46 import xml
.etree
.ElementTree
49 from .compat
import asyncio
, functools
# isort: split
51 compat_etree_fromstring
,
53 compat_HTMLParseError
,
57 from .dependencies
import brotli
, certifi
, websockets
, xattr
58 from .socks
import ProxyType
, sockssocket
61 def register_socks_protocols():
62 # "Register" SOCKS protocols
63 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
64 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
65 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
66 if scheme
not in urllib
.parse
.uses_netloc
:
67 urllib
.parse
.uses_netloc
.append(scheme
)
70 # This is not clearly defined otherwise
71 compiled_regex_type
= type(re
.compile(''))
74 def random_user_agent():
75 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
116 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
119 SUPPORTED_ENCODINGS
= [
123 SUPPORTED_ENCODINGS
.append('br')
126 'User-Agent': random_user_agent(),
127 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
128 'Accept-Language': 'en-us,en;q=0.5',
129 'Sec-Fetch-Mode': 'navigate',
134 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
138 NO_DEFAULT
= object()
139 IDENTITY
= lambda x
: x
141 ENGLISH_MONTH_NAMES
= [
142 'January', 'February', 'March', 'April', 'May', 'June',
143 'July', 'August', 'September', 'October', 'November', 'December']
146 'en': ENGLISH_MONTH_NAMES
,
148 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
149 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
153 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
154 'flv', 'f4v', 'f4a', 'f4b',
155 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
156 'mkv', 'mka', 'mk3d',
165 'f4f', 'f4m', 'm3u8', 'smil')
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([
227 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
228 DATE_FORMATS_MONTH_FIRST
.extend([
236 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
237 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}
)\s
*</script
>'
239 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
243 def preferredencoding():
244 """Get preferred encoding.
246 Returns the best encoding scheme for the system, based on
247 locale.getpreferredencoding() and some further tweaks.
250 pref = locale.getpreferredencoding()
258 def write_json_file(obj, fn):
259 """ Encode obj as JSON and write it to fn, atomically if possible """
261 tf = tempfile.NamedTemporaryFile(
262 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
263 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
267 json.dump(obj, tf, ensure_ascii=False)
268 if sys.platform == 'win32
':
269 # Need to remove existing file on Windows, else os.rename raises
270 # WindowsError or FileExistsError.
271 with contextlib.suppress(OSError):
273 with contextlib.suppress(OSError):
276 os.chmod(tf.name, 0o666 & ~mask)
277 os.rename(tf.name, fn)
279 with contextlib.suppress(OSError):
284 def find_xpath_attr(node, xpath, key, val=None):
285 """ Find the xpath xpath[@key=val] """
286 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
287 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
288 return node.find(expr)
290 # On python2.6 the xml.etree.ElementTree.Element methods don't support
291 # the namespace parameter
294 def xpath_with_ns(path
, ns_map
):
295 components
= [c
.split(':') for c
in path
.split('/')]
299 replaced
.append(c
[0])
302 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
303 return '/'.join(replaced
)
306 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
307 def _find_xpath(xpath
):
308 return node
.find(xpath
)
310 if isinstance(xpath
, str):
311 n
= _find_xpath(xpath
)
319 if default
is not NO_DEFAULT
:
322 name
= xpath
if name
is None else name
323 raise ExtractorError('Could not find XML element %s' % name
)
329 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
330 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
331 if n
is None or n
== default
:
334 if default
is not NO_DEFAULT
:
337 name
= xpath
if name
is None else name
338 raise ExtractorError('Could not find XML element\'s text %s' % name
)
344 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
345 n
= find_xpath_attr(node
, xpath
, key
)
347 if default
is not NO_DEFAULT
:
350 name
= f
'{xpath}[@{key}]' if name
is None else name
351 raise ExtractorError('Could not find XML attribute %s' % name
)
357 def get_element_by_id(id, html
, **kwargs
):
358 """Return the content of the tag with the specified ID in the passed HTML document"""
359 return get_element_by_attribute('id', id, html
, **kwargs
)
362 def get_element_html_by_id(id, html
, **kwargs
):
363 """Return the html of the tag with the specified ID in the passed HTML document"""
364 return get_element_html_by_attribute('id', id, html
, **kwargs
)
367 def get_element_by_class(class_name
, html
):
368 """Return the content of the first tag with the specified class in the passed HTML document"""
369 retval
= get_elements_by_class(class_name
, html
)
370 return retval
[0] if retval
else None
373 def get_element_html_by_class(class_name
, html
):
374 """Return the html of the first tag with the specified class in the passed HTML document"""
375 retval
= get_elements_html_by_class(class_name
, html
)
376 return retval
[0] if retval
else None
379 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
380 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
381 return retval
[0] if retval
else None
384 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
385 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
386 return retval
[0] if retval
else None
389 def get_elements_by_class(class_name
, html
, **kargs
):
390 """Return the content of all tags with the specified class in the passed HTML document as a list"""
391 return get_elements_by_attribute(
392 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
393 html, escape_value=False)
396 def get_elements_html_by_class(class_name, html):
397 """Return the html of all tags with the specified class in the passed HTML document as a list"""
398 return get_elements_html_by_attribute(
399 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
400 html, escape_value=False)
403 def get_elements_by_attribute(*args, **kwargs):
404 """Return the content of the tag with the specified attribute in the passed HTML document"""
405 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
408 def get_elements_html_by_attribute(*args, **kwargs):
409 """Return the html of the tag with the specified attribute in the passed HTML document"""
410 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
413 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
415 Return the text (content) and the html (whole) of the tag with the specified
416 attribute in the passed HTML document
419 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
421 value = re.escape(value) if escape_value else value
423 partial_element_re = rf'''(?x
)
424 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
425 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
426 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
429 for m in re.finditer(partial_element_re, html):
430 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
433 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
438 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
440 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
441 closing tag for the first opening tag it has encountered, and can be used
445 class HTMLBreakOnClosingTagException(Exception):
449 self.tagstack = collections.deque()
450 html.parser.HTMLParser.__init__(self)
455 def __exit__(self, *_):
459 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
460 # so data remains buffered; we no longer have any interest in it, thus
461 # override this method to discard it
464 def handle_starttag(self, tag, _):
465 self.tagstack.append(tag)
467 def handle_endtag(self, tag):
468 if not self.tagstack:
469 raise compat_HTMLParseError('no tags
in the stack
')
471 inner_tag = self.tagstack.pop()
475 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
476 if not self.tagstack:
477 raise self.HTMLBreakOnClosingTagException()
480 def get_element_text_and_html_by_tag(tag, html):
482 For the first element with the specified tag in the passed HTML document
483 return its' content (text
) and the whole
element (html
)
485 def find_or_raise(haystack, needle, exc):
487 return haystack.index(needle)
490 closing_tag = f'</{tag}>'
491 whole_start = find_or_raise(
492 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
493 content_start = find_or_raise(
494 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
495 content_start += whole_start + 1
496 with HTMLBreakOnClosingTagParser() as parser:
497 parser.feed(html[whole_start:content_start])
498 if not parser.tagstack or parser.tagstack[0] != tag:
499 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
500 offset = content_start
501 while offset < len(html):
502 next_closing_tag_start = find_or_raise(
503 html[offset:], closing_tag,
504 compat_HTMLParseError(f'closing {tag} tag not found'))
505 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
507 parser.feed(html[offset:offset + next_closing_tag_end])
508 offset += next_closing_tag_end
509 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
510 return html[content_start:offset + next_closing_tag_start], \
511 html[whole_start:offset + next_closing_tag_end]
512 raise compat_HTMLParseError('unexpected end of html')
515 class HTMLAttributeParser(html.parser.HTMLParser):
516 """Trivial HTML parser to gather the attributes
for a single element
"""
520 html.parser.HTMLParser.__init__(self)
522 def handle_starttag(self, tag, attrs):
523 self.attrs = dict(attrs)
526 class HTMLListAttrsParser(html.parser.HTMLParser):
527 """HTML parser to gather the attributes
for the elements of a
list"""
530 html.parser.HTMLParser.__init__(self)
534 def handle_starttag(self, tag, attrs):
535 if tag == 'li' and self._level == 0:
536 self.items.append(dict(attrs))
539 def handle_endtag(self, tag):
543 def extract_attributes(html_element):
544 """Given a string
for an HTML element such
as
546 a
="foo" B
="bar" c
="&98;az" d
=boz
547 empty
= noval entity
="&"
550 Decode
and return a dictionary of attributes
.
552 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
553 'empty': '', 'noval': None, 'entity': '&',
554 'sq': '"', 'dq': '\''
557 parser = HTMLAttributeParser()
558 with contextlib.suppress(compat_HTMLParseError):
559 parser.feed(html_element)
564 def parse_list(webpage):
565 """Given a string
for an series of HTML
<li
> elements
,
566 return a dictionary of their attributes
"""
567 parser = HTMLListAttrsParser()
573 def clean_html(html):
574 """Clean an HTML snippet into a readable string
"""
576 if html is None: # Convenience for sanitizing descriptions etc.
579 html = re.sub(r'\s+', ' ', html)
580 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
581 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
583 html = re.sub('<.*?>', '', html)
584 # Replace html entities
585 html = unescapeHTML(html)
589 class LenientJSONDecoder(json.JSONDecoder):
590 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
591 self.transform_source, self.ignore_extra = transform_source, ignore_extra
592 super().__init__(*args, **kwargs)
595 if self.transform_source:
596 s = self.transform_source(s)
597 if self.ignore_extra:
598 return self.raw_decode(s.lstrip())[0]
599 return super().decode(s)
602 def sanitize_open(filename, open_mode):
603 """Try to
open the given filename
, and slightly tweak it
if this fails
.
605 Attempts to
open the given filename
. If this fails
, it tries to change
606 the filename slightly
, step by step
, until it
's either able to open it
607 or it fails and raises a final exception, like the standard open()
610 It returns the tuple (stream, definitive_file_name).
613 if sys.platform == 'win32
':
615 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
616 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
618 for attempt in range(2):
621 if sys.platform == 'win32
':
622 # FIXME: An exclusive lock also locks the file from being read.
623 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
624 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
625 raise LockingUnsupportedError()
626 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
628 stream
= open(filename
, open_mode
)
629 return stream
, filename
630 except OSError as err
:
631 if attempt
or err
.errno
in (errno
.EACCES
,):
633 old_filename
, filename
= filename
, sanitize_path(filename
)
634 if old_filename
== filename
:
638 def timeconvert(timestr
):
639 """Convert RFC 2822 defined time string into system timestamp"""
641 timetuple
= email
.utils
.parsedate_tz(timestr
)
642 if timetuple
is not None:
643 timestamp
= email
.utils
.mktime_tz(timetuple
)
647 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
648 """Sanitizes a string so it could be used as part of a filename.
649 @param restricted Use a stricter subset of allowed characters
650 @param is_id Whether this is an ID that should be kept unchanged if possible.
651 If unset, yt-dlp's new sanitization rules are in effect
656 def replace_insane(char
):
657 if restricted
and char
in ACCENT_CHARS
:
658 return ACCENT_CHARS
[char
]
659 elif not restricted
and char
== '\n':
661 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
664 return '' if restricted
else '\''
666 return '\0_\0-' if restricted
else '\0 \0-'
667 elif char
in '\\/|*<>':
669 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
673 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
674 result
= ''.join(map(replace_insane
, s
))
675 if is_id
is NO_DEFAULT
:
676 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
677 STRIP_RE
= r
'(?:\0.|[ _-])*'
678 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
679 result
= result
.replace('\0', '') or '_'
682 while '__' in result
:
683 result
= result
.replace('__', '_')
684 result
= result
.strip('_')
685 # Common case of "Foreign band name - English song title"
686 if restricted
and result
.startswith('-_'):
688 if result
.startswith('-'):
689 result
= '_' + result
[len('-'):]
690 result
= result
.lstrip('.')
696 def sanitize_path(s
, force
=False):
697 """Sanitizes and normalizes path on Windows"""
698 if sys
.platform
== 'win32':
700 drive_or_unc
, _
= os
.path
.splitdrive(s
)
706 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
710 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
711 for path_part
in norm_path
]
713 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
714 elif force
and s
and s
[0] == os
.path
.sep
:
715 sanitized_path
.insert(0, os
.path
.sep
)
716 return os
.path
.join(*sanitized_path
)
719 def sanitize_url(url
):
720 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
721 # the number of unwanted failures due to missing protocol
724 elif url
.startswith('//'):
725 return 'http:%s' % url
726 # Fix some common typos seen so far
728 # https://github.com/ytdl-org/youtube-dl/issues/15649
729 (r
'^httpss://', r
'https://'),
730 # https://bx1.be/lives/direct-tv/
731 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
733 for mistake
, fixup
in COMMON_TYPOS
:
734 if re
.match(mistake
, url
):
735 return re
.sub(mistake
, fixup
, url
)
739 def extract_basic_auth(url
):
740 parts
= urllib
.parse
.urlsplit(url
)
741 if parts
.username
is None:
743 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
744 parts
.hostname
if parts
.port
is None
745 else '%s:%d' % (parts
.hostname
, parts
.port
))))
746 auth_payload
= base64
.b64encode(
747 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
748 return url
, f
'Basic {auth_payload.decode()}'
751 def sanitized_Request(url
, *args
, **kwargs
):
752 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
753 if auth_header
is not None:
754 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
755 headers
['Authorization'] = auth_header
756 return urllib
.request
.Request(url
, *args
, **kwargs
)
760 """Expand shell variables and ~"""
761 return os
.path
.expandvars(compat_expanduser(s
))
764 def orderedSet(iterable
, *, lazy
=False):
765 """Remove all duplicates from the input iterable"""
767 seen
= [] # Do not use set since the items can be unhashable
773 return _iter() if lazy
else list(_iter())
776 def _htmlentity_transform(entity_with_semicolon
):
777 """Transforms an HTML entity to a character."""
778 entity
= entity_with_semicolon
[:-1]
780 # Known non-numeric HTML entity
781 if entity
in html
.entities
.name2codepoint
:
782 return chr(html
.entities
.name2codepoint
[entity
])
784 # TODO: HTML5 allows entities without a semicolon. For example,
785 # 'Éric' should be decoded as 'Éric'.
786 if entity_with_semicolon
in html
.entities
.html5
:
787 return html
.entities
.html5
[entity_with_semicolon
]
789 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
791 numstr
= mobj
.group(1)
792 if numstr
.startswith('x'):
794 numstr
= '0%s' % numstr
797 # See https://github.com/ytdl-org/youtube-dl/issues/7518
798 with contextlib
.suppress(ValueError):
799 return chr(int(numstr
, base
))
801 # Unknown entity in name, return its literal representation
802 return '&%s;' % entity
808 assert isinstance(s
, str)
811 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
814 def escapeHTML(text
):
817 .replace('&', '&')
818 .replace('<', '<')
819 .replace('>', '>')
820 .replace('"', '"')
821 .replace("'", ''')
825 def process_communicate_or_kill(p
, *args
, **kwargs
):
826 write_string('DeprecationWarning: yt_dlp.utils.process_communicate_or_kill is deprecated '
827 'and may be removed in a future version. Use yt_dlp.utils.Popen.communicate_or_kill instead')
828 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
831 class Popen(subprocess
.Popen
):
832 if sys
.platform
== 'win32':
833 _startupinfo
= subprocess
.STARTUPINFO()
834 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
838 def __init__(self
, *args
, text
=False, **kwargs
):
840 kwargs
['universal_newlines'] = True # For 3.6 compatibility
841 kwargs
.setdefault('encoding', 'utf-8')
842 kwargs
.setdefault('errors', 'replace')
843 super().__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
845 def communicate_or_kill(self
, *args
, **kwargs
):
847 return self
.communicate(*args
, **kwargs
)
848 except BaseException
: # Including KeyboardInterrupt
849 self
.kill(timeout
=None)
852 def kill(self
, *, timeout
=0):
855 self
.wait(timeout
=timeout
)
858 def run(cls
, *args
, **kwargs
):
859 with cls(*args
, **kwargs
) as proc
:
860 stdout
, stderr
= proc
.communicate_or_kill()
861 return stdout
or '', stderr
or '', proc
.returncode
864 def get_subprocess_encoding():
865 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
866 # For subprocess calls, encode with locale encoding
867 # Refer to http://stackoverflow.com/a/9951851/35070
868 encoding
= preferredencoding()
870 encoding
= sys
.getfilesystemencoding()
876 def encodeFilename(s
, for_subprocess
=False):
877 assert isinstance(s
, str)
881 def decodeFilename(b
, for_subprocess
=False):
885 def encodeArgument(s
):
886 # Legacy code that uses byte strings
887 # Uncomment the following line after fixing all post processors
888 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
889 return s
if isinstance(s
, str) else s
.decode('ascii')
892 def decodeArgument(b
):
896 def decodeOption(optval
):
899 if isinstance(optval
, bytes):
900 optval
= optval
.decode(preferredencoding())
902 assert isinstance(optval
, str)
906 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
909 def timetuple_from_msec(msec
):
910 secs
, msec
= divmod(msec
, 1000)
911 mins
, secs
= divmod(secs
, 60)
912 hrs
, mins
= divmod(mins
, 60)
913 return _timetuple(hrs
, mins
, secs
, msec
)
916 def formatSeconds(secs
, delim
=':', msec
=False):
917 time
= timetuple_from_msec(secs
* 1000)
919 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
921 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
923 ret
= '%d' % time
.seconds
924 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
927 def _ssl_load_windows_store_certs(ssl_context
, storename
):
928 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
930 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
931 if encoding
== 'x509_asn' and (
932 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
933 except PermissionError
:
936 with contextlib
.suppress(ssl
.SSLError
):
937 ssl_context
.load_verify_locations(cadata
=cert
)
940 def make_HTTPS_handler(params
, **kwargs
):
941 opts_check_certificate
= not params
.get('nocheckcertificate')
942 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
943 context
.check_hostname
= opts_check_certificate
944 if params
.get('legacyserverconnect'):
945 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
946 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
947 context
.set_ciphers('DEFAULT')
949 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
950 if opts_check_certificate
:
951 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
952 context
.load_verify_locations(cafile
=certifi
.where())
955 context
.load_default_certs()
956 # Work around the issue in load_default_certs when there are bad certificates. See:
957 # https://github.com/yt-dlp/yt-dlp/issues/1060,
958 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
960 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
961 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
962 for storename
in ('CA', 'ROOT'):
963 _ssl_load_windows_store_certs(context
, storename
)
964 context
.set_default_verify_paths()
966 client_certfile
= params
.get('client_certificate')
969 context
.load_cert_chain(
970 client_certfile
, keyfile
=params
.get('client_certificate_key'),
971 password
=params
.get('client_certificate_password'))
973 raise YoutubeDLError('Unable to load client certificate')
975 # Some servers may reject requests if ALPN extension is not sent. See:
976 # https://github.com/python/cpython/issues/85140
977 # https://github.com/yt-dlp/yt-dlp/issues/3878
978 with contextlib
.suppress(NotImplementedError):
979 context
.set_alpn_protocols(['http/1.1'])
981 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
984 def bug_reports_message(before
=';'):
985 from .update
import REPOSITORY
987 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
988 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
990 before
= before
.rstrip()
991 if not before
or before
.endswith(('.', '!', '?')):
992 msg
= msg
[0].title() + msg
[1:]
994 return (before
+ ' ' if before
else '') + msg
997 class YoutubeDLError(Exception):
998 """Base exception for YoutubeDL errors."""
1001 def __init__(self
, msg
=None):
1004 elif self
.msg
is None:
1005 self
.msg
= type(self
).__name
__
1006 super().__init
__(self
.msg
)
1009 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1010 if hasattr(ssl
, 'CertificateError'):
1011 network_exceptions
.append(ssl
.CertificateError
)
1012 network_exceptions
= tuple(network_exceptions
)
1015 class ExtractorError(YoutubeDLError
):
1016 """Error during info extraction."""
1018 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1019 """ tb, if given, is the original traceback (so that it can be printed out).
1020 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1022 if sys
.exc_info()[0] in network_exceptions
:
1025 self
.orig_msg
= str(msg
)
1027 self
.expected
= expected
1029 self
.video_id
= video_id
1031 self
.exc_info
= sys
.exc_info() # preserve original exception
1032 if isinstance(self
.exc_info
[1], ExtractorError
):
1033 self
.exc_info
= self
.exc_info
[1].exc_info
1035 super().__init
__(''.join((
1036 format_field(ie
, None, '[%s] '),
1037 format_field(video_id
, None, '%s: '),
1039 format_field(cause
, None, ' (caused by %r)'),
1040 '' if expected
else bug_reports_message())))
1042 def format_traceback(self
):
1043 return join_nonempty(
1044 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1045 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1049 class UnsupportedError(ExtractorError
):
1050 def __init__(self
, url
):
1052 'Unsupported URL: %s' % url
, expected
=True)
1056 class RegexNotFoundError(ExtractorError
):
1057 """Error when a regex didn't match"""
1061 class GeoRestrictedError(ExtractorError
):
1062 """Geographic restriction Error exception.
1064 This exception may be thrown when a video is not available from your
1065 geographic location due to geographic restrictions imposed by a website.
1068 def __init__(self
, msg
, countries
=None, **kwargs
):
1069 kwargs
['expected'] = True
1070 super().__init
__(msg
, **kwargs
)
1071 self
.countries
= countries
1074 class DownloadError(YoutubeDLError
):
1075 """Download Error exception.
1077 This exception may be thrown by FileDownloader objects if they are not
1078 configured to continue on errors. They will contain the appropriate
1082 def __init__(self
, msg
, exc_info
=None):
1083 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1084 super().__init
__(msg
)
1085 self
.exc_info
= exc_info
1088 class EntryNotInPlaylist(YoutubeDLError
):
1089 """Entry not in playlist exception.
1091 This exception will be thrown by YoutubeDL when a requested entry
1092 is not found in the playlist info_dict
1094 msg
= 'Entry not found in info'
1097 class SameFileError(YoutubeDLError
):
1098 """Same File exception.
1100 This exception will be thrown by FileDownloader objects if they detect
1101 multiple files would have to be downloaded to the same file on disk.
1103 msg
= 'Fixed output name but more than one file to download'
1105 def __init__(self
, filename
=None):
1106 if filename
is not None:
1107 self
.msg
+= f
': {filename}'
1108 super().__init
__(self
.msg
)
1111 class PostProcessingError(YoutubeDLError
):
1112 """Post Processing exception.
1114 This exception may be raised by PostProcessor's .run() method to
1115 indicate an error in the postprocessing task.
1119 class DownloadCancelled(YoutubeDLError
):
1120 """ Exception raised when the download queue should be interrupted """
1121 msg
= 'The download was cancelled'
1124 class ExistingVideoReached(DownloadCancelled
):
1125 """ --break-on-existing triggered """
1126 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1129 class RejectedVideoReached(DownloadCancelled
):
1130 """ --break-on-reject triggered """
1131 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1134 class MaxDownloadsReached(DownloadCancelled
):
1135 """ --max-downloads limit has been reached. """
1136 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1139 class ReExtractInfo(YoutubeDLError
):
1140 """ Video info needs to be re-extracted. """
1142 def __init__(self
, msg
, expected
=False):
1143 super().__init
__(msg
)
1144 self
.expected
= expected
1147 class ThrottledDownload(ReExtractInfo
):
1148 """ Download speed below --throttled-rate. """
1149 msg
= 'The download speed is below throttle limit'
1152 super().__init
__(self
.msg
, expected
=False)
1155 class UnavailableVideoError(YoutubeDLError
):
1156 """Unavailable Format exception.
1158 This exception will be thrown when a video is requested
1159 in a format that is not available for that video.
1161 msg
= 'Unable to download video'
1163 def __init__(self
, err
=None):
1165 self
.msg
+= f
': {err}'
1166 super().__init
__(self
.msg
)
1169 class ContentTooShortError(YoutubeDLError
):
1170 """Content Too Short exception.
1172 This exception may be raised by FileDownloader objects when a file they
1173 download is too small for what the server announced first, indicating
1174 the connection was probably interrupted.
1177 def __init__(self
, downloaded
, expected
):
1178 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1180 self
.downloaded
= downloaded
1181 self
.expected
= expected
1184 class XAttrMetadataError(YoutubeDLError
):
1185 def __init__(self
, code
=None, msg
='Unknown error'):
1186 super().__init
__(msg
)
1190 # Parsing code and msg
1191 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1192 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1193 self
.reason
= 'NO_SPACE'
1194 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1195 self
.reason
= 'VALUE_TOO_LONG'
1197 self
.reason
= 'NOT_SUPPORTED'
1200 class XAttrUnavailableError(YoutubeDLError
):
1204 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1205 hc
= http_class(*args
, **kwargs
)
1206 source_address
= ydl_handler
._params
.get('source_address')
1208 if source_address
is not None:
1209 # This is to workaround _create_connection() from socket where it will try all
1210 # address data from getaddrinfo() including IPv6. This filters the result from
1211 # getaddrinfo() based on the source_address value.
1212 # This is based on the cpython socket.create_connection() function.
1213 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1214 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1215 host
, port
= address
1217 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1218 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1219 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1220 if addrs
and not ip_addrs
:
1221 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1223 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1224 % (ip_version
, source_address
[0]))
1225 for res
in ip_addrs
:
1226 af
, socktype
, proto
, canonname
, sa
= res
1229 sock
= socket
.socket(af
, socktype
, proto
)
1230 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1231 sock
.settimeout(timeout
)
1232 sock
.bind(source_address
)
1234 err
= None # Explicitly break reference cycle
1236 except OSError as _
:
1238 if sock
is not None:
1243 raise OSError('getaddrinfo returns an empty list')
1244 if hasattr(hc
, '_create_connection'):
1245 hc
._create
_connection
= _create_connection
1246 hc
.source_address
= (source_address
, 0)
1251 def handle_youtubedl_headers(headers
):
1252 filtered_headers
= headers
1254 if 'Youtubedl-no-compression' in filtered_headers
:
1255 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1256 del filtered_headers
['Youtubedl-no-compression']
1258 return filtered_headers
1261 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1262 """Handler for HTTP requests and responses.
1264 This class, when installed with an OpenerDirector, automatically adds
1265 the standard headers to every HTTP request and handles gzipped and
1266 deflated responses from web servers. If compression is to be avoided in
1267 a particular request, the original request in the program code only has
1268 to include the HTTP header "Youtubedl-no-compression", which will be
1269 removed before making the real request.
1271 Part of this code was copied from:
1273 http://techknack.net/python-urllib2-handlers/
1275 Andrew Rowls, the author of that code, agreed to release it to the
1279 def __init__(self
, params
, *args
, **kwargs
):
1280 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1281 self
._params
= params
1283 def http_open(self
, req
):
1284 conn_class
= http
.client
.HTTPConnection
1286 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1288 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1289 del req
.headers
['Ytdl-socks-proxy']
1291 return self
.do_open(functools
.partial(
1292 _create_http_connection
, self
, conn_class
, False),
1300 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1302 return zlib
.decompress(data
)
1308 return brotli
.decompress(data
)
1310 def http_request(self
, req
):
1311 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1312 # always respected by websites, some tend to give out URLs with non percent-encoded
1313 # non-ASCII characters (see telemb.py, ard.py [#3412])
1314 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1315 # To work around aforementioned issue we will replace request's original URL with
1316 # percent-encoded one
1317 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1318 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1319 url
= req
.get_full_url()
1320 url_escaped
= escape_url(url
)
1322 # Substitute URL if any change after escaping
1323 if url
!= url_escaped
:
1324 req
= update_Request(req
, url
=url_escaped
)
1326 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1327 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1328 # The dict keys are capitalized because of this bug by urllib
1329 if h
.capitalize() not in req
.headers
:
1330 req
.add_header(h
, v
)
1332 if 'Accept-encoding' not in req
.headers
:
1333 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1335 req
.headers
= handle_youtubedl_headers(req
.headers
)
1337 return super().do_request_(req
)
1339 def http_response(self
, req
, resp
):
1342 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1343 content
= resp
.read()
1344 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1346 uncompressed
= io
.BytesIO(gz
.read())
1347 except OSError as original_ioerror
:
1348 # There may be junk add the end of the file
1349 # See http://stackoverflow.com/q/4928560/35070 for details
1350 for i
in range(1, 1024):
1352 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1353 uncompressed
= io
.BytesIO(gz
.read())
1358 raise original_ioerror
1359 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1360 resp
.msg
= old_resp
.msg
1361 del resp
.headers
['Content-encoding']
1363 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1364 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1365 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1366 resp
.msg
= old_resp
.msg
1367 del resp
.headers
['Content-encoding']
1369 if resp
.headers
.get('Content-encoding', '') == 'br':
1370 resp
= urllib
.request
.addinfourl(
1371 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1372 resp
.msg
= old_resp
.msg
1373 del resp
.headers
['Content-encoding']
1374 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1375 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1376 if 300 <= resp
.code
< 400:
1377 location
= resp
.headers
.get('Location')
1379 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1380 location
= location
.encode('iso-8859-1').decode()
1381 location_escaped
= escape_url(location
)
1382 if location
!= location_escaped
:
1383 del resp
.headers
['Location']
1384 resp
.headers
['Location'] = location_escaped
1387 https_request
= http_request
1388 https_response
= http_response
1391 def make_socks_conn_class(base_class
, socks_proxy
):
1392 assert issubclass(base_class
, (
1393 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1395 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1396 if url_components
.scheme
.lower() == 'socks5':
1397 socks_type
= ProxyType
.SOCKS5
1398 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1399 socks_type
= ProxyType
.SOCKS4
1400 elif url_components
.scheme
.lower() == 'socks4a':
1401 socks_type
= ProxyType
.SOCKS4A
1403 def unquote_if_non_empty(s
):
1406 return urllib
.parse
.unquote_plus(s
)
1410 url_components
.hostname
, url_components
.port
or 1080,
1412 unquote_if_non_empty(url_components
.username
),
1413 unquote_if_non_empty(url_components
.password
),
1416 class SocksConnection(base_class
):
1418 self
.sock
= sockssocket()
1419 self
.sock
.setproxy(*proxy_args
)
1420 if isinstance(self
.timeout
, (int, float)):
1421 self
.sock
.settimeout(self
.timeout
)
1422 self
.sock
.connect((self
.host
, self
.port
))
1424 if isinstance(self
, http
.client
.HTTPSConnection
):
1425 if hasattr(self
, '_context'): # Python > 2.6
1426 self
.sock
= self
._context
.wrap_socket(
1427 self
.sock
, server_hostname
=self
.host
)
1429 self
.sock
= ssl
.wrap_socket(self
.sock
)
1431 return SocksConnection
1434 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1435 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1436 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1437 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1438 self
._params
= params
1440 def https_open(self
, req
):
1442 conn_class
= self
._https
_conn
_class
1444 if hasattr(self
, '_context'): # python > 2.6
1445 kwargs
['context'] = self
._context
1446 if hasattr(self
, '_check_hostname'): # python 3.x
1447 kwargs
['check_hostname'] = self
._check
_hostname
1449 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1451 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1452 del req
.headers
['Ytdl-socks-proxy']
1455 return self
.do_open(
1456 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1457 except urllib
.error
.URLError
as e
:
1458 if (isinstance(e
.reason
, ssl
.SSLError
)
1459 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1460 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1464 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1466 See [1] for cookie file format.
1468 1. https://curl.haxx.se/docs/http-cookies.html
1470 _HTTPONLY_PREFIX
= '#HttpOnly_'
1472 _HEADER
= '''# Netscape HTTP Cookie File
1473 # This file is generated by yt-dlp. Do not edit.
1476 _CookieFileEntry
= collections
.namedtuple(
1478 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1480 def __init__(self
, filename
=None, *args
, **kwargs
):
1481 super().__init
__(None, *args
, **kwargs
)
1482 if self
.is_path(filename
):
1483 filename
= os
.fspath(filename
)
1484 self
.filename
= filename
1487 def _true_or_false(cndn
):
1488 return 'TRUE' if cndn
else 'FALSE'
1492 return isinstance(file, (str, bytes, os
.PathLike
))
1494 @contextlib.contextmanager
1495 def open(self
, file, *, write
=False):
1496 if self
.is_path(file):
1497 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1504 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1507 if (not ignore_discard
and cookie
.discard
1508 or not ignore_expires
and cookie
.is_expired(now
)):
1510 name
, value
= cookie
.name
, cookie
.value
1512 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1513 # with no name, whereas http.cookiejar regards it as a
1514 # cookie with no value.
1515 name
, value
= '', name
1516 f
.write('%s\n' % '\t'.join((
1518 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1520 self
._true
_or
_false
(cookie
.secure
),
1521 str_or_none(cookie
.expires
, default
=''),
1525 def save(self
, filename
=None, *args
, **kwargs
):
1527 Save cookies to a file.
1528 Code is taken from CPython 3.6
1529 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1531 if filename
is None:
1532 if self
.filename
is not None:
1533 filename
= self
.filename
1535 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1537 # Store session cookies with `expires` set to 0 instead of an empty string
1539 if cookie
.expires
is None:
1542 with self
.open(filename
, write
=True) as f
:
1543 f
.write(self
._HEADER
)
1544 self
._really
_save
(f
, *args
, **kwargs
)
1546 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1547 """Load cookies from a file."""
1548 if filename
is None:
1549 if self
.filename
is not None:
1550 filename
= self
.filename
1552 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1554 def prepare_line(line
):
1555 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1556 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1557 # comments and empty lines are fine
1558 if line
.startswith('#') or not line
.strip():
1560 cookie_list
= line
.split('\t')
1561 if len(cookie_list
) != self
._ENTRY
_LEN
:
1562 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1563 cookie
= self
._CookieFileEntry
(*cookie_list
)
1564 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1565 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1569 with self
.open(filename
) as f
:
1572 cf
.write(prepare_line(line
))
1573 except http
.cookiejar
.LoadError
as e
:
1574 if f
'{line.strip()} '[0] in '[{"':
1575 raise http
.cookiejar
.LoadError(
1576 'Cookies file must be Netscape formatted, not JSON. See '
1577 'https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl')
1578 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1581 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1582 # Session cookies are denoted by either `expires` field set to
1583 # an empty string or 0. MozillaCookieJar only recognizes the former
1584 # (see [1]). So we need force the latter to be recognized as session
1585 # cookies on our own.
1586 # Session cookies may be important for cookies-based authentication,
1587 # e.g. usually, when user does not check 'Remember me' check box while
1588 # logging in on a site, some important cookies are stored as session
1589 # cookies so that not recognizing them will result in failed login.
1590 # 1. https://bugs.python.org/issue17164
1592 # Treat `expires=0` cookies as session cookies
1593 if cookie
.expires
== 0:
1594 cookie
.expires
= None
1595 cookie
.discard
= True
1598 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1599 def __init__(self
, cookiejar
=None):
1600 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1602 def http_response(self
, request
, response
):
1603 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1605 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1606 https_response
= http_response
1609 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1610 """YoutubeDL redirect handler
1612 The code is based on HTTPRedirectHandler implementation from CPython [1].
1614 This redirect handler solves two issues:
1615 - ensures redirect URL is always unicode under python 2
1616 - introduces support for experimental HTTP response status code
1617 308 Permanent Redirect [2] used by some sites [3]
1619 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1620 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1621 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1624 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1626 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1627 """Return a Request or None in response to a redirect.
1629 This is called by the http_error_30x methods when a
1630 redirection response is received. If a redirection should
1631 take place, return a new Request to allow http_error_30x to
1632 perform the redirect. Otherwise, raise HTTPError if no-one
1633 else should try to handle this url. Return None if you can't
1634 but another Handler might.
1636 m
= req
.get_method()
1637 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1638 or code
in (301, 302, 303) and m
== "POST")):
1639 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1640 # Strictly (according to RFC 2616), 301 or 302 in response to
1641 # a POST MUST NOT cause a redirection without confirmation
1642 # from the user (of urllib.request, in this case). In practice,
1643 # essentially all clients do redirect in this case, so we do
1646 # Be conciliant with URIs containing a space. This is mainly
1647 # redundant with the more complete encoding done in http_error_302(),
1648 # but it is kept for compatibility with other callers.
1649 newurl
= newurl
.replace(' ', '%20')
1651 CONTENT_HEADERS
= ("content-length", "content-type")
1652 # NB: don't use dict comprehension for python 2.6 compatibility
1653 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1655 # A 303 must either use GET or HEAD for subsequent request
1656 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1657 if code
== 303 and m
!= 'HEAD':
1659 # 301 and 302 redirects are commonly turned into a GET from a POST
1660 # for subsequent requests by browsers, so we'll do the same.
1661 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1662 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1663 if code
in (301, 302) and m
== 'POST':
1666 return urllib
.request
.Request(
1667 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1668 unverifiable
=True, method
=m
)
1671 def extract_timezone(date_str
):
1674 ^.{8,}? # >=8 char non-TZ prefix, if present
1675 (?P<tz>Z| # just the UTC Z, or
1676 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1677 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1678 [ ]? # optional space
1679 (?P<sign>\+|-) # +/-
1680 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1684 timezone
= datetime
.timedelta()
1686 date_str
= date_str
[:-len(m
.group('tz'))]
1687 if not m
.group('sign'):
1688 timezone
= datetime
.timedelta()
1690 sign
= 1 if m
.group('sign') == '+' else -1
1691 timezone
= datetime
.timedelta(
1692 hours
=sign
* int(m
.group('hours')),
1693 minutes
=sign
* int(m
.group('minutes')))
1694 return timezone
, date_str
1697 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1698 """ Return a UNIX timestamp from the given date """
1700 if date_str
is None:
1703 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1705 if timezone
is None:
1706 timezone
, date_str
= extract_timezone(date_str
)
1708 with contextlib
.suppress(ValueError):
1709 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1710 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1711 return calendar
.timegm(dt
.timetuple())
1714 def date_formats(day_first
=True):
1715 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1718 def unified_strdate(date_str
, day_first
=True):
1719 """Return a string with the date in the format YYYYMMDD"""
1721 if date_str
is None:
1725 date_str
= date_str
.replace(',', ' ')
1726 # Remove AM/PM + timezone
1727 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1728 _
, date_str
= extract_timezone(date_str
)
1730 for expression
in date_formats(day_first
):
1731 with contextlib
.suppress(ValueError):
1732 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1733 if upload_date
is None:
1734 timetuple
= email
.utils
.parsedate_tz(date_str
)
1736 with contextlib
.suppress(ValueError):
1737 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1738 if upload_date
is not None:
1739 return str(upload_date
)
1742 def unified_timestamp(date_str
, day_first
=True):
1743 if date_str
is None:
1746 date_str
= re
.sub(r
'[,|]', '', date_str
)
1748 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1749 timezone
, date_str
= extract_timezone(date_str
)
1751 # Remove AM/PM + timezone
1752 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1754 # Remove unrecognized timezones from ISO 8601 alike timestamps
1755 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1757 date_str
= date_str
[:-len(m
.group('tz'))]
1759 # Python only supports microseconds, so remove nanoseconds
1760 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]+$', date_str
)
1762 date_str
= m
.group(1)
1764 for expression
in date_formats(day_first
):
1765 with contextlib
.suppress(ValueError):
1766 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1767 return calendar
.timegm(dt
.timetuple())
1768 timetuple
= email
.utils
.parsedate_tz(date_str
)
1770 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1773 def determine_ext(url
, default_ext
='unknown_video'):
1774 if url
is None or '.' not in url
:
1776 guess
= url
.partition('?')[0].rpartition('.')[2]
1777 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1779 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1780 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1781 return guess
.rstrip('/')
1786 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1787 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1790 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1792 Return a datetime object from a string.
1794 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1796 @param format strftime format of DATE
1797 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1798 auto: round to the unit provided in date_str (if applicable).
1800 auto_precision
= False
1801 if precision
== 'auto':
1802 auto_precision
= True
1803 precision
= 'microsecond'
1804 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1805 if date_str
in ('now', 'today'):
1807 if date_str
== 'yesterday':
1808 return today
- datetime
.timedelta(days
=1)
1810 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1812 if match
is not None:
1813 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1814 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1815 unit
= match
.group('unit')
1816 if unit
== 'month' or unit
== 'year':
1817 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1823 delta
= datetime
.timedelta(**{unit + 's': time}
)
1824 new_date
= start_time
+ delta
1826 return datetime_round(new_date
, unit
)
1829 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1832 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1834 Return a date object from a string using datetime_from_str
1836 @param strict Restrict allowed patterns to "YYYYMMDD" and
1837 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1839 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1840 raise ValueError(f
'Invalid date format "{date_str}"')
1841 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1844 def datetime_add_months(dt
, months
):
1845 """Increment/Decrement a datetime object by months."""
1846 month
= dt
.month
+ months
- 1
1847 year
= dt
.year
+ month
// 12
1848 month
= month
% 12 + 1
1849 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1850 return dt
.replace(year
, month
, day
)
1853 def datetime_round(dt
, precision
='day'):
1855 Round a datetime object's time to a specific precision
1857 if precision
== 'microsecond':
1866 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1867 timestamp
= calendar
.timegm(dt
.timetuple())
1868 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1871 def hyphenate_date(date_str
):
1873 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1874 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1875 if match
is not None:
1876 return '-'.join(match
.groups())
1882 """Represents a time interval between two dates"""
1884 def __init__(self
, start
=None, end
=None):
1885 """start and end must be strings in the format accepted by date"""
1886 if start
is not None:
1887 self
.start
= date_from_str(start
, strict
=True)
1889 self
.start
= datetime
.datetime
.min.date()
1891 self
.end
= date_from_str(end
, strict
=True)
1893 self
.end
= datetime
.datetime
.max.date()
1894 if self
.start
> self
.end
:
1895 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1899 """Returns a range that only contains the given day"""
1900 return cls(day
, day
)
1902 def __contains__(self
, date
):
1903 """Check if the date is in the range"""
1904 if not isinstance(date
, datetime
.date
):
1905 date
= date_from_str(date
)
1906 return self
.start
<= date
<= self
.end
1909 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1912 def platform_name():
1913 """ Returns the platform name as a str """
1914 write_string('DeprecationWarning: yt_dlp.utils.platform_name is deprecated, use platform.platform instead')
1915 return platform
.platform()
1919 def system_identifier():
1920 python_implementation
= platform
.python_implementation()
1921 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
1922 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
1924 return 'Python %s (%s %s) - %s %s' % (
1925 platform
.python_version(),
1926 python_implementation
,
1927 platform
.architecture()[0],
1928 platform
.platform(),
1929 format_field(join_nonempty(*platform
.libc_ver(), delim
=' '), None, '(%s)'),
1934 def get_windows_version():
1935 ''' Get Windows version. returns () if it's not running on Windows '''
1936 if compat_os_name
== 'nt':
1937 return version_tuple(platform
.win32_ver()[1])
1942 def write_string(s
, out
=None, encoding
=None):
1943 assert isinstance(s
, str)
1944 out
= out
or sys
.stderr
1946 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
1947 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1949 enc
, buffer = None, out
1950 if 'b' in getattr(out
, 'mode', ''):
1951 enc
= encoding
or preferredencoding()
1952 elif hasattr(out
, 'buffer'):
1954 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1956 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
1960 def bytes_to_intlist(bs
):
1963 if isinstance(bs
[0], int): # Python 3
1966 return [ord(c
) for c
in bs
]
1969 def intlist_to_bytes(xs
):
1972 return struct
.pack('%dB' % len(xs
), *xs
)
1975 class LockingUnsupportedError(OSError):
1976 msg
= 'File locking is not supported'
1979 super().__init
__(self
.msg
)
1982 # Cross-platform file locking
1983 if sys
.platform
== 'win32':
1984 import ctypes
.wintypes
1987 class OVERLAPPED(ctypes
.Structure
):
1989 ('Internal', ctypes
.wintypes
.LPVOID
),
1990 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1991 ('Offset', ctypes
.wintypes
.DWORD
),
1992 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1993 ('hEvent', ctypes
.wintypes
.HANDLE
),
1996 kernel32
= ctypes
.windll
.kernel32
1997 LockFileEx
= kernel32
.LockFileEx
1998 LockFileEx
.argtypes
= [
1999 ctypes
.wintypes
.HANDLE
, # hFile
2000 ctypes
.wintypes
.DWORD
, # dwFlags
2001 ctypes
.wintypes
.DWORD
, # dwReserved
2002 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2003 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2004 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2006 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2007 UnlockFileEx
= kernel32
.UnlockFileEx
2008 UnlockFileEx
.argtypes
= [
2009 ctypes
.wintypes
.HANDLE
, # hFile
2010 ctypes
.wintypes
.DWORD
, # dwReserved
2011 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2012 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2013 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2015 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2016 whole_low
= 0xffffffff
2017 whole_high
= 0x7fffffff
2019 def _lock_file(f
, exclusive
, block
):
2020 overlapped
= OVERLAPPED()
2021 overlapped
.Offset
= 0
2022 overlapped
.OffsetHigh
= 0
2023 overlapped
.hEvent
= 0
2024 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2026 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2027 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2028 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2029 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2030 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2032 def _unlock_file(f
):
2033 assert f
._lock
_file
_overlapped
_p
2034 handle
= msvcrt
.get_osfhandle(f
.fileno())
2035 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2036 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2042 def _lock_file(f
, exclusive
, block
):
2043 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2045 flags |
= fcntl
.LOCK_NB
2047 fcntl
.flock(f
, flags
)
2048 except BlockingIOError
:
2050 except OSError: # AOSP does not have flock()
2051 fcntl
.lockf(f
, flags
)
2053 def _unlock_file(f
):
2055 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2057 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2061 def _lock_file(f
, exclusive
, block
):
2062 raise LockingUnsupportedError()
2064 def _unlock_file(f
):
2065 raise LockingUnsupportedError()
2071 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2072 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2073 raise NotImplementedError(mode
)
2074 self
.mode
, self
.block
= mode
, block
2076 writable
= any(f
in mode
for f
in 'wax+')
2077 readable
= any(f
in mode
for f
in 'r+')
2078 flags
= functools
.reduce(operator
.ior
, (
2079 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2080 getattr(os
, 'O_BINARY', 0), # Windows only
2081 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2082 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2083 os
.O_APPEND
if 'a' in mode
else 0,
2084 os
.O_EXCL
if 'x' in mode
else 0,
2085 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2088 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2090 def __enter__(self
):
2091 exclusive
= 'r' not in self
.mode
2093 _lock_file(self
.f
, exclusive
, self
.block
)
2098 if 'w' in self
.mode
:
2101 except OSError as e
:
2103 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2104 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2113 _unlock_file(self
.f
)
2117 def __exit__(self
, *_
):
2126 def __getattr__(self
, attr
):
2127 return getattr(self
.f
, attr
)
2134 def get_filesystem_encoding():
2135 encoding
= sys
.getfilesystemencoding()
2136 return encoding
if encoding
is not None else 'utf-8'
2139 def shell_quote(args
):
2141 encoding
= get_filesystem_encoding()
2143 if isinstance(a
, bytes):
2144 # We may get a filename encoded with 'encodeFilename'
2145 a
= a
.decode(encoding
)
2146 quoted_args
.append(compat_shlex_quote(a
))
2147 return ' '.join(quoted_args
)
2150 def smuggle_url(url
, data
):
2151 """ Pass additional data in a URL for internal use. """
2153 url
, idata
= unsmuggle_url(url
, {})
2155 sdata
= urllib
.parse
.urlencode(
2156 {'__youtubedl_smuggle': json.dumps(data)}
)
2157 return url
+ '#' + sdata
2160 def unsmuggle_url(smug_url
, default
=None):
2161 if '#__youtubedl_smuggle' not in smug_url
:
2162 return smug_url
, default
2163 url
, _
, sdata
= smug_url
.rpartition('#')
2164 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2165 data
= json
.loads(jsond
)
2169 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2170 """ Formats numbers with decimal sufixes like K, M, etc """
2171 num
, factor
= float_or_none(num
), float(factor
)
2172 if num
is None or num
< 0:
2174 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2175 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2176 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2178 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2179 converted
= num
/ (factor
** exponent
)
2180 return fmt
% (converted
, suffix
)
2183 def format_bytes(bytes):
2184 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2187 def lookup_unit_table(unit_table
, s
):
2188 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2190 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2193 num_str
= m
.group('num').replace(',', '.')
2194 mult
= unit_table
[m
.group('unit')]
2195 return int(float(num_str
) * mult
)
2198 def parse_filesize(s
):
2202 # The lower-case forms are of course incorrect and unofficial,
2203 # but we support those too
2220 'megabytes': 1000 ** 2,
2221 'mebibytes': 1024 ** 2,
2227 'gigabytes': 1000 ** 3,
2228 'gibibytes': 1024 ** 3,
2234 'terabytes': 1000 ** 4,
2235 'tebibytes': 1024 ** 4,
2241 'petabytes': 1000 ** 5,
2242 'pebibytes': 1024 ** 5,
2248 'exabytes': 1000 ** 6,
2249 'exbibytes': 1024 ** 6,
2255 'zettabytes': 1000 ** 7,
2256 'zebibytes': 1024 ** 7,
2262 'yottabytes': 1000 ** 8,
2263 'yobibytes': 1024 ** 8,
2266 return lookup_unit_table(_UNIT_TABLE
, s
)
2273 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2275 if re
.match(r
'^[\d,.]+$', s
):
2276 return str_to_int(s
)
2289 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2293 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2295 return str_to_int(mobj
.group(1))
2298 def parse_resolution(s
, *, lenient
=False):
2303 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2305 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2308 'width': int(mobj
.group('w')),
2309 'height': int(mobj
.group('h')),
2312 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2314 return {'height': int(mobj.group(1))}
2316 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2318 return {'height': int(mobj.group(1)) * 540}
2323 def parse_bitrate(s
):
2324 if not isinstance(s
, str):
2326 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2328 return int(mobj
.group(1))
2331 def month_by_name(name
, lang
='en'):
2332 """ Return the number of a month by (locale-independently) English name """
2334 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2337 return month_names
.index(name
) + 1
2342 def month_by_abbreviation(abbrev
):
2343 """ Return the number of a month by (locale-independently) English
2347 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2352 def fix_xml_ampersands(xml_str
):
2353 """Replace all the '&' by '&' in XML"""
2355 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2360 def setproctitle(title
):
2361 assert isinstance(title
, str)
2363 # ctypes in Jython is not complete
2364 # http://bugs.jython.org/issue2148
2365 if sys
.platform
.startswith('java'):
2369 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2373 # LoadLibrary in Windows Python 2.7.13 only expects
2374 # a bytestring, but since unicode_literals turns
2375 # every string into a unicode string, it fails.
2377 title_bytes
= title
.encode()
2378 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2379 buf
.value
= title_bytes
2381 libc
.prctl(15, buf
, 0, 0, 0)
2382 except AttributeError:
2383 return # Strange libc, just skip this
2386 def remove_start(s
, start
):
2387 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2390 def remove_end(s
, end
):
2391 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2394 def remove_quotes(s
):
2395 if s
is None or len(s
) < 2:
2397 for quote
in ('"', "'", ):
2398 if s
[0] == quote
and s
[-1] == quote
:
2403 def get_domain(url
):
2404 return '.'.join(urllib
.parse
.urlparse(url
).netloc
.rsplit('.', 2)[-2:])
2407 def url_basename(url
):
2408 path
= urllib
.parse
.urlparse(url
).path
2409 return path
.strip('/').split('/')[-1]
2413 return re
.match(r
'https?://[^?#&]+/', url
).group()
2416 def urljoin(base
, path
):
2417 if isinstance(path
, bytes):
2418 path
= path
.decode()
2419 if not isinstance(path
, str) or not path
:
2421 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2423 if isinstance(base
, bytes):
2424 base
= base
.decode()
2425 if not isinstance(base
, str) or not re
.match(
2426 r
'^(?:https?:)?//', base
):
2428 return urllib
.parse
.urljoin(base
, path
)
2431 class HEADRequest(urllib
.request
.Request
):
2432 def get_method(self
):
2436 class PUTRequest(urllib
.request
.Request
):
2437 def get_method(self
):
2441 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2442 if get_attr
and v
is not None:
2443 v
= getattr(v
, get_attr
, None)
2445 return int(v
) * invscale
// scale
2446 except (ValueError, TypeError, OverflowError):
2450 def str_or_none(v
, default
=None):
2451 return default
if v
is None else str(v
)
2454 def str_to_int(int_str
):
2455 """ A more relaxed version of int_or_none """
2456 if isinstance(int_str
, int):
2458 elif isinstance(int_str
, str):
2459 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2460 return int_or_none(int_str
)
2463 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2467 return float(v
) * invscale
/ scale
2468 except (ValueError, TypeError):
2472 def bool_or_none(v
, default
=None):
2473 return v
if isinstance(v
, bool) else default
2476 def strip_or_none(v
, default
=None):
2477 return v
.strip() if isinstance(v
, str) else default
2480 def url_or_none(url
):
2481 if not url
or not isinstance(url
, str):
2484 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2487 def request_to_url(req
):
2488 if isinstance(req
, urllib
.request
.Request
):
2489 return req
.get_full_url()
2494 def strftime_or_none(timestamp
, date_format
, default
=None):
2495 datetime_object
= None
2497 if isinstance(timestamp
, (int, float)): # unix timestamp
2498 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2499 elif isinstance(timestamp
, str): # assume YYYYMMDD
2500 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2501 return datetime_object
.strftime(date_format
)
2502 except (ValueError, TypeError, AttributeError):
2506 def parse_duration(s
):
2507 if not isinstance(s
, str):
2513 days
, hours
, mins
, secs
, ms
= [None] * 5
2514 m
= re
.match(r
'''(?x)
2516 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2517 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2518 (?P<ms>[.:][0-9]+)?Z?$
2521 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2526 [0-9]+\s*y(?:ears?)?,?\s*
2529 [0-9]+\s*m(?:onths?)?,?\s*
2532 [0-9]+\s*w(?:eeks?)?,?\s*
2535 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2539 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2542 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2545 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2548 days
, hours
, mins
, secs
, ms
= m
.groups()
2550 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2552 hours
, mins
= m
.groups()
2557 ms
= ms
.replace(':', '.')
2558 return sum(float(part
or 0) * mult
for part
, mult
in (
2559 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2562 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2563 name
, real_ext
= os
.path
.splitext(filename
)
2565 f
'{name}.{ext}{real_ext}'
2566 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2567 else f
'{filename}.{ext}')
2570 def replace_extension(filename
, ext
, expected_real_ext
=None):
2571 name
, real_ext
= os
.path
.splitext(filename
)
2572 return '{}.{}'.format(
2573 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2577 def check_executable(exe
, args
=[]):
2578 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2579 args can be a list of arguments for a short output (like -version) """
2581 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2587 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2589 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2591 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2592 # SIGTTOU if yt-dlp is run in the background.
2593 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2594 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2595 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2601 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2602 assert isinstance(output
, str)
2603 if version_re
is None:
2604 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2605 m
= re
.search(version_re
, output
)
2612 def get_exe_version(exe
, args
=['--version'],
2613 version_re
=None, unrecognized
='present'):
2614 """ Returns the version of the specified executable,
2615 or False if the executable is not present """
2616 out
= _get_exe_version_output(exe
, args
)
2617 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2620 def frange(start
=0, stop
=None, step
=1):
2623 start
, stop
= 0, start
2624 sign
= [-1, 1][step
> 0] if step
else 0
2625 while sign
* start
< sign
* stop
:
2630 class LazyList(collections
.abc
.Sequence
):
2631 """Lazy immutable list from an iterable
2632 Note that slices of a LazyList are lists and not LazyList"""
2634 class IndexError(IndexError):
2637 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2638 self
._iterable
= iter(iterable
)
2639 self
._cache
= [] if _cache
is None else _cache
2640 self
._reversed
= reverse
2644 # We need to consume the entire iterable to iterate in reverse
2645 yield from self
.exhaust()
2647 yield from self
._cache
2648 for item
in self
._iterable
:
2649 self
._cache
.append(item
)
2653 self
._cache
.extend(self
._iterable
)
2654 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2658 """Evaluate the entire iterable"""
2659 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2662 def _reverse_index(x
):
2663 return None if x
is None else -(x
+ 1)
2665 def __getitem__(self
, idx
):
2666 if isinstance(idx
, slice):
2668 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2669 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2670 elif isinstance(idx
, int):
2672 idx
= self
._reverse
_index
(idx
)
2673 start
, stop
, step
= idx
, idx
, 0
2675 raise TypeError('indices must be integers or slices')
2676 if ((start
or 0) < 0 or (stop
or 0) < 0
2677 or (start
is None and step
< 0)
2678 or (stop
is None and step
> 0)):
2679 # We need to consume the entire iterable to be able to slice from the end
2680 # Obviously, never use this with infinite iterables
2683 return self
._cache
[idx
]
2684 except IndexError as e
:
2685 raise self
.IndexError(e
) from e
2686 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2688 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2690 return self
._cache
[idx
]
2691 except IndexError as e
:
2692 raise self
.IndexError(e
) from e
2696 self
[-1] if self
._reversed
else self
[0]
2697 except self
.IndexError:
2703 return len(self
._cache
)
2705 def __reversed__(self
):
2706 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2709 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2712 # repr and str should mimic a list. So we exhaust the iterable
2713 return repr(self
.exhaust())
2716 return repr(self
.exhaust())
2721 class IndexError(IndexError):
2725 # This is only useful for tests
2726 return len(self
.getslice())
2728 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2729 self
._pagefunc
= pagefunc
2730 self
._pagesize
= pagesize
2731 self
._pagecount
= float('inf')
2732 self
._use
_cache
= use_cache
2735 def getpage(self
, pagenum
):
2736 page_results
= self
._cache
.get(pagenum
)
2737 if page_results
is None:
2738 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2740 self
._cache
[pagenum
] = page_results
2743 def getslice(self
, start
=0, end
=None):
2744 return list(self
._getslice
(start
, end
))
2746 def _getslice(self
, start
, end
):
2747 raise NotImplementedError('This method must be implemented by subclasses')
2749 def __getitem__(self
, idx
):
2750 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2751 if not isinstance(idx
, int) or idx
< 0:
2752 raise TypeError('indices must be non-negative integers')
2753 entries
= self
.getslice(idx
, idx
+ 1)
2755 raise self
.IndexError()
2759 class OnDemandPagedList(PagedList
):
2760 """Download pages until a page with less than maximum results"""
2762 def _getslice(self
, start
, end
):
2763 for pagenum
in itertools
.count(start
// self
._pagesize
):
2764 firstid
= pagenum
* self
._pagesize
2765 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2766 if start
>= nextfirstid
:
2770 start
% self
._pagesize
2771 if firstid
<= start
< nextfirstid
2774 ((end
- 1) % self
._pagesize
) + 1
2775 if (end
is not None and firstid
<= end
<= nextfirstid
)
2779 page_results
= self
.getpage(pagenum
)
2781 self
._pagecount
= pagenum
- 1
2783 if startv
!= 0 or endv
is not None:
2784 page_results
= page_results
[startv
:endv
]
2785 yield from page_results
2787 # A little optimization - if current page is not "full", ie. does
2788 # not contain page_size videos then we can assume that this page
2789 # is the last one - there are no more ids on further pages -
2790 # i.e. no need to query again.
2791 if len(page_results
) + startv
< self
._pagesize
:
2794 # If we got the whole page, but the next page is not interesting,
2795 # break out early as well
2796 if end
== nextfirstid
:
2800 class InAdvancePagedList(PagedList
):
2801 """PagedList with total number of pages known in advance"""
2803 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2804 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2805 self
._pagecount
= pagecount
2807 def _getslice(self
, start
, end
):
2808 start_page
= start
// self
._pagesize
2809 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2810 skip_elems
= start
- start_page
* self
._pagesize
2811 only_more
= None if end
is None else end
- start
2812 for pagenum
in range(start_page
, end_page
):
2813 page_results
= self
.getpage(pagenum
)
2815 page_results
= page_results
[skip_elems
:]
2817 if only_more
is not None:
2818 if len(page_results
) < only_more
:
2819 only_more
-= len(page_results
)
2821 yield from page_results
[:only_more
]
2823 yield from page_results
2826 class PlaylistEntries
:
2827 MissingEntry
= object()
2828 is_exhausted
= False
2830 def __init__(self
, ydl
, info_dict
):
2833 # _entries must be assigned now since infodict can change during iteration
2834 entries
= info_dict
.get('entries')
2836 raise EntryNotInPlaylist('There are no entries')
2837 elif isinstance(entries
, list):
2838 self
.is_exhausted
= True
2840 requested_entries
= info_dict
.get('requested_entries')
2841 self
.is_incomplete
= bool(requested_entries
)
2842 if self
.is_incomplete
:
2843 assert self
.is_exhausted
2844 self
._entries
= [self
.MissingEntry
] * max(requested_entries
)
2845 for i
, entry
in zip(requested_entries
, entries
):
2846 self
._entries
[i
- 1] = entry
2847 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2848 self
._entries
= entries
2850 self
._entries
= LazyList(entries
)
2852 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2853 (?P<start>[+-]?\d+)?
2855 (?P<end>[+-]?\d+|inf(?:inite)?)?
2856 (?::(?P<step>[+-]?\d+))?
2860 def parse_playlist_items(cls
, string
):
2861 for segment
in string
.split(','):
2863 raise ValueError('There is two or more consecutive commas')
2864 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2866 raise ValueError(f
'{segment!r} is not a valid specification')
2867 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2868 if int_or_none(step
) == 0:
2869 raise ValueError(f
'Step in {segment!r} cannot be zero')
2870 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2872 def get_requested_items(self
):
2873 playlist_items
= self
.ydl
.params
.get('playlist_items')
2874 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2875 playlist_end
= self
.ydl
.params
.get('playlistend')
2876 # For backwards compatibility, interpret -1 as whole list
2877 if playlist_end
in (-1, None):
2879 if not playlist_items
:
2880 playlist_items
= f
'{playlist_start}:{playlist_end}'
2881 elif playlist_start
!= 1 or playlist_end
:
2882 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2884 for index
in self
.parse_playlist_items(playlist_items
):
2885 for i
, entry
in self
[index
]:
2890 # TODO: Add auto-generated fields
2891 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
2892 except (ExistingVideoReached
, RejectedVideoReached
):
2895 def get_full_count(self
):
2896 if self
.is_exhausted
and not self
.is_incomplete
:
2898 elif isinstance(self
._entries
, InAdvancePagedList
):
2899 if self
._entries
._pagesize
== 1:
2900 return self
._entries
._pagecount
2902 @functools.cached_property
2904 if isinstance(self
._entries
, list):
2907 entry
= self
._entries
[i
]
2909 entry
= self
.MissingEntry
2910 if not self
.is_incomplete
:
2911 raise self
.IndexError()
2912 if entry
is self
.MissingEntry
:
2913 raise EntryNotInPlaylist(f
'Entry {i} cannot be found')
2918 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
2919 except (LazyList
.IndexError, PagedList
.IndexError):
2920 raise self
.IndexError()
2923 def __getitem__(self
, idx
):
2924 if isinstance(idx
, int):
2925 idx
= slice(idx
, idx
)
2927 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
2928 step
= 1 if idx
.step
is None else idx
.step
2929 if idx
.start
is None:
2930 start
= 0 if step
> 0 else len(self
) - 1
2932 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
2934 # NB: Do not call len(self) when idx == [:]
2935 if idx
.stop
is None:
2936 stop
= 0 if step
< 0 else float('inf')
2938 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
2939 stop
+= [-1, 1][step
> 0]
2941 for i
in frange(start
, stop
, step
):
2945 entry
= self
._getter
(i
)
2946 except self
.IndexError:
2947 self
.is_exhausted
= True
2954 return len(tuple(self
[:]))
2956 class IndexError(IndexError):
2960 def uppercase_escape(s
):
2961 unicode_escape
= codecs
.getdecoder('unicode_escape')
2963 r
'\\U[0-9a-fA-F]{8}',
2964 lambda m
: unicode_escape(m
.group(0))[0],
2968 def lowercase_escape(s
):
2969 unicode_escape
= codecs
.getdecoder('unicode_escape')
2971 r
'\\u[0-9a-fA-F]{4}',
2972 lambda m
: unicode_escape(m
.group(0))[0],
2976 def escape_rfc3986(s
):
2977 """Escape non-ASCII characters as suggested by RFC 3986"""
2978 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2981 def escape_url(url
):
2982 """Escape URL as suggested by RFC 3986"""
2983 url_parsed
= urllib
.parse
.urlparse(url
)
2984 return url_parsed
._replace
(
2985 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2986 path
=escape_rfc3986(url_parsed
.path
),
2987 params
=escape_rfc3986(url_parsed
.params
),
2988 query
=escape_rfc3986(url_parsed
.query
),
2989 fragment
=escape_rfc3986(url_parsed
.fragment
)
2994 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
)
2997 def read_batch_urls(batch_fd
):
2999 if not isinstance(url
, str):
3000 url
= url
.decode('utf-8', 'replace')
3001 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3002 for bom
in BOM_UTF8
:
3003 if url
.startswith(bom
):
3004 url
= url
[len(bom
):]
3006 if not url
or url
.startswith(('#', ';', ']')):
3008 # "#" cannot be stripped out since it is part of the URI
3009 # However, it can be safely stripped out if following a whitespace
3010 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3012 with contextlib
.closing(batch_fd
) as fd
:
3013 return [url
for url
in map(fixup
, fd
) if url
]
3016 def urlencode_postdata(*args
, **kargs
):
3017 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3020 def update_url_query(url
, query
):
3023 parsed_url
= urllib
.parse
.urlparse(url
)
3024 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3026 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3027 query
=urllib
.parse
.urlencode(qs
, True)))
3030 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3031 req_headers
= req
.headers
.copy()
3032 req_headers
.update(headers
or {})
3033 req_data
= data
or req
.data
3034 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3035 req_get_method
= req
.get_method()
3036 if req_get_method
== 'HEAD':
3037 req_type
= HEADRequest
3038 elif req_get_method
== 'PUT':
3039 req_type
= PUTRequest
3041 req_type
= urllib
.request
.Request
3043 req_url
, data
=req_data
, headers
=req_headers
,
3044 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3045 if hasattr(req
, 'timeout'):
3046 new_req
.timeout
= req
.timeout
3050 def _multipart_encode_impl(data
, boundary
):
3051 content_type
= 'multipart/form-data; boundary=%s' % boundary
3054 for k
, v
in data
.items():
3055 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3056 if isinstance(k
, str):
3058 if isinstance(v
, str):
3060 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3061 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3062 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3063 if boundary
.encode('ascii') in content
:
3064 raise ValueError('Boundary overlaps with data')
3067 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3069 return out
, content_type
3072 def multipart_encode(data
, boundary
=None):
3074 Encode a dict to RFC 7578-compliant form-data
3077 A dict where keys and values can be either Unicode or bytes-like
3080 If specified a Unicode object, it's used as the boundary. Otherwise
3081 a random boundary is generated.
3083 Reference: https://tools.ietf.org/html/rfc7578
3085 has_specified_boundary
= boundary
is not None
3088 if boundary
is None:
3089 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3092 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3095 if has_specified_boundary
:
3099 return out
, content_type
3102 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3103 for val
in map(d
.get
, variadic(key_or_keys
)):
3104 if val
is not None and (val
or not skip_false_values
):
3109 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3112 val
= f(*args
, **kwargs
)
3113 except (AttributeError, KeyError, TypeError, IndexError, ZeroDivisionError):
3116 if expected_type
is None or isinstance(val
, expected_type
):
3120 def try_get(src
, getter
, expected_type
=None):
3121 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3124 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3125 return {k: v for k, v in dct.items() if cndn(k, v)}
3128 def merge_dicts(*dicts
):
3130 for a_dict
in dicts
:
3131 for k
, v
in a_dict
.items():
3132 if (v
is not None and k
not in merged
3133 or isinstance(v
, str) and merged
[k
] == ''):
3138 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3139 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3151 TV_PARENTAL_GUIDELINES
= {
3161 def parse_age_limit(s
):
3162 # isinstance(False, int) is True. So type() must be used instead
3163 if type(s
) is int: # noqa: E721
3164 return s
if 0 <= s
<= 21 else None
3165 elif not isinstance(s
, str):
3167 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3169 return int(m
.group('age'))
3172 return US_RATINGS
[s
]
3173 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3175 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3179 def strip_jsonp(code
):
3182 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3183 (?:\s*&&\s*(?P=func_name))?
3184 \s*\(\s*(?P<callback_data>.*)\);?
3185 \s*?(?://[^\n]*)*$''',
3186 r
'\g<callback_data>', code
)
3189 def js_to_json(code
, vars={}):
3190 # vars is a dict of var, val pairs to substitute
3191 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3192 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3194 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3195 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3200 if v
in ('true', 'false', 'null'):
3202 elif v
in ('undefined', 'void 0'):
3204 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3207 if v
[0] in ("'", '"'):
3208 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3213 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3215 for regex
, base
in INTEGER_TABLE
:
3216 im
= re
.match(regex
, v
)
3218 i
= int(im
.group(1), base
)
3219 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3226 def create_map(mobj
):
3227 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3229 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3230 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3232 return re
.sub(r
'''(?sx)
3233 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3234 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3235 {comment}|,(?={skip}[\]}}])|
3236 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3237 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3240 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3243 def qualities(quality_ids
):
3244 """ Get a numeric quality value out of a list of possible values """
3247 return quality_ids
.index(qid
)
3253 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3257 'default': '%(title)s [%(id)s].%(ext)s',
3258 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3264 'description': 'description',
3265 'annotation': 'annotations.xml',
3266 'infojson': 'info.json',
3269 'pl_thumbnail': None,
3270 'pl_description': 'description',
3271 'pl_infojson': 'info.json',
3274 # As of [1] format syntax is:
3275 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3276 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3277 STR_FORMAT_RE_TMPL
= r
'''(?x)
3278 (?<!%)(?P<prefix>(?:%%)*)
3280 (?P<has_key>\((?P<key>{0})\))?
3282 (?P<conversion>[#0\-+ ]+)?
3284 (?P<precision>\.\d+)?
3285 (?P<len_mod>[hlL])? # unused in python
3286 {1} # conversion type
3291 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3294 def limit_length(s
, length
):
3295 """ Add ellipses to overly long strings """
3300 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3304 def version_tuple(v
):
3305 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3308 def is_outdated_version(version
, limit
, assume_new
=True):
3310 return not assume_new
3312 return version_tuple(version
) < version_tuple(limit
)
3314 return not assume_new
3317 def ytdl_is_updateable():
3318 """ Returns if yt-dlp can be updated with -U """
3320 from .update
import is_non_updateable
3322 return not is_non_updateable()
3325 def args_to_str(args
):
3326 # Get a short string representation for a subprocess command
3327 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3330 def error_to_compat_str(err
):
3334 def error_to_str(err
):
3335 return f
'{type(err).__name__}: {err}'
3338 def mimetype2ext(mt
):
3342 mt
, _
, params
= mt
.partition(';')
3347 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3348 # it's the most popular one
3349 'audio/mpeg': 'mp3',
3350 'audio/x-wav': 'wav',
3352 'audio/wave': 'wav',
3355 ext
= FULL_MAP
.get(mt
)
3361 'smptett+xml': 'tt',
3365 'x-mp4-fragmented': 'mp4',
3366 'x-ms-sami': 'sami',
3369 'x-mpegurl': 'm3u8',
3370 'vnd.apple.mpegurl': 'm3u8',
3374 'vnd.ms-sstr+xml': 'ism',
3378 'filmstrip+json': 'fs',
3382 _
, _
, subtype
= mt
.rpartition('/')
3383 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3394 _
, _
, suffix
= subtype
.partition('+')
3395 ext
= SUFFIX_MAP
.get(suffix
)
3399 return subtype
.replace('+', '.')
3402 def ext2mimetype(ext_or_url
):
3405 if '.' not in ext_or_url
:
3406 ext_or_url
= f
'file.{ext_or_url}'
3407 return mimetypes
.guess_type(ext_or_url
)[0]
3410 def parse_codecs(codecs_str
):
3411 # http://tools.ietf.org/html/rfc6381
3414 split_codecs
= list(filter(None, map(
3415 str.strip
, codecs_str
.strip().strip(',').split(','))))
3416 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3417 for full_codec
in split_codecs
:
3418 parts
= full_codec
.split('.')
3419 codec
= parts
[0].replace('0', '')
3420 if codec
in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3421 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3423 vcodec
= '.'.join(parts
[:4]) if codec
in ('vp9', 'av1', 'hvc1') else full_codec
3424 if codec
in ('dvh1', 'dvhe'):
3426 elif codec
== 'av1' and len(parts
) > 3 and parts
[3] == '10':
3428 elif full_codec
.replace('0', '').startswith('vp9.2'):
3430 elif codec
in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3433 elif codec
in ('stpp', 'wvtt',):
3437 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3438 if vcodec
or acodec
or scodec
:
3440 'vcodec': vcodec
or 'none',
3441 'acodec': acodec
or 'none',
3442 'dynamic_range': hdr
,
3443 **({'scodec': scodec}
if scodec
is not None else {}),
3445 elif len(split_codecs
) == 2:
3447 'vcodec': split_codecs
[0],
3448 'acodec': split_codecs
[1],
3453 def urlhandle_detect_ext(url_handle
):
3454 getheader
= url_handle
.headers
.get
3456 cd
= getheader('Content-Disposition')
3458 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3460 e
= determine_ext(m
.group('filename'), default_ext
=None)
3464 return mimetype2ext(getheader('Content-Type'))
3467 def encode_data_uri(data
, mime_type
):
3468 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3471 def age_restricted(content_limit
, age_limit
):
3472 """ Returns True iff the content should be blocked """
3474 if age_limit
is None: # No limit set
3476 if content_limit
is None:
3477 return False # Content available for everyone
3478 return age_limit
< content_limit
3481 def is_html(first_bytes
):
3482 """ Detect whether a file contains HTML by examining its first bytes. """
3485 (b
'\xef\xbb\xbf', 'utf-8'),
3486 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3487 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3488 (b
'\xff\xfe', 'utf-16-le'),
3489 (b
'\xfe\xff', 'utf-16-be'),
3493 for bom
, enc
in BOMS
:
3494 while first_bytes
.startswith(bom
):
3495 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3497 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3500 def determine_protocol(info_dict
):
3501 protocol
= info_dict
.get('protocol')
3502 if protocol
is not None:
3505 url
= sanitize_url(info_dict
['url'])
3506 if url
.startswith('rtmp'):
3508 elif url
.startswith('mms'):
3510 elif url
.startswith('rtsp'):
3513 ext
= determine_ext(url
)
3519 return urllib
.parse
.urlparse(url
).scheme
3522 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3523 """ Render a list of rows, each as a list of values.
3524 Text after a \t will be right aligned """
3526 return len(remove_terminal_sequences(string
).replace('\t', ''))
3528 def get_max_lens(table
):
3529 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3531 def filter_using_list(row
, filterArray
):
3532 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3534 max_lens
= get_max_lens(data
) if hide_empty
else []
3535 header_row
= filter_using_list(header_row
, max_lens
)
3536 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3538 table
= [header_row
] + data
3539 max_lens
= get_max_lens(table
)
3542 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3543 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3545 for pos
, text
in enumerate(map(str, row
)):
3547 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3549 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3550 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3554 def _match_one(filter_part
, dct
, incomplete
):
3555 # TODO: Generalize code with YoutubeDL._build_format_filter
3556 STRING_OPERATORS
= {
3557 '*=': operator
.contains
,
3558 '^=': lambda attr
, value
: attr
.startswith(value
),
3559 '$=': lambda attr
, value
: attr
.endswith(value
),
3560 '~=': lambda attr
, value
: re
.search(value
, attr
),
3562 COMPARISON_OPERATORS
= {
3564 '<=': operator
.le
, # "<=" must be defined above "<"
3571 if isinstance(incomplete
, bool):
3572 is_incomplete
= lambda _
: incomplete
3574 is_incomplete
= lambda k
: k
in incomplete
3576 operator_rex
= re
.compile(r
'''(?x)
3578 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3580 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3583 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3584 m = operator_rex.fullmatch(filter_part.strip())
3587 unnegated_op = COMPARISON_OPERATORS[m['op']]
3589 op = lambda attr, value: not unnegated_op(attr, value)
3592 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3594 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3595 actual_value = dct.get(m['key'])
3596 numeric_comparison = None
3597 if isinstance(actual_value, (int, float)):
3598 # If the original field is a string and matching comparisonvalue is
3599 # a number we should respect the origin of the original field
3600 # and process comparison value as a string (see
3601 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3603 numeric_comparison = int(comparison_value)
3605 numeric_comparison = parse_filesize(comparison_value)
3606 if numeric_comparison is None:
3607 numeric_comparison = parse_filesize(f'{comparison_value}B')
3608 if numeric_comparison is None:
3609 numeric_comparison = parse_duration(comparison_value)
3610 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3611 raise ValueError('Operator %s only supports string values!' % m['op'])
3612 if actual_value is None:
3613 return is_incomplete(m['key']) or m['none_inclusive']
3614 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3617 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3618 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3620 operator_rex = re.compile(r'''(?x
)
3621 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3622 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3623 m = operator_rex.fullmatch(filter_part.strip())
3625 op = UNARY_OPERATORS[m.group('op')]
3626 actual_value = dct.get(m.group('key'))
3627 if is_incomplete(m.group('key')) and actual_value is None:
3629 return op(actual_value)
3631 raise ValueError('Invalid filter part %r' % filter_part)
3634 def match_str(filter_str, dct, incomplete=False):
3635 """ Filter a dictionary with a simple string syntax.
3636 @returns Whether the filter passes
3637 @param incomplete Set of keys that is expected to be missing from dct.
3638 Can be True/False to indicate all/none of the keys may be missing.
3639 All conditions on incomplete keys pass if the key is missing
3642 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3643 for filter_part in re.split(r'(?<!\\)&', filter_str))
3646 def match_filter_func(filters):
3649 filters = set(variadic(filters))
3651 interactive = '-' in filters
3655 def _match_func(info_dict, incomplete=False):
3656 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3657 return NO_DEFAULT if interactive and not incomplete else None
3659 video_title = info_dict.get('title') or info_dict.get('id') or 'video'
3660 filter_str = ') | ('.join(map(str.strip, filters))
3661 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3665 def download_range_func(chapters, ranges):
3666 def inner(info_dict, ydl):
3667 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3668 else 'Cannot match chapters since chapter information is unavailable')
3669 for regex in chapters or []:
3670 for i, chapter in enumerate(info_dict.get('chapters') or []):
3671 if re.search(regex, chapter['title']):
3673 yield {**chapter, 'index': i}
3674 if chapters and warning:
3675 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3677 yield from ({'start_time': start, 'end_time': end} for start, end in ranges or [])
3682 def parse_dfxp_time_expr(time_expr):
3686 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3688 return float(mobj.group('time_offset'))
3690 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3692 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3695 def srt_subtitles_timecode(seconds):
3696 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3699 def ass_subtitles_timecode(seconds):
3700 time = timetuple_from_msec(seconds * 1000)
3701 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3704 def dfxp2srt(dfxp_data):
3706 @param dfxp_data A
bytes-like
object containing DFXP data
3707 @returns A
unicode object containing converted SRT data
3709 LEGACY_NAMESPACES = (
3710 (b'http://www.w3.org/ns/ttml', [
3711 b'http://www.w3.org/2004/11/ttaf1',
3712 b'http://www.w3.org/2006/04/ttaf1',
3713 b'http://www.w3.org/2006/10/ttaf1',
3715 (b'http://www.w3.org/ns/ttml#styling', [
3716 b'http://www.w3.org/ns/ttml#style',
3720 SUPPORTED_STYLING = [
3729 _x = functools.partial(xpath_with_ns, ns_map={
3730 'xml': 'http://www.w3.org/XML/1998/namespace',
3731 'ttml': 'http://www.w3.org/ns/ttml',
3732 'tts': 'http://www.w3.org/ns/ttml#styling',
3738 class TTMLPElementParser:
3740 _unclosed_elements = []
3741 _applied_styles = []
3743 def start(self, tag, attrib):
3744 if tag in (_x('ttml:br'), 'br'):
3747 unclosed_elements = []
3749 element_style_id = attrib.get('style')
3751 style.update(default_style)
3752 if element_style_id:
3753 style.update(styles.get(element_style_id, {}))
3754 for prop in SUPPORTED_STYLING:
3755 prop_val = attrib.get(_x('tts:' + prop))
3757 style[prop] = prop_val
3760 for k, v in sorted(style.items()):
3761 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3764 font += ' color="%s"' % v
3765 elif k == 'fontSize':
3766 font += ' size="%s"' % v
3767 elif k == 'fontFamily':
3768 font += ' face="%s"' % v
3769 elif k == 'fontWeight' and v == 'bold':
3771 unclosed_elements.append('b')
3772 elif k == 'fontStyle' and v == 'italic':
3774 unclosed_elements.append('i')
3775 elif k == 'textDecoration' and v == 'underline':
3777 unclosed_elements.append('u')
3779 self._out += '<font' + font + '>'
3780 unclosed_elements.append('font')
3782 if self._applied_styles:
3783 applied_style.update(self._applied_styles[-1])
3784 applied_style.update(style)
3785 self._applied_styles.append(applied_style)
3786 self._unclosed_elements.append(unclosed_elements)
3789 if tag not in (_x('ttml:br'), 'br'):
3790 unclosed_elements = self._unclosed_elements.pop()
3791 for element in reversed(unclosed_elements):
3792 self._out += '</%s>' % element
3793 if unclosed_elements and self._applied_styles:
3794 self._applied_styles.pop()
3796 def data(self, data):
3800 return self._out.strip()
3802 def parse_node(node):
3803 target = TTMLPElementParser()
3804 parser = xml.etree.ElementTree.XMLParser(target=target)
3805 parser.feed(xml.etree.ElementTree.tostring(node))
3806 return parser.close()
3808 for k, v in LEGACY_NAMESPACES:
3810 dfxp_data = dfxp_data.replace(ns, k)
3812 dfxp = compat_etree_fromstring(dfxp_data)
3814 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3817 raise ValueError('Invalid dfxp/TTML subtitle')
3821 for style in dfxp.findall(_x('.//ttml:style')):
3822 style_id = style.get('id') or style.get(_x('xml:id'))
3825 parent_style_id = style.get('style')
3827 if parent_style_id not in styles:
3830 styles[style_id] = styles[parent_style_id].copy()
3831 for prop in SUPPORTED_STYLING:
3832 prop_val = style.get(_x('tts:' + prop))
3834 styles.setdefault(style_id, {})[prop] = prop_val
3840 for p in ('body', 'div'):
3841 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3844 style = styles.get(ele.get('style'))
3847 default_style.update(style)
3849 for para, index in zip(paras, itertools.count(1)):
3850 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3851 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3852 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3853 if begin_time is None:
3858 end_time = begin_time + dur
3859 out.append('%d\n%s --> %s\n%s\n\n' % (
3861 srt_subtitles_timecode(begin_time),
3862 srt_subtitles_timecode(end_time),
3868 def cli_option(params, command_option, param, separator=None):
3869 param = params.get(param)
3870 return ([] if param is None
3871 else [command_option, str(param)] if separator is None
3872 else [f'{command_option}{separator}{param}'])
3875 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3876 param = params.get(param)
3877 assert param in (True, False, None)
3878 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
3881 def cli_valueless_option(params, command_option, param, expected_value=True):
3882 return [command_option] if params.get(param) == expected_value else []
3885 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3886 if isinstance(argdict, (list, tuple)): # for backward compatibility
3893 assert isinstance(argdict, dict)
3895 assert isinstance(keys, (list, tuple))
3896 for key_list in keys:
3897 arg_list = list(filter(
3898 lambda x: x is not None,
3899 [argdict.get(key.lower()) for key in variadic(key_list)]))
3901 return [arg for args in arg_list for arg in args]
3905 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3906 main_key, exe = main_key.lower(), exe.lower()
3907 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3908 keys = [f'{root_key}{k}' for k in (keys or [''])]
3909 if root_key in keys:
3911 keys.append((main_key, exe))
3912 keys.append('default')
3915 return cli_configuration_args(argdict, keys, default, use_compat)
3919 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
3978 'iw': 'heb', # Replaced by he in 1989 revision
3988 'in': 'ind', # Replaced by id in 1989 revision
4103 'ji': 'yid', # Replaced by yi in 1989 revision
4111 def short2long(cls, code):
4112 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4113 return cls._lang_map.get(code[:2])
4116 def long2short(cls, code):
4117 """Convert language code from ISO 639-2/T to ISO 639-1"""
4118 for short_name, long_name in cls._lang_map.items():
4119 if long_name == code:
4124 # From http://data.okfn.org/data/core/country-list
4126 'AF': 'Afghanistan',
4127 'AX': 'Åland Islands',
4130 'AS': 'American Samoa',
4135 'AG': 'Antigua and Barbuda',
4152 'BO': 'Bolivia, Plurinational State of',
4153 'BQ': 'Bonaire, Sint Eustatius and Saba',
4154 'BA': 'Bosnia and Herzegovina',
4156 'BV': 'Bouvet Island',
4158 'IO': 'British Indian Ocean Territory',
4159 'BN': 'Brunei Darussalam',
4161 'BF': 'Burkina Faso',
4167 'KY': 'Cayman Islands',
4168 'CF': 'Central African Republic',
4172 'CX': 'Christmas Island',
4173 'CC': 'Cocos (Keeling) Islands',
4177 'CD': 'Congo, the Democratic Republic of the',
4178 'CK': 'Cook Islands',
4180 'CI': 'Côte d\'Ivoire',
4185 'CZ': 'Czech Republic',
4189 'DO': 'Dominican Republic',
4192 'SV': 'El Salvador',
4193 'GQ': 'Equatorial Guinea',
4197 'FK': 'Falkland Islands (Malvinas)',
4198 'FO': 'Faroe Islands',
4202 'GF': 'French Guiana',
4203 'PF': 'French Polynesia',
4204 'TF': 'French Southern Territories',
4219 'GW': 'Guinea-Bissau',
4222 'HM': 'Heard Island and McDonald Islands',
4223 'VA': 'Holy See (Vatican City State)',
4230 'IR': 'Iran, Islamic Republic of',
4233 'IM': 'Isle of Man',
4243 'KP': 'Korea, Democratic People\'s Republic of',
4244 'KR': 'Korea, Republic of',
4247 'LA': 'Lao People\'s Democratic Republic',
4253 'LI': 'Liechtenstein',
4257 'MK': 'Macedonia, the Former Yugoslav Republic of',
4264 'MH': 'Marshall Islands',
4270 'FM': 'Micronesia, Federated States of',
4271 'MD': 'Moldova, Republic of',
4282 'NL': 'Netherlands',
4283 'NC': 'New Caledonia',
4284 'NZ': 'New Zealand',
4289 'NF': 'Norfolk Island',
4290 'MP': 'Northern Mariana Islands',
4295 'PS': 'Palestine, State of',
4297 'PG': 'Papua New Guinea',
4300 'PH': 'Philippines',
4304 'PR': 'Puerto Rico',
4308 'RU': 'Russian Federation',
4310 'BL': 'Saint Barthélemy',
4311 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4312 'KN': 'Saint Kitts and Nevis',
4313 'LC': 'Saint Lucia',
4314 'MF': 'Saint Martin (French part)',
4315 'PM': 'Saint Pierre and Miquelon',
4316 'VC': 'Saint Vincent and the Grenadines',
4319 'ST': 'Sao Tome and Principe',
4320 'SA': 'Saudi Arabia',
4324 'SL': 'Sierra Leone',
4326 'SX': 'Sint Maarten (Dutch part)',
4329 'SB': 'Solomon Islands',
4331 'ZA': 'South Africa',
4332 'GS': 'South Georgia and the South Sandwich Islands',
4333 'SS': 'South Sudan',
4338 'SJ': 'Svalbard and Jan Mayen',
4341 'CH': 'Switzerland',
4342 'SY': 'Syrian Arab Republic',
4343 'TW': 'Taiwan, Province of China',
4345 'TZ': 'Tanzania, United Republic of',
4347 'TL': 'Timor-Leste',
4351 'TT': 'Trinidad and Tobago',
4354 'TM': 'Turkmenistan',
4355 'TC': 'Turks and Caicos Islands',
4359 'AE': 'United Arab Emirates',
4360 'GB': 'United Kingdom',
4361 'US': 'United States',
4362 'UM': 'United States Minor Outlying Islands',
4366 'VE': 'Venezuela, Bolivarian Republic of',
4368 'VG': 'Virgin Islands, British',
4369 'VI': 'Virgin Islands, U.S.',
4370 'WF': 'Wallis and Futuna',
4371 'EH': 'Western Sahara',
4375 # Not ISO 3166 codes, but used for IP blocks
4376 'AP': 'Asia/Pacific Region',
4381 def short2full(cls, code):
4382 """Convert an ISO 3166-2 country code to the corresponding full name"""
4383 return cls._country_map.get(code.upper())
4387 # Major IPv4 address blocks per country
4389 'AD': '46.172.224.0/19',
4390 'AE': '94.200.0.0/13',
4391 'AF': '149.54.0.0/17',
4392 'AG': '209.59.64.0/18',
4393 'AI': '204.14.248.0/21',
4394 'AL': '46.99.0.0/16',
4395 'AM': '46.70.0.0/15',
4396 'AO': '105.168.0.0/13',
4397 'AP': '182.50.184.0/21',
4398 'AQ': '23.154.160.0/24',
4399 'AR': '181.0.0.0/12',
4400 'AS': '202.70.112.0/20',
4401 'AT': '77.116.0.0/14',
4402 'AU': '1.128.0.0/11',
4403 'AW': '181.41.0.0/18',
4404 'AX': '185.217.4.0/22',
4405 'AZ': '5.197.0.0/16',
4406 'BA': '31.176.128.0/17',
4407 'BB': '65.48.128.0/17',
4408 'BD': '114.130.0.0/16',
4410 'BF': '102.178.0.0/15',
4411 'BG': '95.42.0.0/15',
4412 'BH': '37.131.0.0/17',
4413 'BI': '154.117.192.0/18',
4414 'BJ': '137.255.0.0/16',
4415 'BL': '185.212.72.0/23',
4416 'BM': '196.12.64.0/18',
4417 'BN': '156.31.0.0/16',
4418 'BO': '161.56.0.0/16',
4419 'BQ': '161.0.80.0/20',
4420 'BR': '191.128.0.0/12',
4421 'BS': '24.51.64.0/18',
4422 'BT': '119.2.96.0/19',
4423 'BW': '168.167.0.0/16',
4424 'BY': '178.120.0.0/13',
4425 'BZ': '179.42.192.0/18',
4426 'CA': '99.224.0.0/11',
4427 'CD': '41.243.0.0/16',
4428 'CF': '197.242.176.0/21',
4429 'CG': '160.113.0.0/16',
4430 'CH': '85.0.0.0/13',
4431 'CI': '102.136.0.0/14',
4432 'CK': '202.65.32.0/19',
4433 'CL': '152.172.0.0/14',
4434 'CM': '102.244.0.0/14',
4435 'CN': '36.128.0.0/10',
4436 'CO': '181.240.0.0/12',
4437 'CR': '201.192.0.0/12',
4438 'CU': '152.206.0.0/15',
4439 'CV': '165.90.96.0/19',
4440 'CW': '190.88.128.0/17',
4441 'CY': '31.153.0.0/16',
4442 'CZ': '88.100.0.0/14',
4444 'DJ': '197.241.0.0/17',
4445 'DK': '87.48.0.0/12',
4446 'DM': '192.243.48.0/20',
4447 'DO': '152.166.0.0/15',
4448 'DZ': '41.96.0.0/12',
4449 'EC': '186.68.0.0/15',
4450 'EE': '90.190.0.0/15',
4451 'EG': '156.160.0.0/11',
4452 'ER': '196.200.96.0/20',
4453 'ES': '88.0.0.0/11',
4454 'ET': '196.188.0.0/14',
4455 'EU': '2.16.0.0/13',
4456 'FI': '91.152.0.0/13',
4457 'FJ': '144.120.0.0/16',
4458 'FK': '80.73.208.0/21',
4459 'FM': '119.252.112.0/20',
4460 'FO': '88.85.32.0/19',
4462 'GA': '41.158.0.0/15',
4464 'GD': '74.122.88.0/21',
4465 'GE': '31.146.0.0/16',
4466 'GF': '161.22.64.0/18',
4467 'GG': '62.68.160.0/19',
4468 'GH': '154.160.0.0/12',
4469 'GI': '95.164.0.0/16',
4470 'GL': '88.83.0.0/19',
4471 'GM': '160.182.0.0/15',
4472 'GN': '197.149.192.0/18',
4473 'GP': '104.250.0.0/19',
4474 'GQ': '105.235.224.0/20',
4475 'GR': '94.64.0.0/13',
4476 'GT': '168.234.0.0/16',
4477 'GU': '168.123.0.0/16',
4478 'GW': '197.214.80.0/20',
4479 'GY': '181.41.64.0/18',
4480 'HK': '113.252.0.0/14',
4481 'HN': '181.210.0.0/16',
4482 'HR': '93.136.0.0/13',
4483 'HT': '148.102.128.0/17',
4484 'HU': '84.0.0.0/14',
4485 'ID': '39.192.0.0/10',
4486 'IE': '87.32.0.0/12',
4487 'IL': '79.176.0.0/13',
4488 'IM': '5.62.80.0/20',
4489 'IN': '117.192.0.0/10',
4490 'IO': '203.83.48.0/21',
4491 'IQ': '37.236.0.0/14',
4492 'IR': '2.176.0.0/12',
4493 'IS': '82.221.0.0/16',
4494 'IT': '79.0.0.0/10',
4495 'JE': '87.244.64.0/18',
4496 'JM': '72.27.0.0/17',
4497 'JO': '176.29.0.0/16',
4498 'JP': '133.0.0.0/8',
4499 'KE': '105.48.0.0/12',
4500 'KG': '158.181.128.0/17',
4501 'KH': '36.37.128.0/17',
4502 'KI': '103.25.140.0/22',
4503 'KM': '197.255.224.0/20',
4504 'KN': '198.167.192.0/19',
4505 'KP': '175.45.176.0/22',
4506 'KR': '175.192.0.0/10',
4507 'KW': '37.36.0.0/14',
4508 'KY': '64.96.0.0/15',
4509 'KZ': '2.72.0.0/13',
4510 'LA': '115.84.64.0/18',
4511 'LB': '178.135.0.0/16',
4512 'LC': '24.92.144.0/20',
4513 'LI': '82.117.0.0/19',
4514 'LK': '112.134.0.0/15',
4515 'LR': '102.183.0.0/16',
4516 'LS': '129.232.0.0/17',
4517 'LT': '78.56.0.0/13',
4518 'LU': '188.42.0.0/16',
4519 'LV': '46.109.0.0/16',
4520 'LY': '41.252.0.0/14',
4521 'MA': '105.128.0.0/11',
4522 'MC': '88.209.64.0/18',
4523 'MD': '37.246.0.0/16',
4524 'ME': '178.175.0.0/17',
4525 'MF': '74.112.232.0/21',
4526 'MG': '154.126.0.0/17',
4527 'MH': '117.103.88.0/21',
4528 'MK': '77.28.0.0/15',
4529 'ML': '154.118.128.0/18',
4530 'MM': '37.111.0.0/17',
4531 'MN': '49.0.128.0/17',
4532 'MO': '60.246.0.0/16',
4533 'MP': '202.88.64.0/20',
4534 'MQ': '109.203.224.0/19',
4535 'MR': '41.188.64.0/18',
4536 'MS': '208.90.112.0/22',
4537 'MT': '46.11.0.0/16',
4538 'MU': '105.16.0.0/12',
4539 'MV': '27.114.128.0/18',
4540 'MW': '102.70.0.0/15',
4541 'MX': '187.192.0.0/11',
4542 'MY': '175.136.0.0/13',
4543 'MZ': '197.218.0.0/15',
4544 'NA': '41.182.0.0/16',
4545 'NC': '101.101.0.0/18',
4546 'NE': '197.214.0.0/18',
4547 'NF': '203.17.240.0/22',
4548 'NG': '105.112.0.0/12',
4549 'NI': '186.76.0.0/15',
4550 'NL': '145.96.0.0/11',
4551 'NO': '84.208.0.0/13',
4552 'NP': '36.252.0.0/15',
4553 'NR': '203.98.224.0/19',
4554 'NU': '49.156.48.0/22',
4555 'NZ': '49.224.0.0/14',
4556 'OM': '5.36.0.0/15',
4557 'PA': '186.72.0.0/15',
4558 'PE': '186.160.0.0/14',
4559 'PF': '123.50.64.0/18',
4560 'PG': '124.240.192.0/19',
4561 'PH': '49.144.0.0/13',
4562 'PK': '39.32.0.0/11',
4563 'PL': '83.0.0.0/11',
4564 'PM': '70.36.0.0/20',
4565 'PR': '66.50.0.0/16',
4566 'PS': '188.161.0.0/16',
4567 'PT': '85.240.0.0/13',
4568 'PW': '202.124.224.0/20',
4569 'PY': '181.120.0.0/14',
4570 'QA': '37.210.0.0/15',
4571 'RE': '102.35.0.0/16',
4572 'RO': '79.112.0.0/13',
4573 'RS': '93.86.0.0/15',
4574 'RU': '5.136.0.0/13',
4575 'RW': '41.186.0.0/16',
4576 'SA': '188.48.0.0/13',
4577 'SB': '202.1.160.0/19',
4578 'SC': '154.192.0.0/11',
4579 'SD': '102.120.0.0/13',
4580 'SE': '78.64.0.0/12',
4581 'SG': '8.128.0.0/10',
4582 'SI': '188.196.0.0/14',
4583 'SK': '78.98.0.0/15',
4584 'SL': '102.143.0.0/17',
4585 'SM': '89.186.32.0/19',
4586 'SN': '41.82.0.0/15',
4587 'SO': '154.115.192.0/18',
4588 'SR': '186.179.128.0/17',
4589 'SS': '105.235.208.0/21',
4590 'ST': '197.159.160.0/19',
4591 'SV': '168.243.0.0/16',
4592 'SX': '190.102.0.0/20',
4594 'SZ': '41.84.224.0/19',
4595 'TC': '65.255.48.0/20',
4596 'TD': '154.68.128.0/19',
4597 'TG': '196.168.0.0/14',
4598 'TH': '171.96.0.0/13',
4599 'TJ': '85.9.128.0/18',
4600 'TK': '27.96.24.0/21',
4601 'TL': '180.189.160.0/20',
4602 'TM': '95.85.96.0/19',
4603 'TN': '197.0.0.0/11',
4604 'TO': '175.176.144.0/21',
4605 'TR': '78.160.0.0/11',
4606 'TT': '186.44.0.0/15',
4607 'TV': '202.2.96.0/19',
4608 'TW': '120.96.0.0/11',
4609 'TZ': '156.156.0.0/14',
4610 'UA': '37.52.0.0/14',
4611 'UG': '102.80.0.0/13',
4613 'UY': '167.56.0.0/13',
4614 'UZ': '84.54.64.0/18',
4615 'VA': '212.77.0.0/19',
4616 'VC': '207.191.240.0/21',
4617 'VE': '186.88.0.0/13',
4618 'VG': '66.81.192.0/20',
4619 'VI': '146.226.0.0/16',
4620 'VN': '14.160.0.0/11',
4621 'VU': '202.80.32.0/20',
4622 'WF': '117.20.32.0/21',
4623 'WS': '202.4.32.0/19',
4624 'YE': '134.35.0.0/16',
4625 'YT': '41.242.116.0/22',
4626 'ZA': '41.0.0.0/11',
4627 'ZM': '102.144.0.0/13',
4628 'ZW': '102.177.192.0/18',
4632 def random_ipv4(cls, code_or_block):
4633 if len(code_or_block) == 2:
4634 block = cls._country_ip_map.get(code_or_block.upper())
4638 block = code_or_block
4639 addr, preflen = block.split('/')
4640 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4641 addr_max = addr_min | (0xffffffff >> int(preflen))
4642 return str(socket.inet_ntoa(
4643 struct.pack('!L', random.randint(addr_min, addr_max))))
4646 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4647 def __init__(self, proxies=None):
4648 # Set default handlers
4649 for type in ('http', 'https'):
4650 setattr(self, '%s_open' % type,
4651 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4652 meth(r, proxy, type))
4653 urllib.request.ProxyHandler.__init__(self, proxies)
4655 def proxy_open(self, req, proxy, type):
4656 req_proxy = req.headers.get('Ytdl-request-proxy')
4657 if req_proxy is not None:
4659 del req.headers['Ytdl-request-proxy']
4661 if proxy == '__noproxy__':
4662 return None # No Proxy
4663 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4664 req.add_header('Ytdl-socks-proxy', proxy)
4665 # yt-dlp's http/https handlers do wrapping the socket with socks
4667 return urllib.request.ProxyHandler.proxy_open(
4668 self, req, proxy, type)
4671 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4672 # released into Public Domain
4673 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4675 def long_to_bytes(n, blocksize=0):
4676 """long_to_bytes(n:long, blocksize:int) : string
4677 Convert a long integer to a byte string.
4679 If optional blocksize is given and greater than zero, pad the front of the
4680 byte string with binary zeros so that the length is a multiple of
4683 # after much testing, this algorithm was deemed to be the fastest
4687 s = struct.pack('>I', n & 0xffffffff) + s
4689 # strip off leading zeros
4690 for i in range(len(s)):
4691 if s[i] != b'\000'[0]:
4694 # only happens when n == 0
4698 # add back some pad bytes. this could be done more efficiently w.r.t. the
4699 # de-padding being done above, but sigh...
4700 if blocksize > 0 and len(s) % blocksize:
4701 s = (blocksize - len(s) % blocksize) * b'\000' + s
4705 def bytes_to_long(s):
4706 """bytes_to_long(string) : long
4707 Convert a byte string to a long integer.
4709 This is (essentially) the inverse of long_to_bytes().
4714 extra = (4 - length % 4)
4715 s = b'\000' * extra + s
4716 length = length + extra
4717 for i in range(0, length, 4):
4718 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4722 def ohdave_rsa_encrypt(data, exponent, modulus):
4724 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4727 data: data to encrypt, bytes-like object
4728 exponent, modulus: parameter e and N of RSA algorithm, both integer
4729 Output: hex string of encrypted data
4731 Limitation: supports one block encryption only
4734 payload = int(binascii.hexlify(data[::-1]), 16)
4735 encrypted = pow(payload, exponent, modulus)
4736 return '%x' % encrypted
4739 def pkcs1pad(data, length):
4741 Padding input data with PKCS#1 scheme
4743 @param {int[]} data input data
4744 @param {int} length target length
4745 @returns {int[]} padded data
4747 if len(data) > length - 11:
4748 raise ValueError('Input data too
long for PKCS
#1 padding')
4750 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4751 return [0, 2] + pseudo_random
+ [0] + data
4754 def _base_n_table(n
, table
):
4755 if not table
and not n
:
4756 raise ValueError('Either table or n must be specified')
4757 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4759 if n
and n
!= len(table
):
4760 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4764 def encode_base_n(num
, n
=None, table
=None):
4765 """Convert given int to a base-n string"""
4766 table
= _base_n_table(n
, table
)
4770 result
, base
= '', len(table
)
4772 result
= table
[num
% base
] + result
4777 def decode_base_n(string
, n
=None, table
=None):
4778 """Convert given base-n string to int"""
4779 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4780 result
, base
= 0, len(table
)
4782 result
= result
* base
+ table
[char
]
4786 def decode_base(value
, digits
):
4787 write_string('DeprecationWarning: yt_dlp.utils.decode_base is deprecated '
4788 'and may be removed in a future version. Use yt_dlp.decode_base_n instead')
4789 return decode_base_n(value
, table
=digits
)
4792 def decode_packed_codes(code
):
4793 mobj
= re
.search(PACKED_CODES_RE
, code
)
4794 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4797 symbols
= symbols
.split('|')
4802 base_n_count
= encode_base_n(count
, base
)
4803 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4806 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4810 def caesar(s
, alphabet
, shift
):
4815 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4820 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4823 def parse_m3u8_attributes(attrib
):
4825 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4826 if val
.startswith('"'):
4832 def urshift(val
, n
):
4833 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4836 # Based on png2str() written by @gdkchan and improved by @yokrysty
4837 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4838 def decode_png(png_data
):
4839 # Reference: https://www.w3.org/TR/PNG/
4840 header
= png_data
[8:]
4842 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4843 raise OSError('Not a valid PNG file.')
4845 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4846 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
4851 length
= unpack_integer(header
[:4])
4854 chunk_type
= header
[:4]
4857 chunk_data
= header
[:length
]
4858 header
= header
[length
:]
4860 header
= header
[4:] # Skip CRC
4868 ihdr
= chunks
[0]['data']
4870 width
= unpack_integer(ihdr
[:4])
4871 height
= unpack_integer(ihdr
[4:8])
4875 for chunk
in chunks
:
4876 if chunk
['type'] == b
'IDAT':
4877 idat
+= chunk
['data']
4880 raise OSError('Unable to read PNG data.')
4882 decompressed_data
= bytearray(zlib
.decompress(idat
))
4887 def _get_pixel(idx
):
4892 for y
in range(height
):
4893 basePos
= y
* (1 + stride
)
4894 filter_type
= decompressed_data
[basePos
]
4898 pixels
.append(current_row
)
4900 for x
in range(stride
):
4901 color
= decompressed_data
[1 + basePos
+ x
]
4902 basex
= y
* stride
+ x
4907 left
= _get_pixel(basex
- 3)
4909 up
= _get_pixel(basex
- stride
)
4911 if filter_type
== 1: # Sub
4912 color
= (color
+ left
) & 0xff
4913 elif filter_type
== 2: # Up
4914 color
= (color
+ up
) & 0xff
4915 elif filter_type
== 3: # Average
4916 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4917 elif filter_type
== 4: # Paeth
4923 c
= _get_pixel(basex
- stride
- 3)
4931 if pa
<= pb
and pa
<= pc
:
4932 color
= (color
+ a
) & 0xff
4934 color
= (color
+ b
) & 0xff
4936 color
= (color
+ c
) & 0xff
4938 current_row
.append(color
)
4940 return width
, height
, pixels
4943 def write_xattr(path
, key
, value
):
4944 # Windows: Write xattrs to NTFS Alternate Data Streams:
4945 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4946 if compat_os_name
== 'nt':
4947 assert ':' not in key
4948 assert os
.path
.exists(path
)
4951 with open(f
'{path}:{key}', 'wb') as f
:
4953 except OSError as e
:
4954 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4957 # UNIX Method 1. Use xattrs/pyxattrs modules
4960 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
4961 # Unicode arguments are not supported in pyxattr until version 0.5.0
4962 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4963 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
4964 setxattr
= xattr
.set
4966 setxattr
= xattr
.setxattr
4970 setxattr(path
, key
, value
)
4971 except OSError as e
:
4972 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4975 # UNIX Method 2. Use setfattr/xattr executables
4976 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
4977 else 'xattr' if check_executable('xattr', ['-h']) else None)
4979 raise XAttrUnavailableError(
4980 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
4981 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
4983 value
= value
.decode()
4985 _
, stderr
, returncode
= Popen
.run(
4986 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
4987 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4988 except OSError as e
:
4989 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4991 raise XAttrMetadataError(returncode
, stderr
)
4994 def random_birthday(year_field
, month_field
, day_field
):
4995 start_date
= datetime
.date(1950, 1, 1)
4996 end_date
= datetime
.date(1995, 12, 31)
4997 offset
= random
.randint(0, (end_date
- start_date
).days
)
4998 random_date
= start_date
+ datetime
.timedelta(offset
)
5000 year_field
: str(random_date
.year
),
5001 month_field
: str(random_date
.month
),
5002 day_field
: str(random_date
.day
),
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
= traverse_obj(obj
, *variadic(field
))
5093 if (not val
and val
!= 0) 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
)
5125 if dn
and not os
.path
.exists(dn
):
5128 except OSError as err
:
5129 if callable(to_screen
) is not None:
5130 to_screen('unable to create directory ' + error_to_compat_str(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 load_plugins(name
, suffix
, namespace
):
5142 with contextlib
.suppress(FileNotFoundError
):
5143 plugins_spec
= importlib
.util
.spec_from_file_location(
5144 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5145 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5146 sys
.modules
[plugins_spec
.name
] = plugins
5147 plugins_spec
.loader
.exec_module(plugins
)
5148 for name
in dir(plugins
):
5149 if name
in namespace
:
5151 if not name
.endswith(suffix
):
5153 klass
= getattr(plugins
, name
)
5154 classes
[name
] = namespace
[name
] = klass
5159 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
5160 casesense
=True, is_user_input
=False, traverse_string
=False):
5161 ''' Traverse nested list/dict/tuple
5162 @param path_list A list of paths which are checked one by one.
5163 Each path is a list of keys where each key is a:
5165 - string: A dictionary key
5166 - int: An index into a list
5167 - tuple: A list of keys all of which will be traversed
5168 - Ellipsis: Fetch all values in the object
5169 - Function: Takes the key and value as arguments
5170 and returns whether the key matches or not
5171 @param default Default value to return
5172 @param expected_type Only accept final value of this type (Can also be any callable)
5173 @param get_all Return all the values obtained from a path or only the first one
5174 @param casesense Whether to consider dictionary keys as case sensitive
5175 @param is_user_input Whether the keys are generated from user input. If True,
5176 strings are converted to int/slice if necessary
5177 @param traverse_string Whether to traverse inside strings. If True, any
5178 non-compatible object will also be converted into a string
5182 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
5183 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
5185 def _traverse_obj(obj
, path
, _current_depth
=0):
5187 path
= tuple(variadic(path
))
5188 for i
, key
in enumerate(path
):
5189 if None in (key
, obj
):
5191 if isinstance(key
, (list, tuple)):
5192 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
5195 obj
= (obj
.values() if isinstance(obj
, dict)
5196 else obj
if isinstance(obj
, (list, tuple, LazyList
))
5197 else str(obj
) if traverse_string
else [])
5199 depth
= max(depth
, _current_depth
)
5200 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
5202 if isinstance(obj
, (list, tuple, LazyList
)):
5203 obj
= enumerate(obj
)
5204 elif isinstance(obj
, dict):
5207 if not traverse_string
:
5211 depth
= max(depth
, _current_depth
)
5212 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if try_call(key
, args
=(k
, v
))]
5213 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
5214 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
5215 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
5218 key
= (int_or_none(key
) if ':' not in key
5219 else slice(*map(int_or_none
, key
.split(':'))))
5220 if key
== slice(None):
5221 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
5222 if not isinstance(key
, (int, slice)):
5224 if not isinstance(obj
, (list, tuple, LazyList
)):
5225 if not traverse_string
:
5234 if isinstance(expected_type
, type):
5235 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5237 type_test
= expected_type
or IDENTITY
5239 for path
in path_list
:
5241 val
= _traverse_obj(obj
, path
)
5244 for _
in range(depth
- 1):
5245 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5246 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5248 return val
if get_all
else val
[0]
5250 val
= type_test(val
)
5256 def traverse_dict(dictn
, keys
, casesense
=True):
5257 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5258 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5259 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5262 def get_first(obj
, keys
, **kwargs
):
5263 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5266 def variadic(x
, allowed_types
=(str, bytes, dict)):
5267 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5270 def time_seconds(**kwargs
):
5271 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5272 return t
.timestamp()
5275 # create a JSON Web Signature (jws) with HS256 algorithm
5276 # the resulting format is in JWS Compact Serialization
5277 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5278 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5279 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5285 header_data
.update(headers
)
5286 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5287 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5288 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5289 signature_b64
= base64
.b64encode(h
.digest())
5290 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5294 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5295 def jwt_decode_hs256(jwt
):
5296 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5297 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5301 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5305 def supports_terminal_sequences(stream
):
5306 if compat_os_name
== 'nt':
5307 if not WINDOWS_VT_MODE
:
5309 elif not os
.getenv('TERM'):
5312 return stream
.isatty()
5313 except BaseException
:
5317 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5318 if get_windows_version() < (10, 0, 10586):
5320 global WINDOWS_VT_MODE
5322 Popen
.run('', shell
=True)
5326 WINDOWS_VT_MODE
= True
5327 supports_terminal_sequences
.cache_clear()
5330 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5333 def remove_terminal_sequences(string
):
5334 return _terminal_sequences_re
.sub('', string
)
5337 def number_of_digits(number
):
5338 return len('%d' % number
)
5341 def join_nonempty(*values
, delim
='-', from_dict
=None):
5342 if from_dict
is not None:
5343 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5344 return delim
.join(map(str, filter(None, values
)))
5347 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5349 Find the largest format dimensions in terms of video width and, for each thumbnail:
5350 * Modify the URL: Match the width with the provided regex and replace with the former width
5353 This function is useful with video services that scale the provided thumbnails on demand
5355 _keys
= ('width', 'height')
5356 max_dimensions
= max(
5357 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5359 if not max_dimensions
[0]:
5363 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5364 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5365 for thumbnail
in thumbnails
5369 def parse_http_range(range):
5370 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5372 return None, None, None
5373 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5375 return None, None, None
5376 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5379 def read_stdin(what
):
5380 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5381 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5389 __initialized
= False
5391 def __init__(self
, parser
, label
=None):
5392 self
.parser
, self
.label
= parser
, label
5393 self
._loaded
_paths
, self
.configs
= set(), []
5395 def init(self
, args
=None, filename
=None):
5396 assert not self
.__initialized
5397 self
.own_args
, self
.filename
= args
, filename
5398 return self
.load_configs()
5400 def load_configs(self
):
5403 location
= os
.path
.realpath(self
.filename
)
5404 directory
= os
.path
.dirname(location
)
5405 if location
in self
._loaded
_paths
:
5407 self
._loaded
_paths
.add(location
)
5409 self
.__initialized
= True
5410 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5411 self
.parsed_args
= self
.own_args
5412 for location
in opts
.config_locations
or []:
5414 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5416 location
= os
.path
.join(directory
, expand_path(location
))
5417 if os
.path
.isdir(location
):
5418 location
= os
.path
.join(location
, 'yt-dlp.conf')
5419 if not os
.path
.exists(location
):
5420 self
.parser
.error(f
'config location {location} does not exist')
5421 self
.append_config(self
.read_file(location
), location
)
5425 label
= join_nonempty(
5426 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5428 return join_nonempty(
5429 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5430 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5434 def read_file(filename
, default
=[]):
5436 optionf
= open(filename
)
5438 return default
# silently skip if file is not present
5440 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5441 contents
= optionf
.read()
5442 res
= shlex
.split(contents
, comments
=True)
5443 except Exception as err
:
5444 raise ValueError(f
'Unable to parse "{filename}": {err}')
5450 def hide_login_info(opts
):
5451 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5452 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5457 return m
.group('key') + '=PRIVATE'
5461 opts
= list(map(_scrub_eq
, opts
))
5462 for idx
, opt
in enumerate(opts
):
5463 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5464 opts
[idx
+ 1] = 'PRIVATE'
5467 def append_config(self
, *args
, label
=None):
5468 config
= type(self
)(self
.parser
, label
)
5469 config
._loaded
_paths
= self
._loaded
_paths
5470 if config
.init(*args
):
5471 self
.configs
.append(config
)
5475 for config
in reversed(self
.configs
):
5476 yield from config
.all_args
5477 yield from self
.parsed_args
or []
5479 def parse_known_args(self
, **kwargs
):
5480 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5482 def parse_args(self
):
5483 return self
.parser
.parse_args(self
.all_args
)
5486 class WebSocketsWrapper():
5487 """Wraps websockets module to use in non-async scopes"""
5490 def __init__(self
, url
, headers
=None, connect
=True):
5491 self
.loop
= asyncio
.new_event_loop()
5492 # XXX: "loop" is deprecated
5493 self
.conn
= websockets
.connect(
5494 url
, extra_headers
=headers
, ping_interval
=None,
5495 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5498 atexit
.register(self
.__exit
__, None, None, None)
5500 def __enter__(self
):
5502 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5505 def send(self
, *args
):
5506 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5508 def recv(self
, *args
):
5509 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5511 def __exit__(self
, type, value
, traceback
):
5513 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5516 self
._cancel
_all
_tasks
(self
.loop
)
5518 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5519 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5521 def run_with_loop(main
, loop
):
5522 if not asyncio
.iscoroutine(main
):
5523 raise ValueError(f
'a coroutine was expected, got {main!r}')
5526 return loop
.run_until_complete(main
)
5528 loop
.run_until_complete(loop
.shutdown_asyncgens())
5529 if hasattr(loop
, 'shutdown_default_executor'):
5530 loop
.run_until_complete(loop
.shutdown_default_executor())
5533 def _cancel_all_tasks(loop
):
5534 to_cancel
= asyncio
.all_tasks(loop
)
5539 for task
in to_cancel
:
5542 # XXX: "loop" is removed in python 3.10+
5543 loop
.run_until_complete(
5544 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5546 for task
in to_cancel
:
5547 if task
.cancelled():
5549 if task
.exception() is not None:
5550 loop
.call_exception_handler({
5551 'message': 'unhandled exception during asyncio.run() shutdown',
5552 'exception': task
.exception(),
5557 def merge_headers(*dicts
):
5558 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5559 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5562 def cached_method(f
):
5563 """Cache a method"""
5564 signature
= inspect
.signature(f
)
5567 def wrapper(self
, *args
, **kwargs
):
5568 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5569 bound_args
.apply_defaults()
5570 key
= tuple(bound_args
.arguments
.values())
5572 if not hasattr(self
, '__cached_method__cache'):
5573 self
.__cached
_method
__cache
= {}
5574 cache
= self
.__cached
_method
__cache
.setdefault(f
.__name
__, {})
5575 if key
not in cache
:
5576 cache
[key
] = f(self
, *args
, **kwargs
)
5581 class classproperty
:
5582 """property access for class methods"""
5584 def __init__(self
, func
):
5585 functools
.update_wrapper(self
, func
)
5588 def __get__(self
, _
, cls
):
5589 return self
.func(cls
)
5592 class Namespace(types
.SimpleNamespace
):
5593 """Immutable namespace"""
5596 return iter(self
.__dict
__.values())
5600 return self
.__dict
__.items()
5604 has_certifi
= bool(certifi
)
5605 has_websockets
= bool(websockets
)