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'],
154 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
156 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
157 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
158 'EST': -5, 'EDT': -4, # Eastern
159 'CST': -6, 'CDT': -5, # Central
160 'MST': -7, 'MDT': -6, # Mountain
161 'PST': -8, 'PDT': -7 # Pacific
164 # needed for sanitizing filenames in restricted mode
165 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
166 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
167 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
197 '%Y-%m-%d %H:%M:%S.%f',
198 '%Y-%m-%d %H:%M:%S:%f',
201 '%Y-%m-%dT%H:%M:%SZ',
202 '%Y-%m-%dT%H:%M:%S.%fZ',
203 '%Y-%m-%dT%H:%M:%S.%f0Z',
205 '%Y-%m-%dT%H:%M:%S.%f',
208 '%b %d %Y at %H:%M:%S',
210 '%B %d %Y at %H:%M:%S',
214 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
215 DATE_FORMATS_DAY_FIRST
.extend([
225 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
226 DATE_FORMATS_MONTH_FIRST
.extend([
234 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
235 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
237 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
241 def preferredencoding():
242 """Get preferred encoding.
244 Returns the best encoding scheme for the system, based on
245 locale.getpreferredencoding() and some further tweaks.
248 pref = locale.getpreferredencoding()
256 def write_json_file(obj, fn):
257 """ Encode obj as JSON and write it to fn, atomically if possible """
259 tf = tempfile.NamedTemporaryFile(
260 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
261 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
265 json.dump(obj, tf, ensure_ascii=False)
266 if sys.platform == 'win32
':
267 # Need to remove existing file on Windows, else os.rename raises
268 # WindowsError or FileExistsError.
269 with contextlib.suppress(OSError):
271 with contextlib.suppress(OSError):
274 os.chmod(tf.name, 0o666 & ~mask)
275 os.rename(tf.name, fn)
277 with contextlib.suppress(OSError):
282 def find_xpath_attr(node, xpath, key, val=None):
283 """ Find the xpath xpath[@key=val] """
284 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
285 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
286 return node.find(expr)
288 # On python2.6 the xml.etree.ElementTree.Element methods don't support
289 # the namespace parameter
292 def xpath_with_ns(path
, ns_map
):
293 components
= [c
.split(':') for c
in path
.split('/')]
297 replaced
.append(c
[0])
300 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
301 return '/'.join(replaced
)
304 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
305 def _find_xpath(xpath
):
306 return node
.find(xpath
)
308 if isinstance(xpath
, str):
309 n
= _find_xpath(xpath
)
317 if default
is not NO_DEFAULT
:
320 name
= xpath
if name
is None else name
321 raise ExtractorError('Could not find XML element %s' % name
)
327 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
328 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
329 if n
is None or n
== default
:
332 if default
is not NO_DEFAULT
:
335 name
= xpath
if name
is None else name
336 raise ExtractorError('Could not find XML element\'s text %s' % name
)
342 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
343 n
= find_xpath_attr(node
, xpath
, key
)
345 if default
is not NO_DEFAULT
:
348 name
= f
'{xpath}[@{key}]' if name
is None else name
349 raise ExtractorError('Could not find XML attribute %s' % name
)
355 def get_element_by_id(id, html
, **kwargs
):
356 """Return the content of the tag with the specified ID in the passed HTML document"""
357 return get_element_by_attribute('id', id, html
, **kwargs
)
360 def get_element_html_by_id(id, html
, **kwargs
):
361 """Return the html of the tag with the specified ID in the passed HTML document"""
362 return get_element_html_by_attribute('id', id, html
, **kwargs
)
365 def get_element_by_class(class_name
, html
):
366 """Return the content of the first tag with the specified class in the passed HTML document"""
367 retval
= get_elements_by_class(class_name
, html
)
368 return retval
[0] if retval
else None
371 def get_element_html_by_class(class_name
, html
):
372 """Return the html of the first tag with the specified class in the passed HTML document"""
373 retval
= get_elements_html_by_class(class_name
, html
)
374 return retval
[0] if retval
else None
377 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
378 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
379 return retval
[0] if retval
else None
382 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
383 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
384 return retval
[0] if retval
else None
387 def get_elements_by_class(class_name
, html
, **kargs
):
388 """Return the content of all tags with the specified class in the passed HTML document as a list"""
389 return get_elements_by_attribute(
390 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
391 html, escape_value=False)
394 def get_elements_html_by_class(class_name, html):
395 """Return the html of all tags with the specified class in the passed HTML document as a list"""
396 return get_elements_html_by_attribute(
397 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
398 html, escape_value=False)
401 def get_elements_by_attribute(*args, **kwargs):
402 """Return the content of the tag with the specified attribute in the passed HTML document"""
403 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
406 def get_elements_html_by_attribute(*args, **kwargs):
407 """Return the html of the tag with the specified attribute in the passed HTML document"""
408 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
411 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
413 Return the text (content) and the html (whole) of the tag with the specified
414 attribute in the passed HTML document
417 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
419 value = re.escape(value) if escape_value else value
421 partial_element_re = rf'''(?x
)
422 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
423 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
424 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
427 for m in re.finditer(partial_element_re, html):
428 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
431 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
436 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
438 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
439 closing tag for the first opening tag it has encountered, and can be used
443 class HTMLBreakOnClosingTagException(Exception):
447 self.tagstack = collections.deque()
448 html.parser.HTMLParser.__init__(self)
453 def __exit__(self, *_):
457 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
458 # so data remains buffered; we no longer have any interest in it, thus
459 # override this method to discard it
462 def handle_starttag(self, tag, _):
463 self.tagstack.append(tag)
465 def handle_endtag(self, tag):
466 if not self.tagstack:
467 raise compat_HTMLParseError('no tags
in the stack
')
469 inner_tag = self.tagstack.pop()
473 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
474 if not self.tagstack:
475 raise self.HTMLBreakOnClosingTagException()
478 def get_element_text_and_html_by_tag(tag, html):
480 For the first element with the specified tag in the passed HTML document
481 return its' content (text
) and the whole
element (html
)
483 def find_or_raise(haystack, needle, exc):
485 return haystack.index(needle)
488 closing_tag = f'</{tag}>'
489 whole_start = find_or_raise(
490 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
491 content_start = find_or_raise(
492 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
493 content_start += whole_start + 1
494 with HTMLBreakOnClosingTagParser() as parser:
495 parser.feed(html[whole_start:content_start])
496 if not parser.tagstack or parser.tagstack[0] != tag:
497 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
498 offset = content_start
499 while offset < len(html):
500 next_closing_tag_start = find_or_raise(
501 html[offset:], closing_tag,
502 compat_HTMLParseError(f'closing {tag} tag not found'))
503 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
505 parser.feed(html[offset:offset + next_closing_tag_end])
506 offset += next_closing_tag_end
507 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
508 return html[content_start:offset + next_closing_tag_start], \
509 html[whole_start:offset + next_closing_tag_end]
510 raise compat_HTMLParseError('unexpected end of html')
513 class HTMLAttributeParser(html.parser.HTMLParser):
514 """Trivial HTML parser to gather the attributes
for a single element
"""
518 html.parser.HTMLParser.__init__(self)
520 def handle_starttag(self, tag, attrs):
521 self.attrs = dict(attrs)
524 class HTMLListAttrsParser(html.parser.HTMLParser):
525 """HTML parser to gather the attributes
for the elements of a
list"""
528 html.parser.HTMLParser.__init__(self)
532 def handle_starttag(self, tag, attrs):
533 if tag == 'li' and self._level == 0:
534 self.items.append(dict(attrs))
537 def handle_endtag(self, tag):
541 def extract_attributes(html_element):
542 """Given a string
for an HTML element such
as
544 a
="foo" B
="bar" c
="&98;az" d
=boz
545 empty
= noval entity
="&"
548 Decode
and return a dictionary of attributes
.
550 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
551 'empty': '', 'noval': None, 'entity': '&',
552 'sq': '"', 'dq': '\''
555 parser = HTMLAttributeParser()
556 with contextlib.suppress(compat_HTMLParseError):
557 parser.feed(html_element)
562 def parse_list(webpage):
563 """Given a string
for an series of HTML
<li
> elements
,
564 return a dictionary of their attributes
"""
565 parser = HTMLListAttrsParser()
571 def clean_html(html):
572 """Clean an HTML snippet into a readable string
"""
574 if html is None: # Convenience for sanitizing descriptions etc.
577 html = re.sub(r'\s+', ' ', html)
578 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
579 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
581 html = re.sub('<.*?>', '', html)
582 # Replace html entities
583 html = unescapeHTML(html)
587 class LenientJSONDecoder(json.JSONDecoder):
588 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
589 self.transform_source, self.ignore_extra = transform_source, ignore_extra
590 super().__init__(*args, **kwargs)
593 if self.transform_source:
594 s = self.transform_source(s)
596 if self.ignore_extra:
597 return self.raw_decode(s.lstrip())[0]
598 return super().decode(s)
599 except json.JSONDecodeError as e:
600 if e.pos is not None:
601 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
605 def sanitize_open(filename, open_mode):
606 """Try to
open the given filename
, and slightly tweak it
if this fails
.
608 Attempts to
open the given filename
. If this fails
, it tries to change
609 the filename slightly
, step by step
, until it
's either able to open it
610 or it fails and raises a final exception, like the standard open()
613 It returns the tuple (stream, definitive_file_name).
616 if sys.platform == 'win32
':
619 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
620 with contextlib.suppress(io.UnsupportedOperation):
621 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
622 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
624 for attempt in range(2):
627 if sys.platform == 'win32
':
628 # FIXME: An exclusive lock also locks the file from being read.
629 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
630 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
631 raise LockingUnsupportedError()
632 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
634 stream
= open(filename
, open_mode
)
635 return stream
, filename
636 except OSError as err
:
637 if attempt
or err
.errno
in (errno
.EACCES
,):
639 old_filename
, filename
= filename
, sanitize_path(filename
)
640 if old_filename
== filename
:
644 def timeconvert(timestr
):
645 """Convert RFC 2822 defined time string into system timestamp"""
647 timetuple
= email
.utils
.parsedate_tz(timestr
)
648 if timetuple
is not None:
649 timestamp
= email
.utils
.mktime_tz(timetuple
)
653 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
654 """Sanitizes a string so it could be used as part of a filename.
655 @param restricted Use a stricter subset of allowed characters
656 @param is_id Whether this is an ID that should be kept unchanged if possible.
657 If unset, yt-dlp's new sanitization rules are in effect
662 def replace_insane(char
):
663 if restricted
and char
in ACCENT_CHARS
:
664 return ACCENT_CHARS
[char
]
665 elif not restricted
and char
== '\n':
667 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
668 # Replace with their full-width unicode counterparts
669 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
670 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
673 return '' if restricted
else '\''
675 return '\0_\0-' if restricted
else '\0 \0-'
676 elif char
in '\\/|*<>':
678 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
682 if restricted
and is_id
is NO_DEFAULT
:
683 s
= unicodedata
.normalize('NFKC', s
)
684 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
685 result
= ''.join(map(replace_insane
, s
))
686 if is_id
is NO_DEFAULT
:
687 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
688 STRIP_RE
= r
'(?:\0.|[ _-])*'
689 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
690 result
= result
.replace('\0', '') or '_'
693 while '__' in result
:
694 result
= result
.replace('__', '_')
695 result
= result
.strip('_')
696 # Common case of "Foreign band name - English song title"
697 if restricted
and result
.startswith('-_'):
699 if result
.startswith('-'):
700 result
= '_' + result
[len('-'):]
701 result
= result
.lstrip('.')
707 def sanitize_path(s
, force
=False):
708 """Sanitizes and normalizes path on Windows"""
709 if sys
.platform
== 'win32':
711 drive_or_unc
, _
= os
.path
.splitdrive(s
)
717 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
721 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
722 for path_part
in norm_path
]
724 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
725 elif force
and s
and s
[0] == os
.path
.sep
:
726 sanitized_path
.insert(0, os
.path
.sep
)
727 return os
.path
.join(*sanitized_path
)
730 def sanitize_url(url
, *, scheme
='http'):
731 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
732 # the number of unwanted failures due to missing protocol
735 elif url
.startswith('//'):
736 return f
'{scheme}:{url}'
737 # Fix some common typos seen so far
739 # https://github.com/ytdl-org/youtube-dl/issues/15649
740 (r
'^httpss://', r
'https://'),
741 # https://bx1.be/lives/direct-tv/
742 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
744 for mistake
, fixup
in COMMON_TYPOS
:
745 if re
.match(mistake
, url
):
746 return re
.sub(mistake
, fixup
, url
)
750 def extract_basic_auth(url
):
751 parts
= urllib
.parse
.urlsplit(url
)
752 if parts
.username
is None:
754 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
755 parts
.hostname
if parts
.port
is None
756 else '%s:%d' % (parts
.hostname
, parts
.port
))))
757 auth_payload
= base64
.b64encode(
758 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
759 return url
, f
'Basic {auth_payload.decode()}'
762 def sanitized_Request(url
, *args
, **kwargs
):
763 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
764 if auth_header
is not None:
765 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
766 headers
['Authorization'] = auth_header
767 return urllib
.request
.Request(url
, *args
, **kwargs
)
771 """Expand shell variables and ~"""
772 return os
.path
.expandvars(compat_expanduser(s
))
775 def orderedSet(iterable
, *, lazy
=False):
776 """Remove all duplicates from the input iterable"""
778 seen
= [] # Do not use set since the items can be unhashable
784 return _iter() if lazy
else list(_iter())
787 def _htmlentity_transform(entity_with_semicolon
):
788 """Transforms an HTML entity to a character."""
789 entity
= entity_with_semicolon
[:-1]
791 # Known non-numeric HTML entity
792 if entity
in html
.entities
.name2codepoint
:
793 return chr(html
.entities
.name2codepoint
[entity
])
795 # TODO: HTML5 allows entities without a semicolon.
796 # E.g. 'Éric' should be decoded as 'Éric'.
797 if entity_with_semicolon
in html
.entities
.html5
:
798 return html
.entities
.html5
[entity_with_semicolon
]
800 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
802 numstr
= mobj
.group(1)
803 if numstr
.startswith('x'):
805 numstr
= '0%s' % numstr
808 # See https://github.com/ytdl-org/youtube-dl/issues/7518
809 with contextlib
.suppress(ValueError):
810 return chr(int(numstr
, base
))
812 # Unknown entity in name, return its literal representation
813 return '&%s;' % entity
819 assert isinstance(s
, str)
822 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
825 def escapeHTML(text
):
828 .replace('&', '&')
829 .replace('<', '<')
830 .replace('>', '>')
831 .replace('"', '"')
832 .replace("'", ''')
836 def process_communicate_or_kill(p
, *args
, **kwargs
):
837 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
838 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
839 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
842 class Popen(subprocess
.Popen
):
843 if sys
.platform
== 'win32':
844 _startupinfo
= subprocess
.STARTUPINFO()
845 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
850 def _fix_pyinstaller_ld_path(env
):
851 """Restore LD_LIBRARY_PATH when using PyInstaller
852 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
853 https://github.com/yt-dlp/yt-dlp/issues/4573
855 if not hasattr(sys
, '_MEIPASS'):
859 orig
= env
.get(f
'{key}_ORIG')
865 _fix('LD_LIBRARY_PATH') # Linux
866 _fix('DYLD_LIBRARY_PATH') # macOS
868 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
870 env
= os
.environ
.copy()
871 self
._fix
_pyinstaller
_ld
_path
(env
)
874 kwargs
['universal_newlines'] = True # For 3.6 compatibility
875 kwargs
.setdefault('encoding', 'utf-8')
876 kwargs
.setdefault('errors', 'replace')
877 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
879 def communicate_or_kill(self
, *args
, **kwargs
):
881 return self
.communicate(*args
, **kwargs
)
882 except BaseException
: # Including KeyboardInterrupt
883 self
.kill(timeout
=None)
886 def kill(self
, *, timeout
=0):
889 self
.wait(timeout
=timeout
)
892 def run(cls
, *args
, timeout
=None, **kwargs
):
893 with cls(*args
, **kwargs
) as proc
:
894 default
= '' if proc
.text_mode
else b
''
895 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
896 return stdout
or default
, stderr
or default
, proc
.returncode
899 def get_subprocess_encoding():
900 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
901 # For subprocess calls, encode with locale encoding
902 # Refer to http://stackoverflow.com/a/9951851/35070
903 encoding
= preferredencoding()
905 encoding
= sys
.getfilesystemencoding()
911 def encodeFilename(s
, for_subprocess
=False):
912 assert isinstance(s
, str)
916 def decodeFilename(b
, for_subprocess
=False):
920 def encodeArgument(s
):
921 # Legacy code that uses byte strings
922 # Uncomment the following line after fixing all post processors
923 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
924 return s
if isinstance(s
, str) else s
.decode('ascii')
927 def decodeArgument(b
):
931 def decodeOption(optval
):
934 if isinstance(optval
, bytes):
935 optval
= optval
.decode(preferredencoding())
937 assert isinstance(optval
, str)
941 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
944 def timetuple_from_msec(msec
):
945 secs
, msec
= divmod(msec
, 1000)
946 mins
, secs
= divmod(secs
, 60)
947 hrs
, mins
= divmod(mins
, 60)
948 return _timetuple(hrs
, mins
, secs
, msec
)
951 def formatSeconds(secs
, delim
=':', msec
=False):
952 time
= timetuple_from_msec(secs
* 1000)
954 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
956 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
958 ret
= '%d' % time
.seconds
959 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
962 def _ssl_load_windows_store_certs(ssl_context
, storename
):
963 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
965 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
966 if encoding
== 'x509_asn' and (
967 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
968 except PermissionError
:
971 with contextlib
.suppress(ssl
.SSLError
):
972 ssl_context
.load_verify_locations(cadata
=cert
)
975 def make_HTTPS_handler(params
, **kwargs
):
976 opts_check_certificate
= not params
.get('nocheckcertificate')
977 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
978 context
.check_hostname
= opts_check_certificate
979 if params
.get('legacyserverconnect'):
980 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
981 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
982 context
.set_ciphers('DEFAULT')
984 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
985 if opts_check_certificate
:
986 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
987 context
.load_verify_locations(cafile
=certifi
.where())
990 context
.load_default_certs()
991 # Work around the issue in load_default_certs when there are bad certificates. See:
992 # https://github.com/yt-dlp/yt-dlp/issues/1060,
993 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
995 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
996 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
997 for storename
in ('CA', 'ROOT'):
998 _ssl_load_windows_store_certs(context
, storename
)
999 context
.set_default_verify_paths()
1001 client_certfile
= params
.get('client_certificate')
1004 context
.load_cert_chain(
1005 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1006 password
=params
.get('client_certificate_password'))
1007 except ssl
.SSLError
:
1008 raise YoutubeDLError('Unable to load client certificate')
1010 # Some servers may reject requests if ALPN extension is not sent. See:
1011 # https://github.com/python/cpython/issues/85140
1012 # https://github.com/yt-dlp/yt-dlp/issues/3878
1013 with contextlib
.suppress(NotImplementedError):
1014 context
.set_alpn_protocols(['http/1.1'])
1016 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1019 def bug_reports_message(before
=';'):
1020 from .update
import REPOSITORY
1022 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1023 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1025 before
= before
.rstrip()
1026 if not before
or before
.endswith(('.', '!', '?')):
1027 msg
= msg
[0].title() + msg
[1:]
1029 return (before
+ ' ' if before
else '') + msg
1032 class YoutubeDLError(Exception):
1033 """Base exception for YoutubeDL errors."""
1036 def __init__(self
, msg
=None):
1039 elif self
.msg
is None:
1040 self
.msg
= type(self
).__name
__
1041 super().__init
__(self
.msg
)
1044 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1045 if hasattr(ssl
, 'CertificateError'):
1046 network_exceptions
.append(ssl
.CertificateError
)
1047 network_exceptions
= tuple(network_exceptions
)
1050 class ExtractorError(YoutubeDLError
):
1051 """Error during info extraction."""
1053 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1054 """ tb, if given, is the original traceback (so that it can be printed out).
1055 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1057 if sys
.exc_info()[0] in network_exceptions
:
1060 self
.orig_msg
= str(msg
)
1062 self
.expected
= expected
1064 self
.video_id
= video_id
1066 self
.exc_info
= sys
.exc_info() # preserve original exception
1067 if isinstance(self
.exc_info
[1], ExtractorError
):
1068 self
.exc_info
= self
.exc_info
[1].exc_info
1070 super().__init
__(''.join((
1071 format_field(ie
, None, '[%s] '),
1072 format_field(video_id
, None, '%s: '),
1074 format_field(cause
, None, ' (caused by %r)'),
1075 '' if expected
else bug_reports_message())))
1077 def format_traceback(self
):
1078 return join_nonempty(
1079 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1080 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1084 class UnsupportedError(ExtractorError
):
1085 def __init__(self
, url
):
1087 'Unsupported URL: %s' % url
, expected
=True)
1091 class RegexNotFoundError(ExtractorError
):
1092 """Error when a regex didn't match"""
1096 class GeoRestrictedError(ExtractorError
):
1097 """Geographic restriction Error exception.
1099 This exception may be thrown when a video is not available from your
1100 geographic location due to geographic restrictions imposed by a website.
1103 def __init__(self
, msg
, countries
=None, **kwargs
):
1104 kwargs
['expected'] = True
1105 super().__init
__(msg
, **kwargs
)
1106 self
.countries
= countries
1109 class UserNotLive(ExtractorError
):
1110 """Error when a channel/user is not live"""
1112 def __init__(self
, msg
=None, **kwargs
):
1113 kwargs
['expected'] = True
1114 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1117 class DownloadError(YoutubeDLError
):
1118 """Download Error exception.
1120 This exception may be thrown by FileDownloader objects if they are not
1121 configured to continue on errors. They will contain the appropriate
1125 def __init__(self
, msg
, exc_info
=None):
1126 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1127 super().__init
__(msg
)
1128 self
.exc_info
= exc_info
1131 class EntryNotInPlaylist(YoutubeDLError
):
1132 """Entry not in playlist exception.
1134 This exception will be thrown by YoutubeDL when a requested entry
1135 is not found in the playlist info_dict
1137 msg
= 'Entry not found in info'
1140 class SameFileError(YoutubeDLError
):
1141 """Same File exception.
1143 This exception will be thrown by FileDownloader objects if they detect
1144 multiple files would have to be downloaded to the same file on disk.
1146 msg
= 'Fixed output name but more than one file to download'
1148 def __init__(self
, filename
=None):
1149 if filename
is not None:
1150 self
.msg
+= f
': {filename}'
1151 super().__init
__(self
.msg
)
1154 class PostProcessingError(YoutubeDLError
):
1155 """Post Processing exception.
1157 This exception may be raised by PostProcessor's .run() method to
1158 indicate an error in the postprocessing task.
1162 class DownloadCancelled(YoutubeDLError
):
1163 """ Exception raised when the download queue should be interrupted """
1164 msg
= 'The download was cancelled'
1167 class ExistingVideoReached(DownloadCancelled
):
1168 """ --break-on-existing triggered """
1169 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1172 class RejectedVideoReached(DownloadCancelled
):
1173 """ --break-on-reject triggered """
1174 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1177 class MaxDownloadsReached(DownloadCancelled
):
1178 """ --max-downloads limit has been reached. """
1179 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1182 class ReExtractInfo(YoutubeDLError
):
1183 """ Video info needs to be re-extracted. """
1185 def __init__(self
, msg
, expected
=False):
1186 super().__init
__(msg
)
1187 self
.expected
= expected
1190 class ThrottledDownload(ReExtractInfo
):
1191 """ Download speed below --throttled-rate. """
1192 msg
= 'The download speed is below throttle limit'
1195 super().__init
__(self
.msg
, expected
=False)
1198 class UnavailableVideoError(YoutubeDLError
):
1199 """Unavailable Format exception.
1201 This exception will be thrown when a video is requested
1202 in a format that is not available for that video.
1204 msg
= 'Unable to download video'
1206 def __init__(self
, err
=None):
1208 self
.msg
+= f
': {err}'
1209 super().__init
__(self
.msg
)
1212 class ContentTooShortError(YoutubeDLError
):
1213 """Content Too Short exception.
1215 This exception may be raised by FileDownloader objects when a file they
1216 download is too small for what the server announced first, indicating
1217 the connection was probably interrupted.
1220 def __init__(self
, downloaded
, expected
):
1221 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1223 self
.downloaded
= downloaded
1224 self
.expected
= expected
1227 class XAttrMetadataError(YoutubeDLError
):
1228 def __init__(self
, code
=None, msg
='Unknown error'):
1229 super().__init
__(msg
)
1233 # Parsing code and msg
1234 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1235 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1236 self
.reason
= 'NO_SPACE'
1237 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1238 self
.reason
= 'VALUE_TOO_LONG'
1240 self
.reason
= 'NOT_SUPPORTED'
1243 class XAttrUnavailableError(YoutubeDLError
):
1247 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1248 hc
= http_class(*args
, **kwargs
)
1249 source_address
= ydl_handler
._params
.get('source_address')
1251 if source_address
is not None:
1252 # This is to workaround _create_connection() from socket where it will try all
1253 # address data from getaddrinfo() including IPv6. This filters the result from
1254 # getaddrinfo() based on the source_address value.
1255 # This is based on the cpython socket.create_connection() function.
1256 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1257 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1258 host
, port
= address
1260 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1261 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1262 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1263 if addrs
and not ip_addrs
:
1264 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1266 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1267 % (ip_version
, source_address
[0]))
1268 for res
in ip_addrs
:
1269 af
, socktype
, proto
, canonname
, sa
= res
1272 sock
= socket
.socket(af
, socktype
, proto
)
1273 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1274 sock
.settimeout(timeout
)
1275 sock
.bind(source_address
)
1277 err
= None # Explicitly break reference cycle
1279 except OSError as _
:
1281 if sock
is not None:
1286 raise OSError('getaddrinfo returns an empty list')
1287 if hasattr(hc
, '_create_connection'):
1288 hc
._create
_connection
= _create_connection
1289 hc
.source_address
= (source_address
, 0)
1294 def handle_youtubedl_headers(headers
):
1295 filtered_headers
= headers
1297 if 'Youtubedl-no-compression' in filtered_headers
:
1298 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1299 del filtered_headers
['Youtubedl-no-compression']
1301 return filtered_headers
1304 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1305 """Handler for HTTP requests and responses.
1307 This class, when installed with an OpenerDirector, automatically adds
1308 the standard headers to every HTTP request and handles gzipped and
1309 deflated responses from web servers. If compression is to be avoided in
1310 a particular request, the original request in the program code only has
1311 to include the HTTP header "Youtubedl-no-compression", which will be
1312 removed before making the real request.
1314 Part of this code was copied from:
1316 http://techknack.net/python-urllib2-handlers/
1318 Andrew Rowls, the author of that code, agreed to release it to the
1322 def __init__(self
, params
, *args
, **kwargs
):
1323 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1324 self
._params
= params
1326 def http_open(self
, req
):
1327 conn_class
= http
.client
.HTTPConnection
1329 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1331 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1332 del req
.headers
['Ytdl-socks-proxy']
1334 return self
.do_open(functools
.partial(
1335 _create_http_connection
, self
, conn_class
, False),
1343 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1345 return zlib
.decompress(data
)
1351 return brotli
.decompress(data
)
1353 def http_request(self
, req
):
1354 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1355 # always respected by websites, some tend to give out URLs with non percent-encoded
1356 # non-ASCII characters (see telemb.py, ard.py [#3412])
1357 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1358 # To work around aforementioned issue we will replace request's original URL with
1359 # percent-encoded one
1360 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1361 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1362 url
= req
.get_full_url()
1363 url_escaped
= escape_url(url
)
1365 # Substitute URL if any change after escaping
1366 if url
!= url_escaped
:
1367 req
= update_Request(req
, url
=url_escaped
)
1369 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1370 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1371 # The dict keys are capitalized because of this bug by urllib
1372 if h
.capitalize() not in req
.headers
:
1373 req
.add_header(h
, v
)
1375 if 'Accept-encoding' not in req
.headers
:
1376 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1378 req
.headers
= handle_youtubedl_headers(req
.headers
)
1380 return super().do_request_(req
)
1382 def http_response(self
, req
, resp
):
1385 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1386 content
= resp
.read()
1387 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1389 uncompressed
= io
.BytesIO(gz
.read())
1390 except OSError as original_ioerror
:
1391 # There may be junk add the end of the file
1392 # See http://stackoverflow.com/q/4928560/35070 for details
1393 for i
in range(1, 1024):
1395 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1396 uncompressed
= io
.BytesIO(gz
.read())
1401 raise original_ioerror
1402 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1403 resp
.msg
= old_resp
.msg
1404 del resp
.headers
['Content-encoding']
1406 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1407 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1408 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1409 resp
.msg
= old_resp
.msg
1410 del resp
.headers
['Content-encoding']
1412 if resp
.headers
.get('Content-encoding', '') == 'br':
1413 resp
= urllib
.request
.addinfourl(
1414 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1415 resp
.msg
= old_resp
.msg
1416 del resp
.headers
['Content-encoding']
1417 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1418 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1419 if 300 <= resp
.code
< 400:
1420 location
= resp
.headers
.get('Location')
1422 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1423 location
= location
.encode('iso-8859-1').decode()
1424 location_escaped
= escape_url(location
)
1425 if location
!= location_escaped
:
1426 del resp
.headers
['Location']
1427 resp
.headers
['Location'] = location_escaped
1430 https_request
= http_request
1431 https_response
= http_response
1434 def make_socks_conn_class(base_class
, socks_proxy
):
1435 assert issubclass(base_class
, (
1436 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1438 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1439 if url_components
.scheme
.lower() == 'socks5':
1440 socks_type
= ProxyType
.SOCKS5
1441 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1442 socks_type
= ProxyType
.SOCKS4
1443 elif url_components
.scheme
.lower() == 'socks4a':
1444 socks_type
= ProxyType
.SOCKS4A
1446 def unquote_if_non_empty(s
):
1449 return urllib
.parse
.unquote_plus(s
)
1453 url_components
.hostname
, url_components
.port
or 1080,
1455 unquote_if_non_empty(url_components
.username
),
1456 unquote_if_non_empty(url_components
.password
),
1459 class SocksConnection(base_class
):
1461 self
.sock
= sockssocket()
1462 self
.sock
.setproxy(*proxy_args
)
1463 if isinstance(self
.timeout
, (int, float)):
1464 self
.sock
.settimeout(self
.timeout
)
1465 self
.sock
.connect((self
.host
, self
.port
))
1467 if isinstance(self
, http
.client
.HTTPSConnection
):
1468 if hasattr(self
, '_context'): # Python > 2.6
1469 self
.sock
= self
._context
.wrap_socket(
1470 self
.sock
, server_hostname
=self
.host
)
1472 self
.sock
= ssl
.wrap_socket(self
.sock
)
1474 return SocksConnection
1477 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1478 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1479 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1480 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1481 self
._params
= params
1483 def https_open(self
, req
):
1485 conn_class
= self
._https
_conn
_class
1487 if hasattr(self
, '_context'): # python > 2.6
1488 kwargs
['context'] = self
._context
1489 if hasattr(self
, '_check_hostname'): # python 3.x
1490 kwargs
['check_hostname'] = self
._check
_hostname
1492 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1494 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1495 del req
.headers
['Ytdl-socks-proxy']
1498 return self
.do_open(
1499 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1500 except urllib
.error
.URLError
as e
:
1501 if (isinstance(e
.reason
, ssl
.SSLError
)
1502 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1503 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1507 def is_path_like(f
):
1508 return isinstance(f
, (str, bytes, os
.PathLike
))
1511 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1513 See [1] for cookie file format.
1515 1. https://curl.haxx.se/docs/http-cookies.html
1517 _HTTPONLY_PREFIX
= '#HttpOnly_'
1519 _HEADER
= '''# Netscape HTTP Cookie File
1520 # This file is generated by yt-dlp. Do not edit.
1523 _CookieFileEntry
= collections
.namedtuple(
1525 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1527 def __init__(self
, filename
=None, *args
, **kwargs
):
1528 super().__init
__(None, *args
, **kwargs
)
1529 if is_path_like(filename
):
1530 filename
= os
.fspath(filename
)
1531 self
.filename
= filename
1534 def _true_or_false(cndn
):
1535 return 'TRUE' if cndn
else 'FALSE'
1537 @contextlib.contextmanager
1538 def open(self
, file, *, write
=False):
1539 if is_path_like(file):
1540 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1547 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1550 if (not ignore_discard
and cookie
.discard
1551 or not ignore_expires
and cookie
.is_expired(now
)):
1553 name
, value
= cookie
.name
, cookie
.value
1555 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1556 # with no name, whereas http.cookiejar regards it as a
1557 # cookie with no value.
1558 name
, value
= '', name
1559 f
.write('%s\n' % '\t'.join((
1561 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1563 self
._true
_or
_false
(cookie
.secure
),
1564 str_or_none(cookie
.expires
, default
=''),
1568 def save(self
, filename
=None, *args
, **kwargs
):
1570 Save cookies to a file.
1571 Code is taken from CPython 3.6
1572 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1574 if filename
is None:
1575 if self
.filename
is not None:
1576 filename
= self
.filename
1578 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1580 # Store session cookies with `expires` set to 0 instead of an empty string
1582 if cookie
.expires
is None:
1585 with self
.open(filename
, write
=True) as f
:
1586 f
.write(self
._HEADER
)
1587 self
._really
_save
(f
, *args
, **kwargs
)
1589 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1590 """Load cookies from a file."""
1591 if filename
is None:
1592 if self
.filename
is not None:
1593 filename
= self
.filename
1595 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1597 def prepare_line(line
):
1598 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1599 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1600 # comments and empty lines are fine
1601 if line
.startswith('#') or not line
.strip():
1603 cookie_list
= line
.split('\t')
1604 if len(cookie_list
) != self
._ENTRY
_LEN
:
1605 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1606 cookie
= self
._CookieFileEntry
(*cookie_list
)
1607 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1608 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1612 with self
.open(filename
) as f
:
1615 cf
.write(prepare_line(line
))
1616 except http
.cookiejar
.LoadError
as e
:
1617 if f
'{line.strip()} '[0] in '[{"':
1618 raise http
.cookiejar
.LoadError(
1619 'Cookies file must be Netscape formatted, not JSON. See '
1620 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1621 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1624 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1625 # Session cookies are denoted by either `expires` field set to
1626 # an empty string or 0. MozillaCookieJar only recognizes the former
1627 # (see [1]). So we need force the latter to be recognized as session
1628 # cookies on our own.
1629 # Session cookies may be important for cookies-based authentication,
1630 # e.g. usually, when user does not check 'Remember me' check box while
1631 # logging in on a site, some important cookies are stored as session
1632 # cookies so that not recognizing them will result in failed login.
1633 # 1. https://bugs.python.org/issue17164
1635 # Treat `expires=0` cookies as session cookies
1636 if cookie
.expires
== 0:
1637 cookie
.expires
= None
1638 cookie
.discard
= True
1641 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1642 def __init__(self
, cookiejar
=None):
1643 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1645 def http_response(self
, request
, response
):
1646 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1648 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1649 https_response
= http_response
1652 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1653 """YoutubeDL redirect handler
1655 The code is based on HTTPRedirectHandler implementation from CPython [1].
1657 This redirect handler solves two issues:
1658 - ensures redirect URL is always unicode under python 2
1659 - introduces support for experimental HTTP response status code
1660 308 Permanent Redirect [2] used by some sites [3]
1662 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1663 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1664 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1667 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1669 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1670 """Return a Request or None in response to a redirect.
1672 This is called by the http_error_30x methods when a
1673 redirection response is received. If a redirection should
1674 take place, return a new Request to allow http_error_30x to
1675 perform the redirect. Otherwise, raise HTTPError if no-one
1676 else should try to handle this url. Return None if you can't
1677 but another Handler might.
1679 m
= req
.get_method()
1680 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1681 or code
in (301, 302, 303) and m
== "POST")):
1682 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1683 # Strictly (according to RFC 2616), 301 or 302 in response to
1684 # a POST MUST NOT cause a redirection without confirmation
1685 # from the user (of urllib.request, in this case). In practice,
1686 # essentially all clients do redirect in this case, so we do
1689 # Be conciliant with URIs containing a space. This is mainly
1690 # redundant with the more complete encoding done in http_error_302(),
1691 # but it is kept for compatibility with other callers.
1692 newurl
= newurl
.replace(' ', '%20')
1694 CONTENT_HEADERS
= ("content-length", "content-type")
1695 # NB: don't use dict comprehension for python 2.6 compatibility
1696 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1698 # A 303 must either use GET or HEAD for subsequent request
1699 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1700 if code
== 303 and m
!= 'HEAD':
1702 # 301 and 302 redirects are commonly turned into a GET from a POST
1703 # for subsequent requests by browsers, so we'll do the same.
1704 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1705 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1706 if code
in (301, 302) and m
== 'POST':
1709 return urllib
.request
.Request(
1710 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1711 unverifiable
=True, method
=m
)
1714 def extract_timezone(date_str
):
1717 ^.{8,}? # >=8 char non-TZ prefix, if present
1718 (?P<tz>Z| # just the UTC Z, or
1719 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1720 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1721 [ ]? # optional space
1722 (?P<sign>\+|-) # +/-
1723 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1727 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1728 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1729 if timezone
is not None:
1730 date_str
= date_str
[:-len(m
.group('tz'))]
1731 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1733 date_str
= date_str
[:-len(m
.group('tz'))]
1734 if not m
.group('sign'):
1735 timezone
= datetime
.timedelta()
1737 sign
= 1 if m
.group('sign') == '+' else -1
1738 timezone
= datetime
.timedelta(
1739 hours
=sign
* int(m
.group('hours')),
1740 minutes
=sign
* int(m
.group('minutes')))
1741 return timezone
, date_str
1744 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1745 """ Return a UNIX timestamp from the given date """
1747 if date_str
is None:
1750 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1752 if timezone
is None:
1753 timezone
, date_str
= extract_timezone(date_str
)
1755 with contextlib
.suppress(ValueError):
1756 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1757 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1758 return calendar
.timegm(dt
.timetuple())
1761 def date_formats(day_first
=True):
1762 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1765 def unified_strdate(date_str
, day_first
=True):
1766 """Return a string with the date in the format YYYYMMDD"""
1768 if date_str
is None:
1772 date_str
= date_str
.replace(',', ' ')
1773 # Remove AM/PM + timezone
1774 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1775 _
, date_str
= extract_timezone(date_str
)
1777 for expression
in date_formats(day_first
):
1778 with contextlib
.suppress(ValueError):
1779 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1780 if upload_date
is None:
1781 timetuple
= email
.utils
.parsedate_tz(date_str
)
1783 with contextlib
.suppress(ValueError):
1784 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1785 if upload_date
is not None:
1786 return str(upload_date
)
1789 def unified_timestamp(date_str
, day_first
=True):
1790 if date_str
is None:
1793 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1794 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1796 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1797 timezone
, date_str
= extract_timezone(date_str
)
1799 # Remove AM/PM + timezone
1800 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1802 # Remove unrecognized timezones from ISO 8601 alike timestamps
1803 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1805 date_str
= date_str
[:-len(m
.group('tz'))]
1807 # Python only supports microseconds, so remove nanoseconds
1808 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1810 date_str
= m
.group(1)
1812 for expression
in date_formats(day_first
):
1813 with contextlib
.suppress(ValueError):
1814 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1815 return calendar
.timegm(dt
.timetuple())
1817 timetuple
= email
.utils
.parsedate_tz(date_str
)
1819 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1822 def determine_ext(url
, default_ext
='unknown_video'):
1823 if url
is None or '.' not in url
:
1825 guess
= url
.partition('?')[0].rpartition('.')[2]
1826 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1828 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1829 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1830 return guess
.rstrip('/')
1835 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1836 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1839 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1841 Return a datetime object from a string.
1843 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1845 @param format strftime format of DATE
1846 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1847 auto: round to the unit provided in date_str (if applicable).
1849 auto_precision
= False
1850 if precision
== 'auto':
1851 auto_precision
= True
1852 precision
= 'microsecond'
1853 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1854 if date_str
in ('now', 'today'):
1856 if date_str
== 'yesterday':
1857 return today
- datetime
.timedelta(days
=1)
1859 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1861 if match
is not None:
1862 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1863 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1864 unit
= match
.group('unit')
1865 if unit
== 'month' or unit
== 'year':
1866 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1872 delta
= datetime
.timedelta(**{unit + 's': time}
)
1873 new_date
= start_time
+ delta
1875 return datetime_round(new_date
, unit
)
1878 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1881 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1883 Return a date object from a string using datetime_from_str
1885 @param strict Restrict allowed patterns to "YYYYMMDD" and
1886 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1888 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1889 raise ValueError(f
'Invalid date format "{date_str}"')
1890 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1893 def datetime_add_months(dt
, months
):
1894 """Increment/Decrement a datetime object by months."""
1895 month
= dt
.month
+ months
- 1
1896 year
= dt
.year
+ month
// 12
1897 month
= month
% 12 + 1
1898 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1899 return dt
.replace(year
, month
, day
)
1902 def datetime_round(dt
, precision
='day'):
1904 Round a datetime object's time to a specific precision
1906 if precision
== 'microsecond':
1915 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1916 timestamp
= calendar
.timegm(dt
.timetuple())
1917 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1920 def hyphenate_date(date_str
):
1922 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1923 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1924 if match
is not None:
1925 return '-'.join(match
.groups())
1931 """Represents a time interval between two dates"""
1933 def __init__(self
, start
=None, end
=None):
1934 """start and end must be strings in the format accepted by date"""
1935 if start
is not None:
1936 self
.start
= date_from_str(start
, strict
=True)
1938 self
.start
= datetime
.datetime
.min.date()
1940 self
.end
= date_from_str(end
, strict
=True)
1942 self
.end
= datetime
.datetime
.max.date()
1943 if self
.start
> self
.end
:
1944 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1948 """Returns a range that only contains the given day"""
1949 return cls(day
, day
)
1951 def __contains__(self
, date
):
1952 """Check if the date is in the range"""
1953 if not isinstance(date
, datetime
.date
):
1954 date
= date_from_str(date
)
1955 return self
.start
<= date
<= self
.end
1958 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1960 def __eq__(self
, other
):
1961 return (isinstance(other
, DateRange
)
1962 and self
.start
== other
.start
and self
.end
== other
.end
)
1965 def platform_name():
1966 """ Returns the platform name as a str """
1967 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
1968 return platform
.platform()
1972 def system_identifier():
1973 python_implementation
= platform
.python_implementation()
1974 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
1975 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
1977 with contextlib
.suppress(OSError): # We may not have access to the executable
1978 libc_ver
= platform
.libc_ver()
1980 return 'Python %s (%s %s) - %s %s' % (
1981 platform
.python_version(),
1982 python_implementation
,
1983 platform
.architecture()[0],
1984 platform
.platform(),
1985 format_field(join_nonempty(*libc_ver
, delim
=' '), None, '(%s)'),
1990 def get_windows_version():
1991 ''' Get Windows version. returns () if it's not running on Windows '''
1992 if compat_os_name
== 'nt':
1993 return version_tuple(platform
.win32_ver()[1])
1998 def write_string(s
, out
=None, encoding
=None):
1999 assert isinstance(s
, str)
2000 out
= out
or sys
.stderr
2002 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2003 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2005 enc
, buffer = None, out
2006 if 'b' in getattr(out
, 'mode', ''):
2007 enc
= encoding
or preferredencoding()
2008 elif hasattr(out
, 'buffer'):
2010 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2012 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2016 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2017 from . import _IN_CLI
2019 if msg
in deprecation_warning
._cache
:
2021 deprecation_warning
._cache
.add(msg
)
2023 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2024 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2027 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2030 deprecation_warning
._cache
= set()
2033 def bytes_to_intlist(bs
):
2036 if isinstance(bs
[0], int): # Python 3
2039 return [ord(c
) for c
in bs
]
2042 def intlist_to_bytes(xs
):
2045 return struct
.pack('%dB' % len(xs
), *xs
)
2048 class LockingUnsupportedError(OSError):
2049 msg
= 'File locking is not supported'
2052 super().__init
__(self
.msg
)
2055 # Cross-platform file locking
2056 if sys
.platform
== 'win32':
2058 import ctypes
.wintypes
2061 class OVERLAPPED(ctypes
.Structure
):
2063 ('Internal', ctypes
.wintypes
.LPVOID
),
2064 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2065 ('Offset', ctypes
.wintypes
.DWORD
),
2066 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2067 ('hEvent', ctypes
.wintypes
.HANDLE
),
2070 kernel32
= ctypes
.windll
.kernel32
2071 LockFileEx
= kernel32
.LockFileEx
2072 LockFileEx
.argtypes
= [
2073 ctypes
.wintypes
.HANDLE
, # hFile
2074 ctypes
.wintypes
.DWORD
, # dwFlags
2075 ctypes
.wintypes
.DWORD
, # dwReserved
2076 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2077 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2078 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2080 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2081 UnlockFileEx
= kernel32
.UnlockFileEx
2082 UnlockFileEx
.argtypes
= [
2083 ctypes
.wintypes
.HANDLE
, # hFile
2084 ctypes
.wintypes
.DWORD
, # dwReserved
2085 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2086 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2087 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2089 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2090 whole_low
= 0xffffffff
2091 whole_high
= 0x7fffffff
2093 def _lock_file(f
, exclusive
, block
):
2094 overlapped
= OVERLAPPED()
2095 overlapped
.Offset
= 0
2096 overlapped
.OffsetHigh
= 0
2097 overlapped
.hEvent
= 0
2098 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2100 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2101 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2102 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2103 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2104 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2106 def _unlock_file(f
):
2107 assert f
._lock
_file
_overlapped
_p
2108 handle
= msvcrt
.get_osfhandle(f
.fileno())
2109 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2110 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2116 def _lock_file(f
, exclusive
, block
):
2117 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2119 flags |
= fcntl
.LOCK_NB
2121 fcntl
.flock(f
, flags
)
2122 except BlockingIOError
:
2124 except OSError: # AOSP does not have flock()
2125 fcntl
.lockf(f
, flags
)
2127 def _unlock_file(f
):
2129 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2131 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2135 def _lock_file(f
, exclusive
, block
):
2136 raise LockingUnsupportedError()
2138 def _unlock_file(f
):
2139 raise LockingUnsupportedError()
2145 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2146 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2147 raise NotImplementedError(mode
)
2148 self
.mode
, self
.block
= mode
, block
2150 writable
= any(f
in mode
for f
in 'wax+')
2151 readable
= any(f
in mode
for f
in 'r+')
2152 flags
= functools
.reduce(operator
.ior
, (
2153 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2154 getattr(os
, 'O_BINARY', 0), # Windows only
2155 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2156 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2157 os
.O_APPEND
if 'a' in mode
else 0,
2158 os
.O_EXCL
if 'x' in mode
else 0,
2159 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2162 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2164 def __enter__(self
):
2165 exclusive
= 'r' not in self
.mode
2167 _lock_file(self
.f
, exclusive
, self
.block
)
2172 if 'w' in self
.mode
:
2175 except OSError as e
:
2177 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2178 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2187 _unlock_file(self
.f
)
2191 def __exit__(self
, *_
):
2200 def __getattr__(self
, attr
):
2201 return getattr(self
.f
, attr
)
2208 def get_filesystem_encoding():
2209 encoding
= sys
.getfilesystemencoding()
2210 return encoding
if encoding
is not None else 'utf-8'
2213 def shell_quote(args
):
2215 encoding
= get_filesystem_encoding()
2217 if isinstance(a
, bytes):
2218 # We may get a filename encoded with 'encodeFilename'
2219 a
= a
.decode(encoding
)
2220 quoted_args
.append(compat_shlex_quote(a
))
2221 return ' '.join(quoted_args
)
2224 def smuggle_url(url
, data
):
2225 """ Pass additional data in a URL for internal use. """
2227 url
, idata
= unsmuggle_url(url
, {})
2229 sdata
= urllib
.parse
.urlencode(
2230 {'__youtubedl_smuggle': json.dumps(data)}
)
2231 return url
+ '#' + sdata
2234 def unsmuggle_url(smug_url
, default
=None):
2235 if '#__youtubedl_smuggle' not in smug_url
:
2236 return smug_url
, default
2237 url
, _
, sdata
= smug_url
.rpartition('#')
2238 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2239 data
= json
.loads(jsond
)
2243 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2244 """ Formats numbers with decimal sufixes like K, M, etc """
2245 num
, factor
= float_or_none(num
), float(factor
)
2246 if num
is None or num
< 0:
2248 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2249 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2250 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2252 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2253 converted
= num
/ (factor
** exponent
)
2254 return fmt
% (converted
, suffix
)
2257 def format_bytes(bytes):
2258 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2261 def lookup_unit_table(unit_table
, s
):
2262 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2264 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2267 num_str
= m
.group('num').replace(',', '.')
2268 mult
= unit_table
[m
.group('unit')]
2269 return int(float(num_str
) * mult
)
2272 def parse_filesize(s
):
2276 # The lower-case forms are of course incorrect and unofficial,
2277 # but we support those too
2294 'megabytes': 1000 ** 2,
2295 'mebibytes': 1024 ** 2,
2301 'gigabytes': 1000 ** 3,
2302 'gibibytes': 1024 ** 3,
2308 'terabytes': 1000 ** 4,
2309 'tebibytes': 1024 ** 4,
2315 'petabytes': 1000 ** 5,
2316 'pebibytes': 1024 ** 5,
2322 'exabytes': 1000 ** 6,
2323 'exbibytes': 1024 ** 6,
2329 'zettabytes': 1000 ** 7,
2330 'zebibytes': 1024 ** 7,
2336 'yottabytes': 1000 ** 8,
2337 'yobibytes': 1024 ** 8,
2340 return lookup_unit_table(_UNIT_TABLE
, s
)
2347 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2349 if re
.match(r
'^[\d,.]+$', s
):
2350 return str_to_int(s
)
2363 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2367 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2369 return str_to_int(mobj
.group(1))
2372 def parse_resolution(s
, *, lenient
=False):
2377 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2379 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2382 'width': int(mobj
.group('w')),
2383 'height': int(mobj
.group('h')),
2386 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2388 return {'height': int(mobj.group(1))}
2390 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2392 return {'height': int(mobj.group(1)) * 540}
2397 def parse_bitrate(s
):
2398 if not isinstance(s
, str):
2400 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2402 return int(mobj
.group(1))
2405 def month_by_name(name
, lang
='en'):
2406 """ Return the number of a month by (locale-independently) English name """
2408 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2411 return month_names
.index(name
) + 1
2416 def month_by_abbreviation(abbrev
):
2417 """ Return the number of a month by (locale-independently) English
2421 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2426 def fix_xml_ampersands(xml_str
):
2427 """Replace all the '&' by '&' in XML"""
2429 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2434 def setproctitle(title
):
2435 assert isinstance(title
, str)
2437 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2444 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2448 # LoadLibrary in Windows Python 2.7.13 only expects
2449 # a bytestring, but since unicode_literals turns
2450 # every string into a unicode string, it fails.
2452 title_bytes
= title
.encode()
2453 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2454 buf
.value
= title_bytes
2456 libc
.prctl(15, buf
, 0, 0, 0)
2457 except AttributeError:
2458 return # Strange libc, just skip this
2461 def remove_start(s
, start
):
2462 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2465 def remove_end(s
, end
):
2466 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2469 def remove_quotes(s
):
2470 if s
is None or len(s
) < 2:
2472 for quote
in ('"', "'", ):
2473 if s
[0] == quote
and s
[-1] == quote
:
2478 def get_domain(url
):
2480 This implementation is inconsistent, but is kept for compatibility.
2481 Use this only for "webpage_url_domain"
2483 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2486 def url_basename(url
):
2487 path
= urllib
.parse
.urlparse(url
).path
2488 return path
.strip('/').split('/')[-1]
2492 return re
.match(r
'https?://[^?#]+/', url
).group()
2495 def urljoin(base
, path
):
2496 if isinstance(path
, bytes):
2497 path
= path
.decode()
2498 if not isinstance(path
, str) or not path
:
2500 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2502 if isinstance(base
, bytes):
2503 base
= base
.decode()
2504 if not isinstance(base
, str) or not re
.match(
2505 r
'^(?:https?:)?//', base
):
2507 return urllib
.parse
.urljoin(base
, path
)
2510 class HEADRequest(urllib
.request
.Request
):
2511 def get_method(self
):
2515 class PUTRequest(urllib
.request
.Request
):
2516 def get_method(self
):
2520 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2521 if get_attr
and v
is not None:
2522 v
= getattr(v
, get_attr
, None)
2524 return int(v
) * invscale
// scale
2525 except (ValueError, TypeError, OverflowError):
2529 def str_or_none(v
, default
=None):
2530 return default
if v
is None else str(v
)
2533 def str_to_int(int_str
):
2534 """ A more relaxed version of int_or_none """
2535 if isinstance(int_str
, int):
2537 elif isinstance(int_str
, str):
2538 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2539 return int_or_none(int_str
)
2542 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2546 return float(v
) * invscale
/ scale
2547 except (ValueError, TypeError):
2551 def bool_or_none(v
, default
=None):
2552 return v
if isinstance(v
, bool) else default
2555 def strip_or_none(v
, default
=None):
2556 return v
.strip() if isinstance(v
, str) else default
2559 def url_or_none(url
):
2560 if not url
or not isinstance(url
, str):
2563 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2566 def request_to_url(req
):
2567 if isinstance(req
, urllib
.request
.Request
):
2568 return req
.get_full_url()
2573 def strftime_or_none(timestamp
, date_format
, default
=None):
2574 datetime_object
= None
2576 if isinstance(timestamp
, (int, float)): # unix timestamp
2577 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2578 elif isinstance(timestamp
, str): # assume YYYYMMDD
2579 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2580 date_format
= re
.sub( # Support %s on windows
2581 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2582 return datetime_object
.strftime(date_format
)
2583 except (ValueError, TypeError, AttributeError):
2587 def parse_duration(s
):
2588 if not isinstance(s
, str):
2594 days
, hours
, mins
, secs
, ms
= [None] * 5
2595 m
= re
.match(r
'''(?x)
2597 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2598 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2599 (?P<ms>[.:][0-9]+)?Z?$
2602 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2607 [0-9]+\s*y(?:ears?)?,?\s*
2610 [0-9]+\s*m(?:onths?)?,?\s*
2613 [0-9]+\s*w(?:eeks?)?,?\s*
2616 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2620 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2623 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2626 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2629 days
, hours
, mins
, secs
, ms
= m
.groups()
2631 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2633 hours
, mins
= m
.groups()
2638 ms
= ms
.replace(':', '.')
2639 return sum(float(part
or 0) * mult
for part
, mult
in (
2640 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2643 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2644 name
, real_ext
= os
.path
.splitext(filename
)
2646 f
'{name}.{ext}{real_ext}'
2647 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2648 else f
'{filename}.{ext}')
2651 def replace_extension(filename
, ext
, expected_real_ext
=None):
2652 name
, real_ext
= os
.path
.splitext(filename
)
2653 return '{}.{}'.format(
2654 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2658 def check_executable(exe
, args
=[]):
2659 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2660 args can be a list of arguments for a short output (like -version) """
2662 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2668 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2670 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2672 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2673 # SIGTTOU if yt-dlp is run in the background.
2674 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2675 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2676 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2682 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2683 assert isinstance(output
, str)
2684 if version_re
is None:
2685 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2686 m
= re
.search(version_re
, output
)
2693 def get_exe_version(exe
, args
=['--version'],
2694 version_re
=None, unrecognized
='present'):
2695 """ Returns the version of the specified executable,
2696 or False if the executable is not present """
2697 out
= _get_exe_version_output(exe
, args
)
2698 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2701 def frange(start
=0, stop
=None, step
=1):
2704 start
, stop
= 0, start
2705 sign
= [-1, 1][step
> 0] if step
else 0
2706 while sign
* start
< sign
* stop
:
2711 class LazyList(collections
.abc
.Sequence
):
2712 """Lazy immutable list from an iterable
2713 Note that slices of a LazyList are lists and not LazyList"""
2715 class IndexError(IndexError):
2718 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2719 self
._iterable
= iter(iterable
)
2720 self
._cache
= [] if _cache
is None else _cache
2721 self
._reversed
= reverse
2725 # We need to consume the entire iterable to iterate in reverse
2726 yield from self
.exhaust()
2728 yield from self
._cache
2729 for item
in self
._iterable
:
2730 self
._cache
.append(item
)
2734 self
._cache
.extend(self
._iterable
)
2735 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2739 """Evaluate the entire iterable"""
2740 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2743 def _reverse_index(x
):
2744 return None if x
is None else ~x
2746 def __getitem__(self
, idx
):
2747 if isinstance(idx
, slice):
2749 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2750 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2751 elif isinstance(idx
, int):
2753 idx
= self
._reverse
_index
(idx
)
2754 start
, stop
, step
= idx
, idx
, 0
2756 raise TypeError('indices must be integers or slices')
2757 if ((start
or 0) < 0 or (stop
or 0) < 0
2758 or (start
is None and step
< 0)
2759 or (stop
is None and step
> 0)):
2760 # We need to consume the entire iterable to be able to slice from the end
2761 # Obviously, never use this with infinite iterables
2764 return self
._cache
[idx
]
2765 except IndexError as e
:
2766 raise self
.IndexError(e
) from e
2767 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2769 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2771 return self
._cache
[idx
]
2772 except IndexError as e
:
2773 raise self
.IndexError(e
) from e
2777 self
[-1] if self
._reversed
else self
[0]
2778 except self
.IndexError:
2784 return len(self
._cache
)
2786 def __reversed__(self
):
2787 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2790 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2793 # repr and str should mimic a list. So we exhaust the iterable
2794 return repr(self
.exhaust())
2797 return repr(self
.exhaust())
2802 class IndexError(IndexError):
2806 # This is only useful for tests
2807 return len(self
.getslice())
2809 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2810 self
._pagefunc
= pagefunc
2811 self
._pagesize
= pagesize
2812 self
._pagecount
= float('inf')
2813 self
._use
_cache
= use_cache
2816 def getpage(self
, pagenum
):
2817 page_results
= self
._cache
.get(pagenum
)
2818 if page_results
is None:
2819 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2821 self
._cache
[pagenum
] = page_results
2824 def getslice(self
, start
=0, end
=None):
2825 return list(self
._getslice
(start
, end
))
2827 def _getslice(self
, start
, end
):
2828 raise NotImplementedError('This method must be implemented by subclasses')
2830 def __getitem__(self
, idx
):
2831 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2832 if not isinstance(idx
, int) or idx
< 0:
2833 raise TypeError('indices must be non-negative integers')
2834 entries
= self
.getslice(idx
, idx
+ 1)
2836 raise self
.IndexError()
2840 class OnDemandPagedList(PagedList
):
2841 """Download pages until a page with less than maximum results"""
2843 def _getslice(self
, start
, end
):
2844 for pagenum
in itertools
.count(start
// self
._pagesize
):
2845 firstid
= pagenum
* self
._pagesize
2846 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2847 if start
>= nextfirstid
:
2851 start
% self
._pagesize
2852 if firstid
<= start
< nextfirstid
2855 ((end
- 1) % self
._pagesize
) + 1
2856 if (end
is not None and firstid
<= end
<= nextfirstid
)
2860 page_results
= self
.getpage(pagenum
)
2862 self
._pagecount
= pagenum
- 1
2864 if startv
!= 0 or endv
is not None:
2865 page_results
= page_results
[startv
:endv
]
2866 yield from page_results
2868 # A little optimization - if current page is not "full", ie. does
2869 # not contain page_size videos then we can assume that this page
2870 # is the last one - there are no more ids on further pages -
2871 # i.e. no need to query again.
2872 if len(page_results
) + startv
< self
._pagesize
:
2875 # If we got the whole page, but the next page is not interesting,
2876 # break out early as well
2877 if end
== nextfirstid
:
2881 class InAdvancePagedList(PagedList
):
2882 """PagedList with total number of pages known in advance"""
2884 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2885 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2886 self
._pagecount
= pagecount
2888 def _getslice(self
, start
, end
):
2889 start_page
= start
// self
._pagesize
2890 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2891 skip_elems
= start
- start_page
* self
._pagesize
2892 only_more
= None if end
is None else end
- start
2893 for pagenum
in range(start_page
, end_page
):
2894 page_results
= self
.getpage(pagenum
)
2896 page_results
= page_results
[skip_elems
:]
2898 if only_more
is not None:
2899 if len(page_results
) < only_more
:
2900 only_more
-= len(page_results
)
2902 yield from page_results
[:only_more
]
2904 yield from page_results
2907 class PlaylistEntries
:
2908 MissingEntry
= object()
2909 is_exhausted
= False
2911 def __init__(self
, ydl
, info_dict
):
2914 # _entries must be assigned now since infodict can change during iteration
2915 entries
= info_dict
.get('entries')
2917 raise EntryNotInPlaylist('There are no entries')
2918 elif isinstance(entries
, list):
2919 self
.is_exhausted
= True
2921 requested_entries
= info_dict
.get('requested_entries')
2922 self
.is_incomplete
= bool(requested_entries
)
2923 if self
.is_incomplete
:
2924 assert self
.is_exhausted
2925 self
._entries
= [self
.MissingEntry
] * max(requested_entries
)
2926 for i
, entry
in zip(requested_entries
, entries
):
2927 self
._entries
[i
- 1] = entry
2928 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2929 self
._entries
= entries
2931 self
._entries
= LazyList(entries
)
2933 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2934 (?P<start>[+-]?\d+)?
2936 (?P<end>[+-]?\d+|inf(?:inite)?)?
2937 (?::(?P<step>[+-]?\d+))?
2941 def parse_playlist_items(cls
, string
):
2942 for segment
in string
.split(','):
2944 raise ValueError('There is two or more consecutive commas')
2945 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2947 raise ValueError(f
'{segment!r} is not a valid specification')
2948 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2949 if int_or_none(step
) == 0:
2950 raise ValueError(f
'Step in {segment!r} cannot be zero')
2951 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2953 def get_requested_items(self
):
2954 playlist_items
= self
.ydl
.params
.get('playlist_items')
2955 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2956 playlist_end
= self
.ydl
.params
.get('playlistend')
2957 # For backwards compatibility, interpret -1 as whole list
2958 if playlist_end
in (-1, None):
2960 if not playlist_items
:
2961 playlist_items
= f
'{playlist_start}:{playlist_end}'
2962 elif playlist_start
!= 1 or playlist_end
:
2963 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2965 for index
in self
.parse_playlist_items(playlist_items
):
2966 for i
, entry
in self
[index
]:
2971 # TODO: Add auto-generated fields
2972 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
2973 except (ExistingVideoReached
, RejectedVideoReached
):
2976 def get_full_count(self
):
2977 if self
.is_exhausted
and not self
.is_incomplete
:
2979 elif isinstance(self
._entries
, InAdvancePagedList
):
2980 if self
._entries
._pagesize
== 1:
2981 return self
._entries
._pagecount
2983 @functools.cached_property
2985 if isinstance(self
._entries
, list):
2988 entry
= self
._entries
[i
]
2990 entry
= self
.MissingEntry
2991 if not self
.is_incomplete
:
2992 raise self
.IndexError()
2993 if entry
is self
.MissingEntry
:
2994 raise EntryNotInPlaylist(f
'Entry {i} cannot be found')
2999 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3000 except (LazyList
.IndexError, PagedList
.IndexError):
3001 raise self
.IndexError()
3004 def __getitem__(self
, idx
):
3005 if isinstance(idx
, int):
3006 idx
= slice(idx
, idx
)
3008 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3009 step
= 1 if idx
.step
is None else idx
.step
3010 if idx
.start
is None:
3011 start
= 0 if step
> 0 else len(self
) - 1
3013 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3015 # NB: Do not call len(self) when idx == [:]
3016 if idx
.stop
is None:
3017 stop
= 0 if step
< 0 else float('inf')
3019 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3020 stop
+= [-1, 1][step
> 0]
3022 for i
in frange(start
, stop
, step
):
3026 entry
= self
._getter
(i
)
3027 except self
.IndexError:
3028 self
.is_exhausted
= True
3035 return len(tuple(self
[:]))
3037 class IndexError(IndexError):
3041 def uppercase_escape(s
):
3042 unicode_escape
= codecs
.getdecoder('unicode_escape')
3044 r
'\\U[0-9a-fA-F]{8}',
3045 lambda m
: unicode_escape(m
.group(0))[0],
3049 def lowercase_escape(s
):
3050 unicode_escape
= codecs
.getdecoder('unicode_escape')
3052 r
'\\u[0-9a-fA-F]{4}',
3053 lambda m
: unicode_escape(m
.group(0))[0],
3057 def escape_rfc3986(s
):
3058 """Escape non-ASCII characters as suggested by RFC 3986"""
3059 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3062 def escape_url(url
):
3063 """Escape URL as suggested by RFC 3986"""
3064 url_parsed
= urllib
.parse
.urlparse(url
)
3065 return url_parsed
._replace
(
3066 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3067 path
=escape_rfc3986(url_parsed
.path
),
3068 params
=escape_rfc3986(url_parsed
.params
),
3069 query
=escape_rfc3986(url_parsed
.query
),
3070 fragment
=escape_rfc3986(url_parsed
.fragment
)
3075 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
)
3078 def read_batch_urls(batch_fd
):
3080 if not isinstance(url
, str):
3081 url
= url
.decode('utf-8', 'replace')
3082 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3083 for bom
in BOM_UTF8
:
3084 if url
.startswith(bom
):
3085 url
= url
[len(bom
):]
3087 if not url
or url
.startswith(('#', ';', ']')):
3089 # "#" cannot be stripped out since it is part of the URI
3090 # However, it can be safely stripped out if following a whitespace
3091 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3093 with contextlib
.closing(batch_fd
) as fd
:
3094 return [url
for url
in map(fixup
, fd
) if url
]
3097 def urlencode_postdata(*args
, **kargs
):
3098 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3101 def update_url_query(url
, query
):
3104 parsed_url
= urllib
.parse
.urlparse(url
)
3105 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3107 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3108 query
=urllib
.parse
.urlencode(qs
, True)))
3111 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3112 req_headers
= req
.headers
.copy()
3113 req_headers
.update(headers
or {})
3114 req_data
= data
or req
.data
3115 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3116 req_get_method
= req
.get_method()
3117 if req_get_method
== 'HEAD':
3118 req_type
= HEADRequest
3119 elif req_get_method
== 'PUT':
3120 req_type
= PUTRequest
3122 req_type
= urllib
.request
.Request
3124 req_url
, data
=req_data
, headers
=req_headers
,
3125 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3126 if hasattr(req
, 'timeout'):
3127 new_req
.timeout
= req
.timeout
3131 def _multipart_encode_impl(data
, boundary
):
3132 content_type
= 'multipart/form-data; boundary=%s' % boundary
3135 for k
, v
in data
.items():
3136 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3137 if isinstance(k
, str):
3139 if isinstance(v
, str):
3141 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3142 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3143 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3144 if boundary
.encode('ascii') in content
:
3145 raise ValueError('Boundary overlaps with data')
3148 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3150 return out
, content_type
3153 def multipart_encode(data
, boundary
=None):
3155 Encode a dict to RFC 7578-compliant form-data
3158 A dict where keys and values can be either Unicode or bytes-like
3161 If specified a Unicode object, it's used as the boundary. Otherwise
3162 a random boundary is generated.
3164 Reference: https://tools.ietf.org/html/rfc7578
3166 has_specified_boundary
= boundary
is not None
3169 if boundary
is None:
3170 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3173 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3176 if has_specified_boundary
:
3180 return out
, content_type
3183 def variadic(x
, allowed_types
=(str, bytes, dict)):
3184 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3187 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3188 for val
in map(d
.get
, variadic(key_or_keys
)):
3189 if val
is not None and (val
or not skip_false_values
):
3194 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3197 val
= f(*args
, **kwargs
)
3198 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3201 if expected_type
is None or isinstance(val
, expected_type
):
3205 def try_get(src
, getter
, expected_type
=None):
3206 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3209 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3210 return {k: v for k, v in dct.items() if cndn(k, v)}
3213 def merge_dicts(*dicts
):
3215 for a_dict
in dicts
:
3216 for k
, v
in a_dict
.items():
3217 if (v
is not None and k
not in merged
3218 or isinstance(v
, str) and merged
[k
] == ''):
3223 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3224 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3236 TV_PARENTAL_GUIDELINES
= {
3246 def parse_age_limit(s
):
3247 # isinstance(False, int) is True. So type() must be used instead
3248 if type(s
) is int: # noqa: E721
3249 return s
if 0 <= s
<= 21 else None
3250 elif not isinstance(s
, str):
3252 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3254 return int(m
.group('age'))
3257 return US_RATINGS
[s
]
3258 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3260 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3264 def strip_jsonp(code
):
3267 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3268 (?:\s*&&\s*(?P=func_name))?
3269 \s*\(\s*(?P<callback_data>.*)\);?
3270 \s*?(?://[^\n]*)*$''',
3271 r
'\g<callback_data>', code
)
3274 def js_to_json(code
, vars={}, *, strict
=False):
3275 # vars is a dict of var, val pairs to substitute
3276 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3277 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3279 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3280 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3285 if v
in ('true', 'false', 'null'):
3287 elif v
in ('undefined', 'void 0'):
3289 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3292 if v
[0] in ("'", '"'):
3293 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3298 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3300 for regex
, base
in INTEGER_TABLE
:
3301 im
= re
.match(regex
, v
)
3303 i
= int(im
.group(1), base
)
3304 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3307 return json
.dumps(vars[v
])
3309 raise ValueError(f
'Unknown value: {v}')
3313 def create_map(mobj
):
3314 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3316 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3318 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3319 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3321 return re
.sub(r
'''(?sx)
3322 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3323 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3324 {comment}|,(?={skip}[\]}}])|
3325 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3326 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3329 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3332 def qualities(quality_ids
):
3333 """ Get a numeric quality value out of a list of possible values """
3336 return quality_ids
.index(qid
)
3342 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3346 'default': '%(title)s [%(id)s].%(ext)s',
3347 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3353 'description': 'description',
3354 'annotation': 'annotations.xml',
3355 'infojson': 'info.json',
3358 'pl_thumbnail': None,
3359 'pl_description': 'description',
3360 'pl_infojson': 'info.json',
3363 # As of [1] format syntax is:
3364 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3365 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3366 STR_FORMAT_RE_TMPL
= r
'''(?x)
3367 (?<!%)(?P<prefix>(?:%%)*)
3369 (?P<has_key>\((?P<key>{0})\))?
3371 (?P<conversion>[#0\-+ ]+)?
3373 (?P<precision>\.\d+)?
3374 (?P<len_mod>[hlL])? # unused in python
3375 {1} # conversion type
3380 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3383 def limit_length(s
, length
):
3384 """ Add ellipses to overly long strings """
3389 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3393 def version_tuple(v
):
3394 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3397 def is_outdated_version(version
, limit
, assume_new
=True):
3399 return not assume_new
3401 return version_tuple(version
) < version_tuple(limit
)
3403 return not assume_new
3406 def ytdl_is_updateable():
3407 """ Returns if yt-dlp can be updated with -U """
3409 from .update
import is_non_updateable
3411 return not is_non_updateable()
3414 def args_to_str(args
):
3415 # Get a short string representation for a subprocess command
3416 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3419 def error_to_compat_str(err
):
3423 def error_to_str(err
):
3424 return f
'{type(err).__name__}: {err}'
3427 def mimetype2ext(mt
):
3431 mt
, _
, params
= mt
.partition(';')
3436 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3437 # it's the most popular one
3438 'audio/mpeg': 'mp3',
3439 'audio/x-wav': 'wav',
3441 'audio/wave': 'wav',
3444 ext
= FULL_MAP
.get(mt
)
3450 'smptett+xml': 'tt',
3454 'x-mp4-fragmented': 'mp4',
3455 'x-ms-sami': 'sami',
3458 'x-mpegurl': 'm3u8',
3459 'vnd.apple.mpegurl': 'm3u8',
3463 'vnd.ms-sstr+xml': 'ism',
3467 'filmstrip+json': 'fs',
3471 _
, _
, subtype
= mt
.rpartition('/')
3472 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3483 _
, _
, suffix
= subtype
.partition('+')
3484 ext
= SUFFIX_MAP
.get(suffix
)
3488 return subtype
.replace('+', '.')
3491 def ext2mimetype(ext_or_url
):
3494 if '.' not in ext_or_url
:
3495 ext_or_url
= f
'file.{ext_or_url}'
3496 return mimetypes
.guess_type(ext_or_url
)[0]
3499 def parse_codecs(codecs_str
):
3500 # http://tools.ietf.org/html/rfc6381
3503 split_codecs
= list(filter(None, map(
3504 str.strip
, codecs_str
.strip().strip(',').split(','))))
3505 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3506 for full_codec
in split_codecs
:
3507 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3508 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3509 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3513 if parts
[0] in ('dvh1', 'dvhe'):
3515 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3517 elif parts
[:2] == ['vp9', '2']:
3519 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3520 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3521 acodec
= acodec
or full_codec
3522 elif parts
[0] in ('stpp', 'wvtt'):
3523 scodec
= scodec
or full_codec
3525 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3526 if vcodec
or acodec
or scodec
:
3528 'vcodec': vcodec
or 'none',
3529 'acodec': acodec
or 'none',
3530 'dynamic_range': hdr
,
3531 **({'scodec': scodec}
if scodec
is not None else {}),
3533 elif len(split_codecs
) == 2:
3535 'vcodec': split_codecs
[0],
3536 'acodec': split_codecs
[1],
3541 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3542 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3544 allow_mkv
= not preferences
or 'mkv' in preferences
3546 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3547 return 'mkv' # TODO: any other format allows this?
3549 # TODO: All codecs supported by parse_codecs isn't handled here
3550 COMPATIBLE_CODECS
= {
3552 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3553 'h264', 'aacl', 'ec-3', # Set in ISM
3556 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3557 'vp9x', 'vp8x', # in the webm spec
3561 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3562 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3564 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3565 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3566 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3570 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3573 for ext
in preferences
or vexts
:
3574 current_exts
= {ext, *vexts, *aexts}
3575 if ext
== 'mkv' or current_exts
== {ext}
or any(
3576 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3578 return 'mkv' if allow_mkv
else preferences
[-1]
3581 def urlhandle_detect_ext(url_handle
):
3582 getheader
= url_handle
.headers
.get
3584 cd
= getheader('Content-Disposition')
3586 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3588 e
= determine_ext(m
.group('filename'), default_ext
=None)
3592 return mimetype2ext(getheader('Content-Type'))
3595 def encode_data_uri(data
, mime_type
):
3596 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3599 def age_restricted(content_limit
, age_limit
):
3600 """ Returns True iff the content should be blocked """
3602 if age_limit
is None: # No limit set
3604 if content_limit
is None:
3605 return False # Content available for everyone
3606 return age_limit
< content_limit
3609 # List of known byte-order-marks (BOM)
3611 (b
'\xef\xbb\xbf', 'utf-8'),
3612 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3613 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3614 (b
'\xff\xfe', 'utf-16-le'),
3615 (b
'\xfe\xff', 'utf-16-be'),
3619 def is_html(first_bytes
):
3620 """ Detect whether a file contains HTML by examining its first bytes. """
3623 for bom
, enc
in BOMS
:
3624 while first_bytes
.startswith(bom
):
3625 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3627 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3630 def determine_protocol(info_dict
):
3631 protocol
= info_dict
.get('protocol')
3632 if protocol
is not None:
3635 url
= sanitize_url(info_dict
['url'])
3636 if url
.startswith('rtmp'):
3638 elif url
.startswith('mms'):
3640 elif url
.startswith('rtsp'):
3643 ext
= determine_ext(url
)
3645 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3649 return urllib
.parse
.urlparse(url
).scheme
3652 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3653 """ Render a list of rows, each as a list of values.
3654 Text after a \t will be right aligned """
3656 return len(remove_terminal_sequences(string
).replace('\t', ''))
3658 def get_max_lens(table
):
3659 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3661 def filter_using_list(row
, filterArray
):
3662 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3664 max_lens
= get_max_lens(data
) if hide_empty
else []
3665 header_row
= filter_using_list(header_row
, max_lens
)
3666 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3668 table
= [header_row
] + data
3669 max_lens
= get_max_lens(table
)
3672 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3673 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3675 for pos
, text
in enumerate(map(str, row
)):
3677 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3679 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3680 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3684 def _match_one(filter_part
, dct
, incomplete
):
3685 # TODO: Generalize code with YoutubeDL._build_format_filter
3686 STRING_OPERATORS
= {
3687 '*=': operator
.contains
,
3688 '^=': lambda attr
, value
: attr
.startswith(value
),
3689 '$=': lambda attr
, value
: attr
.endswith(value
),
3690 '~=': lambda attr
, value
: re
.search(value
, attr
),
3692 COMPARISON_OPERATORS
= {
3694 '<=': operator
.le
, # "<=" must be defined above "<"
3701 if isinstance(incomplete
, bool):
3702 is_incomplete
= lambda _
: incomplete
3704 is_incomplete
= lambda k
: k
in incomplete
3706 operator_rex
= re
.compile(r
'''(?x)
3708 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3710 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3713 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3714 m = operator_rex.fullmatch(filter_part.strip())
3717 unnegated_op = COMPARISON_OPERATORS[m['op']]
3719 op = lambda attr, value: not unnegated_op(attr, value)
3722 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3724 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3725 actual_value = dct.get(m['key'])
3726 numeric_comparison = None
3727 if isinstance(actual_value, (int, float)):
3728 # If the original field is a string and matching comparisonvalue is
3729 # a number we should respect the origin of the original field
3730 # and process comparison value as a string (see
3731 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3733 numeric_comparison = int(comparison_value)
3735 numeric_comparison = parse_filesize(comparison_value)
3736 if numeric_comparison is None:
3737 numeric_comparison = parse_filesize(f'{comparison_value}B')
3738 if numeric_comparison is None:
3739 numeric_comparison = parse_duration(comparison_value)
3740 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3741 raise ValueError('Operator %s only supports string values!' % m['op'])
3742 if actual_value is None:
3743 return is_incomplete(m['key']) or m['none_inclusive']
3744 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3747 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3748 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3750 operator_rex = re.compile(r'''(?x
)
3751 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3752 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3753 m = operator_rex.fullmatch(filter_part.strip())
3755 op = UNARY_OPERATORS[m.group('op')]
3756 actual_value = dct.get(m.group('key'))
3757 if is_incomplete(m.group('key')) and actual_value is None:
3759 return op(actual_value)
3761 raise ValueError('Invalid filter part %r' % filter_part)
3764 def match_str(filter_str, dct, incomplete=False):
3765 """ Filter a dictionary with a simple string syntax.
3766 @returns Whether the filter passes
3767 @param incomplete Set of keys that is expected to be missing from dct.
3768 Can be True/False to indicate all/none of the keys may be missing.
3769 All conditions on incomplete keys pass if the key is missing
3772 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3773 for filter_part in re.split(r'(?<!\\)&', filter_str))
3776 def match_filter_func(filters):
3779 filters = set(variadic(filters))
3781 interactive = '-' in filters
3785 def _match_func(info_dict, incomplete=False):
3786 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3787 return NO_DEFAULT if interactive and not incomplete else None
3789 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3790 filter_str = ') | ('.join(map(str.strip, filters))
3791 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3795 class download_range_func:
3796 def __init__(self, chapters, ranges):
3797 self.chapters, self.ranges = chapters, ranges
3799 def __call__(self, info_dict, ydl):
3800 if not self.ranges and not self.chapters:
3803 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3804 else 'Cannot match chapters since chapter information is unavailable')
3805 for regex in self.chapters or []:
3806 for i, chapter in enumerate(info_dict.get('chapters') or []):
3807 if re.search(regex, chapter['title']):
3809 yield {**chapter, 'index': i}
3810 if self.chapters and warning:
3811 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3813 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3815 def __eq__(self, other):
3816 return (isinstance(other, download_range_func)
3817 and self.chapters == other.chapters and self.ranges == other.ranges)
3820 def parse_dfxp_time_expr(time_expr):
3824 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3826 return float(mobj.group('time_offset'))
3828 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3830 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3833 def srt_subtitles_timecode(seconds):
3834 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3837 def ass_subtitles_timecode(seconds):
3838 time = timetuple_from_msec(seconds * 1000)
3839 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3842 def dfxp2srt(dfxp_data):
3844 @param dfxp_data A
bytes-like
object containing DFXP data
3845 @returns A
unicode object containing converted SRT data
3847 LEGACY_NAMESPACES = (
3848 (b'http://www.w3.org/ns/ttml', [
3849 b'http://www.w3.org/2004/11/ttaf1',
3850 b'http://www.w3.org/2006/04/ttaf1',
3851 b'http://www.w3.org/2006/10/ttaf1',
3853 (b'http://www.w3.org/ns/ttml#styling', [
3854 b'http://www.w3.org/ns/ttml#style',
3858 SUPPORTED_STYLING = [
3867 _x = functools.partial(xpath_with_ns, ns_map={
3868 'xml': 'http://www.w3.org/XML/1998/namespace',
3869 'ttml': 'http://www.w3.org/ns/ttml',
3870 'tts': 'http://www.w3.org/ns/ttml#styling',
3876 class TTMLPElementParser:
3878 _unclosed_elements = []
3879 _applied_styles = []
3881 def start(self, tag, attrib):
3882 if tag in (_x('ttml:br'), 'br'):
3885 unclosed_elements = []
3887 element_style_id = attrib.get('style')
3889 style.update(default_style)
3890 if element_style_id:
3891 style.update(styles.get(element_style_id, {}))
3892 for prop in SUPPORTED_STYLING:
3893 prop_val = attrib.get(_x('tts:' + prop))
3895 style[prop] = prop_val
3898 for k, v in sorted(style.items()):
3899 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3902 font += ' color="%s"' % v
3903 elif k == 'fontSize':
3904 font += ' size="%s"' % v
3905 elif k == 'fontFamily':
3906 font += ' face="%s"' % v
3907 elif k == 'fontWeight' and v == 'bold':
3909 unclosed_elements.append('b')
3910 elif k == 'fontStyle' and v == 'italic':
3912 unclosed_elements.append('i')
3913 elif k == 'textDecoration' and v == 'underline':
3915 unclosed_elements.append('u')
3917 self._out += '<font' + font + '>'
3918 unclosed_elements.append('font')
3920 if self._applied_styles:
3921 applied_style.update(self._applied_styles[-1])
3922 applied_style.update(style)
3923 self._applied_styles.append(applied_style)
3924 self._unclosed_elements.append(unclosed_elements)
3927 if tag not in (_x('ttml:br'), 'br'):
3928 unclosed_elements = self._unclosed_elements.pop()
3929 for element in reversed(unclosed_elements):
3930 self._out += '</%s>' % element
3931 if unclosed_elements and self._applied_styles:
3932 self._applied_styles.pop()
3934 def data(self, data):
3938 return self._out.strip()
3940 def parse_node(node):
3941 target = TTMLPElementParser()
3942 parser = xml.etree.ElementTree.XMLParser(target=target)
3943 parser.feed(xml.etree.ElementTree.tostring(node))
3944 return parser.close()
3946 for k, v in LEGACY_NAMESPACES:
3948 dfxp_data = dfxp_data.replace(ns, k)
3950 dfxp = compat_etree_fromstring(dfxp_data)
3952 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3955 raise ValueError('Invalid dfxp/TTML subtitle')
3959 for style in dfxp.findall(_x('.//ttml:style')):
3960 style_id = style.get('id') or style.get(_x('xml:id'))
3963 parent_style_id = style.get('style')
3965 if parent_style_id not in styles:
3968 styles[style_id] = styles[parent_style_id].copy()
3969 for prop in SUPPORTED_STYLING:
3970 prop_val = style.get(_x('tts:' + prop))
3972 styles.setdefault(style_id, {})[prop] = prop_val
3978 for p in ('body', 'div'):
3979 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3982 style = styles.get(ele.get('style'))
3985 default_style.update(style)
3987 for para, index in zip(paras, itertools.count(1)):
3988 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3989 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3990 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3991 if begin_time is None:
3996 end_time = begin_time + dur
3997 out.append('%d\n%s --> %s\n%s\n\n' % (
3999 srt_subtitles_timecode(begin_time),
4000 srt_subtitles_timecode(end_time),
4006 def cli_option(params, command_option, param, separator=None):
4007 param = params.get(param)
4008 return ([] if param is None
4009 else [command_option, str(param)] if separator is None
4010 else [f'{command_option}{separator}{param}'])
4013 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4014 param = params.get(param)
4015 assert param in (True, False, None)
4016 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4019 def cli_valueless_option(params, command_option, param, expected_value=True):
4020 return [command_option] if params.get(param) == expected_value else []
4023 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4024 if isinstance(argdict, (list, tuple)): # for backward compatibility
4031 assert isinstance(argdict, dict)
4033 assert isinstance(keys, (list, tuple))
4034 for key_list in keys:
4035 arg_list = list(filter(
4036 lambda x: x is not None,
4037 [argdict.get(key.lower()) for key in variadic(key_list)]))
4039 return [arg for args in arg_list for arg in args]
4043 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4044 main_key, exe = main_key.lower(), exe.lower()
4045 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4046 keys = [f'{root_key}{k}' for k in (keys or [''])]
4047 if root_key in keys:
4049 keys.append((main_key, exe))
4050 keys.append('default')
4053 return cli_configuration_args(argdict, keys, default, use_compat)
4057 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4116 'iw': 'heb', # Replaced by he in 1989 revision
4126 'in': 'ind', # Replaced by id in 1989 revision
4241 'ji': 'yid', # Replaced by yi in 1989 revision
4249 def short2long(cls, code):
4250 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4251 return cls._lang_map.get(code[:2])
4254 def long2short(cls, code):
4255 """Convert language code from ISO 639-2/T to ISO 639-1"""
4256 for short_name, long_name in cls._lang_map.items():
4257 if long_name == code:
4262 # From http://data.okfn.org/data/core/country-list
4264 'AF': 'Afghanistan',
4265 'AX': 'Åland Islands',
4268 'AS': 'American Samoa',
4273 'AG': 'Antigua and Barbuda',
4290 'BO': 'Bolivia, Plurinational State of',
4291 'BQ': 'Bonaire, Sint Eustatius and Saba',
4292 'BA': 'Bosnia and Herzegovina',
4294 'BV': 'Bouvet Island',
4296 'IO': 'British Indian Ocean Territory',
4297 'BN': 'Brunei Darussalam',
4299 'BF': 'Burkina Faso',
4305 'KY': 'Cayman Islands',
4306 'CF': 'Central African Republic',
4310 'CX': 'Christmas Island',
4311 'CC': 'Cocos (Keeling) Islands',
4315 'CD': 'Congo, the Democratic Republic of the',
4316 'CK': 'Cook Islands',
4318 'CI': 'Côte d\'Ivoire',
4323 'CZ': 'Czech Republic',
4327 'DO': 'Dominican Republic',
4330 'SV': 'El Salvador',
4331 'GQ': 'Equatorial Guinea',
4335 'FK': 'Falkland Islands (Malvinas)',
4336 'FO': 'Faroe Islands',
4340 'GF': 'French Guiana',
4341 'PF': 'French Polynesia',
4342 'TF': 'French Southern Territories',
4357 'GW': 'Guinea-Bissau',
4360 'HM': 'Heard Island and McDonald Islands',
4361 'VA': 'Holy See (Vatican City State)',
4368 'IR': 'Iran, Islamic Republic of',
4371 'IM': 'Isle of Man',
4381 'KP': 'Korea, Democratic People\'s Republic of',
4382 'KR': 'Korea, Republic of',
4385 'LA': 'Lao People\'s Democratic Republic',
4391 'LI': 'Liechtenstein',
4395 'MK': 'Macedonia, the Former Yugoslav Republic of',
4402 'MH': 'Marshall Islands',
4408 'FM': 'Micronesia, Federated States of',
4409 'MD': 'Moldova, Republic of',
4420 'NL': 'Netherlands',
4421 'NC': 'New Caledonia',
4422 'NZ': 'New Zealand',
4427 'NF': 'Norfolk Island',
4428 'MP': 'Northern Mariana Islands',
4433 'PS': 'Palestine, State of',
4435 'PG': 'Papua New Guinea',
4438 'PH': 'Philippines',
4442 'PR': 'Puerto Rico',
4446 'RU': 'Russian Federation',
4448 'BL': 'Saint Barthélemy',
4449 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4450 'KN': 'Saint Kitts and Nevis',
4451 'LC': 'Saint Lucia',
4452 'MF': 'Saint Martin (French part)',
4453 'PM': 'Saint Pierre and Miquelon',
4454 'VC': 'Saint Vincent and the Grenadines',
4457 'ST': 'Sao Tome and Principe',
4458 'SA': 'Saudi Arabia',
4462 'SL': 'Sierra Leone',
4464 'SX': 'Sint Maarten (Dutch part)',
4467 'SB': 'Solomon Islands',
4469 'ZA': 'South Africa',
4470 'GS': 'South Georgia and the South Sandwich Islands',
4471 'SS': 'South Sudan',
4476 'SJ': 'Svalbard and Jan Mayen',
4479 'CH': 'Switzerland',
4480 'SY': 'Syrian Arab Republic',
4481 'TW': 'Taiwan, Province of China',
4483 'TZ': 'Tanzania, United Republic of',
4485 'TL': 'Timor-Leste',
4489 'TT': 'Trinidad and Tobago',
4492 'TM': 'Turkmenistan',
4493 'TC': 'Turks and Caicos Islands',
4497 'AE': 'United Arab Emirates',
4498 'GB': 'United Kingdom',
4499 'US': 'United States',
4500 'UM': 'United States Minor Outlying Islands',
4504 'VE': 'Venezuela, Bolivarian Republic of',
4506 'VG': 'Virgin Islands, British',
4507 'VI': 'Virgin Islands, U.S.',
4508 'WF': 'Wallis and Futuna',
4509 'EH': 'Western Sahara',
4513 # Not ISO 3166 codes, but used for IP blocks
4514 'AP': 'Asia/Pacific Region',
4519 def short2full(cls, code):
4520 """Convert an ISO 3166-2 country code to the corresponding full name"""
4521 return cls._country_map.get(code.upper())
4525 # Major IPv4 address blocks per country
4527 'AD': '46.172.224.0/19',
4528 'AE': '94.200.0.0/13',
4529 'AF': '149.54.0.0/17',
4530 'AG': '209.59.64.0/18',
4531 'AI': '204.14.248.0/21',
4532 'AL': '46.99.0.0/16',
4533 'AM': '46.70.0.0/15',
4534 'AO': '105.168.0.0/13',
4535 'AP': '182.50.184.0/21',
4536 'AQ': '23.154.160.0/24',
4537 'AR': '181.0.0.0/12',
4538 'AS': '202.70.112.0/20',
4539 'AT': '77.116.0.0/14',
4540 'AU': '1.128.0.0/11',
4541 'AW': '181.41.0.0/18',
4542 'AX': '185.217.4.0/22',
4543 'AZ': '5.197.0.0/16',
4544 'BA': '31.176.128.0/17',
4545 'BB': '65.48.128.0/17',
4546 'BD': '114.130.0.0/16',
4548 'BF': '102.178.0.0/15',
4549 'BG': '95.42.0.0/15',
4550 'BH': '37.131.0.0/17',
4551 'BI': '154.117.192.0/18',
4552 'BJ': '137.255.0.0/16',
4553 'BL': '185.212.72.0/23',
4554 'BM': '196.12.64.0/18',
4555 'BN': '156.31.0.0/16',
4556 'BO': '161.56.0.0/16',
4557 'BQ': '161.0.80.0/20',
4558 'BR': '191.128.0.0/12',
4559 'BS': '24.51.64.0/18',
4560 'BT': '119.2.96.0/19',
4561 'BW': '168.167.0.0/16',
4562 'BY': '178.120.0.0/13',
4563 'BZ': '179.42.192.0/18',
4564 'CA': '99.224.0.0/11',
4565 'CD': '41.243.0.0/16',
4566 'CF': '197.242.176.0/21',
4567 'CG': '160.113.0.0/16',
4568 'CH': '85.0.0.0/13',
4569 'CI': '102.136.0.0/14',
4570 'CK': '202.65.32.0/19',
4571 'CL': '152.172.0.0/14',
4572 'CM': '102.244.0.0/14',
4573 'CN': '36.128.0.0/10',
4574 'CO': '181.240.0.0/12',
4575 'CR': '201.192.0.0/12',
4576 'CU': '152.206.0.0/15',
4577 'CV': '165.90.96.0/19',
4578 'CW': '190.88.128.0/17',
4579 'CY': '31.153.0.0/16',
4580 'CZ': '88.100.0.0/14',
4582 'DJ': '197.241.0.0/17',
4583 'DK': '87.48.0.0/12',
4584 'DM': '192.243.48.0/20',
4585 'DO': '152.166.0.0/15',
4586 'DZ': '41.96.0.0/12',
4587 'EC': '186.68.0.0/15',
4588 'EE': '90.190.0.0/15',
4589 'EG': '156.160.0.0/11',
4590 'ER': '196.200.96.0/20',
4591 'ES': '88.0.0.0/11',
4592 'ET': '196.188.0.0/14',
4593 'EU': '2.16.0.0/13',
4594 'FI': '91.152.0.0/13',
4595 'FJ': '144.120.0.0/16',
4596 'FK': '80.73.208.0/21',
4597 'FM': '119.252.112.0/20',
4598 'FO': '88.85.32.0/19',
4600 'GA': '41.158.0.0/15',
4602 'GD': '74.122.88.0/21',
4603 'GE': '31.146.0.0/16',
4604 'GF': '161.22.64.0/18',
4605 'GG': '62.68.160.0/19',
4606 'GH': '154.160.0.0/12',
4607 'GI': '95.164.0.0/16',
4608 'GL': '88.83.0.0/19',
4609 'GM': '160.182.0.0/15',
4610 'GN': '197.149.192.0/18',
4611 'GP': '104.250.0.0/19',
4612 'GQ': '105.235.224.0/20',
4613 'GR': '94.64.0.0/13',
4614 'GT': '168.234.0.0/16',
4615 'GU': '168.123.0.0/16',
4616 'GW': '197.214.80.0/20',
4617 'GY': '181.41.64.0/18',
4618 'HK': '113.252.0.0/14',
4619 'HN': '181.210.0.0/16',
4620 'HR': '93.136.0.0/13',
4621 'HT': '148.102.128.0/17',
4622 'HU': '84.0.0.0/14',
4623 'ID': '39.192.0.0/10',
4624 'IE': '87.32.0.0/12',
4625 'IL': '79.176.0.0/13',
4626 'IM': '5.62.80.0/20',
4627 'IN': '117.192.0.0/10',
4628 'IO': '203.83.48.0/21',
4629 'IQ': '37.236.0.0/14',
4630 'IR': '2.176.0.0/12',
4631 'IS': '82.221.0.0/16',
4632 'IT': '79.0.0.0/10',
4633 'JE': '87.244.64.0/18',
4634 'JM': '72.27.0.0/17',
4635 'JO': '176.29.0.0/16',
4636 'JP': '133.0.0.0/8',
4637 'KE': '105.48.0.0/12',
4638 'KG': '158.181.128.0/17',
4639 'KH': '36.37.128.0/17',
4640 'KI': '103.25.140.0/22',
4641 'KM': '197.255.224.0/20',
4642 'KN': '198.167.192.0/19',
4643 'KP': '175.45.176.0/22',
4644 'KR': '175.192.0.0/10',
4645 'KW': '37.36.0.0/14',
4646 'KY': '64.96.0.0/15',
4647 'KZ': '2.72.0.0/13',
4648 'LA': '115.84.64.0/18',
4649 'LB': '178.135.0.0/16',
4650 'LC': '24.92.144.0/20',
4651 'LI': '82.117.0.0/19',
4652 'LK': '112.134.0.0/15',
4653 'LR': '102.183.0.0/16',
4654 'LS': '129.232.0.0/17',
4655 'LT': '78.56.0.0/13',
4656 'LU': '188.42.0.0/16',
4657 'LV': '46.109.0.0/16',
4658 'LY': '41.252.0.0/14',
4659 'MA': '105.128.0.0/11',
4660 'MC': '88.209.64.0/18',
4661 'MD': '37.246.0.0/16',
4662 'ME': '178.175.0.0/17',
4663 'MF': '74.112.232.0/21',
4664 'MG': '154.126.0.0/17',
4665 'MH': '117.103.88.0/21',
4666 'MK': '77.28.0.0/15',
4667 'ML': '154.118.128.0/18',
4668 'MM': '37.111.0.0/17',
4669 'MN': '49.0.128.0/17',
4670 'MO': '60.246.0.0/16',
4671 'MP': '202.88.64.0/20',
4672 'MQ': '109.203.224.0/19',
4673 'MR': '41.188.64.0/18',
4674 'MS': '208.90.112.0/22',
4675 'MT': '46.11.0.0/16',
4676 'MU': '105.16.0.0/12',
4677 'MV': '27.114.128.0/18',
4678 'MW': '102.70.0.0/15',
4679 'MX': '187.192.0.0/11',
4680 'MY': '175.136.0.0/13',
4681 'MZ': '197.218.0.0/15',
4682 'NA': '41.182.0.0/16',
4683 'NC': '101.101.0.0/18',
4684 'NE': '197.214.0.0/18',
4685 'NF': '203.17.240.0/22',
4686 'NG': '105.112.0.0/12',
4687 'NI': '186.76.0.0/15',
4688 'NL': '145.96.0.0/11',
4689 'NO': '84.208.0.0/13',
4690 'NP': '36.252.0.0/15',
4691 'NR': '203.98.224.0/19',
4692 'NU': '49.156.48.0/22',
4693 'NZ': '49.224.0.0/14',
4694 'OM': '5.36.0.0/15',
4695 'PA': '186.72.0.0/15',
4696 'PE': '186.160.0.0/14',
4697 'PF': '123.50.64.0/18',
4698 'PG': '124.240.192.0/19',
4699 'PH': '49.144.0.0/13',
4700 'PK': '39.32.0.0/11',
4701 'PL': '83.0.0.0/11',
4702 'PM': '70.36.0.0/20',
4703 'PR': '66.50.0.0/16',
4704 'PS': '188.161.0.0/16',
4705 'PT': '85.240.0.0/13',
4706 'PW': '202.124.224.0/20',
4707 'PY': '181.120.0.0/14',
4708 'QA': '37.210.0.0/15',
4709 'RE': '102.35.0.0/16',
4710 'RO': '79.112.0.0/13',
4711 'RS': '93.86.0.0/15',
4712 'RU': '5.136.0.0/13',
4713 'RW': '41.186.0.0/16',
4714 'SA': '188.48.0.0/13',
4715 'SB': '202.1.160.0/19',
4716 'SC': '154.192.0.0/11',
4717 'SD': '102.120.0.0/13',
4718 'SE': '78.64.0.0/12',
4719 'SG': '8.128.0.0/10',
4720 'SI': '188.196.0.0/14',
4721 'SK': '78.98.0.0/15',
4722 'SL': '102.143.0.0/17',
4723 'SM': '89.186.32.0/19',
4724 'SN': '41.82.0.0/15',
4725 'SO': '154.115.192.0/18',
4726 'SR': '186.179.128.0/17',
4727 'SS': '105.235.208.0/21',
4728 'ST': '197.159.160.0/19',
4729 'SV': '168.243.0.0/16',
4730 'SX': '190.102.0.0/20',
4732 'SZ': '41.84.224.0/19',
4733 'TC': '65.255.48.0/20',
4734 'TD': '154.68.128.0/19',
4735 'TG': '196.168.0.0/14',
4736 'TH': '171.96.0.0/13',
4737 'TJ': '85.9.128.0/18',
4738 'TK': '27.96.24.0/21',
4739 'TL': '180.189.160.0/20',
4740 'TM': '95.85.96.0/19',
4741 'TN': '197.0.0.0/11',
4742 'TO': '175.176.144.0/21',
4743 'TR': '78.160.0.0/11',
4744 'TT': '186.44.0.0/15',
4745 'TV': '202.2.96.0/19',
4746 'TW': '120.96.0.0/11',
4747 'TZ': '156.156.0.0/14',
4748 'UA': '37.52.0.0/14',
4749 'UG': '102.80.0.0/13',
4751 'UY': '167.56.0.0/13',
4752 'UZ': '84.54.64.0/18',
4753 'VA': '212.77.0.0/19',
4754 'VC': '207.191.240.0/21',
4755 'VE': '186.88.0.0/13',
4756 'VG': '66.81.192.0/20',
4757 'VI': '146.226.0.0/16',
4758 'VN': '14.160.0.0/11',
4759 'VU': '202.80.32.0/20',
4760 'WF': '117.20.32.0/21',
4761 'WS': '202.4.32.0/19',
4762 'YE': '134.35.0.0/16',
4763 'YT': '41.242.116.0/22',
4764 'ZA': '41.0.0.0/11',
4765 'ZM': '102.144.0.0/13',
4766 'ZW': '102.177.192.0/18',
4770 def random_ipv4(cls, code_or_block):
4771 if len(code_or_block) == 2:
4772 block = cls._country_ip_map.get(code_or_block.upper())
4776 block = code_or_block
4777 addr, preflen = block.split('/')
4778 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4779 addr_max = addr_min | (0xffffffff >> int(preflen))
4780 return str(socket.inet_ntoa(
4781 struct.pack('!L', random.randint(addr_min, addr_max))))
4784 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4785 def __init__(self, proxies=None):
4786 # Set default handlers
4787 for type in ('http', 'https'):
4788 setattr(self, '%s_open' % type,
4789 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4790 meth(r, proxy, type))
4791 urllib.request.ProxyHandler.__init__(self, proxies)
4793 def proxy_open(self, req, proxy, type):
4794 req_proxy = req.headers.get('Ytdl-request-proxy')
4795 if req_proxy is not None:
4797 del req.headers['Ytdl-request-proxy']
4799 if proxy == '__noproxy__':
4800 return None # No Proxy
4801 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4802 req.add_header('Ytdl-socks-proxy', proxy)
4803 # yt-dlp's http/https handlers do wrapping the socket with socks
4805 return urllib.request.ProxyHandler.proxy_open(
4806 self, req, proxy, type)
4809 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4810 # released into Public Domain
4811 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4813 def long_to_bytes(n, blocksize=0):
4814 """long_to_bytes(n:long, blocksize:int) : string
4815 Convert a long integer to a byte string.
4817 If optional blocksize is given and greater than zero, pad the front of the
4818 byte string with binary zeros so that the length is a multiple of
4821 # after much testing, this algorithm was deemed to be the fastest
4825 s = struct.pack('>I', n & 0xffffffff) + s
4827 # strip off leading zeros
4828 for i in range(len(s)):
4829 if s[i] != b'\000'[0]:
4832 # only happens when n == 0
4836 # add back some pad bytes. this could be done more efficiently w.r.t. the
4837 # de-padding being done above, but sigh...
4838 if blocksize > 0 and len(s) % blocksize:
4839 s = (blocksize - len(s) % blocksize) * b'\000' + s
4843 def bytes_to_long(s):
4844 """bytes_to_long(string) : long
4845 Convert a byte string to a long integer.
4847 This is (essentially) the inverse of long_to_bytes().
4852 extra = (4 - length % 4)
4853 s = b'\000' * extra + s
4854 length = length + extra
4855 for i in range(0, length, 4):
4856 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4860 def ohdave_rsa_encrypt(data, exponent, modulus):
4862 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4865 data: data to encrypt, bytes-like object
4866 exponent, modulus: parameter e and N of RSA algorithm, both integer
4867 Output: hex string of encrypted data
4869 Limitation: supports one block encryption only
4872 payload = int(binascii.hexlify(data[::-1]), 16)
4873 encrypted = pow(payload, exponent, modulus)
4874 return '%x' % encrypted
4877 def pkcs1pad(data, length):
4879 Padding input data with PKCS#1 scheme
4881 @param {int[]} data input data
4882 @param {int} length target length
4883 @returns {int[]} padded data
4885 if len(data) > length - 11:
4886 raise ValueError('Input data too
long for PKCS
#1 padding')
4888 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4889 return [0, 2] + pseudo_random
+ [0] + data
4892 def _base_n_table(n
, table
):
4893 if not table
and not n
:
4894 raise ValueError('Either table or n must be specified')
4895 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4897 if n
and n
!= len(table
):
4898 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4902 def encode_base_n(num
, n
=None, table
=None):
4903 """Convert given int to a base-n string"""
4904 table
= _base_n_table(n
, table
)
4908 result
, base
= '', len(table
)
4910 result
= table
[num
% base
] + result
4915 def decode_base_n(string
, n
=None, table
=None):
4916 """Convert given base-n string to int"""
4917 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4918 result
, base
= 0, len(table
)
4920 result
= result
* base
+ table
[char
]
4924 def decode_base(value
, digits
):
4925 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
4926 f
'in a future version. Use {__name__}.decode_base_n instead')
4927 return decode_base_n(value
, table
=digits
)
4930 def decode_packed_codes(code
):
4931 mobj
= re
.search(PACKED_CODES_RE
, code
)
4932 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4935 symbols
= symbols
.split('|')
4940 base_n_count
= encode_base_n(count
, base
)
4941 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4944 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4948 def caesar(s
, alphabet
, shift
):
4953 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4958 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4961 def parse_m3u8_attributes(attrib
):
4963 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4964 if val
.startswith('"'):
4970 def urshift(val
, n
):
4971 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4974 # Based on png2str() written by @gdkchan and improved by @yokrysty
4975 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4976 def decode_png(png_data
):
4977 # Reference: https://www.w3.org/TR/PNG/
4978 header
= png_data
[8:]
4980 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4981 raise OSError('Not a valid PNG file.')
4983 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4984 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
4989 length
= unpack_integer(header
[:4])
4992 chunk_type
= header
[:4]
4995 chunk_data
= header
[:length
]
4996 header
= header
[length
:]
4998 header
= header
[4:] # Skip CRC
5006 ihdr
= chunks
[0]['data']
5008 width
= unpack_integer(ihdr
[:4])
5009 height
= unpack_integer(ihdr
[4:8])
5013 for chunk
in chunks
:
5014 if chunk
['type'] == b
'IDAT':
5015 idat
+= chunk
['data']
5018 raise OSError('Unable to read PNG data.')
5020 decompressed_data
= bytearray(zlib
.decompress(idat
))
5025 def _get_pixel(idx
):
5030 for y
in range(height
):
5031 basePos
= y
* (1 + stride
)
5032 filter_type
= decompressed_data
[basePos
]
5036 pixels
.append(current_row
)
5038 for x
in range(stride
):
5039 color
= decompressed_data
[1 + basePos
+ x
]
5040 basex
= y
* stride
+ x
5045 left
= _get_pixel(basex
- 3)
5047 up
= _get_pixel(basex
- stride
)
5049 if filter_type
== 1: # Sub
5050 color
= (color
+ left
) & 0xff
5051 elif filter_type
== 2: # Up
5052 color
= (color
+ up
) & 0xff
5053 elif filter_type
== 3: # Average
5054 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5055 elif filter_type
== 4: # Paeth
5061 c
= _get_pixel(basex
- stride
- 3)
5069 if pa
<= pb
and pa
<= pc
:
5070 color
= (color
+ a
) & 0xff
5072 color
= (color
+ b
) & 0xff
5074 color
= (color
+ c
) & 0xff
5076 current_row
.append(color
)
5078 return width
, height
, pixels
5081 def write_xattr(path
, key
, value
):
5082 # Windows: Write xattrs to NTFS Alternate Data Streams:
5083 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5084 if compat_os_name
== 'nt':
5085 assert ':' not in key
5086 assert os
.path
.exists(path
)
5089 with open(f
'{path}:{key}', 'wb') as f
:
5091 except OSError as e
:
5092 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5095 # UNIX Method 1. Use xattrs/pyxattrs modules
5098 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5099 # Unicode arguments are not supported in pyxattr until version 0.5.0
5100 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5101 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5102 setxattr
= xattr
.set
5104 setxattr
= xattr
.setxattr
5108 setxattr(path
, key
, value
)
5109 except OSError as e
:
5110 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5113 # UNIX Method 2. Use setfattr/xattr executables
5114 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5115 else 'xattr' if check_executable('xattr', ['-h']) else None)
5117 raise XAttrUnavailableError(
5118 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5119 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5121 value
= value
.decode()
5123 _
, stderr
, returncode
= Popen
.run(
5124 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5125 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5126 except OSError as e
:
5127 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5129 raise XAttrMetadataError(returncode
, stderr
)
5132 def random_birthday(year_field
, month_field
, day_field
):
5133 start_date
= datetime
.date(1950, 1, 1)
5134 end_date
= datetime
.date(1995, 12, 31)
5135 offset
= random
.randint(0, (end_date
- start_date
).days
)
5136 random_date
= start_date
+ datetime
.timedelta(offset
)
5138 year_field
: str(random_date
.year
),
5139 month_field
: str(random_date
.month
),
5140 day_field
: str(random_date
.day
),
5144 # Templates for internet shortcut files, which are plain text files.
5145 DOT_URL_LINK_TEMPLATE
= '''\
5150 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5151 <?xml version="1.0" encoding="UTF-8"?>
5152 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5153 <plist version="1.0">
5156 \t<string>%(url)s</string>
5161 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5171 'url': DOT_URL_LINK_TEMPLATE
,
5172 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5173 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5177 def iri_to_uri(iri
):
5179 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5181 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.
5184 iri_parts
= urllib
.parse
.urlparse(iri
)
5186 if '[' in iri_parts
.netloc
:
5187 raise ValueError('IPv6 URIs are not, yet, supported.')
5188 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5190 # 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.
5193 if iri_parts
.username
:
5194 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5195 if iri_parts
.password
is not None:
5196 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5199 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5200 # The 'idna' encoding produces ASCII text.
5201 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5202 net_location
+= ':' + str(iri_parts
.port
)
5204 return urllib
.parse
.urlunparse(
5208 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5210 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5211 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5213 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5214 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5216 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5218 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5221 def to_high_limit_path(path
):
5222 if sys
.platform
in ['win32', 'cygwin']:
5223 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5224 return '\\\\?\\' + os
.path
.abspath(path
)
5229 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5230 val
= traverse_obj(obj
, *variadic(field
))
5231 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5233 return template
% func(val
)
5236 def clean_podcast_url(url
):
5237 return re
.sub(r
'''(?x)
5241 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5244 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5247 cn\.co| # https://podcorn.com/analytics-prefix/
5248 st\.fm # https://podsights.com/docs/
5253 _HEX_TABLE
= '0123456789abcdef'
5256 def random_uuidv4():
5257 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5260 def make_dir(path
, to_screen
=None):
5262 dn
= os
.path
.dirname(path
)
5263 if dn
and not os
.path
.exists(dn
):
5266 except OSError as err
:
5267 if callable(to_screen
) is not None:
5268 to_screen('unable to create directory ' + error_to_compat_str(err
))
5272 def get_executable_path():
5273 from .update
import _get_variant_and_executable_path
5275 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5278 def load_plugins(name
, suffix
, namespace
):
5280 with contextlib
.suppress(FileNotFoundError
):
5281 plugins_spec
= importlib
.util
.spec_from_file_location(
5282 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5283 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5284 sys
.modules
[plugins_spec
.name
] = plugins
5285 plugins_spec
.loader
.exec_module(plugins
)
5286 for name
in dir(plugins
):
5287 if name
in namespace
:
5289 if not name
.endswith(suffix
):
5291 klass
= getattr(plugins
, name
)
5292 classes
[name
] = namespace
[name
] = klass
5297 obj
, *paths
, default
=None, expected_type
=None, get_all
=True,
5298 casesense
=True, is_user_input
=False, traverse_string
=False):
5300 Safely traverse nested `dict`s and `Sequence`s
5302 >>> obj = [{}, {"key": "value"}]
5303 >>> traverse_obj(obj, (1, "key"))
5306 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5307 A value of None is treated as the absence of a value.
5309 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5311 The keys in the path can be one of:
5312 - `None`: Return the current object.
5313 - `str`/`int`: Return `obj[key]`.
5314 - `slice`: Branch out and return all values in `obj[key]`.
5315 - `Ellipsis`: Branch out and return a list of all values.
5316 - `tuple`/`list`: Branch out and return a list of all matching values.
5317 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5318 - `function`: Branch out and return values filtered by the function.
5319 Read as: `[value for key, value in obj if function(key, value)]`.
5320 For `Sequence`s, `key` is the index of the value.
5321 - `dict` Transform the current object and return a matching dict.
5322 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5324 `tuple`, `list`, and `dict` all support nested paths and branches
5326 @params paths Paths which to traverse by.
5327 @param default Value to return if the paths do not match.
5328 @param expected_type If a `type`, only accept final values of this type.
5329 If any other callable, try to call the function on each result.
5330 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5331 @param casesense If `False`, consider string dictionary keys as case insensitive.
5333 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5335 @param is_user_input Whether the keys are generated from user input.
5336 If `True` strings get converted to `int`/`slice` if needed.
5337 @param traverse_string Whether to traverse into objects as strings.
5338 If `True`, any non-compatible object will first be
5339 converted into a string and then traversed into.
5342 @returns The result of the object traversal.
5343 If successful, `get_all=True`, and the path branches at least once,
5344 then a list of results is returned instead.
5346 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5347 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5349 if isinstance(expected_type
, type):
5350 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5352 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5354 def apply_key(key
, obj
):
5361 elif isinstance(key
, (list, tuple)):
5363 _
, result
= apply_path(obj
, branch
)
5367 if isinstance(obj
, collections
.abc
.Mapping
):
5368 yield from obj
.values()
5369 elif is_sequence(obj
):
5371 elif traverse_string
:
5375 if is_sequence(obj
):
5376 iter_obj
= enumerate(obj
)
5377 elif isinstance(obj
, collections
.abc
.Mapping
):
5378 iter_obj
= obj
.items()
5379 elif traverse_string
:
5380 iter_obj
= enumerate(str(obj
))
5383 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5385 elif isinstance(key
, dict):
5386 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5387 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5388 if v
is not None or default
is not None}
5390 elif isinstance(obj
, dict):
5391 yield (obj
.get(key
) if casesense
or (key
in obj
)
5392 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5396 key
= (int_or_none(key
) if ':' not in key
5397 else slice(*map(int_or_none
, key
.split(':'))))
5399 if not isinstance(key
, (int, slice)):
5402 if not is_sequence(obj
):
5403 if not traverse_string
:
5407 with contextlib
.suppress(IndexError):
5410 def apply_path(start_obj
, path
):
5412 has_branched
= False
5414 for key
in variadic(path
):
5415 if is_user_input
and key
== ':':
5418 if not casesense
and isinstance(key
, str):
5419 key
= key
.casefold()
5421 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5424 key_func
= functools
.partial(apply_key
, key
)
5425 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5427 return has_branched
, objs
5429 def _traverse_obj(obj
, path
):
5430 has_branched
, results
= apply_path(obj
, path
)
5431 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5433 return results
.exhaust() if get_all
and has_branched
else results
[0]
5436 result
= _traverse_obj(obj
, path
)
5437 if result
is not None:
5443 def traverse_dict(dictn
, keys
, casesense
=True):
5444 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5445 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5446 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5449 def get_first(obj
, keys
, **kwargs
):
5450 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5453 def time_seconds(**kwargs
):
5454 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5455 return t
.timestamp()
5458 # create a JSON Web Signature (jws) with HS256 algorithm
5459 # the resulting format is in JWS Compact Serialization
5460 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5461 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5462 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5468 header_data
.update(headers
)
5469 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5470 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5471 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5472 signature_b64
= base64
.b64encode(h
.digest())
5473 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5477 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5478 def jwt_decode_hs256(jwt
):
5479 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5480 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5484 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5488 def supports_terminal_sequences(stream
):
5489 if compat_os_name
== 'nt':
5490 if not WINDOWS_VT_MODE
:
5492 elif not os
.getenv('TERM'):
5495 return stream
.isatty()
5496 except BaseException
:
5500 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5501 if get_windows_version() < (10, 0, 10586):
5503 global WINDOWS_VT_MODE
5505 Popen
.run('', shell
=True)
5509 WINDOWS_VT_MODE
= True
5510 supports_terminal_sequences
.cache_clear()
5513 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5516 def remove_terminal_sequences(string
):
5517 return _terminal_sequences_re
.sub('', string
)
5520 def number_of_digits(number
):
5521 return len('%d' % number
)
5524 def join_nonempty(*values
, delim
='-', from_dict
=None):
5525 if from_dict
is not None:
5526 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5527 return delim
.join(map(str, filter(None, values
)))
5530 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5532 Find the largest format dimensions in terms of video width and, for each thumbnail:
5533 * Modify the URL: Match the width with the provided regex and replace with the former width
5536 This function is useful with video services that scale the provided thumbnails on demand
5538 _keys
= ('width', 'height')
5539 max_dimensions
= max(
5540 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5542 if not max_dimensions
[0]:
5546 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5547 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5548 for thumbnail
in thumbnails
5552 def parse_http_range(range):
5553 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5555 return None, None, None
5556 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5558 return None, None, None
5559 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5562 def read_stdin(what
):
5563 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5564 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5568 def determine_file_encoding(data
):
5570 Detect the text encoding used
5571 @returns (encoding, bytes to skip)
5574 # BOM marks are given priority over declarations
5575 for bom
, enc
in BOMS
:
5576 if data
.startswith(bom
):
5577 return enc
, len(bom
)
5579 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5580 # We ignore the endianness to get a good enough match
5581 data
= data
.replace(b
'\0', b
'')
5582 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5583 return mobj
.group(1).decode() if mobj
else None, 0
5590 __initialized
= False
5592 def __init__(self
, parser
, label
=None):
5593 self
.parser
, self
.label
= parser
, label
5594 self
._loaded
_paths
, self
.configs
= set(), []
5596 def init(self
, args
=None, filename
=None):
5597 assert not self
.__initialized
5598 self
.own_args
, self
.filename
= args
, filename
5599 return self
.load_configs()
5601 def load_configs(self
):
5604 location
= os
.path
.realpath(self
.filename
)
5605 directory
= os
.path
.dirname(location
)
5606 if location
in self
._loaded
_paths
:
5608 self
._loaded
_paths
.add(location
)
5610 self
.__initialized
= True
5611 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5612 self
.parsed_args
= self
.own_args
5613 for location
in opts
.config_locations
or []:
5615 if location
in self
._loaded
_paths
:
5617 self
._loaded
_paths
.add(location
)
5618 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5620 location
= os
.path
.join(directory
, expand_path(location
))
5621 if os
.path
.isdir(location
):
5622 location
= os
.path
.join(location
, 'yt-dlp.conf')
5623 if not os
.path
.exists(location
):
5624 self
.parser
.error(f
'config location {location} does not exist')
5625 self
.append_config(self
.read_file(location
), location
)
5629 label
= join_nonempty(
5630 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5632 return join_nonempty(
5633 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5634 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5638 def read_file(filename
, default
=[]):
5640 optionf
= open(filename
, 'rb')
5642 return default
# silently skip if file is not present
5644 enc
, skip
= determine_file_encoding(optionf
.read(512))
5645 optionf
.seek(skip
, io
.SEEK_SET
)
5647 enc
= None # silently skip read errors
5649 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5650 contents
= optionf
.read().decode(enc
or preferredencoding())
5651 res
= shlex
.split(contents
, comments
=True)
5652 except Exception as err
:
5653 raise ValueError(f
'Unable to parse "{filename}": {err}')
5659 def hide_login_info(opts
):
5660 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5661 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5666 return m
.group('key') + '=PRIVATE'
5670 opts
= list(map(_scrub_eq
, opts
))
5671 for idx
, opt
in enumerate(opts
):
5672 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5673 opts
[idx
+ 1] = 'PRIVATE'
5676 def append_config(self
, *args
, label
=None):
5677 config
= type(self
)(self
.parser
, label
)
5678 config
._loaded
_paths
= self
._loaded
_paths
5679 if config
.init(*args
):
5680 self
.configs
.append(config
)
5684 for config
in reversed(self
.configs
):
5685 yield from config
.all_args
5686 yield from self
.parsed_args
or []
5688 def parse_known_args(self
, **kwargs
):
5689 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5691 def parse_args(self
):
5692 return self
.parser
.parse_args(self
.all_args
)
5695 class WebSocketsWrapper():
5696 """Wraps websockets module to use in non-async scopes"""
5699 def __init__(self
, url
, headers
=None, connect
=True):
5700 self
.loop
= asyncio
.new_event_loop()
5701 # XXX: "loop" is deprecated
5702 self
.conn
= websockets
.connect(
5703 url
, extra_headers
=headers
, ping_interval
=None,
5704 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5707 atexit
.register(self
.__exit
__, None, None, None)
5709 def __enter__(self
):
5711 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5714 def send(self
, *args
):
5715 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5717 def recv(self
, *args
):
5718 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5720 def __exit__(self
, type, value
, traceback
):
5722 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5725 self
._cancel
_all
_tasks
(self
.loop
)
5727 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5728 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5730 def run_with_loop(main
, loop
):
5731 if not asyncio
.iscoroutine(main
):
5732 raise ValueError(f
'a coroutine was expected, got {main!r}')
5735 return loop
.run_until_complete(main
)
5737 loop
.run_until_complete(loop
.shutdown_asyncgens())
5738 if hasattr(loop
, 'shutdown_default_executor'):
5739 loop
.run_until_complete(loop
.shutdown_default_executor())
5742 def _cancel_all_tasks(loop
):
5743 to_cancel
= asyncio
.all_tasks(loop
)
5748 for task
in to_cancel
:
5751 # XXX: "loop" is removed in python 3.10+
5752 loop
.run_until_complete(
5753 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5755 for task
in to_cancel
:
5756 if task
.cancelled():
5758 if task
.exception() is not None:
5759 loop
.call_exception_handler({
5760 'message': 'unhandled exception during asyncio.run() shutdown',
5761 'exception': task
.exception(),
5766 def merge_headers(*dicts
):
5767 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5768 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5771 def cached_method(f
):
5772 """Cache a method"""
5773 signature
= inspect
.signature(f
)
5776 def wrapper(self
, *args
, **kwargs
):
5777 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5778 bound_args
.apply_defaults()
5779 key
= tuple(bound_args
.arguments
.values())
5781 if not hasattr(self
, '__cached_method__cache'):
5782 self
.__cached
_method
__cache
= {}
5783 cache
= self
.__cached
_method
__cache
.setdefault(f
.__name
__, {})
5784 if key
not in cache
:
5785 cache
[key
] = f(self
, *args
, **kwargs
)
5790 class classproperty
:
5791 """property access for class methods"""
5793 def __init__(self
, func
):
5794 functools
.update_wrapper(self
, func
)
5797 def __get__(self
, _
, cls
):
5798 return self
.func(cls
)
5801 class Namespace(types
.SimpleNamespace
):
5802 """Immutable namespace"""
5805 return iter(self
.__dict
__.values())
5809 return self
.__dict
__.items()
5812 MEDIA_EXTENSIONS
= Namespace(
5813 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5814 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5815 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5816 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5817 thumbnails
=('jpg', 'png', 'webp'),
5818 storyboards
=('mhtml', ),
5819 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5820 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5822 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5823 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5825 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5830 for retry in RetryManager(...):
5833 except SomeException as err:
5837 attempt
, _error
= 0, None
5839 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5840 self
.retries
= _retries
or 0
5841 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5843 def _should_retry(self
):
5844 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5848 if self
._error
is NO_DEFAULT
:
5853 def error(self
, value
):
5857 while self
._should
_retry
():
5858 self
.error
= NO_DEFAULT
5862 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5865 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5866 """Utility function for reporting retries"""
5869 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5874 elif isinstance(e
, ExtractorError
):
5875 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5876 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5878 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5880 info(f
'Sleeping {delay:.2f} seconds ...')
5884 def make_archive_id(ie
, video_id
):
5885 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5886 return f
'{ie_key.lower()} {video_id}'
5889 def truncate_string(s
, left
, right
=0):
5890 assert left
> 3 and right
>= 0
5891 if s
is None or len(s
) <= left
+ right
:
5893 return f
'{s[:left-3]}...{s[-right:]}'
5896 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5897 assert 'all' in alias_dict
, '"all" alias is required'
5898 requested
= list(start
or [])
5900 discard
= val
.startswith('-')
5904 if val
in alias_dict
:
5905 val
= alias_dict
[val
] if not discard
else [
5906 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5907 # NB: Do not allow regex in aliases for performance
5908 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5911 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5912 else [val
] if val
in alias_dict
['all'] else None)
5914 raise ValueError(val
)
5917 for item
in current
:
5918 while item
in requested
:
5919 requested
.remove(item
)
5921 requested
.extend(current
)
5923 return orderedSet(requested
)
5927 has_certifi
= bool(certifi
)
5928 has_websockets
= bool(websockets
)