47 import xml
.etree
.ElementTree
50 from . import traversal
52 from ..compat
import functools
# isort: split
53 from ..compat
import (
54 compat_etree_fromstring
,
56 compat_HTMLParseError
,
60 from ..dependencies
import brotli
, certifi
, websockets
, xattr
61 from ..socks
import ProxyType
, sockssocket
63 __name__
= __name__
.rsplit('.', 1)[0] # Pretend to be the parent module
65 # This is not clearly defined otherwise
66 compiled_regex_type
= type(re
.compile(''))
69 def random_user_agent():
70 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
111 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
114 SUPPORTED_ENCODINGS
= [
118 SUPPORTED_ENCODINGS
.append('br')
121 'User-Agent': random_user_agent(),
122 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
123 'Accept-Language': 'en-us,en;q=0.5',
124 'Sec-Fetch-Mode': 'navigate',
129 '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',
141 ENGLISH_MONTH_NAMES
= [
142 'January', 'February', 'March', 'April', 'May', 'June',
143 'July', 'August', 'September', 'October', 'November', 'December']
146 'en': ENGLISH_MONTH_NAMES
,
148 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
149 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
150 # these follow the genitive grammatical case (dopełniacz)
151 # some websites might be using nominative, which will require another month list
152 # https://en.wikibooks.org/wiki/Polish/Noun_cases
153 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
154 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
157 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
159 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
160 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
161 'EST': -5, 'EDT': -4, # Eastern
162 'CST': -6, 'CDT': -5, # Central
163 'MST': -7, 'MDT': -6, # Mountain
164 'PST': -8, 'PDT': -7 # Pacific
167 # needed for sanitizing filenames in restricted mode
168 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
169 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
170 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
200 '%Y-%m-%d %H:%M:%S.%f',
201 '%Y-%m-%d %H:%M:%S:%f',
204 '%Y-%m-%dT%H:%M:%SZ',
205 '%Y-%m-%dT%H:%M:%S.%fZ',
206 '%Y-%m-%dT%H:%M:%S.%f0Z',
208 '%Y-%m-%dT%H:%M:%S.%f',
211 '%b %d %Y at %H:%M:%S',
213 '%B %d %Y at %H:%M:%S',
217 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
218 DATE_FORMATS_DAY_FIRST
.extend([
229 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
230 DATE_FORMATS_MONTH_FIRST
.extend([
238 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
239 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
241 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
245 def preferredencoding():
246 """Get preferred encoding.
248 Returns the best encoding scheme for the system, based on
249 locale.getpreferredencoding() and some further tweaks.
252 pref = locale.getpreferredencoding()
260 def write_json_file(obj, fn):
261 """ Encode obj as JSON and write it to fn, atomically if possible """
263 tf = tempfile.NamedTemporaryFile(
264 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
265 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
269 json.dump(obj, tf, ensure_ascii=False)
270 if sys.platform == 'win32
':
271 # Need to remove existing file on Windows, else os.rename raises
272 # WindowsError or FileExistsError.
273 with contextlib.suppress(OSError):
275 with contextlib.suppress(OSError):
278 os.chmod(tf.name, 0o666 & ~mask)
279 os.rename(tf.name, fn)
281 with contextlib.suppress(OSError):
286 def find_xpath_attr(node, xpath, key, val=None):
287 """ Find the xpath xpath[@key=val] """
288 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
289 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
290 return node.find(expr)
292 # On python2.6 the xml.etree.ElementTree.Element methods don't support
293 # the namespace parameter
296 def xpath_with_ns(path
, ns_map
):
297 components
= [c
.split(':') for c
in path
.split('/')]
301 replaced
.append(c
[0])
304 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
305 return '/'.join(replaced
)
308 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
309 def _find_xpath(xpath
):
310 return node
.find(xpath
)
312 if isinstance(xpath
, str):
313 n
= _find_xpath(xpath
)
321 if default
is not NO_DEFAULT
:
324 name
= xpath
if name
is None else name
325 raise ExtractorError('Could not find XML element %s' % name
)
331 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
332 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
333 if n
is None or n
== default
:
336 if default
is not NO_DEFAULT
:
339 name
= xpath
if name
is None else name
340 raise ExtractorError('Could not find XML element\'s text %s' % name
)
346 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
347 n
= find_xpath_attr(node
, xpath
, key
)
349 if default
is not NO_DEFAULT
:
352 name
= f
'{xpath}[@{key}]' if name
is None else name
353 raise ExtractorError('Could not find XML attribute %s' % name
)
359 def get_element_by_id(id, html
, **kwargs
):
360 """Return the content of the tag with the specified ID in the passed HTML document"""
361 return get_element_by_attribute('id', id, html
, **kwargs
)
364 def get_element_html_by_id(id, html
, **kwargs
):
365 """Return the html of the tag with the specified ID in the passed HTML document"""
366 return get_element_html_by_attribute('id', id, html
, **kwargs
)
369 def get_element_by_class(class_name
, html
):
370 """Return the content of the first tag with the specified class in the passed HTML document"""
371 retval
= get_elements_by_class(class_name
, html
)
372 return retval
[0] if retval
else None
375 def get_element_html_by_class(class_name
, html
):
376 """Return the html of the first tag with the specified class in the passed HTML document"""
377 retval
= get_elements_html_by_class(class_name
, html
)
378 return retval
[0] if retval
else None
381 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
382 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
383 return retval
[0] if retval
else None
386 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
387 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
388 return retval
[0] if retval
else None
391 def get_elements_by_class(class_name
, html
, **kargs
):
392 """Return the content of all tags with the specified class in the passed HTML document as a list"""
393 return get_elements_by_attribute(
394 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
395 html, escape_value=False)
398 def get_elements_html_by_class(class_name, html):
399 """Return the html of all tags with the specified class in the passed HTML document as a list"""
400 return get_elements_html_by_attribute(
401 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
402 html, escape_value=False)
405 def get_elements_by_attribute(*args, **kwargs):
406 """Return the content of the tag with the specified attribute in the passed HTML document"""
407 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
410 def get_elements_html_by_attribute(*args, **kwargs):
411 """Return the html of the tag with the specified attribute in the passed HTML document"""
412 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
415 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
417 Return the text (content) and the html (whole) of the tag with the specified
418 attribute in the passed HTML document
423 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
425 value = re.escape(value) if escape_value else value
427 partial_element_re = rf'''(?x
)
429 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
430 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
433 for m in re.finditer(partial_element_re, html):
434 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
437 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
442 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
444 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
445 closing tag for the first opening tag it has encountered, and can be used
449 class HTMLBreakOnClosingTagException(Exception):
453 self.tagstack = collections.deque()
454 html.parser.HTMLParser.__init__(self)
459 def __exit__(self, *_):
463 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
464 # so data remains buffered; we no longer have any interest in it, thus
465 # override this method to discard it
468 def handle_starttag(self, tag, _):
469 self.tagstack.append(tag)
471 def handle_endtag(self, tag):
472 if not self.tagstack:
473 raise compat_HTMLParseError('no tags
in the stack
')
475 inner_tag = self.tagstack.pop()
479 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
480 if not self.tagstack:
481 raise self.HTMLBreakOnClosingTagException()
484 # XXX: This should be far less strict
485 def get_element_text_and_html_by_tag(tag, html):
487 For the first element with the specified tag in the passed HTML document
488 return its' content (text
) and the whole
element (html
)
490 def find_or_raise(haystack, needle, exc):
492 return haystack.index(needle)
495 closing_tag = f'</{tag}>'
496 whole_start = find_or_raise(
497 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
498 content_start = find_or_raise(
499 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
500 content_start += whole_start + 1
501 with HTMLBreakOnClosingTagParser() as parser:
502 parser.feed(html[whole_start:content_start])
503 if not parser.tagstack or parser.tagstack[0] != tag:
504 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
505 offset = content_start
506 while offset < len(html):
507 next_closing_tag_start = find_or_raise(
508 html[offset:], closing_tag,
509 compat_HTMLParseError(f'closing {tag} tag not found'))
510 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
512 parser.feed(html[offset:offset + next_closing_tag_end])
513 offset += next_closing_tag_end
514 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
515 return html[content_start:offset + next_closing_tag_start], \
516 html[whole_start:offset + next_closing_tag_end]
517 raise compat_HTMLParseError('unexpected end of html')
520 class HTMLAttributeParser(html.parser.HTMLParser):
521 """Trivial HTML parser to gather the attributes
for a single element
"""
525 html.parser.HTMLParser.__init__(self)
527 def handle_starttag(self, tag, attrs):
528 self.attrs = dict(attrs)
529 raise compat_HTMLParseError('done')
532 class HTMLListAttrsParser(html.parser.HTMLParser):
533 """HTML parser to gather the attributes
for the elements of a
list"""
536 html.parser.HTMLParser.__init__(self)
540 def handle_starttag(self, tag, attrs):
541 if tag == 'li' and self._level == 0:
542 self.items.append(dict(attrs))
545 def handle_endtag(self, tag):
549 def extract_attributes(html_element):
550 """Given a string
for an HTML element such
as
552 a
="foo" B
="bar" c
="&98;az" d
=boz
553 empty
= noval entity
="&"
556 Decode
and return a dictionary of attributes
.
558 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
559 'empty': '', 'noval': None, 'entity': '&',
560 'sq': '"', 'dq': '\''
563 parser = HTMLAttributeParser()
564 with contextlib.suppress(compat_HTMLParseError):
565 parser.feed(html_element)
570 def parse_list(webpage):
571 """Given a string
for an series of HTML
<li
> elements
,
572 return a dictionary of their attributes
"""
573 parser = HTMLListAttrsParser()
579 def clean_html(html):
580 """Clean an HTML snippet into a readable string
"""
582 if html is None: # Convenience for sanitizing descriptions etc.
585 html = re.sub(r'\s+', ' ', html)
586 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
587 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
589 html = re.sub('<.*?>', '', html)
590 # Replace html entities
591 html = unescapeHTML(html)
595 class LenientJSONDecoder(json.JSONDecoder):
597 def __init__(self, *args, transform_source=None, ignore_extra=False, close_objects=0, **kwargs):
598 self.transform_source, self.ignore_extra = transform_source, ignore_extra
599 self._close_attempts = 2 * close_objects
600 super().__init__(*args, **kwargs)
603 def _close_object(err):
604 doc = err.doc[:err.pos]
605 # We need to add comma first to get the correct error message
606 if err.msg.startswith('Expecting \',\''):
608 elif not doc.endswith(','):
611 if err.msg.startswith('Expecting property name'):
612 return doc[:-1] + '}'
613 elif err.msg.startswith('Expecting value'):
614 return doc[:-1] + ']'
617 if self.transform_source:
618 s = self.transform_source(s)
619 for attempt in range(self._close_attempts + 1):
621 if self.ignore_extra:
622 return self.raw_decode(s.lstrip())[0]
623 return super().decode(s)
624 except json.JSONDecodeError as e:
627 elif attempt < self._close_attempts:
628 s = self._close_object(e)
631 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
632 assert False, 'Too many attempts to decode JSON'
635 def sanitize_open(filename, open_mode):
636 """Try to
open the given filename
, and slightly tweak it
if this fails
.
638 Attempts to
open the given filename
. If this fails
, it tries to change
639 the filename slightly
, step by step
, until it
's either able to open it
640 or it fails and raises a final exception, like the standard open()
643 It returns the tuple (stream, definitive_file_name).
646 if sys.platform == 'win32
':
649 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
650 with contextlib.suppress(io.UnsupportedOperation):
651 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
652 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
654 for attempt in range(2):
657 if sys.platform == 'win32
':
658 # FIXME: An exclusive lock also locks the file from being read.
659 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
660 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
661 raise LockingUnsupportedError()
662 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
664 stream
= open(filename
, open_mode
)
665 return stream
, filename
666 except OSError as err
:
667 if attempt
or err
.errno
in (errno
.EACCES
,):
669 old_filename
, filename
= filename
, sanitize_path(filename
)
670 if old_filename
== filename
:
674 def timeconvert(timestr
):
675 """Convert RFC 2822 defined time string into system timestamp"""
677 timetuple
= email
.utils
.parsedate_tz(timestr
)
678 if timetuple
is not None:
679 timestamp
= email
.utils
.mktime_tz(timetuple
)
683 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
684 """Sanitizes a string so it could be used as part of a filename.
685 @param restricted Use a stricter subset of allowed characters
686 @param is_id Whether this is an ID that should be kept unchanged if possible.
687 If unset, yt-dlp's new sanitization rules are in effect
692 def replace_insane(char
):
693 if restricted
and char
in ACCENT_CHARS
:
694 return ACCENT_CHARS
[char
]
695 elif not restricted
and char
== '\n':
697 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
698 # Replace with their full-width unicode counterparts
699 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
700 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
703 return '' if restricted
else '\''
705 return '\0_\0-' if restricted
else '\0 \0-'
706 elif char
in '\\/|*<>':
708 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
712 # Replace look-alike Unicode glyphs
713 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
714 s
= unicodedata
.normalize('NFKC', s
)
715 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
716 result
= ''.join(map(replace_insane
, s
))
717 if is_id
is NO_DEFAULT
:
718 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
719 STRIP_RE
= r
'(?:\0.|[ _-])*'
720 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
721 result
= result
.replace('\0', '') or '_'
724 while '__' in result
:
725 result
= result
.replace('__', '_')
726 result
= result
.strip('_')
727 # Common case of "Foreign band name - English song title"
728 if restricted
and result
.startswith('-_'):
730 if result
.startswith('-'):
731 result
= '_' + result
[len('-'):]
732 result
= result
.lstrip('.')
738 def sanitize_path(s
, force
=False):
739 """Sanitizes and normalizes path on Windows"""
740 if sys
.platform
== 'win32':
742 drive_or_unc
, _
= os
.path
.splitdrive(s
)
748 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
752 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
753 for path_part
in norm_path
]
755 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
756 elif force
and s
and s
[0] == os
.path
.sep
:
757 sanitized_path
.insert(0, os
.path
.sep
)
758 return os
.path
.join(*sanitized_path
)
761 def sanitize_url(url
, *, scheme
='http'):
762 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
763 # the number of unwanted failures due to missing protocol
766 elif url
.startswith('//'):
767 return f
'{scheme}:{url}'
768 # Fix some common typos seen so far
770 # https://github.com/ytdl-org/youtube-dl/issues/15649
771 (r
'^httpss://', r
'https://'),
772 # https://bx1.be/lives/direct-tv/
773 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
775 for mistake
, fixup
in COMMON_TYPOS
:
776 if re
.match(mistake
, url
):
777 return re
.sub(mistake
, fixup
, url
)
781 def extract_basic_auth(url
):
782 parts
= urllib
.parse
.urlsplit(url
)
783 if parts
.username
is None:
785 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
786 parts
.hostname
if parts
.port
is None
787 else '%s:%d' % (parts
.hostname
, parts
.port
))))
788 auth_payload
= base64
.b64encode(
789 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
790 return url
, f
'Basic {auth_payload.decode()}'
793 def sanitized_Request(url
, *args
, **kwargs
):
794 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
795 if auth_header
is not None:
796 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
797 headers
['Authorization'] = auth_header
798 return urllib
.request
.Request(url
, *args
, **kwargs
)
802 """Expand shell variables and ~"""
803 return os
.path
.expandvars(compat_expanduser(s
))
806 def orderedSet(iterable
, *, lazy
=False):
807 """Remove all duplicates from the input iterable"""
809 seen
= [] # Do not use set since the items can be unhashable
815 return _iter() if lazy
else list(_iter())
818 def _htmlentity_transform(entity_with_semicolon
):
819 """Transforms an HTML entity to a character."""
820 entity
= entity_with_semicolon
[:-1]
822 # Known non-numeric HTML entity
823 if entity
in html
.entities
.name2codepoint
:
824 return chr(html
.entities
.name2codepoint
[entity
])
826 # TODO: HTML5 allows entities without a semicolon.
827 # E.g. 'Éric' should be decoded as 'Éric'.
828 if entity_with_semicolon
in html
.entities
.html5
:
829 return html
.entities
.html5
[entity_with_semicolon
]
831 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
833 numstr
= mobj
.group(1)
834 if numstr
.startswith('x'):
836 numstr
= '0%s' % numstr
839 # See https://github.com/ytdl-org/youtube-dl/issues/7518
840 with contextlib
.suppress(ValueError):
841 return chr(int(numstr
, base
))
843 # Unknown entity in name, return its literal representation
844 return '&%s;' % entity
850 assert isinstance(s
, str)
853 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
856 def escapeHTML(text
):
859 .replace('&', '&')
860 .replace('<', '<')
861 .replace('>', '>')
862 .replace('"', '"')
863 .replace("'", ''')
867 def process_communicate_or_kill(p
, *args
, **kwargs
):
868 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
869 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
870 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
873 class Popen(subprocess
.Popen
):
874 if sys
.platform
== 'win32':
875 _startupinfo
= subprocess
.STARTUPINFO()
876 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
881 def _fix_pyinstaller_ld_path(env
):
882 """Restore LD_LIBRARY_PATH when using PyInstaller
883 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
884 https://github.com/yt-dlp/yt-dlp/issues/4573
886 if not hasattr(sys
, '_MEIPASS'):
890 orig
= env
.get(f
'{key}_ORIG')
896 _fix('LD_LIBRARY_PATH') # Linux
897 _fix('DYLD_LIBRARY_PATH') # macOS
899 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
901 env
= os
.environ
.copy()
902 self
._fix
_pyinstaller
_ld
_path
(env
)
904 self
.__text
_mode
= kwargs
.get('encoding') or kwargs
.get('errors') or text
or kwargs
.get('universal_newlines')
906 kwargs
['universal_newlines'] = True # For 3.6 compatibility
907 kwargs
.setdefault('encoding', 'utf-8')
908 kwargs
.setdefault('errors', 'replace')
909 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
911 def communicate_or_kill(self
, *args
, **kwargs
):
913 return self
.communicate(*args
, **kwargs
)
914 except BaseException
: # Including KeyboardInterrupt
915 self
.kill(timeout
=None)
918 def kill(self
, *, timeout
=0):
921 self
.wait(timeout
=timeout
)
924 def run(cls
, *args
, timeout
=None, **kwargs
):
925 with cls(*args
, **kwargs
) as proc
:
926 default
= '' if proc
.__text
_mode
else b
''
927 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
928 return stdout
or default
, stderr
or default
, proc
.returncode
931 def encodeArgument(s
):
932 # Legacy code that uses byte strings
933 # Uncomment the following line after fixing all post processors
934 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
935 return s
if isinstance(s
, str) else s
.decode('ascii')
938 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
941 def timetuple_from_msec(msec
):
942 secs
, msec
= divmod(msec
, 1000)
943 mins
, secs
= divmod(secs
, 60)
944 hrs
, mins
= divmod(mins
, 60)
945 return _timetuple(hrs
, mins
, secs
, msec
)
948 def formatSeconds(secs
, delim
=':', msec
=False):
949 time
= timetuple_from_msec(secs
* 1000)
951 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
953 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
955 ret
= '%d' % time
.seconds
956 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
959 def _ssl_load_windows_store_certs(ssl_context
, storename
):
960 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
962 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
963 if encoding
== 'x509_asn' and (
964 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
965 except PermissionError
:
968 with contextlib
.suppress(ssl
.SSLError
):
969 ssl_context
.load_verify_locations(cadata
=cert
)
972 def make_HTTPS_handler(params
, **kwargs
):
973 opts_check_certificate
= not params
.get('nocheckcertificate')
974 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
975 context
.check_hostname
= opts_check_certificate
976 if params
.get('legacyserverconnect'):
977 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
978 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
979 context
.set_ciphers('DEFAULT')
981 sys
.version_info
< (3, 10)
982 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
983 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
985 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
986 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
987 # in some situations [2][3].
988 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
989 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
990 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
991 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
992 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
993 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
994 # 4. https://peps.python.org/pep-0644/
995 # 5. https://peps.python.org/pep-0644/#libressl-support
996 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
997 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
998 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1000 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1001 if opts_check_certificate
:
1002 if certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1003 context
.load_verify_locations(cafile
=certifi
.where())
1006 context
.load_default_certs()
1007 # Work around the issue in load_default_certs when there are bad certificates. See:
1008 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1009 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1010 except ssl
.SSLError
:
1011 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1012 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1013 for storename
in ('CA', 'ROOT'):
1014 _ssl_load_windows_store_certs(context
, storename
)
1015 context
.set_default_verify_paths()
1017 client_certfile
= params
.get('client_certificate')
1020 context
.load_cert_chain(
1021 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1022 password
=params
.get('client_certificate_password'))
1023 except ssl
.SSLError
:
1024 raise YoutubeDLError('Unable to load client certificate')
1026 # Some servers may reject requests if ALPN extension is not sent. See:
1027 # https://github.com/python/cpython/issues/85140
1028 # https://github.com/yt-dlp/yt-dlp/issues/3878
1029 with contextlib
.suppress(NotImplementedError):
1030 context
.set_alpn_protocols(['http/1.1'])
1032 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1035 def bug_reports_message(before
=';'):
1036 from ..update
import REPOSITORY
1038 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1039 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1041 before
= before
.rstrip()
1042 if not before
or before
.endswith(('.', '!', '?')):
1043 msg
= msg
[0].title() + msg
[1:]
1045 return (before
+ ' ' if before
else '') + msg
1048 class YoutubeDLError(Exception):
1049 """Base exception for YoutubeDL errors."""
1052 def __init__(self
, msg
=None):
1055 elif self
.msg
is None:
1056 self
.msg
= type(self
).__name
__
1057 super().__init
__(self
.msg
)
1060 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1061 if hasattr(ssl
, 'CertificateError'):
1062 network_exceptions
.append(ssl
.CertificateError
)
1063 network_exceptions
= tuple(network_exceptions
)
1066 class ExtractorError(YoutubeDLError
):
1067 """Error during info extraction."""
1069 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1070 """ tb, if given, is the original traceback (so that it can be printed out).
1071 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1073 if sys
.exc_info()[0] in network_exceptions
:
1076 self
.orig_msg
= str(msg
)
1078 self
.expected
= expected
1080 self
.video_id
= video_id
1082 self
.exc_info
= sys
.exc_info() # preserve original exception
1083 if isinstance(self
.exc_info
[1], ExtractorError
):
1084 self
.exc_info
= self
.exc_info
[1].exc_info
1085 super().__init
__(self
.__msg
)
1090 format_field(self
.ie
, None, '[%s] '),
1091 format_field(self
.video_id
, None, '%s: '),
1093 format_field(self
.cause
, None, ' (caused by %r)'),
1094 '' if self
.expected
else bug_reports_message()))
1096 def format_traceback(self
):
1097 return join_nonempty(
1098 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1099 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1102 def __setattr__(self
, name
, value
):
1103 super().__setattr
__(name
, value
)
1104 if getattr(self
, 'msg', None) and name
not in ('msg', 'args'):
1105 self
.msg
= self
.__msg
or type(self
).__name
__
1106 self
.args
= (self
.msg
, ) # Cannot be property
1109 class UnsupportedError(ExtractorError
):
1110 def __init__(self
, url
):
1112 'Unsupported URL: %s' % url
, expected
=True)
1116 class RegexNotFoundError(ExtractorError
):
1117 """Error when a regex didn't match"""
1121 class GeoRestrictedError(ExtractorError
):
1122 """Geographic restriction Error exception.
1124 This exception may be thrown when a video is not available from your
1125 geographic location due to geographic restrictions imposed by a website.
1128 def __init__(self
, msg
, countries
=None, **kwargs
):
1129 kwargs
['expected'] = True
1130 super().__init
__(msg
, **kwargs
)
1131 self
.countries
= countries
1134 class UserNotLive(ExtractorError
):
1135 """Error when a channel/user is not live"""
1137 def __init__(self
, msg
=None, **kwargs
):
1138 kwargs
['expected'] = True
1139 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1142 class DownloadError(YoutubeDLError
):
1143 """Download Error exception.
1145 This exception may be thrown by FileDownloader objects if they are not
1146 configured to continue on errors. They will contain the appropriate
1150 def __init__(self
, msg
, exc_info
=None):
1151 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1152 super().__init
__(msg
)
1153 self
.exc_info
= exc_info
1156 class EntryNotInPlaylist(YoutubeDLError
):
1157 """Entry not in playlist exception.
1159 This exception will be thrown by YoutubeDL when a requested entry
1160 is not found in the playlist info_dict
1162 msg
= 'Entry not found in info'
1165 class SameFileError(YoutubeDLError
):
1166 """Same File exception.
1168 This exception will be thrown by FileDownloader objects if they detect
1169 multiple files would have to be downloaded to the same file on disk.
1171 msg
= 'Fixed output name but more than one file to download'
1173 def __init__(self
, filename
=None):
1174 if filename
is not None:
1175 self
.msg
+= f
': {filename}'
1176 super().__init
__(self
.msg
)
1179 class PostProcessingError(YoutubeDLError
):
1180 """Post Processing exception.
1182 This exception may be raised by PostProcessor's .run() method to
1183 indicate an error in the postprocessing task.
1187 class DownloadCancelled(YoutubeDLError
):
1188 """ Exception raised when the download queue should be interrupted """
1189 msg
= 'The download was cancelled'
1192 class ExistingVideoReached(DownloadCancelled
):
1193 """ --break-on-existing triggered """
1194 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1197 class RejectedVideoReached(DownloadCancelled
):
1198 """ --break-match-filter triggered """
1199 msg
= 'Encountered a video that did not match filter, stopping due to --break-match-filter'
1202 class MaxDownloadsReached(DownloadCancelled
):
1203 """ --max-downloads limit has been reached. """
1204 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1207 class ReExtractInfo(YoutubeDLError
):
1208 """ Video info needs to be re-extracted. """
1210 def __init__(self
, msg
, expected
=False):
1211 super().__init
__(msg
)
1212 self
.expected
= expected
1215 class ThrottledDownload(ReExtractInfo
):
1216 """ Download speed below --throttled-rate. """
1217 msg
= 'The download speed is below throttle limit'
1220 super().__init
__(self
.msg
, expected
=False)
1223 class UnavailableVideoError(YoutubeDLError
):
1224 """Unavailable Format exception.
1226 This exception will be thrown when a video is requested
1227 in a format that is not available for that video.
1229 msg
= 'Unable to download video'
1231 def __init__(self
, err
=None):
1233 self
.msg
+= f
': {err}'
1234 super().__init
__(self
.msg
)
1237 class ContentTooShortError(YoutubeDLError
):
1238 """Content Too Short exception.
1240 This exception may be raised by FileDownloader objects when a file they
1241 download is too small for what the server announced first, indicating
1242 the connection was probably interrupted.
1245 def __init__(self
, downloaded
, expected
):
1246 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1248 self
.downloaded
= downloaded
1249 self
.expected
= expected
1252 class XAttrMetadataError(YoutubeDLError
):
1253 def __init__(self
, code
=None, msg
='Unknown error'):
1254 super().__init
__(msg
)
1258 # Parsing code and msg
1259 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1260 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1261 self
.reason
= 'NO_SPACE'
1262 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1263 self
.reason
= 'VALUE_TOO_LONG'
1265 self
.reason
= 'NOT_SUPPORTED'
1268 class XAttrUnavailableError(YoutubeDLError
):
1272 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1273 hc
= http_class(*args
, **kwargs
)
1274 source_address
= ydl_handler
._params
.get('source_address')
1276 if source_address
is not None:
1277 # This is to workaround _create_connection() from socket where it will try all
1278 # address data from getaddrinfo() including IPv6. This filters the result from
1279 # getaddrinfo() based on the source_address value.
1280 # This is based on the cpython socket.create_connection() function.
1281 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1282 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1283 host
, port
= address
1285 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1286 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1287 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1288 if addrs
and not ip_addrs
:
1289 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1291 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1292 % (ip_version
, source_address
[0]))
1293 for res
in ip_addrs
:
1294 af
, socktype
, proto
, canonname
, sa
= res
1297 sock
= socket
.socket(af
, socktype
, proto
)
1298 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1299 sock
.settimeout(timeout
)
1300 sock
.bind(source_address
)
1302 err
= None # Explicitly break reference cycle
1304 except OSError as _
:
1306 if sock
is not None:
1311 raise OSError('getaddrinfo returns an empty list')
1312 if hasattr(hc
, '_create_connection'):
1313 hc
._create
_connection
= _create_connection
1314 hc
.source_address
= (source_address
, 0)
1319 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1320 """Handler for HTTP requests and responses.
1322 This class, when installed with an OpenerDirector, automatically adds
1323 the standard headers to every HTTP request and handles gzipped, deflated and
1324 brotli responses from web servers.
1326 Part of this code was copied from:
1328 http://techknack.net/python-urllib2-handlers/
1330 Andrew Rowls, the author of that code, agreed to release it to the
1334 def __init__(self
, params
, *args
, **kwargs
):
1335 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1336 self
._params
= params
1338 def http_open(self
, req
):
1339 conn_class
= http
.client
.HTTPConnection
1341 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1343 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1344 del req
.headers
['Ytdl-socks-proxy']
1346 return self
.do_open(functools
.partial(
1347 _create_http_connection
, self
, conn_class
, False),
1355 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1357 return zlib
.decompress(data
)
1363 return brotli
.decompress(data
)
1367 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(data
), mode
='rb')
1370 except OSError as original_oserror
:
1371 # There may be junk add the end of the file
1372 # See http://stackoverflow.com/q/4928560/35070 for details
1373 for i
in range(1, 1024):
1375 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(data
[:-i
]), mode
='rb')
1380 raise original_oserror
1382 def http_request(self
, req
):
1383 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1384 # always respected by websites, some tend to give out URLs with non percent-encoded
1385 # non-ASCII characters (see telemb.py, ard.py [#3412])
1386 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1387 # To work around aforementioned issue we will replace request's original URL with
1388 # percent-encoded one
1389 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1390 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1391 url
= req
.get_full_url()
1392 url_escaped
= escape_url(url
)
1394 # Substitute URL if any change after escaping
1395 if url
!= url_escaped
:
1396 req
= update_Request(req
, url
=url_escaped
)
1398 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1399 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1400 # The dict keys are capitalized because of this bug by urllib
1401 if h
.capitalize() not in req
.headers
:
1402 req
.add_header(h
, v
)
1404 if 'Youtubedl-no-compression' in req
.headers
: # deprecated
1405 req
.headers
.pop('Youtubedl-no-compression', None)
1406 req
.add_header('Accept-encoding', 'identity')
1408 if 'Accept-encoding' not in req
.headers
:
1409 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1411 return super().do_request_(req
)
1413 def http_response(self
, req
, resp
):
1416 # Content-Encoding header lists the encodings in order that they were applied [1].
1417 # To decompress, we simply do the reverse.
1418 # [1]: https://datatracker.ietf.org/doc/html/rfc9110#name-content-encoding
1419 decoded_response
= None
1420 for encoding
in (e
.strip() for e
in reversed(resp
.headers
.get('Content-encoding', '').split(','))):
1421 if encoding
== 'gzip':
1422 decoded_response
= self
.gz(decoded_response
or resp
.read())
1423 elif encoding
== 'deflate':
1424 decoded_response
= self
.deflate(decoded_response
or resp
.read())
1425 elif encoding
== 'br' and brotli
:
1426 decoded_response
= self
.brotli(decoded_response
or resp
.read())
1428 if decoded_response
is not None:
1429 resp
= urllib
.request
.addinfourl(io
.BytesIO(decoded_response
), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1430 resp
.msg
= old_resp
.msg
1431 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1432 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1433 if 300 <= resp
.code
< 400:
1434 location
= resp
.headers
.get('Location')
1436 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1437 location
= location
.encode('iso-8859-1').decode()
1438 location_escaped
= escape_url(location
)
1439 if location
!= location_escaped
:
1440 del resp
.headers
['Location']
1441 resp
.headers
['Location'] = location_escaped
1444 https_request
= http_request
1445 https_response
= http_response
1448 def make_socks_conn_class(base_class
, socks_proxy
):
1449 assert issubclass(base_class
, (
1450 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1452 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1453 if url_components
.scheme
.lower() == 'socks5':
1454 socks_type
= ProxyType
.SOCKS5
1455 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1456 socks_type
= ProxyType
.SOCKS4
1457 elif url_components
.scheme
.lower() == 'socks4a':
1458 socks_type
= ProxyType
.SOCKS4A
1460 def unquote_if_non_empty(s
):
1463 return urllib
.parse
.unquote_plus(s
)
1467 url_components
.hostname
, url_components
.port
or 1080,
1469 unquote_if_non_empty(url_components
.username
),
1470 unquote_if_non_empty(url_components
.password
),
1473 class SocksConnection(base_class
):
1475 self
.sock
= sockssocket()
1476 self
.sock
.setproxy(*proxy_args
)
1477 if isinstance(self
.timeout
, (int, float)):
1478 self
.sock
.settimeout(self
.timeout
)
1479 self
.sock
.connect((self
.host
, self
.port
))
1481 if isinstance(self
, http
.client
.HTTPSConnection
):
1482 if hasattr(self
, '_context'): # Python > 2.6
1483 self
.sock
= self
._context
.wrap_socket(
1484 self
.sock
, server_hostname
=self
.host
)
1486 self
.sock
= ssl
.wrap_socket(self
.sock
)
1488 return SocksConnection
1491 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1492 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1493 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1494 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1495 self
._params
= params
1497 def https_open(self
, req
):
1499 conn_class
= self
._https
_conn
_class
1501 if hasattr(self
, '_context'): # python > 2.6
1502 kwargs
['context'] = self
._context
1503 if hasattr(self
, '_check_hostname'): # python 3.x
1504 kwargs
['check_hostname'] = self
._check
_hostname
1506 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1508 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1509 del req
.headers
['Ytdl-socks-proxy']
1512 return self
.do_open(
1513 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1514 except urllib
.error
.URLError
as e
:
1515 if (isinstance(e
.reason
, ssl
.SSLError
)
1516 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1517 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1521 def is_path_like(f
):
1522 return isinstance(f
, (str, bytes, os
.PathLike
))
1525 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1526 def __init__(self
, cookiejar
=None):
1527 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1529 def http_response(self
, request
, response
):
1530 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1532 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1533 https_response
= http_response
1536 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1537 """YoutubeDL redirect handler
1539 The code is based on HTTPRedirectHandler implementation from CPython [1].
1541 This redirect handler fixes and improves the logic to better align with RFC7261
1542 and what browsers tend to do [2][3]
1544 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1545 2. https://datatracker.ietf.org/doc/html/rfc7231
1546 3. https://github.com/python/cpython/issues/91306
1549 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1551 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1552 if code
not in (301, 302, 303, 307, 308):
1553 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1555 new_method
= req
.get_method()
1558 # A 303 must either use GET or HEAD for subsequent request
1559 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1560 if code
== 303 and req
.get_method() != 'HEAD':
1562 # 301 and 302 redirects are commonly turned into a GET from a POST
1563 # for subsequent requests by browsers, so we'll do the same.
1564 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1565 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1566 elif code
in (301, 302) and req
.get_method() == 'POST':
1569 # only remove payload if method changed (e.g. POST to GET)
1570 if new_method
!= req
.get_method():
1572 remove_headers
.extend(['Content-Length', 'Content-Type'])
1574 new_headers
= {k: v for k, v in req.headers.items() if k.lower() not in remove_headers}
1576 return urllib
.request
.Request(
1577 newurl
, headers
=new_headers
, origin_req_host
=req
.origin_req_host
,
1578 unverifiable
=True, method
=new_method
, data
=new_data
)
1581 def extract_timezone(date_str
):
1584 ^.{8,}? # >=8 char non-TZ prefix, if present
1585 (?P<tz>Z| # just the UTC Z, or
1586 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1587 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1588 [ ]? # optional space
1589 (?P<sign>\+|-) # +/-
1590 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1594 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1595 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1596 if timezone
is not None:
1597 date_str
= date_str
[:-len(m
.group('tz'))]
1598 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1600 date_str
= date_str
[:-len(m
.group('tz'))]
1601 if not m
.group('sign'):
1602 timezone
= datetime
.timedelta()
1604 sign
= 1 if m
.group('sign') == '+' else -1
1605 timezone
= datetime
.timedelta(
1606 hours
=sign
* int(m
.group('hours')),
1607 minutes
=sign
* int(m
.group('minutes')))
1608 return timezone
, date_str
1611 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1612 """ Return a UNIX timestamp from the given date """
1614 if date_str
is None:
1617 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1619 if timezone
is None:
1620 timezone
, date_str
= extract_timezone(date_str
)
1622 with contextlib
.suppress(ValueError):
1623 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1624 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1625 return calendar
.timegm(dt
.timetuple())
1628 def date_formats(day_first
=True):
1629 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1632 def unified_strdate(date_str
, day_first
=True):
1633 """Return a string with the date in the format YYYYMMDD"""
1635 if date_str
is None:
1639 date_str
= date_str
.replace(',', ' ')
1640 # Remove AM/PM + timezone
1641 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1642 _
, date_str
= extract_timezone(date_str
)
1644 for expression
in date_formats(day_first
):
1645 with contextlib
.suppress(ValueError):
1646 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1647 if upload_date
is None:
1648 timetuple
= email
.utils
.parsedate_tz(date_str
)
1650 with contextlib
.suppress(ValueError):
1651 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1652 if upload_date
is not None:
1653 return str(upload_date
)
1656 def unified_timestamp(date_str
, day_first
=True):
1657 if date_str
is None:
1660 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1661 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1663 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1664 timezone
, date_str
= extract_timezone(date_str
)
1666 # Remove AM/PM + timezone
1667 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1669 # Remove unrecognized timezones from ISO 8601 alike timestamps
1670 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1672 date_str
= date_str
[:-len(m
.group('tz'))]
1674 # Python only supports microseconds, so remove nanoseconds
1675 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1677 date_str
= m
.group(1)
1679 for expression
in date_formats(day_first
):
1680 with contextlib
.suppress(ValueError):
1681 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1682 return calendar
.timegm(dt
.timetuple())
1684 timetuple
= email
.utils
.parsedate_tz(date_str
)
1686 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1689 def determine_ext(url
, default_ext
='unknown_video'):
1690 if url
is None or '.' not in url
:
1692 guess
= url
.partition('?')[0].rpartition('.')[2]
1693 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1695 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1696 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1697 return guess
.rstrip('/')
1702 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1703 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1706 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1708 Return a datetime object from a string.
1710 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1712 @param format strftime format of DATE
1713 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1714 auto: round to the unit provided in date_str (if applicable).
1716 auto_precision
= False
1717 if precision
== 'auto':
1718 auto_precision
= True
1719 precision
= 'microsecond'
1720 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1721 if date_str
in ('now', 'today'):
1723 if date_str
== 'yesterday':
1724 return today
- datetime
.timedelta(days
=1)
1726 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1728 if match
is not None:
1729 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1730 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1731 unit
= match
.group('unit')
1732 if unit
== 'month' or unit
== 'year':
1733 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1739 delta
= datetime
.timedelta(**{unit + 's': time}
)
1740 new_date
= start_time
+ delta
1742 return datetime_round(new_date
, unit
)
1745 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1748 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1750 Return a date object from a string using datetime_from_str
1752 @param strict Restrict allowed patterns to "YYYYMMDD" and
1753 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1755 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1756 raise ValueError(f
'Invalid date format "{date_str}"')
1757 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1760 def datetime_add_months(dt
, months
):
1761 """Increment/Decrement a datetime object by months."""
1762 month
= dt
.month
+ months
- 1
1763 year
= dt
.year
+ month
// 12
1764 month
= month
% 12 + 1
1765 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1766 return dt
.replace(year
, month
, day
)
1769 def datetime_round(dt
, precision
='day'):
1771 Round a datetime object's time to a specific precision
1773 if precision
== 'microsecond':
1782 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1783 timestamp
= calendar
.timegm(dt
.timetuple())
1784 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1787 def hyphenate_date(date_str
):
1789 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1790 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1791 if match
is not None:
1792 return '-'.join(match
.groups())
1798 """Represents a time interval between two dates"""
1800 def __init__(self
, start
=None, end
=None):
1801 """start and end must be strings in the format accepted by date"""
1802 if start
is not None:
1803 self
.start
= date_from_str(start
, strict
=True)
1805 self
.start
= datetime
.datetime
.min.date()
1807 self
.end
= date_from_str(end
, strict
=True)
1809 self
.end
= datetime
.datetime
.max.date()
1810 if self
.start
> self
.end
:
1811 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1815 """Returns a range that only contains the given day"""
1816 return cls(day
, day
)
1818 def __contains__(self
, date
):
1819 """Check if the date is in the range"""
1820 if not isinstance(date
, datetime
.date
):
1821 date
= date_from_str(date
)
1822 return self
.start
<= date
<= self
.end
1825 return f
'{__name__}.{type(self).__name__}({self.start.isoformat()!r}, {self.end.isoformat()!r})'
1827 def __eq__(self
, other
):
1828 return (isinstance(other
, DateRange
)
1829 and self
.start
== other
.start
and self
.end
== other
.end
)
1833 def system_identifier():
1834 python_implementation
= platform
.python_implementation()
1835 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
1836 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
1838 with contextlib
.suppress(OSError): # We may not have access to the executable
1839 libc_ver
= platform
.libc_ver()
1841 return 'Python %s (%s %s %s) - %s (%s%s)' % (
1842 platform
.python_version(),
1843 python_implementation
,
1845 platform
.architecture()[0],
1846 platform
.platform(),
1847 ssl
.OPENSSL_VERSION
,
1848 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
1853 def get_windows_version():
1854 ''' Get Windows version. returns () if it's not running on Windows '''
1855 if compat_os_name
== 'nt':
1856 return version_tuple(platform
.win32_ver()[1])
1861 def write_string(s
, out
=None, encoding
=None):
1862 assert isinstance(s
, str)
1863 out
= out
or sys
.stderr
1864 # `sys.stderr` might be `None` (Ref: https://github.com/pyinstaller/pyinstaller/pull/7217)
1868 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
1869 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1871 enc
, buffer = None, out
1872 if 'b' in getattr(out
, 'mode', ''):
1873 enc
= encoding
or preferredencoding()
1874 elif hasattr(out
, 'buffer'):
1876 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1878 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
1882 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
1883 from .. import _IN_CLI
1885 if msg
in deprecation_warning
._cache
:
1887 deprecation_warning
._cache
.add(msg
)
1889 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
1890 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
1893 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
1896 deprecation_warning
._cache
= set()
1899 def bytes_to_intlist(bs
):
1902 if isinstance(bs
[0], int): # Python 3
1905 return [ord(c
) for c
in bs
]
1908 def intlist_to_bytes(xs
):
1911 return struct
.pack('%dB' % len(xs
), *xs
)
1914 class LockingUnsupportedError(OSError):
1915 msg
= 'File locking is not supported'
1918 super().__init
__(self
.msg
)
1921 # Cross-platform file locking
1922 if sys
.platform
== 'win32':
1924 import ctypes
.wintypes
1927 class OVERLAPPED(ctypes
.Structure
):
1929 ('Internal', ctypes
.wintypes
.LPVOID
),
1930 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1931 ('Offset', ctypes
.wintypes
.DWORD
),
1932 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1933 ('hEvent', ctypes
.wintypes
.HANDLE
),
1936 kernel32
= ctypes
.WinDLL('kernel32')
1937 LockFileEx
= kernel32
.LockFileEx
1938 LockFileEx
.argtypes
= [
1939 ctypes
.wintypes
.HANDLE
, # hFile
1940 ctypes
.wintypes
.DWORD
, # dwFlags
1941 ctypes
.wintypes
.DWORD
, # dwReserved
1942 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1943 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1944 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1946 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
1947 UnlockFileEx
= kernel32
.UnlockFileEx
1948 UnlockFileEx
.argtypes
= [
1949 ctypes
.wintypes
.HANDLE
, # hFile
1950 ctypes
.wintypes
.DWORD
, # dwReserved
1951 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1952 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1953 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1955 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
1956 whole_low
= 0xffffffff
1957 whole_high
= 0x7fffffff
1959 def _lock_file(f
, exclusive
, block
):
1960 overlapped
= OVERLAPPED()
1961 overlapped
.Offset
= 0
1962 overlapped
.OffsetHigh
= 0
1963 overlapped
.hEvent
= 0
1964 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
1966 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
1967 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
1968 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
1969 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
1970 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
1972 def _unlock_file(f
):
1973 assert f
._lock
_file
_overlapped
_p
1974 handle
= msvcrt
.get_osfhandle(f
.fileno())
1975 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
1976 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
1982 def _lock_file(f
, exclusive
, block
):
1983 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
1985 flags |
= fcntl
.LOCK_NB
1987 fcntl
.flock(f
, flags
)
1988 except BlockingIOError
:
1990 except OSError: # AOSP does not have flock()
1991 fcntl
.lockf(f
, flags
)
1993 def _unlock_file(f
):
1994 with contextlib
.suppress(OSError):
1995 return fcntl
.flock(f
, fcntl
.LOCK_UN
)
1996 with contextlib
.suppress(OSError):
1997 return fcntl
.lockf(f
, fcntl
.LOCK_UN
) # AOSP does not have flock()
1998 return fcntl
.flock(f
, fcntl
.LOCK_UN | fcntl
.LOCK_NB
) # virtiofs needs LOCK_NB on unlocking
2002 def _lock_file(f
, exclusive
, block
):
2003 raise LockingUnsupportedError()
2005 def _unlock_file(f
):
2006 raise LockingUnsupportedError()
2012 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2013 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2014 raise NotImplementedError(mode
)
2015 self
.mode
, self
.block
= mode
, block
2017 writable
= any(f
in mode
for f
in 'wax+')
2018 readable
= any(f
in mode
for f
in 'r+')
2019 flags
= functools
.reduce(operator
.ior
, (
2020 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2021 getattr(os
, 'O_BINARY', 0), # Windows only
2022 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2023 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2024 os
.O_APPEND
if 'a' in mode
else 0,
2025 os
.O_EXCL
if 'x' in mode
else 0,
2026 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2029 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2031 def __enter__(self
):
2032 exclusive
= 'r' not in self
.mode
2034 _lock_file(self
.f
, exclusive
, self
.block
)
2039 if 'w' in self
.mode
:
2042 except OSError as e
:
2044 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2045 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2054 _unlock_file(self
.f
)
2058 def __exit__(self
, *_
):
2067 def __getattr__(self
, attr
):
2068 return getattr(self
.f
, attr
)
2075 def get_filesystem_encoding():
2076 encoding
= sys
.getfilesystemencoding()
2077 return encoding
if encoding
is not None else 'utf-8'
2080 def shell_quote(args
):
2082 encoding
= get_filesystem_encoding()
2084 if isinstance(a
, bytes):
2085 # We may get a filename encoded with 'encodeFilename'
2086 a
= a
.decode(encoding
)
2087 quoted_args
.append(compat_shlex_quote(a
))
2088 return ' '.join(quoted_args
)
2091 def smuggle_url(url
, data
):
2092 """ Pass additional data in a URL for internal use. """
2094 url
, idata
= unsmuggle_url(url
, {})
2096 sdata
= urllib
.parse
.urlencode(
2097 {'__youtubedl_smuggle': json.dumps(data)}
)
2098 return url
+ '#' + sdata
2101 def unsmuggle_url(smug_url
, default
=None):
2102 if '#__youtubedl_smuggle' not in smug_url
:
2103 return smug_url
, default
2104 url
, _
, sdata
= smug_url
.rpartition('#')
2105 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2106 data
= json
.loads(jsond
)
2110 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2111 """ Formats numbers with decimal sufixes like K, M, etc """
2112 num
, factor
= float_or_none(num
), float(factor
)
2113 if num
is None or num
< 0:
2115 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2116 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2117 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2119 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2120 converted
= num
/ (factor
** exponent
)
2121 return fmt
% (converted
, suffix
)
2124 def format_bytes(bytes):
2125 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2128 def lookup_unit_table(unit_table
, s
, strict
=False):
2129 num_re
= NUMBER_RE
if strict
else NUMBER_RE
.replace(R
'\.', '[,.]')
2130 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2131 m
= (re
.fullmatch
if strict
else re
.match
)(
2132 rf
'(?P<num>{num_re})\s*(?P<unit>{units_re})\b', s
)
2136 num
= float(m
.group('num').replace(',', '.'))
2137 mult
= unit_table
[m
.group('unit')]
2138 return round(num
* mult
)
2142 """Parse a string indicating a byte quantity into an integer"""
2143 return lookup_unit_table(
2144 {u: 1024**i for i, u in enumerate(['', *'KMGTPEZY'])}
,
2145 s
.upper(), strict
=True)
2148 def parse_filesize(s
):
2152 # The lower-case forms are of course incorrect and unofficial,
2153 # but we support those too
2170 'megabytes': 1000 ** 2,
2171 'mebibytes': 1024 ** 2,
2177 'gigabytes': 1000 ** 3,
2178 'gibibytes': 1024 ** 3,
2184 'terabytes': 1000 ** 4,
2185 'tebibytes': 1024 ** 4,
2191 'petabytes': 1000 ** 5,
2192 'pebibytes': 1024 ** 5,
2198 'exabytes': 1000 ** 6,
2199 'exbibytes': 1024 ** 6,
2205 'zettabytes': 1000 ** 7,
2206 'zebibytes': 1024 ** 7,
2212 'yottabytes': 1000 ** 8,
2213 'yobibytes': 1024 ** 8,
2216 return lookup_unit_table(_UNIT_TABLE
, s
)
2223 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2225 if re
.match(r
'^[\d,.]+$', s
):
2226 return str_to_int(s
)
2239 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2243 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2245 return str_to_int(mobj
.group(1))
2248 def parse_resolution(s
, *, lenient
=False):
2253 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2255 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2258 'width': int(mobj
.group('w')),
2259 'height': int(mobj
.group('h')),
2262 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2264 return {'height': int(mobj.group(1))}
2266 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2268 return {'height': int(mobj.group(1)) * 540}
2273 def parse_bitrate(s
):
2274 if not isinstance(s
, str):
2276 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2278 return int(mobj
.group(1))
2281 def month_by_name(name
, lang
='en'):
2282 """ Return the number of a month by (locale-independently) English name """
2284 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2287 return month_names
.index(name
) + 1
2292 def month_by_abbreviation(abbrev
):
2293 """ Return the number of a month by (locale-independently) English
2297 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2302 def fix_xml_ampersands(xml_str
):
2303 """Replace all the '&' by '&' in XML"""
2305 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2310 def setproctitle(title
):
2311 assert isinstance(title
, str)
2313 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2320 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2324 # LoadLibrary in Windows Python 2.7.13 only expects
2325 # a bytestring, but since unicode_literals turns
2326 # every string into a unicode string, it fails.
2328 title_bytes
= title
.encode()
2329 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2330 buf
.value
= title_bytes
2332 libc
.prctl(15, buf
, 0, 0, 0)
2333 except AttributeError:
2334 return # Strange libc, just skip this
2337 def remove_start(s
, start
):
2338 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2341 def remove_end(s
, end
):
2342 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2345 def remove_quotes(s
):
2346 if s
is None or len(s
) < 2:
2348 for quote
in ('"', "'", ):
2349 if s
[0] == quote
and s
[-1] == quote
:
2354 def get_domain(url
):
2356 This implementation is inconsistent, but is kept for compatibility.
2357 Use this only for "webpage_url_domain"
2359 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2362 def url_basename(url
):
2363 path
= urllib
.parse
.urlparse(url
).path
2364 return path
.strip('/').split('/')[-1]
2368 return re
.match(r
'https?://[^?#]+/', url
).group()
2371 def urljoin(base
, path
):
2372 if isinstance(path
, bytes):
2373 path
= path
.decode()
2374 if not isinstance(path
, str) or not path
:
2376 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2378 if isinstance(base
, bytes):
2379 base
= base
.decode()
2380 if not isinstance(base
, str) or not re
.match(
2381 r
'^(?:https?:)?//', base
):
2383 return urllib
.parse
.urljoin(base
, path
)
2386 class HEADRequest(urllib
.request
.Request
):
2387 def get_method(self
):
2391 class PUTRequest(urllib
.request
.Request
):
2392 def get_method(self
):
2396 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2397 if get_attr
and v
is not None:
2398 v
= getattr(v
, get_attr
, None)
2400 return int(v
) * invscale
// scale
2401 except (ValueError, TypeError, OverflowError):
2405 def str_or_none(v
, default
=None):
2406 return default
if v
is None else str(v
)
2409 def str_to_int(int_str
):
2410 """ A more relaxed version of int_or_none """
2411 if isinstance(int_str
, int):
2413 elif isinstance(int_str
, str):
2414 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2415 return int_or_none(int_str
)
2418 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2422 return float(v
) * invscale
/ scale
2423 except (ValueError, TypeError):
2427 def bool_or_none(v
, default
=None):
2428 return v
if isinstance(v
, bool) else default
2431 def strip_or_none(v
, default
=None):
2432 return v
.strip() if isinstance(v
, str) else default
2435 def url_or_none(url
):
2436 if not url
or not isinstance(url
, str):
2439 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2442 def request_to_url(req
):
2443 if isinstance(req
, urllib
.request
.Request
):
2444 return req
.get_full_url()
2449 def strftime_or_none(timestamp
, date_format
, default
=None):
2450 datetime_object
= None
2452 if isinstance(timestamp
, (int, float)): # unix timestamp
2453 # Using naive datetime here can break timestamp() in Windows
2454 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2455 # Also, datetime.datetime.fromtimestamp breaks for negative timestamps
2456 # Ref: https://github.com/yt-dlp/yt-dlp/issues/6706#issuecomment-1496842642
2457 datetime_object
= (datetime
.datetime
.fromtimestamp(0, datetime
.timezone
.utc
)
2458 + datetime
.timedelta(seconds
=timestamp
))
2459 elif isinstance(timestamp
, str): # assume YYYYMMDD
2460 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2461 date_format
= re
.sub( # Support %s on windows
2462 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2463 return datetime_object
.strftime(date_format
)
2464 except (ValueError, TypeError, AttributeError):
2468 def parse_duration(s
):
2469 if not isinstance(s
, str):
2475 days
, hours
, mins
, secs
, ms
= [None] * 5
2476 m
= re
.match(r
'''(?x)
2478 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2479 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2480 (?P<ms>[.:][0-9]+)?Z?$
2483 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2488 [0-9]+\s*y(?:ears?)?,?\s*
2491 [0-9]+\s*m(?:onths?)?,?\s*
2494 [0-9]+\s*w(?:eeks?)?,?\s*
2497 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2501 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2504 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2507 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2510 days
, hours
, mins
, secs
, ms
= m
.groups()
2512 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2514 hours
, mins
= m
.groups()
2519 ms
= ms
.replace(':', '.')
2520 return sum(float(part
or 0) * mult
for part
, mult
in (
2521 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2524 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2525 name
, real_ext
= os
.path
.splitext(filename
)
2527 f
'{name}.{ext}{real_ext}'
2528 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2529 else f
'{filename}.{ext}')
2532 def replace_extension(filename
, ext
, expected_real_ext
=None):
2533 name
, real_ext
= os
.path
.splitext(filename
)
2534 return '{}.{}'.format(
2535 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2539 def check_executable(exe
, args
=[]):
2540 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2541 args can be a list of arguments for a short output (like -version) """
2543 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2549 def _get_exe_version_output(exe
, args
):
2551 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2552 # SIGTTOU if yt-dlp is run in the background.
2553 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2554 stdout
, _
, ret
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2555 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2563 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2564 assert isinstance(output
, str)
2565 if version_re
is None:
2566 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2567 m
= re
.search(version_re
, output
)
2574 def get_exe_version(exe
, args
=['--version'],
2575 version_re
=None, unrecognized
=('present', 'broken')):
2576 """ Returns the version of the specified executable,
2577 or False if the executable is not present """
2578 unrecognized
= variadic(unrecognized
)
2579 assert len(unrecognized
) in (1, 2)
2580 out
= _get_exe_version_output(exe
, args
)
2582 return unrecognized
[-1]
2583 return out
and detect_exe_version(out
, version_re
, unrecognized
[0])
2586 def frange(start
=0, stop
=None, step
=1):
2589 start
, stop
= 0, start
2590 sign
= [-1, 1][step
> 0] if step
else 0
2591 while sign
* start
< sign
* stop
:
2596 class LazyList(collections
.abc
.Sequence
):
2597 """Lazy immutable list from an iterable
2598 Note that slices of a LazyList are lists and not LazyList"""
2600 class IndexError(IndexError):
2603 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2604 self
._iterable
= iter(iterable
)
2605 self
._cache
= [] if _cache
is None else _cache
2606 self
._reversed
= reverse
2610 # We need to consume the entire iterable to iterate in reverse
2611 yield from self
.exhaust()
2613 yield from self
._cache
2614 for item
in self
._iterable
:
2615 self
._cache
.append(item
)
2619 self
._cache
.extend(self
._iterable
)
2620 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2624 """Evaluate the entire iterable"""
2625 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2628 def _reverse_index(x
):
2629 return None if x
is None else ~x
2631 def __getitem__(self
, idx
):
2632 if isinstance(idx
, slice):
2634 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2635 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2636 elif isinstance(idx
, int):
2638 idx
= self
._reverse
_index
(idx
)
2639 start
, stop
, step
= idx
, idx
, 0
2641 raise TypeError('indices must be integers or slices')
2642 if ((start
or 0) < 0 or (stop
or 0) < 0
2643 or (start
is None and step
< 0)
2644 or (stop
is None and step
> 0)):
2645 # We need to consume the entire iterable to be able to slice from the end
2646 # Obviously, never use this with infinite iterables
2649 return self
._cache
[idx
]
2650 except IndexError as e
:
2651 raise self
.IndexError(e
) from e
2652 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2654 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2656 return self
._cache
[idx
]
2657 except IndexError as e
:
2658 raise self
.IndexError(e
) from e
2662 self
[-1] if self
._reversed
else self
[0]
2663 except self
.IndexError:
2669 return len(self
._cache
)
2671 def __reversed__(self
):
2672 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2675 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2678 # repr and str should mimic a list. So we exhaust the iterable
2679 return repr(self
.exhaust())
2682 return repr(self
.exhaust())
2687 class IndexError(IndexError):
2691 # This is only useful for tests
2692 return len(self
.getslice())
2694 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2695 self
._pagefunc
= pagefunc
2696 self
._pagesize
= pagesize
2697 self
._pagecount
= float('inf')
2698 self
._use
_cache
= use_cache
2701 def getpage(self
, pagenum
):
2702 page_results
= self
._cache
.get(pagenum
)
2703 if page_results
is None:
2704 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2706 self
._cache
[pagenum
] = page_results
2709 def getslice(self
, start
=0, end
=None):
2710 return list(self
._getslice
(start
, end
))
2712 def _getslice(self
, start
, end
):
2713 raise NotImplementedError('This method must be implemented by subclasses')
2715 def __getitem__(self
, idx
):
2716 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2717 if not isinstance(idx
, int) or idx
< 0:
2718 raise TypeError('indices must be non-negative integers')
2719 entries
= self
.getslice(idx
, idx
+ 1)
2721 raise self
.IndexError()
2725 class OnDemandPagedList(PagedList
):
2726 """Download pages until a page with less than maximum results"""
2728 def _getslice(self
, start
, end
):
2729 for pagenum
in itertools
.count(start
// self
._pagesize
):
2730 firstid
= pagenum
* self
._pagesize
2731 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2732 if start
>= nextfirstid
:
2736 start
% self
._pagesize
2737 if firstid
<= start
< nextfirstid
2740 ((end
- 1) % self
._pagesize
) + 1
2741 if (end
is not None and firstid
<= end
<= nextfirstid
)
2745 page_results
= self
.getpage(pagenum
)
2747 self
._pagecount
= pagenum
- 1
2749 if startv
!= 0 or endv
is not None:
2750 page_results
= page_results
[startv
:endv
]
2751 yield from page_results
2753 # A little optimization - if current page is not "full", ie. does
2754 # not contain page_size videos then we can assume that this page
2755 # is the last one - there are no more ids on further pages -
2756 # i.e. no need to query again.
2757 if len(page_results
) + startv
< self
._pagesize
:
2760 # If we got the whole page, but the next page is not interesting,
2761 # break out early as well
2762 if end
== nextfirstid
:
2766 class InAdvancePagedList(PagedList
):
2767 """PagedList with total number of pages known in advance"""
2769 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2770 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2771 self
._pagecount
= pagecount
2773 def _getslice(self
, start
, end
):
2774 start_page
= start
// self
._pagesize
2775 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2776 skip_elems
= start
- start_page
* self
._pagesize
2777 only_more
= None if end
is None else end
- start
2778 for pagenum
in range(start_page
, end_page
):
2779 page_results
= self
.getpage(pagenum
)
2781 page_results
= page_results
[skip_elems
:]
2783 if only_more
is not None:
2784 if len(page_results
) < only_more
:
2785 only_more
-= len(page_results
)
2787 yield from page_results
[:only_more
]
2789 yield from page_results
2792 class PlaylistEntries
:
2793 MissingEntry
= object()
2794 is_exhausted
= False
2796 def __init__(self
, ydl
, info_dict
):
2799 # _entries must be assigned now since infodict can change during iteration
2800 entries
= info_dict
.get('entries')
2802 raise EntryNotInPlaylist('There are no entries')
2803 elif isinstance(entries
, list):
2804 self
.is_exhausted
= True
2806 requested_entries
= info_dict
.get('requested_entries')
2807 self
.is_incomplete
= requested_entries
is not None
2808 if self
.is_incomplete
:
2809 assert self
.is_exhausted
2810 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
2811 for i
, entry
in zip(requested_entries
, entries
):
2812 self
._entries
[i
- 1] = entry
2813 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2814 self
._entries
= entries
2816 self
._entries
= LazyList(entries
)
2818 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2819 (?P<start>[+-]?\d+)?
2821 (?P<end>[+-]?\d+|inf(?:inite)?)?
2822 (?::(?P<step>[+-]?\d+))?
2826 def parse_playlist_items(cls
, string
):
2827 for segment
in string
.split(','):
2829 raise ValueError('There is two or more consecutive commas')
2830 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2832 raise ValueError(f
'{segment!r} is not a valid specification')
2833 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2834 if int_or_none(step
) == 0:
2835 raise ValueError(f
'Step in {segment!r} cannot be zero')
2836 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2838 def get_requested_items(self
):
2839 playlist_items
= self
.ydl
.params
.get('playlist_items')
2840 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2841 playlist_end
= self
.ydl
.params
.get('playlistend')
2842 # For backwards compatibility, interpret -1 as whole list
2843 if playlist_end
in (-1, None):
2845 if not playlist_items
:
2846 playlist_items
= f
'{playlist_start}:{playlist_end}'
2847 elif playlist_start
!= 1 or playlist_end
:
2848 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2850 for index
in self
.parse_playlist_items(playlist_items
):
2851 for i
, entry
in self
[index
]:
2856 # The item may have just been added to archive. Don't break due to it
2857 if not self
.ydl
.params
.get('lazy_playlist'):
2858 # TODO: Add auto-generated fields
2859 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
2860 except (ExistingVideoReached
, RejectedVideoReached
):
2863 def get_full_count(self
):
2864 if self
.is_exhausted
and not self
.is_incomplete
:
2866 elif isinstance(self
._entries
, InAdvancePagedList
):
2867 if self
._entries
._pagesize
== 1:
2868 return self
._entries
._pagecount
2870 @functools.cached_property
2872 if isinstance(self
._entries
, list):
2875 entry
= self
._entries
[i
]
2877 entry
= self
.MissingEntry
2878 if not self
.is_incomplete
:
2879 raise self
.IndexError()
2880 if entry
is self
.MissingEntry
:
2881 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
2886 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
2887 except (LazyList
.IndexError, PagedList
.IndexError):
2888 raise self
.IndexError()
2891 def __getitem__(self
, idx
):
2892 if isinstance(idx
, int):
2893 idx
= slice(idx
, idx
)
2895 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
2896 step
= 1 if idx
.step
is None else idx
.step
2897 if idx
.start
is None:
2898 start
= 0 if step
> 0 else len(self
) - 1
2900 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
2902 # NB: Do not call len(self) when idx == [:]
2903 if idx
.stop
is None:
2904 stop
= 0 if step
< 0 else float('inf')
2906 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
2907 stop
+= [-1, 1][step
> 0]
2909 for i
in frange(start
, stop
, step
):
2913 entry
= self
._getter
(i
)
2914 except self
.IndexError:
2915 self
.is_exhausted
= True
2922 return len(tuple(self
[:]))
2924 class IndexError(IndexError):
2928 def uppercase_escape(s
):
2929 unicode_escape
= codecs
.getdecoder('unicode_escape')
2931 r
'\\U[0-9a-fA-F]{8}',
2932 lambda m
: unicode_escape(m
.group(0))[0],
2936 def lowercase_escape(s
):
2937 unicode_escape
= codecs
.getdecoder('unicode_escape')
2939 r
'\\u[0-9a-fA-F]{4}',
2940 lambda m
: unicode_escape(m
.group(0))[0],
2944 def escape_rfc3986(s
):
2945 """Escape non-ASCII characters as suggested by RFC 3986"""
2946 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2949 def escape_url(url
):
2950 """Escape URL as suggested by RFC 3986"""
2951 url_parsed
= urllib
.parse
.urlparse(url
)
2952 return url_parsed
._replace
(
2953 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2954 path
=escape_rfc3986(url_parsed
.path
),
2955 params
=escape_rfc3986(url_parsed
.params
),
2956 query
=escape_rfc3986(url_parsed
.query
),
2957 fragment
=escape_rfc3986(url_parsed
.fragment
)
2961 def parse_qs(url
, **kwargs
):
2962 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
2965 def read_batch_urls(batch_fd
):
2967 if not isinstance(url
, str):
2968 url
= url
.decode('utf-8', 'replace')
2969 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
2970 for bom
in BOM_UTF8
:
2971 if url
.startswith(bom
):
2972 url
= url
[len(bom
):]
2974 if not url
or url
.startswith(('#', ';', ']')):
2976 # "#" cannot be stripped out since it is part of the URI
2977 # However, it can be safely stripped out if following a whitespace
2978 return re
.split(r
'\s#', url
, 1)[0].rstrip()
2980 with contextlib
.closing(batch_fd
) as fd
:
2981 return [url
for url
in map(fixup
, fd
) if url
]
2984 def urlencode_postdata(*args
, **kargs
):
2985 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
2988 def update_url(url
, *, query_update
=None, **kwargs
):
2989 """Replace URL components specified by kwargs
2990 @param url str or parse url tuple
2991 @param query_update update query
2994 if isinstance(url
, str):
2995 if not kwargs
and not query_update
:
2998 url
= urllib
.parse
.urlparse(url
)
3000 assert 'query' not in kwargs
, 'query_update and query cannot be specified at the same time'
3001 kwargs
['query'] = urllib
.parse
.urlencode({
3002 **urllib
.parse
.parse_qs(url
.query
),
3005 return urllib
.parse
.urlunparse(url
._replace
(**kwargs
))
3008 def update_url_query(url
, query
):
3009 return update_url(url
, query_update
=query
)
3012 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3013 req_headers
= req
.headers
.copy()
3014 req_headers
.update(headers
or {})
3015 req_data
= data
or req
.data
3016 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3017 req_get_method
= req
.get_method()
3018 if req_get_method
== 'HEAD':
3019 req_type
= HEADRequest
3020 elif req_get_method
== 'PUT':
3021 req_type
= PUTRequest
3023 req_type
= urllib
.request
.Request
3025 req_url
, data
=req_data
, headers
=req_headers
,
3026 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3027 if hasattr(req
, 'timeout'):
3028 new_req
.timeout
= req
.timeout
3032 def _multipart_encode_impl(data
, boundary
):
3033 content_type
= 'multipart/form-data; boundary=%s' % boundary
3036 for k
, v
in data
.items():
3037 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3038 if isinstance(k
, str):
3040 if isinstance(v
, str):
3042 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3043 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3044 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3045 if boundary
.encode('ascii') in content
:
3046 raise ValueError('Boundary overlaps with data')
3049 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3051 return out
, content_type
3054 def multipart_encode(data
, boundary
=None):
3056 Encode a dict to RFC 7578-compliant form-data
3059 A dict where keys and values can be either Unicode or bytes-like
3062 If specified a Unicode object, it's used as the boundary. Otherwise
3063 a random boundary is generated.
3065 Reference: https://tools.ietf.org/html/rfc7578
3067 has_specified_boundary
= boundary
is not None
3070 if boundary
is None:
3071 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3074 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3077 if has_specified_boundary
:
3081 return out
, content_type
3084 def is_iterable_like(x
, allowed_types
=collections
.abc
.Iterable
, blocked_types
=NO_DEFAULT
):
3085 if blocked_types
is NO_DEFAULT
:
3086 blocked_types
= (str, bytes, collections
.abc
.Mapping
)
3087 return isinstance(x
, allowed_types
) and not isinstance(x
, blocked_types
)
3090 def variadic(x
, allowed_types
=NO_DEFAULT
):
3091 if not isinstance(allowed_types
, (tuple, type)):
3092 deprecation_warning('allowed_types should be a tuple or a type')
3093 allowed_types
= tuple(allowed_types
)
3094 return x
if is_iterable_like(x
, blocked_types
=allowed_types
) else (x
, )
3097 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3100 val
= f(*args
, **kwargs
)
3101 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3104 if expected_type
is None or isinstance(val
, expected_type
):
3108 def try_get(src
, getter
, expected_type
=None):
3109 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3112 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3113 return {k: v for k, v in dct.items() if cndn(k, v)}
3116 def merge_dicts(*dicts
):
3118 for a_dict
in dicts
:
3119 for k
, v
in a_dict
.items():
3120 if (v
is not None and k
not in merged
3121 or isinstance(v
, str) and merged
[k
] == ''):
3126 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3127 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3139 TV_PARENTAL_GUIDELINES
= {
3149 def parse_age_limit(s
):
3150 # isinstance(False, int) is True. So type() must be used instead
3151 if type(s
) is int: # noqa: E721
3152 return s
if 0 <= s
<= 21 else None
3153 elif not isinstance(s
, str):
3155 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3157 return int(m
.group('age'))
3160 return US_RATINGS
[s
]
3161 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3163 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3167 def strip_jsonp(code
):
3170 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3171 (?:\s*&&\s*(?P=func_name))?
3172 \s*\(\s*(?P<callback_data>.*)\);?
3173 \s*?(?://[^\n]*)*$''',
3174 r
'\g<callback_data>', code
)
3177 def js_to_json(code
, vars={}, *, strict
=False):
3178 # vars is a dict of var, val pairs to substitute
3179 STRING_QUOTES
= '\'"`'
3180 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3181 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3182 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3184 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3185 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3188 def process_escape(match
):
3189 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3190 escape
= match
.group(1) or match
.group(2)
3192 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3193 else R
'\u00' if escape
== 'x'
3194 else '' if escape
== '\n'
3197 def template_substitute(match
):
3198 evaluated
= js_to_json(match
.group(1), vars, strict
=strict
)
3199 if evaluated
[0] == '"':
3200 return json
.loads(evaluated
)
3205 if v
in ('true', 'false', 'null'):
3207 elif v
in ('undefined', 'void 0'):
3209 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3212 if v
[0] in STRING_QUOTES
:
3213 v
= re
.sub(r
'(?s)\${([^}]+)}', template_substitute
, v
[1:-1]) if v
[0] == '`' else v
[1:-1]
3214 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
)
3215 return f
'"{escaped}"'
3217 for regex
, base
in INTEGER_TABLE
:
3218 im
= re
.match(regex
, v
)
3220 i
= int(im
.group(1), base
)
3221 return f
'"{i}":' if v
.endswith(':') else str(i
)
3227 except json
.JSONDecodeError
:
3228 return json
.dumps(vars[v
])
3235 raise ValueError(f
'Unknown value: {v}')
3237 def create_map(mobj
):
3238 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3240 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3242 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3243 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3244 code
= re
.sub(r
'parseInt\([^\d]+(\d+)[^\d]+\)', r
'\1', code
)
3245 code
= re
.sub(r
'\(function\([^)]*\)\s*\{[^}]*\}\s*\)\s*\(\s*(["\'][^
)]*["\'])\s*\)', r'\1', code)
3247 return re.sub(rf'''(?sx)
3249 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3250 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3251 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3252 [0-9]+(?={SKIP_RE}:)|
3257 def qualities(quality_ids):
3258 """ Get a numeric quality value out of a list of possible values """
3261 return quality_ids.index(qid)
3267 POSTPROCESS_WHEN = ('pre_process', 'after_filter', 'video', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3271 'default': '%(title)s [%(id)s].%(ext)s',
3272 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3278 'description': 'description',
3279 'annotation': 'annotations.xml',
3280 'infojson': 'info.json',
3283 'pl_thumbnail': None,
3284 'pl_description': 'description',
3285 'pl_infojson': 'info.json',
3288 # As of [1] format syntax is:
3289 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3290 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3291 STR_FORMAT_RE_TMPL = r'''(?x)
3292 (?<!%)(?P<prefix>(?:%%)*)
3294 (?P<has_key>\((?P<key>{0})\))?
3296 (?P<conversion>[#0\-+ ]+)?
3298 (?P<precision>\.\d+)?
3299 (?P<len_mod>[hlL])? # unused in python
3300 {1} # conversion type
3305 STR_FORMAT_TYPES = 'diouxXeEfFgGcrs'
3308 def limit_length(s, length):
3309 """ Add ellipses to overly long strings """
3314 return s[:length - len(ELLIPSES)] + ELLIPSES
3318 def version_tuple(v):
3319 return tuple(int(e) for e in re.split(r'[-.]', v))
3322 def is_outdated_version(version, limit, assume_new=True):
3324 return not assume_new
3326 return version_tuple(version) < version_tuple(limit)
3328 return not assume_new
3331 def ytdl_is_updateable():
3332 """ Returns if yt-dlp can be updated with -U """
3334 from ..update import is_non_updateable
3336 return not is_non_updateable()
3339 def args_to_str(args):
3340 # Get a short string representation for a subprocess command
3341 return ' '.join(compat_shlex_quote(a) for a in args)
3344 def error_to_str(err):
3345 return f'{type(err).__name__}: {err}'
3348 def mimetype2ext(mt, default=NO_DEFAULT):
3349 if not isinstance(mt, str):
3350 if default is not NO_DEFAULT:
3366 'x-matroska': 'mkv',
3368 'x-mp4-fragmented': 'mp4',
3373 # application (streaming playlists)
3377 'vnd.apple.mpegurl': 'm3u8',
3378 'vnd.ms-sstr+xml': 'ism',
3379 'x-mpegurl': 'm3u8',
3383 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3.
3384 # Using .mp3 as it's the most popular one
3385 'audio/mpeg': 'mp3',
3386 'audio/webm': 'webm',
3387 'audio/x-matroska': 'mka',
3388 'audio/x-mpegurl': 'm3u',
3396 'x-realaudio': 'ra',
3407 'vnd.wap.wbmp': 'wbmp',
3414 'filmstrip+json': 'fs',
3415 'smptett+xml': 'tt',
3418 'x-ms-sami': 'sami',
3427 mimetype = mt.partition(';')[0].strip().lower()
3428 _, _, subtype = mimetype.rpartition('/')
3430 ext = traversal.traverse_obj(MAP, mimetype, subtype, subtype.rsplit('+')[-1])
3433 elif default is not NO_DEFAULT:
3435 return subtype.replace('+', '.')
3438 def ext2mimetype(ext_or_url):
3441 if '.' not in ext_or_url:
3442 ext_or_url = f'file.{ext_or_url}'
3443 return mimetypes.guess_type(ext_or_url)[0]
3446 def parse_codecs(codecs_str):
3447 # http://tools.ietf.org/html/rfc6381
3450 split_codecs = list(filter(None, map(
3451 str.strip, codecs_str.strip().strip(',').split(','))))
3452 vcodec, acodec, scodec, hdr = None, None, None, None
3453 for full_codec in split_codecs:
3454 parts = re.sub(r'0+(?=\d)', '', full_codec).split('.')
3455 if parts[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3456 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3460 if parts[0] in ('dvh1', 'dvhe'):
3462 elif parts[0] == 'av1' and traversal.traverse_obj(parts, 3) == '10':
3464 elif parts[:2] == ['vp9', '2']:
3466 elif parts[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-4',
3467 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3468 acodec = acodec or full_codec
3469 elif parts[0] in ('stpp', 'wvtt'):
3470 scodec = scodec or full_codec
3472 write_string(f'WARNING: Unknown codec {full_codec}\n')
3473 if vcodec or acodec or scodec:
3475 'vcodec': vcodec or 'none',
3476 'acodec': acodec or 'none',
3477 'dynamic_range': hdr,
3478 **({'scodec': scodec} if scodec is not None else {}),
3480 elif len(split_codecs) == 2:
3482 'vcodec': split_codecs[0],
3483 'acodec': split_codecs[1],
3488 def get_compatible_ext(*, vcodecs, acodecs, vexts, aexts, preferences=None):
3489 assert len(vcodecs) == len(vexts) and len(acodecs) == len(aexts)
3491 allow_mkv = not preferences or 'mkv' in preferences
3493 if allow_mkv and max(len(acodecs), len(vcodecs)) > 1:
3494 return 'mkv' # TODO: any other format allows this?
3496 # TODO: All codecs supported by parse_codecs isn't handled here
3497 COMPATIBLE_CODECS = {
3499 'av1', 'hevc', 'avc1', 'mp4a', 'ac-4', # fourcc (m3u8, mpd)
3500 'h264', 'aacl', 'ec-3', # Set in ISM
3503 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3504 'vp9x', 'vp8x', # in the webm spec
3508 sanitize_codec = functools.partial(try_get, getter=lambda x: x[0].split('.')[0].replace('0', ''))
3509 vcodec, acodec = sanitize_codec(vcodecs), sanitize_codec(acodecs)
3511 for ext in preferences or COMPATIBLE_CODECS.keys():
3512 codec_set = COMPATIBLE_CODECS.get(ext, set())
3513 if ext == 'mkv' or codec_set.issuperset((vcodec, acodec)):
3517 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'},
3520 for ext in preferences or vexts:
3521 current_exts = {ext, *vexts, *aexts}
3522 if ext == 'mkv' or current_exts == {ext} or any(
3523 ext_sets.issuperset(current_exts) for ext_sets in COMPATIBLE_EXTS):
3525 return 'mkv' if allow_mkv else preferences[-1]
3528 def urlhandle_detect_ext(url_handle, default=NO_DEFAULT):
3529 getheader = url_handle.headers.get
3531 cd = getheader('Content-Disposition')
3533 m = re.match(r'attachment;\s*filename="(?P
<filename
>[^
"]+)"', cd)
3535 e = determine_ext(m.group('filename
'), default_ext=None)
3539 meta_ext = getheader('x
-amz
-meta
-name
')
3541 e = meta_ext.rpartition('.')[2]
3545 return mimetype2ext(getheader('Content
-Type
'), default=default)
3548 def encode_data_uri(data, mime_type):
3549 return 'data
:%s;base64
,%s' % (mime_type, base64.b64encode(data).decode('ascii
'))
3552 def age_restricted(content_limit, age_limit):
3553 """ Returns True iff the content should be blocked """
3555 if age_limit is None: # No limit set
3557 if content_limit is None:
3558 return False # Content available for everyone
3559 return age_limit < content_limit
3562 # List of known byte-order-marks (BOM)
3564 (b'\xef\xbb\xbf', 'utf
-8'),
3565 (b'\x00\x00\xfe\xff', 'utf
-32-be
'),
3566 (b'\xff\xfe\x00\x00', 'utf
-32-le
'),
3567 (b'\xff\xfe', 'utf
-16-le
'),
3568 (b'\xfe\xff', 'utf
-16-be
'),
3572 def is_html(first_bytes):
3573 """ Detect whether a file contains HTML by examining its first bytes. """
3576 for bom, enc in BOMS:
3577 while first_bytes.startswith(bom):
3578 encoding, first_bytes = enc, first_bytes[len(bom):]
3580 return re.match(r'^\s
*<', first_bytes.decode(encoding, 'replace
'))
3583 def determine_protocol(info_dict):
3584 protocol = info_dict.get('protocol
')
3585 if protocol is not None:
3588 url = sanitize_url(info_dict['url
'])
3589 if url.startswith('rtmp
'):
3591 elif url.startswith('mms
'):
3593 elif url.startswith('rtsp
'):
3596 ext = determine_ext(url)
3598 return 'm3u8
' if info_dict.get('is_live
') else 'm3u8_native
'
3602 return urllib.parse.urlparse(url).scheme
3605 def render_table(header_row, data, delim=False, extra_gap=0, hide_empty=False):
3606 """ Render a list of rows, each as a list of values.
3607 Text after a \t will be right aligned """
3609 return len(remove_terminal_sequences(string).replace('\t', ''))
3611 def get_max_lens(table):
3612 return [max(width(str(v)) for v in col) for col in zip(*table)]
3614 def filter_using_list(row, filterArray):
3615 return [col for take, col in itertools.zip_longest(filterArray, row, fillvalue=True) if take]
3617 max_lens = get_max_lens(data) if hide_empty else []
3618 header_row = filter_using_list(header_row, max_lens)
3619 data = [filter_using_list(row, max_lens) for row in data]
3621 table = [header_row] + data
3622 max_lens = get_max_lens(table)
3625 table = [header_row, [delim * (ml + extra_gap) for ml in max_lens]] + data
3626 table[1][-1] = table[1][-1][:-extra_gap * len(delim)] # Remove extra_gap from end of delimiter
3628 for pos, text in enumerate(map(str, row)):
3630 row[pos] = text.replace('\t', ' ' * (max_lens[pos] - width(text))) + ' ' * extra_gap
3632 row[pos] = text + ' ' * (max_lens[pos] - width(text) + extra_gap)
3633 ret = '\n'.join(''.join(row).rstrip() for row in table)
3637 def _match_one(filter_part, dct, incomplete):
3638 # TODO: Generalize code with YoutubeDL._build_format_filter
3639 STRING_OPERATORS = {
3640 '*=': operator.contains,
3641 '^
=': lambda attr, value: attr.startswith(value),
3642 '$
=': lambda attr, value: attr.endswith(value),
3643 '~
=': lambda attr, value: re.search(value, attr),
3645 COMPARISON_OPERATORS = {
3647 '<=': operator.le, # "<=" must be defined above "<"
3654 if isinstance(incomplete, bool):
3655 is_incomplete = lambda _: incomplete
3657 is_incomplete = lambda k: k in incomplete
3659 operator_rex = re.compile(r'''(?x)
3661 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3663 (?P<quote>["\'])(?P<quotedstrval>.+?)(?P=quote)|
3666 ''' % '|
'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3667 m = operator_rex.fullmatch(filter_part.strip())
3670 unnegated_op = COMPARISON_OPERATORS[m['op
']]
3672 op = lambda attr, value: not unnegated_op(attr, value)
3675 comparison_value = m['quotedstrval
'] or m['strval
'] or m['intval
']
3677 comparison_value = comparison_value.replace(r'\
%s' % m['quote
'], m['quote
'])
3678 actual_value = dct.get(m['key
'])
3679 numeric_comparison = None
3680 if isinstance(actual_value, (int, float)):
3681 # If the original field is a string and matching comparisonvalue is
3682 # a number we should respect the origin of the original field
3683 # and process comparison value as a string (see
3684 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3686 numeric_comparison = int(comparison_value)
3688 numeric_comparison = parse_filesize(comparison_value)
3689 if numeric_comparison is None:
3690 numeric_comparison = parse_filesize(f'{comparison_value}B
')
3691 if numeric_comparison is None:
3692 numeric_comparison = parse_duration(comparison_value)
3693 if numeric_comparison is not None and m['op
'] in STRING_OPERATORS:
3694 raise ValueError('Operator
%s only supports string values
!' % m['op
'])
3695 if actual_value is None:
3696 return is_incomplete(m['key
']) or m['none_inclusive
']
3697 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3700 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3701 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3703 operator_rex = re.compile(r'''(?x)
3704 (?P<op>%s)\s*(?P<key>[a-z_]+)
3705 ''' % '|
'.join(map(re.escape, UNARY_OPERATORS.keys())))
3706 m = operator_rex.fullmatch(filter_part.strip())
3708 op = UNARY_OPERATORS[m.group('op
')]
3709 actual_value = dct.get(m.group('key
'))
3710 if is_incomplete(m.group('key
')) and actual_value is None:
3712 return op(actual_value)
3714 raise ValueError('Invalid
filter part
%r' % filter_part)
3717 def match_str(filter_str, dct, incomplete=False):
3718 """ Filter a dictionary with a simple string syntax.
3719 @returns Whether the filter passes
3720 @param incomplete Set of keys that is expected to be missing from dct.
3721 Can be True/False to indicate all/none of the keys may be missing.
3722 All conditions on incomplete keys pass if the key is missing
3725 _match_one(filter_part.replace(r'\
&', '&'), dct, incomplete)
3726 for filter_part in re.split(r'(?
<!\\)&', filter_str))
3729 def match_filter_func(filters, breaking_filters=None):
3730 if not filters and not breaking_filters:
3732 breaking_filters = match_filter_func(breaking_filters) or (lambda _, __: None)
3733 filters = set(variadic(filters or []))
3735 interactive = '-' in filters
3739 def _match_func(info_dict, incomplete=False):
3740 ret = breaking_filters(info_dict, incomplete)
3742 raise RejectedVideoReached(ret)
3744 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3745 return NO_DEFAULT if interactive and not incomplete else None
3747 video_title = info_dict.get('title
') or info_dict.get('id') or 'entry
'
3748 filter_str = ') |
('.join(map(str.strip, filters))
3749 return f'{video_title} does
not pass filter ({filter_str}
), skipping
..'
3753 class download_range_func:
3754 def __init__(self, chapters, ranges):
3755 self.chapters, self.ranges = chapters, ranges
3757 def __call__(self, info_dict, ydl):
3758 if not self.ranges and not self.chapters:
3761 warning = ('There are no chapters matching the regex
' if info_dict.get('chapters
')
3762 else 'Cannot match chapters since chapter information
is unavailable
')
3763 for regex in self.chapters or []:
3764 for i, chapter in enumerate(info_dict.get('chapters
') or []):
3765 if re.search(regex, chapter['title
']):
3767 yield {**chapter, 'index': i}
3768 if self.chapters and warning:
3769 ydl.to_screen(f'[info
] {info_dict["id"]}
: {warning}
')
3771 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3773 def __eq__(self, other):
3774 return (isinstance(other, download_range_func)
3775 and self.chapters == other.chapters and self.ranges == other.ranges)
3778 return f'{__name__}
.{type(self).__name__}
({self.chapters}
, {self.ranges}
)'
3781 def parse_dfxp_time_expr(time_expr):
3785 mobj = re.match(rf'^
(?P
<time_offset
>{NUMBER_RE}
)s?$
', time_expr)
3787 return float(mobj.group('time_offset
'))
3789 mobj = re.match(r'^
(\d
+):(\d\d
):(\d\
d(?
:(?
:\
.|
:)\d
+)?
)$
', time_expr)
3791 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3794 def srt_subtitles_timecode(seconds):
3795 return '%02d
:%02d
:%02d
,%03d
' % timetuple_from_msec(seconds * 1000)
3798 def ass_subtitles_timecode(seconds):
3799 time = timetuple_from_msec(seconds * 1000)
3800 return '%01d
:%02d
:%02d
.%02d
' % (*time[:-1], time.milliseconds / 10)
3803 def dfxp2srt(dfxp_data):
3805 @param dfxp_data A bytes-like object containing DFXP data
3806 @returns A unicode object containing converted SRT data
3808 LEGACY_NAMESPACES = (
3809 (b'http
://www
.w3
.org
/ns
/ttml
', [
3810 b'http
://www
.w3
.org
/2004/11/ttaf1
',
3811 b'http
://www
.w3
.org
/2006/04/ttaf1
',
3812 b'http
://www
.w3
.org
/2006/10/ttaf1
',
3814 (b'http
://www
.w3
.org
/ns
/ttml
#styling', [
3815 b
'http://www.w3.org/ns/ttml#style',
3819 SUPPORTED_STYLING
= [
3828 _x
= functools
.partial(xpath_with_ns
, ns_map
={
3829 'xml': 'http://www.w3.org/XML/1998/namespace',
3830 'ttml': 'http://www.w3.org/ns/ttml',
3831 'tts': 'http://www.w3.org/ns/ttml#styling',
3837 class TTMLPElementParser
:
3839 _unclosed_elements
= []
3840 _applied_styles
= []
3842 def start(self
, tag
, attrib
):
3843 if tag
in (_x('ttml:br'), 'br'):
3846 unclosed_elements
= []
3848 element_style_id
= attrib
.get('style')
3850 style
.update(default_style
)
3851 if element_style_id
:
3852 style
.update(styles
.get(element_style_id
, {}))
3853 for prop
in SUPPORTED_STYLING
:
3854 prop_val
= attrib
.get(_x('tts:' + prop
))
3856 style
[prop
] = prop_val
3859 for k
, v
in sorted(style
.items()):
3860 if self
._applied
_styles
and self
._applied
_styles
[-1].get(k
) == v
:
3863 font
+= ' color="%s"' % v
3864 elif k
== 'fontSize':
3865 font
+= ' size="%s"' % v
3866 elif k
== 'fontFamily':
3867 font
+= ' face="%s"' % v
3868 elif k
== 'fontWeight' and v
== 'bold':
3870 unclosed_elements
.append('b')
3871 elif k
== 'fontStyle' and v
== 'italic':
3873 unclosed_elements
.append('i')
3874 elif k
== 'textDecoration' and v
== 'underline':
3876 unclosed_elements
.append('u')
3878 self
._out
+= '<font' + font
+ '>'
3879 unclosed_elements
.append('font')
3881 if self
._applied
_styles
:
3882 applied_style
.update(self
._applied
_styles
[-1])
3883 applied_style
.update(style
)
3884 self
._applied
_styles
.append(applied_style
)
3885 self
._unclosed
_elements
.append(unclosed_elements
)
3888 if tag
not in (_x('ttml:br'), 'br'):
3889 unclosed_elements
= self
._unclosed
_elements
.pop()
3890 for element
in reversed(unclosed_elements
):
3891 self
._out
+= '</%s>' % element
3892 if unclosed_elements
and self
._applied
_styles
:
3893 self
._applied
_styles
.pop()
3895 def data(self
, data
):
3899 return self
._out
.strip()
3901 # Fix UTF-8 encoded file wrongly marked as UTF-16. See https://github.com/yt-dlp/yt-dlp/issues/6543#issuecomment-1477169870
3902 # This will not trigger false positives since only UTF-8 text is being replaced
3903 dfxp_data
= dfxp_data
.replace(b
'encoding=\'UTF-16\'', b
'encoding=\'UTF-8\'')
3905 def parse_node(node
):
3906 target
= TTMLPElementParser()
3907 parser
= xml
.etree
.ElementTree
.XMLParser(target
=target
)
3908 parser
.feed(xml
.etree
.ElementTree
.tostring(node
))
3909 return parser
.close()
3911 for k
, v
in LEGACY_NAMESPACES
:
3913 dfxp_data
= dfxp_data
.replace(ns
, k
)
3915 dfxp
= compat_etree_fromstring(dfxp_data
)
3917 paras
= dfxp
.findall(_x('.//ttml:p')) or dfxp
.findall('.//p')
3920 raise ValueError('Invalid dfxp/TTML subtitle')
3924 for style
in dfxp
.findall(_x('.//ttml:style')):
3925 style_id
= style
.get('id') or style
.get(_x('xml:id'))
3928 parent_style_id
= style
.get('style')
3930 if parent_style_id
not in styles
:
3933 styles
[style_id
] = styles
[parent_style_id
].copy()
3934 for prop
in SUPPORTED_STYLING
:
3935 prop_val
= style
.get(_x('tts:' + prop
))
3937 styles
.setdefault(style_id
, {})[prop
] = prop_val
3943 for p
in ('body', 'div'):
3944 ele
= xpath_element(dfxp
, [_x('.//ttml:' + p
), './/' + p
])
3947 style
= styles
.get(ele
.get('style'))
3950 default_style
.update(style
)
3952 for para
, index
in zip(paras
, itertools
.count(1)):
3953 begin_time
= parse_dfxp_time_expr(para
.attrib
.get('begin'))
3954 end_time
= parse_dfxp_time_expr(para
.attrib
.get('end'))
3955 dur
= parse_dfxp_time_expr(para
.attrib
.get('dur'))
3956 if begin_time
is None:
3961 end_time
= begin_time
+ dur
3962 out
.append('%d\n%s --> %s\n%s\n\n' % (
3964 srt_subtitles_timecode(begin_time
),
3965 srt_subtitles_timecode(end_time
),
3971 def cli_option(params
, command_option
, param
, separator
=None):
3972 param
= params
.get(param
)
3973 return ([] if param
is None
3974 else [command_option
, str(param
)] if separator
is None
3975 else [f
'{command_option}{separator}{param}'])
3978 def cli_bool_option(params
, command_option
, param
, true_value
='true', false_value
='false', separator
=None):
3979 param
= params
.get(param
)
3980 assert param
in (True, False, None)
3981 return cli_option({True: true_value, False: false_value}
, command_option
, param
, separator
)
3984 def cli_valueless_option(params
, command_option
, param
, expected_value
=True):
3985 return [command_option
] if params
.get(param
) == expected_value
else []
3988 def cli_configuration_args(argdict
, keys
, default
=[], use_compat
=True):
3989 if isinstance(argdict
, (list, tuple)): # for backward compatibility
3996 assert isinstance(argdict
, dict)
3998 assert isinstance(keys
, (list, tuple))
3999 for key_list
in keys
:
4000 arg_list
= list(filter(
4001 lambda x
: x
is not None,
4002 [argdict
.get(key
.lower()) for key
in variadic(key_list
)]))
4004 return [arg
for args
in arg_list
for arg
in args
]
4008 def _configuration_args(main_key
, argdict
, exe
, keys
=None, default
=[], use_compat
=True):
4009 main_key
, exe
= main_key
.lower(), exe
.lower()
4010 root_key
= exe
if main_key
== exe
else f
'{main_key}+{exe}'
4011 keys
= [f
'{root_key}{k}' for k
in (keys
or [''])]
4012 if root_key
in keys
:
4014 keys
.append((main_key
, exe
))
4015 keys
.append('default')
4018 return cli_configuration_args(argdict
, keys
, default
, use_compat
)
4022 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4081 'iw': 'heb', # Replaced by he in 1989 revision
4091 'in': 'ind', # Replaced by id in 1989 revision
4207 'ji': 'yid', # Replaced by yi in 1989 revision
4215 def short2long(cls
, code
):
4216 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4217 return cls
._lang
_map
.get(code
[:2])
4220 def long2short(cls
, code
):
4221 """Convert language code from ISO 639-2/T to ISO 639-1"""
4222 for short_name
, long_name
in cls
._lang
_map
.items():
4223 if long_name
== code
:
4228 # From http://data.okfn.org/data/core/country-list
4230 'AF': 'Afghanistan',
4231 'AX': 'Åland Islands',
4234 'AS': 'American Samoa',
4239 'AG': 'Antigua and Barbuda',
4256 'BO': 'Bolivia, Plurinational State of',
4257 'BQ': 'Bonaire, Sint Eustatius and Saba',
4258 'BA': 'Bosnia and Herzegovina',
4260 'BV': 'Bouvet Island',
4262 'IO': 'British Indian Ocean Territory',
4263 'BN': 'Brunei Darussalam',
4265 'BF': 'Burkina Faso',
4271 'KY': 'Cayman Islands',
4272 'CF': 'Central African Republic',
4276 'CX': 'Christmas Island',
4277 'CC': 'Cocos (Keeling) Islands',
4281 'CD': 'Congo, the Democratic Republic of the',
4282 'CK': 'Cook Islands',
4284 'CI': 'Côte d\'Ivoire',
4289 'CZ': 'Czech Republic',
4293 'DO': 'Dominican Republic',
4296 'SV': 'El Salvador',
4297 'GQ': 'Equatorial Guinea',
4301 'FK': 'Falkland Islands (Malvinas)',
4302 'FO': 'Faroe Islands',
4306 'GF': 'French Guiana',
4307 'PF': 'French Polynesia',
4308 'TF': 'French Southern Territories',
4323 'GW': 'Guinea-Bissau',
4326 'HM': 'Heard Island and McDonald Islands',
4327 'VA': 'Holy See (Vatican City State)',
4334 'IR': 'Iran, Islamic Republic of',
4337 'IM': 'Isle of Man',
4347 'KP': 'Korea, Democratic People\'s Republic of',
4348 'KR': 'Korea, Republic of',
4351 'LA': 'Lao People\'s Democratic Republic',
4357 'LI': 'Liechtenstein',
4361 'MK': 'Macedonia, the Former Yugoslav Republic of',
4368 'MH': 'Marshall Islands',
4374 'FM': 'Micronesia, Federated States of',
4375 'MD': 'Moldova, Republic of',
4386 'NL': 'Netherlands',
4387 'NC': 'New Caledonia',
4388 'NZ': 'New Zealand',
4393 'NF': 'Norfolk Island',
4394 'MP': 'Northern Mariana Islands',
4399 'PS': 'Palestine, State of',
4401 'PG': 'Papua New Guinea',
4404 'PH': 'Philippines',
4408 'PR': 'Puerto Rico',
4412 'RU': 'Russian Federation',
4414 'BL': 'Saint Barthélemy',
4415 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4416 'KN': 'Saint Kitts and Nevis',
4417 'LC': 'Saint Lucia',
4418 'MF': 'Saint Martin (French part)',
4419 'PM': 'Saint Pierre and Miquelon',
4420 'VC': 'Saint Vincent and the Grenadines',
4423 'ST': 'Sao Tome and Principe',
4424 'SA': 'Saudi Arabia',
4428 'SL': 'Sierra Leone',
4430 'SX': 'Sint Maarten (Dutch part)',
4433 'SB': 'Solomon Islands',
4435 'ZA': 'South Africa',
4436 'GS': 'South Georgia and the South Sandwich Islands',
4437 'SS': 'South Sudan',
4442 'SJ': 'Svalbard and Jan Mayen',
4445 'CH': 'Switzerland',
4446 'SY': 'Syrian Arab Republic',
4447 'TW': 'Taiwan, Province of China',
4449 'TZ': 'Tanzania, United Republic of',
4451 'TL': 'Timor-Leste',
4455 'TT': 'Trinidad and Tobago',
4458 'TM': 'Turkmenistan',
4459 'TC': 'Turks and Caicos Islands',
4463 'AE': 'United Arab Emirates',
4464 'GB': 'United Kingdom',
4465 'US': 'United States',
4466 'UM': 'United States Minor Outlying Islands',
4470 'VE': 'Venezuela, Bolivarian Republic of',
4472 'VG': 'Virgin Islands, British',
4473 'VI': 'Virgin Islands, U.S.',
4474 'WF': 'Wallis and Futuna',
4475 'EH': 'Western Sahara',
4479 # Not ISO 3166 codes, but used for IP blocks
4480 'AP': 'Asia/Pacific Region',
4485 def short2full(cls
, code
):
4486 """Convert an ISO 3166-2 country code to the corresponding full name"""
4487 return cls
._country
_map
.get(code
.upper())
4491 # Major IPv4 address blocks per country
4493 'AD': '46.172.224.0/19',
4494 'AE': '94.200.0.0/13',
4495 'AF': '149.54.0.0/17',
4496 'AG': '209.59.64.0/18',
4497 'AI': '204.14.248.0/21',
4498 'AL': '46.99.0.0/16',
4499 'AM': '46.70.0.0/15',
4500 'AO': '105.168.0.0/13',
4501 'AP': '182.50.184.0/21',
4502 'AQ': '23.154.160.0/24',
4503 'AR': '181.0.0.0/12',
4504 'AS': '202.70.112.0/20',
4505 'AT': '77.116.0.0/14',
4506 'AU': '1.128.0.0/11',
4507 'AW': '181.41.0.0/18',
4508 'AX': '185.217.4.0/22',
4509 'AZ': '5.197.0.0/16',
4510 'BA': '31.176.128.0/17',
4511 'BB': '65.48.128.0/17',
4512 'BD': '114.130.0.0/16',
4514 'BF': '102.178.0.0/15',
4515 'BG': '95.42.0.0/15',
4516 'BH': '37.131.0.0/17',
4517 'BI': '154.117.192.0/18',
4518 'BJ': '137.255.0.0/16',
4519 'BL': '185.212.72.0/23',
4520 'BM': '196.12.64.0/18',
4521 'BN': '156.31.0.0/16',
4522 'BO': '161.56.0.0/16',
4523 'BQ': '161.0.80.0/20',
4524 'BR': '191.128.0.0/12',
4525 'BS': '24.51.64.0/18',
4526 'BT': '119.2.96.0/19',
4527 'BW': '168.167.0.0/16',
4528 'BY': '178.120.0.0/13',
4529 'BZ': '179.42.192.0/18',
4530 'CA': '99.224.0.0/11',
4531 'CD': '41.243.0.0/16',
4532 'CF': '197.242.176.0/21',
4533 'CG': '160.113.0.0/16',
4534 'CH': '85.0.0.0/13',
4535 'CI': '102.136.0.0/14',
4536 'CK': '202.65.32.0/19',
4537 'CL': '152.172.0.0/14',
4538 'CM': '102.244.0.0/14',
4539 'CN': '36.128.0.0/10',
4540 'CO': '181.240.0.0/12',
4541 'CR': '201.192.0.0/12',
4542 'CU': '152.206.0.0/15',
4543 'CV': '165.90.96.0/19',
4544 'CW': '190.88.128.0/17',
4545 'CY': '31.153.0.0/16',
4546 'CZ': '88.100.0.0/14',
4548 'DJ': '197.241.0.0/17',
4549 'DK': '87.48.0.0/12',
4550 'DM': '192.243.48.0/20',
4551 'DO': '152.166.0.0/15',
4552 'DZ': '41.96.0.0/12',
4553 'EC': '186.68.0.0/15',
4554 'EE': '90.190.0.0/15',
4555 'EG': '156.160.0.0/11',
4556 'ER': '196.200.96.0/20',
4557 'ES': '88.0.0.0/11',
4558 'ET': '196.188.0.0/14',
4559 'EU': '2.16.0.0/13',
4560 'FI': '91.152.0.0/13',
4561 'FJ': '144.120.0.0/16',
4562 'FK': '80.73.208.0/21',
4563 'FM': '119.252.112.0/20',
4564 'FO': '88.85.32.0/19',
4566 'GA': '41.158.0.0/15',
4568 'GD': '74.122.88.0/21',
4569 'GE': '31.146.0.0/16',
4570 'GF': '161.22.64.0/18',
4571 'GG': '62.68.160.0/19',
4572 'GH': '154.160.0.0/12',
4573 'GI': '95.164.0.0/16',
4574 'GL': '88.83.0.0/19',
4575 'GM': '160.182.0.0/15',
4576 'GN': '197.149.192.0/18',
4577 'GP': '104.250.0.0/19',
4578 'GQ': '105.235.224.0/20',
4579 'GR': '94.64.0.0/13',
4580 'GT': '168.234.0.0/16',
4581 'GU': '168.123.0.0/16',
4582 'GW': '197.214.80.0/20',
4583 'GY': '181.41.64.0/18',
4584 'HK': '113.252.0.0/14',
4585 'HN': '181.210.0.0/16',
4586 'HR': '93.136.0.0/13',
4587 'HT': '148.102.128.0/17',
4588 'HU': '84.0.0.0/14',
4589 'ID': '39.192.0.0/10',
4590 'IE': '87.32.0.0/12',
4591 'IL': '79.176.0.0/13',
4592 'IM': '5.62.80.0/20',
4593 'IN': '117.192.0.0/10',
4594 'IO': '203.83.48.0/21',
4595 'IQ': '37.236.0.0/14',
4596 'IR': '2.176.0.0/12',
4597 'IS': '82.221.0.0/16',
4598 'IT': '79.0.0.0/10',
4599 'JE': '87.244.64.0/18',
4600 'JM': '72.27.0.0/17',
4601 'JO': '176.29.0.0/16',
4602 'JP': '133.0.0.0/8',
4603 'KE': '105.48.0.0/12',
4604 'KG': '158.181.128.0/17',
4605 'KH': '36.37.128.0/17',
4606 'KI': '103.25.140.0/22',
4607 'KM': '197.255.224.0/20',
4608 'KN': '198.167.192.0/19',
4609 'KP': '175.45.176.0/22',
4610 'KR': '175.192.0.0/10',
4611 'KW': '37.36.0.0/14',
4612 'KY': '64.96.0.0/15',
4613 'KZ': '2.72.0.0/13',
4614 'LA': '115.84.64.0/18',
4615 'LB': '178.135.0.0/16',
4616 'LC': '24.92.144.0/20',
4617 'LI': '82.117.0.0/19',
4618 'LK': '112.134.0.0/15',
4619 'LR': '102.183.0.0/16',
4620 'LS': '129.232.0.0/17',
4621 'LT': '78.56.0.0/13',
4622 'LU': '188.42.0.0/16',
4623 'LV': '46.109.0.0/16',
4624 'LY': '41.252.0.0/14',
4625 'MA': '105.128.0.0/11',
4626 'MC': '88.209.64.0/18',
4627 'MD': '37.246.0.0/16',
4628 'ME': '178.175.0.0/17',
4629 'MF': '74.112.232.0/21',
4630 'MG': '154.126.0.0/17',
4631 'MH': '117.103.88.0/21',
4632 'MK': '77.28.0.0/15',
4633 'ML': '154.118.128.0/18',
4634 'MM': '37.111.0.0/17',
4635 'MN': '49.0.128.0/17',
4636 'MO': '60.246.0.0/16',
4637 'MP': '202.88.64.0/20',
4638 'MQ': '109.203.224.0/19',
4639 'MR': '41.188.64.0/18',
4640 'MS': '208.90.112.0/22',
4641 'MT': '46.11.0.0/16',
4642 'MU': '105.16.0.0/12',
4643 'MV': '27.114.128.0/18',
4644 'MW': '102.70.0.0/15',
4645 'MX': '187.192.0.0/11',
4646 'MY': '175.136.0.0/13',
4647 'MZ': '197.218.0.0/15',
4648 'NA': '41.182.0.0/16',
4649 'NC': '101.101.0.0/18',
4650 'NE': '197.214.0.0/18',
4651 'NF': '203.17.240.0/22',
4652 'NG': '105.112.0.0/12',
4653 'NI': '186.76.0.0/15',
4654 'NL': '145.96.0.0/11',
4655 'NO': '84.208.0.0/13',
4656 'NP': '36.252.0.0/15',
4657 'NR': '203.98.224.0/19',
4658 'NU': '49.156.48.0/22',
4659 'NZ': '49.224.0.0/14',
4660 'OM': '5.36.0.0/15',
4661 'PA': '186.72.0.0/15',
4662 'PE': '186.160.0.0/14',
4663 'PF': '123.50.64.0/18',
4664 'PG': '124.240.192.0/19',
4665 'PH': '49.144.0.0/13',
4666 'PK': '39.32.0.0/11',
4667 'PL': '83.0.0.0/11',
4668 'PM': '70.36.0.0/20',
4669 'PR': '66.50.0.0/16',
4670 'PS': '188.161.0.0/16',
4671 'PT': '85.240.0.0/13',
4672 'PW': '202.124.224.0/20',
4673 'PY': '181.120.0.0/14',
4674 'QA': '37.210.0.0/15',
4675 'RE': '102.35.0.0/16',
4676 'RO': '79.112.0.0/13',
4677 'RS': '93.86.0.0/15',
4678 'RU': '5.136.0.0/13',
4679 'RW': '41.186.0.0/16',
4680 'SA': '188.48.0.0/13',
4681 'SB': '202.1.160.0/19',
4682 'SC': '154.192.0.0/11',
4683 'SD': '102.120.0.0/13',
4684 'SE': '78.64.0.0/12',
4685 'SG': '8.128.0.0/10',
4686 'SI': '188.196.0.0/14',
4687 'SK': '78.98.0.0/15',
4688 'SL': '102.143.0.0/17',
4689 'SM': '89.186.32.0/19',
4690 'SN': '41.82.0.0/15',
4691 'SO': '154.115.192.0/18',
4692 'SR': '186.179.128.0/17',
4693 'SS': '105.235.208.0/21',
4694 'ST': '197.159.160.0/19',
4695 'SV': '168.243.0.0/16',
4696 'SX': '190.102.0.0/20',
4698 'SZ': '41.84.224.0/19',
4699 'TC': '65.255.48.0/20',
4700 'TD': '154.68.128.0/19',
4701 'TG': '196.168.0.0/14',
4702 'TH': '171.96.0.0/13',
4703 'TJ': '85.9.128.0/18',
4704 'TK': '27.96.24.0/21',
4705 'TL': '180.189.160.0/20',
4706 'TM': '95.85.96.0/19',
4707 'TN': '197.0.0.0/11',
4708 'TO': '175.176.144.0/21',
4709 'TR': '78.160.0.0/11',
4710 'TT': '186.44.0.0/15',
4711 'TV': '202.2.96.0/19',
4712 'TW': '120.96.0.0/11',
4713 'TZ': '156.156.0.0/14',
4714 'UA': '37.52.0.0/14',
4715 'UG': '102.80.0.0/13',
4717 'UY': '167.56.0.0/13',
4718 'UZ': '84.54.64.0/18',
4719 'VA': '212.77.0.0/19',
4720 'VC': '207.191.240.0/21',
4721 'VE': '186.88.0.0/13',
4722 'VG': '66.81.192.0/20',
4723 'VI': '146.226.0.0/16',
4724 'VN': '14.160.0.0/11',
4725 'VU': '202.80.32.0/20',
4726 'WF': '117.20.32.0/21',
4727 'WS': '202.4.32.0/19',
4728 'YE': '134.35.0.0/16',
4729 'YT': '41.242.116.0/22',
4730 'ZA': '41.0.0.0/11',
4731 'ZM': '102.144.0.0/13',
4732 'ZW': '102.177.192.0/18',
4736 def random_ipv4(cls
, code_or_block
):
4737 if len(code_or_block
) == 2:
4738 block
= cls
._country
_ip
_map
.get(code_or_block
.upper())
4742 block
= code_or_block
4743 addr
, preflen
= block
.split('/')
4744 addr_min
= struct
.unpack('!L', socket
.inet_aton(addr
))[0]
4745 addr_max
= addr_min |
(0xffffffff >> int(preflen
))
4746 return str(socket
.inet_ntoa(
4747 struct
.pack('!L', random
.randint(addr_min
, addr_max
))))
4750 class PerRequestProxyHandler(urllib
.request
.ProxyHandler
):
4751 def __init__(self
, proxies
=None):
4752 # Set default handlers
4753 for type in ('http', 'https'):
4754 setattr(self
, '%s_open' % type,
4755 lambda r
, proxy
='__noproxy__', type=type, meth
=self
.proxy_open
:
4756 meth(r
, proxy
, type))
4757 urllib
.request
.ProxyHandler
.__init
__(self
, proxies
)
4759 def proxy_open(self
, req
, proxy
, type):
4760 req_proxy
= req
.headers
.get('Ytdl-request-proxy')
4761 if req_proxy
is not None:
4763 del req
.headers
['Ytdl-request-proxy']
4765 if proxy
== '__noproxy__':
4766 return None # No Proxy
4767 if urllib
.parse
.urlparse(proxy
).scheme
.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4768 req
.add_header('Ytdl-socks-proxy', proxy
)
4769 # yt-dlp's http/https handlers do wrapping the socket with socks
4771 return urllib
.request
.ProxyHandler
.proxy_open(
4772 self
, req
, proxy
, type)
4775 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4776 # released into Public Domain
4777 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4779 def long_to_bytes(n
, blocksize
=0):
4780 """long_to_bytes(n:long, blocksize:int) : string
4781 Convert a long integer to a byte string.
4783 If optional blocksize is given and greater than zero, pad the front of the
4784 byte string with binary zeros so that the length is a multiple of
4787 # after much testing, this algorithm was deemed to be the fastest
4791 s
= struct
.pack('>I', n
& 0xffffffff) + s
4793 # strip off leading zeros
4794 for i
in range(len(s
)):
4795 if s
[i
] != b
'\000'[0]:
4798 # only happens when n == 0
4802 # add back some pad bytes. this could be done more efficiently w.r.t. the
4803 # de-padding being done above, but sigh...
4804 if blocksize
> 0 and len(s
) % blocksize
:
4805 s
= (blocksize
- len(s
) % blocksize
) * b
'\000' + s
4809 def bytes_to_long(s
):
4810 """bytes_to_long(string) : long
4811 Convert a byte string to a long integer.
4813 This is (essentially) the inverse of long_to_bytes().
4818 extra
= (4 - length
% 4)
4819 s
= b
'\000' * extra
+ s
4820 length
= length
+ extra
4821 for i
in range(0, length
, 4):
4822 acc
= (acc
<< 32) + struct
.unpack('>I', s
[i
:i
+ 4])[0]
4826 def ohdave_rsa_encrypt(data
, exponent
, modulus
):
4828 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
4831 data: data to encrypt, bytes-like object
4832 exponent, modulus: parameter e and N of RSA algorithm, both integer
4833 Output: hex string of encrypted data
4835 Limitation: supports one block encryption only
4838 payload
= int(binascii
.hexlify(data
[::-1]), 16)
4839 encrypted
= pow(payload
, exponent
, modulus
)
4840 return '%x' % encrypted
4843 def pkcs1pad(data
, length
):
4845 Padding input data with PKCS#1 scheme
4847 @param {int[]} data input data
4848 @param {int} length target length
4849 @returns {int[]} padded data
4851 if len(data
) > length
- 11:
4852 raise ValueError('Input data too long for PKCS#1 padding')
4854 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4855 return [0, 2] + pseudo_random
+ [0] + data
4858 def _base_n_table(n
, table
):
4859 if not table
and not n
:
4860 raise ValueError('Either table or n must be specified')
4861 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4863 if n
and n
!= len(table
):
4864 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4868 def encode_base_n(num
, n
=None, table
=None):
4869 """Convert given int to a base-n string"""
4870 table
= _base_n_table(n
, table
)
4874 result
, base
= '', len(table
)
4876 result
= table
[num
% base
] + result
4881 def decode_base_n(string
, n
=None, table
=None):
4882 """Convert given base-n string to int"""
4883 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4884 result
, base
= 0, len(table
)
4886 result
= result
* base
+ table
[char
]
4890 def decode_packed_codes(code
):
4891 mobj
= re
.search(PACKED_CODES_RE
, code
)
4892 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4895 symbols
= symbols
.split('|')
4900 base_n_count
= encode_base_n(count
, base
)
4901 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4904 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4908 def caesar(s
, alphabet
, shift
):
4913 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4918 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4921 def parse_m3u8_attributes(attrib
):
4923 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4924 if val
.startswith('"'):
4930 def urshift(val
, n
):
4931 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4934 def write_xattr(path
, key
, value
):
4935 # Windows: Write xattrs to NTFS Alternate Data Streams:
4936 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4937 if compat_os_name
== 'nt':
4938 assert ':' not in key
4939 assert os
.path
.exists(path
)
4942 with open(f
'{path}:{key}', 'wb') as f
:
4944 except OSError as e
:
4945 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4948 # UNIX Method 1. Use xattrs/pyxattrs modules
4951 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
4952 # Unicode arguments are not supported in pyxattr until version 0.5.0
4953 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4954 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
4955 setxattr
= xattr
.set
4957 setxattr
= xattr
.setxattr
4961 setxattr(path
, key
, value
)
4962 except OSError as e
:
4963 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4966 # UNIX Method 2. Use setfattr/xattr executables
4967 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
4968 else 'xattr' if check_executable('xattr', ['-h']) else None)
4970 raise XAttrUnavailableError(
4971 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
4972 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
4974 value
= value
.decode()
4976 _
, stderr
, returncode
= Popen
.run(
4977 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
4978 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4979 except OSError as e
:
4980 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4982 raise XAttrMetadataError(returncode
, stderr
)
4985 def random_birthday(year_field
, month_field
, day_field
):
4986 start_date
= datetime
.date(1950, 1, 1)
4987 end_date
= datetime
.date(1995, 12, 31)
4988 offset
= random
.randint(0, (end_date
- start_date
).days
)
4989 random_date
= start_date
+ datetime
.timedelta(offset
)
4991 year_field
: str(random_date
.year
),
4992 month_field
: str(random_date
.month
),
4993 day_field
: str(random_date
.day
),
4997 def find_available_port(interface
=''):
4999 with socket
.socket() as sock
:
5000 sock
.bind((interface
, 0))
5001 return sock
.getsockname()[1]
5006 # Templates for internet shortcut files, which are plain text files.
5007 DOT_URL_LINK_TEMPLATE
= '''\
5012 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5013 <?xml version="1.0" encoding="UTF-8"?>
5014 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5015 <plist version="1.0">
5018 \t<string>%(url)s</string>
5023 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5033 'url': DOT_URL_LINK_TEMPLATE
,
5034 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5035 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5039 def iri_to_uri(iri
):
5041 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5043 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.
5046 iri_parts
= urllib
.parse
.urlparse(iri
)
5048 if '[' in iri_parts
.netloc
:
5049 raise ValueError('IPv6 URIs are not, yet, supported.')
5050 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5052 # 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.
5055 if iri_parts
.username
:
5056 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5057 if iri_parts
.password
is not None:
5058 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5061 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5062 # The 'idna' encoding produces ASCII text.
5063 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5064 net_location
+= ':' + str(iri_parts
.port
)
5066 return urllib
.parse
.urlunparse(
5070 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5072 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5073 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5075 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5076 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5078 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5080 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5083 def to_high_limit_path(path
):
5084 if sys
.platform
in ['win32', 'cygwin']:
5085 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5086 return '\\\\?\\' + os
.path
.abspath(path
)
5091 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5092 val
= traversal
.traverse_obj(obj
, *variadic(field
))
5093 if not val
if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5095 return template
% func(val
)
5098 def clean_podcast_url(url
):
5099 return re
.sub(r
'''(?x)
5103 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5106 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5109 cn\.co| # https://podcorn.com/analytics-prefix/
5110 st\.fm # https://podsights.com/docs/
5115 _HEX_TABLE
= '0123456789abcdef'
5118 def random_uuidv4():
5119 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5122 def make_dir(path
, to_screen
=None):
5124 dn
= os
.path
.dirname(path
)
5126 os
.makedirs(dn
, exist_ok
=True)
5128 except OSError as err
:
5129 if callable(to_screen
) is not None:
5130 to_screen(f
'unable to create directory {err}')
5134 def get_executable_path():
5135 from ..update
import _get_variant_and_executable_path
5137 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5140 def get_user_config_dirs(package_name
):
5141 # .config (e.g. ~/.config/package_name)
5142 xdg_config_home
= os
.getenv('XDG_CONFIG_HOME') or compat_expanduser('~/.config')
5143 yield os
.path
.join(xdg_config_home
, package_name
)
5145 # appdata (%APPDATA%/package_name)
5146 appdata_dir
= os
.getenv('appdata')
5148 yield os
.path
.join(appdata_dir
, package_name
)
5150 # home (~/.package_name)
5151 yield os
.path
.join(compat_expanduser('~'), f
'.{package_name}')
5154 def get_system_config_dirs(package_name
):
5156 yield os
.path
.join('/etc', package_name
)
5159 def time_seconds(**kwargs
):
5161 Returns TZ-aware time in seconds since the epoch (1970-01-01T00:00:00Z)
5163 return time
.time() + datetime
.timedelta(**kwargs
).total_seconds()
5166 # create a JSON Web Signature (jws) with HS256 algorithm
5167 # the resulting format is in JWS Compact Serialization
5168 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5169 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5170 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5176 header_data
.update(headers
)
5177 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5178 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5179 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5180 signature_b64
= base64
.b64encode(h
.digest())
5181 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5185 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5186 def jwt_decode_hs256(jwt
):
5187 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5188 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5189 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5193 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5197 def supports_terminal_sequences(stream
):
5198 if compat_os_name
== 'nt':
5199 if not WINDOWS_VT_MODE
:
5201 elif not os
.getenv('TERM'):
5204 return stream
.isatty()
5205 except BaseException
:
5209 def windows_enable_vt_mode():
5210 """Ref: https://bugs.python.org/issue30075 """
5211 if get_windows_version() < (10, 0, 10586):
5215 import ctypes
.wintypes
5218 ENABLE_VIRTUAL_TERMINAL_PROCESSING
= 0x0004
5220 dll
= ctypes
.WinDLL('kernel32', use_last_error
=False)
5221 handle
= os
.open('CONOUT$', os
.O_RDWR
)
5223 h_out
= ctypes
.wintypes
.HANDLE(msvcrt
.get_osfhandle(handle
))
5224 dw_original_mode
= ctypes
.wintypes
.DWORD()
5225 success
= dll
.GetConsoleMode(h_out
, ctypes
.byref(dw_original_mode
))
5227 raise Exception('GetConsoleMode failed')
5229 success
= dll
.SetConsoleMode(h_out
, ctypes
.wintypes
.DWORD(
5230 dw_original_mode
.value | ENABLE_VIRTUAL_TERMINAL_PROCESSING
))
5232 raise Exception('SetConsoleMode failed')
5236 global WINDOWS_VT_MODE
5237 WINDOWS_VT_MODE
= True
5238 supports_terminal_sequences
.cache_clear()
5241 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5244 def remove_terminal_sequences(string
):
5245 return _terminal_sequences_re
.sub('', string
)
5248 def number_of_digits(number
):
5249 return len('%d' % number
)
5252 def join_nonempty(*values
, delim
='-', from_dict
=None):
5253 if from_dict
is not None:
5254 values
= (traversal
.traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5255 return delim
.join(map(str, filter(None, values
)))
5258 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5260 Find the largest format dimensions in terms of video width and, for each thumbnail:
5261 * Modify the URL: Match the width with the provided regex and replace with the former width
5264 This function is useful with video services that scale the provided thumbnails on demand
5266 _keys
= ('width', 'height')
5267 max_dimensions
= max(
5268 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5270 if not max_dimensions
[0]:
5274 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5275 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5276 for thumbnail
in thumbnails
5280 def parse_http_range(range):
5281 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5283 return None, None, None
5284 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5286 return None, None, None
5287 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5290 def read_stdin(what
):
5291 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5292 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5296 def determine_file_encoding(data
):
5298 Detect the text encoding used
5299 @returns (encoding, bytes to skip)
5302 # BOM marks are given priority over declarations
5303 for bom
, enc
in BOMS
:
5304 if data
.startswith(bom
):
5305 return enc
, len(bom
)
5307 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5308 # We ignore the endianness to get a good enough match
5309 data
= data
.replace(b
'\0', b
'')
5310 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5311 return mobj
.group(1).decode() if mobj
else None, 0
5318 __initialized
= False
5320 def __init__(self
, parser
, label
=None):
5321 self
.parser
, self
.label
= parser
, label
5322 self
._loaded
_paths
, self
.configs
= set(), []
5324 def init(self
, args
=None, filename
=None):
5325 assert not self
.__initialized
5326 self
.own_args
, self
.filename
= args
, filename
5327 return self
.load_configs()
5329 def load_configs(self
):
5332 location
= os
.path
.realpath(self
.filename
)
5333 directory
= os
.path
.dirname(location
)
5334 if location
in self
._loaded
_paths
:
5336 self
._loaded
_paths
.add(location
)
5338 self
.__initialized
= True
5339 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5340 self
.parsed_args
= self
.own_args
5341 for location
in opts
.config_locations
or []:
5343 if location
in self
._loaded
_paths
:
5345 self
._loaded
_paths
.add(location
)
5346 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5348 location
= os
.path
.join(directory
, expand_path(location
))
5349 if os
.path
.isdir(location
):
5350 location
= os
.path
.join(location
, 'yt-dlp.conf')
5351 if not os
.path
.exists(location
):
5352 self
.parser
.error(f
'config location {location} does not exist')
5353 self
.append_config(self
.read_file(location
), location
)
5357 label
= join_nonempty(
5358 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5360 return join_nonempty(
5361 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5362 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5366 def read_file(filename
, default
=[]):
5368 optionf
= open(filename
, 'rb')
5370 return default
# silently skip if file is not present
5372 enc
, skip
= determine_file_encoding(optionf
.read(512))
5373 optionf
.seek(skip
, io
.SEEK_SET
)
5375 enc
= None # silently skip read errors
5377 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5378 contents
= optionf
.read().decode(enc
or preferredencoding())
5379 res
= shlex
.split(contents
, comments
=True)
5380 except Exception as err
:
5381 raise ValueError(f
'Unable to parse "{filename}": {err}')
5387 def hide_login_info(opts
):
5388 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5389 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5394 return m
.group('key') + '=PRIVATE'
5398 opts
= list(map(_scrub_eq
, opts
))
5399 for idx
, opt
in enumerate(opts
):
5400 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5401 opts
[idx
+ 1] = 'PRIVATE'
5404 def append_config(self
, *args
, label
=None):
5405 config
= type(self
)(self
.parser
, label
)
5406 config
._loaded
_paths
= self
._loaded
_paths
5407 if config
.init(*args
):
5408 self
.configs
.append(config
)
5412 for config
in reversed(self
.configs
):
5413 yield from config
.all_args
5414 yield from self
.parsed_args
or []
5416 def parse_known_args(self
, **kwargs
):
5417 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5419 def parse_args(self
):
5420 return self
.parser
.parse_args(self
.all_args
)
5423 class WebSocketsWrapper
:
5424 """Wraps websockets module to use in non-async scopes"""
5427 def __init__(self
, url
, headers
=None, connect
=True):
5428 self
.loop
= asyncio
.new_event_loop()
5429 # XXX: "loop" is deprecated
5430 self
.conn
= websockets
.connect(
5431 url
, extra_headers
=headers
, ping_interval
=None,
5432 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5435 atexit
.register(self
.__exit
__, None, None, None)
5437 def __enter__(self
):
5439 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5442 def send(self
, *args
):
5443 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5445 def recv(self
, *args
):
5446 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5448 def __exit__(self
, type, value
, traceback
):
5450 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5453 self
._cancel
_all
_tasks
(self
.loop
)
5455 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5456 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5458 def run_with_loop(main
, loop
):
5459 if not asyncio
.iscoroutine(main
):
5460 raise ValueError(f
'a coroutine was expected, got {main!r}')
5463 return loop
.run_until_complete(main
)
5465 loop
.run_until_complete(loop
.shutdown_asyncgens())
5466 if hasattr(loop
, 'shutdown_default_executor'):
5467 loop
.run_until_complete(loop
.shutdown_default_executor())
5470 def _cancel_all_tasks(loop
):
5471 to_cancel
= asyncio
.all_tasks(loop
)
5476 for task
in to_cancel
:
5479 # XXX: "loop" is removed in python 3.10+
5480 loop
.run_until_complete(
5481 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5483 for task
in to_cancel
:
5484 if task
.cancelled():
5486 if task
.exception() is not None:
5487 loop
.call_exception_handler({
5488 'message': 'unhandled exception during asyncio.run() shutdown',
5489 'exception': task
.exception(),
5494 def merge_headers(*dicts
):
5495 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5496 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5499 def cached_method(f
):
5500 """Cache a method"""
5501 signature
= inspect
.signature(f
)
5504 def wrapper(self
, *args
, **kwargs
):
5505 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5506 bound_args
.apply_defaults()
5507 key
= tuple(bound_args
.arguments
.values())[1:]
5509 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
5510 if key
not in cache
:
5511 cache
[key
] = f(self
, *args
, **kwargs
)
5516 class classproperty
:
5517 """property access for class methods with optional caching"""
5518 def __new__(cls
, func
=None, *args
, **kwargs
):
5520 return functools
.partial(cls
, *args
, **kwargs
)
5521 return super().__new
__(cls
)
5523 def __init__(self
, func
, *, cache
=False):
5524 functools
.update_wrapper(self
, func
)
5526 self
._cache
= {} if cache
else None
5528 def __get__(self
, _
, cls
):
5529 if self
._cache
is None:
5530 return self
.func(cls
)
5531 elif cls
not in self
._cache
:
5532 self
._cache
[cls
] = self
.func(cls
)
5533 return self
._cache
[cls
]
5536 class function_with_repr
:
5537 def __init__(self
, func
, repr_
=None):
5538 functools
.update_wrapper(self
, func
)
5539 self
.func
, self
.__repr
= func
, repr_
5541 def __call__(self
, *args
, **kwargs
):
5542 return self
.func(*args
, **kwargs
)
5547 return f
'{self.func.__module__}.{self.func.__qualname__}'
5550 class Namespace(types
.SimpleNamespace
):
5551 """Immutable namespace"""
5554 return iter(self
.__dict
__.values())
5558 return self
.__dict
__.items()
5561 MEDIA_EXTENSIONS
= Namespace(
5562 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5563 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5564 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5565 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma', 'weba'),
5566 thumbnails
=('jpg', 'png', 'webp'),
5567 storyboards
=('mhtml', ),
5568 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5569 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5571 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5572 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5574 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5579 for retry in RetryManager(...):
5582 except SomeException as err:
5586 attempt
, _error
= 0, None
5588 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5589 self
.retries
= _retries
or 0
5590 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5592 def _should_retry(self
):
5593 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5597 if self
._error
is NO_DEFAULT
:
5602 def error(self
, value
):
5606 while self
._should
_retry
():
5607 self
.error
= NO_DEFAULT
5611 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5614 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5615 """Utility function for reporting retries"""
5618 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5623 elif isinstance(e
, ExtractorError
):
5624 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5625 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5627 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5629 info(f
'Sleeping {delay:.2f} seconds ...')
5633 def make_archive_id(ie
, video_id
):
5634 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5635 return f
'{ie_key.lower()} {video_id}'
5638 def truncate_string(s
, left
, right
=0):
5639 assert left
> 3 and right
>= 0
5640 if s
is None or len(s
) <= left
+ right
:
5642 return f
'{s[:left-3]}...{s[-right:] if right else ""}'
5645 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5646 assert 'all' in alias_dict
, '"all" alias is required'
5647 requested
= list(start
or [])
5649 discard
= val
.startswith('-')
5653 if val
in alias_dict
:
5654 val
= alias_dict
[val
] if not discard
else [
5655 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5656 # NB: Do not allow regex in aliases for performance
5657 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5660 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5661 else [val
] if val
in alias_dict
['all'] else None)
5663 raise ValueError(val
)
5666 for item
in current
:
5667 while item
in requested
:
5668 requested
.remove(item
)
5670 requested
.extend(current
)
5672 return orderedSet(requested
)
5677 regex
= r
' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
5679 default
= ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
5680 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
5681 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
5682 ytdl_default
= ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
5683 'height', 'width', 'proto', 'vext', 'abr', 'aext',
5684 'fps', 'fs_approx', 'source', 'id')
5687 'vcodec': {'type': 'ordered', 'regex': True,
5688 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
5689 'acodec': {'type': 'ordered', 'regex': True,
5690 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'ac-?4', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
5691 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
5692 'order': ['dv', '(hdr)?12', r
'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
5693 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
5694 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
5695 'vext': {'type': 'ordered', 'field': 'video_ext',
5696 'order': ('mp4', 'mov', 'webm', 'flv', '', 'none'),
5697 'order_free': ('webm', 'mp4', 'mov', 'flv', '', 'none')},
5698 'aext': {'type': 'ordered', 'regex': True, 'field': 'audio_ext',
5699 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'web[am]', '', 'none'),
5700 'order_free': ('ogg', 'opus', 'web[am]', 'mp3', 'm4a', 'aac', '', 'none')},
5701 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}
,
5702 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
5703 'field': ('vcodec', 'acodec'),
5704 'function': lambda it
: int(any(v
!= 'none' for v
in it
))},
5705 'ie_pref': {'priority': True, 'type': 'extractor'}
,
5706 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
5707 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
5708 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}
,
5709 'quality': {'convert': 'float', 'default': -1}
,
5710 'filesize': {'convert': 'bytes'}
,
5711 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}
,
5712 'id': {'convert': 'string', 'field': 'format_id'}
,
5713 'height': {'convert': 'float_none'}
,
5714 'width': {'convert': 'float_none'}
,
5715 'fps': {'convert': 'float_none'}
,
5716 'channels': {'convert': 'float_none', 'field': 'audio_channels'}
,
5717 'tbr': {'convert': 'float_none'}
,
5718 'vbr': {'convert': 'float_none'}
,
5719 'abr': {'convert': 'float_none'}
,
5720 'asr': {'convert': 'float_none'}
,
5721 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}
,
5723 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}
,
5724 'br': {'type': 'multiple', 'field': ('tbr', 'vbr', 'abr'),
5725 'function': lambda it
: next(filter(None, it
), None)},
5726 'size': {'type': 'multiple', 'field': ('filesize', 'fs_approx'),
5727 'function': lambda it
: next(filter(None, it
), None)},
5728 'ext': {'type': 'combined', 'field': ('vext', 'aext')}
,
5729 'res': {'type': 'multiple', 'field': ('height', 'width'),
5730 'function': lambda it
: (lambda l
: min(l
) if l
else 0)(tuple(filter(None, it
)))},
5732 # Actual field names
5733 'format_id': {'type': 'alias', 'field': 'id'}
,
5734 'preference': {'type': 'alias', 'field': 'ie_pref'}
,
5735 'language_preference': {'type': 'alias', 'field': 'lang'}
,
5736 'source_preference': {'type': 'alias', 'field': 'source'}
,
5737 'protocol': {'type': 'alias', 'field': 'proto'}
,
5738 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}
,
5739 'audio_channels': {'type': 'alias', 'field': 'channels'}
,
5742 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
5743 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
5744 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}
,
5745 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}
,
5746 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}
,
5747 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}
,
5748 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}
,
5749 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}
,
5750 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}
,
5751 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}
,
5752 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}
,
5753 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}
,
5754 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}
,
5755 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}
,
5756 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
5757 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
5758 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
5759 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
5760 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
5761 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
5764 def __init__(self
, ydl
, field_preference
):
5767 self
.evaluate_params(self
.ydl
.params
, field_preference
)
5768 if ydl
.params
.get('verbose'):
5769 self
.print_verbose_info(self
.ydl
.write_debug
)
5771 def _get_field_setting(self
, field
, key
):
5772 if field
not in self
.settings
:
5773 if key
in ('forced', 'priority'):
5775 self
.ydl
.deprecated_feature(f
'Using arbitrary fields ({field}) for format sorting is '
5776 'deprecated and may be removed in a future version')
5777 self
.settings
[field
] = {}
5778 propObj
= self
.settings
[field
]
5779 if key
not in propObj
:
5780 type = propObj
.get('type')
5782 default
= 'preference' if type == 'extractor' else (field
,) if type in ('combined', 'multiple') else field
5783 elif key
== 'convert':
5784 default
= 'order' if type == 'ordered' else 'float_string' if field
else 'ignore'
5786 default
= {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}
.get(key
, None)
5787 propObj
[key
] = default
5790 def _resolve_field_value(self
, field
, value
, convertNone
=False):
5795 value
= value
.lower()
5796 conversion
= self
._get
_field
_setting
(field
, 'convert')
5797 if conversion
== 'ignore':
5799 if conversion
== 'string':
5801 elif conversion
== 'float_none':
5802 return float_or_none(value
)
5803 elif conversion
== 'bytes':
5804 return parse_bytes(value
)
5805 elif conversion
== 'order':
5806 order_list
= (self
._use
_free
_order
and self
._get
_field
_setting
(field
, 'order_free')) or self
._get
_field
_setting
(field
, 'order')
5807 use_regex
= self
._get
_field
_setting
(field
, 'regex')
5808 list_length
= len(order_list
)
5809 empty_pos
= order_list
.index('') if '' in order_list
else list_length
+ 1
5810 if use_regex
and value
is not None:
5811 for i
, regex
in enumerate(order_list
):
5812 if regex
and re
.match(regex
, value
):
5813 return list_length
- i
5814 return list_length
- empty_pos
# not in list
5815 else: # not regex or value = None
5816 return list_length
- (order_list
.index(value
) if value
in order_list
else empty_pos
)
5818 if value
.isnumeric():
5821 self
.settings
[field
]['convert'] = 'string'
5824 def evaluate_params(self
, params
, sort_extractor
):
5825 self
._use
_free
_order
= params
.get('prefer_free_formats', False)
5826 self
._sort
_user
= params
.get('format_sort', [])
5827 self
._sort
_extractor
= sort_extractor
5829 def add_item(field
, reverse
, closest
, limit_text
):
5830 field
= field
.lower()
5831 if field
in self
._order
:
5833 self
._order
.append(field
)
5834 limit
= self
._resolve
_field
_value
(field
, limit_text
)
5837 'closest': False if limit
is None else closest
,
5838 'limit_text': limit_text
,
5840 if field
in self
.settings
:
5841 self
.settings
[field
].update(data
)
5843 self
.settings
[field
] = data
5846 tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'forced'))
5847 + (tuple() if params
.get('format_sort_force', False)
5848 else tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'priority')))
5849 + tuple(self
._sort
_user
) + tuple(sort_extractor
) + self
.default
)
5851 for item
in sort_list
:
5852 match
= re
.match(self
.regex
, item
)
5854 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item
)
5855 field
= match
.group('field')
5858 if self
._get
_field
_setting
(field
, 'type') == 'alias':
5859 alias
, field
= field
, self
._get
_field
_setting
(field
, 'field')
5860 if self
._get
_field
_setting
(alias
, 'deprecated'):
5861 self
.ydl
.deprecated_feature(f
'Format sorting alias {alias} is deprecated and may '
5862 f
'be removed in a future version. Please use {field} instead')
5863 reverse
= match
.group('reverse') is not None
5864 closest
= match
.group('separator') == '~'
5865 limit_text
= match
.group('limit')
5867 has_limit
= limit_text
is not None
5868 has_multiple_fields
= self
._get
_field
_setting
(field
, 'type') == 'combined'
5869 has_multiple_limits
= has_limit
and has_multiple_fields
and not self
._get
_field
_setting
(field
, 'same_limit')
5871 fields
= self
._get
_field
_setting
(field
, 'field') if has_multiple_fields
else (field
,)
5872 limits
= limit_text
.split(':') if has_multiple_limits
else (limit_text
,) if has_limit
else tuple()
5873 limit_count
= len(limits
)
5874 for (i
, f
) in enumerate(fields
):
5875 add_item(f
, reverse
, closest
,
5876 limits
[i
] if i
< limit_count
5877 else limits
[0] if has_limit
and not has_multiple_limits
5880 def print_verbose_info(self
, write_debug
):
5882 write_debug('Sort order given by user: %s' % ', '.join(self
._sort
_user
))
5883 if self
._sort
_extractor
:
5884 write_debug('Sort order given by extractor: %s' % ', '.join(self
._sort
_extractor
))
5885 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
5886 '+' if self
._get
_field
_setting
(field
, 'reverse') else '', field
,
5887 '%s%s(%s)' % ('~' if self
._get
_field
_setting
(field
, 'closest') else ':',
5888 self
._get
_field
_setting
(field
, 'limit_text'),
5889 self
._get
_field
_setting
(field
, 'limit'))
5890 if self
._get
_field
_setting
(field
, 'limit_text') is not None else '')
5891 for field
in self
._order
if self
._get
_field
_setting
(field
, 'visible')]))
5893 def _calculate_field_preference_from_value(self
, format
, field
, type, value
):
5894 reverse
= self
._get
_field
_setting
(field
, 'reverse')
5895 closest
= self
._get
_field
_setting
(field
, 'closest')
5896 limit
= self
._get
_field
_setting
(field
, 'limit')
5898 if type == 'extractor':
5899 maximum
= self
._get
_field
_setting
(field
, 'max')
5900 if value
is None or (maximum
is not None and value
>= maximum
):
5902 elif type == 'boolean':
5903 in_list
= self
._get
_field
_setting
(field
, 'in_list')
5904 not_in_list
= self
._get
_field
_setting
(field
, 'not_in_list')
5905 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
5906 elif type == 'ordered':
5907 value
= self
._resolve
_field
_value
(field
, value
, True)
5909 # try to convert to number
5910 val_num
= float_or_none(value
, default
=self
._get
_field
_setting
(field
, 'default'))
5911 is_num
= self
._get
_field
_setting
(field
, 'convert') != 'string' and val_num
is not None
5915 return ((-10, 0) if value
is None
5916 else (1, value
, 0) if not is_num
# if a field has mixed strings and numbers, strings are sorted higher
5917 else (0, -abs(value
- limit
), value
- limit
if reverse
else limit
- value
) if closest
5918 else (0, value
, 0) if not reverse
and (limit
is None or value
<= limit
)
5919 else (0, -value
, 0) if limit
is None or (reverse
and value
== limit
) or value
> limit
5920 else (-1, value
, 0))
5922 def _calculate_field_preference(self
, format
, field
):
5923 type = self
._get
_field
_setting
(field
, 'type') # extractor, boolean, ordered, field, multiple
5924 get_value
= lambda f
: format
.get(self
._get
_field
_setting
(f
, 'field'))
5925 if type == 'multiple':
5926 type = 'field' # Only 'field' is allowed in multiple for now
5927 actual_fields
= self
._get
_field
_setting
(field
, 'field')
5929 value
= self
._get
_field
_setting
(field
, 'function')(get_value(f
) for f
in actual_fields
)
5931 value
= get_value(field
)
5932 return self
._calculate
_field
_preference
_from
_value
(format
, field
, type, value
)
5934 def calculate_preference(self
, format
):
5935 # Determine missing protocol
5936 if not format
.get('protocol'):
5937 format
['protocol'] = determine_protocol(format
)
5939 # Determine missing ext
5940 if not format
.get('ext') and 'url' in format
:
5941 format
['ext'] = determine_ext(format
['url'])
5942 if format
.get('vcodec') == 'none':
5943 format
['audio_ext'] = format
['ext'] if format
.get('acodec') != 'none' else 'none'
5944 format
['video_ext'] = 'none'
5946 format
['video_ext'] = format
['ext']
5947 format
['audio_ext'] = 'none'
5948 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
5949 # format['preference'] = -1000
5951 if format
.get('preference') is None and format
.get('ext') == 'flv' and re
.match('[hx]265|he?vc?', format
.get('vcodec') or ''):
5952 # HEVC-over-FLV is out-of-spec by FLV's original spec
5953 # ref. https://trac.ffmpeg.org/ticket/6389
5954 # ref. https://github.com/yt-dlp/yt-dlp/pull/5821
5955 format
['preference'] = -100
5957 # Determine missing bitrates
5958 if format
.get('vcodec') == 'none':
5960 if format
.get('acodec') == 'none':
5962 if not format
.get('vbr') and format
.get('vcodec') != 'none':
5963 format
['vbr'] = try_call(lambda: format
['tbr'] - format
['abr']) or None
5964 if not format
.get('abr') and format
.get('acodec') != 'none':
5965 format
['abr'] = try_call(lambda: format
['tbr'] - format
['vbr']) or None
5966 if not format
.get('tbr'):
5967 format
['tbr'] = try_call(lambda: format
['vbr'] + format
['abr']) or None
5969 return tuple(self
._calculate
_field
_preference
(format
, field
) for field
in self
._order
)