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-match-filter triggered """
1234 msg
= 'Encountered a video that did not match filter, stopping due to --break-match-filter'
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
2060 # `sys.stderr` might be `None` (Ref: https://github.com/pyinstaller/pyinstaller/pull/7217)
2064 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2065 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2067 enc
, buffer = None, out
2068 if 'b' in getattr(out
, 'mode', ''):
2069 enc
= encoding
or preferredencoding()
2070 elif hasattr(out
, 'buffer'):
2072 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2074 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2078 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2079 from . import _IN_CLI
2081 if msg
in deprecation_warning
._cache
:
2083 deprecation_warning
._cache
.add(msg
)
2085 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2086 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2089 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2092 deprecation_warning
._cache
= set()
2095 def bytes_to_intlist(bs
):
2098 if isinstance(bs
[0], int): # Python 3
2101 return [ord(c
) for c
in bs
]
2104 def intlist_to_bytes(xs
):
2107 return struct
.pack('%dB' % len(xs
), *xs
)
2110 class LockingUnsupportedError(OSError):
2111 msg
= 'File locking is not supported'
2114 super().__init
__(self
.msg
)
2117 # Cross-platform file locking
2118 if sys
.platform
== 'win32':
2120 import ctypes
.wintypes
2123 class OVERLAPPED(ctypes
.Structure
):
2125 ('Internal', ctypes
.wintypes
.LPVOID
),
2126 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2127 ('Offset', ctypes
.wintypes
.DWORD
),
2128 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2129 ('hEvent', ctypes
.wintypes
.HANDLE
),
2132 kernel32
= ctypes
.WinDLL('kernel32')
2133 LockFileEx
= kernel32
.LockFileEx
2134 LockFileEx
.argtypes
= [
2135 ctypes
.wintypes
.HANDLE
, # hFile
2136 ctypes
.wintypes
.DWORD
, # dwFlags
2137 ctypes
.wintypes
.DWORD
, # dwReserved
2138 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2139 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2140 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2142 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2143 UnlockFileEx
= kernel32
.UnlockFileEx
2144 UnlockFileEx
.argtypes
= [
2145 ctypes
.wintypes
.HANDLE
, # hFile
2146 ctypes
.wintypes
.DWORD
, # dwReserved
2147 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2148 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2149 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2151 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2152 whole_low
= 0xffffffff
2153 whole_high
= 0x7fffffff
2155 def _lock_file(f
, exclusive
, block
):
2156 overlapped
= OVERLAPPED()
2157 overlapped
.Offset
= 0
2158 overlapped
.OffsetHigh
= 0
2159 overlapped
.hEvent
= 0
2160 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2162 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2163 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2164 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2165 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2166 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2168 def _unlock_file(f
):
2169 assert f
._lock
_file
_overlapped
_p
2170 handle
= msvcrt
.get_osfhandle(f
.fileno())
2171 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2172 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2178 def _lock_file(f
, exclusive
, block
):
2179 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2181 flags |
= fcntl
.LOCK_NB
2183 fcntl
.flock(f
, flags
)
2184 except BlockingIOError
:
2186 except OSError: # AOSP does not have flock()
2187 fcntl
.lockf(f
, flags
)
2189 def _unlock_file(f
):
2190 with contextlib
.suppress(OSError):
2191 return fcntl
.flock(f
, fcntl
.LOCK_UN
)
2192 with contextlib
.suppress(OSError):
2193 return fcntl
.lockf(f
, fcntl
.LOCK_UN
) # AOSP does not have flock()
2194 return fcntl
.flock(f
, fcntl
.LOCK_UN | fcntl
.LOCK_NB
) # virtiofs needs LOCK_NB on unlocking
2198 def _lock_file(f
, exclusive
, block
):
2199 raise LockingUnsupportedError()
2201 def _unlock_file(f
):
2202 raise LockingUnsupportedError()
2208 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2209 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2210 raise NotImplementedError(mode
)
2211 self
.mode
, self
.block
= mode
, block
2213 writable
= any(f
in mode
for f
in 'wax+')
2214 readable
= any(f
in mode
for f
in 'r+')
2215 flags
= functools
.reduce(operator
.ior
, (
2216 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2217 getattr(os
, 'O_BINARY', 0), # Windows only
2218 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2219 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2220 os
.O_APPEND
if 'a' in mode
else 0,
2221 os
.O_EXCL
if 'x' in mode
else 0,
2222 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2225 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2227 def __enter__(self
):
2228 exclusive
= 'r' not in self
.mode
2230 _lock_file(self
.f
, exclusive
, self
.block
)
2235 if 'w' in self
.mode
:
2238 except OSError as e
:
2240 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2241 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2250 _unlock_file(self
.f
)
2254 def __exit__(self
, *_
):
2263 def __getattr__(self
, attr
):
2264 return getattr(self
.f
, attr
)
2271 def get_filesystem_encoding():
2272 encoding
= sys
.getfilesystemencoding()
2273 return encoding
if encoding
is not None else 'utf-8'
2276 def shell_quote(args
):
2278 encoding
= get_filesystem_encoding()
2280 if isinstance(a
, bytes):
2281 # We may get a filename encoded with 'encodeFilename'
2282 a
= a
.decode(encoding
)
2283 quoted_args
.append(compat_shlex_quote(a
))
2284 return ' '.join(quoted_args
)
2287 def smuggle_url(url
, data
):
2288 """ Pass additional data in a URL for internal use. """
2290 url
, idata
= unsmuggle_url(url
, {})
2292 sdata
= urllib
.parse
.urlencode(
2293 {'__youtubedl_smuggle': json.dumps(data)}
)
2294 return url
+ '#' + sdata
2297 def unsmuggle_url(smug_url
, default
=None):
2298 if '#__youtubedl_smuggle' not in smug_url
:
2299 return smug_url
, default
2300 url
, _
, sdata
= smug_url
.rpartition('#')
2301 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2302 data
= json
.loads(jsond
)
2306 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2307 """ Formats numbers with decimal sufixes like K, M, etc """
2308 num
, factor
= float_or_none(num
), float(factor
)
2309 if num
is None or num
< 0:
2311 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2312 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2313 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2315 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2316 converted
= num
/ (factor
** exponent
)
2317 return fmt
% (converted
, suffix
)
2320 def format_bytes(bytes):
2321 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2324 def lookup_unit_table(unit_table
, s
, strict
=False):
2325 num_re
= NUMBER_RE
if strict
else NUMBER_RE
.replace(R
'\.', '[,.]')
2326 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2327 m
= (re
.fullmatch
if strict
else re
.match
)(
2328 rf
'(?P<num>{num_re})\s*(?P<unit>{units_re})\b', s
)
2332 num
= float(m
.group('num').replace(',', '.'))
2333 mult
= unit_table
[m
.group('unit')]
2334 return round(num
* mult
)
2338 """Parse a string indicating a byte quantity into an integer"""
2339 return lookup_unit_table(
2340 {u: 1024**i for i, u in enumerate(['', *'KMGTPEZY'])}
,
2341 s
.upper(), strict
=True)
2344 def parse_filesize(s
):
2348 # The lower-case forms are of course incorrect and unofficial,
2349 # but we support those too
2366 'megabytes': 1000 ** 2,
2367 'mebibytes': 1024 ** 2,
2373 'gigabytes': 1000 ** 3,
2374 'gibibytes': 1024 ** 3,
2380 'terabytes': 1000 ** 4,
2381 'tebibytes': 1024 ** 4,
2387 'petabytes': 1000 ** 5,
2388 'pebibytes': 1024 ** 5,
2394 'exabytes': 1000 ** 6,
2395 'exbibytes': 1024 ** 6,
2401 'zettabytes': 1000 ** 7,
2402 'zebibytes': 1024 ** 7,
2408 'yottabytes': 1000 ** 8,
2409 'yobibytes': 1024 ** 8,
2412 return lookup_unit_table(_UNIT_TABLE
, s
)
2419 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2421 if re
.match(r
'^[\d,.]+$', s
):
2422 return str_to_int(s
)
2435 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2439 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2441 return str_to_int(mobj
.group(1))
2444 def parse_resolution(s
, *, lenient
=False):
2449 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2451 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2454 'width': int(mobj
.group('w')),
2455 'height': int(mobj
.group('h')),
2458 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2460 return {'height': int(mobj.group(1))}
2462 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2464 return {'height': int(mobj.group(1)) * 540}
2469 def parse_bitrate(s
):
2470 if not isinstance(s
, str):
2472 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2474 return int(mobj
.group(1))
2477 def month_by_name(name
, lang
='en'):
2478 """ Return the number of a month by (locale-independently) English name """
2480 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2483 return month_names
.index(name
) + 1
2488 def month_by_abbreviation(abbrev
):
2489 """ Return the number of a month by (locale-independently) English
2493 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2498 def fix_xml_ampersands(xml_str
):
2499 """Replace all the '&' by '&' in XML"""
2501 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2506 def setproctitle(title
):
2507 assert isinstance(title
, str)
2509 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2516 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2520 # LoadLibrary in Windows Python 2.7.13 only expects
2521 # a bytestring, but since unicode_literals turns
2522 # every string into a unicode string, it fails.
2524 title_bytes
= title
.encode()
2525 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2526 buf
.value
= title_bytes
2528 libc
.prctl(15, buf
, 0, 0, 0)
2529 except AttributeError:
2530 return # Strange libc, just skip this
2533 def remove_start(s
, start
):
2534 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2537 def remove_end(s
, end
):
2538 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2541 def remove_quotes(s
):
2542 if s
is None or len(s
) < 2:
2544 for quote
in ('"', "'", ):
2545 if s
[0] == quote
and s
[-1] == quote
:
2550 def get_domain(url
):
2552 This implementation is inconsistent, but is kept for compatibility.
2553 Use this only for "webpage_url_domain"
2555 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2558 def url_basename(url
):
2559 path
= urllib
.parse
.urlparse(url
).path
2560 return path
.strip('/').split('/')[-1]
2564 return re
.match(r
'https?://[^?#]+/', url
).group()
2567 def urljoin(base
, path
):
2568 if isinstance(path
, bytes):
2569 path
= path
.decode()
2570 if not isinstance(path
, str) or not path
:
2572 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2574 if isinstance(base
, bytes):
2575 base
= base
.decode()
2576 if not isinstance(base
, str) or not re
.match(
2577 r
'^(?:https?:)?//', base
):
2579 return urllib
.parse
.urljoin(base
, path
)
2582 class HEADRequest(urllib
.request
.Request
):
2583 def get_method(self
):
2587 class PUTRequest(urllib
.request
.Request
):
2588 def get_method(self
):
2592 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2593 if get_attr
and v
is not None:
2594 v
= getattr(v
, get_attr
, None)
2596 return int(v
) * invscale
// scale
2597 except (ValueError, TypeError, OverflowError):
2601 def str_or_none(v
, default
=None):
2602 return default
if v
is None else str(v
)
2605 def str_to_int(int_str
):
2606 """ A more relaxed version of int_or_none """
2607 if isinstance(int_str
, int):
2609 elif isinstance(int_str
, str):
2610 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2611 return int_or_none(int_str
)
2614 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2618 return float(v
) * invscale
/ scale
2619 except (ValueError, TypeError):
2623 def bool_or_none(v
, default
=None):
2624 return v
if isinstance(v
, bool) else default
2627 def strip_or_none(v
, default
=None):
2628 return v
.strip() if isinstance(v
, str) else default
2631 def url_or_none(url
):
2632 if not url
or not isinstance(url
, str):
2635 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2638 def request_to_url(req
):
2639 if isinstance(req
, urllib
.request
.Request
):
2640 return req
.get_full_url()
2645 def strftime_or_none(timestamp
, date_format
, default
=None):
2646 datetime_object
= None
2648 if isinstance(timestamp
, (int, float)): # unix timestamp
2649 # Using naive datetime here can break timestamp() in Windows
2650 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2651 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2652 elif isinstance(timestamp
, str): # assume YYYYMMDD
2653 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2654 date_format
= re
.sub( # Support %s on windows
2655 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2656 return datetime_object
.strftime(date_format
)
2657 except (ValueError, TypeError, AttributeError):
2661 def parse_duration(s
):
2662 if not isinstance(s
, str):
2668 days
, hours
, mins
, secs
, ms
= [None] * 5
2669 m
= re
.match(r
'''(?x)
2671 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2672 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2673 (?P<ms>[.:][0-9]+)?Z?$
2676 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2681 [0-9]+\s*y(?:ears?)?,?\s*
2684 [0-9]+\s*m(?:onths?)?,?\s*
2687 [0-9]+\s*w(?:eeks?)?,?\s*
2690 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2694 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2697 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2700 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2703 days
, hours
, mins
, secs
, ms
= m
.groups()
2705 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2707 hours
, mins
= m
.groups()
2712 ms
= ms
.replace(':', '.')
2713 return sum(float(part
or 0) * mult
for part
, mult
in (
2714 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2717 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2718 name
, real_ext
= os
.path
.splitext(filename
)
2720 f
'{name}.{ext}{real_ext}'
2721 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2722 else f
'{filename}.{ext}')
2725 def replace_extension(filename
, ext
, expected_real_ext
=None):
2726 name
, real_ext
= os
.path
.splitext(filename
)
2727 return '{}.{}'.format(
2728 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2732 def check_executable(exe
, args
=[]):
2733 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2734 args can be a list of arguments for a short output (like -version) """
2736 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2742 def _get_exe_version_output(exe
, args
):
2744 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2745 # SIGTTOU if yt-dlp is run in the background.
2746 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2747 stdout
, _
, ret
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2748 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2756 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2757 assert isinstance(output
, str)
2758 if version_re
is None:
2759 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2760 m
= re
.search(version_re
, output
)
2767 def get_exe_version(exe
, args
=['--version'],
2768 version_re
=None, unrecognized
=('present', 'broken')):
2769 """ Returns the version of the specified executable,
2770 or False if the executable is not present """
2771 unrecognized
= variadic(unrecognized
)
2772 assert len(unrecognized
) in (1, 2)
2773 out
= _get_exe_version_output(exe
, args
)
2775 return unrecognized
[-1]
2776 return out
and detect_exe_version(out
, version_re
, unrecognized
[0])
2779 def frange(start
=0, stop
=None, step
=1):
2782 start
, stop
= 0, start
2783 sign
= [-1, 1][step
> 0] if step
else 0
2784 while sign
* start
< sign
* stop
:
2789 class LazyList(collections
.abc
.Sequence
):
2790 """Lazy immutable list from an iterable
2791 Note that slices of a LazyList are lists and not LazyList"""
2793 class IndexError(IndexError):
2796 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2797 self
._iterable
= iter(iterable
)
2798 self
._cache
= [] if _cache
is None else _cache
2799 self
._reversed
= reverse
2803 # We need to consume the entire iterable to iterate in reverse
2804 yield from self
.exhaust()
2806 yield from self
._cache
2807 for item
in self
._iterable
:
2808 self
._cache
.append(item
)
2812 self
._cache
.extend(self
._iterable
)
2813 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2817 """Evaluate the entire iterable"""
2818 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2821 def _reverse_index(x
):
2822 return None if x
is None else ~x
2824 def __getitem__(self
, idx
):
2825 if isinstance(idx
, slice):
2827 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2828 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2829 elif isinstance(idx
, int):
2831 idx
= self
._reverse
_index
(idx
)
2832 start
, stop
, step
= idx
, idx
, 0
2834 raise TypeError('indices must be integers or slices')
2835 if ((start
or 0) < 0 or (stop
or 0) < 0
2836 or (start
is None and step
< 0)
2837 or (stop
is None and step
> 0)):
2838 # We need to consume the entire iterable to be able to slice from the end
2839 # Obviously, never use this with infinite iterables
2842 return self
._cache
[idx
]
2843 except IndexError as e
:
2844 raise self
.IndexError(e
) from e
2845 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2847 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2849 return self
._cache
[idx
]
2850 except IndexError as e
:
2851 raise self
.IndexError(e
) from e
2855 self
[-1] if self
._reversed
else self
[0]
2856 except self
.IndexError:
2862 return len(self
._cache
)
2864 def __reversed__(self
):
2865 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2868 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2871 # repr and str should mimic a list. So we exhaust the iterable
2872 return repr(self
.exhaust())
2875 return repr(self
.exhaust())
2880 class IndexError(IndexError):
2884 # This is only useful for tests
2885 return len(self
.getslice())
2887 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2888 self
._pagefunc
= pagefunc
2889 self
._pagesize
= pagesize
2890 self
._pagecount
= float('inf')
2891 self
._use
_cache
= use_cache
2894 def getpage(self
, pagenum
):
2895 page_results
= self
._cache
.get(pagenum
)
2896 if page_results
is None:
2897 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2899 self
._cache
[pagenum
] = page_results
2902 def getslice(self
, start
=0, end
=None):
2903 return list(self
._getslice
(start
, end
))
2905 def _getslice(self
, start
, end
):
2906 raise NotImplementedError('This method must be implemented by subclasses')
2908 def __getitem__(self
, idx
):
2909 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2910 if not isinstance(idx
, int) or idx
< 0:
2911 raise TypeError('indices must be non-negative integers')
2912 entries
= self
.getslice(idx
, idx
+ 1)
2914 raise self
.IndexError()
2918 class OnDemandPagedList(PagedList
):
2919 """Download pages until a page with less than maximum results"""
2921 def _getslice(self
, start
, end
):
2922 for pagenum
in itertools
.count(start
// self
._pagesize
):
2923 firstid
= pagenum
* self
._pagesize
2924 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2925 if start
>= nextfirstid
:
2929 start
% self
._pagesize
2930 if firstid
<= start
< nextfirstid
2933 ((end
- 1) % self
._pagesize
) + 1
2934 if (end
is not None and firstid
<= end
<= nextfirstid
)
2938 page_results
= self
.getpage(pagenum
)
2940 self
._pagecount
= pagenum
- 1
2942 if startv
!= 0 or endv
is not None:
2943 page_results
= page_results
[startv
:endv
]
2944 yield from page_results
2946 # A little optimization - if current page is not "full", ie. does
2947 # not contain page_size videos then we can assume that this page
2948 # is the last one - there are no more ids on further pages -
2949 # i.e. no need to query again.
2950 if len(page_results
) + startv
< self
._pagesize
:
2953 # If we got the whole page, but the next page is not interesting,
2954 # break out early as well
2955 if end
== nextfirstid
:
2959 class InAdvancePagedList(PagedList
):
2960 """PagedList with total number of pages known in advance"""
2962 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2963 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2964 self
._pagecount
= pagecount
2966 def _getslice(self
, start
, end
):
2967 start_page
= start
// self
._pagesize
2968 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2969 skip_elems
= start
- start_page
* self
._pagesize
2970 only_more
= None if end
is None else end
- start
2971 for pagenum
in range(start_page
, end_page
):
2972 page_results
= self
.getpage(pagenum
)
2974 page_results
= page_results
[skip_elems
:]
2976 if only_more
is not None:
2977 if len(page_results
) < only_more
:
2978 only_more
-= len(page_results
)
2980 yield from page_results
[:only_more
]
2982 yield from page_results
2985 class PlaylistEntries
:
2986 MissingEntry
= object()
2987 is_exhausted
= False
2989 def __init__(self
, ydl
, info_dict
):
2992 # _entries must be assigned now since infodict can change during iteration
2993 entries
= info_dict
.get('entries')
2995 raise EntryNotInPlaylist('There are no entries')
2996 elif isinstance(entries
, list):
2997 self
.is_exhausted
= True
2999 requested_entries
= info_dict
.get('requested_entries')
3000 self
.is_incomplete
= requested_entries
is not None
3001 if self
.is_incomplete
:
3002 assert self
.is_exhausted
3003 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
3004 for i
, entry
in zip(requested_entries
, entries
):
3005 self
._entries
[i
- 1] = entry
3006 elif isinstance(entries
, (list, PagedList
, LazyList
)):
3007 self
._entries
= entries
3009 self
._entries
= LazyList(entries
)
3011 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
3012 (?P<start>[+-]?\d+)?
3014 (?P<end>[+-]?\d+|inf(?:inite)?)?
3015 (?::(?P<step>[+-]?\d+))?
3019 def parse_playlist_items(cls
, string
):
3020 for segment
in string
.split(','):
3022 raise ValueError('There is two or more consecutive commas')
3023 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
3025 raise ValueError(f
'{segment!r} is not a valid specification')
3026 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
3027 if int_or_none(step
) == 0:
3028 raise ValueError(f
'Step in {segment!r} cannot be zero')
3029 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
3031 def get_requested_items(self
):
3032 playlist_items
= self
.ydl
.params
.get('playlist_items')
3033 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
3034 playlist_end
= self
.ydl
.params
.get('playlistend')
3035 # For backwards compatibility, interpret -1 as whole list
3036 if playlist_end
in (-1, None):
3038 if not playlist_items
:
3039 playlist_items
= f
'{playlist_start}:{playlist_end}'
3040 elif playlist_start
!= 1 or playlist_end
:
3041 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
3043 for index
in self
.parse_playlist_items(playlist_items
):
3044 for i
, entry
in self
[index
]:
3049 # The item may have just been added to archive. Don't break due to it
3050 if not self
.ydl
.params
.get('lazy_playlist'):
3051 # TODO: Add auto-generated fields
3052 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3053 except (ExistingVideoReached
, RejectedVideoReached
):
3056 def get_full_count(self
):
3057 if self
.is_exhausted
and not self
.is_incomplete
:
3059 elif isinstance(self
._entries
, InAdvancePagedList
):
3060 if self
._entries
._pagesize
== 1:
3061 return self
._entries
._pagecount
3063 @functools.cached_property
3065 if isinstance(self
._entries
, list):
3068 entry
= self
._entries
[i
]
3070 entry
= self
.MissingEntry
3071 if not self
.is_incomplete
:
3072 raise self
.IndexError()
3073 if entry
is self
.MissingEntry
:
3074 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
3079 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3080 except (LazyList
.IndexError, PagedList
.IndexError):
3081 raise self
.IndexError()
3084 def __getitem__(self
, idx
):
3085 if isinstance(idx
, int):
3086 idx
= slice(idx
, idx
)
3088 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3089 step
= 1 if idx
.step
is None else idx
.step
3090 if idx
.start
is None:
3091 start
= 0 if step
> 0 else len(self
) - 1
3093 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3095 # NB: Do not call len(self) when idx == [:]
3096 if idx
.stop
is None:
3097 stop
= 0 if step
< 0 else float('inf')
3099 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3100 stop
+= [-1, 1][step
> 0]
3102 for i
in frange(start
, stop
, step
):
3106 entry
= self
._getter
(i
)
3107 except self
.IndexError:
3108 self
.is_exhausted
= True
3115 return len(tuple(self
[:]))
3117 class IndexError(IndexError):
3121 def uppercase_escape(s
):
3122 unicode_escape
= codecs
.getdecoder('unicode_escape')
3124 r
'\\U[0-9a-fA-F]{8}',
3125 lambda m
: unicode_escape(m
.group(0))[0],
3129 def lowercase_escape(s
):
3130 unicode_escape
= codecs
.getdecoder('unicode_escape')
3132 r
'\\u[0-9a-fA-F]{4}',
3133 lambda m
: unicode_escape(m
.group(0))[0],
3137 def escape_rfc3986(s
):
3138 """Escape non-ASCII characters as suggested by RFC 3986"""
3139 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3142 def escape_url(url
):
3143 """Escape URL as suggested by RFC 3986"""
3144 url_parsed
= urllib
.parse
.urlparse(url
)
3145 return url_parsed
._replace
(
3146 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3147 path
=escape_rfc3986(url_parsed
.path
),
3148 params
=escape_rfc3986(url_parsed
.params
),
3149 query
=escape_rfc3986(url_parsed
.query
),
3150 fragment
=escape_rfc3986(url_parsed
.fragment
)
3154 def parse_qs(url
, **kwargs
):
3155 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3158 def read_batch_urls(batch_fd
):
3160 if not isinstance(url
, str):
3161 url
= url
.decode('utf-8', 'replace')
3162 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3163 for bom
in BOM_UTF8
:
3164 if url
.startswith(bom
):
3165 url
= url
[len(bom
):]
3167 if not url
or url
.startswith(('#', ';', ']')):
3169 # "#" cannot be stripped out since it is part of the URI
3170 # However, it can be safely stripped out if following a whitespace
3171 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3173 with contextlib
.closing(batch_fd
) as fd
:
3174 return [url
for url
in map(fixup
, fd
) if url
]
3177 def urlencode_postdata(*args
, **kargs
):
3178 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3181 def update_url(url
, *, query_update
=None, **kwargs
):
3182 """Replace URL components specified by kwargs
3183 @param url str or parse url tuple
3184 @param query_update update query
3187 if isinstance(url
, str):
3188 if not kwargs
and not query_update
:
3191 url
= urllib
.parse
.urlparse(url
)
3193 assert 'query' not in kwargs
, 'query_update and query cannot be specified at the same time'
3194 kwargs
['query'] = urllib
.parse
.urlencode({
3195 **urllib
.parse
.parse_qs(url
.query
),
3198 return urllib
.parse
.urlunparse(url
._replace
(**kwargs
))
3201 def update_url_query(url
, query
):
3202 return update_url(url
, query_update
=query
)
3205 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3206 req_headers
= req
.headers
.copy()
3207 req_headers
.update(headers
or {})
3208 req_data
= data
or req
.data
3209 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3210 req_get_method
= req
.get_method()
3211 if req_get_method
== 'HEAD':
3212 req_type
= HEADRequest
3213 elif req_get_method
== 'PUT':
3214 req_type
= PUTRequest
3216 req_type
= urllib
.request
.Request
3218 req_url
, data
=req_data
, headers
=req_headers
,
3219 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3220 if hasattr(req
, 'timeout'):
3221 new_req
.timeout
= req
.timeout
3225 def _multipart_encode_impl(data
, boundary
):
3226 content_type
= 'multipart/form-data; boundary=%s' % boundary
3229 for k
, v
in data
.items():
3230 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3231 if isinstance(k
, str):
3233 if isinstance(v
, str):
3235 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3236 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3237 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3238 if boundary
.encode('ascii') in content
:
3239 raise ValueError('Boundary overlaps with data')
3242 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3244 return out
, content_type
3247 def multipart_encode(data
, boundary
=None):
3249 Encode a dict to RFC 7578-compliant form-data
3252 A dict where keys and values can be either Unicode or bytes-like
3255 If specified a Unicode object, it's used as the boundary. Otherwise
3256 a random boundary is generated.
3258 Reference: https://tools.ietf.org/html/rfc7578
3260 has_specified_boundary
= boundary
is not None
3263 if boundary
is None:
3264 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3267 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3270 if has_specified_boundary
:
3274 return out
, content_type
3277 def is_iterable_like(x
, allowed_types
=collections
.abc
.Iterable
, blocked_types
=NO_DEFAULT
):
3278 if blocked_types
is NO_DEFAULT
:
3279 blocked_types
= (str, bytes, collections
.abc
.Mapping
)
3280 return isinstance(x
, allowed_types
) and not isinstance(x
, blocked_types
)
3283 def variadic(x
, allowed_types
=NO_DEFAULT
):
3284 return x
if is_iterable_like(x
, blocked_types
=allowed_types
) else (x
, )
3287 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3288 for val
in map(d
.get
, variadic(key_or_keys
)):
3289 if val
is not None and (val
or not skip_false_values
):
3294 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3297 val
= f(*args
, **kwargs
)
3298 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3301 if expected_type
is None or isinstance(val
, expected_type
):
3305 def try_get(src
, getter
, expected_type
=None):
3306 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3309 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3310 return {k: v for k, v in dct.items() if cndn(k, v)}
3313 def merge_dicts(*dicts
):
3315 for a_dict
in dicts
:
3316 for k
, v
in a_dict
.items():
3317 if (v
is not None and k
not in merged
3318 or isinstance(v
, str) and merged
[k
] == ''):
3323 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3324 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3336 TV_PARENTAL_GUIDELINES
= {
3346 def parse_age_limit(s
):
3347 # isinstance(False, int) is True. So type() must be used instead
3348 if type(s
) is int: # noqa: E721
3349 return s
if 0 <= s
<= 21 else None
3350 elif not isinstance(s
, str):
3352 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3354 return int(m
.group('age'))
3357 return US_RATINGS
[s
]
3358 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3360 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3364 def strip_jsonp(code
):
3367 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3368 (?:\s*&&\s*(?P=func_name))?
3369 \s*\(\s*(?P<callback_data>.*)\);?
3370 \s*?(?://[^\n]*)*$''',
3371 r
'\g<callback_data>', code
)
3374 def js_to_json(code
, vars={}, *, strict
=False):
3375 # vars is a dict of var, val pairs to substitute
3376 STRING_QUOTES
= '\'"`'
3377 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3378 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3379 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3381 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3382 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3385 def process_escape(match
):
3386 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3387 escape
= match
.group(1) or match
.group(2)
3389 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3390 else R
'\u00' if escape
== 'x'
3391 else '' if escape
== '\n'
3394 def template_substitute(match
):
3395 evaluated
= js_to_json(match
.group(1), vars, strict
=strict
)
3396 if evaluated
[0] == '"':
3397 return json
.loads(evaluated
)
3402 if v
in ('true', 'false', 'null'):
3404 elif v
in ('undefined', 'void 0'):
3406 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3409 if v
[0] in STRING_QUOTES
:
3410 v
= re
.sub(r
'(?s)\${([^}]+)}', template_substitute
, v
[1:-1]) if v
[0] == '`' else v
[1:-1]
3411 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
)
3412 return f
'"{escaped}"'
3414 for regex
, base
in INTEGER_TABLE
:
3415 im
= re
.match(regex
, v
)
3417 i
= int(im
.group(1), base
)
3418 return f
'"{i}":' if v
.endswith(':') else str(i
)
3424 except json
.JSONDecodeError
:
3425 return json
.dumps(vars[v
])
3432 raise ValueError(f
'Unknown value: {v}')
3434 def create_map(mobj
):
3435 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3437 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3439 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3440 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3441 code
= re
.sub(r
'parseInt\([^\d]+(\d+)[^\d]+\)', r
'\1', code
)
3442 code
= re
.sub(r
'\(function\([^)]*\)\s*\{[^}]*\}\s*\)\s*\(\s*(["\'][^
)]*["\'])\s*\)', r'\1', code)
3444 return re.sub(rf'''(?sx)
3446 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3447 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3448 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3449 [0-9]+(?={SKIP_RE}:)|
3454 def qualities(quality_ids):
3455 """ Get a numeric quality value out of a list of possible values """
3458 return quality_ids.index(qid)
3464 POSTPROCESS_WHEN = ('pre_process', 'after_filter', 'video', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3468 'default': '%(title)s [%(id)s].%(ext)s',
3469 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3475 'description': 'description',
3476 'annotation': 'annotations.xml',
3477 'infojson': 'info.json',
3480 'pl_thumbnail': None,
3481 'pl_description': 'description',
3482 'pl_infojson': 'info.json',
3485 # As of [1] format syntax is:
3486 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3487 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3488 STR_FORMAT_RE_TMPL = r'''(?x)
3489 (?<!%)(?P<prefix>(?:%%)*)
3491 (?P<has_key>\((?P<key>{0})\))?
3493 (?P<conversion>[#0\-+ ]+)?
3495 (?P<precision>\.\d+)?
3496 (?P<len_mod>[hlL])? # unused in python
3497 {1} # conversion type
3502 STR_FORMAT_TYPES = 'diouxXeEfFgGcrs'
3505 def limit_length(s, length):
3506 """ Add ellipses to overly long strings """
3511 return s[:length - len(ELLIPSES)] + ELLIPSES
3515 def version_tuple(v):
3516 return tuple(int(e) for e in re.split(r'[-.]', v))
3519 def is_outdated_version(version, limit, assume_new=True):
3521 return not assume_new
3523 return version_tuple(version) < version_tuple(limit)
3525 return not assume_new
3528 def ytdl_is_updateable():
3529 """ Returns if yt-dlp can be updated with -U """
3531 from .update import is_non_updateable
3533 return not is_non_updateable()
3536 def args_to_str(args):
3537 # Get a short string representation for a subprocess command
3538 return ' '.join(compat_shlex_quote(a) for a in args)
3541 def error_to_compat_str(err):
3545 def error_to_str(err):
3546 return f'{type(err).__name__}: {err}'
3549 def mimetype2ext(mt, default=NO_DEFAULT):
3550 if not isinstance(mt, str):
3551 if default is not NO_DEFAULT:
3567 'x-matroska': 'mkv',
3569 'x-mp4-fragmented': 'mp4',
3574 # application (streaming playlists)
3578 'vnd.apple.mpegurl': 'm3u8',
3579 'vnd.ms-sstr+xml': 'ism',
3580 'x-mpegurl': 'm3u8',
3584 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3.
3585 # Using .mp3 as it's the most popular one
3586 'audio/mpeg': 'mp3',
3587 'audio/webm': 'webm',
3588 'audio/x-matroska': 'mka',
3589 'audio/x-mpegurl': 'm3u',
3597 'x-realaudio': 'ra',
3608 'vnd.wap.wbmp': 'wbmp',
3615 'filmstrip+json': 'fs',
3616 'smptett+xml': 'tt',
3619 'x-ms-sami': 'sami',
3628 mimetype = mt.partition(';')[0].strip().lower()
3629 _, _, subtype = mimetype.rpartition('/')
3631 ext = traverse_obj(MAP, mimetype, subtype, subtype.rsplit('+')[-1])
3634 elif default is not NO_DEFAULT:
3636 return subtype.replace('+', '.')
3639 def ext2mimetype(ext_or_url):
3642 if '.' not in ext_or_url:
3643 ext_or_url = f'file.{ext_or_url}'
3644 return mimetypes.guess_type(ext_or_url)[0]
3647 def parse_codecs(codecs_str):
3648 # http://tools.ietf.org/html/rfc6381
3651 split_codecs = list(filter(None, map(
3652 str.strip, codecs_str.strip().strip(',').split(','))))
3653 vcodec, acodec, scodec, hdr = None, None, None, None
3654 for full_codec in split_codecs:
3655 parts = re.sub(r'0+(?=\d)', '', full_codec).split('.')
3656 if parts[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3657 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3661 if parts[0] in ('dvh1', 'dvhe'):
3663 elif parts[0] == 'av1' and traverse_obj(parts, 3) == '10':
3665 elif parts[:2] == ['vp9', '2']:
3667 elif parts[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-4',
3668 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3669 acodec = acodec or full_codec
3670 elif parts[0] in ('stpp', 'wvtt'):
3671 scodec = scodec or full_codec
3673 write_string(f'WARNING: Unknown codec {full_codec}\n')
3674 if vcodec or acodec or scodec:
3676 'vcodec': vcodec or 'none',
3677 'acodec': acodec or 'none',
3678 'dynamic_range': hdr,
3679 **({'scodec': scodec} if scodec is not None else {}),
3681 elif len(split_codecs) == 2:
3683 'vcodec': split_codecs[0],
3684 'acodec': split_codecs[1],
3689 def get_compatible_ext(*, vcodecs, acodecs, vexts, aexts, preferences=None):
3690 assert len(vcodecs) == len(vexts) and len(acodecs) == len(aexts)
3692 allow_mkv = not preferences or 'mkv' in preferences
3694 if allow_mkv and max(len(acodecs), len(vcodecs)) > 1:
3695 return 'mkv' # TODO: any other format allows this?
3697 # TODO: All codecs supported by parse_codecs isn't handled here
3698 COMPATIBLE_CODECS = {
3700 'av1', 'hevc', 'avc1', 'mp4a', 'ac-4', # fourcc (m3u8, mpd)
3701 'h264', 'aacl', 'ec-3', # Set in ISM
3704 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3705 'vp9x', 'vp8x', # in the webm spec
3709 sanitize_codec = functools.partial(
3710 try_get, getter=lambda x: x[0].split('.')[0].replace('0', '').lower())
3711 vcodec, acodec = sanitize_codec(vcodecs), sanitize_codec(acodecs)
3713 for ext in preferences or COMPATIBLE_CODECS.keys():
3714 codec_set = COMPATIBLE_CODECS.get(ext, set())
3715 if ext == 'mkv' or codec_set.issuperset((vcodec, acodec)):
3719 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'},
3722 for ext in preferences or vexts:
3723 current_exts = {ext, *vexts, *aexts}
3724 if ext == 'mkv' or current_exts == {ext} or any(
3725 ext_sets.issuperset(current_exts) for ext_sets in COMPATIBLE_EXTS):
3727 return 'mkv' if allow_mkv else preferences[-1]
3730 def urlhandle_detect_ext(url_handle, default=NO_DEFAULT):
3731 getheader = url_handle.headers.get
3733 cd = getheader('Content-Disposition')
3735 m = re.match(r'attachment;\s*filename="(?P
<filename
>[^
"]+)"', cd)
3737 e = determine_ext(m.group('filename
'), default_ext=None)
3741 meta_ext = getheader('x
-amz
-meta
-name
')
3743 e = meta_ext.rpartition('.')[2]
3747 return mimetype2ext(getheader('Content
-Type
'), default=default)
3750 def encode_data_uri(data, mime_type):
3751 return 'data
:%s;base64
,%s' % (mime_type, base64.b64encode(data).decode('ascii
'))
3754 def age_restricted(content_limit, age_limit):
3755 """ Returns True iff the content should be blocked """
3757 if age_limit is None: # No limit set
3759 if content_limit is None:
3760 return False # Content available for everyone
3761 return age_limit < content_limit
3764 # List of known byte-order-marks (BOM)
3766 (b'\xef\xbb\xbf', 'utf
-8'),
3767 (b'\x00\x00\xfe\xff', 'utf
-32-be
'),
3768 (b'\xff\xfe\x00\x00', 'utf
-32-le
'),
3769 (b'\xff\xfe', 'utf
-16-le
'),
3770 (b'\xfe\xff', 'utf
-16-be
'),
3774 def is_html(first_bytes):
3775 """ Detect whether a file contains HTML by examining its first bytes. """
3778 for bom, enc in BOMS:
3779 while first_bytes.startswith(bom):
3780 encoding, first_bytes = enc, first_bytes[len(bom):]
3782 return re.match(r'^\s
*<', first_bytes.decode(encoding, 'replace
'))
3785 def determine_protocol(info_dict):
3786 protocol = info_dict.get('protocol
')
3787 if protocol is not None:
3790 url = sanitize_url(info_dict['url
'])
3791 if url.startswith('rtmp
'):
3793 elif url.startswith('mms
'):
3795 elif url.startswith('rtsp
'):
3798 ext = determine_ext(url)
3800 return 'm3u8
' if info_dict.get('is_live
') else 'm3u8_native
'
3804 return urllib.parse.urlparse(url).scheme
3807 def render_table(header_row, data, delim=False, extra_gap=0, hide_empty=False):
3808 """ Render a list of rows, each as a list of values.
3809 Text after a \t will be right aligned """
3811 return len(remove_terminal_sequences(string).replace('\t', ''))
3813 def get_max_lens(table):
3814 return [max(width(str(v)) for v in col) for col in zip(*table)]
3816 def filter_using_list(row, filterArray):
3817 return [col for take, col in itertools.zip_longest(filterArray, row, fillvalue=True) if take]
3819 max_lens = get_max_lens(data) if hide_empty else []
3820 header_row = filter_using_list(header_row, max_lens)
3821 data = [filter_using_list(row, max_lens) for row in data]
3823 table = [header_row] + data
3824 max_lens = get_max_lens(table)
3827 table = [header_row, [delim * (ml + extra_gap) for ml in max_lens]] + data
3828 table[1][-1] = table[1][-1][:-extra_gap * len(delim)] # Remove extra_gap from end of delimiter
3830 for pos, text in enumerate(map(str, row)):
3832 row[pos] = text.replace('\t', ' ' * (max_lens[pos] - width(text))) + ' ' * extra_gap
3834 row[pos] = text + ' ' * (max_lens[pos] - width(text) + extra_gap)
3835 ret = '\n'.join(''.join(row).rstrip() for row in table)
3839 def _match_one(filter_part, dct, incomplete):
3840 # TODO: Generalize code with YoutubeDL._build_format_filter
3841 STRING_OPERATORS = {
3842 '*=': operator.contains,
3843 '^
=': lambda attr, value: attr.startswith(value),
3844 '$
=': lambda attr, value: attr.endswith(value),
3845 '~
=': lambda attr, value: re.search(value, attr),
3847 COMPARISON_OPERATORS = {
3849 '<=': operator.le, # "<=" must be defined above "<"
3856 if isinstance(incomplete, bool):
3857 is_incomplete = lambda _: incomplete
3859 is_incomplete = lambda k: k in incomplete
3861 operator_rex = re.compile(r'''(?x)
3863 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3865 (?P<quote>["\'])(?P<quotedstrval>.+?)(?P=quote)|
3868 ''' % '|
'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3869 m = operator_rex.fullmatch(filter_part.strip())
3872 unnegated_op = COMPARISON_OPERATORS[m['op
']]
3874 op = lambda attr, value: not unnegated_op(attr, value)
3877 comparison_value = m['quotedstrval
'] or m['strval
'] or m['intval
']
3879 comparison_value = comparison_value.replace(r'\
%s' % m['quote
'], m['quote
'])
3880 actual_value = dct.get(m['key
'])
3881 numeric_comparison = None
3882 if isinstance(actual_value, (int, float)):
3883 # If the original field is a string and matching comparisonvalue is
3884 # a number we should respect the origin of the original field
3885 # and process comparison value as a string (see
3886 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3888 numeric_comparison = int(comparison_value)
3890 numeric_comparison = parse_filesize(comparison_value)
3891 if numeric_comparison is None:
3892 numeric_comparison = parse_filesize(f'{comparison_value}B
')
3893 if numeric_comparison is None:
3894 numeric_comparison = parse_duration(comparison_value)
3895 if numeric_comparison is not None and m['op
'] in STRING_OPERATORS:
3896 raise ValueError('Operator
%s only supports string values
!' % m['op
'])
3897 if actual_value is None:
3898 return is_incomplete(m['key
']) or m['none_inclusive
']
3899 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3902 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3903 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3905 operator_rex = re.compile(r'''(?x)
3906 (?P<op>%s)\s*(?P<key>[a-z_]+)
3907 ''' % '|
'.join(map(re.escape, UNARY_OPERATORS.keys())))
3908 m = operator_rex.fullmatch(filter_part.strip())
3910 op = UNARY_OPERATORS[m.group('op
')]
3911 actual_value = dct.get(m.group('key
'))
3912 if is_incomplete(m.group('key
')) and actual_value is None:
3914 return op(actual_value)
3916 raise ValueError('Invalid
filter part
%r' % filter_part)
3919 def match_str(filter_str, dct, incomplete=False):
3920 """ Filter a dictionary with a simple string syntax.
3921 @returns Whether the filter passes
3922 @param incomplete Set of keys that is expected to be missing from dct.
3923 Can be True/False to indicate all/none of the keys may be missing.
3924 All conditions on incomplete keys pass if the key is missing
3927 _match_one(filter_part.replace(r'\
&', '&'), dct, incomplete)
3928 for filter_part in re.split(r'(?
<!\\)&', filter_str))
3931 def match_filter_func(filters, breaking_filters=None):
3932 if not filters and not breaking_filters:
3934 breaking_filters = match_filter_func(breaking_filters) or (lambda _, __: None)
3935 filters = set(variadic(filters or []))
3937 interactive = '-' in filters
3941 def _match_func(info_dict, incomplete=False):
3942 ret = breaking_filters(info_dict, incomplete)
3944 raise RejectedVideoReached(ret)
3946 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3947 return NO_DEFAULT if interactive and not incomplete else None
3949 video_title = info_dict.get('title
') or info_dict.get('id') or 'entry
'
3950 filter_str = ') |
('.join(map(str.strip, filters))
3951 return f'{video_title} does
not pass filter ({filter_str}
), skipping
..'
3955 class download_range_func:
3956 def __init__(self, chapters, ranges):
3957 self.chapters, self.ranges = chapters, ranges
3959 def __call__(self, info_dict, ydl):
3960 if not self.ranges and not self.chapters:
3963 warning = ('There are no chapters matching the regex
' if info_dict.get('chapters
')
3964 else 'Cannot match chapters since chapter information
is unavailable
')
3965 for regex in self.chapters or []:
3966 for i, chapter in enumerate(info_dict.get('chapters
') or []):
3967 if re.search(regex, chapter['title
']):
3969 yield {**chapter, 'index': i}
3970 if self.chapters and warning:
3971 ydl.to_screen(f'[info
] {info_dict["id"]}
: {warning}
')
3973 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3975 def __eq__(self, other):
3976 return (isinstance(other, download_range_func)
3977 and self.chapters == other.chapters and self.ranges == other.ranges)
3980 return f'{__name__}
.{type(self).__name__}
({self.chapters}
, {self.ranges}
)'
3983 def parse_dfxp_time_expr(time_expr):
3987 mobj = re.match(rf'^
(?P
<time_offset
>{NUMBER_RE}
)s?$
', time_expr)
3989 return float(mobj.group('time_offset
'))
3991 mobj = re.match(r'^
(\d
+):(\d\d
):(\d\
d(?
:(?
:\
.|
:)\d
+)?
)$
', time_expr)
3993 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3996 def srt_subtitles_timecode(seconds):
3997 return '%02d
:%02d
:%02d
,%03d
' % timetuple_from_msec(seconds * 1000)
4000 def ass_subtitles_timecode(seconds):
4001 time = timetuple_from_msec(seconds * 1000)
4002 return '%01d
:%02d
:%02d
.%02d
' % (*time[:-1], time.milliseconds / 10)
4005 def dfxp2srt(dfxp_data):
4007 @param dfxp_data A bytes-like object containing DFXP data
4008 @returns A unicode object containing converted SRT data
4010 LEGACY_NAMESPACES = (
4011 (b'http
://www
.w3
.org
/ns
/ttml
', [
4012 b'http
://www
.w3
.org
/2004/11/ttaf1
',
4013 b'http
://www
.w3
.org
/2006/04/ttaf1
',
4014 b'http
://www
.w3
.org
/2006/10/ttaf1
',
4016 (b'http
://www
.w3
.org
/ns
/ttml
#styling', [
4017 b
'http://www.w3.org/ns/ttml#style',
4021 SUPPORTED_STYLING
= [
4030 _x
= functools
.partial(xpath_with_ns
, ns_map
={
4031 'xml': 'http://www.w3.org/XML/1998/namespace',
4032 'ttml': 'http://www.w3.org/ns/ttml',
4033 'tts': 'http://www.w3.org/ns/ttml#styling',
4039 class TTMLPElementParser
:
4041 _unclosed_elements
= []
4042 _applied_styles
= []
4044 def start(self
, tag
, attrib
):
4045 if tag
in (_x('ttml:br'), 'br'):
4048 unclosed_elements
= []
4050 element_style_id
= attrib
.get('style')
4052 style
.update(default_style
)
4053 if element_style_id
:
4054 style
.update(styles
.get(element_style_id
, {}))
4055 for prop
in SUPPORTED_STYLING
:
4056 prop_val
= attrib
.get(_x('tts:' + prop
))
4058 style
[prop
] = prop_val
4061 for k
, v
in sorted(style
.items()):
4062 if self
._applied
_styles
and self
._applied
_styles
[-1].get(k
) == v
:
4065 font
+= ' color="%s"' % v
4066 elif k
== 'fontSize':
4067 font
+= ' size="%s"' % v
4068 elif k
== 'fontFamily':
4069 font
+= ' face="%s"' % v
4070 elif k
== 'fontWeight' and v
== 'bold':
4072 unclosed_elements
.append('b')
4073 elif k
== 'fontStyle' and v
== 'italic':
4075 unclosed_elements
.append('i')
4076 elif k
== 'textDecoration' and v
== 'underline':
4078 unclosed_elements
.append('u')
4080 self
._out
+= '<font' + font
+ '>'
4081 unclosed_elements
.append('font')
4083 if self
._applied
_styles
:
4084 applied_style
.update(self
._applied
_styles
[-1])
4085 applied_style
.update(style
)
4086 self
._applied
_styles
.append(applied_style
)
4087 self
._unclosed
_elements
.append(unclosed_elements
)
4090 if tag
not in (_x('ttml:br'), 'br'):
4091 unclosed_elements
= self
._unclosed
_elements
.pop()
4092 for element
in reversed(unclosed_elements
):
4093 self
._out
+= '</%s>' % element
4094 if unclosed_elements
and self
._applied
_styles
:
4095 self
._applied
_styles
.pop()
4097 def data(self
, data
):
4101 return self
._out
.strip()
4103 # Fix UTF-8 encoded file wrongly marked as UTF-16. See https://github.com/yt-dlp/yt-dlp/issues/6543#issuecomment-1477169870
4104 # This will not trigger false positives since only UTF-8 text is being replaced
4105 dfxp_data
= dfxp_data
.replace(b
'encoding=\'UTF-16\'', b
'encoding=\'UTF-8\'')
4107 def parse_node(node
):
4108 target
= TTMLPElementParser()
4109 parser
= xml
.etree
.ElementTree
.XMLParser(target
=target
)
4110 parser
.feed(xml
.etree
.ElementTree
.tostring(node
))
4111 return parser
.close()
4113 for k
, v
in LEGACY_NAMESPACES
:
4115 dfxp_data
= dfxp_data
.replace(ns
, k
)
4117 dfxp
= compat_etree_fromstring(dfxp_data
)
4119 paras
= dfxp
.findall(_x('.//ttml:p')) or dfxp
.findall('.//p')
4122 raise ValueError('Invalid dfxp/TTML subtitle')
4126 for style
in dfxp
.findall(_x('.//ttml:style')):
4127 style_id
= style
.get('id') or style
.get(_x('xml:id'))
4130 parent_style_id
= style
.get('style')
4132 if parent_style_id
not in styles
:
4135 styles
[style_id
] = styles
[parent_style_id
].copy()
4136 for prop
in SUPPORTED_STYLING
:
4137 prop_val
= style
.get(_x('tts:' + prop
))
4139 styles
.setdefault(style_id
, {})[prop
] = prop_val
4145 for p
in ('body', 'div'):
4146 ele
= xpath_element(dfxp
, [_x('.//ttml:' + p
), './/' + p
])
4149 style
= styles
.get(ele
.get('style'))
4152 default_style
.update(style
)
4154 for para
, index
in zip(paras
, itertools
.count(1)):
4155 begin_time
= parse_dfxp_time_expr(para
.attrib
.get('begin'))
4156 end_time
= parse_dfxp_time_expr(para
.attrib
.get('end'))
4157 dur
= parse_dfxp_time_expr(para
.attrib
.get('dur'))
4158 if begin_time
is None:
4163 end_time
= begin_time
+ dur
4164 out
.append('%d\n%s --> %s\n%s\n\n' % (
4166 srt_subtitles_timecode(begin_time
),
4167 srt_subtitles_timecode(end_time
),
4173 def cli_option(params
, command_option
, param
, separator
=None):
4174 param
= params
.get(param
)
4175 return ([] if param
is None
4176 else [command_option
, str(param
)] if separator
is None
4177 else [f
'{command_option}{separator}{param}'])
4180 def cli_bool_option(params
, command_option
, param
, true_value
='true', false_value
='false', separator
=None):
4181 param
= params
.get(param
)
4182 assert param
in (True, False, None)
4183 return cli_option({True: true_value, False: false_value}
, command_option
, param
, separator
)
4186 def cli_valueless_option(params
, command_option
, param
, expected_value
=True):
4187 return [command_option
] if params
.get(param
) == expected_value
else []
4190 def cli_configuration_args(argdict
, keys
, default
=[], use_compat
=True):
4191 if isinstance(argdict
, (list, tuple)): # for backward compatibility
4198 assert isinstance(argdict
, dict)
4200 assert isinstance(keys
, (list, tuple))
4201 for key_list
in keys
:
4202 arg_list
= list(filter(
4203 lambda x
: x
is not None,
4204 [argdict
.get(key
.lower()) for key
in variadic(key_list
)]))
4206 return [arg
for args
in arg_list
for arg
in args
]
4210 def _configuration_args(main_key
, argdict
, exe
, keys
=None, default
=[], use_compat
=True):
4211 main_key
, exe
= main_key
.lower(), exe
.lower()
4212 root_key
= exe
if main_key
== exe
else f
'{main_key}+{exe}'
4213 keys
= [f
'{root_key}{k}' for k
in (keys
or [''])]
4214 if root_key
in keys
:
4216 keys
.append((main_key
, exe
))
4217 keys
.append('default')
4220 return cli_configuration_args(argdict
, keys
, default
, use_compat
)
4224 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4283 'iw': 'heb', # Replaced by he in 1989 revision
4293 'in': 'ind', # Replaced by id in 1989 revision
4408 'ji': 'yid', # Replaced by yi in 1989 revision
4416 def short2long(cls
, code
):
4417 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4418 return cls
._lang
_map
.get(code
[:2])
4421 def long2short(cls
, code
):
4422 """Convert language code from ISO 639-2/T to ISO 639-1"""
4423 for short_name
, long_name
in cls
._lang
_map
.items():
4424 if long_name
== code
:
4429 # From http://data.okfn.org/data/core/country-list
4431 'AF': 'Afghanistan',
4432 'AX': 'Åland Islands',
4435 'AS': 'American Samoa',
4440 'AG': 'Antigua and Barbuda',
4457 'BO': 'Bolivia, Plurinational State of',
4458 'BQ': 'Bonaire, Sint Eustatius and Saba',
4459 'BA': 'Bosnia and Herzegovina',
4461 'BV': 'Bouvet Island',
4463 'IO': 'British Indian Ocean Territory',
4464 'BN': 'Brunei Darussalam',
4466 'BF': 'Burkina Faso',
4472 'KY': 'Cayman Islands',
4473 'CF': 'Central African Republic',
4477 'CX': 'Christmas Island',
4478 'CC': 'Cocos (Keeling) Islands',
4482 'CD': 'Congo, the Democratic Republic of the',
4483 'CK': 'Cook Islands',
4485 'CI': 'Côte d\'Ivoire',
4490 'CZ': 'Czech Republic',
4494 'DO': 'Dominican Republic',
4497 'SV': 'El Salvador',
4498 'GQ': 'Equatorial Guinea',
4502 'FK': 'Falkland Islands (Malvinas)',
4503 'FO': 'Faroe Islands',
4507 'GF': 'French Guiana',
4508 'PF': 'French Polynesia',
4509 'TF': 'French Southern Territories',
4524 'GW': 'Guinea-Bissau',
4527 'HM': 'Heard Island and McDonald Islands',
4528 'VA': 'Holy See (Vatican City State)',
4535 'IR': 'Iran, Islamic Republic of',
4538 'IM': 'Isle of Man',
4548 'KP': 'Korea, Democratic People\'s Republic of',
4549 'KR': 'Korea, Republic of',
4552 'LA': 'Lao People\'s Democratic Republic',
4558 'LI': 'Liechtenstein',
4562 'MK': 'Macedonia, the Former Yugoslav Republic of',
4569 'MH': 'Marshall Islands',
4575 'FM': 'Micronesia, Federated States of',
4576 'MD': 'Moldova, Republic of',
4587 'NL': 'Netherlands',
4588 'NC': 'New Caledonia',
4589 'NZ': 'New Zealand',
4594 'NF': 'Norfolk Island',
4595 'MP': 'Northern Mariana Islands',
4600 'PS': 'Palestine, State of',
4602 'PG': 'Papua New Guinea',
4605 'PH': 'Philippines',
4609 'PR': 'Puerto Rico',
4613 'RU': 'Russian Federation',
4615 'BL': 'Saint Barthélemy',
4616 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4617 'KN': 'Saint Kitts and Nevis',
4618 'LC': 'Saint Lucia',
4619 'MF': 'Saint Martin (French part)',
4620 'PM': 'Saint Pierre and Miquelon',
4621 'VC': 'Saint Vincent and the Grenadines',
4624 'ST': 'Sao Tome and Principe',
4625 'SA': 'Saudi Arabia',
4629 'SL': 'Sierra Leone',
4631 'SX': 'Sint Maarten (Dutch part)',
4634 'SB': 'Solomon Islands',
4636 'ZA': 'South Africa',
4637 'GS': 'South Georgia and the South Sandwich Islands',
4638 'SS': 'South Sudan',
4643 'SJ': 'Svalbard and Jan Mayen',
4646 'CH': 'Switzerland',
4647 'SY': 'Syrian Arab Republic',
4648 'TW': 'Taiwan, Province of China',
4650 'TZ': 'Tanzania, United Republic of',
4652 'TL': 'Timor-Leste',
4656 'TT': 'Trinidad and Tobago',
4659 'TM': 'Turkmenistan',
4660 'TC': 'Turks and Caicos Islands',
4664 'AE': 'United Arab Emirates',
4665 'GB': 'United Kingdom',
4666 'US': 'United States',
4667 'UM': 'United States Minor Outlying Islands',
4671 'VE': 'Venezuela, Bolivarian Republic of',
4673 'VG': 'Virgin Islands, British',
4674 'VI': 'Virgin Islands, U.S.',
4675 'WF': 'Wallis and Futuna',
4676 'EH': 'Western Sahara',
4680 # Not ISO 3166 codes, but used for IP blocks
4681 'AP': 'Asia/Pacific Region',
4686 def short2full(cls
, code
):
4687 """Convert an ISO 3166-2 country code to the corresponding full name"""
4688 return cls
._country
_map
.get(code
.upper())
4692 # Major IPv4 address blocks per country
4694 'AD': '46.172.224.0/19',
4695 'AE': '94.200.0.0/13',
4696 'AF': '149.54.0.0/17',
4697 'AG': '209.59.64.0/18',
4698 'AI': '204.14.248.0/21',
4699 'AL': '46.99.0.0/16',
4700 'AM': '46.70.0.0/15',
4701 'AO': '105.168.0.0/13',
4702 'AP': '182.50.184.0/21',
4703 'AQ': '23.154.160.0/24',
4704 'AR': '181.0.0.0/12',
4705 'AS': '202.70.112.0/20',
4706 'AT': '77.116.0.0/14',
4707 'AU': '1.128.0.0/11',
4708 'AW': '181.41.0.0/18',
4709 'AX': '185.217.4.0/22',
4710 'AZ': '5.197.0.0/16',
4711 'BA': '31.176.128.0/17',
4712 'BB': '65.48.128.0/17',
4713 'BD': '114.130.0.0/16',
4715 'BF': '102.178.0.0/15',
4716 'BG': '95.42.0.0/15',
4717 'BH': '37.131.0.0/17',
4718 'BI': '154.117.192.0/18',
4719 'BJ': '137.255.0.0/16',
4720 'BL': '185.212.72.0/23',
4721 'BM': '196.12.64.0/18',
4722 'BN': '156.31.0.0/16',
4723 'BO': '161.56.0.0/16',
4724 'BQ': '161.0.80.0/20',
4725 'BR': '191.128.0.0/12',
4726 'BS': '24.51.64.0/18',
4727 'BT': '119.2.96.0/19',
4728 'BW': '168.167.0.0/16',
4729 'BY': '178.120.0.0/13',
4730 'BZ': '179.42.192.0/18',
4731 'CA': '99.224.0.0/11',
4732 'CD': '41.243.0.0/16',
4733 'CF': '197.242.176.0/21',
4734 'CG': '160.113.0.0/16',
4735 'CH': '85.0.0.0/13',
4736 'CI': '102.136.0.0/14',
4737 'CK': '202.65.32.0/19',
4738 'CL': '152.172.0.0/14',
4739 'CM': '102.244.0.0/14',
4740 'CN': '36.128.0.0/10',
4741 'CO': '181.240.0.0/12',
4742 'CR': '201.192.0.0/12',
4743 'CU': '152.206.0.0/15',
4744 'CV': '165.90.96.0/19',
4745 'CW': '190.88.128.0/17',
4746 'CY': '31.153.0.0/16',
4747 'CZ': '88.100.0.0/14',
4749 'DJ': '197.241.0.0/17',
4750 'DK': '87.48.0.0/12',
4751 'DM': '192.243.48.0/20',
4752 'DO': '152.166.0.0/15',
4753 'DZ': '41.96.0.0/12',
4754 'EC': '186.68.0.0/15',
4755 'EE': '90.190.0.0/15',
4756 'EG': '156.160.0.0/11',
4757 'ER': '196.200.96.0/20',
4758 'ES': '88.0.0.0/11',
4759 'ET': '196.188.0.0/14',
4760 'EU': '2.16.0.0/13',
4761 'FI': '91.152.0.0/13',
4762 'FJ': '144.120.0.0/16',
4763 'FK': '80.73.208.0/21',
4764 'FM': '119.252.112.0/20',
4765 'FO': '88.85.32.0/19',
4767 'GA': '41.158.0.0/15',
4769 'GD': '74.122.88.0/21',
4770 'GE': '31.146.0.0/16',
4771 'GF': '161.22.64.0/18',
4772 'GG': '62.68.160.0/19',
4773 'GH': '154.160.0.0/12',
4774 'GI': '95.164.0.0/16',
4775 'GL': '88.83.0.0/19',
4776 'GM': '160.182.0.0/15',
4777 'GN': '197.149.192.0/18',
4778 'GP': '104.250.0.0/19',
4779 'GQ': '105.235.224.0/20',
4780 'GR': '94.64.0.0/13',
4781 'GT': '168.234.0.0/16',
4782 'GU': '168.123.0.0/16',
4783 'GW': '197.214.80.0/20',
4784 'GY': '181.41.64.0/18',
4785 'HK': '113.252.0.0/14',
4786 'HN': '181.210.0.0/16',
4787 'HR': '93.136.0.0/13',
4788 'HT': '148.102.128.0/17',
4789 'HU': '84.0.0.0/14',
4790 'ID': '39.192.0.0/10',
4791 'IE': '87.32.0.0/12',
4792 'IL': '79.176.0.0/13',
4793 'IM': '5.62.80.0/20',
4794 'IN': '117.192.0.0/10',
4795 'IO': '203.83.48.0/21',
4796 'IQ': '37.236.0.0/14',
4797 'IR': '2.176.0.0/12',
4798 'IS': '82.221.0.0/16',
4799 'IT': '79.0.0.0/10',
4800 'JE': '87.244.64.0/18',
4801 'JM': '72.27.0.0/17',
4802 'JO': '176.29.0.0/16',
4803 'JP': '133.0.0.0/8',
4804 'KE': '105.48.0.0/12',
4805 'KG': '158.181.128.0/17',
4806 'KH': '36.37.128.0/17',
4807 'KI': '103.25.140.0/22',
4808 'KM': '197.255.224.0/20',
4809 'KN': '198.167.192.0/19',
4810 'KP': '175.45.176.0/22',
4811 'KR': '175.192.0.0/10',
4812 'KW': '37.36.0.0/14',
4813 'KY': '64.96.0.0/15',
4814 'KZ': '2.72.0.0/13',
4815 'LA': '115.84.64.0/18',
4816 'LB': '178.135.0.0/16',
4817 'LC': '24.92.144.0/20',
4818 'LI': '82.117.0.0/19',
4819 'LK': '112.134.0.0/15',
4820 'LR': '102.183.0.0/16',
4821 'LS': '129.232.0.0/17',
4822 'LT': '78.56.0.0/13',
4823 'LU': '188.42.0.0/16',
4824 'LV': '46.109.0.0/16',
4825 'LY': '41.252.0.0/14',
4826 'MA': '105.128.0.0/11',
4827 'MC': '88.209.64.0/18',
4828 'MD': '37.246.0.0/16',
4829 'ME': '178.175.0.0/17',
4830 'MF': '74.112.232.0/21',
4831 'MG': '154.126.0.0/17',
4832 'MH': '117.103.88.0/21',
4833 'MK': '77.28.0.0/15',
4834 'ML': '154.118.128.0/18',
4835 'MM': '37.111.0.0/17',
4836 'MN': '49.0.128.0/17',
4837 'MO': '60.246.0.0/16',
4838 'MP': '202.88.64.0/20',
4839 'MQ': '109.203.224.0/19',
4840 'MR': '41.188.64.0/18',
4841 'MS': '208.90.112.0/22',
4842 'MT': '46.11.0.0/16',
4843 'MU': '105.16.0.0/12',
4844 'MV': '27.114.128.0/18',
4845 'MW': '102.70.0.0/15',
4846 'MX': '187.192.0.0/11',
4847 'MY': '175.136.0.0/13',
4848 'MZ': '197.218.0.0/15',
4849 'NA': '41.182.0.0/16',
4850 'NC': '101.101.0.0/18',
4851 'NE': '197.214.0.0/18',
4852 'NF': '203.17.240.0/22',
4853 'NG': '105.112.0.0/12',
4854 'NI': '186.76.0.0/15',
4855 'NL': '145.96.0.0/11',
4856 'NO': '84.208.0.0/13',
4857 'NP': '36.252.0.0/15',
4858 'NR': '203.98.224.0/19',
4859 'NU': '49.156.48.0/22',
4860 'NZ': '49.224.0.0/14',
4861 'OM': '5.36.0.0/15',
4862 'PA': '186.72.0.0/15',
4863 'PE': '186.160.0.0/14',
4864 'PF': '123.50.64.0/18',
4865 'PG': '124.240.192.0/19',
4866 'PH': '49.144.0.0/13',
4867 'PK': '39.32.0.0/11',
4868 'PL': '83.0.0.0/11',
4869 'PM': '70.36.0.0/20',
4870 'PR': '66.50.0.0/16',
4871 'PS': '188.161.0.0/16',
4872 'PT': '85.240.0.0/13',
4873 'PW': '202.124.224.0/20',
4874 'PY': '181.120.0.0/14',
4875 'QA': '37.210.0.0/15',
4876 'RE': '102.35.0.0/16',
4877 'RO': '79.112.0.0/13',
4878 'RS': '93.86.0.0/15',
4879 'RU': '5.136.0.0/13',
4880 'RW': '41.186.0.0/16',
4881 'SA': '188.48.0.0/13',
4882 'SB': '202.1.160.0/19',
4883 'SC': '154.192.0.0/11',
4884 'SD': '102.120.0.0/13',
4885 'SE': '78.64.0.0/12',
4886 'SG': '8.128.0.0/10',
4887 'SI': '188.196.0.0/14',
4888 'SK': '78.98.0.0/15',
4889 'SL': '102.143.0.0/17',
4890 'SM': '89.186.32.0/19',
4891 'SN': '41.82.0.0/15',
4892 'SO': '154.115.192.0/18',
4893 'SR': '186.179.128.0/17',
4894 'SS': '105.235.208.0/21',
4895 'ST': '197.159.160.0/19',
4896 'SV': '168.243.0.0/16',
4897 'SX': '190.102.0.0/20',
4899 'SZ': '41.84.224.0/19',
4900 'TC': '65.255.48.0/20',
4901 'TD': '154.68.128.0/19',
4902 'TG': '196.168.0.0/14',
4903 'TH': '171.96.0.0/13',
4904 'TJ': '85.9.128.0/18',
4905 'TK': '27.96.24.0/21',
4906 'TL': '180.189.160.0/20',
4907 'TM': '95.85.96.0/19',
4908 'TN': '197.0.0.0/11',
4909 'TO': '175.176.144.0/21',
4910 'TR': '78.160.0.0/11',
4911 'TT': '186.44.0.0/15',
4912 'TV': '202.2.96.0/19',
4913 'TW': '120.96.0.0/11',
4914 'TZ': '156.156.0.0/14',
4915 'UA': '37.52.0.0/14',
4916 'UG': '102.80.0.0/13',
4918 'UY': '167.56.0.0/13',
4919 'UZ': '84.54.64.0/18',
4920 'VA': '212.77.0.0/19',
4921 'VC': '207.191.240.0/21',
4922 'VE': '186.88.0.0/13',
4923 'VG': '66.81.192.0/20',
4924 'VI': '146.226.0.0/16',
4925 'VN': '14.160.0.0/11',
4926 'VU': '202.80.32.0/20',
4927 'WF': '117.20.32.0/21',
4928 'WS': '202.4.32.0/19',
4929 'YE': '134.35.0.0/16',
4930 'YT': '41.242.116.0/22',
4931 'ZA': '41.0.0.0/11',
4932 'ZM': '102.144.0.0/13',
4933 'ZW': '102.177.192.0/18',
4937 def random_ipv4(cls
, code_or_block
):
4938 if len(code_or_block
) == 2:
4939 block
= cls
._country
_ip
_map
.get(code_or_block
.upper())
4943 block
= code_or_block
4944 addr
, preflen
= block
.split('/')
4945 addr_min
= struct
.unpack('!L', socket
.inet_aton(addr
))[0]
4946 addr_max
= addr_min |
(0xffffffff >> int(preflen
))
4947 return str(socket
.inet_ntoa(
4948 struct
.pack('!L', random
.randint(addr_min
, addr_max
))))
4951 class PerRequestProxyHandler(urllib
.request
.ProxyHandler
):
4952 def __init__(self
, proxies
=None):
4953 # Set default handlers
4954 for type in ('http', 'https'):
4955 setattr(self
, '%s_open' % type,
4956 lambda r
, proxy
='__noproxy__', type=type, meth
=self
.proxy_open
:
4957 meth(r
, proxy
, type))
4958 urllib
.request
.ProxyHandler
.__init
__(self
, proxies
)
4960 def proxy_open(self
, req
, proxy
, type):
4961 req_proxy
= req
.headers
.get('Ytdl-request-proxy')
4962 if req_proxy
is not None:
4964 del req
.headers
['Ytdl-request-proxy']
4966 if proxy
== '__noproxy__':
4967 return None # No Proxy
4968 if urllib
.parse
.urlparse(proxy
).scheme
.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4969 req
.add_header('Ytdl-socks-proxy', proxy
)
4970 # yt-dlp's http/https handlers do wrapping the socket with socks
4972 return urllib
.request
.ProxyHandler
.proxy_open(
4973 self
, req
, proxy
, type)
4976 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4977 # released into Public Domain
4978 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4980 def long_to_bytes(n
, blocksize
=0):
4981 """long_to_bytes(n:long, blocksize:int) : string
4982 Convert a long integer to a byte string.
4984 If optional blocksize is given and greater than zero, pad the front of the
4985 byte string with binary zeros so that the length is a multiple of
4988 # after much testing, this algorithm was deemed to be the fastest
4992 s
= struct
.pack('>I', n
& 0xffffffff) + s
4994 # strip off leading zeros
4995 for i
in range(len(s
)):
4996 if s
[i
] != b
'\000'[0]:
4999 # only happens when n == 0
5003 # add back some pad bytes. this could be done more efficiently w.r.t. the
5004 # de-padding being done above, but sigh...
5005 if blocksize
> 0 and len(s
) % blocksize
:
5006 s
= (blocksize
- len(s
) % blocksize
) * b
'\000' + s
5010 def bytes_to_long(s
):
5011 """bytes_to_long(string) : long
5012 Convert a byte string to a long integer.
5014 This is (essentially) the inverse of long_to_bytes().
5019 extra
= (4 - length
% 4)
5020 s
= b
'\000' * extra
+ s
5021 length
= length
+ extra
5022 for i
in range(0, length
, 4):
5023 acc
= (acc
<< 32) + struct
.unpack('>I', s
[i
:i
+ 4])[0]
5027 def ohdave_rsa_encrypt(data
, exponent
, modulus
):
5029 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
5032 data: data to encrypt, bytes-like object
5033 exponent, modulus: parameter e and N of RSA algorithm, both integer
5034 Output: hex string of encrypted data
5036 Limitation: supports one block encryption only
5039 payload
= int(binascii
.hexlify(data
[::-1]), 16)
5040 encrypted
= pow(payload
, exponent
, modulus
)
5041 return '%x' % encrypted
5044 def pkcs1pad(data
, length
):
5046 Padding input data with PKCS#1 scheme
5048 @param {int[]} data input data
5049 @param {int} length target length
5050 @returns {int[]} padded data
5052 if len(data
) > length
- 11:
5053 raise ValueError('Input data too long for PKCS#1 padding')
5055 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
5056 return [0, 2] + pseudo_random
+ [0] + data
5059 def _base_n_table(n
, table
):
5060 if not table
and not n
:
5061 raise ValueError('Either table or n must be specified')
5062 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
5064 if n
and n
!= len(table
):
5065 raise ValueError(f
'base {n} exceeds table length {len(table)}')
5069 def encode_base_n(num
, n
=None, table
=None):
5070 """Convert given int to a base-n string"""
5071 table
= _base_n_table(n
, table
)
5075 result
, base
= '', len(table
)
5077 result
= table
[num
% base
] + result
5082 def decode_base_n(string
, n
=None, table
=None):
5083 """Convert given base-n string to int"""
5084 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
5085 result
, base
= 0, len(table
)
5087 result
= result
* base
+ table
[char
]
5091 def decode_base(value
, digits
):
5092 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
5093 f
'in a future version. Use {__name__}.decode_base_n instead')
5094 return decode_base_n(value
, table
=digits
)
5097 def decode_packed_codes(code
):
5098 mobj
= re
.search(PACKED_CODES_RE
, code
)
5099 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
5102 symbols
= symbols
.split('|')
5107 base_n_count
= encode_base_n(count
, base
)
5108 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
5111 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
5115 def caesar(s
, alphabet
, shift
):
5120 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
5125 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
5128 def parse_m3u8_attributes(attrib
):
5130 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5131 if val
.startswith('"'):
5137 def urshift(val
, n
):
5138 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5141 # Based on png2str() written by @gdkchan and improved by @yokrysty
5142 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5143 def decode_png(png_data
):
5144 # Reference: https://www.w3.org/TR/PNG/
5145 header
= png_data
[8:]
5147 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5148 raise OSError('Not a valid PNG file.')
5150 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5151 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5156 length
= unpack_integer(header
[:4])
5159 chunk_type
= header
[:4]
5162 chunk_data
= header
[:length
]
5163 header
= header
[length
:]
5165 header
= header
[4:] # Skip CRC
5173 ihdr
= chunks
[0]['data']
5175 width
= unpack_integer(ihdr
[:4])
5176 height
= unpack_integer(ihdr
[4:8])
5180 for chunk
in chunks
:
5181 if chunk
['type'] == b
'IDAT':
5182 idat
+= chunk
['data']
5185 raise OSError('Unable to read PNG data.')
5187 decompressed_data
= bytearray(zlib
.decompress(idat
))
5192 def _get_pixel(idx
):
5197 for y
in range(height
):
5198 basePos
= y
* (1 + stride
)
5199 filter_type
= decompressed_data
[basePos
]
5203 pixels
.append(current_row
)
5205 for x
in range(stride
):
5206 color
= decompressed_data
[1 + basePos
+ x
]
5207 basex
= y
* stride
+ x
5212 left
= _get_pixel(basex
- 3)
5214 up
= _get_pixel(basex
- stride
)
5216 if filter_type
== 1: # Sub
5217 color
= (color
+ left
) & 0xff
5218 elif filter_type
== 2: # Up
5219 color
= (color
+ up
) & 0xff
5220 elif filter_type
== 3: # Average
5221 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5222 elif filter_type
== 4: # Paeth
5228 c
= _get_pixel(basex
- stride
- 3)
5236 if pa
<= pb
and pa
<= pc
:
5237 color
= (color
+ a
) & 0xff
5239 color
= (color
+ b
) & 0xff
5241 color
= (color
+ c
) & 0xff
5243 current_row
.append(color
)
5245 return width
, height
, pixels
5248 def write_xattr(path
, key
, value
):
5249 # Windows: Write xattrs to NTFS Alternate Data Streams:
5250 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5251 if compat_os_name
== 'nt':
5252 assert ':' not in key
5253 assert os
.path
.exists(path
)
5256 with open(f
'{path}:{key}', 'wb') as f
:
5258 except OSError as e
:
5259 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5262 # UNIX Method 1. Use xattrs/pyxattrs modules
5265 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5266 # Unicode arguments are not supported in pyxattr until version 0.5.0
5267 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5268 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5269 setxattr
= xattr
.set
5271 setxattr
= xattr
.setxattr
5275 setxattr(path
, key
, value
)
5276 except OSError as e
:
5277 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5280 # UNIX Method 2. Use setfattr/xattr executables
5281 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5282 else 'xattr' if check_executable('xattr', ['-h']) else None)
5284 raise XAttrUnavailableError(
5285 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5286 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5288 value
= value
.decode()
5290 _
, stderr
, returncode
= Popen
.run(
5291 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5292 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5293 except OSError as e
:
5294 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5296 raise XAttrMetadataError(returncode
, stderr
)
5299 def random_birthday(year_field
, month_field
, day_field
):
5300 start_date
= datetime
.date(1950, 1, 1)
5301 end_date
= datetime
.date(1995, 12, 31)
5302 offset
= random
.randint(0, (end_date
- start_date
).days
)
5303 random_date
= start_date
+ datetime
.timedelta(offset
)
5305 year_field
: str(random_date
.year
),
5306 month_field
: str(random_date
.month
),
5307 day_field
: str(random_date
.day
),
5311 def find_available_port(interface
=''):
5313 with socket
.socket() as sock
:
5314 sock
.bind((interface
, 0))
5315 return sock
.getsockname()[1]
5320 # Templates for internet shortcut files, which are plain text files.
5321 DOT_URL_LINK_TEMPLATE
= '''\
5326 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5327 <?xml version="1.0" encoding="UTF-8"?>
5328 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5329 <plist version="1.0">
5332 \t<string>%(url)s</string>
5337 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5347 'url': DOT_URL_LINK_TEMPLATE
,
5348 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5349 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5353 def iri_to_uri(iri
):
5355 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5357 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.
5360 iri_parts
= urllib
.parse
.urlparse(iri
)
5362 if '[' in iri_parts
.netloc
:
5363 raise ValueError('IPv6 URIs are not, yet, supported.')
5364 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5366 # 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.
5369 if iri_parts
.username
:
5370 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5371 if iri_parts
.password
is not None:
5372 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5375 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5376 # The 'idna' encoding produces ASCII text.
5377 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5378 net_location
+= ':' + str(iri_parts
.port
)
5380 return urllib
.parse
.urlunparse(
5384 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5386 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5387 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5389 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5390 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5392 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5394 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5397 def to_high_limit_path(path
):
5398 if sys
.platform
in ['win32', 'cygwin']:
5399 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5400 return '\\\\?\\' + os
.path
.abspath(path
)
5405 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5406 val
= traverse_obj(obj
, *variadic(field
))
5407 if not val
if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5409 return template
% func(val
)
5412 def clean_podcast_url(url
):
5413 return re
.sub(r
'''(?x)
5417 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5420 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5423 cn\.co| # https://podcorn.com/analytics-prefix/
5424 st\.fm # https://podsights.com/docs/
5429 _HEX_TABLE
= '0123456789abcdef'
5432 def random_uuidv4():
5433 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5436 def make_dir(path
, to_screen
=None):
5438 dn
= os
.path
.dirname(path
)
5440 os
.makedirs(dn
, exist_ok
=True)
5442 except OSError as err
:
5443 if callable(to_screen
) is not None:
5444 to_screen('unable to create directory ' + error_to_compat_str(err
))
5448 def get_executable_path():
5449 from .update
import _get_variant_and_executable_path
5451 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5454 def get_user_config_dirs(package_name
):
5455 # .config (e.g. ~/.config/package_name)
5456 xdg_config_home
= os
.getenv('XDG_CONFIG_HOME') or compat_expanduser('~/.config')
5457 yield os
.path
.join(xdg_config_home
, package_name
)
5459 # appdata (%APPDATA%/package_name)
5460 appdata_dir
= os
.getenv('appdata')
5462 yield os
.path
.join(appdata_dir
, package_name
)
5464 # home (~/.package_name)
5465 yield os
.path
.join(compat_expanduser('~'), f
'.{package_name}')
5468 def get_system_config_dirs(package_name
):
5470 yield os
.path
.join('/etc', package_name
)
5474 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5475 casesense
=True, is_user_input
=False, traverse_string
=False):
5477 Safely traverse nested `dict`s and `Iterable`s
5479 >>> obj = [{}, {"key": "value"}]
5480 >>> traverse_obj(obj, (1, "key"))
5483 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5484 The next path will also be tested if the path branched but no results could be found.
5485 Supported values for traversal are `Mapping`, `Iterable` and `re.Match`.
5486 Unhelpful values (`{}`, `None`) are treated as the absence of a value and discarded.
5488 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5490 The keys in the path can be one of:
5491 - `None`: Return the current object.
5492 - `set`: Requires the only item in the set to be a type or function,
5493 like `{type}`/`{func}`. If a `type`, returns only values
5494 of this type. If a function, returns `func(obj)`.
5495 - `str`/`int`: Return `obj[key]`. For `re.Match`, return `obj.group(key)`.
5496 - `slice`: Branch out and return all values in `obj[key]`.
5497 - `Ellipsis`: Branch out and return a list of all values.
5498 - `tuple`/`list`: Branch out and return a list of all matching values.
5499 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5500 - `function`: Branch out and return values filtered by the function.
5501 Read as: `[value for key, value in obj if function(key, value)]`.
5502 For `Iterable`s, `key` is the index of the value.
5503 For `re.Match`es, `key` is the group number (0 = full match)
5504 as well as additionally any group names, if given.
5505 - `dict` Transform the current object and return a matching dict.
5506 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5508 `tuple`, `list`, and `dict` all support nested paths and branches.
5510 @params paths Paths which to traverse by.
5511 @param default Value to return if the paths do not match.
5512 If the last key in the path is a `dict`, it will apply to each value inside
5513 the dict instead, depth first. Try to avoid if using nested `dict` keys.
5514 @param expected_type If a `type`, only accept final values of this type.
5515 If any other callable, try to call the function on each result.
5516 If the last key in the path is a `dict`, it will apply to each value inside
5517 the dict instead, recursively. This does respect branching paths.
5518 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5519 @param casesense If `False`, consider string dictionary keys as case insensitive.
5521 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5523 @param is_user_input Whether the keys are generated from user input.
5524 If `True` strings get converted to `int`/`slice` if needed.
5525 @param traverse_string Whether to traverse into objects as strings.
5526 If `True`, any non-compatible object will first be
5527 converted into a string and then traversed into.
5528 The return value of that path will be a string instead,
5529 not respecting any further branching.
5532 @returns The result of the object traversal.
5533 If successful, `get_all=True`, and the path branches at least once,
5534 then a list of results is returned instead.
5535 If no `default` is given and the last path branches, a `list` of results
5536 is always returned. If a path ends on a `dict` that result will always be a `dict`.
5538 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5540 if isinstance(expected_type
, type):
5541 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5543 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5545 def apply_key(key
, obj
, is_last
):
5549 if obj
is None and traverse_string
:
5550 if key
is ... or callable(key
) or isinstance(key
, slice):
5557 elif isinstance(key
, set):
5558 assert len(key
) == 1, 'Set should only be used to wrap a single item'
5559 item
= next(iter(key
))
5560 if isinstance(item
, type):
5561 if isinstance(obj
, item
):
5564 result
= try_call(item
, args
=(obj
,))
5566 elif isinstance(key
, (list, tuple)):
5568 result
= itertools
.chain
.from_iterable(
5569 apply_path(obj
, branch
, is_last
)[0] for branch
in key
)
5573 if isinstance(obj
, collections
.abc
.Mapping
):
5574 result
= obj
.values()
5575 elif is_iterable_like(obj
):
5577 elif isinstance(obj
, re
.Match
):
5578 result
= obj
.groups()
5579 elif traverse_string
:
5587 if isinstance(obj
, collections
.abc
.Mapping
):
5588 iter_obj
= obj
.items()
5589 elif is_iterable_like(obj
):
5590 iter_obj
= enumerate(obj
)
5591 elif isinstance(obj
, re
.Match
):
5592 iter_obj
= itertools
.chain(
5593 enumerate((obj
.group(), *obj
.groups())),
5594 obj
.groupdict().items())
5595 elif traverse_string
:
5597 iter_obj
= enumerate(str(obj
))
5601 result
= (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5602 if not branching
: # string traversal
5603 result
= ''.join(result
)
5605 elif isinstance(key
, dict):
5606 iter_obj
= ((k
, _traverse_obj(obj
, v
, False, is_last
)) for k
, v
in key
.items())
5608 k
: v
if v
is not None else default
for k
, v
in iter_obj
5609 if v
is not None or default
is not NO_DEFAULT
5612 elif isinstance(obj
, collections
.abc
.Mapping
):
5613 result
= (try_call(obj
.get
, args
=(key
,)) if casesense
or try_call(obj
.__contains
__, args
=(key
,)) else
5614 next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5616 elif isinstance(obj
, re
.Match
):
5617 if isinstance(key
, int) or casesense
:
5618 with contextlib
.suppress(IndexError):
5619 result
= obj
.group(key
)
5621 elif isinstance(key
, str):
5622 result
= next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5624 elif isinstance(key
, (int, slice)):
5625 if is_iterable_like(obj
, collections
.abc
.Sequence
):
5626 branching
= isinstance(key
, slice)
5627 with contextlib
.suppress(IndexError):
5629 elif traverse_string
:
5630 with contextlib
.suppress(IndexError):
5631 result
= str(obj
)[key
]
5633 return branching
, result
if branching
else (result
,)
5635 def lazy_last(iterable
):
5636 iterator
= iter(iterable
)
5637 prev
= next(iterator
, NO_DEFAULT
)
5638 if prev
is NO_DEFAULT
:
5641 for item
in iterator
:
5647 def apply_path(start_obj
, path
, test_type
):
5649 has_branched
= False
5652 for last
, key
in lazy_last(variadic(path
, (str, bytes, dict, set))):
5653 if is_user_input
and isinstance(key
, str):
5657 key
= slice(*map(int_or_none
, key
.split(':')))
5658 elif int_or_none(key
) is not None:
5661 if not casesense
and isinstance(key
, str):
5662 key
= key
.casefold()
5664 if __debug__
and callable(key
):
5665 # Verify function signature
5666 inspect
.signature(key
).bind(None, None)
5670 branching
, results
= apply_key(key
, obj
, last
)
5671 has_branched |
= branching
5672 new_objs
.append(results
)
5674 objs
= itertools
.chain
.from_iterable(new_objs
)
5676 if test_type
and not isinstance(key
, (dict, list, tuple)):
5677 objs
= map(type_test
, objs
)
5679 return objs
, has_branched
, isinstance(key
, dict)
5681 def _traverse_obj(obj
, path
, allow_empty
, test_type
):
5682 results
, has_branched
, is_dict
= apply_path(obj
, path
, test_type
)
5683 results
= LazyList(item
for item
in results
if item
not in (None, {}))
5684 if get_all
and has_branched
:
5686 return results
.exhaust()
5688 return [] if default
is NO_DEFAULT
else default
5691 return results
[0] if results
else {} if allow_empty
and is_dict
else None
5693 for index
, path
in enumerate(paths
, 1):
5694 result
= _traverse_obj(obj
, path
, index
== len(paths
), True)
5695 if result
is not None:
5698 return None if default
is NO_DEFAULT
else default
5701 def traverse_dict(dictn
, keys
, casesense
=True):
5702 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5703 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5704 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5707 def get_first(obj
, *paths
, **kwargs
):
5708 return traverse_obj(obj
, *((..., *variadic(keys
)) for keys
in paths
), **kwargs
, get_all
=False)
5711 def time_seconds(**kwargs
):
5713 Returns TZ-aware time in seconds since the epoch (1970-01-01T00:00:00Z)
5715 return time
.time() + datetime
.timedelta(**kwargs
).total_seconds()
5718 # create a JSON Web Signature (jws) with HS256 algorithm
5719 # the resulting format is in JWS Compact Serialization
5720 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5721 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5722 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5728 header_data
.update(headers
)
5729 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5730 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5731 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5732 signature_b64
= base64
.b64encode(h
.digest())
5733 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5737 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5738 def jwt_decode_hs256(jwt
):
5739 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5740 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5741 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5745 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5749 def supports_terminal_sequences(stream
):
5750 if compat_os_name
== 'nt':
5751 if not WINDOWS_VT_MODE
:
5753 elif not os
.getenv('TERM'):
5756 return stream
.isatty()
5757 except BaseException
:
5761 def windows_enable_vt_mode():
5762 """Ref: https://bugs.python.org/issue30075 """
5763 if get_windows_version() < (10, 0, 10586):
5767 import ctypes
.wintypes
5770 ENABLE_VIRTUAL_TERMINAL_PROCESSING
= 0x0004
5772 dll
= ctypes
.WinDLL('kernel32', use_last_error
=False)
5773 handle
= os
.open('CONOUT$', os
.O_RDWR
)
5775 h_out
= ctypes
.wintypes
.HANDLE(msvcrt
.get_osfhandle(handle
))
5776 dw_original_mode
= ctypes
.wintypes
.DWORD()
5777 success
= dll
.GetConsoleMode(h_out
, ctypes
.byref(dw_original_mode
))
5779 raise Exception('GetConsoleMode failed')
5781 success
= dll
.SetConsoleMode(h_out
, ctypes
.wintypes
.DWORD(
5782 dw_original_mode
.value | ENABLE_VIRTUAL_TERMINAL_PROCESSING
))
5784 raise Exception('SetConsoleMode failed')
5788 global WINDOWS_VT_MODE
5789 WINDOWS_VT_MODE
= True
5790 supports_terminal_sequences
.cache_clear()
5793 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5796 def remove_terminal_sequences(string
):
5797 return _terminal_sequences_re
.sub('', string
)
5800 def number_of_digits(number
):
5801 return len('%d' % number
)
5804 def join_nonempty(*values
, delim
='-', from_dict
=None):
5805 if from_dict
is not None:
5806 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5807 return delim
.join(map(str, filter(None, values
)))
5810 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5812 Find the largest format dimensions in terms of video width and, for each thumbnail:
5813 * Modify the URL: Match the width with the provided regex and replace with the former width
5816 This function is useful with video services that scale the provided thumbnails on demand
5818 _keys
= ('width', 'height')
5819 max_dimensions
= max(
5820 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5822 if not max_dimensions
[0]:
5826 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5827 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5828 for thumbnail
in thumbnails
5832 def parse_http_range(range):
5833 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5835 return None, None, None
5836 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5838 return None, None, None
5839 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5842 def read_stdin(what
):
5843 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5844 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5848 def determine_file_encoding(data
):
5850 Detect the text encoding used
5851 @returns (encoding, bytes to skip)
5854 # BOM marks are given priority over declarations
5855 for bom
, enc
in BOMS
:
5856 if data
.startswith(bom
):
5857 return enc
, len(bom
)
5859 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5860 # We ignore the endianness to get a good enough match
5861 data
= data
.replace(b
'\0', b
'')
5862 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5863 return mobj
.group(1).decode() if mobj
else None, 0
5870 __initialized
= False
5872 def __init__(self
, parser
, label
=None):
5873 self
.parser
, self
.label
= parser
, label
5874 self
._loaded
_paths
, self
.configs
= set(), []
5876 def init(self
, args
=None, filename
=None):
5877 assert not self
.__initialized
5878 self
.own_args
, self
.filename
= args
, filename
5879 return self
.load_configs()
5881 def load_configs(self
):
5884 location
= os
.path
.realpath(self
.filename
)
5885 directory
= os
.path
.dirname(location
)
5886 if location
in self
._loaded
_paths
:
5888 self
._loaded
_paths
.add(location
)
5890 self
.__initialized
= True
5891 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5892 self
.parsed_args
= self
.own_args
5893 for location
in opts
.config_locations
or []:
5895 if location
in self
._loaded
_paths
:
5897 self
._loaded
_paths
.add(location
)
5898 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5900 location
= os
.path
.join(directory
, expand_path(location
))
5901 if os
.path
.isdir(location
):
5902 location
= os
.path
.join(location
, 'yt-dlp.conf')
5903 if not os
.path
.exists(location
):
5904 self
.parser
.error(f
'config location {location} does not exist')
5905 self
.append_config(self
.read_file(location
), location
)
5909 label
= join_nonempty(
5910 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5912 return join_nonempty(
5913 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5914 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5918 def read_file(filename
, default
=[]):
5920 optionf
= open(filename
, 'rb')
5922 return default
# silently skip if file is not present
5924 enc
, skip
= determine_file_encoding(optionf
.read(512))
5925 optionf
.seek(skip
, io
.SEEK_SET
)
5927 enc
= None # silently skip read errors
5929 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5930 contents
= optionf
.read().decode(enc
or preferredencoding())
5931 res
= shlex
.split(contents
, comments
=True)
5932 except Exception as err
:
5933 raise ValueError(f
'Unable to parse "{filename}": {err}')
5939 def hide_login_info(opts
):
5940 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5941 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5946 return m
.group('key') + '=PRIVATE'
5950 opts
= list(map(_scrub_eq
, opts
))
5951 for idx
, opt
in enumerate(opts
):
5952 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5953 opts
[idx
+ 1] = 'PRIVATE'
5956 def append_config(self
, *args
, label
=None):
5957 config
= type(self
)(self
.parser
, label
)
5958 config
._loaded
_paths
= self
._loaded
_paths
5959 if config
.init(*args
):
5960 self
.configs
.append(config
)
5964 for config
in reversed(self
.configs
):
5965 yield from config
.all_args
5966 yield from self
.parsed_args
or []
5968 def parse_known_args(self
, **kwargs
):
5969 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5971 def parse_args(self
):
5972 return self
.parser
.parse_args(self
.all_args
)
5975 class WebSocketsWrapper
:
5976 """Wraps websockets module to use in non-async scopes"""
5979 def __init__(self
, url
, headers
=None, connect
=True):
5980 self
.loop
= asyncio
.new_event_loop()
5981 # XXX: "loop" is deprecated
5982 self
.conn
= websockets
.connect(
5983 url
, extra_headers
=headers
, ping_interval
=None,
5984 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5987 atexit
.register(self
.__exit
__, None, None, None)
5989 def __enter__(self
):
5991 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5994 def send(self
, *args
):
5995 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5997 def recv(self
, *args
):
5998 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
6000 def __exit__(self
, type, value
, traceback
):
6002 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
6005 self
._cancel
_all
_tasks
(self
.loop
)
6007 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
6008 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
6010 def run_with_loop(main
, loop
):
6011 if not asyncio
.iscoroutine(main
):
6012 raise ValueError(f
'a coroutine was expected, got {main!r}')
6015 return loop
.run_until_complete(main
)
6017 loop
.run_until_complete(loop
.shutdown_asyncgens())
6018 if hasattr(loop
, 'shutdown_default_executor'):
6019 loop
.run_until_complete(loop
.shutdown_default_executor())
6022 def _cancel_all_tasks(loop
):
6023 to_cancel
= asyncio
.all_tasks(loop
)
6028 for task
in to_cancel
:
6031 # XXX: "loop" is removed in python 3.10+
6032 loop
.run_until_complete(
6033 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
6035 for task
in to_cancel
:
6036 if task
.cancelled():
6038 if task
.exception() is not None:
6039 loop
.call_exception_handler({
6040 'message': 'unhandled exception during asyncio.run() shutdown',
6041 'exception': task
.exception(),
6046 def merge_headers(*dicts
):
6047 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
6048 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
6051 def cached_method(f
):
6052 """Cache a method"""
6053 signature
= inspect
.signature(f
)
6056 def wrapper(self
, *args
, **kwargs
):
6057 bound_args
= signature
.bind(self
, *args
, **kwargs
)
6058 bound_args
.apply_defaults()
6059 key
= tuple(bound_args
.arguments
.values())[1:]
6061 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
6062 if key
not in cache
:
6063 cache
[key
] = f(self
, *args
, **kwargs
)
6068 class classproperty
:
6069 """property access for class methods with optional caching"""
6070 def __new__(cls
, func
=None, *args
, **kwargs
):
6072 return functools
.partial(cls
, *args
, **kwargs
)
6073 return super().__new
__(cls
)
6075 def __init__(self
, func
, *, cache
=False):
6076 functools
.update_wrapper(self
, func
)
6078 self
._cache
= {} if cache
else None
6080 def __get__(self
, _
, cls
):
6081 if self
._cache
is None:
6082 return self
.func(cls
)
6083 elif cls
not in self
._cache
:
6084 self
._cache
[cls
] = self
.func(cls
)
6085 return self
._cache
[cls
]
6088 class function_with_repr
:
6089 def __init__(self
, func
, repr_
=None):
6090 functools
.update_wrapper(self
, func
)
6091 self
.func
, self
.__repr
= func
, repr_
6093 def __call__(self
, *args
, **kwargs
):
6094 return self
.func(*args
, **kwargs
)
6099 return f
'{self.func.__module__}.{self.func.__qualname__}'
6102 class Namespace(types
.SimpleNamespace
):
6103 """Immutable namespace"""
6106 return iter(self
.__dict
__.values())
6110 return self
.__dict
__.items()
6113 MEDIA_EXTENSIONS
= Namespace(
6114 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
6115 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
6116 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
6117 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma', 'weba'),
6118 thumbnails
=('jpg', 'png', 'webp'),
6119 storyboards
=('mhtml', ),
6120 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
6121 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
6123 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
6124 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
6126 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
6131 for retry in RetryManager(...):
6134 except SomeException as err:
6138 attempt
, _error
= 0, None
6140 def __init__(self
, _retries
, _error_callback
, **kwargs
):
6141 self
.retries
= _retries
or 0
6142 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
6144 def _should_retry(self
):
6145 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
6149 if self
._error
is NO_DEFAULT
:
6154 def error(self
, value
):
6158 while self
._should
_retry
():
6159 self
.error
= NO_DEFAULT
6163 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
6166 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
6167 """Utility function for reporting retries"""
6170 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
6175 elif isinstance(e
, ExtractorError
):
6176 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
6177 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
6179 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
6181 info(f
'Sleeping {delay:.2f} seconds ...')
6185 def make_archive_id(ie
, video_id
):
6186 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
6187 return f
'{ie_key.lower()} {video_id}'
6190 def truncate_string(s
, left
, right
=0):
6191 assert left
> 3 and right
>= 0
6192 if s
is None or len(s
) <= left
+ right
:
6194 return f
'{s[:left-3]}...{s[-right:] if right else ""}'
6197 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
6198 assert 'all' in alias_dict
, '"all" alias is required'
6199 requested
= list(start
or [])
6201 discard
= val
.startswith('-')
6205 if val
in alias_dict
:
6206 val
= alias_dict
[val
] if not discard
else [
6207 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
6208 # NB: Do not allow regex in aliases for performance
6209 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
6212 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
6213 else [val
] if val
in alias_dict
['all'] else None)
6215 raise ValueError(val
)
6218 for item
in current
:
6219 while item
in requested
:
6220 requested
.remove(item
)
6222 requested
.extend(current
)
6224 return orderedSet(requested
)
6228 regex
= r
' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
6230 default
= ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
6231 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
6232 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
6233 ytdl_default
= ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
6234 'height', 'width', 'proto', 'vext', 'abr', 'aext',
6235 'fps', 'fs_approx', 'source', 'id')
6238 'vcodec': {'type': 'ordered', 'regex': True,
6239 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
6240 'acodec': {'type': 'ordered', 'regex': True,
6241 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'ac-?4', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
6242 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
6243 'order': ['dv', '(hdr)?12', r
'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
6244 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
6245 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
6246 'vext': {'type': 'ordered', 'field': 'video_ext',
6247 'order': ('mp4', 'mov', 'webm', 'flv', '', 'none'),
6248 'order_free': ('webm', 'mp4', 'mov', 'flv', '', 'none')},
6249 'aext': {'type': 'ordered', 'regex': True, 'field': 'audio_ext',
6250 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'web[am]', '', 'none'),
6251 'order_free': ('ogg', 'opus', 'web[am]', 'mp3', 'm4a', 'aac', '', 'none')},
6252 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}
,
6253 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
6254 'field': ('vcodec', 'acodec'),
6255 'function': lambda it
: int(any(v
!= 'none' for v
in it
))},
6256 'ie_pref': {'priority': True, 'type': 'extractor'}
,
6257 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6258 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6259 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}
,
6260 'quality': {'convert': 'float', 'default': -1}
,
6261 'filesize': {'convert': 'bytes'}
,
6262 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}
,
6263 'id': {'convert': 'string', 'field': 'format_id'}
,
6264 'height': {'convert': 'float_none'}
,
6265 'width': {'convert': 'float_none'}
,
6266 'fps': {'convert': 'float_none'}
,
6267 'channels': {'convert': 'float_none', 'field': 'audio_channels'}
,
6268 'tbr': {'convert': 'float_none'}
,
6269 'vbr': {'convert': 'float_none'}
,
6270 'abr': {'convert': 'float_none'}
,
6271 'asr': {'convert': 'float_none'}
,
6272 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}
,
6274 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}
,
6275 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True}
,
6276 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')}
,
6277 'ext': {'type': 'combined', 'field': ('vext', 'aext')}
,
6278 'res': {'type': 'multiple', 'field': ('height', 'width'),
6279 'function': lambda it
: (lambda l
: min(l
) if l
else 0)(tuple(filter(None, it
)))},
6281 # Actual field names
6282 'format_id': {'type': 'alias', 'field': 'id'}
,
6283 'preference': {'type': 'alias', 'field': 'ie_pref'}
,
6284 'language_preference': {'type': 'alias', 'field': 'lang'}
,
6285 'source_preference': {'type': 'alias', 'field': 'source'}
,
6286 'protocol': {'type': 'alias', 'field': 'proto'}
,
6287 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}
,
6288 'audio_channels': {'type': 'alias', 'field': 'channels'}
,
6291 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6292 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6293 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}
,
6294 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}
,
6295 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}
,
6296 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}
,
6297 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}
,
6298 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}
,
6299 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}
,
6300 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}
,
6301 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}
,
6302 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}
,
6303 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}
,
6304 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}
,
6305 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6306 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6307 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6308 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6309 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6310 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6313 def __init__(self
, ydl
, field_preference
):
6316 self
.evaluate_params(self
.ydl
.params
, field_preference
)
6317 if ydl
.params
.get('verbose'):
6318 self
.print_verbose_info(self
.ydl
.write_debug
)
6320 def _get_field_setting(self
, field
, key
):
6321 if field
not in self
.settings
:
6322 if key
in ('forced', 'priority'):
6324 self
.ydl
.deprecated_feature(f
'Using arbitrary fields ({field}) for format sorting is '
6325 'deprecated and may be removed in a future version')
6326 self
.settings
[field
] = {}
6327 propObj
= self
.settings
[field
]
6328 if key
not in propObj
:
6329 type = propObj
.get('type')
6331 default
= 'preference' if type == 'extractor' else (field
,) if type in ('combined', 'multiple') else field
6332 elif key
== 'convert':
6333 default
= 'order' if type == 'ordered' else 'float_string' if field
else 'ignore'
6335 default
= {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}
.get(key
, None)
6336 propObj
[key
] = default
6339 def _resolve_field_value(self
, field
, value
, convertNone
=False):
6344 value
= value
.lower()
6345 conversion
= self
._get
_field
_setting
(field
, 'convert')
6346 if conversion
== 'ignore':
6348 if conversion
== 'string':
6350 elif conversion
== 'float_none':
6351 return float_or_none(value
)
6352 elif conversion
== 'bytes':
6353 return parse_bytes(value
)
6354 elif conversion
== 'order':
6355 order_list
= (self
._use
_free
_order
and self
._get
_field
_setting
(field
, 'order_free')) or self
._get
_field
_setting
(field
, 'order')
6356 use_regex
= self
._get
_field
_setting
(field
, 'regex')
6357 list_length
= len(order_list
)
6358 empty_pos
= order_list
.index('') if '' in order_list
else list_length
+ 1
6359 if use_regex
and value
is not None:
6360 for i
, regex
in enumerate(order_list
):
6361 if regex
and re
.match(regex
, value
):
6362 return list_length
- i
6363 return list_length
- empty_pos
# not in list
6364 else: # not regex or value = None
6365 return list_length
- (order_list
.index(value
) if value
in order_list
else empty_pos
)
6367 if value
.isnumeric():
6370 self
.settings
[field
]['convert'] = 'string'
6373 def evaluate_params(self
, params
, sort_extractor
):
6374 self
._use
_free
_order
= params
.get('prefer_free_formats', False)
6375 self
._sort
_user
= params
.get('format_sort', [])
6376 self
._sort
_extractor
= sort_extractor
6378 def add_item(field
, reverse
, closest
, limit_text
):
6379 field
= field
.lower()
6380 if field
in self
._order
:
6382 self
._order
.append(field
)
6383 limit
= self
._resolve
_field
_value
(field
, limit_text
)
6386 'closest': False if limit
is None else closest
,
6387 'limit_text': limit_text
,
6389 if field
in self
.settings
:
6390 self
.settings
[field
].update(data
)
6392 self
.settings
[field
] = data
6395 tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'forced'))
6396 + (tuple() if params
.get('format_sort_force', False)
6397 else tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'priority')))
6398 + tuple(self
._sort
_user
) + tuple(sort_extractor
) + self
.default
)
6400 for item
in sort_list
:
6401 match
= re
.match(self
.regex
, item
)
6403 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item
)
6404 field
= match
.group('field')
6407 if self
._get
_field
_setting
(field
, 'type') == 'alias':
6408 alias
, field
= field
, self
._get
_field
_setting
(field
, 'field')
6409 if self
._get
_field
_setting
(alias
, 'deprecated'):
6410 self
.ydl
.deprecated_feature(f
'Format sorting alias {alias} is deprecated and may '
6411 f
'be removed in a future version. Please use {field} instead')
6412 reverse
= match
.group('reverse') is not None
6413 closest
= match
.group('separator') == '~'
6414 limit_text
= match
.group('limit')
6416 has_limit
= limit_text
is not None
6417 has_multiple_fields
= self
._get
_field
_setting
(field
, 'type') == 'combined'
6418 has_multiple_limits
= has_limit
and has_multiple_fields
and not self
._get
_field
_setting
(field
, 'same_limit')
6420 fields
= self
._get
_field
_setting
(field
, 'field') if has_multiple_fields
else (field
,)
6421 limits
= limit_text
.split(':') if has_multiple_limits
else (limit_text
,) if has_limit
else tuple()
6422 limit_count
= len(limits
)
6423 for (i
, f
) in enumerate(fields
):
6424 add_item(f
, reverse
, closest
,
6425 limits
[i
] if i
< limit_count
6426 else limits
[0] if has_limit
and not has_multiple_limits
6429 def print_verbose_info(self
, write_debug
):
6431 write_debug('Sort order given by user: %s' % ', '.join(self
._sort
_user
))
6432 if self
._sort
_extractor
:
6433 write_debug('Sort order given by extractor: %s' % ', '.join(self
._sort
_extractor
))
6434 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
6435 '+' if self
._get
_field
_setting
(field
, 'reverse') else '', field
,
6436 '%s%s(%s)' % ('~' if self
._get
_field
_setting
(field
, 'closest') else ':',
6437 self
._get
_field
_setting
(field
, 'limit_text'),
6438 self
._get
_field
_setting
(field
, 'limit'))
6439 if self
._get
_field
_setting
(field
, 'limit_text') is not None else '')
6440 for field
in self
._order
if self
._get
_field
_setting
(field
, 'visible')]))
6442 def _calculate_field_preference_from_value(self
, format
, field
, type, value
):
6443 reverse
= self
._get
_field
_setting
(field
, 'reverse')
6444 closest
= self
._get
_field
_setting
(field
, 'closest')
6445 limit
= self
._get
_field
_setting
(field
, 'limit')
6447 if type == 'extractor':
6448 maximum
= self
._get
_field
_setting
(field
, 'max')
6449 if value
is None or (maximum
is not None and value
>= maximum
):
6451 elif type == 'boolean':
6452 in_list
= self
._get
_field
_setting
(field
, 'in_list')
6453 not_in_list
= self
._get
_field
_setting
(field
, 'not_in_list')
6454 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
6455 elif type == 'ordered':
6456 value
= self
._resolve
_field
_value
(field
, value
, True)
6458 # try to convert to number
6459 val_num
= float_or_none(value
, default
=self
._get
_field
_setting
(field
, 'default'))
6460 is_num
= self
._get
_field
_setting
(field
, 'convert') != 'string' and val_num
is not None
6464 return ((-10, 0) if value
is None
6465 else (1, value
, 0) if not is_num
# if a field has mixed strings and numbers, strings are sorted higher
6466 else (0, -abs(value
- limit
), value
- limit
if reverse
else limit
- value
) if closest
6467 else (0, value
, 0) if not reverse
and (limit
is None or value
<= limit
)
6468 else (0, -value
, 0) if limit
is None or (reverse
and value
== limit
) or value
> limit
6469 else (-1, value
, 0))
6471 def _calculate_field_preference(self
, format
, field
):
6472 type = self
._get
_field
_setting
(field
, 'type') # extractor, boolean, ordered, field, multiple
6473 get_value
= lambda f
: format
.get(self
._get
_field
_setting
(f
, 'field'))
6474 if type == 'multiple':
6475 type = 'field' # Only 'field' is allowed in multiple for now
6476 actual_fields
= self
._get
_field
_setting
(field
, 'field')
6478 value
= self
._get
_field
_setting
(field
, 'function')(get_value(f
) for f
in actual_fields
)
6480 value
= get_value(field
)
6481 return self
._calculate
_field
_preference
_from
_value
(format
, field
, type, value
)
6483 def calculate_preference(self
, format
):
6484 # Determine missing protocol
6485 if not format
.get('protocol'):
6486 format
['protocol'] = determine_protocol(format
)
6488 # Determine missing ext
6489 if not format
.get('ext') and 'url' in format
:
6490 format
['ext'] = determine_ext(format
['url'])
6491 if format
.get('vcodec') == 'none':
6492 format
['audio_ext'] = format
['ext'] if format
.get('acodec') != 'none' else 'none'
6493 format
['video_ext'] = 'none'
6495 format
['video_ext'] = format
['ext']
6496 format
['audio_ext'] = 'none'
6497 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
6498 # format['preference'] = -1000
6500 if format
.get('preference') is None and format
.get('ext') == 'flv' and re
.match('[hx]265|he?vc?', format
.get('vcodec') or ''):
6501 # HEVC-over-FLV is out-of-spec by FLV's original spec
6502 # ref. https://trac.ffmpeg.org/ticket/6389
6503 # ref. https://github.com/yt-dlp/yt-dlp/pull/5821
6504 format
['preference'] = -100
6506 # Determine missing bitrates
6507 if format
.get('tbr') is None:
6508 if format
.get('vbr') is not None and format
.get('abr') is not None:
6509 format
['tbr'] = format
.get('vbr', 0) + format
.get('abr', 0)
6511 if format
.get('vcodec') != 'none' and format
.get('vbr') is None:
6512 format
['vbr'] = format
.get('tbr') - format
.get('abr', 0)
6513 if format
.get('acodec') != 'none' and format
.get('abr') is None:
6514 format
['abr'] = format
.get('tbr') - format
.get('vbr', 0)
6516 return tuple(self
._calculate
_field
_preference
(format
, field
) for field
in self
._order
)
6520 has_certifi
= bool(certifi
)
6521 has_websockets
= bool(websockets
)
6524 def load_plugins(name
, suffix
, namespace
):
6525 from .plugins
import load_plugins
6526 ret
= load_plugins(name
, suffix
)
6527 namespace
.update(ret
)