyHandler
):
4924 def __init__(self
, proxies
=None):
4925 # Set default handlers
4926 for type in ('http', 'https'):
4927 setattr(self
, '%s_open' % type,
4928 lambda r
, proxy
='__noproxy__', type=type, meth
=self
.proxy_open
:
4929 meth(r
, proxy
, type))
4930 urllib
.request
.ProxyHandler
.__init
__(self
, proxies
)
4932 def proxy_open(self
, req
, proxy
, type):
4933 req_proxy
= req
.headers
.get('Ytdl-request-proxy')
4934 if req_proxy
is not None:
4936 del req
.headers
['Ytdl-request-proxy']
4938 if proxy
== '__noproxy__':
4939 return None # No Proxy
4940 if urllib
.parse
.urlparse(proxy
).scheme
.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4941 req
.add_header('Ytdl-socks-proxy', proxy
)
4942 # yt-dlp's http/https handlers do wrapping the socket with socks
4944 return urllib
.request
.ProxyHandler
.proxy_open(
4945 self
, req
, proxy
, type)
4948 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4949 # released into Public Domain
4950 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4952 def long_to_bytes(n
, blocksize
=0):
4953 """long_to_bytes(n:long, blocksize:int) : string
4954 Convert a long integer to a byte string.
4956 If optional blocksize is given and greater than zero, pad the front of the
4957 byte string with binary zeros so that the length is a multiple of
4960 # after much testing, this algorithm was deemed to be the fastest
4964 s
= struct
.pack('>I', n
& 0xffffffff) + s
4966 # strip off leading zeros
4967 for i
in range(len(s
)):
4968 if s
[i
] != b
'\000'[0]:
4971 # only happens when n == 0
4975 # add back some pad bytes. this could be done more efficiently w.r.t. the
4976 # de-padding being done above, but sigh...
4977 if blocksize
> 0 and len(s
) % blocksize
:
4978 s
= (blocksize
- len(s
) % blocksize
) * b
'\000' + s
4982 def bytes_to_long(s
):
4983 """bytes_to_long(string) : long
4984 Convert a byte string to a long integer.
4986 This is (essentially) the inverse of long_to_bytes().
4991 extra
= (4 - length
% 4)
4992 s
= b
'\000' * extra
+ s
4993 length
= length
+ extra
4994 for i
in range(0, length
, 4):
4995 acc
= (acc
<< 32) + struct
.unpack('>I', s
[i
:i
+ 4])[0]
4999 def ohdave_rsa_encrypt(data
, exponent
, modulus
):
5001 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
5004 data: data to encrypt, bytes-like object
5005 exponent, modulus: parameter e and N of RSA algorithm, both integer
5006 Output: hex string of encrypted data
5008 Limitation: supports one block encryption only
5011 payload
= int(binascii
.hexlify(data
[::-1]), 16)
5012 encrypted
= pow(payload
, exponent
, modulus
)
5013 return '%x' % encrypted
5016 def pkcs1pad(data
, length
):
5018 Padding input data with PKCS#1 scheme
5020 @param {int[]} data input data
5021 @param {int} length target length
5022 @returns {int[]} padded data
5024 if len(data
) > length
- 11:
5025 raise ValueError('Input data too long for PKCS#1 padding')
5027 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
5028 return [0, 2] + pseudo_random
+ [0] + data
5031 def _base_n_table(n
, table
):
5032 if not table
and not n
:
5033 raise ValueError('Either table or n must be specified')
5034 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
5036 if n
and n
!= len(table
):
5037 raise ValueError(f
'base {n} exceeds table length {len(table)}')
5041 def encode_base_n(num
, n
=None, table
=None):
5042 """Convert given int to a base-n string"""
5043 table
= _base_n_table(n
, table
)
5047 result
, base
= '', len(table
)
5049 result
= table
[num
% base
] + result
5054 def decode_base_n(string
, n
=None, table
=None):
5055 """Convert given base-n string to int"""
5056 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
5057 result
, base
= 0, len(table
)
5059 result
= result
* base
+ table
[char
]
5063 def decode_base(value
, digits
):
5064 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
5065 f
'in a future version. Use {__name__}.decode_base_n instead')
5066 return decode_base_n(value
, table
=digits
)
5069 def decode_packed_codes(code
):
5070 mobj
= re
.search(PACKED_CODES_RE
, code
)
5071 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
5074 symbols
= symbols
.split('|')
5079 base_n_count
= encode_base_n(count
, base
)
5080 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
5083 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
5087 def caesar(s
, alphabet
, shift
):
5092 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
5097 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
5100 def parse_m3u8_attributes(attrib
):
5102 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5103 if val
.startswith('"'):
5109 def urshift(val
, n
):
5110 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5113 # Based on png2str() written by @gdkchan and improved by @yokrysty
5114 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5115 def decode_png(png_data
):
5116 # Reference: https://www.w3.org/TR/PNG/
5117 header
= png_data
[8:]
5119 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5120 raise OSError('Not a valid PNG file.')
5122 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5123 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5128 length
= unpack_integer(header
[:4])
5131 chunk_type
= header
[:4]
5134 chunk_data
= header
[:length
]
5135 header
= header
[length
:]
5137 header
= header
[4:] # Skip CRC
5145 ihdr
= chunks
[0]['data']
5147 width
= unpack_integer(ihdr
[:4])
5148 height
= unpack_integer(ihdr
[4:8])
5152 for chunk
in chunks
:
5153 if chunk
['type'] == b
'IDAT':
5154 idat
+= chunk
['data']
5157 raise OSError('Unable to read PNG data.')
5159 decompressed_data
= bytearray(zlib
.decompress(idat
))
5164 def _get_pixel(idx
):
5169 for y
in range(height
):
5170 basePos
= y
* (1 + stride
)
5171 filter_type
= decompressed_data
[basePos
]
5175 pixels
.append(current_row
)
5177 for x
in range(stride
):
5178 color
= decompressed_data
[1 + basePos
+ x
]
5179 basex
= y
* stride
+ x
5184 left
= _get_pixel(basex
- 3)
5186 up
= _get_pixel(basex
- stride
)
5188 if filter_type
== 1: # Sub
5189 color
= (color
+ left
) & 0xff
5190 elif filter_type
== 2: # Up
5191 color
= (color
+ up
) & 0xff
5192 elif filter_type
== 3: # Average
5193 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5194 elif filter_type
== 4: # Paeth
5200 c
= _get_pixel(basex
- stride
- 3)
5208 if pa
<= pb
and pa
<= pc
:
5209 color
= (color
+ a
) & 0xff
5211 color
= (color
+ b
) & 0xff
5213 color
= (color
+ c
) & 0xff
5215 current_row
.append(color
)
5217 return width
, height
, pixels
5220 def write_xattr(path
, key
, value
):
5221 # Windows: Write xattrs to NTFS Alternate Data Streams:
5222 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5223 if compat_os_name
== 'nt':
5224 assert ':' not in key
5225 assert os
.path
.exists(path
)
5228 with open(f
'{path}:{key}', 'wb') as f
:
5230 except OSError as e
:
5231 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5234 # UNIX Method 1. Use xattrs/pyxattrs modules
5237 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5238 # Unicode arguments are not supported in pyxattr until version 0.5.0
5239 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5240 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5241 setxattr
= xattr
.set
5243 setxattr
= xattr
.setxattr
5247 setxattr(path
, key
, value
)
5248 except OSError as e
:
5249 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5252 # UNIX Method 2. Use setfattr/xattr executables
5253 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5254 else 'xattr' if check_executable('xattr', ['-h']) else None)
5256 raise XAttrUnavailableError(
5257 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5258 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5260 value
= value
.decode()
5262 _
, stderr
, returncode
= Popen
.run(
5263 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5264 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5265 except OSError as e
:
5266 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5268 raise XAttrMetadataError(returncode
, stderr
)
5271 def random_birthday(year_field
, month_field
, day_field
):
5272 start_date
= datetime
.date(1950, 1, 1)
5273 end_date
= datetime
.date(1995, 12, 31)
5274 offset
= random
.randint(0, (end_date
- start_date
).days
)
5275 random_date
= start_date
+ datetime
.timedelta(offset
)
5277 year_field
: str(random_date
.year
),
5278 month_field
: str(random_date
.month
),
5279 day_field
: str(random_date
.day
),
5283 def find_available_port(interface
=''):
5285 with socket
.socket() as sock
:
5286 sock
.bind((interface
, 0))
5287 return sock
.getsockname()[1]
5292 # Templates for internet shortcut files, which are plain text files.
5293 DOT_URL_LINK_TEMPLATE
= '''\
5298 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5299 <?xml version="1.0" encoding="UTF-8"?>
5300 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5301 <plist version="1.0">
5304 \t<string>%(url)s</string>
5309 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5319 'url': DOT_URL_LINK_TEMPLATE
,
5320 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5321 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5325 def iri_to_uri(iri
):
5327 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5329 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.
5332 iri_parts
= urllib
.parse
.urlparse(iri
)
5334 if '[' in iri_parts
.netloc
:
5335 raise ValueError('IPv6 URIs are not, yet, supported.')
5336 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5338 # 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.
5341 if iri_parts
.username
:
5342 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5343 if iri_parts
.password
is not None:
5344 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5347 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5348 # The 'idna' encoding produces ASCII text.
5349 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5350 net_location
+= ':' + str(iri_parts
.port
)
5352 return urllib
.parse
.urlunparse(
5356 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5358 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5359 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5361 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5362 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5364 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5366 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5369 def to_high_limit_path(path
):
5370 if sys
.platform
in ['win32', 'cygwin']:
5371 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5372 return '\\\\?\\' + os
.path
.abspath(path
)
5377 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5378 val
= traverse_obj(obj
, *variadic(field
))
5379 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5381 return template
% func(val
)
5384 def clean_podcast_url(url
):
5385 return re
.sub(r
'''(?x)
5389 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5392 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5395 cn\.co| # https://podcorn.com/analytics-prefix/
5396 st\.fm # https://podsights.com/docs/
5401 _HEX_TABLE
= '0123456789abcdef'
5404 def random_uuidv4():
5405 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5408 def make_dir(path
, to_screen
=None):
5410 dn
= os
.path
.dirname(path
)
5412 os
.makedirs(dn
, exist_ok
=True)
5414 except OSError as err
:
5415 if callable(to_screen
) is not None:
5416 to_screen('unable to create directory ' + error_to_compat_str(err
))
5420 def get_executable_path():
5421 from .update
import _get_variant_and_executable_path
5423 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5426 def get_user_config_dirs(package_name
):
5427 # .config (e.g. ~/.config/package_name)
5428 xdg_config_home
= os
.getenv('XDG_CONFIG_HOME') or compat_expanduser('~/.config')
5429 yield os
.path
.join(xdg_config_home
, package_name
)
5431 # appdata (%APPDATA%/package_name)
5432 appdata_dir
= os
.getenv('appdata')
5434 yield os
.path
.join(appdata_dir
, package_name
)
5436 # home (~/.package_name)
5437 yield os
.path
.join(compat_expanduser('~'), f
'.{package_name}')
5440 def get_system_config_dirs(package_name
):
5442 yield os
.path
.join('/etc', package_name
)
5446 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5447 casesense
=True, is_user_input
=False, traverse_string
=False):
5449 Safely traverse nested `dict`s and `Sequence`s
5451 >>> obj = [{}, {"key": "value"}]
5452 >>> traverse_obj(obj, (1, "key"))
5455 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5456 The next path will also be tested if the path branched but no results could be found.
5457 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5458 Unhelpful values (`{}`, `None`) are treated as the absence of a value and discarded.
5460 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5462 The keys in the path can be one of:
5463 - `None`: Return the current object.
5464 - `set`: Requires the only item in the set to be a type or function,
5465 like `{type}`/`{func}`. If a `type`, returns only values
5466 of this type. If a function, returns `func(obj)`.
5467 - `str`/`int`: Return `obj[key]`. For `re.Match`, return `obj.group(key)`.
5468 - `slice`: Branch out and return all values in `obj[key]`.
5469 - `Ellipsis`: Branch out and return a list of all values.
5470 - `tuple`/`list`: Branch out and return a list of all matching values.
5471 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5472 - `function`: Branch out and return values filtered by the function.
5473 Read as: `[value for key, value in obj if function(key, value)]`.
5474 For `Sequence`s, `key` is the index of the value.
5475 For `re.Match`es, `key` is the group number (0 = full match)
5476 as well as additionally any group names, if given.
5477 - `dict` Transform the current object and return a matching dict.
5478 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5480 `tuple`, `list`, and `dict` all support nested paths and branches.
5482 @params paths Paths which to traverse by.
5483 @param default Value to return if the paths do not match.
5484 If the last key in the path is a `dict`, it will apply to each value inside
5485 the dict instead, depth first. Try to avoid if using nested `dict` keys.
5486 @param expected_type If a `type`, only accept final values of this type.
5487 If any other callable, try to call the function on each result.
5488 If the last key in the path is a `dict`, it will apply to each value inside
5489 the dict instead, recursively. This does respect branching paths.
5490 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5491 @param casesense If `False`, consider string dictionary keys as case insensitive.
5493 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5495 @param is_user_input Whether the keys are generated from user input.
5496 If `True` strings get converted to `int`/`slice` if needed.
5497 @param traverse_string Whether to traverse into objects as strings.
5498 If `True`, any non-compatible object will first be
5499 converted into a string and then traversed into.
5500 The return value of that path will be a string instead,
5501 not respecting any further branching.
5504 @returns The result of the object traversal.
5505 If successful, `get_all=True`, and the path branches at least once,
5506 then a list of results is returned instead.
5507 If no `default` is given and the last path branches, a `list` of results
5508 is always returned. If a path ends on a `dict` that result will always be a `dict`.
5510 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5511 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5513 if isinstance(expected_type
, type):
5514 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5516 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5518 def apply_key(key
, obj
, is_last
):
5522 if obj
is None and traverse_string
:
5528 elif isinstance(key
, set):
5529 assert len(key
) == 1, 'Set should only be used to wrap a single item'
5530 item
= next(iter(key
))
5531 if isinstance(item
, type):
5532 if isinstance(obj
, item
):
5535 result
= try_call(item
, args
=(obj
,))
5537 elif isinstance(key
, (list, tuple)):
5539 result
= itertools
.chain
.from_iterable(
5540 apply_path(obj
, branch
, is_last
)[0] for branch
in key
)
5544 if isinstance(obj
, collections
.abc
.Mapping
):
5545 result
= obj
.values()
5546 elif is_sequence(obj
):
5548 elif isinstance(obj
, re
.Match
):
5549 result
= obj
.groups()
5550 elif traverse_string
:
5558 if isinstance(obj
, collections
.abc
.Mapping
):
5559 iter_obj
= obj
.items()
5560 elif is_sequence(obj
):
5561 iter_obj
= enumerate(obj
)
5562 elif isinstance(obj
, re
.Match
):
5563 iter_obj
= itertools
.chain(
5564 enumerate((obj
.group(), *obj
.groups())),
5565 obj
.groupdict().items())
5566 elif traverse_string
:
5568 iter_obj
= enumerate(str(obj
))
5572 result
= (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5573 if not branching
: # string traversal
5574 result
= ''.join(result
)
5576 elif isinstance(key
, dict):
5577 iter_obj
= ((k
, _traverse_obj(obj
, v
, False, is_last
)) for k
, v
in key
.items())
5579 k
: v
if v
is not None else default
for k
, v
in iter_obj
5580 if v
is not None or default
is not NO_DEFAULT
5583 elif isinstance(obj
, collections
.abc
.Mapping
):
5584 result
= (obj
.get(key
) if casesense
or (key
in obj
) else
5585 next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5587 elif isinstance(obj
, re
.Match
):
5588 if isinstance(key
, int) or casesense
:
5589 with contextlib
.suppress(IndexError):
5590 result
= obj
.group(key
)
5592 elif isinstance(key
, str):
5593 result
= next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5595 elif isinstance(key
, (int, slice)):
5596 if is_sequence(obj
):
5597 branching
= isinstance(key
, slice)
5598 with contextlib
.suppress(IndexError):
5600 elif traverse_string
:
5601 with contextlib
.suppress(IndexError):
5602 result
= str(obj
)[key
]
5604 return branching
, result
if branching
else (result
,)
5606 def lazy_last(iterable
):
5607 iterator
= iter(iterable
)
5608 prev
= next(iterator
, NO_DEFAULT
)
5609 if prev
is NO_DEFAULT
:
5612 for item
in iterator
:
5618 def apply_path(start_obj
, path
, test_type
):
5620 has_branched
= False
5623 for last
, key
in lazy_last(variadic(path
, (str, bytes, dict, set))):
5624 if is_user_input
and isinstance(key
, str):
5628 key
= slice(*map(int_or_none
, key
.split(':')))
5629 elif int_or_none(key
) is not None:
5632 if not casesense
and isinstance(key
, str):
5633 key
= key
.casefold()
5635 if __debug__
and callable(key
):
5636 # Verify function signature
5637 inspect
.signature(key
).bind(None, None)
5641 branching
, results
= apply_key(key
, obj
, last
)
5642 has_branched |
= branching
5643 new_objs
.append(results
)
5645 objs
= itertools
.chain
.from_iterable(new_objs
)
5647 if test_type
and not isinstance(key
, (dict, list, tuple)):
5648 objs
= map(type_test
, objs
)
5650 return objs
, has_branched
, isinstance(key
, dict)
5652 def _traverse_obj(obj
, path
, allow_empty
, test_type
):
5653 results
, has_branched
, is_dict
= apply_path(obj
, path
, test_type
)
5654 results
= LazyList(item
for item
in results
if item
not in (None, {}))
5655 if get_all
and has_branched
:
5657 return results
.exhaust()
5659 return [] if default
is NO_DEFAULT
else default
5662 return results
[0] if results
else {} if allow_empty
and is_dict
else None
5664 for index
, path
in enumerate(paths
, 1):
5665 result
= _traverse_obj(obj
, path
, index
== len(paths
), True)
5666 if result
is not None:
5669 return None if default
is NO_DEFAULT
else default
5672 def traverse_dict(dictn
, keys
, casesense
=True):
5673 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5674 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5675 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5678 def get_first(obj
, keys
, **kwargs
):
5679 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5682 def time_seconds(**kwargs
):
5684 Returns TZ-aware time in seconds since the epoch (1970-01-01T00:00:00Z)
5686 return time
.time() + datetime
.timedelta(**kwargs
).total_seconds()
5689 # create a JSON Web Signature (jws) with HS256 algorithm
5690 # the resulting format is in JWS Compact Serialization
5691 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5692 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5693 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5699 header_data
.update(headers
)
5700 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5701 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5702 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5703 signature_b64
= base64
.b64encode(h
.digest())
5704 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5708 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5709 def jwt_decode_hs256(jwt
):
5710 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5711 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5712 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5716 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5720 def supports_terminal_sequences(stream
):
5721 if compat_os_name
== 'nt':
5722 if not WINDOWS_VT_MODE
:
5724 elif not os
.getenv('TERM'):
5727 return stream
.isatty()
5728 except BaseException
:
5732 def windows_enable_vt_mode():
5733 """Ref: https://bugs.python.org/issue30075 """
5734 if get_windows_version() < (10, 0, 10586):
5738 import ctypes
.wintypes
5741 ENABLE_VIRTUAL_TERMINAL_PROCESSING
= 0x0004
5743 dll
= ctypes
.WinDLL('kernel32', use_last_error
=False)
5744 handle
= os
.open('CONOUT$', os
.O_RDWR
)
5746 h_out
= ctypes
.wintypes
.HANDLE(msvcrt
.get_osfhandle(handle
))
5747 dw_original_mode
= ctypes
.wintypes
.DWORD()
5748 success
= dll
.GetConsoleMode(h_out
, ctypes
.byref(dw_original_mode
))
5750 raise Exception('GetConsoleMode failed')
5752 success
= dll
.SetConsoleMode(h_out
, ctypes
.wintypes
.DWORD(
5753 dw_original_mode
.value | ENABLE_VIRTUAL_TERMINAL_PROCESSING
))
5755 raise Exception('SetConsoleMode failed')
5759 global WINDOWS_VT_MODE
5760 WINDOWS_VT_MODE
= True
5761 supports_terminal_sequences
.cache_clear()
5764 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5767 def remove_terminal_sequences(string
):
5768 return _terminal_sequences_re
.sub('', string
)
5771 def number_of_digits(number
):
5772 return len('%d' % number
)
5775 def join_nonempty(*values
, delim
='-', from_dict
=None):
5776 if from_dict
is not None:
5777 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5778 return delim
.join(map(str, filter(None, values
)))
5781 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5783 Find the largest format dimensions in terms of video width and, for each thumbnail:
5784 * Modify the URL: Match the width with the provided regex and replace with the former width
5787 This function is useful with video services that scale the provided thumbnails on demand
5789 _keys
= ('width', 'height')
5790 max_dimensions
= max(
5791 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5793 if not max_dimensions
[0]:
5797 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5798 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5799 for thumbnail
in thumbnails
5803 def parse_http_range(range):
5804 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5806 return None, None, None
5807 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5809 return None, None, None
5810 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5813 def read_stdin(what
):
5814 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5815 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5819 def determine_file_encoding(data
):
5821 Detect the text encoding used
5822 @returns (encoding, bytes to skip)
5825 # BOM marks are given priority over declarations
5826 for bom
, enc
in BOMS
:
5827 if data
.startswith(bom
):
5828 return enc
, len(bom
)
5830 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5831 # We ignore the endianness to get a good enough match
5832 data
= data
.replace(b
'\0', b
'')
5833 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5834 return mobj
.group(1).decode() if mobj
else None, 0
5841 __initialized
= False
5843 def __init__(self
, parser
, label
=None):
5844 self
.parser
, self
.label
= parser
, label
5845 self
._loaded
_paths
, self
.configs
= set(), []
5847 def init(self
, args
=None, filename
=None):
5848 assert not self
.__initialized
5849 self
.own_args
, self
.filename
= args
, filename
5850 return self
.load_configs()
5852 def load_configs(self
):
5855 location
= os
.path
.realpath(self
.filename
)
5856 directory
= os
.path
.dirname(location
)
5857 if location
in self
._loaded
_paths
:
5859 self
._loaded
_paths
.add(location
)
5861 self
.__initialized
= True
5862 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5863 self
.parsed_args
= self
.own_args
5864 for location
in opts
.config_locations
or []:
5866 if location
in self
._loaded
_paths
:
5868 self
._loaded
_paths
.add(location
)
5869 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5871 location
= os
.path
.join(directory
, expand_path(location
))
5872 if os
.path
.isdir(location
):
5873 location
= os
.path
.join(location
, 'yt-dlp.conf')
5874 if not os
.path
.exists(location
):
5875 self
.parser
.error(f
'config location {location} does not exist')
5876 self
.append_config(self
.read_file(location
), location
)
5880 label
= join_nonempty(
5881 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5883 return join_nonempty(
5884 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5885 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5889 def read_file(filename
, default
=[]):
5891 optionf
= open(filename
, 'rb')
5893 return default
# silently skip if file is not present
5895 enc
, skip
= determine_file_encoding(optionf
.read(512))
5896 optionf
.seek(skip
, io
.SEEK_SET
)
5898 enc
= None # silently skip read errors
5900 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5901 contents
= optionf
.read().decode(enc
or preferredencoding())
5902 res
= shlex
.split(contents
, comments
=True)
5903 except Exception as err
:
5904 raise ValueError(f
'Unable to parse "{filename}": {err}')
5910 def hide_login_info(opts
):
5911 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5912 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5917 return m
.group('key') + '=PRIVATE'
5921 opts
= list(map(_scrub_eq
, opts
))
5922 for idx
, opt
in enumerate(opts
):
5923 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5924 opts
[idx
+ 1] = 'PRIVATE'
5927 def append_config(self
, *args
, label
=None):
5928 config
= type(self
)(self
.parser
, label
)
5929 config
._loaded
_paths
= self
._loaded
_paths
5930 if config
.init(*args
):
5931 self
.configs
.append(config
)
5935 for config
in reversed(self
.configs
):
5936 yield from config
.all_args
5937 yield from self
.parsed_args
or []
5939 def parse_known_args(self
, **kwargs
):
5940 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5942 def parse_args(self
):
5943 return self
.parser
.parse_args(self
.all_args
)
5946 class WebSocketsWrapper
:
5947 """Wraps websockets module to use in non-async scopes"""
5950 def __init__(self
, url
, headers
=None, connect
=True):
5951 self
.loop
= asyncio
.new_event_loop()
5952 # XXX: "loop" is deprecated
5953 self
.conn
= websockets
.connect(
5954 url
, extra_headers
=headers
, ping_interval
=None,
5955 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5958 atexit
.register(self
.__exit
__, None, None, None)
5960 def __enter__(self
):
5962 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5965 def send(self
, *args
):
5966 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5968 def recv(self
, *args
):
5969 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5971 def __exit__(self
, type, value
, traceback
):
5973 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5976 self
._cancel
_all
_tasks
(self
.loop
)
5978 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5979 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5981 def run_with_loop(main
, loop
):
5982 if not asyncio
.iscoroutine(main
):
5983 raise ValueError(f
'a coroutine was expected, got {main!r}')
5986 return loop
.run_until_complete(main
)
5988 loop
.run_until_complete(loop
.shutdown_asyncgens())
5989 if hasattr(loop
, 'shutdown_default_executor'):
5990 loop
.run_until_complete(loop
.shutdown_default_executor())
5993 def _cancel_all_tasks(loop
):
5994 to_cancel
= asyncio
.all_tasks(loop
)
5999 for task
in to_cancel
:
6002 # XXX: "loop" is removed in python 3.10+
6003 loop
.run_until_complete(
6004 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
6006 for task
in to_cancel
:
6007 if task
.cancelled():
6009 if task
.exception() is not None:
6010 loop
.call_exception_handler({
6011 'message': 'unhandled exception during asyncio.run() shutdown',
6012 'exception': task
.exception(),
6017 def merge_headers(*dicts
):
6018 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
6019 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
6022 def cached_method(f
):
6023 """Cache a method"""
6024 signature
= inspect
.signature(f
)
6027 def wrapper(self
, *args
, **kwargs
):
6028 bound_args
= signature
.bind(self
, *args
, **kwargs
)
6029 bound_args
.apply_defaults()
6030 key
= tuple(bound_args
.arguments
.values())[1:]
6032 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
6033 if key
not in cache
:
6034 cache
[key
] = f(self
, *args
, **kwargs
)
6039 class classproperty
:
6040 """property access for class methods with optional caching"""
6041 def __new__(cls
, func
=None, *args
, **kwargs
):
6043 return functools
.partial(cls
, *args
, **kwargs
)
6044 return super().__new
__(cls
)
6046 def __init__(self
, func
, *, cache
=False):
6047 functools
.update_wrapper(self
, func
)
6049 self
._cache
= {} if cache
else None
6051 def __get__(self
, _
, cls
):
6052 if self
._cache
is None:
6053 return self
.func(cls
)
6054 elif cls
not in self
._cache
:
6055 self
._cache
[cls
] = self
.func(cls
)
6056 return self
._cache
[cls
]
6059 class function_with_repr
:
6060 def __init__(self
, func
):
6061 functools
.update_wrapper(self
, func
)
6064 def __call__(self
, *args
, **kwargs
):
6065 return self
.func(*args
, **kwargs
)
6068 return f
'{self.func.__module__}.{self.func.__qualname__}'
6071 class Namespace(types
.SimpleNamespace
):
6072 """Immutable namespace"""
6075 return iter(self
.__dict
__.values())
6079 return self
.__dict
__.items()
6082 MEDIA_EXTENSIONS
= Namespace(
6083 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
6084 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
6085 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
6086 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma', 'weba'),
6087 thumbnails
=('jpg', 'png', 'webp'),
6088 storyboards
=('mhtml', ),
6089 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
6090 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
6092 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
6093 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
6095 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
6100 for retry in RetryManager(...):
6103 except SomeException as err:
6107 attempt
, _error
= 0, None
6109 def __init__(self
, _retries
, _error_callback
, **kwargs
):
6110 self
.retries
= _retries
or 0
6111 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
6113 def _should_retry(self
):
6114 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
6118 if self
._error
is NO_DEFAULT
:
6123 def error(self
, value
):
6127 while self
._should
_retry
():
6128 self
.error
= NO_DEFAULT
6132 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
6135 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
6136 """Utility function for reporting retries"""
6139 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
6144 elif isinstance(e
, ExtractorError
):
6145 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
6146 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
6148 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
6150 info(f
'Sleeping {delay:.2f} seconds ...')
6154 def make_archive_id(ie
, video_id
):
6155 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
6156 return f
'{ie_key.lower()} {video_id}'
6159 def truncate_string(s
, left
, right
=0):
6160 assert left
> 3 and right
>= 0
6161 if s
is None or len(s
) <= left
+ right
:
6163 return f
'{s[:left-3]}...{s[-right:] if right else ""}'
6166 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
6167 assert 'all' in alias_dict
, '"all" alias is required'
6168 requested
= list(start
or [])
6170 discard
= val
.startswith('-')
6174 if val
in alias_dict
:
6175 val
= alias_dict
[val
] if not discard
else [
6176 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
6177 # NB: Do not allow regex in aliases for performance
6178 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
6181 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
6182 else [val
] if val
in alias_dict
['all'] else None)
6184 raise ValueError(val
)
6187 for item
in current
:
6188 while item
in requested
:
6189 requested
.remove(item
)
6191 requested
.extend(current
)
6193 return orderedSet(requested
)
6197 regex
= r
' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
6199 default
= ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
6200 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
6201 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
6202 ytdl_default
= ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
6203 'height', 'width', 'proto', 'vext', 'abr', 'aext',
6204 'fps', 'fs_approx', 'source', 'id')
6207 'vcodec': {'type': 'ordered', 'regex': True,
6208 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
6209 'acodec': {'type': 'ordered', 'regex': True,
6210 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'ac-?4', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
6211 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
6212 'order': ['dv', '(hdr)?12', r
'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
6213 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
6214 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
6215 'vext': {'type': 'ordered', 'field': 'video_ext',
6216 'order': ('mp4', 'mov', 'webm', 'flv', '', 'none'),
6217 'order_free': ('webm', 'mp4', 'mov', 'flv', '', 'none')},
6218 'aext': {'type': 'ordered', 'regex': True, 'field': 'audio_ext',
6219 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'web[am]', '', 'none'),
6220 'order_free': ('ogg', 'opus', 'web[am]', 'mp3', 'm4a', 'aac', '', 'none')},
6221 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}
,
6222 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
6223 'field': ('vcodec', 'acodec'),
6224 'function': lambda it
: int(any(v
!= 'none' for v
in it
))},
6225 'ie_pref': {'priority': True, 'type': 'extractor'}
,
6226 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6227 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6228 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}
,
6229 'quality': {'convert': 'float', 'default': -1}
,
6230 'filesize': {'convert': 'bytes'}
,
6231 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}
,
6232 'id': {'convert': 'string', 'field': 'format_id'}
,
6233 'height': {'convert': 'float_none'}
,
6234 'width': {'convert': 'float_none'}
,
6235 'fps': {'convert': 'float_none'}
,
6236 'channels': {'convert': 'float_none', 'field': 'audio_channels'}
,
6237 'tbr': {'convert': 'float_none'}
,
6238 'vbr': {'convert': 'float_none'}
,
6239 'abr': {'convert': 'float_none'}
,
6240 'asr': {'convert': 'float_none'}
,
6241 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}
,
6243 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}
,
6244 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True}
,
6245 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')}
,
6246 'ext': {'type': 'combined', 'field': ('vext', 'aext')}
,
6247 'res': {'type': 'multiple', 'field': ('height', 'width'),
6248 'function': lambda it
: (lambda l
: min(l
) if l
else 0)(tuple(filter(None, it
)))},
6250 # Actual field names
6251 'format_id': {'type': 'alias', 'field': 'id'}
,
6252 'preference': {'type': 'alias', 'field': 'ie_pref'}
,
6253 'language_preference': {'type': 'alias', 'field': 'lang'}
,
6254 'source_preference': {'type': 'alias', 'field': 'source'}
,
6255 'protocol': {'type': 'alias', 'field': 'proto'}
,
6256 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}
,
6257 'audio_channels': {'type': 'alias', 'field': 'channels'}
,
6260 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6261 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6262 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}
,
6263 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}
,
6264 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}
,
6265 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}
,
6266 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}
,
6267 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}
,
6268 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}
,
6269 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}
,
6270 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}
,
6271 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}
,
6272 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}
,
6273 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}
,
6274 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6275 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6276 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6277 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6278 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6279 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6282 def __init__(self
, ydl
, field_preference
):
6285 self
.evaluate_params(self
.ydl
.params
, field_preference
)
6286 if ydl
.params
.get('verbose'):
6287 self
.print_verbose_info(self
.ydl
.write_debug
)
6289 def _get_field_setting(self
, field
, key
):
6290 if field
not in self
.settings
:
6291 if key
in ('forced', 'priority'):
6293 self
.ydl
.deprecated_feature(f
'Using arbitrary fields ({field}) for format sorting is '
6294 'deprecated and may be removed in a future version')
6295 self
.settings
[field
] = {}
6296 propObj
= self
.settings
[field
]
6297 if key
not in propObj
:
6298 type = propObj
.get('type')
6300 default
= 'preference' if type == 'extractor' else (field
,) if type in ('combined', 'multiple') else field
6301 elif key
== 'convert':
6302 default
= 'order' if type == 'ordered' else 'float_string' if field
else 'ignore'
6304 default
= {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}
.get(key
, None)
6305 propObj
[key
] = default
6308 def _resolve_field_value(self
, field
, value
, convertNone
=False):
6313 value
= value
.lower()
6314 conversion
= self
._get
_field
_setting
(field
, 'convert')
6315 if conversion
== 'ignore':
6317 if conversion
== 'string':
6319 elif conversion
== 'float_none':
6320 return float_or_none(value
)
6321 elif conversion
== 'bytes':
6322 return parse_bytes(value
)
6323 elif conversion
== 'order':
6324 order_list
= (self
._use
_free
_order
and self
._get
_field
_setting
(field
, 'order_free')) or self
._get
_field
_setting
(field
, 'order')
6325 use_regex
= self
._get
_field
_setting
(field
, 'regex')
6326 list_length
= len(order_list
)
6327 empty_pos
= order_list
.index('') if '' in order_list
else list_length
+ 1
6328 if use_regex
and value
is not None:
6329 for i
, regex
in enumerate(order_list
):
6330 if regex
and re
.match(regex
, value
):
6331 return list_length
- i
6332 return list_length
- empty_pos
# not in list
6333 else: # not regex or value = None
6334 return list_length
- (order_list
.index(value
) if value
in order_list
else empty_pos
)
6336 if value
.isnumeric():
6339 self
.settings
[field
]['convert'] = 'string'
6342 def evaluate_params(self
, params
, sort_extractor
):
6343 self
._use
_free
_order
= params
.get('prefer_free_formats', False)
6344 self
._sort
_user
= params
.get('format_sort', [])
6345 self
._sort
_extractor
= sort_extractor
6347 def add_item(field
, reverse
, closest
, limit_text
):
6348 field
= field
.lower()
6349 if field
in self
._order
:
6351 self
._order
.append(field
)
6352 limit
= self
._resolve
_field
_value
(field
, limit_text
)
6355 'closest': False if limit
is None else closest
,
6356 'limit_text': limit_text
,
6358 if field
in self
.settings
:
6359 self
.settings
[field
].update(data
)
6361 self
.settings
[field
] = data
6364 tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'forced'))
6365 + (tuple() if params
.get('format_sort_force', False)
6366 else tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'priority')))
6367 + tuple(self
._sort
_user
) + tuple(sort_extractor
) + self
.default
)
6369 for item
in sort_list
:
6370 match
= re
.match(self
.regex
, item
)
6372 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item
)
6373 field
= match
.group('field')
6376 if self
._get
_field
_setting
(field
, 'type') == 'alias':
6377 alias
, field
= field
, self
._get
_field
_setting
(field
, 'field')
6378 if self
._get
_field
_setting
(alias
, 'deprecated'):
6379 self
.ydl
.deprecated_feature(f
'Format sorting alias {alias} is deprecated and may '
6380 f
'be removed in a future version. Please use {field} instead')
6381 reverse
= match
.group('reverse') is not None
6382 closest
= match
.group('separator') == '~'
6383 limit_text
= match
.group('limit')
6385 has_limit
= limit_text
is not None
6386 has_multiple_fields
= self
._get
_field
_setting
(field
, 'type') == 'combined'
6387 has_multiple_limits
= has_limit
and has_multiple_fields
and not self
._get
_field
_setting
(field
, 'same_limit')
6389 fields
= self
._get
_field
_setting
(field
, 'field') if has_multiple_fields
else (field
,)
6390 limits
= limit_text
.split(':') if has_multiple_limits
else (limit_text
,) if has_limit
else tuple()
6391 limit_count
= len(limits
)
6392 for (i
, f
) in enumerate(fields
):
6393 add_item(f
, reverse
, closest
,
6394 limits
[i
] if i
< limit_count
6395 else limits
[0] if has_limit
and not has_multiple_limits
6398 def print_verbose_info(self
, write_debug
):
6400 write_debug('Sort order given by user: %s' % ', '.join(self
._sort
_user
))
6401 if self
._sort
_extractor
:
6402 write_debug('Sort order given by extractor: %s' % ', '.join(self
._sort
_extractor
))
6403 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
6404 '+' if self
._get
_field
_setting
(field
, 'reverse') else '', field
,
6405 '%s%s(%s)' % ('~' if self
._get
_field
_setting
(field
, 'closest') else ':',
6406 self
._get
_field
_setting
(field
, 'limit_text'),
6407 self
._get
_field
_setting
(field
, 'limit'))
6408 if self
._get
_field
_setting
(field
, 'limit_text') is not None else '')
6409 for field
in self
._order
if self
._get
_field
_setting
(field
, 'visible')]))
6411 def _calculate_field_preference_from_value(self
, format
, field
, type, value
):
6412 reverse
= self
._get
_field
_setting
(field
, 'reverse')
6413 closest
= self
._get
_field
_setting
(field
, 'closest')
6414 limit
= self
._get
_field
_setting
(field
, 'limit')
6416 if type == 'extractor':
6417 maximum
= self
._get
_field
_setting
(field
, 'max')
6418 if value
is None or (maximum
is not None and value
>= maximum
):
6420 elif type == 'boolean':
6421 in_list
= self
._get
_field
_setting
(field
, 'in_list')
6422 not_in_list
= self
._get
_field
_setting
(field
, 'not_in_list')
6423 value
= 0 if ((in_list
is None or value
in in_list
) and (not_in_list
is None or value
not in not_in_list
)) else -1
6424 elif type == 'ordered':
6425 value
= self
._resolve
_field
_value
(field
, value
, True)
6427 # try to convert to number
6428 val_num
= float_or_none(value
, default
=self
._get
_field
_setting
(field
, 'default'))
6429 is_num
= self
._get
_field
_setting
(field
, 'convert') != 'string' and val_num
is not None
6433 return ((-10, 0) if value
is None
6434 else (1, value
, 0) if not is_num
# if a field has mixed strings and numbers, strings are sorted higher
6435 else (0, -abs(value
- limit
), value
- limit
if reverse
else limit
- value
) if closest
6436 else (0, value
, 0) if not reverse
and (limit
is None or value
<= limit
)
6437 else (0, -value
, 0) if limit
is None or (reverse
and value
== limit
) or value
> limit
6438 else (-1, value
, 0))
6440 def _calculate_field_preference(self
, format
, field
):
6441 type = self
._get
_field
_setting
(field
, 'type') # extractor, boolean, ordered, field, multiple
6442 get_value
= lambda f
: format
.get(self
._get
_field
_setting
(f
, 'field'))
6443 if type == 'multiple':
6444 type = 'field' # Only 'field' is allowed in multiple for now
6445 actual_fields
= self
._get
_field
_setting
(field
, 'field')
6447 value
= self
._get
_field
_setting
(field
, 'function')(get_value(f
) for f
in actual_fields
)
6449 value
= get_value(field
)
6450 return self
._calculate
_field
_preference
_from
_value
(format
, field
, type, value
)
6452 def calculate_preference(self
, format
):
6453 # Determine missing protocol
6454 if not format
.get('protocol'):
6455 format
['protocol'] = determine_protocol(format
)
6457 # Determine missing ext
6458 if not format
.get('ext') and 'url' in format
:
6459 format
['ext'] = determine_ext(format
['url'])
6460 if format
.get('vcodec') == 'none':
6461 format
['audio_ext'] = format
['ext'] if format
.get('acodec') != 'none' else 'none'
6462 format
['video_ext'] = 'none'
6464 format
['video_ext'] = format
['ext']
6465 format
['audio_ext'] = 'none'
6466 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
6467 # format['preference'] = -1000
6469 if format
.get('preference') is None and format
.get('ext') == 'flv' and re
.match('[hx]265|he?vc?', format
.get('vcodec') or ''):
6470 # HEVC-over-FLV is out-of-spec by FLV's original spec
6471 # ref. https://trac.ffmpeg.org/ticket/6389
6472 # ref. https://github.com/yt-dlp/yt-dlp/pull/5821
6473 format
['preference'] = -100
6475 # Determine missing bitrates
6476 if format
.get('tbr') is None:
6477 if format
.get('vbr') is not None and format
.get('abr') is not None:
6478 format
['tbr'] = format
.get('vbr', 0) + format
.get('abr', 0)
6480 if format
.get('vcodec') != 'none' and format
.get('vbr') is None:
6481 format
['vbr'] = format
.get('tbr') - format
.get('abr', 0)
6482 if format
.get('acodec') != 'none' and format
.get('abr') is None:
6483 format
['abr'] = format
.get('tbr') - format
.get('vbr', 0)
6485 return tuple(self
._calculate
_field
_preference
(format
, field
) for field
in self
._order
)
6489 has_certifi
= bool(certifi
)
6490 has_websockets
= bool(websockets
)
6493 def load_plugins(name
, suffix
, namespace
):
6494 from .plugins
import load_plugins
6495 ret
= load_plugins(name
, suffix
)
6496 namespace
.update(ret
)