48 import xml
.etree
.ElementTree
51 from .compat
import functools
# isort: split
53 compat_etree_fromstring
,
55 compat_HTMLParseError
,
59 from .dependencies
import brotli
, certifi
, websockets
, xattr
60 from .socks
import ProxyType
, sockssocket
63 def register_socks_protocols():
64 # "Register" SOCKS protocols
65 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
66 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
67 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
68 if scheme
not in urllib
.parse
.uses_netloc
:
69 urllib
.parse
.uses_netloc
.append(scheme
)
72 # This is not clearly defined otherwise
73 compiled_regex_type
= type(re
.compile(''))
76 def random_user_agent():
77 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
118 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
121 SUPPORTED_ENCODINGS
= [
125 SUPPORTED_ENCODINGS
.append('br')
128 'User-Agent': random_user_agent(),
129 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
130 'Accept-Language': 'en-us,en;q=0.5',
131 'Sec-Fetch-Mode': 'navigate',
136 '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',
140 NO_DEFAULT
= object()
141 IDENTITY
= lambda x
: x
143 ENGLISH_MONTH_NAMES
= [
144 'January', 'February', 'March', 'April', 'May', 'June',
145 'July', 'August', 'September', 'October', 'November', 'December']
148 'en': ENGLISH_MONTH_NAMES
,
150 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
151 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
152 # these follow the genitive grammatical case (dopełniacz)
153 # some websites might be using nominative, which will require another month list
154 # https://en.wikibooks.org/wiki/Polish/Noun_cases
155 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
156 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
159 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
161 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
162 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
163 'EST': -5, 'EDT': -4, # Eastern
164 'CST': -6, 'CDT': -5, # Central
165 'MST': -7, 'MDT': -6, # Mountain
166 'PST': -8, 'PDT': -7 # Pacific
169 # needed for sanitizing filenames in restricted mode
170 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
171 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
172 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
202 '%Y-%m-%d %H:%M:%S.%f',
203 '%Y-%m-%d %H:%M:%S:%f',
206 '%Y-%m-%dT%H:%M:%SZ',
207 '%Y-%m-%dT%H:%M:%S.%fZ',
208 '%Y-%m-%dT%H:%M:%S.%f0Z',
210 '%Y-%m-%dT%H:%M:%S.%f',
213 '%b %d %Y at %H:%M:%S',
215 '%B %d %Y at %H:%M:%S',
219 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
220 DATE_FORMATS_DAY_FIRST
.extend([
230 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
231 DATE_FORMATS_MONTH_FIRST
.extend([
239 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
240 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
242 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
246 def preferredencoding():
247 """Get preferred encoding.
249 Returns the best encoding scheme for the system, based on
250 locale.getpreferredencoding() and some further tweaks.
253 pref = locale.getpreferredencoding()
261 def write_json_file(obj, fn):
262 """ Encode obj as JSON and write it to fn, atomically if possible """
264 tf = tempfile.NamedTemporaryFile(
265 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
266 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
270 json.dump(obj, tf, ensure_ascii=False)
271 if sys.platform == 'win32
':
272 # Need to remove existing file on Windows, else os.rename raises
273 # WindowsError or FileExistsError.
274 with contextlib.suppress(OSError):
276 with contextlib.suppress(OSError):
279 os.chmod(tf.name, 0o666 & ~mask)
280 os.rename(tf.name, fn)
282 with contextlib.suppress(OSError):
287 def find_xpath_attr(node, xpath, key, val=None):
288 """ Find the xpath xpath[@key=val] """
289 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
290 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
291 return node.find(expr)
293 # On python2.6 the xml.etree.ElementTree.Element methods don't support
294 # the namespace parameter
297 def xpath_with_ns(path
, ns_map
):
298 components
= [c
.split(':') for c
in path
.split('/')]
302 replaced
.append(c
[0])
305 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
306 return '/'.join(replaced
)
309 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
310 def _find_xpath(xpath
):
311 return node
.find(xpath
)
313 if isinstance(xpath
, str):
314 n
= _find_xpath(xpath
)
322 if default
is not NO_DEFAULT
:
325 name
= xpath
if name
is None else name
326 raise ExtractorError('Could not find XML element %s' % name
)
332 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
333 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
334 if n
is None or n
== default
:
337 if default
is not NO_DEFAULT
:
340 name
= xpath
if name
is None else name
341 raise ExtractorError('Could not find XML element\'s text %s' % name
)
347 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
348 n
= find_xpath_attr(node
, xpath
, key
)
350 if default
is not NO_DEFAULT
:
353 name
= f
'{xpath}[@{key}]' if name
is None else name
354 raise ExtractorError('Could not find XML attribute %s' % name
)
360 def get_element_by_id(id, html
, **kwargs
):
361 """Return the content of the tag with the specified ID in the passed HTML document"""
362 return get_element_by_attribute('id', id, html
, **kwargs
)
365 def get_element_html_by_id(id, html
, **kwargs
):
366 """Return the html of the tag with the specified ID in the passed HTML document"""
367 return get_element_html_by_attribute('id', id, html
, **kwargs
)
370 def get_element_by_class(class_name
, html
):
371 """Return the content of the first tag with the specified class in the passed HTML document"""
372 retval
= get_elements_by_class(class_name
, html
)
373 return retval
[0] if retval
else None
376 def get_element_html_by_class(class_name
, html
):
377 """Return the html of the first tag with the specified class in the passed HTML document"""
378 retval
= get_elements_html_by_class(class_name
, html
)
379 return retval
[0] if retval
else None
382 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
383 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
384 return retval
[0] if retval
else None
387 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
388 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
389 return retval
[0] if retval
else None
392 def get_elements_by_class(class_name
, html
, **kargs
):
393 """Return the content of all tags with the specified class in the passed HTML document as a list"""
394 return get_elements_by_attribute(
395 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
396 html, escape_value=False)
399 def get_elements_html_by_class(class_name, html):
400 """Return the html of all tags with the specified class in the passed HTML document as a list"""
401 return get_elements_html_by_attribute(
402 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
403 html, escape_value=False)
406 def get_elements_by_attribute(*args, **kwargs):
407 """Return the content of the tag with the specified attribute in the passed HTML document"""
408 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
411 def get_elements_html_by_attribute(*args, **kwargs):
412 """Return the html of the tag with the specified attribute in the passed HTML document"""
413 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
416 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
418 Return the text (content) and the html (whole) of the tag with the specified
419 attribute in the passed HTML document
424 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
426 value = re.escape(value) if escape_value else value
428 partial_element_re = rf'''(?x
)
430 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
431 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
434 for m in re.finditer(partial_element_re, html):
435 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
438 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
443 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
445 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
446 closing tag for the first opening tag it has encountered, and can be used
450 class HTMLBreakOnClosingTagException(Exception):
454 self.tagstack = collections.deque()
455 html.parser.HTMLParser.__init__(self)
460 def __exit__(self, *_):
464 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
465 # so data remains buffered; we no longer have any interest in it, thus
466 # override this method to discard it
469 def handle_starttag(self, tag, _):
470 self.tagstack.append(tag)
472 def handle_endtag(self, tag):
473 if not self.tagstack:
474 raise compat_HTMLParseError('no tags
in the stack
')
476 inner_tag = self.tagstack.pop()
480 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
481 if not self.tagstack:
482 raise self.HTMLBreakOnClosingTagException()
485 # XXX: This should be far less strict
486 def get_element_text_and_html_by_tag(tag, html):
488 For the first element with the specified tag in the passed HTML document
489 return its' content (text
) and the whole
element (html
)
491 def find_or_raise(haystack, needle, exc):
493 return haystack.index(needle)
496 closing_tag = f'</{tag}>'
497 whole_start = find_or_raise(
498 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
499 content_start = find_or_raise(
500 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
501 content_start += whole_start + 1
502 with HTMLBreakOnClosingTagParser() as parser:
503 parser.feed(html[whole_start:content_start])
504 if not parser.tagstack or parser.tagstack[0] != tag:
505 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
506 offset = content_start
507 while offset < len(html):
508 next_closing_tag_start = find_or_raise(
509 html[offset:], closing_tag,
510 compat_HTMLParseError(f'closing {tag} tag not found'))
511 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
513 parser.feed(html[offset:offset + next_closing_tag_end])
514 offset += next_closing_tag_end
515 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
516 return html[content_start:offset + next_closing_tag_start], \
517 html[whole_start:offset + next_closing_tag_end]
518 raise compat_HTMLParseError('unexpected end of html')
521 class HTMLAttributeParser(html.parser.HTMLParser):
522 """Trivial HTML parser to gather the attributes
for a single element
"""
526 html.parser.HTMLParser.__init__(self)
528 def handle_starttag(self, tag, attrs):
529 self.attrs = dict(attrs)
530 raise compat_HTMLParseError('done')
533 class HTMLListAttrsParser(html.parser.HTMLParser):
534 """HTML parser to gather the attributes
for the elements of a
list"""
537 html.parser.HTMLParser.__init__(self)
541 def handle_starttag(self, tag, attrs):
542 if tag == 'li' and self._level == 0:
543 self.items.append(dict(attrs))
546 def handle_endtag(self, tag):
550 def extract_attributes(html_element):
551 """Given a string
for an HTML element such
as
553 a
="foo" B
="bar" c
="&98;az" d
=boz
554 empty
= noval entity
="&"
557 Decode
and return a dictionary of attributes
.
559 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
560 'empty': '', 'noval': None, 'entity': '&',
561 'sq': '"', 'dq': '\''
564 parser = HTMLAttributeParser()
565 with contextlib.suppress(compat_HTMLParseError):
566 parser.feed(html_element)
571 def parse_list(webpage):
572 """Given a string
for an series of HTML
<li
> elements
,
573 return a dictionary of their attributes
"""
574 parser = HTMLListAttrsParser()
580 def clean_html(html):
581 """Clean an HTML snippet into a readable string
"""
583 if html is None: # Convenience for sanitizing descriptions etc.
586 html = re.sub(r'\s+', ' ', html)
587 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
588 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
590 html = re.sub('<.*?>', '', html)
591 # Replace html entities
592 html = unescapeHTML(html)
596 class LenientJSONDecoder(json.JSONDecoder):
597 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
598 self.transform_source, self.ignore_extra = transform_source, ignore_extra
599 super().__init__(*args, **kwargs)
602 if self.transform_source:
603 s = self.transform_source(s)
605 if self.ignore_extra:
606 return self.raw_decode(s.lstrip())[0]
607 return super().decode(s)
608 except json.JSONDecodeError as e:
609 if e.pos is not None:
610 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
614 def sanitize_open(filename, open_mode):
615 """Try to
open the given filename
, and slightly tweak it
if this fails
.
617 Attempts to
open the given filename
. If this fails
, it tries to change
618 the filename slightly
, step by step
, until it
's either able to open it
619 or it fails and raises a final exception, like the standard open()
622 It returns the tuple (stream, definitive_file_name).
625 if sys.platform == 'win32
':
628 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
629 with contextlib.suppress(io.UnsupportedOperation):
630 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
631 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
633 for attempt in range(2):
636 if sys.platform == 'win32
':
637 # FIXME: An exclusive lock also locks the file from being read.
638 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
639 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
640 raise LockingUnsupportedError()
641 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
643 stream
= open(filename
, open_mode
)
644 return stream
, filename
645 except OSError as err
:
646 if attempt
or err
.errno
in (errno
.EACCES
,):
648 old_filename
, filename
= filename
, sanitize_path(filename
)
649 if old_filename
== filename
:
653 def timeconvert(timestr
):
654 """Convert RFC 2822 defined time string into system timestamp"""
656 timetuple
= email
.utils
.parsedate_tz(timestr
)
657 if timetuple
is not None:
658 timestamp
= email
.utils
.mktime_tz(timetuple
)
662 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
663 """Sanitizes a string so it could be used as part of a filename.
664 @param restricted Use a stricter subset of allowed characters
665 @param is_id Whether this is an ID that should be kept unchanged if possible.
666 If unset, yt-dlp's new sanitization rules are in effect
671 def replace_insane(char
):
672 if restricted
and char
in ACCENT_CHARS
:
673 return ACCENT_CHARS
[char
]
674 elif not restricted
and char
== '\n':
676 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
677 # Replace with their full-width unicode counterparts
678 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
679 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
682 return '' if restricted
else '\''
684 return '\0_\0-' if restricted
else '\0 \0-'
685 elif char
in '\\/|*<>':
687 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
691 # Replace look-alike Unicode glyphs
692 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
693 s
= unicodedata
.normalize('NFKC', s
)
694 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
695 result
= ''.join(map(replace_insane
, s
))
696 if is_id
is NO_DEFAULT
:
697 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
698 STRIP_RE
= r
'(?:\0.|[ _-])*'
699 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
700 result
= result
.replace('\0', '') or '_'
703 while '__' in result
:
704 result
= result
.replace('__', '_')
705 result
= result
.strip('_')
706 # Common case of "Foreign band name - English song title"
707 if restricted
and result
.startswith('-_'):
709 if result
.startswith('-'):
710 result
= '_' + result
[len('-'):]
711 result
= result
.lstrip('.')
717 def sanitize_path(s
, force
=False):
718 """Sanitizes and normalizes path on Windows"""
719 if sys
.platform
== 'win32':
721 drive_or_unc
, _
= os
.path
.splitdrive(s
)
727 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
731 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
732 for path_part
in norm_path
]
734 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
735 elif force
and s
and s
[0] == os
.path
.sep
:
736 sanitized_path
.insert(0, os
.path
.sep
)
737 return os
.path
.join(*sanitized_path
)
740 def sanitize_url(url
, *, scheme
='http'):
741 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
742 # the number of unwanted failures due to missing protocol
745 elif url
.startswith('//'):
746 return f
'{scheme}:{url}'
747 # Fix some common typos seen so far
749 # https://github.com/ytdl-org/youtube-dl/issues/15649
750 (r
'^httpss://', r
'https://'),
751 # https://bx1.be/lives/direct-tv/
752 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
754 for mistake
, fixup
in COMMON_TYPOS
:
755 if re
.match(mistake
, url
):
756 return re
.sub(mistake
, fixup
, url
)
760 def extract_basic_auth(url
):
761 parts
= urllib
.parse
.urlsplit(url
)
762 if parts
.username
is None:
764 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
765 parts
.hostname
if parts
.port
is None
766 else '%s:%d' % (parts
.hostname
, parts
.port
))))
767 auth_payload
= base64
.b64encode(
768 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
769 return url
, f
'Basic {auth_payload.decode()}'
772 def sanitized_Request(url
, *args
, **kwargs
):
773 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
774 if auth_header
is not None:
775 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
776 headers
['Authorization'] = auth_header
777 return urllib
.request
.Request(url
, *args
, **kwargs
)
781 """Expand shell variables and ~"""
782 return os
.path
.expandvars(compat_expanduser(s
))
785 def orderedSet(iterable
, *, lazy
=False):
786 """Remove all duplicates from the input iterable"""
788 seen
= [] # Do not use set since the items can be unhashable
794 return _iter() if lazy
else list(_iter())
797 def _htmlentity_transform(entity_with_semicolon
):
798 """Transforms an HTML entity to a character."""
799 entity
= entity_with_semicolon
[:-1]
801 # Known non-numeric HTML entity
802 if entity
in html
.entities
.name2codepoint
:
803 return chr(html
.entities
.name2codepoint
[entity
])
805 # TODO: HTML5 allows entities without a semicolon.
806 # E.g. 'Éric' should be decoded as 'Éric'.
807 if entity_with_semicolon
in html
.entities
.html5
:
808 return html
.entities
.html5
[entity_with_semicolon
]
810 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
812 numstr
= mobj
.group(1)
813 if numstr
.startswith('x'):
815 numstr
= '0%s' % numstr
818 # See https://github.com/ytdl-org/youtube-dl/issues/7518
819 with contextlib
.suppress(ValueError):
820 return chr(int(numstr
, base
))
822 # Unknown entity in name, return its literal representation
823 return '&%s;' % entity
829 assert isinstance(s
, str)
832 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
835 def escapeHTML(text
):
838 .replace('&', '&')
839 .replace('<', '<')
840 .replace('>', '>')
841 .replace('"', '"')
842 .replace("'", ''')
846 def process_communicate_or_kill(p
, *args
, **kwargs
):
847 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
848 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
849 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
852 class Popen(subprocess
.Popen
):
853 if sys
.platform
== 'win32':
854 _startupinfo
= subprocess
.STARTUPINFO()
855 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
860 def _fix_pyinstaller_ld_path(env
):
861 """Restore LD_LIBRARY_PATH when using PyInstaller
862 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
863 https://github.com/yt-dlp/yt-dlp/issues/4573
865 if not hasattr(sys
, '_MEIPASS'):
869 orig
= env
.get(f
'{key}_ORIG')
875 _fix('LD_LIBRARY_PATH') # Linux
876 _fix('DYLD_LIBRARY_PATH') # macOS
878 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
880 env
= os
.environ
.copy()
881 self
._fix
_pyinstaller
_ld
_path
(env
)
884 kwargs
['universal_newlines'] = True # For 3.6 compatibility
885 kwargs
.setdefault('encoding', 'utf-8')
886 kwargs
.setdefault('errors', 'replace')
887 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
889 def communicate_or_kill(self
, *args
, **kwargs
):
891 return self
.communicate(*args
, **kwargs
)
892 except BaseException
: # Including KeyboardInterrupt
893 self
.kill(timeout
=None)
896 def kill(self
, *, timeout
=0):
899 self
.wait(timeout
=timeout
)
902 def run(cls
, *args
, timeout
=None, **kwargs
):
903 with cls(*args
, **kwargs
) as proc
:
904 default
= '' if proc
.text_mode
else b
''
905 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
906 return stdout
or default
, stderr
or default
, proc
.returncode
909 def get_subprocess_encoding():
910 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
911 # For subprocess calls, encode with locale encoding
912 # Refer to http://stackoverflow.com/a/9951851/35070
913 encoding
= preferredencoding()
915 encoding
= sys
.getfilesystemencoding()
921 def encodeFilename(s
, for_subprocess
=False):
922 assert isinstance(s
, str)
926 def decodeFilename(b
, for_subprocess
=False):
930 def encodeArgument(s
):
931 # Legacy code that uses byte strings
932 # Uncomment the following line after fixing all post processors
933 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
934 return s
if isinstance(s
, str) else s
.decode('ascii')
937 def decodeArgument(b
):
941 def decodeOption(optval
):
944 if isinstance(optval
, bytes):
945 optval
= optval
.decode(preferredencoding())
947 assert isinstance(optval
, str)
951 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
954 def timetuple_from_msec(msec
):
955 secs
, msec
= divmod(msec
, 1000)
956 mins
, secs
= divmod(secs
, 60)
957 hrs
, mins
= divmod(mins
, 60)
958 return _timetuple(hrs
, mins
, secs
, msec
)
961 def formatSeconds(secs
, delim
=':', msec
=False):
962 time
= timetuple_from_msec(secs
* 1000)
964 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
966 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
968 ret
= '%d' % time
.seconds
969 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
972 def _ssl_load_windows_store_certs(ssl_context
, storename
):
973 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
975 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
976 if encoding
== 'x509_asn' and (
977 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
978 except PermissionError
:
981 with contextlib
.suppress(ssl
.SSLError
):
982 ssl_context
.load_verify_locations(cadata
=cert
)
985 def make_HTTPS_handler(params
, **kwargs
):
986 opts_check_certificate
= not params
.get('nocheckcertificate')
987 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
988 context
.check_hostname
= opts_check_certificate
989 if params
.get('legacyserverconnect'):
990 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
991 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
992 context
.set_ciphers('DEFAULT')
994 sys
.version_info
< (3, 10)
995 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
996 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
998 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
999 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
1000 # in some situations [2][3].
1001 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
1002 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1003 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1004 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1005 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1006 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1007 # 4. https://peps.python.org/pep-0644/
1008 # 5. https://peps.python.org/pep-0644/#libressl-support
1009 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1010 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1011 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1013 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1014 if opts_check_certificate
:
1015 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1016 context
.load_verify_locations(cafile
=certifi
.where())
1019 context
.load_default_certs()
1020 # Work around the issue in load_default_certs when there are bad certificates. See:
1021 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1022 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1023 except ssl
.SSLError
:
1024 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1025 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1026 for storename
in ('CA', 'ROOT'):
1027 _ssl_load_windows_store_certs(context
, storename
)
1028 context
.set_default_verify_paths()
1030 client_certfile
= params
.get('client_certificate')
1033 context
.load_cert_chain(
1034 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1035 password
=params
.get('client_certificate_password'))
1036 except ssl
.SSLError
:
1037 raise YoutubeDLError('Unable to load client certificate')
1039 # Some servers may reject requests if ALPN extension is not sent. See:
1040 # https://github.com/python/cpython/issues/85140
1041 # https://github.com/yt-dlp/yt-dlp/issues/3878
1042 with contextlib
.suppress(NotImplementedError):
1043 context
.set_alpn_protocols(['http/1.1'])
1045 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1048 def bug_reports_message(before
=';'):
1049 from .update
import REPOSITORY
1051 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1052 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1054 before
= before
.rstrip()
1055 if not before
or before
.endswith(('.', '!', '?')):
1056 msg
= msg
[0].title() + msg
[1:]
1058 return (before
+ ' ' if before
else '') + msg
1061 class YoutubeDLError(Exception):
1062 """Base exception for YoutubeDL errors."""
1065 def __init__(self
, msg
=None):
1068 elif self
.msg
is None:
1069 self
.msg
= type(self
).__name
__
1070 super().__init
__(self
.msg
)
1073 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1074 if hasattr(ssl
, 'CertificateError'):
1075 network_exceptions
.append(ssl
.CertificateError
)
1076 network_exceptions
= tuple(network_exceptions
)
1079 class ExtractorError(YoutubeDLError
):
1080 """Error during info extraction."""
1082 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1083 """ tb, if given, is the original traceback (so that it can be printed out).
1084 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1086 if sys
.exc_info()[0] in network_exceptions
:
1089 self
.orig_msg
= str(msg
)
1091 self
.expected
= expected
1093 self
.video_id
= video_id
1095 self
.exc_info
= sys
.exc_info() # preserve original exception
1096 if isinstance(self
.exc_info
[1], ExtractorError
):
1097 self
.exc_info
= self
.exc_info
[1].exc_info
1098 super().__init
__(self
.__msg
)
1103 format_field(self
.ie
, None, '[%s] '),
1104 format_field(self
.video_id
, None, '%s: '),
1106 format_field(self
.cause
, None, ' (caused by %r)'),
1107 '' if self
.expected
else bug_reports_message()))
1109 def format_traceback(self
):
1110 return join_nonempty(
1111 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1112 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1115 def __setattr__(self
, name
, value
):
1116 super().__setattr
__(name
, value
)
1117 if getattr(self
, 'msg', None) and name
not in ('msg', 'args'):
1118 self
.msg
= self
.__msg
or type(self
).__name
__
1119 self
.args
= (self
.msg
, ) # Cannot be property
1122 class UnsupportedError(ExtractorError
):
1123 def __init__(self
, url
):
1125 'Unsupported URL: %s' % url
, expected
=True)
1129 class RegexNotFoundError(ExtractorError
):
1130 """Error when a regex didn't match"""
1134 class GeoRestrictedError(ExtractorError
):
1135 """Geographic restriction Error exception.
1137 This exception may be thrown when a video is not available from your
1138 geographic location due to geographic restrictions imposed by a website.
1141 def __init__(self
, msg
, countries
=None, **kwargs
):
1142 kwargs
['expected'] = True
1143 super().__init
__(msg
, **kwargs
)
1144 self
.countries
= countries
1147 class UserNotLive(ExtractorError
):
1148 """Error when a channel/user is not live"""
1150 def __init__(self
, msg
=None, **kwargs
):
1151 kwargs
['expected'] = True
1152 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1155 class DownloadError(YoutubeDLError
):
1156 """Download Error exception.
1158 This exception may be thrown by FileDownloader objects if they are not
1159 configured to continue on errors. They will contain the appropriate
1163 def __init__(self
, msg
, exc_info
=None):
1164 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1165 super().__init
__(msg
)
1166 self
.exc_info
= exc_info
1169 class EntryNotInPlaylist(YoutubeDLError
):
1170 """Entry not in playlist exception.
1172 This exception will be thrown by YoutubeDL when a requested entry
1173 is not found in the playlist info_dict
1175 msg
= 'Entry not found in info'
1178 class SameFileError(YoutubeDLError
):
1179 """Same File exception.
1181 This exception will be thrown by FileDownloader objects if they detect
1182 multiple files would have to be downloaded to the same file on disk.
1184 msg
= 'Fixed output name but more than one file to download'
1186 def __init__(self
, filename
=None):
1187 if filename
is not None:
1188 self
.msg
+= f
': {filename}'
1189 super().__init
__(self
.msg
)
1192 class PostProcessingError(YoutubeDLError
):
1193 """Post Processing exception.
1195 This exception may be raised by PostProcessor's .run() method to
1196 indicate an error in the postprocessing task.
1200 class DownloadCancelled(YoutubeDLError
):
1201 """ Exception raised when the download queue should be interrupted """
1202 msg
= 'The download was cancelled'
1205 class ExistingVideoReached(DownloadCancelled
):
1206 """ --break-on-existing triggered """
1207 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1210 class RejectedVideoReached(DownloadCancelled
):
1211 """ --break-on-reject triggered """
1212 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1215 class MaxDownloadsReached(DownloadCancelled
):
1216 """ --max-downloads limit has been reached. """
1217 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1220 class ReExtractInfo(YoutubeDLError
):
1221 """ Video info needs to be re-extracted. """
1223 def __init__(self
, msg
, expected
=False):
1224 super().__init
__(msg
)
1225 self
.expected
= expected
1228 class ThrottledDownload(ReExtractInfo
):
1229 """ Download speed below --throttled-rate. """
1230 msg
= 'The download speed is below throttle limit'
1233 super().__init
__(self
.msg
, expected
=False)
1236 class UnavailableVideoError(YoutubeDLError
):
1237 """Unavailable Format exception.
1239 This exception will be thrown when a video is requested
1240 in a format that is not available for that video.
1242 msg
= 'Unable to download video'
1244 def __init__(self
, err
=None):
1246 self
.msg
+= f
': {err}'
1247 super().__init
__(self
.msg
)
1250 class ContentTooShortError(YoutubeDLError
):
1251 """Content Too Short exception.
1253 This exception may be raised by FileDownloader objects when a file they
1254 download is too small for what the server announced first, indicating
1255 the connection was probably interrupted.
1258 def __init__(self
, downloaded
, expected
):
1259 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1261 self
.downloaded
= downloaded
1262 self
.expected
= expected
1265 class XAttrMetadataError(YoutubeDLError
):
1266 def __init__(self
, code
=None, msg
='Unknown error'):
1267 super().__init
__(msg
)
1271 # Parsing code and msg
1272 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1273 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1274 self
.reason
= 'NO_SPACE'
1275 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1276 self
.reason
= 'VALUE_TOO_LONG'
1278 self
.reason
= 'NOT_SUPPORTED'
1281 class XAttrUnavailableError(YoutubeDLError
):
1285 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1286 hc
= http_class(*args
, **kwargs
)
1287 source_address
= ydl_handler
._params
.get('source_address')
1289 if source_address
is not None:
1290 # This is to workaround _create_connection() from socket where it will try all
1291 # address data from getaddrinfo() including IPv6. This filters the result from
1292 # getaddrinfo() based on the source_address value.
1293 # This is based on the cpython socket.create_connection() function.
1294 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1295 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1296 host
, port
= address
1298 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1299 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1300 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1301 if addrs
and not ip_addrs
:
1302 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1304 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1305 % (ip_version
, source_address
[0]))
1306 for res
in ip_addrs
:
1307 af
, socktype
, proto
, canonname
, sa
= res
1310 sock
= socket
.socket(af
, socktype
, proto
)
1311 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1312 sock
.settimeout(timeout
)
1313 sock
.bind(source_address
)
1315 err
= None # Explicitly break reference cycle
1317 except OSError as _
:
1319 if sock
is not None:
1324 raise OSError('getaddrinfo returns an empty list')
1325 if hasattr(hc
, '_create_connection'):
1326 hc
._create
_connection
= _create_connection
1327 hc
.source_address
= (source_address
, 0)
1332 def handle_youtubedl_headers(headers
):
1333 filtered_headers
= headers
1335 if 'Youtubedl-no-compression' in filtered_headers
:
1336 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1337 del filtered_headers
['Youtubedl-no-compression']
1339 return filtered_headers
1342 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1343 """Handler for HTTP requests and responses.
1345 This class, when installed with an OpenerDirector, automatically adds
1346 the standard headers to every HTTP request and handles gzipped and
1347 deflated responses from web servers. If compression is to be avoided in
1348 a particular request, the original request in the program code only has
1349 to include the HTTP header "Youtubedl-no-compression", which will be
1350 removed before making the real request.
1352 Part of this code was copied from:
1354 http://techknack.net/python-urllib2-handlers/
1356 Andrew Rowls, the author of that code, agreed to release it to the
1360 def __init__(self
, params
, *args
, **kwargs
):
1361 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1362 self
._params
= params
1364 def http_open(self
, req
):
1365 conn_class
= http
.client
.HTTPConnection
1367 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1369 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1370 del req
.headers
['Ytdl-socks-proxy']
1372 return self
.do_open(functools
.partial(
1373 _create_http_connection
, self
, conn_class
, False),
1381 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1383 return zlib
.decompress(data
)
1389 return brotli
.decompress(data
)
1391 def http_request(self
, req
):
1392 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1393 # always respected by websites, some tend to give out URLs with non percent-encoded
1394 # non-ASCII characters (see telemb.py, ard.py [#3412])
1395 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1396 # To work around aforementioned issue we will replace request's original URL with
1397 # percent-encoded one
1398 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1399 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1400 url
= req
.get_full_url()
1401 url_escaped
= escape_url(url
)
1403 # Substitute URL if any change after escaping
1404 if url
!= url_escaped
:
1405 req
= update_Request(req
, url
=url_escaped
)
1407 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1408 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1409 # The dict keys are capitalized because of this bug by urllib
1410 if h
.capitalize() not in req
.headers
:
1411 req
.add_header(h
, v
)
1413 if 'Accept-encoding' not in req
.headers
:
1414 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1416 req
.headers
= handle_youtubedl_headers(req
.headers
)
1418 return super().do_request_(req
)
1420 def http_response(self
, req
, resp
):
1423 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1424 content
= resp
.read()
1425 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1427 uncompressed
= io
.BytesIO(gz
.read())
1428 except OSError as original_ioerror
:
1429 # There may be junk add the end of the file
1430 # See http://stackoverflow.com/q/4928560/35070 for details
1431 for i
in range(1, 1024):
1433 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1434 uncompressed
= io
.BytesIO(gz
.read())
1439 raise original_ioerror
1440 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1441 resp
.msg
= old_resp
.msg
1442 del resp
.headers
['Content-encoding']
1444 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1445 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1446 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1447 resp
.msg
= old_resp
.msg
1448 del resp
.headers
['Content-encoding']
1450 if resp
.headers
.get('Content-encoding', '') == 'br':
1451 resp
= urllib
.request
.addinfourl(
1452 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1453 resp
.msg
= old_resp
.msg
1454 del resp
.headers
['Content-encoding']
1455 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1456 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1457 if 300 <= resp
.code
< 400:
1458 location
= resp
.headers
.get('Location')
1460 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1461 location
= location
.encode('iso-8859-1').decode()
1462 location_escaped
= escape_url(location
)
1463 if location
!= location_escaped
:
1464 del resp
.headers
['Location']
1465 resp
.headers
['Location'] = location_escaped
1468 https_request
= http_request
1469 https_response
= http_response
1472 def make_socks_conn_class(base_class
, socks_proxy
):
1473 assert issubclass(base_class
, (
1474 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1476 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1477 if url_components
.scheme
.lower() == 'socks5':
1478 socks_type
= ProxyType
.SOCKS5
1479 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1480 socks_type
= ProxyType
.SOCKS4
1481 elif url_components
.scheme
.lower() == 'socks4a':
1482 socks_type
= ProxyType
.SOCKS4A
1484 def unquote_if_non_empty(s
):
1487 return urllib
.parse
.unquote_plus(s
)
1491 url_components
.hostname
, url_components
.port
or 1080,
1493 unquote_if_non_empty(url_components
.username
),
1494 unquote_if_non_empty(url_components
.password
),
1497 class SocksConnection(base_class
):
1499 self
.sock
= sockssocket()
1500 self
.sock
.setproxy(*proxy_args
)
1501 if isinstance(self
.timeout
, (int, float)):
1502 self
.sock
.settimeout(self
.timeout
)
1503 self
.sock
.connect((self
.host
, self
.port
))
1505 if isinstance(self
, http
.client
.HTTPSConnection
):
1506 if hasattr(self
, '_context'): # Python > 2.6
1507 self
.sock
= self
._context
.wrap_socket(
1508 self
.sock
, server_hostname
=self
.host
)
1510 self
.sock
= ssl
.wrap_socket(self
.sock
)
1512 return SocksConnection
1515 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1516 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1517 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1518 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1519 self
._params
= params
1521 def https_open(self
, req
):
1523 conn_class
= self
._https
_conn
_class
1525 if hasattr(self
, '_context'): # python > 2.6
1526 kwargs
['context'] = self
._context
1527 if hasattr(self
, '_check_hostname'): # python 3.x
1528 kwargs
['check_hostname'] = self
._check
_hostname
1530 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1532 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1533 del req
.headers
['Ytdl-socks-proxy']
1536 return self
.do_open(
1537 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1538 except urllib
.error
.URLError
as e
:
1539 if (isinstance(e
.reason
, ssl
.SSLError
)
1540 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1541 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1545 def is_path_like(f
):
1546 return isinstance(f
, (str, bytes, os
.PathLike
))
1549 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1551 See [1] for cookie file format.
1553 1. https://curl.haxx.se/docs/http-cookies.html
1555 _HTTPONLY_PREFIX
= '#HttpOnly_'
1557 _HEADER
= '''# Netscape HTTP Cookie File
1558 # This file is generated by yt-dlp. Do not edit.
1561 _CookieFileEntry
= collections
.namedtuple(
1563 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1565 def __init__(self
, filename
=None, *args
, **kwargs
):
1566 super().__init
__(None, *args
, **kwargs
)
1567 if is_path_like(filename
):
1568 filename
= os
.fspath(filename
)
1569 self
.filename
= filename
1572 def _true_or_false(cndn
):
1573 return 'TRUE' if cndn
else 'FALSE'
1575 @contextlib.contextmanager
1576 def open(self
, file, *, write
=False):
1577 if is_path_like(file):
1578 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1585 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1588 if (not ignore_discard
and cookie
.discard
1589 or not ignore_expires
and cookie
.is_expired(now
)):
1591 name
, value
= cookie
.name
, cookie
.value
1593 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1594 # with no name, whereas http.cookiejar regards it as a
1595 # cookie with no value.
1596 name
, value
= '', name
1597 f
.write('%s\n' % '\t'.join((
1599 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1601 self
._true
_or
_false
(cookie
.secure
),
1602 str_or_none(cookie
.expires
, default
=''),
1606 def save(self
, filename
=None, *args
, **kwargs
):
1608 Save cookies to a file.
1609 Code is taken from CPython 3.6
1610 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1612 if filename
is None:
1613 if self
.filename
is not None:
1614 filename
= self
.filename
1616 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1618 # Store session cookies with `expires` set to 0 instead of an empty string
1620 if cookie
.expires
is None:
1623 with self
.open(filename
, write
=True) as f
:
1624 f
.write(self
._HEADER
)
1625 self
._really
_save
(f
, *args
, **kwargs
)
1627 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1628 """Load cookies from a file."""
1629 if filename
is None:
1630 if self
.filename
is not None:
1631 filename
= self
.filename
1633 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1635 def prepare_line(line
):
1636 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1637 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1638 # comments and empty lines are fine
1639 if line
.startswith('#') or not line
.strip():
1641 cookie_list
= line
.split('\t')
1642 if len(cookie_list
) != self
._ENTRY
_LEN
:
1643 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1644 cookie
= self
._CookieFileEntry
(*cookie_list
)
1645 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1646 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1650 with self
.open(filename
) as f
:
1653 cf
.write(prepare_line(line
))
1654 except http
.cookiejar
.LoadError
as e
:
1655 if f
'{line.strip()} '[0] in '[{"':
1656 raise http
.cookiejar
.LoadError(
1657 'Cookies file must be Netscape formatted, not JSON. See '
1658 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1659 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1662 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1663 # Session cookies are denoted by either `expires` field set to
1664 # an empty string or 0. MozillaCookieJar only recognizes the former
1665 # (see [1]). So we need force the latter to be recognized as session
1666 # cookies on our own.
1667 # Session cookies may be important for cookies-based authentication,
1668 # e.g. usually, when user does not check 'Remember me' check box while
1669 # logging in on a site, some important cookies are stored as session
1670 # cookies so that not recognizing them will result in failed login.
1671 # 1. https://bugs.python.org/issue17164
1673 # Treat `expires=0` cookies as session cookies
1674 if cookie
.expires
== 0:
1675 cookie
.expires
= None
1676 cookie
.discard
= True
1679 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1680 def __init__(self
, cookiejar
=None):
1681 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1683 def http_response(self
, request
, response
):
1684 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1686 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1687 https_response
= http_response
1690 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1691 """YoutubeDL redirect handler
1693 The code is based on HTTPRedirectHandler implementation from CPython [1].
1695 This redirect handler solves two issues:
1696 - ensures redirect URL is always unicode under python 2
1697 - introduces support for experimental HTTP response status code
1698 308 Permanent Redirect [2] used by some sites [3]
1700 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1701 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1702 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1705 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1707 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1708 """Return a Request or None in response to a redirect.
1710 This is called by the http_error_30x methods when a
1711 redirection response is received. If a redirection should
1712 take place, return a new Request to allow http_error_30x to
1713 perform the redirect. Otherwise, raise HTTPError if no-one
1714 else should try to handle this url. Return None if you can't
1715 but another Handler might.
1717 m
= req
.get_method()
1718 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1719 or code
in (301, 302, 303) and m
== "POST")):
1720 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1721 # Strictly (according to RFC 2616), 301 or 302 in response to
1722 # a POST MUST NOT cause a redirection without confirmation
1723 # from the user (of urllib.request, in this case). In practice,
1724 # essentially all clients do redirect in this case, so we do
1727 # Be conciliant with URIs containing a space. This is mainly
1728 # redundant with the more complete encoding done in http_error_302(),
1729 # but it is kept for compatibility with other callers.
1730 newurl
= newurl
.replace(' ', '%20')
1732 CONTENT_HEADERS
= ("content-length", "content-type")
1733 # NB: don't use dict comprehension for python 2.6 compatibility
1734 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1736 # A 303 must either use GET or HEAD for subsequent request
1737 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1738 if code
== 303 and m
!= 'HEAD':
1740 # 301 and 302 redirects are commonly turned into a GET from a POST
1741 # for subsequent requests by browsers, so we'll do the same.
1742 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1743 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1744 if code
in (301, 302) and m
== 'POST':
1747 return urllib
.request
.Request(
1748 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1749 unverifiable
=True, method
=m
)
1752 def extract_timezone(date_str
):
1755 ^.{8,}? # >=8 char non-TZ prefix, if present
1756 (?P<tz>Z| # just the UTC Z, or
1757 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1758 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1759 [ ]? # optional space
1760 (?P<sign>\+|-) # +/-
1761 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1765 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1766 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1767 if timezone
is not None:
1768 date_str
= date_str
[:-len(m
.group('tz'))]
1769 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1771 date_str
= date_str
[:-len(m
.group('tz'))]
1772 if not m
.group('sign'):
1773 timezone
= datetime
.timedelta()
1775 sign
= 1 if m
.group('sign') == '+' else -1
1776 timezone
= datetime
.timedelta(
1777 hours
=sign
* int(m
.group('hours')),
1778 minutes
=sign
* int(m
.group('minutes')))
1779 return timezone
, date_str
1782 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1783 """ Return a UNIX timestamp from the given date """
1785 if date_str
is None:
1788 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1790 if timezone
is None:
1791 timezone
, date_str
= extract_timezone(date_str
)
1793 with contextlib
.suppress(ValueError):
1794 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1795 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1796 return calendar
.timegm(dt
.timetuple())
1799 def date_formats(day_first
=True):
1800 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1803 def unified_strdate(date_str
, day_first
=True):
1804 """Return a string with the date in the format YYYYMMDD"""
1806 if date_str
is None:
1810 date_str
= date_str
.replace(',', ' ')
1811 # Remove AM/PM + timezone
1812 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1813 _
, date_str
= extract_timezone(date_str
)
1815 for expression
in date_formats(day_first
):
1816 with contextlib
.suppress(ValueError):
1817 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1818 if upload_date
is None:
1819 timetuple
= email
.utils
.parsedate_tz(date_str
)
1821 with contextlib
.suppress(ValueError):
1822 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1823 if upload_date
is not None:
1824 return str(upload_date
)
1827 def unified_timestamp(date_str
, day_first
=True):
1828 if date_str
is None:
1831 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1832 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1834 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1835 timezone
, date_str
= extract_timezone(date_str
)
1837 # Remove AM/PM + timezone
1838 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1840 # Remove unrecognized timezones from ISO 8601 alike timestamps
1841 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1843 date_str
= date_str
[:-len(m
.group('tz'))]
1845 # Python only supports microseconds, so remove nanoseconds
1846 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1848 date_str
= m
.group(1)
1850 for expression
in date_formats(day_first
):
1851 with contextlib
.suppress(ValueError):
1852 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1853 return calendar
.timegm(dt
.timetuple())
1855 timetuple
= email
.utils
.parsedate_tz(date_str
)
1857 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1860 def determine_ext(url
, default_ext
='unknown_video'):
1861 if url
is None or '.' not in url
:
1863 guess
= url
.partition('?')[0].rpartition('.')[2]
1864 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1866 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1867 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1868 return guess
.rstrip('/')
1873 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1874 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1877 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1879 Return a datetime object from a string.
1881 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1883 @param format strftime format of DATE
1884 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1885 auto: round to the unit provided in date_str (if applicable).
1887 auto_precision
= False
1888 if precision
== 'auto':
1889 auto_precision
= True
1890 precision
= 'microsecond'
1891 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1892 if date_str
in ('now', 'today'):
1894 if date_str
== 'yesterday':
1895 return today
- datetime
.timedelta(days
=1)
1897 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1899 if match
is not None:
1900 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1901 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1902 unit
= match
.group('unit')
1903 if unit
== 'month' or unit
== 'year':
1904 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1910 delta
= datetime
.timedelta(**{unit + 's': time}
)
1911 new_date
= start_time
+ delta
1913 return datetime_round(new_date
, unit
)
1916 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1919 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1921 Return a date object from a string using datetime_from_str
1923 @param strict Restrict allowed patterns to "YYYYMMDD" and
1924 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1926 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1927 raise ValueError(f
'Invalid date format "{date_str}"')
1928 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1931 def datetime_add_months(dt
, months
):
1932 """Increment/Decrement a datetime object by months."""
1933 month
= dt
.month
+ months
- 1
1934 year
= dt
.year
+ month
// 12
1935 month
= month
% 12 + 1
1936 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1937 return dt
.replace(year
, month
, day
)
1940 def datetime_round(dt
, precision
='day'):
1942 Round a datetime object's time to a specific precision
1944 if precision
== 'microsecond':
1953 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1954 timestamp
= calendar
.timegm(dt
.timetuple())
1955 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1958 def hyphenate_date(date_str
):
1960 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1961 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1962 if match
is not None:
1963 return '-'.join(match
.groups())
1969 """Represents a time interval between two dates"""
1971 def __init__(self
, start
=None, end
=None):
1972 """start and end must be strings in the format accepted by date"""
1973 if start
is not None:
1974 self
.start
= date_from_str(start
, strict
=True)
1976 self
.start
= datetime
.datetime
.min.date()
1978 self
.end
= date_from_str(end
, strict
=True)
1980 self
.end
= datetime
.datetime
.max.date()
1981 if self
.start
> self
.end
:
1982 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1986 """Returns a range that only contains the given day"""
1987 return cls(day
, day
)
1989 def __contains__(self
, date
):
1990 """Check if the date is in the range"""
1991 if not isinstance(date
, datetime
.date
):
1992 date
= date_from_str(date
)
1993 return self
.start
<= date
<= self
.end
1996 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1998 def __eq__(self
, other
):
1999 return (isinstance(other
, DateRange
)
2000 and self
.start
== other
.start
and self
.end
== other
.end
)
2003 def platform_name():
2004 """ Returns the platform name as a str """
2005 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
2006 return platform
.platform()
2010 def system_identifier():
2011 python_implementation
= platform
.python_implementation()
2012 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2013 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2015 with contextlib
.suppress(OSError): # We may not have access to the executable
2016 libc_ver
= platform
.libc_ver()
2018 return 'Python %s (%s %s %s) - %s (%s%s)' % (
2019 platform
.python_version(),
2020 python_implementation
,
2022 platform
.architecture()[0],
2023 platform
.platform(),
2024 ssl
.OPENSSL_VERSION
,
2025 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2030 def get_windows_version():
2031 ''' Get Windows version. returns () if it's not running on Windows '''
2032 if compat_os_name
== 'nt':
2033 return version_tuple(platform
.win32_ver()[1])
2038 def write_string(s
, out
=None, encoding
=None):
2039 assert isinstance(s
, str)
2040 out
= out
or sys
.stderr
2042 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2043 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2045 enc
, buffer = None, out
2046 if 'b' in getattr(out
, 'mode', ''):
2047 enc
= encoding
or preferredencoding()
2048 elif hasattr(out
, 'buffer'):
2050 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2052 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2056 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2057 from . import _IN_CLI
2059 if msg
in deprecation_warning
._cache
:
2061 deprecation_warning
._cache
.add(msg
)
2063 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2064 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2067 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2070 deprecation_warning
._cache
= set()
2073 def bytes_to_intlist(bs
):
2076 if isinstance(bs
[0], int): # Python 3
2079 return [ord(c
) for c
in bs
]
2082 def intlist_to_bytes(xs
):
2085 return struct
.pack('%dB' % len(xs
), *xs
)
2088 class LockingUnsupportedError(OSError):
2089 msg
= 'File locking is not supported'
2092 super().__init
__(self
.msg
)
2095 # Cross-platform file locking
2096 if sys
.platform
== 'win32':
2098 import ctypes
.wintypes
2101 class OVERLAPPED(ctypes
.Structure
):
2103 ('Internal', ctypes
.wintypes
.LPVOID
),
2104 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2105 ('Offset', ctypes
.wintypes
.DWORD
),
2106 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2107 ('hEvent', ctypes
.wintypes
.HANDLE
),
2110 kernel32
= ctypes
.windll
.kernel32
2111 LockFileEx
= kernel32
.LockFileEx
2112 LockFileEx
.argtypes
= [
2113 ctypes
.wintypes
.HANDLE
, # hFile
2114 ctypes
.wintypes
.DWORD
, # dwFlags
2115 ctypes
.wintypes
.DWORD
, # dwReserved
2116 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2117 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2118 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2120 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2121 UnlockFileEx
= kernel32
.UnlockFileEx
2122 UnlockFileEx
.argtypes
= [
2123 ctypes
.wintypes
.HANDLE
, # hFile
2124 ctypes
.wintypes
.DWORD
, # dwReserved
2125 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2126 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2127 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2129 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2130 whole_low
= 0xffffffff
2131 whole_high
= 0x7fffffff
2133 def _lock_file(f
, exclusive
, block
):
2134 overlapped
= OVERLAPPED()
2135 overlapped
.Offset
= 0
2136 overlapped
.OffsetHigh
= 0
2137 overlapped
.hEvent
= 0
2138 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2140 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2141 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2142 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2143 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2144 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2146 def _unlock_file(f
):
2147 assert f
._lock
_file
_overlapped
_p
2148 handle
= msvcrt
.get_osfhandle(f
.fileno())
2149 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2150 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2156 def _lock_file(f
, exclusive
, block
):
2157 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2159 flags |
= fcntl
.LOCK_NB
2161 fcntl
.flock(f
, flags
)
2162 except BlockingIOError
:
2164 except OSError: # AOSP does not have flock()
2165 fcntl
.lockf(f
, flags
)
2167 def _unlock_file(f
):
2169 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2171 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2175 def _lock_file(f
, exclusive
, block
):
2176 raise LockingUnsupportedError()
2178 def _unlock_file(f
):
2179 raise LockingUnsupportedError()
2185 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2186 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2187 raise NotImplementedError(mode
)
2188 self
.mode
, self
.block
= mode
, block
2190 writable
= any(f
in mode
for f
in 'wax+')
2191 readable
= any(f
in mode
for f
in 'r+')
2192 flags
= functools
.reduce(operator
.ior
, (
2193 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2194 getattr(os
, 'O_BINARY', 0), # Windows only
2195 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2196 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2197 os
.O_APPEND
if 'a' in mode
else 0,
2198 os
.O_EXCL
if 'x' in mode
else 0,
2199 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2202 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2204 def __enter__(self
):
2205 exclusive
= 'r' not in self
.mode
2207 _lock_file(self
.f
, exclusive
, self
.block
)
2212 if 'w' in self
.mode
:
2215 except OSError as e
:
2217 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2218 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2227 _unlock_file(self
.f
)
2231 def __exit__(self
, *_
):
2240 def __getattr__(self
, attr
):
2241 return getattr(self
.f
, attr
)
2248 def get_filesystem_encoding():
2249 encoding
= sys
.getfilesystemencoding()
2250 return encoding
if encoding
is not None else 'utf-8'
2253 def shell_quote(args
):
2255 encoding
= get_filesystem_encoding()
2257 if isinstance(a
, bytes):
2258 # We may get a filename encoded with 'encodeFilename'
2259 a
= a
.decode(encoding
)
2260 quoted_args
.append(compat_shlex_quote(a
))
2261 return ' '.join(quoted_args
)
2264 def smuggle_url(url
, data
):
2265 """ Pass additional data in a URL for internal use. """
2267 url
, idata
= unsmuggle_url(url
, {})
2269 sdata
= urllib
.parse
.urlencode(
2270 {'__youtubedl_smuggle': json.dumps(data)}
)
2271 return url
+ '#' + sdata
2274 def unsmuggle_url(smug_url
, default
=None):
2275 if '#__youtubedl_smuggle' not in smug_url
:
2276 return smug_url
, default
2277 url
, _
, sdata
= smug_url
.rpartition('#')
2278 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2279 data
= json
.loads(jsond
)
2283 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2284 """ Formats numbers with decimal sufixes like K, M, etc """
2285 num
, factor
= float_or_none(num
), float(factor
)
2286 if num
is None or num
< 0:
2288 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2289 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2290 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2292 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2293 converted
= num
/ (factor
** exponent
)
2294 return fmt
% (converted
, suffix
)
2297 def format_bytes(bytes):
2298 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2301 def lookup_unit_table(unit_table
, s
, strict
=False):
2302 num_re
= NUMBER_RE
if strict
else NUMBER_RE
.replace(R
'\.', '[,.]')
2303 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2304 m
= (re
.fullmatch
if strict
else re
.match
)(
2305 rf
'(?P<num>{num_re})\s*(?P<unit>{units_re})\b', s
)
2309 num
= float(m
.group('num').replace(',', '.'))
2310 mult
= unit_table
[m
.group('unit')]
2311 return round(num
* mult
)
2315 """Parse a string indicating a byte quantity into an integer"""
2316 return lookup_unit_table(
2317 {u: 1024**i for i, u in enumerate(['', *'KMGTPEZY'])}
,
2318 s
.upper(), strict
=True)
2321 def parse_filesize(s
):
2325 # The lower-case forms are of course incorrect and unofficial,
2326 # but we support those too
2343 'megabytes': 1000 ** 2,
2344 'mebibytes': 1024 ** 2,
2350 'gigabytes': 1000 ** 3,
2351 'gibibytes': 1024 ** 3,
2357 'terabytes': 1000 ** 4,
2358 'tebibytes': 1024 ** 4,
2364 'petabytes': 1000 ** 5,
2365 'pebibytes': 1024 ** 5,
2371 'exabytes': 1000 ** 6,
2372 'exbibytes': 1024 ** 6,
2378 'zettabytes': 1000 ** 7,
2379 'zebibytes': 1024 ** 7,
2385 'yottabytes': 1000 ** 8,
2386 'yobibytes': 1024 ** 8,
2389 return lookup_unit_table(_UNIT_TABLE
, s
)
2396 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2398 if re
.match(r
'^[\d,.]+$', s
):
2399 return str_to_int(s
)
2412 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2416 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2418 return str_to_int(mobj
.group(1))
2421 def parse_resolution(s
, *, lenient
=False):
2426 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2428 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2431 'width': int(mobj
.group('w')),
2432 'height': int(mobj
.group('h')),
2435 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2437 return {'height': int(mobj.group(1))}
2439 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2441 return {'height': int(mobj.group(1)) * 540}
2446 def parse_bitrate(s
):
2447 if not isinstance(s
, str):
2449 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2451 return int(mobj
.group(1))
2454 def month_by_name(name
, lang
='en'):
2455 """ Return the number of a month by (locale-independently) English name """
2457 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2460 return month_names
.index(name
) + 1
2465 def month_by_abbreviation(abbrev
):
2466 """ Return the number of a month by (locale-independently) English
2470 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2475 def fix_xml_ampersands(xml_str
):
2476 """Replace all the '&' by '&' in XML"""
2478 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2483 def setproctitle(title
):
2484 assert isinstance(title
, str)
2486 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2493 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2497 # LoadLibrary in Windows Python 2.7.13 only expects
2498 # a bytestring, but since unicode_literals turns
2499 # every string into a unicode string, it fails.
2501 title_bytes
= title
.encode()
2502 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2503 buf
.value
= title_bytes
2505 libc
.prctl(15, buf
, 0, 0, 0)
2506 except AttributeError:
2507 return # Strange libc, just skip this
2510 def remove_start(s
, start
):
2511 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2514 def remove_end(s
, end
):
2515 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2518 def remove_quotes(s
):
2519 if s
is None or len(s
) < 2:
2521 for quote
in ('"', "'", ):
2522 if s
[0] == quote
and s
[-1] == quote
:
2527 def get_domain(url
):
2529 This implementation is inconsistent, but is kept for compatibility.
2530 Use this only for "webpage_url_domain"
2532 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2535 def url_basename(url
):
2536 path
= urllib
.parse
.urlparse(url
).path
2537 return path
.strip('/').split('/')[-1]
2541 return re
.match(r
'https?://[^?#]+/', url
).group()
2544 def urljoin(base
, path
):
2545 if isinstance(path
, bytes):
2546 path
= path
.decode()
2547 if not isinstance(path
, str) or not path
:
2549 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2551 if isinstance(base
, bytes):
2552 base
= base
.decode()
2553 if not isinstance(base
, str) or not re
.match(
2554 r
'^(?:https?:)?//', base
):
2556 return urllib
.parse
.urljoin(base
, path
)
2559 class HEADRequest(urllib
.request
.Request
):
2560 def get_method(self
):
2564 class PUTRequest(urllib
.request
.Request
):
2565 def get_method(self
):
2569 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2570 if get_attr
and v
is not None:
2571 v
= getattr(v
, get_attr
, None)
2573 return int(v
) * invscale
// scale
2574 except (ValueError, TypeError, OverflowError):
2578 def str_or_none(v
, default
=None):
2579 return default
if v
is None else str(v
)
2582 def str_to_int(int_str
):
2583 """ A more relaxed version of int_or_none """
2584 if isinstance(int_str
, int):
2586 elif isinstance(int_str
, str):
2587 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2588 return int_or_none(int_str
)
2591 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2595 return float(v
) * invscale
/ scale
2596 except (ValueError, TypeError):
2600 def bool_or_none(v
, default
=None):
2601 return v
if isinstance(v
, bool) else default
2604 def strip_or_none(v
, default
=None):
2605 return v
.strip() if isinstance(v
, str) else default
2608 def url_or_none(url
):
2609 if not url
or not isinstance(url
, str):
2612 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2615 def request_to_url(req
):
2616 if isinstance(req
, urllib
.request
.Request
):
2617 return req
.get_full_url()
2622 def strftime_or_none(timestamp
, date_format
, default
=None):
2623 datetime_object
= None
2625 if isinstance(timestamp
, (int, float)): # unix timestamp
2626 # Using naive datetime here can break timestamp() in Windows
2627 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2628 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2629 elif isinstance(timestamp
, str): # assume YYYYMMDD
2630 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2631 date_format
= re
.sub( # Support %s on windows
2632 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2633 return datetime_object
.strftime(date_format
)
2634 except (ValueError, TypeError, AttributeError):
2638 def parse_duration(s
):
2639 if not isinstance(s
, str):
2645 days
, hours
, mins
, secs
, ms
= [None] * 5
2646 m
= re
.match(r
'''(?x)
2648 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2649 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2650 (?P<ms>[.:][0-9]+)?Z?$
2653 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2658 [0-9]+\s*y(?:ears?)?,?\s*
2661 [0-9]+\s*m(?:onths?)?,?\s*
2664 [0-9]+\s*w(?:eeks?)?,?\s*
2667 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2671 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2674 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2677 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2680 days
, hours
, mins
, secs
, ms
= m
.groups()
2682 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2684 hours
, mins
= m
.groups()
2689 ms
= ms
.replace(':', '.')
2690 return sum(float(part
or 0) * mult
for part
, mult
in (
2691 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2694 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2695 name
, real_ext
= os
.path
.splitext(filename
)
2697 f
'{name}.{ext}{real_ext}'
2698 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2699 else f
'{filename}.{ext}')
2702 def replace_extension(filename
, ext
, expected_real_ext
=None):
2703 name
, real_ext
= os
.path
.splitext(filename
)
2704 return '{}.{}'.format(
2705 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2709 def check_executable(exe
, args
=[]):
2710 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2711 args can be a list of arguments for a short output (like -version) """
2713 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2719 def _get_exe_version_output(exe
, args
):
2721 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2722 # SIGTTOU if yt-dlp is run in the background.
2723 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2724 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2725 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2731 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2732 assert isinstance(output
, str)
2733 if version_re
is None:
2734 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2735 m
= re
.search(version_re
, output
)
2742 def get_exe_version(exe
, args
=['--version'],
2743 version_re
=None, unrecognized
='present'):
2744 """ Returns the version of the specified executable,
2745 or False if the executable is not present """
2746 out
= _get_exe_version_output(exe
, args
)
2747 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2750 def frange(start
=0, stop
=None, step
=1):
2753 start
, stop
= 0, start
2754 sign
= [-1, 1][step
> 0] if step
else 0
2755 while sign
* start
< sign
* stop
:
2760 class LazyList(collections
.abc
.Sequence
):
2761 """Lazy immutable list from an iterable
2762 Note that slices of a LazyList are lists and not LazyList"""
2764 class IndexError(IndexError):
2767 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2768 self
._iterable
= iter(iterable
)
2769 self
._cache
= [] if _cache
is None else _cache
2770 self
._reversed
= reverse
2774 # We need to consume the entire iterable to iterate in reverse
2775 yield from self
.exhaust()
2777 yield from self
._cache
2778 for item
in self
._iterable
:
2779 self
._cache
.append(item
)
2783 self
._cache
.extend(self
._iterable
)
2784 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2788 """Evaluate the entire iterable"""
2789 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2792 def _reverse_index(x
):
2793 return None if x
is None else ~x
2795 def __getitem__(self
, idx
):
2796 if isinstance(idx
, slice):
2798 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2799 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2800 elif isinstance(idx
, int):
2802 idx
= self
._reverse
_index
(idx
)
2803 start
, stop
, step
= idx
, idx
, 0
2805 raise TypeError('indices must be integers or slices')
2806 if ((start
or 0) < 0 or (stop
or 0) < 0
2807 or (start
is None and step
< 0)
2808 or (stop
is None and step
> 0)):
2809 # We need to consume the entire iterable to be able to slice from the end
2810 # Obviously, never use this with infinite iterables
2813 return self
._cache
[idx
]
2814 except IndexError as e
:
2815 raise self
.IndexError(e
) from e
2816 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2818 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2820 return self
._cache
[idx
]
2821 except IndexError as e
:
2822 raise self
.IndexError(e
) from e
2826 self
[-1] if self
._reversed
else self
[0]
2827 except self
.IndexError:
2833 return len(self
._cache
)
2835 def __reversed__(self
):
2836 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2839 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2842 # repr and str should mimic a list. So we exhaust the iterable
2843 return repr(self
.exhaust())
2846 return repr(self
.exhaust())
2851 class IndexError(IndexError):
2855 # This is only useful for tests
2856 return len(self
.getslice())
2858 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2859 self
._pagefunc
= pagefunc
2860 self
._pagesize
= pagesize
2861 self
._pagecount
= float('inf')
2862 self
._use
_cache
= use_cache
2865 def getpage(self
, pagenum
):
2866 page_results
= self
._cache
.get(pagenum
)
2867 if page_results
is None:
2868 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2870 self
._cache
[pagenum
] = page_results
2873 def getslice(self
, start
=0, end
=None):
2874 return list(self
._getslice
(start
, end
))
2876 def _getslice(self
, start
, end
):
2877 raise NotImplementedError('This method must be implemented by subclasses')
2879 def __getitem__(self
, idx
):
2880 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2881 if not isinstance(idx
, int) or idx
< 0:
2882 raise TypeError('indices must be non-negative integers')
2883 entries
= self
.getslice(idx
, idx
+ 1)
2885 raise self
.IndexError()
2889 class OnDemandPagedList(PagedList
):
2890 """Download pages until a page with less than maximum results"""
2892 def _getslice(self
, start
, end
):
2893 for pagenum
in itertools
.count(start
// self
._pagesize
):
2894 firstid
= pagenum
* self
._pagesize
2895 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2896 if start
>= nextfirstid
:
2900 start
% self
._pagesize
2901 if firstid
<= start
< nextfirstid
2904 ((end
- 1) % self
._pagesize
) + 1
2905 if (end
is not None and firstid
<= end
<= nextfirstid
)
2909 page_results
= self
.getpage(pagenum
)
2911 self
._pagecount
= pagenum
- 1
2913 if startv
!= 0 or endv
is not None:
2914 page_results
= page_results
[startv
:endv
]
2915 yield from page_results
2917 # A little optimization - if current page is not "full", ie. does
2918 # not contain page_size videos then we can assume that this page
2919 # is the last one - there are no more ids on further pages -
2920 # i.e. no need to query again.
2921 if len(page_results
) + startv
< self
._pagesize
:
2924 # If we got the whole page, but the next page is not interesting,
2925 # break out early as well
2926 if end
== nextfirstid
:
2930 class InAdvancePagedList(PagedList
):
2931 """PagedList with total number of pages known in advance"""
2933 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2934 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2935 self
._pagecount
= pagecount
2937 def _getslice(self
, start
, end
):
2938 start_page
= start
// self
._pagesize
2939 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2940 skip_elems
= start
- start_page
* self
._pagesize
2941 only_more
= None if end
is None else end
- start
2942 for pagenum
in range(start_page
, end_page
):
2943 page_results
= self
.getpage(pagenum
)
2945 page_results
= page_results
[skip_elems
:]
2947 if only_more
is not None:
2948 if len(page_results
) < only_more
:
2949 only_more
-= len(page_results
)
2951 yield from page_results
[:only_more
]
2953 yield from page_results
2956 class PlaylistEntries
:
2957 MissingEntry
= object()
2958 is_exhausted
= False
2960 def __init__(self
, ydl
, info_dict
):
2963 # _entries must be assigned now since infodict can change during iteration
2964 entries
= info_dict
.get('entries')
2966 raise EntryNotInPlaylist('There are no entries')
2967 elif isinstance(entries
, list):
2968 self
.is_exhausted
= True
2970 requested_entries
= info_dict
.get('requested_entries')
2971 self
.is_incomplete
= requested_entries
is not None
2972 if self
.is_incomplete
:
2973 assert self
.is_exhausted
2974 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
2975 for i
, entry
in zip(requested_entries
, entries
):
2976 self
._entries
[i
- 1] = entry
2977 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2978 self
._entries
= entries
2980 self
._entries
= LazyList(entries
)
2982 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2983 (?P<start>[+-]?\d+)?
2985 (?P<end>[+-]?\d+|inf(?:inite)?)?
2986 (?::(?P<step>[+-]?\d+))?
2990 def parse_playlist_items(cls
, string
):
2991 for segment
in string
.split(','):
2993 raise ValueError('There is two or more consecutive commas')
2994 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2996 raise ValueError(f
'{segment!r} is not a valid specification')
2997 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2998 if int_or_none(step
) == 0:
2999 raise ValueError(f
'Step in {segment!r} cannot be zero')
3000 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
3002 def get_requested_items(self
):
3003 playlist_items
= self
.ydl
.params
.get('playlist_items')
3004 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
3005 playlist_end
= self
.ydl
.params
.get('playlistend')
3006 # For backwards compatibility, interpret -1 as whole list
3007 if playlist_end
in (-1, None):
3009 if not playlist_items
:
3010 playlist_items
= f
'{playlist_start}:{playlist_end}'
3011 elif playlist_start
!= 1 or playlist_end
:
3012 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
3014 for index
in self
.parse_playlist_items(playlist_items
):
3015 for i
, entry
in self
[index
]:
3020 # TODO: Add auto-generated fields
3021 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3022 except (ExistingVideoReached
, RejectedVideoReached
):
3025 def get_full_count(self
):
3026 if self
.is_exhausted
and not self
.is_incomplete
:
3028 elif isinstance(self
._entries
, InAdvancePagedList
):
3029 if self
._entries
._pagesize
== 1:
3030 return self
._entries
._pagecount
3032 @functools.cached_property
3034 if isinstance(self
._entries
, list):
3037 entry
= self
._entries
[i
]
3039 entry
= self
.MissingEntry
3040 if not self
.is_incomplete
:
3041 raise self
.IndexError()
3042 if entry
is self
.MissingEntry
:
3043 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
3048 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3049 except (LazyList
.IndexError, PagedList
.IndexError):
3050 raise self
.IndexError()
3053 def __getitem__(self
, idx
):
3054 if isinstance(idx
, int):
3055 idx
= slice(idx
, idx
)
3057 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3058 step
= 1 if idx
.step
is None else idx
.step
3059 if idx
.start
is None:
3060 start
= 0 if step
> 0 else len(self
) - 1
3062 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3064 # NB: Do not call len(self) when idx == [:]
3065 if idx
.stop
is None:
3066 stop
= 0 if step
< 0 else float('inf')
3068 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3069 stop
+= [-1, 1][step
> 0]
3071 for i
in frange(start
, stop
, step
):
3075 entry
= self
._getter
(i
)
3076 except self
.IndexError:
3077 self
.is_exhausted
= True
3084 return len(tuple(self
[:]))
3086 class IndexError(IndexError):
3090 def uppercase_escape(s
):
3091 unicode_escape
= codecs
.getdecoder('unicode_escape')
3093 r
'\\U[0-9a-fA-F]{8}',
3094 lambda m
: unicode_escape(m
.group(0))[0],
3098 def lowercase_escape(s
):
3099 unicode_escape
= codecs
.getdecoder('unicode_escape')
3101 r
'\\u[0-9a-fA-F]{4}',
3102 lambda m
: unicode_escape(m
.group(0))[0],
3106 def escape_rfc3986(s
):
3107 """Escape non-ASCII characters as suggested by RFC 3986"""
3108 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3111 def escape_url(url
):
3112 """Escape URL as suggested by RFC 3986"""
3113 url_parsed
= urllib
.parse
.urlparse(url
)
3114 return url_parsed
._replace
(
3115 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3116 path
=escape_rfc3986(url_parsed
.path
),
3117 params
=escape_rfc3986(url_parsed
.params
),
3118 query
=escape_rfc3986(url_parsed
.query
),
3119 fragment
=escape_rfc3986(url_parsed
.fragment
)
3123 def parse_qs(url
, **kwargs
):
3124 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3127 def read_batch_urls(batch_fd
):
3129 if not isinstance(url
, str):
3130 url
= url
.decode('utf-8', 'replace')
3131 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3132 for bom
in BOM_UTF8
:
3133 if url
.startswith(bom
):
3134 url
= url
[len(bom
):]
3136 if not url
or url
.startswith(('#', ';', ']')):
3138 # "#" cannot be stripped out since it is part of the URI
3139 # However, it can be safely stripped out if following a whitespace
3140 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3142 with contextlib
.closing(batch_fd
) as fd
:
3143 return [url
for url
in map(fixup
, fd
) if url
]
3146 def urlencode_postdata(*args
, **kargs
):
3147 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3150 def update_url_query(url
, query
):
3153 parsed_url
= urllib
.parse
.urlparse(url
)
3154 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3156 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3157 query
=urllib
.parse
.urlencode(qs
, True)))
3160 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3161 req_headers
= req
.headers
.copy()
3162 req_headers
.update(headers
or {})
3163 req_data
= data
or req
.data
3164 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3165 req_get_method
= req
.get_method()
3166 if req_get_method
== 'HEAD':
3167 req_type
= HEADRequest
3168 elif req_get_method
== 'PUT':
3169 req_type
= PUTRequest
3171 req_type
= urllib
.request
.Request
3173 req_url
, data
=req_data
, headers
=req_headers
,
3174 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3175 if hasattr(req
, 'timeout'):
3176 new_req
.timeout
= req
.timeout
3180 def _multipart_encode_impl(data
, boundary
):
3181 content_type
= 'multipart/form-data; boundary=%s' % boundary
3184 for k
, v
in data
.items():
3185 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3186 if isinstance(k
, str):
3188 if isinstance(v
, str):
3190 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3191 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3192 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3193 if boundary
.encode('ascii') in content
:
3194 raise ValueError('Boundary overlaps with data')
3197 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3199 return out
, content_type
3202 def multipart_encode(data
, boundary
=None):
3204 Encode a dict to RFC 7578-compliant form-data
3207 A dict where keys and values can be either Unicode or bytes-like
3210 If specified a Unicode object, it's used as the boundary. Otherwise
3211 a random boundary is generated.
3213 Reference: https://tools.ietf.org/html/rfc7578
3215 has_specified_boundary
= boundary
is not None
3218 if boundary
is None:
3219 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3222 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3225 if has_specified_boundary
:
3229 return out
, content_type
3232 def variadic(x
, allowed_types
=(str, bytes, dict)):
3233 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3236 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3237 for val
in map(d
.get
, variadic(key_or_keys
)):
3238 if val
is not None and (val
or not skip_false_values
):
3243 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3246 val
= f(*args
, **kwargs
)
3247 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3250 if expected_type
is None or isinstance(val
, expected_type
):
3254 def try_get(src
, getter
, expected_type
=None):
3255 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3258 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3259 return {k: v for k, v in dct.items() if cndn(k, v)}
3262 def merge_dicts(*dicts
):
3264 for a_dict
in dicts
:
3265 for k
, v
in a_dict
.items():
3266 if (v
is not None and k
not in merged
3267 or isinstance(v
, str) and merged
[k
] == ''):
3272 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3273 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3285 TV_PARENTAL_GUIDELINES
= {
3295 def parse_age_limit(s
):
3296 # isinstance(False, int) is True. So type() must be used instead
3297 if type(s
) is int: # noqa: E721
3298 return s
if 0 <= s
<= 21 else None
3299 elif not isinstance(s
, str):
3301 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3303 return int(m
.group('age'))
3306 return US_RATINGS
[s
]
3307 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3309 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3313 def strip_jsonp(code
):
3316 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3317 (?:\s*&&\s*(?P=func_name))?
3318 \s*\(\s*(?P<callback_data>.*)\);?
3319 \s*?(?://[^\n]*)*$''',
3320 r
'\g<callback_data>', code
)
3323 def js_to_json(code
, vars={}, *, strict
=False):
3324 # vars is a dict of var, val pairs to substitute
3325 STRING_QUOTES
= '\'"'
3326 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3327 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3328 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3330 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3331 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3334 def process_escape(match
):
3335 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3336 escape
= match
.group(1) or match
.group(2)
3338 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3339 else R
'\u00' if escape
== 'x'
3340 else '' if escape
== '\n'
3345 if v
in ('true', 'false', 'null'):
3347 elif v
in ('undefined', 'void 0'):
3349 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3352 if v
[0] in STRING_QUOTES
:
3353 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3354 return f
'"{escaped}"'
3356 for regex
, base
in INTEGER_TABLE
:
3357 im
= re
.match(regex
, v
)
3359 i
= int(im
.group(1), base
)
3360 return f
'"{i}":' if v
.endswith(':') else str(i
)
3363 return json
.dumps(vars[v
])
3368 raise ValueError(f
'Unknown value: {v}')
3370 def create_map(mobj
):
3371 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3373 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3375 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3376 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3378 return re
.sub(rf
'''(?sx)
3380 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3381 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3382 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3383 [0-9]+(?={SKIP_RE}:)|
3388 def qualities(quality_ids
):
3389 """ Get a numeric quality value out of a list of possible values """
3392 return quality_ids
.index(qid
)
3398 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3402 'default': '%(title)s [%(id)s].%(ext)s',
3403 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3409 'description': 'description',
3410 'annotation': 'annotations.xml',
3411 'infojson': 'info.json',
3414 'pl_thumbnail': None,
3415 'pl_description': 'description',
3416 'pl_infojson': 'info.json',
3419 # As of [1] format syntax is:
3420 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3421 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3422 STR_FORMAT_RE_TMPL
= r
'''(?x)
3423 (?<!%)(?P<prefix>(?:%%)*)
3425 (?P<has_key>\((?P<key>{0})\))?
3427 (?P<conversion>[#0\-+ ]+)?
3429 (?P<precision>\.\d+)?
3430 (?P<len_mod>[hlL])? # unused in python
3431 {1} # conversion type
3436 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3439 def limit_length(s
, length
):
3440 """ Add ellipses to overly long strings """
3445 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3449 def version_tuple(v
):
3450 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3453 def is_outdated_version(version
, limit
, assume_new
=True):
3455 return not assume_new
3457 return version_tuple(version
) < version_tuple(limit
)
3459 return not assume_new
3462 def ytdl_is_updateable():
3463 """ Returns if yt-dlp can be updated with -U """
3465 from .update
import is_non_updateable
3467 return not is_non_updateable()
3470 def args_to_str(args
):
3471 # Get a short string representation for a subprocess command
3472 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3475 def error_to_compat_str(err
):
3479 def error_to_str(err
):
3480 return f
'{type(err).__name__}: {err}'
3483 def mimetype2ext(mt
):
3487 mt
, _
, params
= mt
.partition(';')
3492 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3493 # it's the most popular one
3494 'audio/mpeg': 'mp3',
3495 'audio/x-wav': 'wav',
3497 'audio/wave': 'wav',
3500 ext
= FULL_MAP
.get(mt
)
3506 'smptett+xml': 'tt',
3510 'x-mp4-fragmented': 'mp4',
3511 'x-ms-sami': 'sami',
3514 'x-mpegurl': 'm3u8',
3515 'vnd.apple.mpegurl': 'm3u8',
3519 'vnd.ms-sstr+xml': 'ism',
3523 'filmstrip+json': 'fs',
3527 _
, _
, subtype
= mt
.rpartition('/')
3528 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3539 _
, _
, suffix
= subtype
.partition('+')
3540 ext
= SUFFIX_MAP
.get(suffix
)
3544 return subtype
.replace('+', '.')
3547 def ext2mimetype(ext_or_url
):
3550 if '.' not in ext_or_url
:
3551 ext_or_url
= f
'file.{ext_or_url}'
3552 return mimetypes
.guess_type(ext_or_url
)[0]
3555 def parse_codecs(codecs_str
):
3556 # http://tools.ietf.org/html/rfc6381
3559 split_codecs
= list(filter(None, map(
3560 str.strip
, codecs_str
.strip().strip(',').split(','))))
3561 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3562 for full_codec
in split_codecs
:
3563 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3564 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3565 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3569 if parts
[0] in ('dvh1', 'dvhe'):
3571 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3573 elif parts
[:2] == ['vp9', '2']:
3575 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3576 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3577 acodec
= acodec
or full_codec
3578 elif parts
[0] in ('stpp', 'wvtt'):
3579 scodec
= scodec
or full_codec
3581 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3582 if vcodec
or acodec
or scodec
:
3584 'vcodec': vcodec
or 'none',
3585 'acodec': acodec
or 'none',
3586 'dynamic_range': hdr
,
3587 **({'scodec': scodec}
if scodec
is not None else {}),
3589 elif len(split_codecs
) == 2:
3591 'vcodec': split_codecs
[0],
3592 'acodec': split_codecs
[1],
3597 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3598 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3600 allow_mkv
= not preferences
or 'mkv' in preferences
3602 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3603 return 'mkv' # TODO: any other format allows this?
3605 # TODO: All codecs supported by parse_codecs isn't handled here
3606 COMPATIBLE_CODECS
= {
3608 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3609 'h264', 'aacl', 'ec-3', # Set in ISM
3612 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3613 'vp9x', 'vp8x', # in the webm spec
3617 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3618 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3620 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3621 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3622 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3626 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3629 for ext
in preferences
or vexts
:
3630 current_exts
= {ext, *vexts, *aexts}
3631 if ext
== 'mkv' or current_exts
== {ext}
or any(
3632 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3634 return 'mkv' if allow_mkv
else preferences
[-1]
3637 def urlhandle_detect_ext(url_handle
):
3638 getheader
= url_handle
.headers
.get
3640 cd
= getheader('Content-Disposition')
3642 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3644 e
= determine_ext(m
.group('filename'), default_ext
=None)
3648 return mimetype2ext(getheader('Content-Type'))
3651 def encode_data_uri(data
, mime_type
):
3652 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3655 def age_restricted(content_limit
, age_limit
):
3656 """ Returns True iff the content should be blocked """
3658 if age_limit
is None: # No limit set
3660 if content_limit
is None:
3661 return False # Content available for everyone
3662 return age_limit
< content_limit
3665 # List of known byte-order-marks (BOM)
3667 (b
'\xef\xbb\xbf', 'utf-8'),
3668 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3669 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3670 (b
'\xff\xfe', 'utf-16-le'),
3671 (b
'\xfe\xff', 'utf-16-be'),
3675 def is_html(first_bytes
):
3676 """ Detect whether a file contains HTML by examining its first bytes. """
3679 for bom
, enc
in BOMS
:
3680 while first_bytes
.startswith(bom
):
3681 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3683 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3686 def determine_protocol(info_dict
):
3687 protocol
= info_dict
.get('protocol')
3688 if protocol
is not None:
3691 url
= sanitize_url(info_dict
['url'])
3692 if url
.startswith('rtmp'):
3694 elif url
.startswith('mms'):
3696 elif url
.startswith('rtsp'):
3699 ext
= determine_ext(url
)
3701 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3705 return urllib
.parse
.urlparse(url
).scheme
3708 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3709 """ Render a list of rows, each as a list of values.
3710 Text after a \t will be right aligned """
3712 return len(remove_terminal_sequences(string
).replace('\t', ''))
3714 def get_max_lens(table
):
3715 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3717 def filter_using_list(row
, filterArray
):
3718 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3720 max_lens
= get_max_lens(data
) if hide_empty
else []
3721 header_row
= filter_using_list(header_row
, max_lens
)
3722 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3724 table
= [header_row
] + data
3725 max_lens
= get_max_lens(table
)
3728 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3729 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3731 for pos
, text
in enumerate(map(str, row
)):
3733 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3735 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3736 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3740 def _match_one(filter_part
, dct
, incomplete
):
3741 # TODO: Generalize code with YoutubeDL._build_format_filter
3742 STRING_OPERATORS
= {
3743 '*=': operator
.contains
,
3744 '^=': lambda attr
, value
: attr
.startswith(value
),
3745 '$=': lambda attr
, value
: attr
.endswith(value
),
3746 '~=': lambda attr
, value
: re
.search(value
, attr
),
3748 COMPARISON_OPERATORS
= {
3750 '<=': operator
.le
, # "<=" must be defined above "<"
3757 if isinstance(incomplete
, bool):
3758 is_incomplete
= lambda _
: incomplete
3760 is_incomplete
= lambda k
: k
in incomplete
3762 operator_rex
= re
.compile(r
'''(?x)
3764 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3766 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3769 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3770 m = operator_rex.fullmatch(filter_part.strip())
3773 unnegated_op = COMPARISON_OPERATORS[m['op']]
3775 op = lambda attr, value: not unnegated_op(attr, value)
3778 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3780 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3781 actual_value = dct.get(m['key'])
3782 numeric_comparison = None
3783 if isinstance(actual_value, (int, float)):
3784 # If the original field is a string and matching comparisonvalue is
3785 # a number we should respect the origin of the original field
3786 # and process comparison value as a string (see
3787 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3789 numeric_comparison = int(comparison_value)
3791 numeric_comparison = parse_filesize(comparison_value)
3792 if numeric_comparison is None:
3793 numeric_comparison = parse_filesize(f'{comparison_value}B')
3794 if numeric_comparison is None:
3795 numeric_comparison = parse_duration(comparison_value)
3796 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3797 raise ValueError('Operator %s only supports string values!' % m['op'])
3798 if actual_value is None:
3799 return is_incomplete(m['key']) or m['none_inclusive']
3800 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3803 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3804 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3806 operator_rex = re.compile(r'''(?x
)
3807 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3808 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3809 m = operator_rex.fullmatch(filter_part.strip())
3811 op = UNARY_OPERATORS[m.group('op')]
3812 actual_value = dct.get(m.group('key'))
3813 if is_incomplete(m.group('key')) and actual_value is None:
3815 return op(actual_value)
3817 raise ValueError('Invalid filter part %r' % filter_part)
3820 def match_str(filter_str, dct, incomplete=False):
3821 """ Filter a dictionary with a simple string syntax.
3822 @returns Whether the filter passes
3823 @param incomplete Set of keys that is expected to be missing from dct.
3824 Can be True/False to indicate all/none of the keys may be missing.
3825 All conditions on incomplete keys pass if the key is missing
3828 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3829 for filter_part in re.split(r'(?<!\\)&', filter_str))
3832 def match_filter_func(filters):
3835 filters = set(variadic(filters))
3837 interactive = '-' in filters
3841 def _match_func(info_dict, incomplete=False):
3842 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3843 return NO_DEFAULT if interactive and not incomplete else None
3845 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3846 filter_str = ') | ('.join(map(str.strip, filters))
3847 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3851 class download_range_func:
3852 def __init__(self, chapters, ranges):
3853 self.chapters, self.ranges = chapters, ranges
3855 def __call__(self, info_dict, ydl):
3856 if not self.ranges and not self.chapters:
3859 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3860 else 'Cannot match chapters since chapter information is unavailable')
3861 for regex in self.chapters or []:
3862 for i, chapter in enumerate(info_dict.get('chapters') or []):
3863 if re.search(regex, chapter['title']):
3865 yield {**chapter, 'index': i}
3866 if self.chapters and warning:
3867 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3869 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3871 def __eq__(self, other):
3872 return (isinstance(other, download_range_func)
3873 and self.chapters == other.chapters and self.ranges == other.ranges)
3876 def parse_dfxp_time_expr(time_expr):
3880 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3882 return float(mobj.group('time_offset'))
3884 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3886 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3889 def srt_subtitles_timecode(seconds):
3890 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3893 def ass_subtitles_timecode(seconds):
3894 time = timetuple_from_msec(seconds * 1000)
3895 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3898 def dfxp2srt(dfxp_data):
3900 @param dfxp_data A
bytes-like
object containing DFXP data
3901 @returns A
unicode object containing converted SRT data
3903 LEGACY_NAMESPACES = (
3904 (b'http://www.w3.org/ns/ttml', [
3905 b'http://www.w3.org/2004/11/ttaf1',
3906 b'http://www.w3.org/2006/04/ttaf1',
3907 b'http://www.w3.org/2006/10/ttaf1',
3909 (b'http://www.w3.org/ns/ttml#styling', [
3910 b'http://www.w3.org/ns/ttml#style',
3914 SUPPORTED_STYLING = [
3923 _x = functools.partial(xpath_with_ns, ns_map={
3924 'xml': 'http://www.w3.org/XML/1998/namespace',
3925 'ttml': 'http://www.w3.org/ns/ttml',
3926 'tts': 'http://www.w3.org/ns/ttml#styling',
3932 class TTMLPElementParser:
3934 _unclosed_elements = []
3935 _applied_styles = []
3937 def start(self, tag, attrib):
3938 if tag in (_x('ttml:br'), 'br'):
3941 unclosed_elements = []
3943 element_style_id = attrib.get('style')
3945 style.update(default_style)
3946 if element_style_id:
3947 style.update(styles.get(element_style_id, {}))
3948 for prop in SUPPORTED_STYLING:
3949 prop_val = attrib.get(_x('tts:' + prop))
3951 style[prop] = prop_val
3954 for k, v in sorted(style.items()):
3955 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3958 font += ' color="%s"' % v
3959 elif k == 'fontSize':
3960 font += ' size="%s"' % v
3961 elif k == 'fontFamily':
3962 font += ' face="%s"' % v
3963 elif k == 'fontWeight' and v == 'bold':
3965 unclosed_elements.append('b')
3966 elif k == 'fontStyle' and v == 'italic':
3968 unclosed_elements.append('i')
3969 elif k == 'textDecoration' and v == 'underline':
3971 unclosed_elements.append('u')
3973 self._out += '<font' + font + '>'
3974 unclosed_elements.append('font')
3976 if self._applied_styles:
3977 applied_style.update(self._applied_styles[-1])
3978 applied_style.update(style)
3979 self._applied_styles.append(applied_style)
3980 self._unclosed_elements.append(unclosed_elements)
3983 if tag not in (_x('ttml:br'), 'br'):
3984 unclosed_elements = self._unclosed_elements.pop()
3985 for element in reversed(unclosed_elements):
3986 self._out += '</%s>' % element
3987 if unclosed_elements and self._applied_styles:
3988 self._applied_styles.pop()
3990 def data(self, data):
3994 return self._out.strip()
3996 def parse_node(node):
3997 target = TTMLPElementParser()
3998 parser = xml.etree.ElementTree.XMLParser(target=target)
3999 parser.feed(xml.etree.ElementTree.tostring(node))
4000 return parser.close()
4002 for k, v in LEGACY_NAMESPACES:
4004 dfxp_data = dfxp_data.replace(ns, k)
4006 dfxp = compat_etree_fromstring(dfxp_data)
4008 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
4011 raise ValueError('Invalid dfxp/TTML subtitle')
4015 for style in dfxp.findall(_x('.//ttml:style')):
4016 style_id = style.get('id') or style.get(_x('xml:id'))
4019 parent_style_id = style.get('style')
4021 if parent_style_id not in styles:
4024 styles[style_id] = styles[parent_style_id].copy()
4025 for prop in SUPPORTED_STYLING:
4026 prop_val = style.get(_x('tts:' + prop))
4028 styles.setdefault(style_id, {})[prop] = prop_val
4034 for p in ('body', 'div'):
4035 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
4038 style = styles.get(ele.get('style'))
4041 default_style.update(style)
4043 for para, index in zip(paras, itertools.count(1)):
4044 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
4045 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
4046 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
4047 if begin_time is None:
4052 end_time = begin_time + dur
4053 out.append('%d\n%s --> %s\n%s\n\n' % (
4055 srt_subtitles_timecode(begin_time),
4056 srt_subtitles_timecode(end_time),
4062 def cli_option(params, command_option, param, separator=None):
4063 param = params.get(param)
4064 return ([] if param is None
4065 else [command_option, str(param)] if separator is None
4066 else [f'{command_option}{separator}{param}'])
4069 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4070 param = params.get(param)
4071 assert param in (True, False, None)
4072 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4075 def cli_valueless_option(params, command_option, param, expected_value=True):
4076 return [command_option] if params.get(param) == expected_value else []
4079 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4080 if isinstance(argdict, (list, tuple)): # for backward compatibility
4087 assert isinstance(argdict, dict)
4089 assert isinstance(keys, (list, tuple))
4090 for key_list in keys:
4091 arg_list = list(filter(
4092 lambda x: x is not None,
4093 [argdict.get(key.lower()) for key in variadic(key_list)]))
4095 return [arg for args in arg_list for arg in args]
4099 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4100 main_key, exe = main_key.lower(), exe.lower()
4101 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4102 keys = [f'{root_key}{k}' for k in (keys or [''])]
4103 if root_key in keys:
4105 keys.append((main_key, exe))
4106 keys.append('default')
4109 return cli_configuration_args(argdict, keys, default, use_compat)
4113 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4172 'iw': 'heb', # Replaced by he in 1989 revision
4182 'in': 'ind', # Replaced by id in 1989 revision
4297 'ji': 'yid', # Replaced by yi in 1989 revision
4305 def short2long(cls, code):
4306 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4307 return cls._lang_map.get(code[:2])
4310 def long2short(cls, code):
4311 """Convert language code from ISO 639-2/T to ISO 639-1"""
4312 for short_name, long_name in cls._lang_map.items():
4313 if long_name == code:
4318 # From http://data.okfn.org/data/core/country-list
4320 'AF': 'Afghanistan',
4321 'AX': 'Åland Islands',
4324 'AS': 'American Samoa',
4329 'AG': 'Antigua and Barbuda',
4346 'BO': 'Bolivia, Plurinational State of',
4347 'BQ': 'Bonaire, Sint Eustatius and Saba',
4348 'BA': 'Bosnia and Herzegovina',
4350 'BV': 'Bouvet Island',
4352 'IO': 'British Indian Ocean Territory',
4353 'BN': 'Brunei Darussalam',
4355 'BF': 'Burkina Faso',
4361 'KY': 'Cayman Islands',
4362 'CF': 'Central African Republic',
4366 'CX': 'Christmas Island',
4367 'CC': 'Cocos (Keeling) Islands',
4371 'CD': 'Congo, the Democratic Republic of the',
4372 'CK': 'Cook Islands',
4374 'CI': 'Côte d\'Ivoire',
4379 'CZ': 'Czech Republic',
4383 'DO': 'Dominican Republic',
4386 'SV': 'El Salvador',
4387 'GQ': 'Equatorial Guinea',
4391 'FK': 'Falkland Islands (Malvinas)',
4392 'FO': 'Faroe Islands',
4396 'GF': 'French Guiana',
4397 'PF': 'French Polynesia',
4398 'TF': 'French Southern Territories',
4413 'GW': 'Guinea-Bissau',
4416 'HM': 'Heard Island and McDonald Islands',
4417 'VA': 'Holy See (Vatican City State)',
4424 'IR': 'Iran, Islamic Republic of',
4427 'IM': 'Isle of Man',
4437 'KP': 'Korea, Democratic People\'s Republic of',
4438 'KR': 'Korea, Republic of',
4441 'LA': 'Lao People\'s Democratic Republic',
4447 'LI': 'Liechtenstein',
4451 'MK': 'Macedonia, the Former Yugoslav Republic of',
4458 'MH': 'Marshall Islands',
4464 'FM': 'Micronesia, Federated States of',
4465 'MD': 'Moldova, Republic of',
4476 'NL': 'Netherlands',
4477 'NC': 'New Caledonia',
4478 'NZ': 'New Zealand',
4483 'NF': 'Norfolk Island',
4484 'MP': 'Northern Mariana Islands',
4489 'PS': 'Palestine, State of',
4491 'PG': 'Papua New Guinea',
4494 'PH': 'Philippines',
4498 'PR': 'Puerto Rico',
4502 'RU': 'Russian Federation',
4504 'BL': 'Saint Barthélemy',
4505 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4506 'KN': 'Saint Kitts and Nevis',
4507 'LC': 'Saint Lucia',
4508 'MF': 'Saint Martin (French part)',
4509 'PM': 'Saint Pierre and Miquelon',
4510 'VC': 'Saint Vincent and the Grenadines',
4513 'ST': 'Sao Tome and Principe',
4514 'SA': 'Saudi Arabia',
4518 'SL': 'Sierra Leone',
4520 'SX': 'Sint Maarten (Dutch part)',
4523 'SB': 'Solomon Islands',
4525 'ZA': 'South Africa',
4526 'GS': 'South Georgia and the South Sandwich Islands',
4527 'SS': 'South Sudan',
4532 'SJ': 'Svalbard and Jan Mayen',
4535 'CH': 'Switzerland',
4536 'SY': 'Syrian Arab Republic',
4537 'TW': 'Taiwan, Province of China',
4539 'TZ': 'Tanzania, United Republic of',
4541 'TL': 'Timor-Leste',
4545 'TT': 'Trinidad and Tobago',
4548 'TM': 'Turkmenistan',
4549 'TC': 'Turks and Caicos Islands',
4553 'AE': 'United Arab Emirates',
4554 'GB': 'United Kingdom',
4555 'US': 'United States',
4556 'UM': 'United States Minor Outlying Islands',
4560 'VE': 'Venezuela, Bolivarian Republic of',
4562 'VG': 'Virgin Islands, British',
4563 'VI': 'Virgin Islands, U.S.',
4564 'WF': 'Wallis and Futuna',
4565 'EH': 'Western Sahara',
4569 # Not ISO 3166 codes, but used for IP blocks
4570 'AP': 'Asia/Pacific Region',
4575 def short2full(cls, code):
4576 """Convert an ISO 3166-2 country code to the corresponding full name"""
4577 return cls._country_map.get(code.upper())
4581 # Major IPv4 address blocks per country
4583 'AD': '46.172.224.0/19',
4584 'AE': '94.200.0.0/13',
4585 'AF': '149.54.0.0/17',
4586 'AG': '209.59.64.0/18',
4587 'AI': '204.14.248.0/21',
4588 'AL': '46.99.0.0/16',
4589 'AM': '46.70.0.0/15',
4590 'AO': '105.168.0.0/13',
4591 'AP': '182.50.184.0/21',
4592 'AQ': '23.154.160.0/24',
4593 'AR': '181.0.0.0/12',
4594 'AS': '202.70.112.0/20',
4595 'AT': '77.116.0.0/14',
4596 'AU': '1.128.0.0/11',
4597 'AW': '181.41.0.0/18',
4598 'AX': '185.217.4.0/22',
4599 'AZ': '5.197.0.0/16',
4600 'BA': '31.176.128.0/17',
4601 'BB': '65.48.128.0/17',
4602 'BD': '114.130.0.0/16',
4604 'BF': '102.178.0.0/15',
4605 'BG': '95.42.0.0/15',
4606 'BH': '37.131.0.0/17',
4607 'BI': '154.117.192.0/18',
4608 'BJ': '137.255.0.0/16',
4609 'BL': '185.212.72.0/23',
4610 'BM': '196.12.64.0/18',
4611 'BN': '156.31.0.0/16',
4612 'BO': '161.56.0.0/16',
4613 'BQ': '161.0.80.0/20',
4614 'BR': '191.128.0.0/12',
4615 'BS': '24.51.64.0/18',
4616 'BT': '119.2.96.0/19',
4617 'BW': '168.167.0.0/16',
4618 'BY': '178.120.0.0/13',
4619 'BZ': '179.42.192.0/18',
4620 'CA': '99.224.0.0/11',
4621 'CD': '41.243.0.0/16',
4622 'CF': '197.242.176.0/21',
4623 'CG': '160.113.0.0/16',
4624 'CH': '85.0.0.0/13',
4625 'CI': '102.136.0.0/14',
4626 'CK': '202.65.32.0/19',
4627 'CL': '152.172.0.0/14',
4628 'CM': '102.244.0.0/14',
4629 'CN': '36.128.0.0/10',
4630 'CO': '181.240.0.0/12',
4631 'CR': '201.192.0.0/12',
4632 'CU': '152.206.0.0/15',
4633 'CV': '165.90.96.0/19',
4634 'CW': '190.88.128.0/17',
4635 'CY': '31.153.0.0/16',
4636 'CZ': '88.100.0.0/14',
4638 'DJ': '197.241.0.0/17',
4639 'DK': '87.48.0.0/12',
4640 'DM': '192.243.48.0/20',
4641 'DO': '152.166.0.0/15',
4642 'DZ': '41.96.0.0/12',
4643 'EC': '186.68.0.0/15',
4644 'EE': '90.190.0.0/15',
4645 'EG': '156.160.0.0/11',
4646 'ER': '196.200.96.0/20',
4647 'ES': '88.0.0.0/11',
4648 'ET': '196.188.0.0/14',
4649 'EU': '2.16.0.0/13',
4650 'FI': '91.152.0.0/13',
4651 'FJ': '144.120.0.0/16',
4652 'FK': '80.73.208.0/21',
4653 'FM': '119.252.112.0/20',
4654 'FO': '88.85.32.0/19',
4656 'GA': '41.158.0.0/15',
4658 'GD': '74.122.88.0/21',
4659 'GE': '31.146.0.0/16',
4660 'GF': '161.22.64.0/18',
4661 'GG': '62.68.160.0/19',
4662 'GH': '154.160.0.0/12',
4663 'GI': '95.164.0.0/16',
4664 'GL': '88.83.0.0/19',
4665 'GM': '160.182.0.0/15',
4666 'GN': '197.149.192.0/18',
4667 'GP': '104.250.0.0/19',
4668 'GQ': '105.235.224.0/20',
4669 'GR': '94.64.0.0/13',
4670 'GT': '168.234.0.0/16',
4671 'GU': '168.123.0.0/16',
4672 'GW': '197.214.80.0/20',
4673 'GY': '181.41.64.0/18',
4674 'HK': '113.252.0.0/14',
4675 'HN': '181.210.0.0/16',
4676 'HR': '93.136.0.0/13',
4677 'HT': '148.102.128.0/17',
4678 'HU': '84.0.0.0/14',
4679 'ID': '39.192.0.0/10',
4680 'IE': '87.32.0.0/12',
4681 'IL': '79.176.0.0/13',
4682 'IM': '5.62.80.0/20',
4683 'IN': '117.192.0.0/10',
4684 'IO': '203.83.48.0/21',
4685 'IQ': '37.236.0.0/14',
4686 'IR': '2.176.0.0/12',
4687 'IS': '82.221.0.0/16',
4688 'IT': '79.0.0.0/10',
4689 'JE': '87.244.64.0/18',
4690 'JM': '72.27.0.0/17',
4691 'JO': '176.29.0.0/16',
4692 'JP': '133.0.0.0/8',
4693 'KE': '105.48.0.0/12',
4694 'KG': '158.181.128.0/17',
4695 'KH': '36.37.128.0/17',
4696 'KI': '103.25.140.0/22',
4697 'KM': '197.255.224.0/20',
4698 'KN': '198.167.192.0/19',
4699 'KP': '175.45.176.0/22',
4700 'KR': '175.192.0.0/10',
4701 'KW': '37.36.0.0/14',
4702 'KY': '64.96.0.0/15',
4703 'KZ': '2.72.0.0/13',
4704 'LA': '115.84.64.0/18',
4705 'LB': '178.135.0.0/16',
4706 'LC': '24.92.144.0/20',
4707 'LI': '82.117.0.0/19',
4708 'LK': '112.134.0.0/15',
4709 'LR': '102.183.0.0/16',
4710 'LS': '129.232.0.0/17',
4711 'LT': '78.56.0.0/13',
4712 'LU': '188.42.0.0/16',
4713 'LV': '46.109.0.0/16',
4714 'LY': '41.252.0.0/14',
4715 'MA': '105.128.0.0/11',
4716 'MC': '88.209.64.0/18',
4717 'MD': '37.246.0.0/16',
4718 'ME': '178.175.0.0/17',
4719 'MF': '74.112.232.0/21',
4720 'MG': '154.126.0.0/17',
4721 'MH': '117.103.88.0/21',
4722 'MK': '77.28.0.0/15',
4723 'ML': '154.118.128.0/18',
4724 'MM': '37.111.0.0/17',
4725 'MN': '49.0.128.0/17',
4726 'MO': '60.246.0.0/16',
4727 'MP': '202.88.64.0/20',
4728 'MQ': '109.203.224.0/19',
4729 'MR': '41.188.64.0/18',
4730 'MS': '208.90.112.0/22',
4731 'MT': '46.11.0.0/16',
4732 'MU': '105.16.0.0/12',
4733 'MV': '27.114.128.0/18',
4734 'MW': '102.70.0.0/15',
4735 'MX': '187.192.0.0/11',
4736 'MY': '175.136.0.0/13',
4737 'MZ': '197.218.0.0/15',
4738 'NA': '41.182.0.0/16',
4739 'NC': '101.101.0.0/18',
4740 'NE': '197.214.0.0/18',
4741 'NF': '203.17.240.0/22',
4742 'NG': '105.112.0.0/12',
4743 'NI': '186.76.0.0/15',
4744 'NL': '145.96.0.0/11',
4745 'NO': '84.208.0.0/13',
4746 'NP': '36.252.0.0/15',
4747 'NR': '203.98.224.0/19',
4748 'NU': '49.156.48.0/22',
4749 'NZ': '49.224.0.0/14',
4750 'OM': '5.36.0.0/15',
4751 'PA': '186.72.0.0/15',
4752 'PE': '186.160.0.0/14',
4753 'PF': '123.50.64.0/18',
4754 'PG': '124.240.192.0/19',
4755 'PH': '49.144.0.0/13',
4756 'PK': '39.32.0.0/11',
4757 'PL': '83.0.0.0/11',
4758 'PM': '70.36.0.0/20',
4759 'PR': '66.50.0.0/16',
4760 'PS': '188.161.0.0/16',
4761 'PT': '85.240.0.0/13',
4762 'PW': '202.124.224.0/20',
4763 'PY': '181.120.0.0/14',
4764 'QA': '37.210.0.0/15',
4765 'RE': '102.35.0.0/16',
4766 'RO': '79.112.0.0/13',
4767 'RS': '93.86.0.0/15',
4768 'RU': '5.136.0.0/13',
4769 'RW': '41.186.0.0/16',
4770 'SA': '188.48.0.0/13',
4771 'SB': '202.1.160.0/19',
4772 'SC': '154.192.0.0/11',
4773 'SD': '102.120.0.0/13',
4774 'SE': '78.64.0.0/12',
4775 'SG': '8.128.0.0/10',
4776 'SI': '188.196.0.0/14',
4777 'SK': '78.98.0.0/15',
4778 'SL': '102.143.0.0/17',
4779 'SM': '89.186.32.0/19',
4780 'SN': '41.82.0.0/15',
4781 'SO': '154.115.192.0/18',
4782 'SR': '186.179.128.0/17',
4783 'SS': '105.235.208.0/21',
4784 'ST': '197.159.160.0/19',
4785 'SV': '168.243.0.0/16',
4786 'SX': '190.102.0.0/20',
4788 'SZ': '41.84.224.0/19',
4789 'TC': '65.255.48.0/20',
4790 'TD': '154.68.128.0/19',
4791 'TG': '196.168.0.0/14',
4792 'TH': '171.96.0.0/13',
4793 'TJ': '85.9.128.0/18',
4794 'TK': '27.96.24.0/21',
4795 'TL': '180.189.160.0/20',
4796 'TM': '95.85.96.0/19',
4797 'TN': '197.0.0.0/11',
4798 'TO': '175.176.144.0/21',
4799 'TR': '78.160.0.0/11',
4800 'TT': '186.44.0.0/15',
4801 'TV': '202.2.96.0/19',
4802 'TW': '120.96.0.0/11',
4803 'TZ': '156.156.0.0/14',
4804 'UA': '37.52.0.0/14',
4805 'UG': '102.80.0.0/13',
4807 'UY': '167.56.0.0/13',
4808 'UZ': '84.54.64.0/18',
4809 'VA': '212.77.0.0/19',
4810 'VC': '207.191.240.0/21',
4811 'VE': '186.88.0.0/13',
4812 'VG': '66.81.192.0/20',
4813 'VI': '146.226.0.0/16',
4814 'VN': '14.160.0.0/11',
4815 'VU': '202.80.32.0/20',
4816 'WF': '117.20.32.0/21',
4817 'WS': '202.4.32.0/19',
4818 'YE': '134.35.0.0/16',
4819 'YT': '41.242.116.0/22',
4820 'ZA': '41.0.0.0/11',
4821 'ZM': '102.144.0.0/13',
4822 'ZW': '102.177.192.0/18',
4826 def random_ipv4(cls, code_or_block):
4827 if len(code_or_block) == 2:
4828 block = cls._country_ip_map.get(code_or_block.upper())
4832 block = code_or_block
4833 addr, preflen = block.split('/')
4834 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4835 addr_max = addr_min | (0xffffffff >> int(preflen))
4836 return str(socket.inet_ntoa(
4837 struct.pack('!L', random.randint(addr_min, addr_max))))
4840 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4841 def __init__(self, proxies=None):
4842 # Set default handlers
4843 for type in ('http', 'https'):
4844 setattr(self, '%s_open' % type,
4845 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4846 meth(r, proxy, type))
4847 urllib.request.ProxyHandler.__init__(self, proxies)
4849 def proxy_open(self, req, proxy, type):
4850 req_proxy = req.headers.get('Ytdl-request-proxy')
4851 if req_proxy is not None:
4853 del req.headers['Ytdl-request-proxy']
4855 if proxy == '__noproxy__':
4856 return None # No Proxy
4857 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4858 req.add_header('Ytdl-socks-proxy', proxy)
4859 # yt-dlp's http/https handlers do wrapping the socket with socks
4861 return urllib.request.ProxyHandler.proxy_open(
4862 self, req, proxy, type)
4865 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4866 # released into Public Domain
4867 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4869 def long_to_bytes(n, blocksize=0):
4870 """long_to_bytes(n:long, blocksize:int) : string
4871 Convert a long integer to a byte string.
4873 If optional blocksize is given and greater than zero, pad the front of the
4874 byte string with binary zeros so that the length is a multiple of
4877 # after much testing, this algorithm was deemed to be the fastest
4881 s = struct.pack('>I', n & 0xffffffff) + s
4883 # strip off leading zeros
4884 for i in range(len(s)):
4885 if s[i] != b'\000'[0]:
4888 # only happens when n == 0
4892 # add back some pad bytes. this could be done more efficiently w.r.t. the
4893 # de-padding being done above, but sigh...
4894 if blocksize > 0 and len(s) % blocksize:
4895 s = (blocksize - len(s) % blocksize) * b'\000' + s
4899 def bytes_to_long(s):
4900 """bytes_to_long(string) : long
4901 Convert a byte string to a long integer.
4903 This is (essentially) the inverse of long_to_bytes().
4908 extra = (4 - length % 4)
4909 s = b'\000' * extra + s
4910 length = length + extra
4911 for i in range(0, length, 4):
4912 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4916 def ohdave_rsa_encrypt(data, exponent, modulus):
4918 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4921 data: data to encrypt, bytes-like object
4922 exponent, modulus: parameter e and N of RSA algorithm, both integer
4923 Output: hex string of encrypted data
4925 Limitation: supports one block encryption only
4928 payload = int(binascii.hexlify(data[::-1]), 16)
4929 encrypted = pow(payload, exponent, modulus)
4930 return '%x' % encrypted
4933 def pkcs1pad(data, length):
4935 Padding input data with PKCS#1 scheme
4937 @param {int[]} data input data
4938 @param {int} length target length
4939 @returns {int[]} padded data
4941 if len(data) > length - 11:
4942 raise ValueError('Input data too
long for PKCS
#1 padding')
4944 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4945 return [0, 2] + pseudo_random
+ [0] + data
4948 def _base_n_table(n
, table
):
4949 if not table
and not n
:
4950 raise ValueError('Either table or n must be specified')
4951 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4953 if n
and n
!= len(table
):
4954 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4958 def encode_base_n(num
, n
=None, table
=None):
4959 """Convert given int to a base-n string"""
4960 table
= _base_n_table(n
, table
)
4964 result
, base
= '', len(table
)
4966 result
= table
[num
% base
] + result
4971 def decode_base_n(string
, n
=None, table
=None):
4972 """Convert given base-n string to int"""
4973 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4974 result
, base
= 0, len(table
)
4976 result
= result
* base
+ table
[char
]
4980 def decode_base(value
, digits
):
4981 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
4982 f
'in a future version. Use {__name__}.decode_base_n instead')
4983 return decode_base_n(value
, table
=digits
)
4986 def decode_packed_codes(code
):
4987 mobj
= re
.search(PACKED_CODES_RE
, code
)
4988 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4991 symbols
= symbols
.split('|')
4996 base_n_count
= encode_base_n(count
, base
)
4997 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
5000 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
5004 def caesar(s
, alphabet
, shift
):
5009 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
5014 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
5017 def parse_m3u8_attributes(attrib
):
5019 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5020 if val
.startswith('"'):
5026 def urshift(val
, n
):
5027 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5030 # Based on png2str() written by @gdkchan and improved by @yokrysty
5031 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5032 def decode_png(png_data
):
5033 # Reference: https://www.w3.org/TR/PNG/
5034 header
= png_data
[8:]
5036 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5037 raise OSError('Not a valid PNG file.')
5039 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5040 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5045 length
= unpack_integer(header
[:4])
5048 chunk_type
= header
[:4]
5051 chunk_data
= header
[:length
]
5052 header
= header
[length
:]
5054 header
= header
[4:] # Skip CRC
5062 ihdr
= chunks
[0]['data']
5064 width
= unpack_integer(ihdr
[:4])
5065 height
= unpack_integer(ihdr
[4:8])
5069 for chunk
in chunks
:
5070 if chunk
['type'] == b
'IDAT':
5071 idat
+= chunk
['data']
5074 raise OSError('Unable to read PNG data.')
5076 decompressed_data
= bytearray(zlib
.decompress(idat
))
5081 def _get_pixel(idx
):
5086 for y
in range(height
):
5087 basePos
= y
* (1 + stride
)
5088 filter_type
= decompressed_data
[basePos
]
5092 pixels
.append(current_row
)
5094 for x
in range(stride
):
5095 color
= decompressed_data
[1 + basePos
+ x
]
5096 basex
= y
* stride
+ x
5101 left
= _get_pixel(basex
- 3)
5103 up
= _get_pixel(basex
- stride
)
5105 if filter_type
== 1: # Sub
5106 color
= (color
+ left
) & 0xff
5107 elif filter_type
== 2: # Up
5108 color
= (color
+ up
) & 0xff
5109 elif filter_type
== 3: # Average
5110 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5111 elif filter_type
== 4: # Paeth
5117 c
= _get_pixel(basex
- stride
- 3)
5125 if pa
<= pb
and pa
<= pc
:
5126 color
= (color
+ a
) & 0xff
5128 color
= (color
+ b
) & 0xff
5130 color
= (color
+ c
) & 0xff
5132 current_row
.append(color
)
5134 return width
, height
, pixels
5137 def write_xattr(path
, key
, value
):
5138 # Windows: Write xattrs to NTFS Alternate Data Streams:
5139 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5140 if compat_os_name
== 'nt':
5141 assert ':' not in key
5142 assert os
.path
.exists(path
)
5145 with open(f
'{path}:{key}', 'wb') as f
:
5147 except OSError as e
:
5148 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5151 # UNIX Method 1. Use xattrs/pyxattrs modules
5154 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5155 # Unicode arguments are not supported in pyxattr until version 0.5.0
5156 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5157 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5158 setxattr
= xattr
.set
5160 setxattr
= xattr
.setxattr
5164 setxattr(path
, key
, value
)
5165 except OSError as e
:
5166 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5169 # UNIX Method 2. Use setfattr/xattr executables
5170 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5171 else 'xattr' if check_executable('xattr', ['-h']) else None)
5173 raise XAttrUnavailableError(
5174 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5175 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5177 value
= value
.decode()
5179 _
, stderr
, returncode
= Popen
.run(
5180 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5181 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5182 except OSError as e
:
5183 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5185 raise XAttrMetadataError(returncode
, stderr
)
5188 def random_birthday(year_field
, month_field
, day_field
):
5189 start_date
= datetime
.date(1950, 1, 1)
5190 end_date
= datetime
.date(1995, 12, 31)
5191 offset
= random
.randint(0, (end_date
- start_date
).days
)
5192 random_date
= start_date
+ datetime
.timedelta(offset
)
5194 year_field
: str(random_date
.year
),
5195 month_field
: str(random_date
.month
),
5196 day_field
: str(random_date
.day
),
5200 # Templates for internet shortcut files, which are plain text files.
5201 DOT_URL_LINK_TEMPLATE
= '''\
5206 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5207 <?xml version="1.0" encoding="UTF-8"?>
5208 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5209 <plist version="1.0">
5212 \t<string>%(url)s</string>
5217 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5227 'url': DOT_URL_LINK_TEMPLATE
,
5228 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5229 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5233 def iri_to_uri(iri
):
5235 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5237 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.
5240 iri_parts
= urllib
.parse
.urlparse(iri
)
5242 if '[' in iri_parts
.netloc
:
5243 raise ValueError('IPv6 URIs are not, yet, supported.')
5244 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5246 # 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.
5249 if iri_parts
.username
:
5250 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5251 if iri_parts
.password
is not None:
5252 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5255 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5256 # The 'idna' encoding produces ASCII text.
5257 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5258 net_location
+= ':' + str(iri_parts
.port
)
5260 return urllib
.parse
.urlunparse(
5264 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5266 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5267 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5269 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5270 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5272 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5274 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5277 def to_high_limit_path(path
):
5278 if sys
.platform
in ['win32', 'cygwin']:
5279 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5280 return '\\\\?\\' + os
.path
.abspath(path
)
5285 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5286 val
= traverse_obj(obj
, *variadic(field
))
5287 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5289 return template
% func(val
)
5292 def clean_podcast_url(url
):
5293 return re
.sub(r
'''(?x)
5297 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5300 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5303 cn\.co| # https://podcorn.com/analytics-prefix/
5304 st\.fm # https://podsights.com/docs/
5309 _HEX_TABLE
= '0123456789abcdef'
5312 def random_uuidv4():
5313 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5316 def make_dir(path
, to_screen
=None):
5318 dn
= os
.path
.dirname(path
)
5319 if dn
and not os
.path
.exists(dn
):
5322 except OSError as err
:
5323 if callable(to_screen
) is not None:
5324 to_screen('unable to create directory ' + error_to_compat_str(err
))
5328 def get_executable_path():
5329 from .update
import _get_variant_and_executable_path
5331 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5334 def load_plugins(name
, suffix
, namespace
):
5336 with contextlib
.suppress(FileNotFoundError
):
5337 plugins_spec
= importlib
.util
.spec_from_file_location(
5338 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5339 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5340 sys
.modules
[plugins_spec
.name
] = plugins
5341 plugins_spec
.loader
.exec_module(plugins
)
5342 for name
in dir(plugins
):
5343 if name
in namespace
:
5345 if not name
.endswith(suffix
):
5347 klass
= getattr(plugins
, name
)
5348 classes
[name
] = namespace
[name
] = klass
5353 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5354 casesense
=True, is_user_input
=False, traverse_string
=False):
5356 Safely traverse nested `dict`s and `Sequence`s
5358 >>> obj = [{}, {"key": "value"}]
5359 >>> traverse_obj(obj, (1, "key"))
5362 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5363 The next path will also be tested if the path branched but no results could be found.
5364 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5365 A value of None is treated as the absence of a value.
5367 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5369 The keys in the path can be one of:
5370 - `None`: Return the current object.
5371 - `str`/`int`: Return `obj[key]`. For `re.Match, return `obj.group(key)`.
5372 - `slice`: Branch out and return all values in `obj[key]`.
5373 - `Ellipsis`: Branch out and return a list of all values.
5374 - `tuple`/`list`: Branch out and return a list of all matching values.
5375 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5376 - `function`: Branch out and return values filtered by the function.
5377 Read as: `[value for key, value in obj if function(key, value)]`.
5378 For `Sequence`s, `key` is the index of the value.
5379 - `dict` Transform the current object and return a matching dict.
5380 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5382 `tuple`, `list`, and `dict` all support nested paths and branches.
5384 @params paths Paths which to traverse by.
5385 @param default Value to return if the paths do not match.
5386 @param expected_type If a `type`, only accept final values of this type.
5387 If any other callable, try to call the function on each result.
5388 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5389 @param casesense If `False`, consider string dictionary keys as case insensitive.
5391 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5393 @param is_user_input Whether the keys are generated from user input.
5394 If `True` strings get converted to `int`/`slice` if needed.
5395 @param traverse_string Whether to traverse into objects as strings.
5396 If `True`, any non-compatible object will first be
5397 converted into a string and then traversed into.
5400 @returns The result of the object traversal.
5401 If successful, `get_all=True`, and the path branches at least once,
5402 then a list of results is returned instead.
5403 A list is always returned if the last path branches and no `default` is given.
5405 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5406 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5408 if isinstance(expected_type
, type):
5409 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5411 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5413 def apply_key(key
, obj
):
5420 elif isinstance(key
, (list, tuple)):
5422 _
, result
= apply_path(obj
, branch
)
5426 if isinstance(obj
, collections
.abc
.Mapping
):
5427 yield from obj
.values()
5428 elif is_sequence(obj
):
5430 elif isinstance(obj
, re
.Match
):
5431 yield from obj
.groups()
5432 elif traverse_string
:
5436 if is_sequence(obj
):
5437 iter_obj
= enumerate(obj
)
5438 elif isinstance(obj
, collections
.abc
.Mapping
):
5439 iter_obj
= obj
.items()
5440 elif isinstance(obj
, re
.Match
):
5441 iter_obj
= enumerate((obj
.group(), *obj
.groups()))
5442 elif traverse_string
:
5443 iter_obj
= enumerate(str(obj
))
5446 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5448 elif isinstance(key
, dict):
5449 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5450 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5451 if v
is not None or default
is not NO_DEFAULT
}
5453 elif isinstance(obj
, collections
.abc
.Mapping
):
5454 yield (obj
.get(key
) if casesense
or (key
in obj
)
5455 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5457 elif isinstance(obj
, re
.Match
):
5458 if isinstance(key
, int) or casesense
:
5459 with contextlib
.suppress(IndexError):
5460 yield obj
.group(key
)
5463 if not isinstance(key
, str):
5466 yield next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5470 key
= (int_or_none(key
) if ':' not in key
5471 else slice(*map(int_or_none
, key
.split(':'))))
5473 if not isinstance(key
, (int, slice)):
5476 if not is_sequence(obj
):
5477 if not traverse_string
:
5481 with contextlib
.suppress(IndexError):
5484 def apply_path(start_obj
, path
):
5486 has_branched
= False
5488 for key
in variadic(path
):
5489 if is_user_input
and key
== ':':
5492 if not casesense
and isinstance(key
, str):
5493 key
= key
.casefold()
5495 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5498 key_func
= functools
.partial(apply_key
, key
)
5499 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5501 return has_branched
, objs
5503 def _traverse_obj(obj
, path
, use_list
=True):
5504 has_branched
, results
= apply_path(obj
, path
)
5505 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5507 if get_all
and has_branched
:
5508 return results
.exhaust() if results
or use_list
else None
5510 return results
[0] if results
else None
5512 for index
, path
in enumerate(paths
, 1):
5513 use_list
= default
is NO_DEFAULT
and index
== len(paths
)
5514 result
= _traverse_obj(obj
, path
, use_list
)
5515 if result
is not None:
5518 return None if default
is NO_DEFAULT
else default
5521 def traverse_dict(dictn
, keys
, casesense
=True):
5522 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5523 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5524 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5527 def get_first(obj
, keys
, **kwargs
):
5528 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5531 def time_seconds(**kwargs
):
5532 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5533 return t
.timestamp()
5536 # create a JSON Web Signature (jws) with HS256 algorithm
5537 # the resulting format is in JWS Compact Serialization
5538 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5539 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5540 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5546 header_data
.update(headers
)
5547 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5548 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5549 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5550 signature_b64
= base64
.b64encode(h
.digest())
5551 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5555 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5556 def jwt_decode_hs256(jwt
):
5557 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5558 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5559 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5563 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5567 def supports_terminal_sequences(stream
):
5568 if compat_os_name
== 'nt':
5569 if not WINDOWS_VT_MODE
:
5571 elif not os
.getenv('TERM'):
5574 return stream
.isatty()
5575 except BaseException
:
5579 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5580 if get_windows_version() < (10, 0, 10586):
5582 global WINDOWS_VT_MODE
5584 Popen
.run('', shell
=True)
5588 WINDOWS_VT_MODE
= True
5589 supports_terminal_sequences
.cache_clear()
5592 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5595 def remove_terminal_sequences(string
):
5596 return _terminal_sequences_re
.sub('', string
)
5599 def number_of_digits(number
):
5600 return len('%d' % number
)
5603 def join_nonempty(*values
, delim
='-', from_dict
=None):
5604 if from_dict
is not None:
5605 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5606 return delim
.join(map(str, filter(None, values
)))
5609 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5611 Find the largest format dimensions in terms of video width and, for each thumbnail:
5612 * Modify the URL: Match the width with the provided regex and replace with the former width
5615 This function is useful with video services that scale the provided thumbnails on demand
5617 _keys
= ('width', 'height')
5618 max_dimensions
= max(
5619 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5621 if not max_dimensions
[0]:
5625 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5626 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5627 for thumbnail
in thumbnails
5631 def parse_http_range(range):
5632 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5634 return None, None, None
5635 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5637 return None, None, None
5638 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5641 def read_stdin(what
):
5642 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5643 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5647 def determine_file_encoding(data
):
5649 Detect the text encoding used
5650 @returns (encoding, bytes to skip)
5653 # BOM marks are given priority over declarations
5654 for bom
, enc
in BOMS
:
5655 if data
.startswith(bom
):
5656 return enc
, len(bom
)
5658 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5659 # We ignore the endianness to get a good enough match
5660 data
= data
.replace(b
'\0', b
'')
5661 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5662 return mobj
.group(1).decode() if mobj
else None, 0
5669 __initialized
= False
5671 def __init__(self
, parser
, label
=None):
5672 self
.parser
, self
.label
= parser
, label
5673 self
._loaded
_paths
, self
.configs
= set(), []
5675 def init(self
, args
=None, filename
=None):
5676 assert not self
.__initialized
5677 self
.own_args
, self
.filename
= args
, filename
5678 return self
.load_configs()
5680 def load_configs(self
):
5683 location
= os
.path
.realpath(self
.filename
)
5684 directory
= os
.path
.dirname(location
)
5685 if location
in self
._loaded
_paths
:
5687 self
._loaded
_paths
.add(location
)
5689 self
.__initialized
= True
5690 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5691 self
.parsed_args
= self
.own_args
5692 for location
in opts
.config_locations
or []:
5694 if location
in self
._loaded
_paths
:
5696 self
._loaded
_paths
.add(location
)
5697 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5699 location
= os
.path
.join(directory
, expand_path(location
))
5700 if os
.path
.isdir(location
):
5701 location
= os
.path
.join(location
, 'yt-dlp.conf')
5702 if not os
.path
.exists(location
):
5703 self
.parser
.error(f
'config location {location} does not exist')
5704 self
.append_config(self
.read_file(location
), location
)
5708 label
= join_nonempty(
5709 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5711 return join_nonempty(
5712 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5713 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5717 def read_file(filename
, default
=[]):
5719 optionf
= open(filename
, 'rb')
5721 return default
# silently skip if file is not present
5723 enc
, skip
= determine_file_encoding(optionf
.read(512))
5724 optionf
.seek(skip
, io
.SEEK_SET
)
5726 enc
= None # silently skip read errors
5728 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5729 contents
= optionf
.read().decode(enc
or preferredencoding())
5730 res
= shlex
.split(contents
, comments
=True)
5731 except Exception as err
:
5732 raise ValueError(f
'Unable to parse "{filename}": {err}')
5738 def hide_login_info(opts
):
5739 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5740 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5745 return m
.group('key') + '=PRIVATE'
5749 opts
= list(map(_scrub_eq
, opts
))
5750 for idx
, opt
in enumerate(opts
):
5751 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5752 opts
[idx
+ 1] = 'PRIVATE'
5755 def append_config(self
, *args
, label
=None):
5756 config
= type(self
)(self
.parser
, label
)
5757 config
._loaded
_paths
= self
._loaded
_paths
5758 if config
.init(*args
):
5759 self
.configs
.append(config
)
5763 for config
in reversed(self
.configs
):
5764 yield from config
.all_args
5765 yield from self
.parsed_args
or []
5767 def parse_known_args(self
, **kwargs
):
5768 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5770 def parse_args(self
):
5771 return self
.parser
.parse_args(self
.all_args
)
5774 class WebSocketsWrapper
:
5775 """Wraps websockets module to use in non-async scopes"""
5778 def __init__(self
, url
, headers
=None, connect
=True):
5779 self
.loop
= asyncio
.new_event_loop()
5780 # XXX: "loop" is deprecated
5781 self
.conn
= websockets
.connect(
5782 url
, extra_headers
=headers
, ping_interval
=None,
5783 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5786 atexit
.register(self
.__exit
__, None, None, None)
5788 def __enter__(self
):
5790 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5793 def send(self
, *args
):
5794 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5796 def recv(self
, *args
):
5797 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5799 def __exit__(self
, type, value
, traceback
):
5801 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5804 self
._cancel
_all
_tasks
(self
.loop
)
5806 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5807 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5809 def run_with_loop(main
, loop
):
5810 if not asyncio
.iscoroutine(main
):
5811 raise ValueError(f
'a coroutine was expected, got {main!r}')
5814 return loop
.run_until_complete(main
)
5816 loop
.run_until_complete(loop
.shutdown_asyncgens())
5817 if hasattr(loop
, 'shutdown_default_executor'):
5818 loop
.run_until_complete(loop
.shutdown_default_executor())
5821 def _cancel_all_tasks(loop
):
5822 to_cancel
= asyncio
.all_tasks(loop
)
5827 for task
in to_cancel
:
5830 # XXX: "loop" is removed in python 3.10+
5831 loop
.run_until_complete(
5832 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5834 for task
in to_cancel
:
5835 if task
.cancelled():
5837 if task
.exception() is not None:
5838 loop
.call_exception_handler({
5839 'message': 'unhandled exception during asyncio.run() shutdown',
5840 'exception': task
.exception(),
5845 def merge_headers(*dicts
):
5846 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5847 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5850 def cached_method(f
):
5851 """Cache a method"""
5852 signature
= inspect
.signature(f
)
5855 def wrapper(self
, *args
, **kwargs
):
5856 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5857 bound_args
.apply_defaults()
5858 key
= tuple(bound_args
.arguments
.values())[1:]
5860 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
5861 if key
not in cache
:
5862 cache
[key
] = f(self
, *args
, **kwargs
)
5867 class classproperty
:
5868 """property access for class methods with optional caching"""
5869 def __new__(cls
, func
=None, *args
, **kwargs
):
5871 return functools
.partial(cls
, *args
, **kwargs
)
5872 return super().__new
__(cls
)
5874 def __init__(self
, func
, *, cache
=False):
5875 functools
.update_wrapper(self
, func
)
5877 self
._cache
= {} if cache
else None
5879 def __get__(self
, _
, cls
):
5880 if self
._cache
is None:
5881 return self
.func(cls
)
5882 elif cls
not in self
._cache
:
5883 self
._cache
[cls
] = self
.func(cls
)
5884 return self
._cache
[cls
]
5887 class Namespace(types
.SimpleNamespace
):
5888 """Immutable namespace"""
5891 return iter(self
.__dict
__.values())
5895 return self
.__dict
__.items()
5898 MEDIA_EXTENSIONS
= Namespace(
5899 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5900 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5901 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5902 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5903 thumbnails
=('jpg', 'png', 'webp'),
5904 storyboards
=('mhtml', ),
5905 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5906 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5908 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5909 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5911 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5916 for retry in RetryManager(...):
5919 except SomeException as err:
5923 attempt
, _error
= 0, None
5925 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5926 self
.retries
= _retries
or 0
5927 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5929 def _should_retry(self
):
5930 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5934 if self
._error
is NO_DEFAULT
:
5939 def error(self
, value
):
5943 while self
._should
_retry
():
5944 self
.error
= NO_DEFAULT
5948 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5951 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5952 """Utility function for reporting retries"""
5955 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5960 elif isinstance(e
, ExtractorError
):
5961 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5962 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5964 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5966 info(f
'Sleeping {delay:.2f} seconds ...')
5970 def make_archive_id(ie
, video_id
):
5971 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5972 return f
'{ie_key.lower()} {video_id}'
5975 def truncate_string(s
, left
, right
=0):
5976 assert left
> 3 and right
>= 0
5977 if s
is None or len(s
) <= left
+ right
:
5979 return f
'{s[:left-3]}...{s[-right:]}'
5982 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5983 assert 'all' in alias_dict
, '"all" alias is required'
5984 requested
= list(start
or [])
5986 discard
= val
.startswith('-')
5990 if val
in alias_dict
:
5991 val
= alias_dict
[val
] if not discard
else [
5992 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5993 # NB: Do not allow regex in aliases for performance
5994 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5997 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5998 else [val
] if val
in alias_dict
['all'] else None)
6000 raise ValueError(val
)
6003 for item
in current
:
6004 while item
in requested
:
6005 requested
.remove(item
)
6007 requested
.extend(current
)
6009 return orderedSet(requested
)
6013 regex
= r
' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
6015 default
= ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
6016 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
6017 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
6018 ytdl_default
= ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
6019 'height', 'width', 'proto', 'vext', 'abr', 'aext',
6020 'fps', 'fs_approx', 'source', 'id')
6023 'vcodec': {'type': 'ordered', 'regex': True,
6024 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
6025 'acodec': {'type': 'ordered', 'regex': True,
6026 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
6027 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
6028 'order': ['dv', '(hdr)?12', r
'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
6029 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
6030 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
6031 'vext': {'type': 'ordered', 'field': 'video_ext',
6032 'order': ('mp4', 'mov', 'webm', 'flv', '', 'none'),
6033 'order_free': ('webm', 'mp4', 'mov', 'flv', '', 'none')},
6034 'aext': {'type': 'ordered', 'field': 'audio_ext',
6035 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'webm', '', 'none'),
6036 'order_free': ('ogg', 'opus', 'webm', 'mp3', 'm4a', 'aac', '', 'none')},
6037 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}
,
6038 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
6039 'field': ('vcodec', 'acodec'),
6040 'function': lambda it
: int(any(v
!= 'none' for v
in it
))},
6041 'ie_pref': {'priority': True, 'type': 'extractor'}
,
6042 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6043 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6044 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}
,
6045 'quality': {'convert': 'float', 'default': -1}
,
6046 'filesize': {'convert': 'bytes'}
,
6047 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}
,
6048 'id': {'convert': 'string', 'field': 'format_id'}
,
6049 'height': {'convert': 'float_none'}
,
6050 'width': {'convert': 'float_none'}
,
6051 'fps': {'convert': 'float_none'}
,
6052 'channels': {'convert': 'float_none', 'field': 'audio_channels'}
,
6053 'tbr': {'convert': 'float_none'}
,
6054 'vbr': {'convert': 'float_none'}
,
6055 'abr': {'convert': 'float_none'}
,
6056 'asr': {'convert': 'float_none'}
,
6057 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}
,
6059 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}
,
6060 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True}
,
6061 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')}
,
6062 'ext': {'type': 'combined', 'field': ('vext', 'aext')}
,
6063 'res': {'type': 'multiple', 'field': ('height', 'width'),
6064 'function': lambda it
: (lambda l
: min(l
) if l
else 0)(tuple(filter(None, it
)))},
6066 # Actual field names
6067 'format_id': {'type': 'alias', 'field': 'id'}
,
6068 'preference': {'type': 'alias', 'field': 'ie_pref'}
,
6069 'language_preference': {'type': 'alias', 'field': 'lang'}
,
6070 'source_preference': {'type': 'alias', 'field': 'source'}
,
6071 'protocol': {'type': 'alias', 'field': 'proto'}
,
6072 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}
,
6073 'audio_channels': {'type': 'alias', 'field': 'channels'}
,
6076 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6077 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6078 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}
,
6079 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}
,
6080 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}
,
6081 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}
,
6082 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}
,
6083 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}
,
6084 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}
,
6085 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}
,
6086 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}
,
6087 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}
,
6088 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}
,
6089 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}
,
6090 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6091 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6092 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6093 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6094 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6095 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6098 def __init__(self
, ydl
, field_preference
):
6101 self
.evaluate_params(self
.ydl
.params
, field_preference
)
6102 if ydl
.params
.get('verbose'):
6103 self
.print_verbose_info(self
.ydl
.write_debug
)
6105 def _get_field_setting(self
, field
, key
):
6106 if field
not in self
.settings
:
6107 if key
in ('forced', 'priority'):
6109 self
.ydl
.deprecated_feature(f
'Using arbitrary fields ({field}) for format sorting is '
6110 'deprecated and may be removed in a future version')
6111 self
.settings
[field
] = {}
6112 propObj
= self
.settings
[field
]
6113 if key
not in propObj
:
6114 type = propObj
.get('type')
6116 default
= 'preference' if type == 'extractor' else (field
,) if type in ('combined', 'multiple') else field
6117 elif key
== 'convert':
6118 default
= 'order' if type == 'ordered' else 'float_string' if field
else 'ignore'
6120 default
= {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}
.get(key
, None)
6121 propObj
[key
] = default
6124 def _resolve_field_value(self
, field
, value
, convertNone
=False):
6129 value
= value
.lower()
6130 conversion
= self
._get
_field
_setting
(field
, 'convert')
6131 if conversion
== 'ignore':
6133 if conversion
== 'string':
6135 elif conversion
== 'float_none':
6136 return float_or_none(value
)
6137 elif conversion
== 'bytes':
6138 return parse_bytes(value
)
6139 elif conversion
== 'order':
6140 order_list
= (self
._use
_free
_order
and self
._get
_field
_setting
(field
, 'order_free')) or self
._get
_field
_setting
(field
, 'order')
6141 use_regex
= self
._get
_field
_setting
(field
, 'regex')
6142 list_length
= len(order_list
)
6143 empty_pos
= order_list
.index('') if '' in order_list
else list_length
+ 1
6144 if use_regex
and value
is not None:
6145 for i
, regex
in enumerate(order_list
):
6146 if regex
and re
.match(regex
, value
):
6147 return list_length
- i
6148 return list_length
- empty_pos
# not in list
6149 else: # not regex or value = None
6150 return list_length
- (order_list
.index(value
) if value
in order_list
else empty_pos
)
6152 if value
.isnumeric():
6155 self
.settings
[field
]['convert'] = 'string'
6158 def evaluate_params(self
, params
, sort_extractor
):
6159 self
._use
_free
_order
= params
.get('prefer_free_formats', False)
6160 self
._sort
_user
= params
.get('format_sort', [])
6161 self
._sort
_extractor
= sort_extractor
6163 def add_item(field
, reverse
, closest
, limit_text
):
6164 field
= field
.lower()
6165 if field
in self
._order
:
6167 self
._order
.append(field
)
6168 limit
= self
._resolve
_field
_value
(field
, limit_text
)
6171 'closest': False if limit
is None else closest
,
6172 'limit_text': limit_text
,
6174 if field
in self
.settings
:
6175 self
.settings
[field
].update(data
)
6177 self
.settings
[field
] = data
6180 tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'forced'))
6181 + (tuple() if params
.get('format_sort_force', False)
6182 else tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'priority')))
6183 + tuple(self
._sort
_user
) + tuple(sort_extractor
) + self
.default
)
6185 for item
in sort_list
:
6186 match
= re
.match(self
.regex
, item
)
6188 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item
)
6189 field
= match
.group('field')
6192 if self
._get
_field
_setting
(field
, 'type') == 'alias':
6193 alias
, field
= field
, self
._get
_field
_setting
(field
, 'field')
6194 if self
._get
_field
_setting
(alias
, 'deprecated'):
6195 self
.ydl
.deprecated_feature(f
'Format sorting alias {alias} is deprecated and may '
6196 f
'be removed in a future version. Please use {field} instead')
6197 reverse
= match
.group('reverse') is not None
6198 closest
= match
.group('separator') == '~'
6199 limit_text
= match
.group('limit')
6201 has_limit
= limit_text
is not None
6202 has_multiple_fields
= self
._get
_field
_setting
(field
, 'type') == 'combined'
6203 has_multiple_limits
= has_limit
and has_multiple_fields
and not self
._get
_field
_setting
(field
, 'same_limit')
6205 fields
= self
._get
_field
_setting
(field
, 'field') if has_multiple_fields
else (field
,)
6206 limits
= limit_text
.split(':') if has_multiple_limits
else (limit_text
,) if has_limit
else tuple()
6207 limit_count
= len(limits
)
6208 for (i
, f
) in enumerate(fields
):
6209 add_item(f
, reverse
, closest
,
6210 limits
[i
] if i
< limit_count
6211 else limits
[0] if has_limit
and not has_multiple_limits
6214 def print_verbose_info(self
, write_debug
):
6216 write_debug('Sort order given by user: %s' % ', '.join(self
._sort
_user
))
6217 if self
._sort
_extractor
:
6218 write_debug('Sort order given by extractor: %s' % ', '.join(self
._sort
_extractor
))
6219 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
6220 '+' if self
._get
_field
_setting
(field
, 'reverse') else '', field
,
6221 '%s%s(%s)' % ('~' if self
._get
_field
_setting
(field
, 'closest') else ':',
6222 self
._get
_field
_setting
(field
, 'limit_text'),
6223 self
._get
_field
_setting
(field
, 'limit'))
6224 if self
._get
_field
_setting
(field
, 'limit_text') is not None else '')
6225 for field
in self
._order
if self
._get
_field
_setting
(field
, 'visible')]))
6227 def _calculate_field_preference_from_value(self
, format
, field
, type, value
):
6228 reverse
= self
._get
_field
_setting
(field
, 'reverse')
6229 closest
= self
._get
_field
_setting
(field
, 'closest')
6230 limit
= self
._get
_field
_setting
(field
, 'limit')
6232 if type == 'extractor':
6233 maximum
= self
._get
_field
_setting
(field
, 'max')
6234 if value
is None or (maximum
is not None and value
>= maximum
):
6236 elif type == 'boolean':
6237 in_list
= self
._get
_field
_setting
(field
, 'in_list')
6238 not_in_list
= self
._get
_field
_setting
(field
, 'not_in_list')
6239 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
6240 elif type == 'ordered':
6241 value
= self
._resolve
_field
_value
(field
, value
, True)
6243 # try to convert to number
6244 val_num
= float_or_none(value
, default
=self
._get
_field
_setting
(field
, 'default'))
6245 is_num
= self
._get
_field
_setting
(field
, 'convert') != 'string' and val_num
is not None
6249 return ((-10, 0) if value
is None
6250 else (1, value
, 0) if not is_num
# if a field has mixed strings and numbers, strings are sorted higher
6251 else (0, -abs(value
- limit
), value
- limit
if reverse
else limit
- value
) if closest
6252 else (0, value
, 0) if not reverse
and (limit
is None or value
<= limit
)
6253 else (0, -value
, 0) if limit
is None or (reverse
and value
== limit
) or value
> limit
6254 else (-1, value
, 0))
6256 def _calculate_field_preference(self
, format
, field
):
6257 type = self
._get
_field
_setting
(field
, 'type') # extractor, boolean, ordered, field, multiple
6258 get_value
= lambda f
: format
.get(self
._get
_field
_setting
(f
, 'field'))
6259 if type == 'multiple':
6260 type = 'field' # Only 'field' is allowed in multiple for now
6261 actual_fields
= self
._get
_field
_setting
(field
, 'field')
6263 value
= self
._get
_field
_setting
(field
, 'function')(get_value(f
) for f
in actual_fields
)
6265 value
= get_value(field
)
6266 return self
._calculate
_field
_preference
_from
_value
(format
, field
, type, value
)
6268 def calculate_preference(self
, format
):
6269 # Determine missing protocol
6270 if not format
.get('protocol'):
6271 format
['protocol'] = determine_protocol(format
)
6273 # Determine missing ext
6274 if not format
.get('ext') and 'url' in format
:
6275 format
['ext'] = determine_ext(format
['url'])
6276 if format
.get('vcodec') == 'none':
6277 format
['audio_ext'] = format
['ext'] if format
.get('acodec') != 'none' else 'none'
6278 format
['video_ext'] = 'none'
6280 format
['video_ext'] = format
['ext']
6281 format
['audio_ext'] = 'none'
6282 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
6283 # format['preference'] = -1000
6285 # Determine missing bitrates
6286 if format
.get('tbr') is None:
6287 if format
.get('vbr') is not None and format
.get('abr') is not None:
6288 format
['tbr'] = format
.get('vbr', 0) + format
.get('abr', 0)
6290 if format
.get('vcodec') != 'none' and format
.get('vbr') is None:
6291 format
['vbr'] = format
.get('tbr') - format
.get('abr', 0)
6292 if format
.get('acodec') != 'none' and format
.get('abr') is None:
6293 format
['abr'] = format
.get('tbr') - format
.get('vbr', 0)
6295 return tuple(self
._calculate
_field
_preference
(format
, field
) for field
in self
._order
)
6299 has_certifi
= bool(certifi
)
6300 has_websockets
= bool(websockets
)