48 import xml
.etree
.ElementTree
51 from .compat
import functools
# isort: split
53 compat_etree_fromstring
,
55 compat_HTMLParseError
,
59 from .dependencies
import brotli
, certifi
, websockets
, xattr
60 from .socks
import ProxyType
, sockssocket
63 def register_socks_protocols():
64 # "Register" SOCKS protocols
65 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
66 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
67 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
68 if scheme
not in urllib
.parse
.uses_netloc
:
69 urllib
.parse
.uses_netloc
.append(scheme
)
72 # This is not clearly defined otherwise
73 compiled_regex_type
= type(re
.compile(''))
76 def random_user_agent():
77 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
118 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
121 SUPPORTED_ENCODINGS
= [
125 SUPPORTED_ENCODINGS
.append('br')
128 'User-Agent': random_user_agent(),
129 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
130 'Accept-Language': 'en-us,en;q=0.5',
131 'Sec-Fetch-Mode': 'navigate',
136 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
140 NO_DEFAULT
= object()
141 IDENTITY
= lambda x
: x
143 ENGLISH_MONTH_NAMES
= [
144 'January', 'February', 'March', 'April', 'May', 'June',
145 'July', 'August', 'September', 'October', 'November', 'December']
148 'en': ENGLISH_MONTH_NAMES
,
150 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
151 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
152 # these follow the genitive grammatical case (dopełniacz)
153 # some websites might be using nominative, which will require another month list
154 # https://en.wikibooks.org/wiki/Polish/Noun_cases
155 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
156 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
159 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
161 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
162 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
163 'EST': -5, 'EDT': -4, # Eastern
164 'CST': -6, 'CDT': -5, # Central
165 'MST': -7, 'MDT': -6, # Mountain
166 'PST': -8, 'PDT': -7 # Pacific
169 # needed for sanitizing filenames in restricted mode
170 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
171 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
172 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
202 '%Y-%m-%d %H:%M:%S.%f',
203 '%Y-%m-%d %H:%M:%S:%f',
206 '%Y-%m-%dT%H:%M:%SZ',
207 '%Y-%m-%dT%H:%M:%S.%fZ',
208 '%Y-%m-%dT%H:%M:%S.%f0Z',
210 '%Y-%m-%dT%H:%M:%S.%f',
213 '%b %d %Y at %H:%M:%S',
215 '%B %d %Y at %H:%M:%S',
219 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
220 DATE_FORMATS_DAY_FIRST
.extend([
230 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
231 DATE_FORMATS_MONTH_FIRST
.extend([
239 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
240 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
242 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
246 def preferredencoding():
247 """Get preferred encoding.
249 Returns the best encoding scheme for the system, based on
250 locale.getpreferredencoding() and some further tweaks.
253 pref = locale.getpreferredencoding()
261 def write_json_file(obj, fn):
262 """ Encode obj as JSON and write it to fn, atomically if possible """
264 tf = tempfile.NamedTemporaryFile(
265 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
266 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
270 json.dump(obj, tf, ensure_ascii=False)
271 if sys.platform == 'win32
':
272 # Need to remove existing file on Windows, else os.rename raises
273 # WindowsError or FileExistsError.
274 with contextlib.suppress(OSError):
276 with contextlib.suppress(OSError):
279 os.chmod(tf.name, 0o666 & ~mask)
280 os.rename(tf.name, fn)
282 with contextlib.suppress(OSError):
287 def find_xpath_attr(node, xpath, key, val=None):
288 """ Find the xpath xpath[@key=val] """
289 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
290 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
291 return node.find(expr)
293 # On python2.6 the xml.etree.ElementTree.Element methods don't support
294 # the namespace parameter
297 def xpath_with_ns(path
, ns_map
):
298 components
= [c
.split(':') for c
in path
.split('/')]
302 replaced
.append(c
[0])
305 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
306 return '/'.join(replaced
)
309 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
310 def _find_xpath(xpath
):
311 return node
.find(xpath
)
313 if isinstance(xpath
, str):
314 n
= _find_xpath(xpath
)
322 if default
is not NO_DEFAULT
:
325 name
= xpath
if name
is None else name
326 raise ExtractorError('Could not find XML element %s' % name
)
332 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
333 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
334 if n
is None or n
== default
:
337 if default
is not NO_DEFAULT
:
340 name
= xpath
if name
is None else name
341 raise ExtractorError('Could not find XML element\'s text %s' % name
)
347 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
348 n
= find_xpath_attr(node
, xpath
, key
)
350 if default
is not NO_DEFAULT
:
353 name
= f
'{xpath}[@{key}]' if name
is None else name
354 raise ExtractorError('Could not find XML attribute %s' % name
)
360 def get_element_by_id(id, html
, **kwargs
):
361 """Return the content of the tag with the specified ID in the passed HTML document"""
362 return get_element_by_attribute('id', id, html
, **kwargs
)
365 def get_element_html_by_id(id, html
, **kwargs
):
366 """Return the html of the tag with the specified ID in the passed HTML document"""
367 return get_element_html_by_attribute('id', id, html
, **kwargs
)
370 def get_element_by_class(class_name
, html
):
371 """Return the content of the first tag with the specified class in the passed HTML document"""
372 retval
= get_elements_by_class(class_name
, html
)
373 return retval
[0] if retval
else None
376 def get_element_html_by_class(class_name
, html
):
377 """Return the html of the first tag with the specified class in the passed HTML document"""
378 retval
= get_elements_html_by_class(class_name
, html
)
379 return retval
[0] if retval
else None
382 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
383 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
384 return retval
[0] if retval
else None
387 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
388 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
389 return retval
[0] if retval
else None
392 def get_elements_by_class(class_name
, html
, **kargs
):
393 """Return the content of all tags with the specified class in the passed HTML document as a list"""
394 return get_elements_by_attribute(
395 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
396 html, escape_value=False)
399 def get_elements_html_by_class(class_name, html):
400 """Return the html of all tags with the specified class in the passed HTML document as a list"""
401 return get_elements_html_by_attribute(
402 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
403 html, escape_value=False)
406 def get_elements_by_attribute(*args, **kwargs):
407 """Return the content of the tag with the specified attribute in the passed HTML document"""
408 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
411 def get_elements_html_by_attribute(*args, **kwargs):
412 """Return the html of the tag with the specified attribute in the passed HTML document"""
413 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
416 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
418 Return the text (content) and the html (whole) of the tag with the specified
419 attribute in the passed HTML document
424 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
426 value = re.escape(value) if escape_value else value
428 partial_element_re = rf'''(?x
)
430 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
431 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
434 for m in re.finditer(partial_element_re, html):
435 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
438 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
443 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
445 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
446 closing tag for the first opening tag it has encountered, and can be used
450 class HTMLBreakOnClosingTagException(Exception):
454 self.tagstack = collections.deque()
455 html.parser.HTMLParser.__init__(self)
460 def __exit__(self, *_):
464 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
465 # so data remains buffered; we no longer have any interest in it, thus
466 # override this method to discard it
469 def handle_starttag(self, tag, _):
470 self.tagstack.append(tag)
472 def handle_endtag(self, tag):
473 if not self.tagstack:
474 raise compat_HTMLParseError('no tags
in the stack
')
476 inner_tag = self.tagstack.pop()
480 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
481 if not self.tagstack:
482 raise self.HTMLBreakOnClosingTagException()
485 # XXX: This should be far less strict
486 def get_element_text_and_html_by_tag(tag, html):
488 For the first element with the specified tag in the passed HTML document
489 return its' content (text
) and the whole
element (html
)
491 def find_or_raise(haystack, needle, exc):
493 return haystack.index(needle)
496 closing_tag = f'</{tag}>'
497 whole_start = find_or_raise(
498 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
499 content_start = find_or_raise(
500 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
501 content_start += whole_start + 1
502 with HTMLBreakOnClosingTagParser() as parser:
503 parser.feed(html[whole_start:content_start])
504 if not parser.tagstack or parser.tagstack[0] != tag:
505 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
506 offset = content_start
507 while offset < len(html):
508 next_closing_tag_start = find_or_raise(
509 html[offset:], closing_tag,
510 compat_HTMLParseError(f'closing {tag} tag not found'))
511 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
513 parser.feed(html[offset:offset + next_closing_tag_end])
514 offset += next_closing_tag_end
515 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
516 return html[content_start:offset + next_closing_tag_start], \
517 html[whole_start:offset + next_closing_tag_end]
518 raise compat_HTMLParseError('unexpected end of html')
521 class HTMLAttributeParser(html.parser.HTMLParser):
522 """Trivial HTML parser to gather the attributes
for a single element
"""
526 html.parser.HTMLParser.__init__(self)
528 def handle_starttag(self, tag, attrs):
529 self.attrs = dict(attrs)
530 raise compat_HTMLParseError('done')
533 class HTMLListAttrsParser(html.parser.HTMLParser):
534 """HTML parser to gather the attributes
for the elements of a
list"""
537 html.parser.HTMLParser.__init__(self)
541 def handle_starttag(self, tag, attrs):
542 if tag == 'li' and self._level == 0:
543 self.items.append(dict(attrs))
546 def handle_endtag(self, tag):
550 def extract_attributes(html_element):
551 """Given a string
for an HTML element such
as
553 a
="foo" B
="bar" c
="&98;az" d
=boz
554 empty
= noval entity
="&"
557 Decode
and return a dictionary of attributes
.
559 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
560 'empty': '', 'noval': None, 'entity': '&',
561 'sq': '"', 'dq': '\''
564 parser = HTMLAttributeParser()
565 with contextlib.suppress(compat_HTMLParseError):
566 parser.feed(html_element)
571 def parse_list(webpage):
572 """Given a string
for an series of HTML
<li
> elements
,
573 return a dictionary of their attributes
"""
574 parser = HTMLListAttrsParser()
580 def clean_html(html):
581 """Clean an HTML snippet into a readable string
"""
583 if html is None: # Convenience for sanitizing descriptions etc.
586 html = re.sub(r'\s+', ' ', html)
587 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
588 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
590 html = re.sub('<.*?>', '', html)
591 # Replace html entities
592 html = unescapeHTML(html)
596 class LenientJSONDecoder(json.JSONDecoder):
597 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
598 self.transform_source, self.ignore_extra = transform_source, ignore_extra
599 super().__init__(*args, **kwargs)
602 if self.transform_source:
603 s = self.transform_source(s)
605 if self.ignore_extra:
606 return self.raw_decode(s.lstrip())[0]
607 return super().decode(s)
608 except json.JSONDecodeError as e:
609 if e.pos is not None:
610 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
614 def sanitize_open(filename, open_mode):
615 """Try to
open the given filename
, and slightly tweak it
if this fails
.
617 Attempts to
open the given filename
. If this fails
, it tries to change
618 the filename slightly
, step by step
, until it
's either able to open it
619 or it fails and raises a final exception, like the standard open()
622 It returns the tuple (stream, definitive_file_name).
625 if sys.platform == 'win32
':
628 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
629 with contextlib.suppress(io.UnsupportedOperation):
630 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
631 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
633 for attempt in range(2):
636 if sys.platform == 'win32
':
637 # FIXME: An exclusive lock also locks the file from being read.
638 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
639 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
640 raise LockingUnsupportedError()
641 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
643 stream
= open(filename
, open_mode
)
644 return stream
, filename
645 except OSError as err
:
646 if attempt
or err
.errno
in (errno
.EACCES
,):
648 old_filename
, filename
= filename
, sanitize_path(filename
)
649 if old_filename
== filename
:
653 def timeconvert(timestr
):
654 """Convert RFC 2822 defined time string into system timestamp"""
656 timetuple
= email
.utils
.parsedate_tz(timestr
)
657 if timetuple
is not None:
658 timestamp
= email
.utils
.mktime_tz(timetuple
)
662 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
663 """Sanitizes a string so it could be used as part of a filename.
664 @param restricted Use a stricter subset of allowed characters
665 @param is_id Whether this is an ID that should be kept unchanged if possible.
666 If unset, yt-dlp's new sanitization rules are in effect
671 def replace_insane(char
):
672 if restricted
and char
in ACCENT_CHARS
:
673 return ACCENT_CHARS
[char
]
674 elif not restricted
and char
== '\n':
676 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
677 # Replace with their full-width unicode counterparts
678 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
679 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
682 return '' if restricted
else '\''
684 return '\0_\0-' if restricted
else '\0 \0-'
685 elif char
in '\\/|*<>':
687 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
691 # Replace look-alike Unicode glyphs
692 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
693 s
= unicodedata
.normalize('NFKC', s
)
694 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
695 result
= ''.join(map(replace_insane
, s
))
696 if is_id
is NO_DEFAULT
:
697 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
698 STRIP_RE
= r
'(?:\0.|[ _-])*'
699 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
700 result
= result
.replace('\0', '') or '_'
703 while '__' in result
:
704 result
= result
.replace('__', '_')
705 result
= result
.strip('_')
706 # Common case of "Foreign band name - English song title"
707 if restricted
and result
.startswith('-_'):
709 if result
.startswith('-'):
710 result
= '_' + result
[len('-'):]
711 result
= result
.lstrip('.')
717 def sanitize_path(s
, force
=False):
718 """Sanitizes and normalizes path on Windows"""
719 if sys
.platform
== 'win32':
721 drive_or_unc
, _
= os
.path
.splitdrive(s
)
727 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
731 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
732 for path_part
in norm_path
]
734 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
735 elif force
and s
and s
[0] == os
.path
.sep
:
736 sanitized_path
.insert(0, os
.path
.sep
)
737 return os
.path
.join(*sanitized_path
)
740 def sanitize_url(url
, *, scheme
='http'):
741 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
742 # the number of unwanted failures due to missing protocol
745 elif url
.startswith('//'):
746 return f
'{scheme}:{url}'
747 # Fix some common typos seen so far
749 # https://github.com/ytdl-org/youtube-dl/issues/15649
750 (r
'^httpss://', r
'https://'),
751 # https://bx1.be/lives/direct-tv/
752 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
754 for mistake
, fixup
in COMMON_TYPOS
:
755 if re
.match(mistake
, url
):
756 return re
.sub(mistake
, fixup
, url
)
760 def extract_basic_auth(url
):
761 parts
= urllib
.parse
.urlsplit(url
)
762 if parts
.username
is None:
764 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
765 parts
.hostname
if parts
.port
is None
766 else '%s:%d' % (parts
.hostname
, parts
.port
))))
767 auth_payload
= base64
.b64encode(
768 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
769 return url
, f
'Basic {auth_payload.decode()}'
772 def sanitized_Request(url
, *args
, **kwargs
):
773 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
774 if auth_header
is not None:
775 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
776 headers
['Authorization'] = auth_header
777 return urllib
.request
.Request(url
, *args
, **kwargs
)
781 """Expand shell variables and ~"""
782 return os
.path
.expandvars(compat_expanduser(s
))
785 def orderedSet(iterable
, *, lazy
=False):
786 """Remove all duplicates from the input iterable"""
788 seen
= [] # Do not use set since the items can be unhashable
794 return _iter() if lazy
else list(_iter())
797 def _htmlentity_transform(entity_with_semicolon
):
798 """Transforms an HTML entity to a character."""
799 entity
= entity_with_semicolon
[:-1]
801 # Known non-numeric HTML entity
802 if entity
in html
.entities
.name2codepoint
:
803 return chr(html
.entities
.name2codepoint
[entity
])
805 # TODO: HTML5 allows entities without a semicolon.
806 # E.g. 'Éric' should be decoded as 'Éric'.
807 if entity_with_semicolon
in html
.entities
.html5
:
808 return html
.entities
.html5
[entity_with_semicolon
]
810 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
812 numstr
= mobj
.group(1)
813 if numstr
.startswith('x'):
815 numstr
= '0%s' % numstr
818 # See https://github.com/ytdl-org/youtube-dl/issues/7518
819 with contextlib
.suppress(ValueError):
820 return chr(int(numstr
, base
))
822 # Unknown entity in name, return its literal representation
823 return '&%s;' % entity
829 assert isinstance(s
, str)
832 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
835 def escapeHTML(text
):
838 .replace('&', '&')
839 .replace('<', '<')
840 .replace('>', '>')
841 .replace('"', '"')
842 .replace("'", ''')
846 def process_communicate_or_kill(p
, *args
, **kwargs
):
847 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
848 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
849 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
852 class Popen(subprocess
.Popen
):
853 if sys
.platform
== 'win32':
854 _startupinfo
= subprocess
.STARTUPINFO()
855 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
860 def _fix_pyinstaller_ld_path(env
):
861 """Restore LD_LIBRARY_PATH when using PyInstaller
862 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
863 https://github.com/yt-dlp/yt-dlp/issues/4573
865 if not hasattr(sys
, '_MEIPASS'):
869 orig
= env
.get(f
'{key}_ORIG')
875 _fix('LD_LIBRARY_PATH') # Linux
876 _fix('DYLD_LIBRARY_PATH') # macOS
878 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
880 env
= os
.environ
.copy()
881 self
._fix
_pyinstaller
_ld
_path
(env
)
884 kwargs
['universal_newlines'] = True # For 3.6 compatibility
885 kwargs
.setdefault('encoding', 'utf-8')
886 kwargs
.setdefault('errors', 'replace')
887 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
889 def communicate_or_kill(self
, *args
, **kwargs
):
891 return self
.communicate(*args
, **kwargs
)
892 except BaseException
: # Including KeyboardInterrupt
893 self
.kill(timeout
=None)
896 def kill(self
, *, timeout
=0):
899 self
.wait(timeout
=timeout
)
902 def run(cls
, *args
, timeout
=None, **kwargs
):
903 with cls(*args
, **kwargs
) as proc
:
904 default
= '' if proc
.text_mode
else b
''
905 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
906 return stdout
or default
, stderr
or default
, proc
.returncode
909 def get_subprocess_encoding():
910 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
911 # For subprocess calls, encode with locale encoding
912 # Refer to http://stackoverflow.com/a/9951851/35070
913 encoding
= preferredencoding()
915 encoding
= sys
.getfilesystemencoding()
921 def encodeFilename(s
, for_subprocess
=False):
922 assert isinstance(s
, str)
926 def decodeFilename(b
, for_subprocess
=False):
930 def encodeArgument(s
):
931 # Legacy code that uses byte strings
932 # Uncomment the following line after fixing all post processors
933 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
934 return s
if isinstance(s
, str) else s
.decode('ascii')
937 def decodeArgument(b
):
941 def decodeOption(optval
):
944 if isinstance(optval
, bytes):
945 optval
= optval
.decode(preferredencoding())
947 assert isinstance(optval
, str)
951 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
954 def timetuple_from_msec(msec
):
955 secs
, msec
= divmod(msec
, 1000)
956 mins
, secs
= divmod(secs
, 60)
957 hrs
, mins
= divmod(mins
, 60)
958 return _timetuple(hrs
, mins
, secs
, msec
)
961 def formatSeconds(secs
, delim
=':', msec
=False):
962 time
= timetuple_from_msec(secs
* 1000)
964 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
966 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
968 ret
= '%d' % time
.seconds
969 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
972 def _ssl_load_windows_store_certs(ssl_context
, storename
):
973 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
975 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
976 if encoding
== 'x509_asn' and (
977 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
978 except PermissionError
:
981 with contextlib
.suppress(ssl
.SSLError
):
982 ssl_context
.load_verify_locations(cadata
=cert
)
985 def make_HTTPS_handler(params
, **kwargs
):
986 opts_check_certificate
= not params
.get('nocheckcertificate')
987 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
988 context
.check_hostname
= opts_check_certificate
989 if params
.get('legacyserverconnect'):
990 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
991 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
992 context
.set_ciphers('DEFAULT')
994 sys
.version_info
< (3, 10)
995 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
996 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
998 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
999 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
1000 # in some situations [2][3].
1001 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
1002 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1003 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1004 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1005 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1006 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1007 # 4. https://peps.python.org/pep-0644/
1008 # 5. https://peps.python.org/pep-0644/#libressl-support
1009 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1010 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1011 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1013 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1014 if opts_check_certificate
:
1015 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1016 context
.load_verify_locations(cafile
=certifi
.where())
1019 context
.load_default_certs()
1020 # Work around the issue in load_default_certs when there are bad certificates. See:
1021 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1022 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1023 except ssl
.SSLError
:
1024 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1025 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1026 for storename
in ('CA', 'ROOT'):
1027 _ssl_load_windows_store_certs(context
, storename
)
1028 context
.set_default_verify_paths()
1030 client_certfile
= params
.get('client_certificate')
1033 context
.load_cert_chain(
1034 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1035 password
=params
.get('client_certificate_password'))
1036 except ssl
.SSLError
:
1037 raise YoutubeDLError('Unable to load client certificate')
1039 # Some servers may reject requests if ALPN extension is not sent. See:
1040 # https://github.com/python/cpython/issues/85140
1041 # https://github.com/yt-dlp/yt-dlp/issues/3878
1042 with contextlib
.suppress(NotImplementedError):
1043 context
.set_alpn_protocols(['http/1.1'])
1045 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1048 def bug_reports_message(before
=';'):
1049 from .update
import REPOSITORY
1051 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1052 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1054 before
= before
.rstrip()
1055 if not before
or before
.endswith(('.', '!', '?')):
1056 msg
= msg
[0].title() + msg
[1:]
1058 return (before
+ ' ' if before
else '') + msg
1061 class YoutubeDLError(Exception):
1062 """Base exception for YoutubeDL errors."""
1065 def __init__(self
, msg
=None):
1068 elif self
.msg
is None:
1069 self
.msg
= type(self
).__name
__
1070 super().__init
__(self
.msg
)
1073 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1074 if hasattr(ssl
, 'CertificateError'):
1075 network_exceptions
.append(ssl
.CertificateError
)
1076 network_exceptions
= tuple(network_exceptions
)
1079 class ExtractorError(YoutubeDLError
):
1080 """Error during info extraction."""
1082 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1083 """ tb, if given, is the original traceback (so that it can be printed out).
1084 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1086 if sys
.exc_info()[0] in network_exceptions
:
1089 self
.orig_msg
= str(msg
)
1091 self
.expected
= expected
1093 self
.video_id
= video_id
1095 self
.exc_info
= sys
.exc_info() # preserve original exception
1096 if isinstance(self
.exc_info
[1], ExtractorError
):
1097 self
.exc_info
= self
.exc_info
[1].exc_info
1099 super().__init
__(''.join((
1100 format_field(ie
, None, '[%s] '),
1101 format_field(video_id
, None, '%s: '),
1103 format_field(cause
, None, ' (caused by %r)'),
1104 '' if expected
else bug_reports_message())))
1106 def format_traceback(self
):
1107 return join_nonempty(
1108 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1109 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1113 class UnsupportedError(ExtractorError
):
1114 def __init__(self
, url
):
1116 'Unsupported URL: %s' % url
, expected
=True)
1120 class RegexNotFoundError(ExtractorError
):
1121 """Error when a regex didn't match"""
1125 class GeoRestrictedError(ExtractorError
):
1126 """Geographic restriction Error exception.
1128 This exception may be thrown when a video is not available from your
1129 geographic location due to geographic restrictions imposed by a website.
1132 def __init__(self
, msg
, countries
=None, **kwargs
):
1133 kwargs
['expected'] = True
1134 super().__init
__(msg
, **kwargs
)
1135 self
.countries
= countries
1138 class UserNotLive(ExtractorError
):
1139 """Error when a channel/user is not live"""
1141 def __init__(self
, msg
=None, **kwargs
):
1142 kwargs
['expected'] = True
1143 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1146 class DownloadError(YoutubeDLError
):
1147 """Download Error exception.
1149 This exception may be thrown by FileDownloader objects if they are not
1150 configured to continue on errors. They will contain the appropriate
1154 def __init__(self
, msg
, exc_info
=None):
1155 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1156 super().__init
__(msg
)
1157 self
.exc_info
= exc_info
1160 class EntryNotInPlaylist(YoutubeDLError
):
1161 """Entry not in playlist exception.
1163 This exception will be thrown by YoutubeDL when a requested entry
1164 is not found in the playlist info_dict
1166 msg
= 'Entry not found in info'
1169 class SameFileError(YoutubeDLError
):
1170 """Same File exception.
1172 This exception will be thrown by FileDownloader objects if they detect
1173 multiple files would have to be downloaded to the same file on disk.
1175 msg
= 'Fixed output name but more than one file to download'
1177 def __init__(self
, filename
=None):
1178 if filename
is not None:
1179 self
.msg
+= f
': {filename}'
1180 super().__init
__(self
.msg
)
1183 class PostProcessingError(YoutubeDLError
):
1184 """Post Processing exception.
1186 This exception may be raised by PostProcessor's .run() method to
1187 indicate an error in the postprocessing task.
1191 class DownloadCancelled(YoutubeDLError
):
1192 """ Exception raised when the download queue should be interrupted """
1193 msg
= 'The download was cancelled'
1196 class ExistingVideoReached(DownloadCancelled
):
1197 """ --break-on-existing triggered """
1198 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1201 class RejectedVideoReached(DownloadCancelled
):
1202 """ --break-on-reject triggered """
1203 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1206 class MaxDownloadsReached(DownloadCancelled
):
1207 """ --max-downloads limit has been reached. """
1208 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1211 class ReExtractInfo(YoutubeDLError
):
1212 """ Video info needs to be re-extracted. """
1214 def __init__(self
, msg
, expected
=False):
1215 super().__init
__(msg
)
1216 self
.expected
= expected
1219 class ThrottledDownload(ReExtractInfo
):
1220 """ Download speed below --throttled-rate. """
1221 msg
= 'The download speed is below throttle limit'
1224 super().__init
__(self
.msg
, expected
=False)
1227 class UnavailableVideoError(YoutubeDLError
):
1228 """Unavailable Format exception.
1230 This exception will be thrown when a video is requested
1231 in a format that is not available for that video.
1233 msg
= 'Unable to download video'
1235 def __init__(self
, err
=None):
1237 self
.msg
+= f
': {err}'
1238 super().__init
__(self
.msg
)
1241 class ContentTooShortError(YoutubeDLError
):
1242 """Content Too Short exception.
1244 This exception may be raised by FileDownloader objects when a file they
1245 download is too small for what the server announced first, indicating
1246 the connection was probably interrupted.
1249 def __init__(self
, downloaded
, expected
):
1250 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1252 self
.downloaded
= downloaded
1253 self
.expected
= expected
1256 class XAttrMetadataError(YoutubeDLError
):
1257 def __init__(self
, code
=None, msg
='Unknown error'):
1258 super().__init
__(msg
)
1262 # Parsing code and msg
1263 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1264 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1265 self
.reason
= 'NO_SPACE'
1266 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1267 self
.reason
= 'VALUE_TOO_LONG'
1269 self
.reason
= 'NOT_SUPPORTED'
1272 class XAttrUnavailableError(YoutubeDLError
):
1276 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1277 hc
= http_class(*args
, **kwargs
)
1278 source_address
= ydl_handler
._params
.get('source_address')
1280 if source_address
is not None:
1281 # This is to workaround _create_connection() from socket where it will try all
1282 # address data from getaddrinfo() including IPv6. This filters the result from
1283 # getaddrinfo() based on the source_address value.
1284 # This is based on the cpython socket.create_connection() function.
1285 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1286 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1287 host
, port
= address
1289 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1290 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1291 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1292 if addrs
and not ip_addrs
:
1293 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1295 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1296 % (ip_version
, source_address
[0]))
1297 for res
in ip_addrs
:
1298 af
, socktype
, proto
, canonname
, sa
= res
1301 sock
= socket
.socket(af
, socktype
, proto
)
1302 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1303 sock
.settimeout(timeout
)
1304 sock
.bind(source_address
)
1306 err
= None # Explicitly break reference cycle
1308 except OSError as _
:
1310 if sock
is not None:
1315 raise OSError('getaddrinfo returns an empty list')
1316 if hasattr(hc
, '_create_connection'):
1317 hc
._create
_connection
= _create_connection
1318 hc
.source_address
= (source_address
, 0)
1323 def handle_youtubedl_headers(headers
):
1324 filtered_headers
= headers
1326 if 'Youtubedl-no-compression' in filtered_headers
:
1327 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1328 del filtered_headers
['Youtubedl-no-compression']
1330 return filtered_headers
1333 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1334 """Handler for HTTP requests and responses.
1336 This class, when installed with an OpenerDirector, automatically adds
1337 the standard headers to every HTTP request and handles gzipped and
1338 deflated responses from web servers. If compression is to be avoided in
1339 a particular request, the original request in the program code only has
1340 to include the HTTP header "Youtubedl-no-compression", which will be
1341 removed before making the real request.
1343 Part of this code was copied from:
1345 http://techknack.net/python-urllib2-handlers/
1347 Andrew Rowls, the author of that code, agreed to release it to the
1351 def __init__(self
, params
, *args
, **kwargs
):
1352 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1353 self
._params
= params
1355 def http_open(self
, req
):
1356 conn_class
= http
.client
.HTTPConnection
1358 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1360 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1361 del req
.headers
['Ytdl-socks-proxy']
1363 return self
.do_open(functools
.partial(
1364 _create_http_connection
, self
, conn_class
, False),
1372 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1374 return zlib
.decompress(data
)
1380 return brotli
.decompress(data
)
1382 def http_request(self
, req
):
1383 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1384 # always respected by websites, some tend to give out URLs with non percent-encoded
1385 # non-ASCII characters (see telemb.py, ard.py [#3412])
1386 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1387 # To work around aforementioned issue we will replace request's original URL with
1388 # percent-encoded one
1389 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1390 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1391 url
= req
.get_full_url()
1392 url_escaped
= escape_url(url
)
1394 # Substitute URL if any change after escaping
1395 if url
!= url_escaped
:
1396 req
= update_Request(req
, url
=url_escaped
)
1398 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1399 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1400 # The dict keys are capitalized because of this bug by urllib
1401 if h
.capitalize() not in req
.headers
:
1402 req
.add_header(h
, v
)
1404 if 'Accept-encoding' not in req
.headers
:
1405 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1407 req
.headers
= handle_youtubedl_headers(req
.headers
)
1409 return super().do_request_(req
)
1411 def http_response(self
, req
, resp
):
1414 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1415 content
= resp
.read()
1416 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1418 uncompressed
= io
.BytesIO(gz
.read())
1419 except OSError as original_ioerror
:
1420 # There may be junk add the end of the file
1421 # See http://stackoverflow.com/q/4928560/35070 for details
1422 for i
in range(1, 1024):
1424 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1425 uncompressed
= io
.BytesIO(gz
.read())
1430 raise original_ioerror
1431 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1432 resp
.msg
= old_resp
.msg
1433 del resp
.headers
['Content-encoding']
1435 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1436 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1437 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1438 resp
.msg
= old_resp
.msg
1439 del resp
.headers
['Content-encoding']
1441 if resp
.headers
.get('Content-encoding', '') == 'br':
1442 resp
= urllib
.request
.addinfourl(
1443 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1444 resp
.msg
= old_resp
.msg
1445 del resp
.headers
['Content-encoding']
1446 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1447 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1448 if 300 <= resp
.code
< 400:
1449 location
= resp
.headers
.get('Location')
1451 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1452 location
= location
.encode('iso-8859-1').decode()
1453 location_escaped
= escape_url(location
)
1454 if location
!= location_escaped
:
1455 del resp
.headers
['Location']
1456 resp
.headers
['Location'] = location_escaped
1459 https_request
= http_request
1460 https_response
= http_response
1463 def make_socks_conn_class(base_class
, socks_proxy
):
1464 assert issubclass(base_class
, (
1465 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1467 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1468 if url_components
.scheme
.lower() == 'socks5':
1469 socks_type
= ProxyType
.SOCKS5
1470 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1471 socks_type
= ProxyType
.SOCKS4
1472 elif url_components
.scheme
.lower() == 'socks4a':
1473 socks_type
= ProxyType
.SOCKS4A
1475 def unquote_if_non_empty(s
):
1478 return urllib
.parse
.unquote_plus(s
)
1482 url_components
.hostname
, url_components
.port
or 1080,
1484 unquote_if_non_empty(url_components
.username
),
1485 unquote_if_non_empty(url_components
.password
),
1488 class SocksConnection(base_class
):
1490 self
.sock
= sockssocket()
1491 self
.sock
.setproxy(*proxy_args
)
1492 if isinstance(self
.timeout
, (int, float)):
1493 self
.sock
.settimeout(self
.timeout
)
1494 self
.sock
.connect((self
.host
, self
.port
))
1496 if isinstance(self
, http
.client
.HTTPSConnection
):
1497 if hasattr(self
, '_context'): # Python > 2.6
1498 self
.sock
= self
._context
.wrap_socket(
1499 self
.sock
, server_hostname
=self
.host
)
1501 self
.sock
= ssl
.wrap_socket(self
.sock
)
1503 return SocksConnection
1506 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1507 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1508 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1509 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1510 self
._params
= params
1512 def https_open(self
, req
):
1514 conn_class
= self
._https
_conn
_class
1516 if hasattr(self
, '_context'): # python > 2.6
1517 kwargs
['context'] = self
._context
1518 if hasattr(self
, '_check_hostname'): # python 3.x
1519 kwargs
['check_hostname'] = self
._check
_hostname
1521 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1523 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1524 del req
.headers
['Ytdl-socks-proxy']
1527 return self
.do_open(
1528 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1529 except urllib
.error
.URLError
as e
:
1530 if (isinstance(e
.reason
, ssl
.SSLError
)
1531 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1532 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1536 def is_path_like(f
):
1537 return isinstance(f
, (str, bytes, os
.PathLike
))
1540 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1542 See [1] for cookie file format.
1544 1. https://curl.haxx.se/docs/http-cookies.html
1546 _HTTPONLY_PREFIX
= '#HttpOnly_'
1548 _HEADER
= '''# Netscape HTTP Cookie File
1549 # This file is generated by yt-dlp. Do not edit.
1552 _CookieFileEntry
= collections
.namedtuple(
1554 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1556 def __init__(self
, filename
=None, *args
, **kwargs
):
1557 super().__init
__(None, *args
, **kwargs
)
1558 if is_path_like(filename
):
1559 filename
= os
.fspath(filename
)
1560 self
.filename
= filename
1563 def _true_or_false(cndn
):
1564 return 'TRUE' if cndn
else 'FALSE'
1566 @contextlib.contextmanager
1567 def open(self
, file, *, write
=False):
1568 if is_path_like(file):
1569 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1576 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1579 if (not ignore_discard
and cookie
.discard
1580 or not ignore_expires
and cookie
.is_expired(now
)):
1582 name
, value
= cookie
.name
, cookie
.value
1584 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1585 # with no name, whereas http.cookiejar regards it as a
1586 # cookie with no value.
1587 name
, value
= '', name
1588 f
.write('%s\n' % '\t'.join((
1590 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1592 self
._true
_or
_false
(cookie
.secure
),
1593 str_or_none(cookie
.expires
, default
=''),
1597 def save(self
, filename
=None, *args
, **kwargs
):
1599 Save cookies to a file.
1600 Code is taken from CPython 3.6
1601 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1603 if filename
is None:
1604 if self
.filename
is not None:
1605 filename
= self
.filename
1607 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1609 # Store session cookies with `expires` set to 0 instead of an empty string
1611 if cookie
.expires
is None:
1614 with self
.open(filename
, write
=True) as f
:
1615 f
.write(self
._HEADER
)
1616 self
._really
_save
(f
, *args
, **kwargs
)
1618 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1619 """Load cookies from a file."""
1620 if filename
is None:
1621 if self
.filename
is not None:
1622 filename
= self
.filename
1624 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1626 def prepare_line(line
):
1627 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1628 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1629 # comments and empty lines are fine
1630 if line
.startswith('#') or not line
.strip():
1632 cookie_list
= line
.split('\t')
1633 if len(cookie_list
) != self
._ENTRY
_LEN
:
1634 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1635 cookie
= self
._CookieFileEntry
(*cookie_list
)
1636 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1637 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1641 with self
.open(filename
) as f
:
1644 cf
.write(prepare_line(line
))
1645 except http
.cookiejar
.LoadError
as e
:
1646 if f
'{line.strip()} '[0] in '[{"':
1647 raise http
.cookiejar
.LoadError(
1648 'Cookies file must be Netscape formatted, not JSON. See '
1649 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1650 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1653 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1654 # Session cookies are denoted by either `expires` field set to
1655 # an empty string or 0. MozillaCookieJar only recognizes the former
1656 # (see [1]). So we need force the latter to be recognized as session
1657 # cookies on our own.
1658 # Session cookies may be important for cookies-based authentication,
1659 # e.g. usually, when user does not check 'Remember me' check box while
1660 # logging in on a site, some important cookies are stored as session
1661 # cookies so that not recognizing them will result in failed login.
1662 # 1. https://bugs.python.org/issue17164
1664 # Treat `expires=0` cookies as session cookies
1665 if cookie
.expires
== 0:
1666 cookie
.expires
= None
1667 cookie
.discard
= True
1670 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1671 def __init__(self
, cookiejar
=None):
1672 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1674 def http_response(self
, request
, response
):
1675 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1677 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1678 https_response
= http_response
1681 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1682 """YoutubeDL redirect handler
1684 The code is based on HTTPRedirectHandler implementation from CPython [1].
1686 This redirect handler solves two issues:
1687 - ensures redirect URL is always unicode under python 2
1688 - introduces support for experimental HTTP response status code
1689 308 Permanent Redirect [2] used by some sites [3]
1691 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1692 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1693 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1696 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1698 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1699 """Return a Request or None in response to a redirect.
1701 This is called by the http_error_30x methods when a
1702 redirection response is received. If a redirection should
1703 take place, return a new Request to allow http_error_30x to
1704 perform the redirect. Otherwise, raise HTTPError if no-one
1705 else should try to handle this url. Return None if you can't
1706 but another Handler might.
1708 m
= req
.get_method()
1709 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1710 or code
in (301, 302, 303) and m
== "POST")):
1711 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1712 # Strictly (according to RFC 2616), 301 or 302 in response to
1713 # a POST MUST NOT cause a redirection without confirmation
1714 # from the user (of urllib.request, in this case). In practice,
1715 # essentially all clients do redirect in this case, so we do
1718 # Be conciliant with URIs containing a space. This is mainly
1719 # redundant with the more complete encoding done in http_error_302(),
1720 # but it is kept for compatibility with other callers.
1721 newurl
= newurl
.replace(' ', '%20')
1723 CONTENT_HEADERS
= ("content-length", "content-type")
1724 # NB: don't use dict comprehension for python 2.6 compatibility
1725 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1727 # A 303 must either use GET or HEAD for subsequent request
1728 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1729 if code
== 303 and m
!= 'HEAD':
1731 # 301 and 302 redirects are commonly turned into a GET from a POST
1732 # for subsequent requests by browsers, so we'll do the same.
1733 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1734 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1735 if code
in (301, 302) and m
== 'POST':
1738 return urllib
.request
.Request(
1739 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1740 unverifiable
=True, method
=m
)
1743 def extract_timezone(date_str
):
1746 ^.{8,}? # >=8 char non-TZ prefix, if present
1747 (?P<tz>Z| # just the UTC Z, or
1748 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1749 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1750 [ ]? # optional space
1751 (?P<sign>\+|-) # +/-
1752 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1756 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1757 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1758 if timezone
is not None:
1759 date_str
= date_str
[:-len(m
.group('tz'))]
1760 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1762 date_str
= date_str
[:-len(m
.group('tz'))]
1763 if not m
.group('sign'):
1764 timezone
= datetime
.timedelta()
1766 sign
= 1 if m
.group('sign') == '+' else -1
1767 timezone
= datetime
.timedelta(
1768 hours
=sign
* int(m
.group('hours')),
1769 minutes
=sign
* int(m
.group('minutes')))
1770 return timezone
, date_str
1773 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1774 """ Return a UNIX timestamp from the given date """
1776 if date_str
is None:
1779 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1781 if timezone
is None:
1782 timezone
, date_str
= extract_timezone(date_str
)
1784 with contextlib
.suppress(ValueError):
1785 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1786 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1787 return calendar
.timegm(dt
.timetuple())
1790 def date_formats(day_first
=True):
1791 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1794 def unified_strdate(date_str
, day_first
=True):
1795 """Return a string with the date in the format YYYYMMDD"""
1797 if date_str
is None:
1801 date_str
= date_str
.replace(',', ' ')
1802 # Remove AM/PM + timezone
1803 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1804 _
, date_str
= extract_timezone(date_str
)
1806 for expression
in date_formats(day_first
):
1807 with contextlib
.suppress(ValueError):
1808 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1809 if upload_date
is None:
1810 timetuple
= email
.utils
.parsedate_tz(date_str
)
1812 with contextlib
.suppress(ValueError):
1813 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1814 if upload_date
is not None:
1815 return str(upload_date
)
1818 def unified_timestamp(date_str
, day_first
=True):
1819 if date_str
is None:
1822 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1823 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1825 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1826 timezone
, date_str
= extract_timezone(date_str
)
1828 # Remove AM/PM + timezone
1829 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1831 # Remove unrecognized timezones from ISO 8601 alike timestamps
1832 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1834 date_str
= date_str
[:-len(m
.group('tz'))]
1836 # Python only supports microseconds, so remove nanoseconds
1837 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1839 date_str
= m
.group(1)
1841 for expression
in date_formats(day_first
):
1842 with contextlib
.suppress(ValueError):
1843 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1844 return calendar
.timegm(dt
.timetuple())
1846 timetuple
= email
.utils
.parsedate_tz(date_str
)
1848 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1851 def determine_ext(url
, default_ext
='unknown_video'):
1852 if url
is None or '.' not in url
:
1854 guess
= url
.partition('?')[0].rpartition('.')[2]
1855 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1857 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1858 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1859 return guess
.rstrip('/')
1864 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1865 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1868 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1870 Return a datetime object from a string.
1872 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1874 @param format strftime format of DATE
1875 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1876 auto: round to the unit provided in date_str (if applicable).
1878 auto_precision
= False
1879 if precision
== 'auto':
1880 auto_precision
= True
1881 precision
= 'microsecond'
1882 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1883 if date_str
in ('now', 'today'):
1885 if date_str
== 'yesterday':
1886 return today
- datetime
.timedelta(days
=1)
1888 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1890 if match
is not None:
1891 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1892 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1893 unit
= match
.group('unit')
1894 if unit
== 'month' or unit
== 'year':
1895 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1901 delta
= datetime
.timedelta(**{unit + 's': time}
)
1902 new_date
= start_time
+ delta
1904 return datetime_round(new_date
, unit
)
1907 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1910 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1912 Return a date object from a string using datetime_from_str
1914 @param strict Restrict allowed patterns to "YYYYMMDD" and
1915 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1917 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1918 raise ValueError(f
'Invalid date format "{date_str}"')
1919 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1922 def datetime_add_months(dt
, months
):
1923 """Increment/Decrement a datetime object by months."""
1924 month
= dt
.month
+ months
- 1
1925 year
= dt
.year
+ month
// 12
1926 month
= month
% 12 + 1
1927 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1928 return dt
.replace(year
, month
, day
)
1931 def datetime_round(dt
, precision
='day'):
1933 Round a datetime object's time to a specific precision
1935 if precision
== 'microsecond':
1944 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1945 timestamp
= calendar
.timegm(dt
.timetuple())
1946 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1949 def hyphenate_date(date_str
):
1951 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1952 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1953 if match
is not None:
1954 return '-'.join(match
.groups())
1960 """Represents a time interval between two dates"""
1962 def __init__(self
, start
=None, end
=None):
1963 """start and end must be strings in the format accepted by date"""
1964 if start
is not None:
1965 self
.start
= date_from_str(start
, strict
=True)
1967 self
.start
= datetime
.datetime
.min.date()
1969 self
.end
= date_from_str(end
, strict
=True)
1971 self
.end
= datetime
.datetime
.max.date()
1972 if self
.start
> self
.end
:
1973 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1977 """Returns a range that only contains the given day"""
1978 return cls(day
, day
)
1980 def __contains__(self
, date
):
1981 """Check if the date is in the range"""
1982 if not isinstance(date
, datetime
.date
):
1983 date
= date_from_str(date
)
1984 return self
.start
<= date
<= self
.end
1987 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1989 def __eq__(self
, other
):
1990 return (isinstance(other
, DateRange
)
1991 and self
.start
== other
.start
and self
.end
== other
.end
)
1994 def platform_name():
1995 """ Returns the platform name as a str """
1996 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
1997 return platform
.platform()
2001 def system_identifier():
2002 python_implementation
= platform
.python_implementation()
2003 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2004 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2006 with contextlib
.suppress(OSError): # We may not have access to the executable
2007 libc_ver
= platform
.libc_ver()
2009 return 'Python %s (%s %s) - %s (%s%s)' % (
2010 platform
.python_version(),
2011 python_implementation
,
2012 platform
.architecture()[0],
2013 platform
.platform(),
2014 ssl
.OPENSSL_VERSION
,
2015 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2020 def get_windows_version():
2021 ''' Get Windows version. returns () if it's not running on Windows '''
2022 if compat_os_name
== 'nt':
2023 return version_tuple(platform
.win32_ver()[1])
2028 def write_string(s
, out
=None, encoding
=None):
2029 assert isinstance(s
, str)
2030 out
= out
or sys
.stderr
2032 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2033 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2035 enc
, buffer = None, out
2036 if 'b' in getattr(out
, 'mode', ''):
2037 enc
= encoding
or preferredencoding()
2038 elif hasattr(out
, 'buffer'):
2040 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2042 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2046 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2047 from . import _IN_CLI
2049 if msg
in deprecation_warning
._cache
:
2051 deprecation_warning
._cache
.add(msg
)
2053 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2054 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2057 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2060 deprecation_warning
._cache
= set()
2063 def bytes_to_intlist(bs
):
2066 if isinstance(bs
[0], int): # Python 3
2069 return [ord(c
) for c
in bs
]
2072 def intlist_to_bytes(xs
):
2075 return struct
.pack('%dB' % len(xs
), *xs
)
2078 class LockingUnsupportedError(OSError):
2079 msg
= 'File locking is not supported'
2082 super().__init
__(self
.msg
)
2085 # Cross-platform file locking
2086 if sys
.platform
== 'win32':
2088 import ctypes
.wintypes
2091 class OVERLAPPED(ctypes
.Structure
):
2093 ('Internal', ctypes
.wintypes
.LPVOID
),
2094 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2095 ('Offset', ctypes
.wintypes
.DWORD
),
2096 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2097 ('hEvent', ctypes
.wintypes
.HANDLE
),
2100 kernel32
= ctypes
.windll
.kernel32
2101 LockFileEx
= kernel32
.LockFileEx
2102 LockFileEx
.argtypes
= [
2103 ctypes
.wintypes
.HANDLE
, # hFile
2104 ctypes
.wintypes
.DWORD
, # dwFlags
2105 ctypes
.wintypes
.DWORD
, # dwReserved
2106 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2107 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2108 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2110 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2111 UnlockFileEx
= kernel32
.UnlockFileEx
2112 UnlockFileEx
.argtypes
= [
2113 ctypes
.wintypes
.HANDLE
, # hFile
2114 ctypes
.wintypes
.DWORD
, # dwReserved
2115 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2116 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2117 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2119 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2120 whole_low
= 0xffffffff
2121 whole_high
= 0x7fffffff
2123 def _lock_file(f
, exclusive
, block
):
2124 overlapped
= OVERLAPPED()
2125 overlapped
.Offset
= 0
2126 overlapped
.OffsetHigh
= 0
2127 overlapped
.hEvent
= 0
2128 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2130 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2131 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2132 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2133 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2134 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2136 def _unlock_file(f
):
2137 assert f
._lock
_file
_overlapped
_p
2138 handle
= msvcrt
.get_osfhandle(f
.fileno())
2139 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2140 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2146 def _lock_file(f
, exclusive
, block
):
2147 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2149 flags |
= fcntl
.LOCK_NB
2151 fcntl
.flock(f
, flags
)
2152 except BlockingIOError
:
2154 except OSError: # AOSP does not have flock()
2155 fcntl
.lockf(f
, flags
)
2157 def _unlock_file(f
):
2159 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2161 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2165 def _lock_file(f
, exclusive
, block
):
2166 raise LockingUnsupportedError()
2168 def _unlock_file(f
):
2169 raise LockingUnsupportedError()
2175 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2176 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2177 raise NotImplementedError(mode
)
2178 self
.mode
, self
.block
= mode
, block
2180 writable
= any(f
in mode
for f
in 'wax+')
2181 readable
= any(f
in mode
for f
in 'r+')
2182 flags
= functools
.reduce(operator
.ior
, (
2183 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2184 getattr(os
, 'O_BINARY', 0), # Windows only
2185 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2186 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2187 os
.O_APPEND
if 'a' in mode
else 0,
2188 os
.O_EXCL
if 'x' in mode
else 0,
2189 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2192 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2194 def __enter__(self
):
2195 exclusive
= 'r' not in self
.mode
2197 _lock_file(self
.f
, exclusive
, self
.block
)
2202 if 'w' in self
.mode
:
2205 except OSError as e
:
2207 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2208 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2217 _unlock_file(self
.f
)
2221 def __exit__(self
, *_
):
2230 def __getattr__(self
, attr
):
2231 return getattr(self
.f
, attr
)
2238 def get_filesystem_encoding():
2239 encoding
= sys
.getfilesystemencoding()
2240 return encoding
if encoding
is not None else 'utf-8'
2243 def shell_quote(args
):
2245 encoding
= get_filesystem_encoding()
2247 if isinstance(a
, bytes):
2248 # We may get a filename encoded with 'encodeFilename'
2249 a
= a
.decode(encoding
)
2250 quoted_args
.append(compat_shlex_quote(a
))
2251 return ' '.join(quoted_args
)
2254 def smuggle_url(url
, data
):
2255 """ Pass additional data in a URL for internal use. """
2257 url
, idata
= unsmuggle_url(url
, {})
2259 sdata
= urllib
.parse
.urlencode(
2260 {'__youtubedl_smuggle': json.dumps(data)}
)
2261 return url
+ '#' + sdata
2264 def unsmuggle_url(smug_url
, default
=None):
2265 if '#__youtubedl_smuggle' not in smug_url
:
2266 return smug_url
, default
2267 url
, _
, sdata
= smug_url
.rpartition('#')
2268 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2269 data
= json
.loads(jsond
)
2273 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2274 """ Formats numbers with decimal sufixes like K, M, etc """
2275 num
, factor
= float_or_none(num
), float(factor
)
2276 if num
is None or num
< 0:
2278 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2279 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2280 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2282 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2283 converted
= num
/ (factor
** exponent
)
2284 return fmt
% (converted
, suffix
)
2287 def format_bytes(bytes):
2288 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2291 def lookup_unit_table(unit_table
, s
):
2292 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2294 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2297 num_str
= m
.group('num').replace(',', '.')
2298 mult
= unit_table
[m
.group('unit')]
2299 return int(float(num_str
) * mult
)
2302 def parse_filesize(s
):
2306 # The lower-case forms are of course incorrect and unofficial,
2307 # but we support those too
2324 'megabytes': 1000 ** 2,
2325 'mebibytes': 1024 ** 2,
2331 'gigabytes': 1000 ** 3,
2332 'gibibytes': 1024 ** 3,
2338 'terabytes': 1000 ** 4,
2339 'tebibytes': 1024 ** 4,
2345 'petabytes': 1000 ** 5,
2346 'pebibytes': 1024 ** 5,
2352 'exabytes': 1000 ** 6,
2353 'exbibytes': 1024 ** 6,
2359 'zettabytes': 1000 ** 7,
2360 'zebibytes': 1024 ** 7,
2366 'yottabytes': 1000 ** 8,
2367 'yobibytes': 1024 ** 8,
2370 return lookup_unit_table(_UNIT_TABLE
, s
)
2377 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2379 if re
.match(r
'^[\d,.]+$', s
):
2380 return str_to_int(s
)
2393 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2397 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2399 return str_to_int(mobj
.group(1))
2402 def parse_resolution(s
, *, lenient
=False):
2407 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2409 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2412 'width': int(mobj
.group('w')),
2413 'height': int(mobj
.group('h')),
2416 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2418 return {'height': int(mobj.group(1))}
2420 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2422 return {'height': int(mobj.group(1)) * 540}
2427 def parse_bitrate(s
):
2428 if not isinstance(s
, str):
2430 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2432 return int(mobj
.group(1))
2435 def month_by_name(name
, lang
='en'):
2436 """ Return the number of a month by (locale-independently) English name """
2438 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2441 return month_names
.index(name
) + 1
2446 def month_by_abbreviation(abbrev
):
2447 """ Return the number of a month by (locale-independently) English
2451 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2456 def fix_xml_ampersands(xml_str
):
2457 """Replace all the '&' by '&' in XML"""
2459 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2464 def setproctitle(title
):
2465 assert isinstance(title
, str)
2467 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2474 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2478 # LoadLibrary in Windows Python 2.7.13 only expects
2479 # a bytestring, but since unicode_literals turns
2480 # every string into a unicode string, it fails.
2482 title_bytes
= title
.encode()
2483 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2484 buf
.value
= title_bytes
2486 libc
.prctl(15, buf
, 0, 0, 0)
2487 except AttributeError:
2488 return # Strange libc, just skip this
2491 def remove_start(s
, start
):
2492 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2495 def remove_end(s
, end
):
2496 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2499 def remove_quotes(s
):
2500 if s
is None or len(s
) < 2:
2502 for quote
in ('"', "'", ):
2503 if s
[0] == quote
and s
[-1] == quote
:
2508 def get_domain(url
):
2510 This implementation is inconsistent, but is kept for compatibility.
2511 Use this only for "webpage_url_domain"
2513 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2516 def url_basename(url
):
2517 path
= urllib
.parse
.urlparse(url
).path
2518 return path
.strip('/').split('/')[-1]
2522 return re
.match(r
'https?://[^?#]+/', url
).group()
2525 def urljoin(base
, path
):
2526 if isinstance(path
, bytes):
2527 path
= path
.decode()
2528 if not isinstance(path
, str) or not path
:
2530 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2532 if isinstance(base
, bytes):
2533 base
= base
.decode()
2534 if not isinstance(base
, str) or not re
.match(
2535 r
'^(?:https?:)?//', base
):
2537 return urllib
.parse
.urljoin(base
, path
)
2540 class HEADRequest(urllib
.request
.Request
):
2541 def get_method(self
):
2545 class PUTRequest(urllib
.request
.Request
):
2546 def get_method(self
):
2550 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2551 if get_attr
and v
is not None:
2552 v
= getattr(v
, get_attr
, None)
2554 return int(v
) * invscale
// scale
2555 except (ValueError, TypeError, OverflowError):
2559 def str_or_none(v
, default
=None):
2560 return default
if v
is None else str(v
)
2563 def str_to_int(int_str
):
2564 """ A more relaxed version of int_or_none """
2565 if isinstance(int_str
, int):
2567 elif isinstance(int_str
, str):
2568 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2569 return int_or_none(int_str
)
2572 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2576 return float(v
) * invscale
/ scale
2577 except (ValueError, TypeError):
2581 def bool_or_none(v
, default
=None):
2582 return v
if isinstance(v
, bool) else default
2585 def strip_or_none(v
, default
=None):
2586 return v
.strip() if isinstance(v
, str) else default
2589 def url_or_none(url
):
2590 if not url
or not isinstance(url
, str):
2593 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2596 def request_to_url(req
):
2597 if isinstance(req
, urllib
.request
.Request
):
2598 return req
.get_full_url()
2603 def strftime_or_none(timestamp
, date_format
, default
=None):
2604 datetime_object
= None
2606 if isinstance(timestamp
, (int, float)): # unix timestamp
2607 # Using naive datetime here can break timestamp() in Windows
2608 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2609 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2610 elif isinstance(timestamp
, str): # assume YYYYMMDD
2611 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2612 date_format
= re
.sub( # Support %s on windows
2613 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2614 return datetime_object
.strftime(date_format
)
2615 except (ValueError, TypeError, AttributeError):
2619 def parse_duration(s
):
2620 if not isinstance(s
, str):
2626 days
, hours
, mins
, secs
, ms
= [None] * 5
2627 m
= re
.match(r
'''(?x)
2629 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2630 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2631 (?P<ms>[.:][0-9]+)?Z?$
2634 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2639 [0-9]+\s*y(?:ears?)?,?\s*
2642 [0-9]+\s*m(?:onths?)?,?\s*
2645 [0-9]+\s*w(?:eeks?)?,?\s*
2648 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2652 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2655 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2658 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2661 days
, hours
, mins
, secs
, ms
= m
.groups()
2663 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2665 hours
, mins
= m
.groups()
2670 ms
= ms
.replace(':', '.')
2671 return sum(float(part
or 0) * mult
for part
, mult
in (
2672 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2675 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2676 name
, real_ext
= os
.path
.splitext(filename
)
2678 f
'{name}.{ext}{real_ext}'
2679 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2680 else f
'{filename}.{ext}')
2683 def replace_extension(filename
, ext
, expected_real_ext
=None):
2684 name
, real_ext
= os
.path
.splitext(filename
)
2685 return '{}.{}'.format(
2686 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2690 def check_executable(exe
, args
=[]):
2691 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2692 args can be a list of arguments for a short output (like -version) """
2694 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2700 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2702 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2704 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2705 # SIGTTOU if yt-dlp is run in the background.
2706 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2707 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2708 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2714 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2715 assert isinstance(output
, str)
2716 if version_re
is None:
2717 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2718 m
= re
.search(version_re
, output
)
2725 def get_exe_version(exe
, args
=['--version'],
2726 version_re
=None, unrecognized
='present'):
2727 """ Returns the version of the specified executable,
2728 or False if the executable is not present """
2729 out
= _get_exe_version_output(exe
, args
)
2730 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2733 def frange(start
=0, stop
=None, step
=1):
2736 start
, stop
= 0, start
2737 sign
= [-1, 1][step
> 0] if step
else 0
2738 while sign
* start
< sign
* stop
:
2743 class LazyList(collections
.abc
.Sequence
):
2744 """Lazy immutable list from an iterable
2745 Note that slices of a LazyList are lists and not LazyList"""
2747 class IndexError(IndexError):
2750 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2751 self
._iterable
= iter(iterable
)
2752 self
._cache
= [] if _cache
is None else _cache
2753 self
._reversed
= reverse
2757 # We need to consume the entire iterable to iterate in reverse
2758 yield from self
.exhaust()
2760 yield from self
._cache
2761 for item
in self
._iterable
:
2762 self
._cache
.append(item
)
2766 self
._cache
.extend(self
._iterable
)
2767 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2771 """Evaluate the entire iterable"""
2772 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2775 def _reverse_index(x
):
2776 return None if x
is None else ~x
2778 def __getitem__(self
, idx
):
2779 if isinstance(idx
, slice):
2781 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2782 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2783 elif isinstance(idx
, int):
2785 idx
= self
._reverse
_index
(idx
)
2786 start
, stop
, step
= idx
, idx
, 0
2788 raise TypeError('indices must be integers or slices')
2789 if ((start
or 0) < 0 or (stop
or 0) < 0
2790 or (start
is None and step
< 0)
2791 or (stop
is None and step
> 0)):
2792 # We need to consume the entire iterable to be able to slice from the end
2793 # Obviously, never use this with infinite iterables
2796 return self
._cache
[idx
]
2797 except IndexError as e
:
2798 raise self
.IndexError(e
) from e
2799 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2801 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2803 return self
._cache
[idx
]
2804 except IndexError as e
:
2805 raise self
.IndexError(e
) from e
2809 self
[-1] if self
._reversed
else self
[0]
2810 except self
.IndexError:
2816 return len(self
._cache
)
2818 def __reversed__(self
):
2819 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2822 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2825 # repr and str should mimic a list. So we exhaust the iterable
2826 return repr(self
.exhaust())
2829 return repr(self
.exhaust())
2834 class IndexError(IndexError):
2838 # This is only useful for tests
2839 return len(self
.getslice())
2841 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2842 self
._pagefunc
= pagefunc
2843 self
._pagesize
= pagesize
2844 self
._pagecount
= float('inf')
2845 self
._use
_cache
= use_cache
2848 def getpage(self
, pagenum
):
2849 page_results
= self
._cache
.get(pagenum
)
2850 if page_results
is None:
2851 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2853 self
._cache
[pagenum
] = page_results
2856 def getslice(self
, start
=0, end
=None):
2857 return list(self
._getslice
(start
, end
))
2859 def _getslice(self
, start
, end
):
2860 raise NotImplementedError('This method must be implemented by subclasses')
2862 def __getitem__(self
, idx
):
2863 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2864 if not isinstance(idx
, int) or idx
< 0:
2865 raise TypeError('indices must be non-negative integers')
2866 entries
= self
.getslice(idx
, idx
+ 1)
2868 raise self
.IndexError()
2872 class OnDemandPagedList(PagedList
):
2873 """Download pages until a page with less than maximum results"""
2875 def _getslice(self
, start
, end
):
2876 for pagenum
in itertools
.count(start
// self
._pagesize
):
2877 firstid
= pagenum
* self
._pagesize
2878 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2879 if start
>= nextfirstid
:
2883 start
% self
._pagesize
2884 if firstid
<= start
< nextfirstid
2887 ((end
- 1) % self
._pagesize
) + 1
2888 if (end
is not None and firstid
<= end
<= nextfirstid
)
2892 page_results
= self
.getpage(pagenum
)
2894 self
._pagecount
= pagenum
- 1
2896 if startv
!= 0 or endv
is not None:
2897 page_results
= page_results
[startv
:endv
]
2898 yield from page_results
2900 # A little optimization - if current page is not "full", ie. does
2901 # not contain page_size videos then we can assume that this page
2902 # is the last one - there are no more ids on further pages -
2903 # i.e. no need to query again.
2904 if len(page_results
) + startv
< self
._pagesize
:
2907 # If we got the whole page, but the next page is not interesting,
2908 # break out early as well
2909 if end
== nextfirstid
:
2913 class InAdvancePagedList(PagedList
):
2914 """PagedList with total number of pages known in advance"""
2916 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2917 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2918 self
._pagecount
= pagecount
2920 def _getslice(self
, start
, end
):
2921 start_page
= start
// self
._pagesize
2922 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2923 skip_elems
= start
- start_page
* self
._pagesize
2924 only_more
= None if end
is None else end
- start
2925 for pagenum
in range(start_page
, end_page
):
2926 page_results
= self
.getpage(pagenum
)
2928 page_results
= page_results
[skip_elems
:]
2930 if only_more
is not None:
2931 if len(page_results
) < only_more
:
2932 only_more
-= len(page_results
)
2934 yield from page_results
[:only_more
]
2936 yield from page_results
2939 class PlaylistEntries
:
2940 MissingEntry
= object()
2941 is_exhausted
= False
2943 def __init__(self
, ydl
, info_dict
):
2946 # _entries must be assigned now since infodict can change during iteration
2947 entries
= info_dict
.get('entries')
2949 raise EntryNotInPlaylist('There are no entries')
2950 elif isinstance(entries
, list):
2951 self
.is_exhausted
= True
2953 requested_entries
= info_dict
.get('requested_entries')
2954 self
.is_incomplete
= bool(requested_entries
)
2955 if self
.is_incomplete
:
2956 assert self
.is_exhausted
2957 self
._entries
= [self
.MissingEntry
] * max(requested_entries
)
2958 for i
, entry
in zip(requested_entries
, entries
):
2959 self
._entries
[i
- 1] = entry
2960 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2961 self
._entries
= entries
2963 self
._entries
= LazyList(entries
)
2965 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2966 (?P<start>[+-]?\d+)?
2968 (?P<end>[+-]?\d+|inf(?:inite)?)?
2969 (?::(?P<step>[+-]?\d+))?
2973 def parse_playlist_items(cls
, string
):
2974 for segment
in string
.split(','):
2976 raise ValueError('There is two or more consecutive commas')
2977 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2979 raise ValueError(f
'{segment!r} is not a valid specification')
2980 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2981 if int_or_none(step
) == 0:
2982 raise ValueError(f
'Step in {segment!r} cannot be zero')
2983 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2985 def get_requested_items(self
):
2986 playlist_items
= self
.ydl
.params
.get('playlist_items')
2987 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2988 playlist_end
= self
.ydl
.params
.get('playlistend')
2989 # For backwards compatibility, interpret -1 as whole list
2990 if playlist_end
in (-1, None):
2992 if not playlist_items
:
2993 playlist_items
= f
'{playlist_start}:{playlist_end}'
2994 elif playlist_start
!= 1 or playlist_end
:
2995 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2997 for index
in self
.parse_playlist_items(playlist_items
):
2998 for i
, entry
in self
[index
]:
3003 # TODO: Add auto-generated fields
3004 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3005 except (ExistingVideoReached
, RejectedVideoReached
):
3008 def get_full_count(self
):
3009 if self
.is_exhausted
and not self
.is_incomplete
:
3011 elif isinstance(self
._entries
, InAdvancePagedList
):
3012 if self
._entries
._pagesize
== 1:
3013 return self
._entries
._pagecount
3015 @functools.cached_property
3017 if isinstance(self
._entries
, list):
3020 entry
= self
._entries
[i
]
3022 entry
= self
.MissingEntry
3023 if not self
.is_incomplete
:
3024 raise self
.IndexError()
3025 if entry
is self
.MissingEntry
:
3026 raise EntryNotInPlaylist(f
'Entry {i} cannot be found')
3031 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3032 except (LazyList
.IndexError, PagedList
.IndexError):
3033 raise self
.IndexError()
3036 def __getitem__(self
, idx
):
3037 if isinstance(idx
, int):
3038 idx
= slice(idx
, idx
)
3040 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3041 step
= 1 if idx
.step
is None else idx
.step
3042 if idx
.start
is None:
3043 start
= 0 if step
> 0 else len(self
) - 1
3045 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3047 # NB: Do not call len(self) when idx == [:]
3048 if idx
.stop
is None:
3049 stop
= 0 if step
< 0 else float('inf')
3051 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3052 stop
+= [-1, 1][step
> 0]
3054 for i
in frange(start
, stop
, step
):
3058 entry
= self
._getter
(i
)
3059 except self
.IndexError:
3060 self
.is_exhausted
= True
3067 return len(tuple(self
[:]))
3069 class IndexError(IndexError):
3073 def uppercase_escape(s
):
3074 unicode_escape
= codecs
.getdecoder('unicode_escape')
3076 r
'\\U[0-9a-fA-F]{8}',
3077 lambda m
: unicode_escape(m
.group(0))[0],
3081 def lowercase_escape(s
):
3082 unicode_escape
= codecs
.getdecoder('unicode_escape')
3084 r
'\\u[0-9a-fA-F]{4}',
3085 lambda m
: unicode_escape(m
.group(0))[0],
3089 def escape_rfc3986(s
):
3090 """Escape non-ASCII characters as suggested by RFC 3986"""
3091 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3094 def escape_url(url
):
3095 """Escape URL as suggested by RFC 3986"""
3096 url_parsed
= urllib
.parse
.urlparse(url
)
3097 return url_parsed
._replace
(
3098 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3099 path
=escape_rfc3986(url_parsed
.path
),
3100 params
=escape_rfc3986(url_parsed
.params
),
3101 query
=escape_rfc3986(url_parsed
.query
),
3102 fragment
=escape_rfc3986(url_parsed
.fragment
)
3106 def parse_qs(url
, **kwargs
):
3107 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3110 def read_batch_urls(batch_fd
):
3112 if not isinstance(url
, str):
3113 url
= url
.decode('utf-8', 'replace')
3114 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3115 for bom
in BOM_UTF8
:
3116 if url
.startswith(bom
):
3117 url
= url
[len(bom
):]
3119 if not url
or url
.startswith(('#', ';', ']')):
3121 # "#" cannot be stripped out since it is part of the URI
3122 # However, it can be safely stripped out if following a whitespace
3123 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3125 with contextlib
.closing(batch_fd
) as fd
:
3126 return [url
for url
in map(fixup
, fd
) if url
]
3129 def urlencode_postdata(*args
, **kargs
):
3130 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3133 def update_url_query(url
, query
):
3136 parsed_url
= urllib
.parse
.urlparse(url
)
3137 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3139 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3140 query
=urllib
.parse
.urlencode(qs
, True)))
3143 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3144 req_headers
= req
.headers
.copy()
3145 req_headers
.update(headers
or {})
3146 req_data
= data
or req
.data
3147 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3148 req_get_method
= req
.get_method()
3149 if req_get_method
== 'HEAD':
3150 req_type
= HEADRequest
3151 elif req_get_method
== 'PUT':
3152 req_type
= PUTRequest
3154 req_type
= urllib
.request
.Request
3156 req_url
, data
=req_data
, headers
=req_headers
,
3157 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3158 if hasattr(req
, 'timeout'):
3159 new_req
.timeout
= req
.timeout
3163 def _multipart_encode_impl(data
, boundary
):
3164 content_type
= 'multipart/form-data; boundary=%s' % boundary
3167 for k
, v
in data
.items():
3168 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3169 if isinstance(k
, str):
3171 if isinstance(v
, str):
3173 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3174 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3175 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3176 if boundary
.encode('ascii') in content
:
3177 raise ValueError('Boundary overlaps with data')
3180 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3182 return out
, content_type
3185 def multipart_encode(data
, boundary
=None):
3187 Encode a dict to RFC 7578-compliant form-data
3190 A dict where keys and values can be either Unicode or bytes-like
3193 If specified a Unicode object, it's used as the boundary. Otherwise
3194 a random boundary is generated.
3196 Reference: https://tools.ietf.org/html/rfc7578
3198 has_specified_boundary
= boundary
is not None
3201 if boundary
is None:
3202 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3205 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3208 if has_specified_boundary
:
3212 return out
, content_type
3215 def variadic(x
, allowed_types
=(str, bytes, dict)):
3216 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3219 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3220 for val
in map(d
.get
, variadic(key_or_keys
)):
3221 if val
is not None and (val
or not skip_false_values
):
3226 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3229 val
= f(*args
, **kwargs
)
3230 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3233 if expected_type
is None or isinstance(val
, expected_type
):
3237 def try_get(src
, getter
, expected_type
=None):
3238 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3241 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3242 return {k: v for k, v in dct.items() if cndn(k, v)}
3245 def merge_dicts(*dicts
):
3247 for a_dict
in dicts
:
3248 for k
, v
in a_dict
.items():
3249 if (v
is not None and k
not in merged
3250 or isinstance(v
, str) and merged
[k
] == ''):
3255 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3256 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3268 TV_PARENTAL_GUIDELINES
= {
3278 def parse_age_limit(s
):
3279 # isinstance(False, int) is True. So type() must be used instead
3280 if type(s
) is int: # noqa: E721
3281 return s
if 0 <= s
<= 21 else None
3282 elif not isinstance(s
, str):
3284 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3286 return int(m
.group('age'))
3289 return US_RATINGS
[s
]
3290 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3292 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3296 def strip_jsonp(code
):
3299 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3300 (?:\s*&&\s*(?P=func_name))?
3301 \s*\(\s*(?P<callback_data>.*)\);?
3302 \s*?(?://[^\n]*)*$''',
3303 r
'\g<callback_data>', code
)
3306 def js_to_json(code
, vars={}, *, strict
=False):
3307 # vars is a dict of var, val pairs to substitute
3308 STRING_QUOTES
= '\'"'
3309 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3310 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3311 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3313 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3314 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3317 def process_escape(match
):
3318 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3319 escape
= match
.group(1) or match
.group(2)
3321 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3322 else R
'\u00' if escape
== 'x'
3323 else '' if escape
== '\n'
3328 if v
in ('true', 'false', 'null'):
3330 elif v
in ('undefined', 'void 0'):
3332 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3335 if v
[0] in STRING_QUOTES
:
3336 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3337 return f
'"{escaped}"'
3339 for regex
, base
in INTEGER_TABLE
:
3340 im
= re
.match(regex
, v
)
3342 i
= int(im
.group(1), base
)
3343 return f
'"{i}":' if v
.endswith(':') else str(i
)
3346 return json
.dumps(vars[v
])
3351 raise ValueError(f
'Unknown value: {v}')
3353 def create_map(mobj
):
3354 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3356 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3358 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3359 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3361 return re
.sub(rf
'''(?sx)
3363 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3364 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3365 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3366 [0-9]+(?={SKIP_RE}:)|
3371 def qualities(quality_ids
):
3372 """ Get a numeric quality value out of a list of possible values """
3375 return quality_ids
.index(qid
)
3381 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3385 'default': '%(title)s [%(id)s].%(ext)s',
3386 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3392 'description': 'description',
3393 'annotation': 'annotations.xml',
3394 'infojson': 'info.json',
3397 'pl_thumbnail': None,
3398 'pl_description': 'description',
3399 'pl_infojson': 'info.json',
3402 # As of [1] format syntax is:
3403 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3404 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3405 STR_FORMAT_RE_TMPL
= r
'''(?x)
3406 (?<!%)(?P<prefix>(?:%%)*)
3408 (?P<has_key>\((?P<key>{0})\))?
3410 (?P<conversion>[#0\-+ ]+)?
3412 (?P<precision>\.\d+)?
3413 (?P<len_mod>[hlL])? # unused in python
3414 {1} # conversion type
3419 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3422 def limit_length(s
, length
):
3423 """ Add ellipses to overly long strings """
3428 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3432 def version_tuple(v
):
3433 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3436 def is_outdated_version(version
, limit
, assume_new
=True):
3438 return not assume_new
3440 return version_tuple(version
) < version_tuple(limit
)
3442 return not assume_new
3445 def ytdl_is_updateable():
3446 """ Returns if yt-dlp can be updated with -U """
3448 from .update
import is_non_updateable
3450 return not is_non_updateable()
3453 def args_to_str(args
):
3454 # Get a short string representation for a subprocess command
3455 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3458 def error_to_compat_str(err
):
3462 def error_to_str(err
):
3463 return f
'{type(err).__name__}: {err}'
3466 def mimetype2ext(mt
):
3470 mt
, _
, params
= mt
.partition(';')
3475 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3476 # it's the most popular one
3477 'audio/mpeg': 'mp3',
3478 'audio/x-wav': 'wav',
3480 'audio/wave': 'wav',
3483 ext
= FULL_MAP
.get(mt
)
3489 'smptett+xml': 'tt',
3493 'x-mp4-fragmented': 'mp4',
3494 'x-ms-sami': 'sami',
3497 'x-mpegurl': 'm3u8',
3498 'vnd.apple.mpegurl': 'm3u8',
3502 'vnd.ms-sstr+xml': 'ism',
3506 'filmstrip+json': 'fs',
3510 _
, _
, subtype
= mt
.rpartition('/')
3511 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3522 _
, _
, suffix
= subtype
.partition('+')
3523 ext
= SUFFIX_MAP
.get(suffix
)
3527 return subtype
.replace('+', '.')
3530 def ext2mimetype(ext_or_url
):
3533 if '.' not in ext_or_url
:
3534 ext_or_url
= f
'file.{ext_or_url}'
3535 return mimetypes
.guess_type(ext_or_url
)[0]
3538 def parse_codecs(codecs_str
):
3539 # http://tools.ietf.org/html/rfc6381
3542 split_codecs
= list(filter(None, map(
3543 str.strip
, codecs_str
.strip().strip(',').split(','))))
3544 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3545 for full_codec
in split_codecs
:
3546 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3547 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3548 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3552 if parts
[0] in ('dvh1', 'dvhe'):
3554 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3556 elif parts
[:2] == ['vp9', '2']:
3558 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3559 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3560 acodec
= acodec
or full_codec
3561 elif parts
[0] in ('stpp', 'wvtt'):
3562 scodec
= scodec
or full_codec
3564 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3565 if vcodec
or acodec
or scodec
:
3567 'vcodec': vcodec
or 'none',
3568 'acodec': acodec
or 'none',
3569 'dynamic_range': hdr
,
3570 **({'scodec': scodec}
if scodec
is not None else {}),
3572 elif len(split_codecs
) == 2:
3574 'vcodec': split_codecs
[0],
3575 'acodec': split_codecs
[1],
3580 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3581 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3583 allow_mkv
= not preferences
or 'mkv' in preferences
3585 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3586 return 'mkv' # TODO: any other format allows this?
3588 # TODO: All codecs supported by parse_codecs isn't handled here
3589 COMPATIBLE_CODECS
= {
3591 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3592 'h264', 'aacl', 'ec-3', # Set in ISM
3595 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3596 'vp9x', 'vp8x', # in the webm spec
3600 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3601 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3603 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3604 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3605 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3609 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3612 for ext
in preferences
or vexts
:
3613 current_exts
= {ext, *vexts, *aexts}
3614 if ext
== 'mkv' or current_exts
== {ext}
or any(
3615 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3617 return 'mkv' if allow_mkv
else preferences
[-1]
3620 def urlhandle_detect_ext(url_handle
):
3621 getheader
= url_handle
.headers
.get
3623 cd
= getheader('Content-Disposition')
3625 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3627 e
= determine_ext(m
.group('filename'), default_ext
=None)
3631 return mimetype2ext(getheader('Content-Type'))
3634 def encode_data_uri(data
, mime_type
):
3635 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3638 def age_restricted(content_limit
, age_limit
):
3639 """ Returns True iff the content should be blocked """
3641 if age_limit
is None: # No limit set
3643 if content_limit
is None:
3644 return False # Content available for everyone
3645 return age_limit
< content_limit
3648 # List of known byte-order-marks (BOM)
3650 (b
'\xef\xbb\xbf', 'utf-8'),
3651 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3652 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3653 (b
'\xff\xfe', 'utf-16-le'),
3654 (b
'\xfe\xff', 'utf-16-be'),
3658 def is_html(first_bytes
):
3659 """ Detect whether a file contains HTML by examining its first bytes. """
3662 for bom
, enc
in BOMS
:
3663 while first_bytes
.startswith(bom
):
3664 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3666 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3669 def determine_protocol(info_dict
):
3670 protocol
= info_dict
.get('protocol')
3671 if protocol
is not None:
3674 url
= sanitize_url(info_dict
['url'])
3675 if url
.startswith('rtmp'):
3677 elif url
.startswith('mms'):
3679 elif url
.startswith('rtsp'):
3682 ext
= determine_ext(url
)
3684 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3688 return urllib
.parse
.urlparse(url
).scheme
3691 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3692 """ Render a list of rows, each as a list of values.
3693 Text after a \t will be right aligned """
3695 return len(remove_terminal_sequences(string
).replace('\t', ''))
3697 def get_max_lens(table
):
3698 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3700 def filter_using_list(row
, filterArray
):
3701 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3703 max_lens
= get_max_lens(data
) if hide_empty
else []
3704 header_row
= filter_using_list(header_row
, max_lens
)
3705 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3707 table
= [header_row
] + data
3708 max_lens
= get_max_lens(table
)
3711 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3712 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3714 for pos
, text
in enumerate(map(str, row
)):
3716 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3718 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3719 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3723 def _match_one(filter_part
, dct
, incomplete
):
3724 # TODO: Generalize code with YoutubeDL._build_format_filter
3725 STRING_OPERATORS
= {
3726 '*=': operator
.contains
,
3727 '^=': lambda attr
, value
: attr
.startswith(value
),
3728 '$=': lambda attr
, value
: attr
.endswith(value
),
3729 '~=': lambda attr
, value
: re
.search(value
, attr
),
3731 COMPARISON_OPERATORS
= {
3733 '<=': operator
.le
, # "<=" must be defined above "<"
3740 if isinstance(incomplete
, bool):
3741 is_incomplete
= lambda _
: incomplete
3743 is_incomplete
= lambda k
: k
in incomplete
3745 operator_rex
= re
.compile(r
'''(?x)
3747 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3749 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3752 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3753 m = operator_rex.fullmatch(filter_part.strip())
3756 unnegated_op = COMPARISON_OPERATORS[m['op']]
3758 op = lambda attr, value: not unnegated_op(attr, value)
3761 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3763 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3764 actual_value = dct.get(m['key'])
3765 numeric_comparison = None
3766 if isinstance(actual_value, (int, float)):
3767 # If the original field is a string and matching comparisonvalue is
3768 # a number we should respect the origin of the original field
3769 # and process comparison value as a string (see
3770 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3772 numeric_comparison = int(comparison_value)
3774 numeric_comparison = parse_filesize(comparison_value)
3775 if numeric_comparison is None:
3776 numeric_comparison = parse_filesize(f'{comparison_value}B')
3777 if numeric_comparison is None:
3778 numeric_comparison = parse_duration(comparison_value)
3779 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3780 raise ValueError('Operator %s only supports string values!' % m['op'])
3781 if actual_value is None:
3782 return is_incomplete(m['key']) or m['none_inclusive']
3783 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3786 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3787 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3789 operator_rex = re.compile(r'''(?x
)
3790 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3791 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3792 m = operator_rex.fullmatch(filter_part.strip())
3794 op = UNARY_OPERATORS[m.group('op')]
3795 actual_value = dct.get(m.group('key'))
3796 if is_incomplete(m.group('key')) and actual_value is None:
3798 return op(actual_value)
3800 raise ValueError('Invalid filter part %r' % filter_part)
3803 def match_str(filter_str, dct, incomplete=False):
3804 """ Filter a dictionary with a simple string syntax.
3805 @returns Whether the filter passes
3806 @param incomplete Set of keys that is expected to be missing from dct.
3807 Can be True/False to indicate all/none of the keys may be missing.
3808 All conditions on incomplete keys pass if the key is missing
3811 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3812 for filter_part in re.split(r'(?<!\\)&', filter_str))
3815 def match_filter_func(filters):
3818 filters = set(variadic(filters))
3820 interactive = '-' in filters
3824 def _match_func(info_dict, incomplete=False):
3825 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3826 return NO_DEFAULT if interactive and not incomplete else None
3828 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3829 filter_str = ') | ('.join(map(str.strip, filters))
3830 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3834 class download_range_func:
3835 def __init__(self, chapters, ranges):
3836 self.chapters, self.ranges = chapters, ranges
3838 def __call__(self, info_dict, ydl):
3839 if not self.ranges and not self.chapters:
3842 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3843 else 'Cannot match chapters since chapter information is unavailable')
3844 for regex in self.chapters or []:
3845 for i, chapter in enumerate(info_dict.get('chapters') or []):
3846 if re.search(regex, chapter['title']):
3848 yield {**chapter, 'index': i}
3849 if self.chapters and warning:
3850 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3852 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3854 def __eq__(self, other):
3855 return (isinstance(other, download_range_func)
3856 and self.chapters == other.chapters and self.ranges == other.ranges)
3859 def parse_dfxp_time_expr(time_expr):
3863 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3865 return float(mobj.group('time_offset'))
3867 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3869 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3872 def srt_subtitles_timecode(seconds):
3873 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3876 def ass_subtitles_timecode(seconds):
3877 time = timetuple_from_msec(seconds * 1000)
3878 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3881 def dfxp2srt(dfxp_data):
3883 @param dfxp_data A
bytes-like
object containing DFXP data
3884 @returns A
unicode object containing converted SRT data
3886 LEGACY_NAMESPACES = (
3887 (b'http://www.w3.org/ns/ttml', [
3888 b'http://www.w3.org/2004/11/ttaf1',
3889 b'http://www.w3.org/2006/04/ttaf1',
3890 b'http://www.w3.org/2006/10/ttaf1',
3892 (b'http://www.w3.org/ns/ttml#styling', [
3893 b'http://www.w3.org/ns/ttml#style',
3897 SUPPORTED_STYLING = [
3906 _x = functools.partial(xpath_with_ns, ns_map={
3907 'xml': 'http://www.w3.org/XML/1998/namespace',
3908 'ttml': 'http://www.w3.org/ns/ttml',
3909 'tts': 'http://www.w3.org/ns/ttml#styling',
3915 class TTMLPElementParser:
3917 _unclosed_elements = []
3918 _applied_styles = []
3920 def start(self, tag, attrib):
3921 if tag in (_x('ttml:br'), 'br'):
3924 unclosed_elements = []
3926 element_style_id = attrib.get('style')
3928 style.update(default_style)
3929 if element_style_id:
3930 style.update(styles.get(element_style_id, {}))
3931 for prop in SUPPORTED_STYLING:
3932 prop_val = attrib.get(_x('tts:' + prop))
3934 style[prop] = prop_val
3937 for k, v in sorted(style.items()):
3938 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3941 font += ' color="%s"' % v
3942 elif k == 'fontSize':
3943 font += ' size="%s"' % v
3944 elif k == 'fontFamily':
3945 font += ' face="%s"' % v
3946 elif k == 'fontWeight' and v == 'bold':
3948 unclosed_elements.append('b')
3949 elif k == 'fontStyle' and v == 'italic':
3951 unclosed_elements.append('i')
3952 elif k == 'textDecoration' and v == 'underline':
3954 unclosed_elements.append('u')
3956 self._out += '<font' + font + '>'
3957 unclosed_elements.append('font')
3959 if self._applied_styles:
3960 applied_style.update(self._applied_styles[-1])
3961 applied_style.update(style)
3962 self._applied_styles.append(applied_style)
3963 self._unclosed_elements.append(unclosed_elements)
3966 if tag not in (_x('ttml:br'), 'br'):
3967 unclosed_elements = self._unclosed_elements.pop()
3968 for element in reversed(unclosed_elements):
3969 self._out += '</%s>' % element
3970 if unclosed_elements and self._applied_styles:
3971 self._applied_styles.pop()
3973 def data(self, data):
3977 return self._out.strip()
3979 def parse_node(node):
3980 target = TTMLPElementParser()
3981 parser = xml.etree.ElementTree.XMLParser(target=target)
3982 parser.feed(xml.etree.ElementTree.tostring(node))
3983 return parser.close()
3985 for k, v in LEGACY_NAMESPACES:
3987 dfxp_data = dfxp_data.replace(ns, k)
3989 dfxp = compat_etree_fromstring(dfxp_data)
3991 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3994 raise ValueError('Invalid dfxp/TTML subtitle')
3998 for style in dfxp.findall(_x('.//ttml:style')):
3999 style_id = style.get('id') or style.get(_x('xml:id'))
4002 parent_style_id = style.get('style')
4004 if parent_style_id not in styles:
4007 styles[style_id] = styles[parent_style_id].copy()
4008 for prop in SUPPORTED_STYLING:
4009 prop_val = style.get(_x('tts:' + prop))
4011 styles.setdefault(style_id, {})[prop] = prop_val
4017 for p in ('body', 'div'):
4018 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
4021 style = styles.get(ele.get('style'))
4024 default_style.update(style)
4026 for para, index in zip(paras, itertools.count(1)):
4027 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
4028 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
4029 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
4030 if begin_time is None:
4035 end_time = begin_time + dur
4036 out.append('%d\n%s --> %s\n%s\n\n' % (
4038 srt_subtitles_timecode(begin_time),
4039 srt_subtitles_timecode(end_time),
4045 def cli_option(params, command_option, param, separator=None):
4046 param = params.get(param)
4047 return ([] if param is None
4048 else [command_option, str(param)] if separator is None
4049 else [f'{command_option}{separator}{param}'])
4052 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4053 param = params.get(param)
4054 assert param in (True, False, None)
4055 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4058 def cli_valueless_option(params, command_option, param, expected_value=True):
4059 return [command_option] if params.get(param) == expected_value else []
4062 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4063 if isinstance(argdict, (list, tuple)): # for backward compatibility
4070 assert isinstance(argdict, dict)
4072 assert isinstance(keys, (list, tuple))
4073 for key_list in keys:
4074 arg_list = list(filter(
4075 lambda x: x is not None,
4076 [argdict.get(key.lower()) for key in variadic(key_list)]))
4078 return [arg for args in arg_list for arg in args]
4082 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4083 main_key, exe = main_key.lower(), exe.lower()
4084 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4085 keys = [f'{root_key}{k}' for k in (keys or [''])]
4086 if root_key in keys:
4088 keys.append((main_key, exe))
4089 keys.append('default')
4092 return cli_configuration_args(argdict, keys, default, use_compat)
4096 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4155 'iw': 'heb', # Replaced by he in 1989 revision
4165 'in': 'ind', # Replaced by id in 1989 revision
4280 'ji': 'yid', # Replaced by yi in 1989 revision
4288 def short2long(cls, code):
4289 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4290 return cls._lang_map.get(code[:2])
4293 def long2short(cls, code):
4294 """Convert language code from ISO 639-2/T to ISO 639-1"""
4295 for short_name, long_name in cls._lang_map.items():
4296 if long_name == code:
4301 # From http://data.okfn.org/data/core/country-list
4303 'AF': 'Afghanistan',
4304 'AX': 'Åland Islands',
4307 'AS': 'American Samoa',
4312 'AG': 'Antigua and Barbuda',
4329 'BO': 'Bolivia, Plurinational State of',
4330 'BQ': 'Bonaire, Sint Eustatius and Saba',
4331 'BA': 'Bosnia and Herzegovina',
4333 'BV': 'Bouvet Island',
4335 'IO': 'British Indian Ocean Territory',
4336 'BN': 'Brunei Darussalam',
4338 'BF': 'Burkina Faso',
4344 'KY': 'Cayman Islands',
4345 'CF': 'Central African Republic',
4349 'CX': 'Christmas Island',
4350 'CC': 'Cocos (Keeling) Islands',
4354 'CD': 'Congo, the Democratic Republic of the',
4355 'CK': 'Cook Islands',
4357 'CI': 'Côte d\'Ivoire',
4362 'CZ': 'Czech Republic',
4366 'DO': 'Dominican Republic',
4369 'SV': 'El Salvador',
4370 'GQ': 'Equatorial Guinea',
4374 'FK': 'Falkland Islands (Malvinas)',
4375 'FO': 'Faroe Islands',
4379 'GF': 'French Guiana',
4380 'PF': 'French Polynesia',
4381 'TF': 'French Southern Territories',
4396 'GW': 'Guinea-Bissau',
4399 'HM': 'Heard Island and McDonald Islands',
4400 'VA': 'Holy See (Vatican City State)',
4407 'IR': 'Iran, Islamic Republic of',
4410 'IM': 'Isle of Man',
4420 'KP': 'Korea, Democratic People\'s Republic of',
4421 'KR': 'Korea, Republic of',
4424 'LA': 'Lao People\'s Democratic Republic',
4430 'LI': 'Liechtenstein',
4434 'MK': 'Macedonia, the Former Yugoslav Republic of',
4441 'MH': 'Marshall Islands',
4447 'FM': 'Micronesia, Federated States of',
4448 'MD': 'Moldova, Republic of',
4459 'NL': 'Netherlands',
4460 'NC': 'New Caledonia',
4461 'NZ': 'New Zealand',
4466 'NF': 'Norfolk Island',
4467 'MP': 'Northern Mariana Islands',
4472 'PS': 'Palestine, State of',
4474 'PG': 'Papua New Guinea',
4477 'PH': 'Philippines',
4481 'PR': 'Puerto Rico',
4485 'RU': 'Russian Federation',
4487 'BL': 'Saint Barthélemy',
4488 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4489 'KN': 'Saint Kitts and Nevis',
4490 'LC': 'Saint Lucia',
4491 'MF': 'Saint Martin (French part)',
4492 'PM': 'Saint Pierre and Miquelon',
4493 'VC': 'Saint Vincent and the Grenadines',
4496 'ST': 'Sao Tome and Principe',
4497 'SA': 'Saudi Arabia',
4501 'SL': 'Sierra Leone',
4503 'SX': 'Sint Maarten (Dutch part)',
4506 'SB': 'Solomon Islands',
4508 'ZA': 'South Africa',
4509 'GS': 'South Georgia and the South Sandwich Islands',
4510 'SS': 'South Sudan',
4515 'SJ': 'Svalbard and Jan Mayen',
4518 'CH': 'Switzerland',
4519 'SY': 'Syrian Arab Republic',
4520 'TW': 'Taiwan, Province of China',
4522 'TZ': 'Tanzania, United Republic of',
4524 'TL': 'Timor-Leste',
4528 'TT': 'Trinidad and Tobago',
4531 'TM': 'Turkmenistan',
4532 'TC': 'Turks and Caicos Islands',
4536 'AE': 'United Arab Emirates',
4537 'GB': 'United Kingdom',
4538 'US': 'United States',
4539 'UM': 'United States Minor Outlying Islands',
4543 'VE': 'Venezuela, Bolivarian Republic of',
4545 'VG': 'Virgin Islands, British',
4546 'VI': 'Virgin Islands, U.S.',
4547 'WF': 'Wallis and Futuna',
4548 'EH': 'Western Sahara',
4552 # Not ISO 3166 codes, but used for IP blocks
4553 'AP': 'Asia/Pacific Region',
4558 def short2full(cls, code):
4559 """Convert an ISO 3166-2 country code to the corresponding full name"""
4560 return cls._country_map.get(code.upper())
4564 # Major IPv4 address blocks per country
4566 'AD': '46.172.224.0/19',
4567 'AE': '94.200.0.0/13',
4568 'AF': '149.54.0.0/17',
4569 'AG': '209.59.64.0/18',
4570 'AI': '204.14.248.0/21',
4571 'AL': '46.99.0.0/16',
4572 'AM': '46.70.0.0/15',
4573 'AO': '105.168.0.0/13',
4574 'AP': '182.50.184.0/21',
4575 'AQ': '23.154.160.0/24',
4576 'AR': '181.0.0.0/12',
4577 'AS': '202.70.112.0/20',
4578 'AT': '77.116.0.0/14',
4579 'AU': '1.128.0.0/11',
4580 'AW': '181.41.0.0/18',
4581 'AX': '185.217.4.0/22',
4582 'AZ': '5.197.0.0/16',
4583 'BA': '31.176.128.0/17',
4584 'BB': '65.48.128.0/17',
4585 'BD': '114.130.0.0/16',
4587 'BF': '102.178.0.0/15',
4588 'BG': '95.42.0.0/15',
4589 'BH': '37.131.0.0/17',
4590 'BI': '154.117.192.0/18',
4591 'BJ': '137.255.0.0/16',
4592 'BL': '185.212.72.0/23',
4593 'BM': '196.12.64.0/18',
4594 'BN': '156.31.0.0/16',
4595 'BO': '161.56.0.0/16',
4596 'BQ': '161.0.80.0/20',
4597 'BR': '191.128.0.0/12',
4598 'BS': '24.51.64.0/18',
4599 'BT': '119.2.96.0/19',
4600 'BW': '168.167.0.0/16',
4601 'BY': '178.120.0.0/13',
4602 'BZ': '179.42.192.0/18',
4603 'CA': '99.224.0.0/11',
4604 'CD': '41.243.0.0/16',
4605 'CF': '197.242.176.0/21',
4606 'CG': '160.113.0.0/16',
4607 'CH': '85.0.0.0/13',
4608 'CI': '102.136.0.0/14',
4609 'CK': '202.65.32.0/19',
4610 'CL': '152.172.0.0/14',
4611 'CM': '102.244.0.0/14',
4612 'CN': '36.128.0.0/10',
4613 'CO': '181.240.0.0/12',
4614 'CR': '201.192.0.0/12',
4615 'CU': '152.206.0.0/15',
4616 'CV': '165.90.96.0/19',
4617 'CW': '190.88.128.0/17',
4618 'CY': '31.153.0.0/16',
4619 'CZ': '88.100.0.0/14',
4621 'DJ': '197.241.0.0/17',
4622 'DK': '87.48.0.0/12',
4623 'DM': '192.243.48.0/20',
4624 'DO': '152.166.0.0/15',
4625 'DZ': '41.96.0.0/12',
4626 'EC': '186.68.0.0/15',
4627 'EE': '90.190.0.0/15',
4628 'EG': '156.160.0.0/11',
4629 'ER': '196.200.96.0/20',
4630 'ES': '88.0.0.0/11',
4631 'ET': '196.188.0.0/14',
4632 'EU': '2.16.0.0/13',
4633 'FI': '91.152.0.0/13',
4634 'FJ': '144.120.0.0/16',
4635 'FK': '80.73.208.0/21',
4636 'FM': '119.252.112.0/20',
4637 'FO': '88.85.32.0/19',
4639 'GA': '41.158.0.0/15',
4641 'GD': '74.122.88.0/21',
4642 'GE': '31.146.0.0/16',
4643 'GF': '161.22.64.0/18',
4644 'GG': '62.68.160.0/19',
4645 'GH': '154.160.0.0/12',
4646 'GI': '95.164.0.0/16',
4647 'GL': '88.83.0.0/19',
4648 'GM': '160.182.0.0/15',
4649 'GN': '197.149.192.0/18',
4650 'GP': '104.250.0.0/19',
4651 'GQ': '105.235.224.0/20',
4652 'GR': '94.64.0.0/13',
4653 'GT': '168.234.0.0/16',
4654 'GU': '168.123.0.0/16',
4655 'GW': '197.214.80.0/20',
4656 'GY': '181.41.64.0/18',
4657 'HK': '113.252.0.0/14',
4658 'HN': '181.210.0.0/16',
4659 'HR': '93.136.0.0/13',
4660 'HT': '148.102.128.0/17',
4661 'HU': '84.0.0.0/14',
4662 'ID': '39.192.0.0/10',
4663 'IE': '87.32.0.0/12',
4664 'IL': '79.176.0.0/13',
4665 'IM': '5.62.80.0/20',
4666 'IN': '117.192.0.0/10',
4667 'IO': '203.83.48.0/21',
4668 'IQ': '37.236.0.0/14',
4669 'IR': '2.176.0.0/12',
4670 'IS': '82.221.0.0/16',
4671 'IT': '79.0.0.0/10',
4672 'JE': '87.244.64.0/18',
4673 'JM': '72.27.0.0/17',
4674 'JO': '176.29.0.0/16',
4675 'JP': '133.0.0.0/8',
4676 'KE': '105.48.0.0/12',
4677 'KG': '158.181.128.0/17',
4678 'KH': '36.37.128.0/17',
4679 'KI': '103.25.140.0/22',
4680 'KM': '197.255.224.0/20',
4681 'KN': '198.167.192.0/19',
4682 'KP': '175.45.176.0/22',
4683 'KR': '175.192.0.0/10',
4684 'KW': '37.36.0.0/14',
4685 'KY': '64.96.0.0/15',
4686 'KZ': '2.72.0.0/13',
4687 'LA': '115.84.64.0/18',
4688 'LB': '178.135.0.0/16',
4689 'LC': '24.92.144.0/20',
4690 'LI': '82.117.0.0/19',
4691 'LK': '112.134.0.0/15',
4692 'LR': '102.183.0.0/16',
4693 'LS': '129.232.0.0/17',
4694 'LT': '78.56.0.0/13',
4695 'LU': '188.42.0.0/16',
4696 'LV': '46.109.0.0/16',
4697 'LY': '41.252.0.0/14',
4698 'MA': '105.128.0.0/11',
4699 'MC': '88.209.64.0/18',
4700 'MD': '37.246.0.0/16',
4701 'ME': '178.175.0.0/17',
4702 'MF': '74.112.232.0/21',
4703 'MG': '154.126.0.0/17',
4704 'MH': '117.103.88.0/21',
4705 'MK': '77.28.0.0/15',
4706 'ML': '154.118.128.0/18',
4707 'MM': '37.111.0.0/17',
4708 'MN': '49.0.128.0/17',
4709 'MO': '60.246.0.0/16',
4710 'MP': '202.88.64.0/20',
4711 'MQ': '109.203.224.0/19',
4712 'MR': '41.188.64.0/18',
4713 'MS': '208.90.112.0/22',
4714 'MT': '46.11.0.0/16',
4715 'MU': '105.16.0.0/12',
4716 'MV': '27.114.128.0/18',
4717 'MW': '102.70.0.0/15',
4718 'MX': '187.192.0.0/11',
4719 'MY': '175.136.0.0/13',
4720 'MZ': '197.218.0.0/15',
4721 'NA': '41.182.0.0/16',
4722 'NC': '101.101.0.0/18',
4723 'NE': '197.214.0.0/18',
4724 'NF': '203.17.240.0/22',
4725 'NG': '105.112.0.0/12',
4726 'NI': '186.76.0.0/15',
4727 'NL': '145.96.0.0/11',
4728 'NO': '84.208.0.0/13',
4729 'NP': '36.252.0.0/15',
4730 'NR': '203.98.224.0/19',
4731 'NU': '49.156.48.0/22',
4732 'NZ': '49.224.0.0/14',
4733 'OM': '5.36.0.0/15',
4734 'PA': '186.72.0.0/15',
4735 'PE': '186.160.0.0/14',
4736 'PF': '123.50.64.0/18',
4737 'PG': '124.240.192.0/19',
4738 'PH': '49.144.0.0/13',
4739 'PK': '39.32.0.0/11',
4740 'PL': '83.0.0.0/11',
4741 'PM': '70.36.0.0/20',
4742 'PR': '66.50.0.0/16',
4743 'PS': '188.161.0.0/16',
4744 'PT': '85.240.0.0/13',
4745 'PW': '202.124.224.0/20',
4746 'PY': '181.120.0.0/14',
4747 'QA': '37.210.0.0/15',
4748 'RE': '102.35.0.0/16',
4749 'RO': '79.112.0.0/13',
4750 'RS': '93.86.0.0/15',
4751 'RU': '5.136.0.0/13',
4752 'RW': '41.186.0.0/16',
4753 'SA': '188.48.0.0/13',
4754 'SB': '202.1.160.0/19',
4755 'SC': '154.192.0.0/11',
4756 'SD': '102.120.0.0/13',
4757 'SE': '78.64.0.0/12',
4758 'SG': '8.128.0.0/10',
4759 'SI': '188.196.0.0/14',
4760 'SK': '78.98.0.0/15',
4761 'SL': '102.143.0.0/17',
4762 'SM': '89.186.32.0/19',
4763 'SN': '41.82.0.0/15',
4764 'SO': '154.115.192.0/18',
4765 'SR': '186.179.128.0/17',
4766 'SS': '105.235.208.0/21',
4767 'ST': '197.159.160.0/19',
4768 'SV': '168.243.0.0/16',
4769 'SX': '190.102.0.0/20',
4771 'SZ': '41.84.224.0/19',
4772 'TC': '65.255.48.0/20',
4773 'TD': '154.68.128.0/19',
4774 'TG': '196.168.0.0/14',
4775 'TH': '171.96.0.0/13',
4776 'TJ': '85.9.128.0/18',
4777 'TK': '27.96.24.0/21',
4778 'TL': '180.189.160.0/20',
4779 'TM': '95.85.96.0/19',
4780 'TN': '197.0.0.0/11',
4781 'TO': '175.176.144.0/21',
4782 'TR': '78.160.0.0/11',
4783 'TT': '186.44.0.0/15',
4784 'TV': '202.2.96.0/19',
4785 'TW': '120.96.0.0/11',
4786 'TZ': '156.156.0.0/14',
4787 'UA': '37.52.0.0/14',
4788 'UG': '102.80.0.0/13',
4790 'UY': '167.56.0.0/13',
4791 'UZ': '84.54.64.0/18',
4792 'VA': '212.77.0.0/19',
4793 'VC': '207.191.240.0/21',
4794 'VE': '186.88.0.0/13',
4795 'VG': '66.81.192.0/20',
4796 'VI': '146.226.0.0/16',
4797 'VN': '14.160.0.0/11',
4798 'VU': '202.80.32.0/20',
4799 'WF': '117.20.32.0/21',
4800 'WS': '202.4.32.0/19',
4801 'YE': '134.35.0.0/16',
4802 'YT': '41.242.116.0/22',
4803 'ZA': '41.0.0.0/11',
4804 'ZM': '102.144.0.0/13',
4805 'ZW': '102.177.192.0/18',
4809 def random_ipv4(cls, code_or_block):
4810 if len(code_or_block) == 2:
4811 block = cls._country_ip_map.get(code_or_block.upper())
4815 block = code_or_block
4816 addr, preflen = block.split('/')
4817 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4818 addr_max = addr_min | (0xffffffff >> int(preflen))
4819 return str(socket.inet_ntoa(
4820 struct.pack('!L', random.randint(addr_min, addr_max))))
4823 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4824 def __init__(self, proxies=None):
4825 # Set default handlers
4826 for type in ('http', 'https'):
4827 setattr(self, '%s_open' % type,
4828 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4829 meth(r, proxy, type))
4830 urllib.request.ProxyHandler.__init__(self, proxies)
4832 def proxy_open(self, req, proxy, type):
4833 req_proxy = req.headers.get('Ytdl-request-proxy')
4834 if req_proxy is not None:
4836 del req.headers['Ytdl-request-proxy']
4838 if proxy == '__noproxy__':
4839 return None # No Proxy
4840 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4841 req.add_header('Ytdl-socks-proxy', proxy)
4842 # yt-dlp's http/https handlers do wrapping the socket with socks
4844 return urllib.request.ProxyHandler.proxy_open(
4845 self, req, proxy, type)
4848 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4849 # released into Public Domain
4850 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4852 def long_to_bytes(n, blocksize=0):
4853 """long_to_bytes(n:long, blocksize:int) : string
4854 Convert a long integer to a byte string.
4856 If optional blocksize is given and greater than zero, pad the front of the
4857 byte string with binary zeros so that the length is a multiple of
4860 # after much testing, this algorithm was deemed to be the fastest
4864 s = struct.pack('>I', n & 0xffffffff) + s
4866 # strip off leading zeros
4867 for i in range(len(s)):
4868 if s[i] != b'\000'[0]:
4871 # only happens when n == 0
4875 # add back some pad bytes. this could be done more efficiently w.r.t. the
4876 # de-padding being done above, but sigh...
4877 if blocksize > 0 and len(s) % blocksize:
4878 s = (blocksize - len(s) % blocksize) * b'\000' + s
4882 def bytes_to_long(s):
4883 """bytes_to_long(string) : long
4884 Convert a byte string to a long integer.
4886 This is (essentially) the inverse of long_to_bytes().
4891 extra = (4 - length % 4)
4892 s = b'\000' * extra + s
4893 length = length + extra
4894 for i in range(0, length, 4):
4895 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4899 def ohdave_rsa_encrypt(data, exponent, modulus):
4901 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4904 data: data to encrypt, bytes-like object
4905 exponent, modulus: parameter e and N of RSA algorithm, both integer
4906 Output: hex string of encrypted data
4908 Limitation: supports one block encryption only
4911 payload = int(binascii.hexlify(data[::-1]), 16)
4912 encrypted = pow(payload, exponent, modulus)
4913 return '%x' % encrypted
4916 def pkcs1pad(data, length):
4918 Padding input data with PKCS#1 scheme
4920 @param {int[]} data input data
4921 @param {int} length target length
4922 @returns {int[]} padded data
4924 if len(data) > length - 11:
4925 raise ValueError('Input data too
long for PKCS
#1 padding')
4927 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4928 return [0, 2] + pseudo_random
+ [0] + data
4931 def _base_n_table(n
, table
):
4932 if not table
and not n
:
4933 raise ValueError('Either table or n must be specified')
4934 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4936 if n
and n
!= len(table
):
4937 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4941 def encode_base_n(num
, n
=None, table
=None):
4942 """Convert given int to a base-n string"""
4943 table
= _base_n_table(n
, table
)
4947 result
, base
= '', len(table
)
4949 result
= table
[num
% base
] + result
4954 def decode_base_n(string
, n
=None, table
=None):
4955 """Convert given base-n string to int"""
4956 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4957 result
, base
= 0, len(table
)
4959 result
= result
* base
+ table
[char
]
4963 def decode_base(value
, digits
):
4964 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
4965 f
'in a future version. Use {__name__}.decode_base_n instead')
4966 return decode_base_n(value
, table
=digits
)
4969 def decode_packed_codes(code
):
4970 mobj
= re
.search(PACKED_CODES_RE
, code
)
4971 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4974 symbols
= symbols
.split('|')
4979 base_n_count
= encode_base_n(count
, base
)
4980 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4983 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4987 def caesar(s
, alphabet
, shift
):
4992 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4997 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
5000 def parse_m3u8_attributes(attrib
):
5002 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5003 if val
.startswith('"'):
5009 def urshift(val
, n
):
5010 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5013 # Based on png2str() written by @gdkchan and improved by @yokrysty
5014 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5015 def decode_png(png_data
):
5016 # Reference: https://www.w3.org/TR/PNG/
5017 header
= png_data
[8:]
5019 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5020 raise OSError('Not a valid PNG file.')
5022 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5023 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5028 length
= unpack_integer(header
[:4])
5031 chunk_type
= header
[:4]
5034 chunk_data
= header
[:length
]
5035 header
= header
[length
:]
5037 header
= header
[4:] # Skip CRC
5045 ihdr
= chunks
[0]['data']
5047 width
= unpack_integer(ihdr
[:4])
5048 height
= unpack_integer(ihdr
[4:8])
5052 for chunk
in chunks
:
5053 if chunk
['type'] == b
'IDAT':
5054 idat
+= chunk
['data']
5057 raise OSError('Unable to read PNG data.')
5059 decompressed_data
= bytearray(zlib
.decompress(idat
))
5064 def _get_pixel(idx
):
5069 for y
in range(height
):
5070 basePos
= y
* (1 + stride
)
5071 filter_type
= decompressed_data
[basePos
]
5075 pixels
.append(current_row
)
5077 for x
in range(stride
):
5078 color
= decompressed_data
[1 + basePos
+ x
]
5079 basex
= y
* stride
+ x
5084 left
= _get_pixel(basex
- 3)
5086 up
= _get_pixel(basex
- stride
)
5088 if filter_type
== 1: # Sub
5089 color
= (color
+ left
) & 0xff
5090 elif filter_type
== 2: # Up
5091 color
= (color
+ up
) & 0xff
5092 elif filter_type
== 3: # Average
5093 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5094 elif filter_type
== 4: # Paeth
5100 c
= _get_pixel(basex
- stride
- 3)
5108 if pa
<= pb
and pa
<= pc
:
5109 color
= (color
+ a
) & 0xff
5111 color
= (color
+ b
) & 0xff
5113 color
= (color
+ c
) & 0xff
5115 current_row
.append(color
)
5117 return width
, height
, pixels
5120 def write_xattr(path
, key
, value
):
5121 # Windows: Write xattrs to NTFS Alternate Data Streams:
5122 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5123 if compat_os_name
== 'nt':
5124 assert ':' not in key
5125 assert os
.path
.exists(path
)
5128 with open(f
'{path}:{key}', 'wb') as f
:
5130 except OSError as e
:
5131 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5134 # UNIX Method 1. Use xattrs/pyxattrs modules
5137 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5138 # Unicode arguments are not supported in pyxattr until version 0.5.0
5139 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5140 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5141 setxattr
= xattr
.set
5143 setxattr
= xattr
.setxattr
5147 setxattr(path
, key
, value
)
5148 except OSError as e
:
5149 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5152 # UNIX Method 2. Use setfattr/xattr executables
5153 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5154 else 'xattr' if check_executable('xattr', ['-h']) else None)
5156 raise XAttrUnavailableError(
5157 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5158 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5160 value
= value
.decode()
5162 _
, stderr
, returncode
= Popen
.run(
5163 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5164 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5165 except OSError as e
:
5166 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5168 raise XAttrMetadataError(returncode
, stderr
)
5171 def random_birthday(year_field
, month_field
, day_field
):
5172 start_date
= datetime
.date(1950, 1, 1)
5173 end_date
= datetime
.date(1995, 12, 31)
5174 offset
= random
.randint(0, (end_date
- start_date
).days
)
5175 random_date
= start_date
+ datetime
.timedelta(offset
)
5177 year_field
: str(random_date
.year
),
5178 month_field
: str(random_date
.month
),
5179 day_field
: str(random_date
.day
),
5183 # Templates for internet shortcut files, which are plain text files.
5184 DOT_URL_LINK_TEMPLATE
= '''\
5189 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5190 <?xml version="1.0" encoding="UTF-8"?>
5191 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5192 <plist version="1.0">
5195 \t<string>%(url)s</string>
5200 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5210 'url': DOT_URL_LINK_TEMPLATE
,
5211 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5212 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5216 def iri_to_uri(iri
):
5218 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5220 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.
5223 iri_parts
= urllib
.parse
.urlparse(iri
)
5225 if '[' in iri_parts
.netloc
:
5226 raise ValueError('IPv6 URIs are not, yet, supported.')
5227 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5229 # 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.
5232 if iri_parts
.username
:
5233 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5234 if iri_parts
.password
is not None:
5235 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5238 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5239 # The 'idna' encoding produces ASCII text.
5240 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5241 net_location
+= ':' + str(iri_parts
.port
)
5243 return urllib
.parse
.urlunparse(
5247 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5249 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5250 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5252 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5253 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5255 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5257 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5260 def to_high_limit_path(path
):
5261 if sys
.platform
in ['win32', 'cygwin']:
5262 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5263 return '\\\\?\\' + os
.path
.abspath(path
)
5268 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5269 val
= traverse_obj(obj
, *variadic(field
))
5270 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5272 return template
% func(val
)
5275 def clean_podcast_url(url
):
5276 return re
.sub(r
'''(?x)
5280 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5283 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5286 cn\.co| # https://podcorn.com/analytics-prefix/
5287 st\.fm # https://podsights.com/docs/
5292 _HEX_TABLE
= '0123456789abcdef'
5295 def random_uuidv4():
5296 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5299 def make_dir(path
, to_screen
=None):
5301 dn
= os
.path
.dirname(path
)
5302 if dn
and not os
.path
.exists(dn
):
5305 except OSError as err
:
5306 if callable(to_screen
) is not None:
5307 to_screen('unable to create directory ' + error_to_compat_str(err
))
5311 def get_executable_path():
5312 from .update
import _get_variant_and_executable_path
5314 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5317 def load_plugins(name
, suffix
, namespace
):
5319 with contextlib
.suppress(FileNotFoundError
):
5320 plugins_spec
= importlib
.util
.spec_from_file_location(
5321 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5322 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5323 sys
.modules
[plugins_spec
.name
] = plugins
5324 plugins_spec
.loader
.exec_module(plugins
)
5325 for name
in dir(plugins
):
5326 if name
in namespace
:
5328 if not name
.endswith(suffix
):
5330 klass
= getattr(plugins
, name
)
5331 classes
[name
] = namespace
[name
] = klass
5336 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5337 casesense
=True, is_user_input
=False, traverse_string
=False):
5339 Safely traverse nested `dict`s and `Sequence`s
5341 >>> obj = [{}, {"key": "value"}]
5342 >>> traverse_obj(obj, (1, "key"))
5345 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5346 The next path will also be tested if the path branched but no results could be found.
5347 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5348 A value of None is treated as the absence of a value.
5350 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5352 The keys in the path can be one of:
5353 - `None`: Return the current object.
5354 - `str`/`int`: Return `obj[key]`. For `re.Match, return `obj.group(key)`.
5355 - `slice`: Branch out and return all values in `obj[key]`.
5356 - `Ellipsis`: Branch out and return a list of all values.
5357 - `tuple`/`list`: Branch out and return a list of all matching values.
5358 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5359 - `function`: Branch out and return values filtered by the function.
5360 Read as: `[value for key, value in obj if function(key, value)]`.
5361 For `Sequence`s, `key` is the index of the value.
5362 - `dict` Transform the current object and return a matching dict.
5363 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5365 `tuple`, `list`, and `dict` all support nested paths and branches.
5367 @params paths Paths which to traverse by.
5368 @param default Value to return if the paths do not match.
5369 @param expected_type If a `type`, only accept final values of this type.
5370 If any other callable, try to call the function on each result.
5371 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5372 @param casesense If `False`, consider string dictionary keys as case insensitive.
5374 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5376 @param is_user_input Whether the keys are generated from user input.
5377 If `True` strings get converted to `int`/`slice` if needed.
5378 @param traverse_string Whether to traverse into objects as strings.
5379 If `True`, any non-compatible object will first be
5380 converted into a string and then traversed into.
5383 @returns The result of the object traversal.
5384 If successful, `get_all=True`, and the path branches at least once,
5385 then a list of results is returned instead.
5386 A list is always returned if the last path branches and no `default` is given.
5388 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5389 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5391 if isinstance(expected_type
, type):
5392 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5394 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5396 def apply_key(key
, obj
):
5403 elif isinstance(key
, (list, tuple)):
5405 _
, result
= apply_path(obj
, branch
)
5409 if isinstance(obj
, collections
.abc
.Mapping
):
5410 yield from obj
.values()
5411 elif is_sequence(obj
):
5413 elif isinstance(obj
, re
.Match
):
5414 yield from obj
.groups()
5415 elif traverse_string
:
5419 if is_sequence(obj
):
5420 iter_obj
= enumerate(obj
)
5421 elif isinstance(obj
, collections
.abc
.Mapping
):
5422 iter_obj
= obj
.items()
5423 elif isinstance(obj
, re
.Match
):
5424 iter_obj
= enumerate((obj
.group(), *obj
.groups()))
5425 elif traverse_string
:
5426 iter_obj
= enumerate(str(obj
))
5429 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5431 elif isinstance(key
, dict):
5432 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5433 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5434 if v
is not None or default
is not NO_DEFAULT
}
5436 elif isinstance(obj
, collections
.abc
.Mapping
):
5437 yield (obj
.get(key
) if casesense
or (key
in obj
)
5438 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5440 elif isinstance(obj
, re
.Match
):
5441 if isinstance(key
, int) or casesense
:
5442 with contextlib
.suppress(IndexError):
5443 yield obj
.group(key
)
5446 if not isinstance(key
, str):
5449 yield next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5453 key
= (int_or_none(key
) if ':' not in key
5454 else slice(*map(int_or_none
, key
.split(':'))))
5456 if not isinstance(key
, (int, slice)):
5459 if not is_sequence(obj
):
5460 if not traverse_string
:
5464 with contextlib
.suppress(IndexError):
5467 def apply_path(start_obj
, path
):
5469 has_branched
= False
5471 for key
in variadic(path
):
5472 if is_user_input
and key
== ':':
5475 if not casesense
and isinstance(key
, str):
5476 key
= key
.casefold()
5478 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5481 key_func
= functools
.partial(apply_key
, key
)
5482 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5484 return has_branched
, objs
5486 def _traverse_obj(obj
, path
, use_list
=True):
5487 has_branched
, results
= apply_path(obj
, path
)
5488 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5490 if get_all
and has_branched
:
5491 return results
.exhaust() if results
or use_list
else None
5493 return results
[0] if results
else None
5495 for index
, path
in enumerate(paths
, 1):
5496 use_list
= default
is NO_DEFAULT
and index
== len(paths
)
5497 result
= _traverse_obj(obj
, path
, use_list
)
5498 if result
is not None:
5501 return None if default
is NO_DEFAULT
else default
5504 def traverse_dict(dictn
, keys
, casesense
=True):
5505 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5506 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5507 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5510 def get_first(obj
, keys
, **kwargs
):
5511 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5514 def time_seconds(**kwargs
):
5515 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5516 return t
.timestamp()
5519 # create a JSON Web Signature (jws) with HS256 algorithm
5520 # the resulting format is in JWS Compact Serialization
5521 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5522 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5523 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5529 header_data
.update(headers
)
5530 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5531 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5532 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5533 signature_b64
= base64
.b64encode(h
.digest())
5534 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5538 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5539 def jwt_decode_hs256(jwt
):
5540 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5541 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5542 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5546 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5550 def supports_terminal_sequences(stream
):
5551 if compat_os_name
== 'nt':
5552 if not WINDOWS_VT_MODE
:
5554 elif not os
.getenv('TERM'):
5557 return stream
.isatty()
5558 except BaseException
:
5562 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5563 if get_windows_version() < (10, 0, 10586):
5565 global WINDOWS_VT_MODE
5567 Popen
.run('', shell
=True)
5571 WINDOWS_VT_MODE
= True
5572 supports_terminal_sequences
.cache_clear()
5575 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5578 def remove_terminal_sequences(string
):
5579 return _terminal_sequences_re
.sub('', string
)
5582 def number_of_digits(number
):
5583 return len('%d' % number
)
5586 def join_nonempty(*values
, delim
='-', from_dict
=None):
5587 if from_dict
is not None:
5588 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5589 return delim
.join(map(str, filter(None, values
)))
5592 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5594 Find the largest format dimensions in terms of video width and, for each thumbnail:
5595 * Modify the URL: Match the width with the provided regex and replace with the former width
5598 This function is useful with video services that scale the provided thumbnails on demand
5600 _keys
= ('width', 'height')
5601 max_dimensions
= max(
5602 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5604 if not max_dimensions
[0]:
5608 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5609 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5610 for thumbnail
in thumbnails
5614 def parse_http_range(range):
5615 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5617 return None, None, None
5618 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5620 return None, None, None
5621 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5624 def read_stdin(what
):
5625 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5626 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5630 def determine_file_encoding(data
):
5632 Detect the text encoding used
5633 @returns (encoding, bytes to skip)
5636 # BOM marks are given priority over declarations
5637 for bom
, enc
in BOMS
:
5638 if data
.startswith(bom
):
5639 return enc
, len(bom
)
5641 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5642 # We ignore the endianness to get a good enough match
5643 data
= data
.replace(b
'\0', b
'')
5644 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5645 return mobj
.group(1).decode() if mobj
else None, 0
5652 __initialized
= False
5654 def __init__(self
, parser
, label
=None):
5655 self
.parser
, self
.label
= parser
, label
5656 self
._loaded
_paths
, self
.configs
= set(), []
5658 def init(self
, args
=None, filename
=None):
5659 assert not self
.__initialized
5660 self
.own_args
, self
.filename
= args
, filename
5661 return self
.load_configs()
5663 def load_configs(self
):
5666 location
= os
.path
.realpath(self
.filename
)
5667 directory
= os
.path
.dirname(location
)
5668 if location
in self
._loaded
_paths
:
5670 self
._loaded
_paths
.add(location
)
5672 self
.__initialized
= True
5673 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5674 self
.parsed_args
= self
.own_args
5675 for location
in opts
.config_locations
or []:
5677 if location
in self
._loaded
_paths
:
5679 self
._loaded
_paths
.add(location
)
5680 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5682 location
= os
.path
.join(directory
, expand_path(location
))
5683 if os
.path
.isdir(location
):
5684 location
= os
.path
.join(location
, 'yt-dlp.conf')
5685 if not os
.path
.exists(location
):
5686 self
.parser
.error(f
'config location {location} does not exist')
5687 self
.append_config(self
.read_file(location
), location
)
5691 label
= join_nonempty(
5692 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5694 return join_nonempty(
5695 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5696 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5700 def read_file(filename
, default
=[]):
5702 optionf
= open(filename
, 'rb')
5704 return default
# silently skip if file is not present
5706 enc
, skip
= determine_file_encoding(optionf
.read(512))
5707 optionf
.seek(skip
, io
.SEEK_SET
)
5709 enc
= None # silently skip read errors
5711 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5712 contents
= optionf
.read().decode(enc
or preferredencoding())
5713 res
= shlex
.split(contents
, comments
=True)
5714 except Exception as err
:
5715 raise ValueError(f
'Unable to parse "{filename}": {err}')
5721 def hide_login_info(opts
):
5722 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5723 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5728 return m
.group('key') + '=PRIVATE'
5732 opts
= list(map(_scrub_eq
, opts
))
5733 for idx
, opt
in enumerate(opts
):
5734 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5735 opts
[idx
+ 1] = 'PRIVATE'
5738 def append_config(self
, *args
, label
=None):
5739 config
= type(self
)(self
.parser
, label
)
5740 config
._loaded
_paths
= self
._loaded
_paths
5741 if config
.init(*args
):
5742 self
.configs
.append(config
)
5746 for config
in reversed(self
.configs
):
5747 yield from config
.all_args
5748 yield from self
.parsed_args
or []
5750 def parse_known_args(self
, **kwargs
):
5751 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5753 def parse_args(self
):
5754 return self
.parser
.parse_args(self
.all_args
)
5757 class WebSocketsWrapper
:
5758 """Wraps websockets module to use in non-async scopes"""
5761 def __init__(self
, url
, headers
=None, connect
=True):
5762 self
.loop
= asyncio
.new_event_loop()
5763 # XXX: "loop" is deprecated
5764 self
.conn
= websockets
.connect(
5765 url
, extra_headers
=headers
, ping_interval
=None,
5766 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5769 atexit
.register(self
.__exit
__, None, None, None)
5771 def __enter__(self
):
5773 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5776 def send(self
, *args
):
5777 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5779 def recv(self
, *args
):
5780 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5782 def __exit__(self
, type, value
, traceback
):
5784 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5787 self
._cancel
_all
_tasks
(self
.loop
)
5789 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5790 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5792 def run_with_loop(main
, loop
):
5793 if not asyncio
.iscoroutine(main
):
5794 raise ValueError(f
'a coroutine was expected, got {main!r}')
5797 return loop
.run_until_complete(main
)
5799 loop
.run_until_complete(loop
.shutdown_asyncgens())
5800 if hasattr(loop
, 'shutdown_default_executor'):
5801 loop
.run_until_complete(loop
.shutdown_default_executor())
5804 def _cancel_all_tasks(loop
):
5805 to_cancel
= asyncio
.all_tasks(loop
)
5810 for task
in to_cancel
:
5813 # XXX: "loop" is removed in python 3.10+
5814 loop
.run_until_complete(
5815 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5817 for task
in to_cancel
:
5818 if task
.cancelled():
5820 if task
.exception() is not None:
5821 loop
.call_exception_handler({
5822 'message': 'unhandled exception during asyncio.run() shutdown',
5823 'exception': task
.exception(),
5828 def merge_headers(*dicts
):
5829 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5830 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5833 def cached_method(f
):
5834 """Cache a method"""
5835 signature
= inspect
.signature(f
)
5838 def wrapper(self
, *args
, **kwargs
):
5839 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5840 bound_args
.apply_defaults()
5841 key
= tuple(bound_args
.arguments
.values())[1:]
5843 cache
= vars(self
).setdefault('__cached_method__cache', {}).setdefault(f.__name__, {}
)
5844 if key
not in cache
:
5845 cache
[key
] = f(self
, *args
, **kwargs
)
5850 class classproperty
:
5851 """property access for class methods"""
5853 def __init__(self
, func
):
5854 functools
.update_wrapper(self
, func
)
5857 def __get__(self
, _
, cls
):
5858 return self
.func(cls
)
5861 class Namespace(types
.SimpleNamespace
):
5862 """Immutable namespace"""
5865 return iter(self
.__dict
__.values())
5869 return self
.__dict
__.items()
5872 MEDIA_EXTENSIONS
= Namespace(
5873 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5874 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5875 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5876 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5877 thumbnails
=('jpg', 'png', 'webp'),
5878 storyboards
=('mhtml', ),
5879 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5880 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5882 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5883 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5885 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5890 for retry in RetryManager(...):
5893 except SomeException as err:
5897 attempt
, _error
= 0, None
5899 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5900 self
.retries
= _retries
or 0
5901 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5903 def _should_retry(self
):
5904 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5908 if self
._error
is NO_DEFAULT
:
5913 def error(self
, value
):
5917 while self
._should
_retry
():
5918 self
.error
= NO_DEFAULT
5922 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5925 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5926 """Utility function for reporting retries"""
5929 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5934 elif isinstance(e
, ExtractorError
):
5935 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5936 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5938 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5940 info(f
'Sleeping {delay:.2f} seconds ...')
5944 def make_archive_id(ie
, video_id
):
5945 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5946 return f
'{ie_key.lower()} {video_id}'
5949 def truncate_string(s
, left
, right
=0):
5950 assert left
> 3 and right
>= 0
5951 if s
is None or len(s
) <= left
+ right
:
5953 return f
'{s[:left-3]}...{s[-right:]}'
5956 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5957 assert 'all' in alias_dict
, '"all" alias is required'
5958 requested
= list(start
or [])
5960 discard
= val
.startswith('-')
5964 if val
in alias_dict
:
5965 val
= alias_dict
[val
] if not discard
else [
5966 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5967 # NB: Do not allow regex in aliases for performance
5968 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5971 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5972 else [val
] if val
in alias_dict
['all'] else None)
5974 raise ValueError(val
)
5977 for item
in current
:
5978 while item
in requested
:
5979 requested
.remove(item
)
5981 requested
.extend(current
)
5983 return orderedSet(requested
)
5987 has_certifi
= bool(certifi
)
5988 has_websockets
= bool(websockets
)