48 import xml
.etree
.ElementTree
51 from .compat
import functools
# isort: split
53 compat_etree_fromstring
,
55 compat_HTMLParseError
,
59 from .dependencies
import brotli
, certifi
, websockets
, xattr
60 from .socks
import ProxyType
, sockssocket
63 def register_socks_protocols():
64 # "Register" SOCKS protocols
65 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
66 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
67 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
68 if scheme
not in urllib
.parse
.uses_netloc
:
69 urllib
.parse
.uses_netloc
.append(scheme
)
72 # This is not clearly defined otherwise
73 compiled_regex_type
= type(re
.compile(''))
76 def random_user_agent():
77 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
118 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
121 SUPPORTED_ENCODINGS
= [
125 SUPPORTED_ENCODINGS
.append('br')
128 'User-Agent': random_user_agent(),
129 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
130 'Accept-Language': 'en-us,en;q=0.5',
131 'Sec-Fetch-Mode': 'navigate',
136 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
140 NO_DEFAULT
= object()
141 IDENTITY
= lambda x
: x
143 ENGLISH_MONTH_NAMES
= [
144 'January', 'February', 'March', 'April', 'May', 'June',
145 'July', 'August', 'September', 'October', 'November', 'December']
148 'en': ENGLISH_MONTH_NAMES
,
150 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
151 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
152 # these follow the genitive grammatical case (dopełniacz)
153 # some websites might be using nominative, which will require another month list
154 # https://en.wikibooks.org/wiki/Polish/Noun_cases
155 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
156 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
159 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
161 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
162 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
163 'EST': -5, 'EDT': -4, # Eastern
164 'CST': -6, 'CDT': -5, # Central
165 'MST': -7, 'MDT': -6, # Mountain
166 'PST': -8, 'PDT': -7 # Pacific
169 # needed for sanitizing filenames in restricted mode
170 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
171 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
172 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
202 '%Y-%m-%d %H:%M:%S.%f',
203 '%Y-%m-%d %H:%M:%S:%f',
206 '%Y-%m-%dT%H:%M:%SZ',
207 '%Y-%m-%dT%H:%M:%S.%fZ',
208 '%Y-%m-%dT%H:%M:%S.%f0Z',
210 '%Y-%m-%dT%H:%M:%S.%f',
213 '%b %d %Y at %H:%M:%S',
215 '%B %d %Y at %H:%M:%S',
219 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
220 DATE_FORMATS_DAY_FIRST
.extend([
230 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
231 DATE_FORMATS_MONTH_FIRST
.extend([
239 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
240 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
242 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
246 def preferredencoding():
247 """Get preferred encoding.
249 Returns the best encoding scheme for the system, based on
250 locale.getpreferredencoding() and some further tweaks.
253 pref = locale.getpreferredencoding()
261 def write_json_file(obj, fn):
262 """ Encode obj as JSON and write it to fn, atomically if possible """
264 tf = tempfile.NamedTemporaryFile(
265 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
266 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
270 json.dump(obj, tf, ensure_ascii=False)
271 if sys.platform == 'win32
':
272 # Need to remove existing file on Windows, else os.rename raises
273 # WindowsError or FileExistsError.
274 with contextlib.suppress(OSError):
276 with contextlib.suppress(OSError):
279 os.chmod(tf.name, 0o666 & ~mask)
280 os.rename(tf.name, fn)
282 with contextlib.suppress(OSError):
287 def find_xpath_attr(node, xpath, key, val=None):
288 """ Find the xpath xpath[@key=val] """
289 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
290 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
291 return node.find(expr)
293 # On python2.6 the xml.etree.ElementTree.Element methods don't support
294 # the namespace parameter
297 def xpath_with_ns(path
, ns_map
):
298 components
= [c
.split(':') for c
in path
.split('/')]
302 replaced
.append(c
[0])
305 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
306 return '/'.join(replaced
)
309 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
310 def _find_xpath(xpath
):
311 return node
.find(xpath
)
313 if isinstance(xpath
, str):
314 n
= _find_xpath(xpath
)
322 if default
is not NO_DEFAULT
:
325 name
= xpath
if name
is None else name
326 raise ExtractorError('Could not find XML element %s' % name
)
332 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
333 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
334 if n
is None or n
== default
:
337 if default
is not NO_DEFAULT
:
340 name
= xpath
if name
is None else name
341 raise ExtractorError('Could not find XML element\'s text %s' % name
)
347 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
348 n
= find_xpath_attr(node
, xpath
, key
)
350 if default
is not NO_DEFAULT
:
353 name
= f
'{xpath}[@{key}]' if name
is None else name
354 raise ExtractorError('Could not find XML attribute %s' % name
)
360 def get_element_by_id(id, html
, **kwargs
):
361 """Return the content of the tag with the specified ID in the passed HTML document"""
362 return get_element_by_attribute('id', id, html
, **kwargs
)
365 def get_element_html_by_id(id, html
, **kwargs
):
366 """Return the html of the tag with the specified ID in the passed HTML document"""
367 return get_element_html_by_attribute('id', id, html
, **kwargs
)
370 def get_element_by_class(class_name
, html
):
371 """Return the content of the first tag with the specified class in the passed HTML document"""
372 retval
= get_elements_by_class(class_name
, html
)
373 return retval
[0] if retval
else None
376 def get_element_html_by_class(class_name
, html
):
377 """Return the html of the first tag with the specified class in the passed HTML document"""
378 retval
= get_elements_html_by_class(class_name
, html
)
379 return retval
[0] if retval
else None
382 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
383 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
384 return retval
[0] if retval
else None
387 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
388 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
389 return retval
[0] if retval
else None
392 def get_elements_by_class(class_name
, html
, **kargs
):
393 """Return the content of all tags with the specified class in the passed HTML document as a list"""
394 return get_elements_by_attribute(
395 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
396 html, escape_value=False)
399 def get_elements_html_by_class(class_name, html):
400 """Return the html of all tags with the specified class in the passed HTML document as a list"""
401 return get_elements_html_by_attribute(
402 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
403 html, escape_value=False)
406 def get_elements_by_attribute(*args, **kwargs):
407 """Return the content of the tag with the specified attribute in the passed HTML document"""
408 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
411 def get_elements_html_by_attribute(*args, **kwargs):
412 """Return the html of the tag with the specified attribute in the passed HTML document"""
413 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
416 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
418 Return the text (content) and the html (whole) of the tag with the specified
419 attribute in the passed HTML document
424 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
426 value = re.escape(value) if escape_value else value
428 partial_element_re = rf'''(?x
)
430 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
431 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
434 for m in re.finditer(partial_element_re, html):
435 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
438 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
443 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
445 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
446 closing tag for the first opening tag it has encountered, and can be used
450 class HTMLBreakOnClosingTagException(Exception):
454 self.tagstack = collections.deque()
455 html.parser.HTMLParser.__init__(self)
460 def __exit__(self, *_):
464 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
465 # so data remains buffered; we no longer have any interest in it, thus
466 # override this method to discard it
469 def handle_starttag(self, tag, _):
470 self.tagstack.append(tag)
472 def handle_endtag(self, tag):
473 if not self.tagstack:
474 raise compat_HTMLParseError('no tags
in the stack
')
476 inner_tag = self.tagstack.pop()
480 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
481 if not self.tagstack:
482 raise self.HTMLBreakOnClosingTagException()
485 # XXX: This should be far less strict
486 def get_element_text_and_html_by_tag(tag, html):
488 For the first element with the specified tag in the passed HTML document
489 return its' content (text
) and the whole
element (html
)
491 def find_or_raise(haystack, needle, exc):
493 return haystack.index(needle)
496 closing_tag = f'</{tag}>'
497 whole_start = find_or_raise(
498 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
499 content_start = find_or_raise(
500 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
501 content_start += whole_start + 1
502 with HTMLBreakOnClosingTagParser() as parser:
503 parser.feed(html[whole_start:content_start])
504 if not parser.tagstack or parser.tagstack[0] != tag:
505 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
506 offset = content_start
507 while offset < len(html):
508 next_closing_tag_start = find_or_raise(
509 html[offset:], closing_tag,
510 compat_HTMLParseError(f'closing {tag} tag not found'))
511 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
513 parser.feed(html[offset:offset + next_closing_tag_end])
514 offset += next_closing_tag_end
515 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
516 return html[content_start:offset + next_closing_tag_start], \
517 html[whole_start:offset + next_closing_tag_end]
518 raise compat_HTMLParseError('unexpected end of html')
521 class HTMLAttributeParser(html.parser.HTMLParser):
522 """Trivial HTML parser to gather the attributes
for a single element
"""
526 html.parser.HTMLParser.__init__(self)
528 def handle_starttag(self, tag, attrs):
529 self.attrs = dict(attrs)
530 raise compat_HTMLParseError('done')
533 class HTMLListAttrsParser(html.parser.HTMLParser):
534 """HTML parser to gather the attributes
for the elements of a
list"""
537 html.parser.HTMLParser.__init__(self)
541 def handle_starttag(self, tag, attrs):
542 if tag == 'li' and self._level == 0:
543 self.items.append(dict(attrs))
546 def handle_endtag(self, tag):
550 def extract_attributes(html_element):
551 """Given a string
for an HTML element such
as
553 a
="foo" B
="bar" c
="&98;az" d
=boz
554 empty
= noval entity
="&"
557 Decode
and return a dictionary of attributes
.
559 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
560 'empty': '', 'noval': None, 'entity': '&',
561 'sq': '"', 'dq': '\''
564 parser = HTMLAttributeParser()
565 with contextlib.suppress(compat_HTMLParseError):
566 parser.feed(html_element)
571 def parse_list(webpage):
572 """Given a string
for an series of HTML
<li
> elements
,
573 return a dictionary of their attributes
"""
574 parser = HTMLListAttrsParser()
580 def clean_html(html):
581 """Clean an HTML snippet into a readable string
"""
583 if html is None: # Convenience for sanitizing descriptions etc.
586 html = re.sub(r'\s+', ' ', html)
587 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
588 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
590 html = re.sub('<.*?>', '', html)
591 # Replace html entities
592 html = unescapeHTML(html)
596 class LenientJSONDecoder(json.JSONDecoder):
597 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
598 self.transform_source, self.ignore_extra = transform_source, ignore_extra
599 super().__init__(*args, **kwargs)
602 if self.transform_source:
603 s = self.transform_source(s)
605 if self.ignore_extra:
606 return self.raw_decode(s.lstrip())[0]
607 return super().decode(s)
608 except json.JSONDecodeError as e:
609 if e.pos is not None:
610 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
614 def sanitize_open(filename, open_mode):
615 """Try to
open the given filename
, and slightly tweak it
if this fails
.
617 Attempts to
open the given filename
. If this fails
, it tries to change
618 the filename slightly
, step by step
, until it
's either able to open it
619 or it fails and raises a final exception, like the standard open()
622 It returns the tuple (stream, definitive_file_name).
625 if sys.platform == 'win32
':
628 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
629 with contextlib.suppress(io.UnsupportedOperation):
630 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
631 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
633 for attempt in range(2):
636 if sys.platform == 'win32
':
637 # FIXME: An exclusive lock also locks the file from being read.
638 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
639 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
640 raise LockingUnsupportedError()
641 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
643 stream
= open(filename
, open_mode
)
644 return stream
, filename
645 except OSError as err
:
646 if attempt
or err
.errno
in (errno
.EACCES
,):
648 old_filename
, filename
= filename
, sanitize_path(filename
)
649 if old_filename
== filename
:
653 def timeconvert(timestr
):
654 """Convert RFC 2822 defined time string into system timestamp"""
656 timetuple
= email
.utils
.parsedate_tz(timestr
)
657 if timetuple
is not None:
658 timestamp
= email
.utils
.mktime_tz(timetuple
)
662 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
663 """Sanitizes a string so it could be used as part of a filename.
664 @param restricted Use a stricter subset of allowed characters
665 @param is_id Whether this is an ID that should be kept unchanged if possible.
666 If unset, yt-dlp's new sanitization rules are in effect
671 def replace_insane(char
):
672 if restricted
and char
in ACCENT_CHARS
:
673 return ACCENT_CHARS
[char
]
674 elif not restricted
and char
== '\n':
676 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
677 # Replace with their full-width unicode counterparts
678 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
679 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
682 return '' if restricted
else '\''
684 return '\0_\0-' if restricted
else '\0 \0-'
685 elif char
in '\\/|*<>':
687 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
691 # Replace look-alike Unicode glyphs
692 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
693 s
= unicodedata
.normalize('NFKC', s
)
694 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
695 result
= ''.join(map(replace_insane
, s
))
696 if is_id
is NO_DEFAULT
:
697 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
698 STRIP_RE
= r
'(?:\0.|[ _-])*'
699 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
700 result
= result
.replace('\0', '') or '_'
703 while '__' in result
:
704 result
= result
.replace('__', '_')
705 result
= result
.strip('_')
706 # Common case of "Foreign band name - English song title"
707 if restricted
and result
.startswith('-_'):
709 if result
.startswith('-'):
710 result
= '_' + result
[len('-'):]
711 result
= result
.lstrip('.')
717 def sanitize_path(s
, force
=False):
718 """Sanitizes and normalizes path on Windows"""
719 if sys
.platform
== 'win32':
721 drive_or_unc
, _
= os
.path
.splitdrive(s
)
727 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
731 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
732 for path_part
in norm_path
]
734 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
735 elif force
and s
and s
[0] == os
.path
.sep
:
736 sanitized_path
.insert(0, os
.path
.sep
)
737 return os
.path
.join(*sanitized_path
)
740 def sanitize_url(url
, *, scheme
='http'):
741 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
742 # the number of unwanted failures due to missing protocol
745 elif url
.startswith('//'):
746 return f
'{scheme}:{url}'
747 # Fix some common typos seen so far
749 # https://github.com/ytdl-org/youtube-dl/issues/15649
750 (r
'^httpss://', r
'https://'),
751 # https://bx1.be/lives/direct-tv/
752 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
754 for mistake
, fixup
in COMMON_TYPOS
:
755 if re
.match(mistake
, url
):
756 return re
.sub(mistake
, fixup
, url
)
760 def extract_basic_auth(url
):
761 parts
= urllib
.parse
.urlsplit(url
)
762 if parts
.username
is None:
764 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
765 parts
.hostname
if parts
.port
is None
766 else '%s:%d' % (parts
.hostname
, parts
.port
))))
767 auth_payload
= base64
.b64encode(
768 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
769 return url
, f
'Basic {auth_payload.decode()}'
772 def sanitized_Request(url
, *args
, **kwargs
):
773 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
774 if auth_header
is not None:
775 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
776 headers
['Authorization'] = auth_header
777 return urllib
.request
.Request(url
, *args
, **kwargs
)
781 """Expand shell variables and ~"""
782 return os
.path
.expandvars(compat_expanduser(s
))
785 def orderedSet(iterable
, *, lazy
=False):
786 """Remove all duplicates from the input iterable"""
788 seen
= [] # Do not use set since the items can be unhashable
794 return _iter() if lazy
else list(_iter())
797 def _htmlentity_transform(entity_with_semicolon
):
798 """Transforms an HTML entity to a character."""
799 entity
= entity_with_semicolon
[:-1]
801 # Known non-numeric HTML entity
802 if entity
in html
.entities
.name2codepoint
:
803 return chr(html
.entities
.name2codepoint
[entity
])
805 # TODO: HTML5 allows entities without a semicolon.
806 # E.g. 'Éric' should be decoded as 'Éric'.
807 if entity_with_semicolon
in html
.entities
.html5
:
808 return html
.entities
.html5
[entity_with_semicolon
]
810 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
812 numstr
= mobj
.group(1)
813 if numstr
.startswith('x'):
815 numstr
= '0%s' % numstr
818 # See https://github.com/ytdl-org/youtube-dl/issues/7518
819 with contextlib
.suppress(ValueError):
820 return chr(int(numstr
, base
))
822 # Unknown entity in name, return its literal representation
823 return '&%s;' % entity
829 assert isinstance(s
, str)
832 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
835 def escapeHTML(text
):
838 .replace('&', '&')
839 .replace('<', '<')
840 .replace('>', '>')
841 .replace('"', '"')
842 .replace("'", ''')
846 def process_communicate_or_kill(p
, *args
, **kwargs
):
847 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
848 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
849 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
852 class Popen(subprocess
.Popen
):
853 if sys
.platform
== 'win32':
854 _startupinfo
= subprocess
.STARTUPINFO()
855 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
860 def _fix_pyinstaller_ld_path(env
):
861 """Restore LD_LIBRARY_PATH when using PyInstaller
862 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
863 https://github.com/yt-dlp/yt-dlp/issues/4573
865 if not hasattr(sys
, '_MEIPASS'):
869 orig
= env
.get(f
'{key}_ORIG')
875 _fix('LD_LIBRARY_PATH') # Linux
876 _fix('DYLD_LIBRARY_PATH') # macOS
878 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
880 env
= os
.environ
.copy()
881 self
._fix
_pyinstaller
_ld
_path
(env
)
884 kwargs
['universal_newlines'] = True # For 3.6 compatibility
885 kwargs
.setdefault('encoding', 'utf-8')
886 kwargs
.setdefault('errors', 'replace')
887 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
889 def communicate_or_kill(self
, *args
, **kwargs
):
891 return self
.communicate(*args
, **kwargs
)
892 except BaseException
: # Including KeyboardInterrupt
893 self
.kill(timeout
=None)
896 def kill(self
, *, timeout
=0):
899 self
.wait(timeout
=timeout
)
902 def run(cls
, *args
, timeout
=None, **kwargs
):
903 with cls(*args
, **kwargs
) as proc
:
904 default
= '' if proc
.text_mode
else b
''
905 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
906 return stdout
or default
, stderr
or default
, proc
.returncode
909 def get_subprocess_encoding():
910 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
911 # For subprocess calls, encode with locale encoding
912 # Refer to http://stackoverflow.com/a/9951851/35070
913 encoding
= preferredencoding()
915 encoding
= sys
.getfilesystemencoding()
921 def encodeFilename(s
, for_subprocess
=False):
922 assert isinstance(s
, str)
926 def decodeFilename(b
, for_subprocess
=False):
930 def encodeArgument(s
):
931 # Legacy code that uses byte strings
932 # Uncomment the following line after fixing all post processors
933 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
934 return s
if isinstance(s
, str) else s
.decode('ascii')
937 def decodeArgument(b
):
941 def decodeOption(optval
):
944 if isinstance(optval
, bytes):
945 optval
= optval
.decode(preferredencoding())
947 assert isinstance(optval
, str)
951 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
954 def timetuple_from_msec(msec
):
955 secs
, msec
= divmod(msec
, 1000)
956 mins
, secs
= divmod(secs
, 60)
957 hrs
, mins
= divmod(mins
, 60)
958 return _timetuple(hrs
, mins
, secs
, msec
)
961 def formatSeconds(secs
, delim
=':', msec
=False):
962 time
= timetuple_from_msec(secs
* 1000)
964 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
966 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
968 ret
= '%d' % time
.seconds
969 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
972 def _ssl_load_windows_store_certs(ssl_context
, storename
):
973 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
975 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
976 if encoding
== 'x509_asn' and (
977 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
978 except PermissionError
:
981 with contextlib
.suppress(ssl
.SSLError
):
982 ssl_context
.load_verify_locations(cadata
=cert
)
985 def make_HTTPS_handler(params
, **kwargs
):
986 opts_check_certificate
= not params
.get('nocheckcertificate')
987 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
988 context
.check_hostname
= opts_check_certificate
989 if params
.get('legacyserverconnect'):
990 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
991 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
992 context
.set_ciphers('DEFAULT')
994 sys
.version_info
< (3, 10)
995 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
996 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
998 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
999 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
1000 # in some situations [2][3].
1001 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
1002 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1003 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1004 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1005 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1006 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1007 # 4. https://peps.python.org/pep-0644/
1008 # 5. https://peps.python.org/pep-0644/#libressl-support
1009 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1010 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1011 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1013 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1014 if opts_check_certificate
:
1015 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1016 context
.load_verify_locations(cafile
=certifi
.where())
1019 context
.load_default_certs()
1020 # Work around the issue in load_default_certs when there are bad certificates. See:
1021 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1022 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1023 except ssl
.SSLError
:
1024 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1025 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1026 for storename
in ('CA', 'ROOT'):
1027 _ssl_load_windows_store_certs(context
, storename
)
1028 context
.set_default_verify_paths()
1030 client_certfile
= params
.get('client_certificate')
1033 context
.load_cert_chain(
1034 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1035 password
=params
.get('client_certificate_password'))
1036 except ssl
.SSLError
:
1037 raise YoutubeDLError('Unable to load client certificate')
1039 # Some servers may reject requests if ALPN extension is not sent. See:
1040 # https://github.com/python/cpython/issues/85140
1041 # https://github.com/yt-dlp/yt-dlp/issues/3878
1042 with contextlib
.suppress(NotImplementedError):
1043 context
.set_alpn_protocols(['http/1.1'])
1045 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1048 def bug_reports_message(before
=';'):
1049 from .update
import REPOSITORY
1051 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1052 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1054 before
= before
.rstrip()
1055 if not before
or before
.endswith(('.', '!', '?')):
1056 msg
= msg
[0].title() + msg
[1:]
1058 return (before
+ ' ' if before
else '') + msg
1061 class YoutubeDLError(Exception):
1062 """Base exception for YoutubeDL errors."""
1065 def __init__(self
, msg
=None):
1068 elif self
.msg
is None:
1069 self
.msg
= type(self
).__name
__
1070 super().__init
__(self
.msg
)
1073 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1074 if hasattr(ssl
, 'CertificateError'):
1075 network_exceptions
.append(ssl
.CertificateError
)
1076 network_exceptions
= tuple(network_exceptions
)
1079 class ExtractorError(YoutubeDLError
):
1080 """Error during info extraction."""
1082 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1083 """ tb, if given, is the original traceback (so that it can be printed out).
1084 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1086 if sys
.exc_info()[0] in network_exceptions
:
1089 self
.orig_msg
= str(msg
)
1091 self
.expected
= expected
1093 self
.video_id
= video_id
1095 self
.exc_info
= sys
.exc_info() # preserve original exception
1096 if isinstance(self
.exc_info
[1], ExtractorError
):
1097 self
.exc_info
= self
.exc_info
[1].exc_info
1099 super().__init
__(''.join((
1100 format_field(ie
, None, '[%s] '),
1101 format_field(video_id
, None, '%s: '),
1103 format_field(cause
, None, ' (caused by %r)'),
1104 '' if expected
else bug_reports_message())))
1106 def format_traceback(self
):
1107 return join_nonempty(
1108 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1109 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1113 class UnsupportedError(ExtractorError
):
1114 def __init__(self
, url
):
1116 'Unsupported URL: %s' % url
, expected
=True)
1120 class RegexNotFoundError(ExtractorError
):
1121 """Error when a regex didn't match"""
1125 class GeoRestrictedError(ExtractorError
):
1126 """Geographic restriction Error exception.
1128 This exception may be thrown when a video is not available from your
1129 geographic location due to geographic restrictions imposed by a website.
1132 def __init__(self
, msg
, countries
=None, **kwargs
):
1133 kwargs
['expected'] = True
1134 super().__init
__(msg
, **kwargs
)
1135 self
.countries
= countries
1138 class UserNotLive(ExtractorError
):
1139 """Error when a channel/user is not live"""
1141 def __init__(self
, msg
=None, **kwargs
):
1142 kwargs
['expected'] = True
1143 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1146 class DownloadError(YoutubeDLError
):
1147 """Download Error exception.
1149 This exception may be thrown by FileDownloader objects if they are not
1150 configured to continue on errors. They will contain the appropriate
1154 def __init__(self
, msg
, exc_info
=None):
1155 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1156 super().__init
__(msg
)
1157 self
.exc_info
= exc_info
1160 class EntryNotInPlaylist(YoutubeDLError
):
1161 """Entry not in playlist exception.
1163 This exception will be thrown by YoutubeDL when a requested entry
1164 is not found in the playlist info_dict
1166 msg
= 'Entry not found in info'
1169 class SameFileError(YoutubeDLError
):
1170 """Same File exception.
1172 This exception will be thrown by FileDownloader objects if they detect
1173 multiple files would have to be downloaded to the same file on disk.
1175 msg
= 'Fixed output name but more than one file to download'
1177 def __init__(self
, filename
=None):
1178 if filename
is not None:
1179 self
.msg
+= f
': {filename}'
1180 super().__init
__(self
.msg
)
1183 class PostProcessingError(YoutubeDLError
):
1184 """Post Processing exception.
1186 This exception may be raised by PostProcessor's .run() method to
1187 indicate an error in the postprocessing task.
1191 class DownloadCancelled(YoutubeDLError
):
1192 """ Exception raised when the download queue should be interrupted """
1193 msg
= 'The download was cancelled'
1196 class ExistingVideoReached(DownloadCancelled
):
1197 """ --break-on-existing triggered """
1198 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1201 class RejectedVideoReached(DownloadCancelled
):
1202 """ --break-on-reject triggered """
1203 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1206 class MaxDownloadsReached(DownloadCancelled
):
1207 """ --max-downloads limit has been reached. """
1208 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1211 class ReExtractInfo(YoutubeDLError
):
1212 """ Video info needs to be re-extracted. """
1214 def __init__(self
, msg
, expected
=False):
1215 super().__init
__(msg
)
1216 self
.expected
= expected
1219 class ThrottledDownload(ReExtractInfo
):
1220 """ Download speed below --throttled-rate. """
1221 msg
= 'The download speed is below throttle limit'
1224 super().__init
__(self
.msg
, expected
=False)
1227 class UnavailableVideoError(YoutubeDLError
):
1228 """Unavailable Format exception.
1230 This exception will be thrown when a video is requested
1231 in a format that is not available for that video.
1233 msg
= 'Unable to download video'
1235 def __init__(self
, err
=None):
1237 self
.msg
+= f
': {err}'
1238 super().__init
__(self
.msg
)
1241 class ContentTooShortError(YoutubeDLError
):
1242 """Content Too Short exception.
1244 This exception may be raised by FileDownloader objects when a file they
1245 download is too small for what the server announced first, indicating
1246 the connection was probably interrupted.
1249 def __init__(self
, downloaded
, expected
):
1250 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1252 self
.downloaded
= downloaded
1253 self
.expected
= expected
1256 class XAttrMetadataError(YoutubeDLError
):
1257 def __init__(self
, code
=None, msg
='Unknown error'):
1258 super().__init
__(msg
)
1262 # Parsing code and msg
1263 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1264 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1265 self
.reason
= 'NO_SPACE'
1266 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1267 self
.reason
= 'VALUE_TOO_LONG'
1269 self
.reason
= 'NOT_SUPPORTED'
1272 class XAttrUnavailableError(YoutubeDLError
):
1276 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1277 hc
= http_class(*args
, **kwargs
)
1278 source_address
= ydl_handler
._params
.get('source_address')
1280 if source_address
is not None:
1281 # This is to workaround _create_connection() from socket where it will try all
1282 # address data from getaddrinfo() including IPv6. This filters the result from
1283 # getaddrinfo() based on the source_address value.
1284 # This is based on the cpython socket.create_connection() function.
1285 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1286 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1287 host
, port
= address
1289 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1290 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1291 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1292 if addrs
and not ip_addrs
:
1293 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1295 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1296 % (ip_version
, source_address
[0]))
1297 for res
in ip_addrs
:
1298 af
, socktype
, proto
, canonname
, sa
= res
1301 sock
= socket
.socket(af
, socktype
, proto
)
1302 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1303 sock
.settimeout(timeout
)
1304 sock
.bind(source_address
)
1306 err
= None # Explicitly break reference cycle
1308 except OSError as _
:
1310 if sock
is not None:
1315 raise OSError('getaddrinfo returns an empty list')
1316 if hasattr(hc
, '_create_connection'):
1317 hc
._create
_connection
= _create_connection
1318 hc
.source_address
= (source_address
, 0)
1323 def handle_youtubedl_headers(headers
):
1324 filtered_headers
= headers
1326 if 'Youtubedl-no-compression' in filtered_headers
:
1327 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1328 del filtered_headers
['Youtubedl-no-compression']
1330 return filtered_headers
1333 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1334 """Handler for HTTP requests and responses.
1336 This class, when installed with an OpenerDirector, automatically adds
1337 the standard headers to every HTTP request and handles gzipped and
1338 deflated responses from web servers. If compression is to be avoided in
1339 a particular request, the original request in the program code only has
1340 to include the HTTP header "Youtubedl-no-compression", which will be
1341 removed before making the real request.
1343 Part of this code was copied from:
1345 http://techknack.net/python-urllib2-handlers/
1347 Andrew Rowls, the author of that code, agreed to release it to the
1351 def __init__(self
, params
, *args
, **kwargs
):
1352 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1353 self
._params
= params
1355 def http_open(self
, req
):
1356 conn_class
= http
.client
.HTTPConnection
1358 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1360 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1361 del req
.headers
['Ytdl-socks-proxy']
1363 return self
.do_open(functools
.partial(
1364 _create_http_connection
, self
, conn_class
, False),
1372 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1374 return zlib
.decompress(data
)
1380 return brotli
.decompress(data
)
1382 def http_request(self
, req
):
1383 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1384 # always respected by websites, some tend to give out URLs with non percent-encoded
1385 # non-ASCII characters (see telemb.py, ard.py [#3412])
1386 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1387 # To work around aforementioned issue we will replace request's original URL with
1388 # percent-encoded one
1389 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1390 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1391 url
= req
.get_full_url()
1392 url_escaped
= escape_url(url
)
1394 # Substitute URL if any change after escaping
1395 if url
!= url_escaped
:
1396 req
= update_Request(req
, url
=url_escaped
)
1398 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1399 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1400 # The dict keys are capitalized because of this bug by urllib
1401 if h
.capitalize() not in req
.headers
:
1402 req
.add_header(h
, v
)
1404 if 'Accept-encoding' not in req
.headers
:
1405 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1407 req
.headers
= handle_youtubedl_headers(req
.headers
)
1409 return super().do_request_(req
)
1411 def http_response(self
, req
, resp
):
1414 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1415 content
= resp
.read()
1416 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1418 uncompressed
= io
.BytesIO(gz
.read())
1419 except OSError as original_ioerror
:
1420 # There may be junk add the end of the file
1421 # See http://stackoverflow.com/q/4928560/35070 for details
1422 for i
in range(1, 1024):
1424 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1425 uncompressed
= io
.BytesIO(gz
.read())
1430 raise original_ioerror
1431 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1432 resp
.msg
= old_resp
.msg
1433 del resp
.headers
['Content-encoding']
1435 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1436 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1437 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1438 resp
.msg
= old_resp
.msg
1439 del resp
.headers
['Content-encoding']
1441 if resp
.headers
.get('Content-encoding', '') == 'br':
1442 resp
= urllib
.request
.addinfourl(
1443 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1444 resp
.msg
= old_resp
.msg
1445 del resp
.headers
['Content-encoding']
1446 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1447 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1448 if 300 <= resp
.code
< 400:
1449 location
= resp
.headers
.get('Location')
1451 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1452 location
= location
.encode('iso-8859-1').decode()
1453 location_escaped
= escape_url(location
)
1454 if location
!= location_escaped
:
1455 del resp
.headers
['Location']
1456 resp
.headers
['Location'] = location_escaped
1459 https_request
= http_request
1460 https_response
= http_response
1463 def make_socks_conn_class(base_class
, socks_proxy
):
1464 assert issubclass(base_class
, (
1465 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1467 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1468 if url_components
.scheme
.lower() == 'socks5':
1469 socks_type
= ProxyType
.SOCKS5
1470 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1471 socks_type
= ProxyType
.SOCKS4
1472 elif url_components
.scheme
.lower() == 'socks4a':
1473 socks_type
= ProxyType
.SOCKS4A
1475 def unquote_if_non_empty(s
):
1478 return urllib
.parse
.unquote_plus(s
)
1482 url_components
.hostname
, url_components
.port
or 1080,
1484 unquote_if_non_empty(url_components
.username
),
1485 unquote_if_non_empty(url_components
.password
),
1488 class SocksConnection(base_class
):
1490 self
.sock
= sockssocket()
1491 self
.sock
.setproxy(*proxy_args
)
1492 if isinstance(self
.timeout
, (int, float)):
1493 self
.sock
.settimeout(self
.timeout
)
1494 self
.sock
.connect((self
.host
, self
.port
))
1496 if isinstance(self
, http
.client
.HTTPSConnection
):
1497 if hasattr(self
, '_context'): # Python > 2.6
1498 self
.sock
= self
._context
.wrap_socket(
1499 self
.sock
, server_hostname
=self
.host
)
1501 self
.sock
= ssl
.wrap_socket(self
.sock
)
1503 return SocksConnection
1506 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1507 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1508 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1509 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1510 self
._params
= params
1512 def https_open(self
, req
):
1514 conn_class
= self
._https
_conn
_class
1516 if hasattr(self
, '_context'): # python > 2.6
1517 kwargs
['context'] = self
._context
1518 if hasattr(self
, '_check_hostname'): # python 3.x
1519 kwargs
['check_hostname'] = self
._check
_hostname
1521 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1523 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1524 del req
.headers
['Ytdl-socks-proxy']
1527 return self
.do_open(
1528 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1529 except urllib
.error
.URLError
as e
:
1530 if (isinstance(e
.reason
, ssl
.SSLError
)
1531 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1532 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1536 def is_path_like(f
):
1537 return isinstance(f
, (str, bytes, os
.PathLike
))
1540 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1542 See [1] for cookie file format.
1544 1. https://curl.haxx.se/docs/http-cookies.html
1546 _HTTPONLY_PREFIX
= '#HttpOnly_'
1548 _HEADER
= '''# Netscape HTTP Cookie File
1549 # This file is generated by yt-dlp. Do not edit.
1552 _CookieFileEntry
= collections
.namedtuple(
1554 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1556 def __init__(self
, filename
=None, *args
, **kwargs
):
1557 super().__init
__(None, *args
, **kwargs
)
1558 if is_path_like(filename
):
1559 filename
= os
.fspath(filename
)
1560 self
.filename
= filename
1563 def _true_or_false(cndn
):
1564 return 'TRUE' if cndn
else 'FALSE'
1566 @contextlib.contextmanager
1567 def open(self
, file, *, write
=False):
1568 if is_path_like(file):
1569 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1576 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1579 if (not ignore_discard
and cookie
.discard
1580 or not ignore_expires
and cookie
.is_expired(now
)):
1582 name
, value
= cookie
.name
, cookie
.value
1584 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1585 # with no name, whereas http.cookiejar regards it as a
1586 # cookie with no value.
1587 name
, value
= '', name
1588 f
.write('%s\n' % '\t'.join((
1590 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1592 self
._true
_or
_false
(cookie
.secure
),
1593 str_or_none(cookie
.expires
, default
=''),
1597 def save(self
, filename
=None, *args
, **kwargs
):
1599 Save cookies to a file.
1600 Code is taken from CPython 3.6
1601 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1603 if filename
is None:
1604 if self
.filename
is not None:
1605 filename
= self
.filename
1607 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1609 # Store session cookies with `expires` set to 0 instead of an empty string
1611 if cookie
.expires
is None:
1614 with self
.open(filename
, write
=True) as f
:
1615 f
.write(self
._HEADER
)
1616 self
._really
_save
(f
, *args
, **kwargs
)
1618 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1619 """Load cookies from a file."""
1620 if filename
is None:
1621 if self
.filename
is not None:
1622 filename
= self
.filename
1624 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1626 def prepare_line(line
):
1627 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1628 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1629 # comments and empty lines are fine
1630 if line
.startswith('#') or not line
.strip():
1632 cookie_list
= line
.split('\t')
1633 if len(cookie_list
) != self
._ENTRY
_LEN
:
1634 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1635 cookie
= self
._CookieFileEntry
(*cookie_list
)
1636 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1637 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1641 with self
.open(filename
) as f
:
1644 cf
.write(prepare_line(line
))
1645 except http
.cookiejar
.LoadError
as e
:
1646 if f
'{line.strip()} '[0] in '[{"':
1647 raise http
.cookiejar
.LoadError(
1648 'Cookies file must be Netscape formatted, not JSON. See '
1649 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1650 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1653 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1654 # Session cookies are denoted by either `expires` field set to
1655 # an empty string or 0. MozillaCookieJar only recognizes the former
1656 # (see [1]). So we need force the latter to be recognized as session
1657 # cookies on our own.
1658 # Session cookies may be important for cookies-based authentication,
1659 # e.g. usually, when user does not check 'Remember me' check box while
1660 # logging in on a site, some important cookies are stored as session
1661 # cookies so that not recognizing them will result in failed login.
1662 # 1. https://bugs.python.org/issue17164
1664 # Treat `expires=0` cookies as session cookies
1665 if cookie
.expires
== 0:
1666 cookie
.expires
= None
1667 cookie
.discard
= True
1670 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1671 def __init__(self
, cookiejar
=None):
1672 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1674 def http_response(self
, request
, response
):
1675 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1677 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1678 https_response
= http_response
1681 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1682 """YoutubeDL redirect handler
1684 The code is based on HTTPRedirectHandler implementation from CPython [1].
1686 This redirect handler solves two issues:
1687 - ensures redirect URL is always unicode under python 2
1688 - introduces support for experimental HTTP response status code
1689 308 Permanent Redirect [2] used by some sites [3]
1691 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1692 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1693 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1696 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1698 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1699 """Return a Request or None in response to a redirect.
1701 This is called by the http_error_30x methods when a
1702 redirection response is received. If a redirection should
1703 take place, return a new Request to allow http_error_30x to
1704 perform the redirect. Otherwise, raise HTTPError if no-one
1705 else should try to handle this url. Return None if you can't
1706 but another Handler might.
1708 m
= req
.get_method()
1709 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1710 or code
in (301, 302, 303) and m
== "POST")):
1711 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1712 # Strictly (according to RFC 2616), 301 or 302 in response to
1713 # a POST MUST NOT cause a redirection without confirmation
1714 # from the user (of urllib.request, in this case). In practice,
1715 # essentially all clients do redirect in this case, so we do
1718 # Be conciliant with URIs containing a space. This is mainly
1719 # redundant with the more complete encoding done in http_error_302(),
1720 # but it is kept for compatibility with other callers.
1721 newurl
= newurl
.replace(' ', '%20')
1723 CONTENT_HEADERS
= ("content-length", "content-type")
1724 # NB: don't use dict comprehension for python 2.6 compatibility
1725 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1727 # A 303 must either use GET or HEAD for subsequent request
1728 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1729 if code
== 303 and m
!= 'HEAD':
1731 # 301 and 302 redirects are commonly turned into a GET from a POST
1732 # for subsequent requests by browsers, so we'll do the same.
1733 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1734 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1735 if code
in (301, 302) and m
== 'POST':
1738 return urllib
.request
.Request(
1739 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1740 unverifiable
=True, method
=m
)
1743 def extract_timezone(date_str
):
1746 ^.{8,}? # >=8 char non-TZ prefix, if present
1747 (?P<tz>Z| # just the UTC Z, or
1748 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1749 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1750 [ ]? # optional space
1751 (?P<sign>\+|-) # +/-
1752 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1756 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1757 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1758 if timezone
is not None:
1759 date_str
= date_str
[:-len(m
.group('tz'))]
1760 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1762 date_str
= date_str
[:-len(m
.group('tz'))]
1763 if not m
.group('sign'):
1764 timezone
= datetime
.timedelta()
1766 sign
= 1 if m
.group('sign') == '+' else -1
1767 timezone
= datetime
.timedelta(
1768 hours
=sign
* int(m
.group('hours')),
1769 minutes
=sign
* int(m
.group('minutes')))
1770 return timezone
, date_str
1773 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1774 """ Return a UNIX timestamp from the given date """
1776 if date_str
is None:
1779 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1781 if timezone
is None:
1782 timezone
, date_str
= extract_timezone(date_str
)
1784 with contextlib
.suppress(ValueError):
1785 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1786 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1787 return calendar
.timegm(dt
.timetuple())
1790 def date_formats(day_first
=True):
1791 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1794 def unified_strdate(date_str
, day_first
=True):
1795 """Return a string with the date in the format YYYYMMDD"""
1797 if date_str
is None:
1801 date_str
= date_str
.replace(',', ' ')
1802 # Remove AM/PM + timezone
1803 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1804 _
, date_str
= extract_timezone(date_str
)
1806 for expression
in date_formats(day_first
):
1807 with contextlib
.suppress(ValueError):
1808 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1809 if upload_date
is None:
1810 timetuple
= email
.utils
.parsedate_tz(date_str
)
1812 with contextlib
.suppress(ValueError):
1813 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1814 if upload_date
is not None:
1815 return str(upload_date
)
1818 def unified_timestamp(date_str
, day_first
=True):
1819 if date_str
is None:
1822 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1823 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1825 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1826 timezone
, date_str
= extract_timezone(date_str
)
1828 # Remove AM/PM + timezone
1829 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1831 # Remove unrecognized timezones from ISO 8601 alike timestamps
1832 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1834 date_str
= date_str
[:-len(m
.group('tz'))]
1836 # Python only supports microseconds, so remove nanoseconds
1837 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1839 date_str
= m
.group(1)
1841 for expression
in date_formats(day_first
):
1842 with contextlib
.suppress(ValueError):
1843 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1844 return calendar
.timegm(dt
.timetuple())
1846 timetuple
= email
.utils
.parsedate_tz(date_str
)
1848 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1851 def determine_ext(url
, default_ext
='unknown_video'):
1852 if url
is None or '.' not in url
:
1854 guess
= url
.partition('?')[0].rpartition('.')[2]
1855 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1857 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1858 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1859 return guess
.rstrip('/')
1864 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1865 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1868 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1870 Return a datetime object from a string.
1872 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1874 @param format strftime format of DATE
1875 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1876 auto: round to the unit provided in date_str (if applicable).
1878 auto_precision
= False
1879 if precision
== 'auto':
1880 auto_precision
= True
1881 precision
= 'microsecond'
1882 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1883 if date_str
in ('now', 'today'):
1885 if date_str
== 'yesterday':
1886 return today
- datetime
.timedelta(days
=1)
1888 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1890 if match
is not None:
1891 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1892 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1893 unit
= match
.group('unit')
1894 if unit
== 'month' or unit
== 'year':
1895 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1901 delta
= datetime
.timedelta(**{unit + 's': time}
)
1902 new_date
= start_time
+ delta
1904 return datetime_round(new_date
, unit
)
1907 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1910 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1912 Return a date object from a string using datetime_from_str
1914 @param strict Restrict allowed patterns to "YYYYMMDD" and
1915 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1917 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1918 raise ValueError(f
'Invalid date format "{date_str}"')
1919 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1922 def datetime_add_months(dt
, months
):
1923 """Increment/Decrement a datetime object by months."""
1924 month
= dt
.month
+ months
- 1
1925 year
= dt
.year
+ month
// 12
1926 month
= month
% 12 + 1
1927 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1928 return dt
.replace(year
, month
, day
)
1931 def datetime_round(dt
, precision
='day'):
1933 Round a datetime object's time to a specific precision
1935 if precision
== 'microsecond':
1944 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1945 timestamp
= calendar
.timegm(dt
.timetuple())
1946 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1949 def hyphenate_date(date_str
):
1951 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1952 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1953 if match
is not None:
1954 return '-'.join(match
.groups())
1960 """Represents a time interval between two dates"""
1962 def __init__(self
, start
=None, end
=None):
1963 """start and end must be strings in the format accepted by date"""
1964 if start
is not None:
1965 self
.start
= date_from_str(start
, strict
=True)
1967 self
.start
= datetime
.datetime
.min.date()
1969 self
.end
= date_from_str(end
, strict
=True)
1971 self
.end
= datetime
.datetime
.max.date()
1972 if self
.start
> self
.end
:
1973 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1977 """Returns a range that only contains the given day"""
1978 return cls(day
, day
)
1980 def __contains__(self
, date
):
1981 """Check if the date is in the range"""
1982 if not isinstance(date
, datetime
.date
):
1983 date
= date_from_str(date
)
1984 return self
.start
<= date
<= self
.end
1987 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1989 def __eq__(self
, other
):
1990 return (isinstance(other
, DateRange
)
1991 and self
.start
== other
.start
and self
.end
== other
.end
)
1994 def platform_name():
1995 """ Returns the platform name as a str """
1996 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
1997 return platform
.platform()
2001 def system_identifier():
2002 python_implementation
= platform
.python_implementation()
2003 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2004 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2006 with contextlib
.suppress(OSError): # We may not have access to the executable
2007 libc_ver
= platform
.libc_ver()
2009 return 'Python %s (%s %s %s) - %s (%s%s)' % (
2010 platform
.python_version(),
2011 python_implementation
,
2013 platform
.architecture()[0],
2014 platform
.platform(),
2015 ssl
.OPENSSL_VERSION
,
2016 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2021 def get_windows_version():
2022 ''' Get Windows version. returns () if it's not running on Windows '''
2023 if compat_os_name
== 'nt':
2024 return version_tuple(platform
.win32_ver()[1])
2029 def write_string(s
, out
=None, encoding
=None):
2030 assert isinstance(s
, str)
2031 out
= out
or sys
.stderr
2033 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2034 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2036 enc
, buffer = None, out
2037 if 'b' in getattr(out
, 'mode', ''):
2038 enc
= encoding
or preferredencoding()
2039 elif hasattr(out
, 'buffer'):
2041 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2043 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2047 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2048 from . import _IN_CLI
2050 if msg
in deprecation_warning
._cache
:
2052 deprecation_warning
._cache
.add(msg
)
2054 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2055 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2058 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2061 deprecation_warning
._cache
= set()
2064 def bytes_to_intlist(bs
):
2067 if isinstance(bs
[0], int): # Python 3
2070 return [ord(c
) for c
in bs
]
2073 def intlist_to_bytes(xs
):
2076 return struct
.pack('%dB' % len(xs
), *xs
)
2079 class LockingUnsupportedError(OSError):
2080 msg
= 'File locking is not supported'
2083 super().__init
__(self
.msg
)
2086 # Cross-platform file locking
2087 if sys
.platform
== 'win32':
2089 import ctypes
.wintypes
2092 class OVERLAPPED(ctypes
.Structure
):
2094 ('Internal', ctypes
.wintypes
.LPVOID
),
2095 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2096 ('Offset', ctypes
.wintypes
.DWORD
),
2097 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2098 ('hEvent', ctypes
.wintypes
.HANDLE
),
2101 kernel32
= ctypes
.windll
.kernel32
2102 LockFileEx
= kernel32
.LockFileEx
2103 LockFileEx
.argtypes
= [
2104 ctypes
.wintypes
.HANDLE
, # hFile
2105 ctypes
.wintypes
.DWORD
, # dwFlags
2106 ctypes
.wintypes
.DWORD
, # dwReserved
2107 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2108 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2109 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2111 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2112 UnlockFileEx
= kernel32
.UnlockFileEx
2113 UnlockFileEx
.argtypes
= [
2114 ctypes
.wintypes
.HANDLE
, # hFile
2115 ctypes
.wintypes
.DWORD
, # dwReserved
2116 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2117 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2118 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2120 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2121 whole_low
= 0xffffffff
2122 whole_high
= 0x7fffffff
2124 def _lock_file(f
, exclusive
, block
):
2125 overlapped
= OVERLAPPED()
2126 overlapped
.Offset
= 0
2127 overlapped
.OffsetHigh
= 0
2128 overlapped
.hEvent
= 0
2129 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2131 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2132 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2133 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2134 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2135 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2137 def _unlock_file(f
):
2138 assert f
._lock
_file
_overlapped
_p
2139 handle
= msvcrt
.get_osfhandle(f
.fileno())
2140 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2141 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2147 def _lock_file(f
, exclusive
, block
):
2148 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2150 flags |
= fcntl
.LOCK_NB
2152 fcntl
.flock(f
, flags
)
2153 except BlockingIOError
:
2155 except OSError: # AOSP does not have flock()
2156 fcntl
.lockf(f
, flags
)
2158 def _unlock_file(f
):
2160 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2162 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2166 def _lock_file(f
, exclusive
, block
):
2167 raise LockingUnsupportedError()
2169 def _unlock_file(f
):
2170 raise LockingUnsupportedError()
2176 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2177 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2178 raise NotImplementedError(mode
)
2179 self
.mode
, self
.block
= mode
, block
2181 writable
= any(f
in mode
for f
in 'wax+')
2182 readable
= any(f
in mode
for f
in 'r+')
2183 flags
= functools
.reduce(operator
.ior
, (
2184 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2185 getattr(os
, 'O_BINARY', 0), # Windows only
2186 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2187 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2188 os
.O_APPEND
if 'a' in mode
else 0,
2189 os
.O_EXCL
if 'x' in mode
else 0,
2190 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2193 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2195 def __enter__(self
):
2196 exclusive
= 'r' not in self
.mode
2198 _lock_file(self
.f
, exclusive
, self
.block
)
2203 if 'w' in self
.mode
:
2206 except OSError as e
:
2208 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2209 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2218 _unlock_file(self
.f
)
2222 def __exit__(self
, *_
):
2231 def __getattr__(self
, attr
):
2232 return getattr(self
.f
, attr
)
2239 def get_filesystem_encoding():
2240 encoding
= sys
.getfilesystemencoding()
2241 return encoding
if encoding
is not None else 'utf-8'
2244 def shell_quote(args
):
2246 encoding
= get_filesystem_encoding()
2248 if isinstance(a
, bytes):
2249 # We may get a filename encoded with 'encodeFilename'
2250 a
= a
.decode(encoding
)
2251 quoted_args
.append(compat_shlex_quote(a
))
2252 return ' '.join(quoted_args
)
2255 def smuggle_url(url
, data
):
2256 """ Pass additional data in a URL for internal use. """
2258 url
, idata
= unsmuggle_url(url
, {})
2260 sdata
= urllib
.parse
.urlencode(
2261 {'__youtubedl_smuggle': json.dumps(data)}
)
2262 return url
+ '#' + sdata
2265 def unsmuggle_url(smug_url
, default
=None):
2266 if '#__youtubedl_smuggle' not in smug_url
:
2267 return smug_url
, default
2268 url
, _
, sdata
= smug_url
.rpartition('#')
2269 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2270 data
= json
.loads(jsond
)
2274 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2275 """ Formats numbers with decimal sufixes like K, M, etc """
2276 num
, factor
= float_or_none(num
), float(factor
)
2277 if num
is None or num
< 0:
2279 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2280 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2281 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2283 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2284 converted
= num
/ (factor
** exponent
)
2285 return fmt
% (converted
, suffix
)
2288 def format_bytes(bytes):
2289 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2292 def lookup_unit_table(unit_table
, s
):
2293 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2295 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2298 num_str
= m
.group('num').replace(',', '.')
2299 mult
= unit_table
[m
.group('unit')]
2300 return int(float(num_str
) * mult
)
2303 def parse_filesize(s
):
2307 # The lower-case forms are of course incorrect and unofficial,
2308 # but we support those too
2325 'megabytes': 1000 ** 2,
2326 'mebibytes': 1024 ** 2,
2332 'gigabytes': 1000 ** 3,
2333 'gibibytes': 1024 ** 3,
2339 'terabytes': 1000 ** 4,
2340 'tebibytes': 1024 ** 4,
2346 'petabytes': 1000 ** 5,
2347 'pebibytes': 1024 ** 5,
2353 'exabytes': 1000 ** 6,
2354 'exbibytes': 1024 ** 6,
2360 'zettabytes': 1000 ** 7,
2361 'zebibytes': 1024 ** 7,
2367 'yottabytes': 1000 ** 8,
2368 'yobibytes': 1024 ** 8,
2371 return lookup_unit_table(_UNIT_TABLE
, s
)
2378 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2380 if re
.match(r
'^[\d,.]+$', s
):
2381 return str_to_int(s
)
2394 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2398 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2400 return str_to_int(mobj
.group(1))
2403 def parse_resolution(s
, *, lenient
=False):
2408 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2410 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2413 'width': int(mobj
.group('w')),
2414 'height': int(mobj
.group('h')),
2417 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2419 return {'height': int(mobj.group(1))}
2421 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2423 return {'height': int(mobj.group(1)) * 540}
2428 def parse_bitrate(s
):
2429 if not isinstance(s
, str):
2431 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2433 return int(mobj
.group(1))
2436 def month_by_name(name
, lang
='en'):
2437 """ Return the number of a month by (locale-independently) English name """
2439 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2442 return month_names
.index(name
) + 1
2447 def month_by_abbreviation(abbrev
):
2448 """ Return the number of a month by (locale-independently) English
2452 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2457 def fix_xml_ampersands(xml_str
):
2458 """Replace all the '&' by '&' in XML"""
2460 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2465 def setproctitle(title
):
2466 assert isinstance(title
, str)
2468 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2475 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2479 # LoadLibrary in Windows Python 2.7.13 only expects
2480 # a bytestring, but since unicode_literals turns
2481 # every string into a unicode string, it fails.
2483 title_bytes
= title
.encode()
2484 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2485 buf
.value
= title_bytes
2487 libc
.prctl(15, buf
, 0, 0, 0)
2488 except AttributeError:
2489 return # Strange libc, just skip this
2492 def remove_start(s
, start
):
2493 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2496 def remove_end(s
, end
):
2497 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2500 def remove_quotes(s
):
2501 if s
is None or len(s
) < 2:
2503 for quote
in ('"', "'", ):
2504 if s
[0] == quote
and s
[-1] == quote
:
2509 def get_domain(url
):
2511 This implementation is inconsistent, but is kept for compatibility.
2512 Use this only for "webpage_url_domain"
2514 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2517 def url_basename(url
):
2518 path
= urllib
.parse
.urlparse(url
).path
2519 return path
.strip('/').split('/')[-1]
2523 return re
.match(r
'https?://[^?#]+/', url
).group()
2526 def urljoin(base
, path
):
2527 if isinstance(path
, bytes):
2528 path
= path
.decode()
2529 if not isinstance(path
, str) or not path
:
2531 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2533 if isinstance(base
, bytes):
2534 base
= base
.decode()
2535 if not isinstance(base
, str) or not re
.match(
2536 r
'^(?:https?:)?//', base
):
2538 return urllib
.parse
.urljoin(base
, path
)
2541 class HEADRequest(urllib
.request
.Request
):
2542 def get_method(self
):
2546 class PUTRequest(urllib
.request
.Request
):
2547 def get_method(self
):
2551 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2552 if get_attr
and v
is not None:
2553 v
= getattr(v
, get_attr
, None)
2555 return int(v
) * invscale
// scale
2556 except (ValueError, TypeError, OverflowError):
2560 def str_or_none(v
, default
=None):
2561 return default
if v
is None else str(v
)
2564 def str_to_int(int_str
):
2565 """ A more relaxed version of int_or_none """
2566 if isinstance(int_str
, int):
2568 elif isinstance(int_str
, str):
2569 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2570 return int_or_none(int_str
)
2573 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2577 return float(v
) * invscale
/ scale
2578 except (ValueError, TypeError):
2582 def bool_or_none(v
, default
=None):
2583 return v
if isinstance(v
, bool) else default
2586 def strip_or_none(v
, default
=None):
2587 return v
.strip() if isinstance(v
, str) else default
2590 def url_or_none(url
):
2591 if not url
or not isinstance(url
, str):
2594 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2597 def request_to_url(req
):
2598 if isinstance(req
, urllib
.request
.Request
):
2599 return req
.get_full_url()
2604 def strftime_or_none(timestamp
, date_format
, default
=None):
2605 datetime_object
= None
2607 if isinstance(timestamp
, (int, float)): # unix timestamp
2608 # Using naive datetime here can break timestamp() in Windows
2609 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2610 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2611 elif isinstance(timestamp
, str): # assume YYYYMMDD
2612 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2613 date_format
= re
.sub( # Support %s on windows
2614 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2615 return datetime_object
.strftime(date_format
)
2616 except (ValueError, TypeError, AttributeError):
2620 def parse_duration(s
):
2621 if not isinstance(s
, str):
2627 days
, hours
, mins
, secs
, ms
= [None] * 5
2628 m
= re
.match(r
'''(?x)
2630 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2631 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2632 (?P<ms>[.:][0-9]+)?Z?$
2635 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2640 [0-9]+\s*y(?:ears?)?,?\s*
2643 [0-9]+\s*m(?:onths?)?,?\s*
2646 [0-9]+\s*w(?:eeks?)?,?\s*
2649 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2653 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2656 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2659 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2662 days
, hours
, mins
, secs
, ms
= m
.groups()
2664 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2666 hours
, mins
= m
.groups()
2671 ms
= ms
.replace(':', '.')
2672 return sum(float(part
or 0) * mult
for part
, mult
in (
2673 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2676 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2677 name
, real_ext
= os
.path
.splitext(filename
)
2679 f
'{name}.{ext}{real_ext}'
2680 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2681 else f
'{filename}.{ext}')
2684 def replace_extension(filename
, ext
, expected_real_ext
=None):
2685 name
, real_ext
= os
.path
.splitext(filename
)
2686 return '{}.{}'.format(
2687 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2691 def check_executable(exe
, args
=[]):
2692 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2693 args can be a list of arguments for a short output (like -version) """
2695 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2701 def _get_exe_version_output(exe
, args
):
2703 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2704 # SIGTTOU if yt-dlp is run in the background.
2705 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2706 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2707 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2713 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2714 assert isinstance(output
, str)
2715 if version_re
is None:
2716 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2717 m
= re
.search(version_re
, output
)
2724 def get_exe_version(exe
, args
=['--version'],
2725 version_re
=None, unrecognized
='present'):
2726 """ Returns the version of the specified executable,
2727 or False if the executable is not present """
2728 out
= _get_exe_version_output(exe
, args
)
2729 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2732 def frange(start
=0, stop
=None, step
=1):
2735 start
, stop
= 0, start
2736 sign
= [-1, 1][step
> 0] if step
else 0
2737 while sign
* start
< sign
* stop
:
2742 class LazyList(collections
.abc
.Sequence
):
2743 """Lazy immutable list from an iterable
2744 Note that slices of a LazyList are lists and not LazyList"""
2746 class IndexError(IndexError):
2749 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2750 self
._iterable
= iter(iterable
)
2751 self
._cache
= [] if _cache
is None else _cache
2752 self
._reversed
= reverse
2756 # We need to consume the entire iterable to iterate in reverse
2757 yield from self
.exhaust()
2759 yield from self
._cache
2760 for item
in self
._iterable
:
2761 self
._cache
.append(item
)
2765 self
._cache
.extend(self
._iterable
)
2766 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2770 """Evaluate the entire iterable"""
2771 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2774 def _reverse_index(x
):
2775 return None if x
is None else ~x
2777 def __getitem__(self
, idx
):
2778 if isinstance(idx
, slice):
2780 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2781 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2782 elif isinstance(idx
, int):
2784 idx
= self
._reverse
_index
(idx
)
2785 start
, stop
, step
= idx
, idx
, 0
2787 raise TypeError('indices must be integers or slices')
2788 if ((start
or 0) < 0 or (stop
or 0) < 0
2789 or (start
is None and step
< 0)
2790 or (stop
is None and step
> 0)):
2791 # We need to consume the entire iterable to be able to slice from the end
2792 # Obviously, never use this with infinite iterables
2795 return self
._cache
[idx
]
2796 except IndexError as e
:
2797 raise self
.IndexError(e
) from e
2798 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2800 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2802 return self
._cache
[idx
]
2803 except IndexError as e
:
2804 raise self
.IndexError(e
) from e
2808 self
[-1] if self
._reversed
else self
[0]
2809 except self
.IndexError:
2815 return len(self
._cache
)
2817 def __reversed__(self
):
2818 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2821 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2824 # repr and str should mimic a list. So we exhaust the iterable
2825 return repr(self
.exhaust())
2828 return repr(self
.exhaust())
2833 class IndexError(IndexError):
2837 # This is only useful for tests
2838 return len(self
.getslice())
2840 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2841 self
._pagefunc
= pagefunc
2842 self
._pagesize
= pagesize
2843 self
._pagecount
= float('inf')
2844 self
._use
_cache
= use_cache
2847 def getpage(self
, pagenum
):
2848 page_results
= self
._cache
.get(pagenum
)
2849 if page_results
is None:
2850 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2852 self
._cache
[pagenum
] = page_results
2855 def getslice(self
, start
=0, end
=None):
2856 return list(self
._getslice
(start
, end
))
2858 def _getslice(self
, start
, end
):
2859 raise NotImplementedError('This method must be implemented by subclasses')
2861 def __getitem__(self
, idx
):
2862 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2863 if not isinstance(idx
, int) or idx
< 0:
2864 raise TypeError('indices must be non-negative integers')
2865 entries
= self
.getslice(idx
, idx
+ 1)
2867 raise self
.IndexError()
2871 class OnDemandPagedList(PagedList
):
2872 """Download pages until a page with less than maximum results"""
2874 def _getslice(self
, start
, end
):
2875 for pagenum
in itertools
.count(start
// self
._pagesize
):
2876 firstid
= pagenum
* self
._pagesize
2877 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2878 if start
>= nextfirstid
:
2882 start
% self
._pagesize
2883 if firstid
<= start
< nextfirstid
2886 ((end
- 1) % self
._pagesize
) + 1
2887 if (end
is not None and firstid
<= end
<= nextfirstid
)
2891 page_results
= self
.getpage(pagenum
)
2893 self
._pagecount
= pagenum
- 1
2895 if startv
!= 0 or endv
is not None:
2896 page_results
= page_results
[startv
:endv
]
2897 yield from page_results
2899 # A little optimization - if current page is not "full", ie. does
2900 # not contain page_size videos then we can assume that this page
2901 # is the last one - there are no more ids on further pages -
2902 # i.e. no need to query again.
2903 if len(page_results
) + startv
< self
._pagesize
:
2906 # If we got the whole page, but the next page is not interesting,
2907 # break out early as well
2908 if end
== nextfirstid
:
2912 class InAdvancePagedList(PagedList
):
2913 """PagedList with total number of pages known in advance"""
2915 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2916 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2917 self
._pagecount
= pagecount
2919 def _getslice(self
, start
, end
):
2920 start_page
= start
// self
._pagesize
2921 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2922 skip_elems
= start
- start_page
* self
._pagesize
2923 only_more
= None if end
is None else end
- start
2924 for pagenum
in range(start_page
, end_page
):
2925 page_results
= self
.getpage(pagenum
)
2927 page_results
= page_results
[skip_elems
:]
2929 if only_more
is not None:
2930 if len(page_results
) < only_more
:
2931 only_more
-= len(page_results
)
2933 yield from page_results
[:only_more
]
2935 yield from page_results
2938 class PlaylistEntries
:
2939 MissingEntry
= object()
2940 is_exhausted
= False
2942 def __init__(self
, ydl
, info_dict
):
2945 # _entries must be assigned now since infodict can change during iteration
2946 entries
= info_dict
.get('entries')
2948 raise EntryNotInPlaylist('There are no entries')
2949 elif isinstance(entries
, list):
2950 self
.is_exhausted
= True
2952 requested_entries
= info_dict
.get('requested_entries')
2953 self
.is_incomplete
= requested_entries
is not None
2954 if self
.is_incomplete
:
2955 assert self
.is_exhausted
2956 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
2957 for i
, entry
in zip(requested_entries
, entries
):
2958 self
._entries
[i
- 1] = entry
2959 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2960 self
._entries
= entries
2962 self
._entries
= LazyList(entries
)
2964 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2965 (?P<start>[+-]?\d+)?
2967 (?P<end>[+-]?\d+|inf(?:inite)?)?
2968 (?::(?P<step>[+-]?\d+))?
2972 def parse_playlist_items(cls
, string
):
2973 for segment
in string
.split(','):
2975 raise ValueError('There is two or more consecutive commas')
2976 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2978 raise ValueError(f
'{segment!r} is not a valid specification')
2979 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2980 if int_or_none(step
) == 0:
2981 raise ValueError(f
'Step in {segment!r} cannot be zero')
2982 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2984 def get_requested_items(self
):
2985 playlist_items
= self
.ydl
.params
.get('playlist_items')
2986 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2987 playlist_end
= self
.ydl
.params
.get('playlistend')
2988 # For backwards compatibility, interpret -1 as whole list
2989 if playlist_end
in (-1, None):
2991 if not playlist_items
:
2992 playlist_items
= f
'{playlist_start}:{playlist_end}'
2993 elif playlist_start
!= 1 or playlist_end
:
2994 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2996 for index
in self
.parse_playlist_items(playlist_items
):
2997 for i
, entry
in self
[index
]:
3002 # TODO: Add auto-generated fields
3003 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3004 except (ExistingVideoReached
, RejectedVideoReached
):
3007 def get_full_count(self
):
3008 if self
.is_exhausted
and not self
.is_incomplete
:
3010 elif isinstance(self
._entries
, InAdvancePagedList
):
3011 if self
._entries
._pagesize
== 1:
3012 return self
._entries
._pagecount
3014 @functools.cached_property
3016 if isinstance(self
._entries
, list):
3019 entry
= self
._entries
[i
]
3021 entry
= self
.MissingEntry
3022 if not self
.is_incomplete
:
3023 raise self
.IndexError()
3024 if entry
is self
.MissingEntry
:
3025 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
3030 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3031 except (LazyList
.IndexError, PagedList
.IndexError):
3032 raise self
.IndexError()
3035 def __getitem__(self
, idx
):
3036 if isinstance(idx
, int):
3037 idx
= slice(idx
, idx
)
3039 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3040 step
= 1 if idx
.step
is None else idx
.step
3041 if idx
.start
is None:
3042 start
= 0 if step
> 0 else len(self
) - 1
3044 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3046 # NB: Do not call len(self) when idx == [:]
3047 if idx
.stop
is None:
3048 stop
= 0 if step
< 0 else float('inf')
3050 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3051 stop
+= [-1, 1][step
> 0]
3053 for i
in frange(start
, stop
, step
):
3057 entry
= self
._getter
(i
)
3058 except self
.IndexError:
3059 self
.is_exhausted
= True
3066 return len(tuple(self
[:]))
3068 class IndexError(IndexError):
3072 def uppercase_escape(s
):
3073 unicode_escape
= codecs
.getdecoder('unicode_escape')
3075 r
'\\U[0-9a-fA-F]{8}',
3076 lambda m
: unicode_escape(m
.group(0))[0],
3080 def lowercase_escape(s
):
3081 unicode_escape
= codecs
.getdecoder('unicode_escape')
3083 r
'\\u[0-9a-fA-F]{4}',
3084 lambda m
: unicode_escape(m
.group(0))[0],
3088 def escape_rfc3986(s
):
3089 """Escape non-ASCII characters as suggested by RFC 3986"""
3090 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3093 def escape_url(url
):
3094 """Escape URL as suggested by RFC 3986"""
3095 url_parsed
= urllib
.parse
.urlparse(url
)
3096 return url_parsed
._replace
(
3097 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3098 path
=escape_rfc3986(url_parsed
.path
),
3099 params
=escape_rfc3986(url_parsed
.params
),
3100 query
=escape_rfc3986(url_parsed
.query
),
3101 fragment
=escape_rfc3986(url_parsed
.fragment
)
3105 def parse_qs(url
, **kwargs
):
3106 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3109 def read_batch_urls(batch_fd
):
3111 if not isinstance(url
, str):
3112 url
= url
.decode('utf-8', 'replace')
3113 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3114 for bom
in BOM_UTF8
:
3115 if url
.startswith(bom
):
3116 url
= url
[len(bom
):]
3118 if not url
or url
.startswith(('#', ';', ']')):
3120 # "#" cannot be stripped out since it is part of the URI
3121 # However, it can be safely stripped out if following a whitespace
3122 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3124 with contextlib
.closing(batch_fd
) as fd
:
3125 return [url
for url
in map(fixup
, fd
) if url
]
3128 def urlencode_postdata(*args
, **kargs
):
3129 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3132 def update_url_query(url
, query
):
3135 parsed_url
= urllib
.parse
.urlparse(url
)
3136 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3138 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3139 query
=urllib
.parse
.urlencode(qs
, True)))
3142 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3143 req_headers
= req
.headers
.copy()
3144 req_headers
.update(headers
or {})
3145 req_data
= data
or req
.data
3146 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3147 req_get_method
= req
.get_method()
3148 if req_get_method
== 'HEAD':
3149 req_type
= HEADRequest
3150 elif req_get_method
== 'PUT':
3151 req_type
= PUTRequest
3153 req_type
= urllib
.request
.Request
3155 req_url
, data
=req_data
, headers
=req_headers
,
3156 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3157 if hasattr(req
, 'timeout'):
3158 new_req
.timeout
= req
.timeout
3162 def _multipart_encode_impl(data
, boundary
):
3163 content_type
= 'multipart/form-data; boundary=%s' % boundary
3166 for k
, v
in data
.items():
3167 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3168 if isinstance(k
, str):
3170 if isinstance(v
, str):
3172 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3173 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3174 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3175 if boundary
.encode('ascii') in content
:
3176 raise ValueError('Boundary overlaps with data')
3179 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3181 return out
, content_type
3184 def multipart_encode(data
, boundary
=None):
3186 Encode a dict to RFC 7578-compliant form-data
3189 A dict where keys and values can be either Unicode or bytes-like
3192 If specified a Unicode object, it's used as the boundary. Otherwise
3193 a random boundary is generated.
3195 Reference: https://tools.ietf.org/html/rfc7578
3197 has_specified_boundary
= boundary
is not None
3200 if boundary
is None:
3201 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3204 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3207 if has_specified_boundary
:
3211 return out
, content_type
3214 def variadic(x
, allowed_types
=(str, bytes, dict)):
3215 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3218 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3219 for val
in map(d
.get
, variadic(key_or_keys
)):
3220 if val
is not None and (val
or not skip_false_values
):
3225 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3228 val
= f(*args
, **kwargs
)
3229 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3232 if expected_type
is None or isinstance(val
, expected_type
):
3236 def try_get(src
, getter
, expected_type
=None):
3237 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3240 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3241 return {k: v for k, v in dct.items() if cndn(k, v)}
3244 def merge_dicts(*dicts
):
3246 for a_dict
in dicts
:
3247 for k
, v
in a_dict
.items():
3248 if (v
is not None and k
not in merged
3249 or isinstance(v
, str) and merged
[k
] == ''):
3254 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3255 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3267 TV_PARENTAL_GUIDELINES
= {
3277 def parse_age_limit(s
):
3278 # isinstance(False, int) is True. So type() must be used instead
3279 if type(s
) is int: # noqa: E721
3280 return s
if 0 <= s
<= 21 else None
3281 elif not isinstance(s
, str):
3283 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3285 return int(m
.group('age'))
3288 return US_RATINGS
[s
]
3289 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3291 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3295 def strip_jsonp(code
):
3298 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3299 (?:\s*&&\s*(?P=func_name))?
3300 \s*\(\s*(?P<callback_data>.*)\);?
3301 \s*?(?://[^\n]*)*$''',
3302 r
'\g<callback_data>', code
)
3305 def js_to_json(code
, vars={}, *, strict
=False):
3306 # vars is a dict of var, val pairs to substitute
3307 STRING_QUOTES
= '\'"'
3308 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3309 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3310 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3312 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3313 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3316 def process_escape(match
):
3317 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3318 escape
= match
.group(1) or match
.group(2)
3320 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3321 else R
'\u00' if escape
== 'x'
3322 else '' if escape
== '\n'
3327 if v
in ('true', 'false', 'null'):
3329 elif v
in ('undefined', 'void 0'):
3331 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3334 if v
[0] in STRING_QUOTES
:
3335 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3336 return f
'"{escaped}"'
3338 for regex
, base
in INTEGER_TABLE
:
3339 im
= re
.match(regex
, v
)
3341 i
= int(im
.group(1), base
)
3342 return f
'"{i}":' if v
.endswith(':') else str(i
)
3345 return json
.dumps(vars[v
])
3350 raise ValueError(f
'Unknown value: {v}')
3352 def create_map(mobj
):
3353 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3355 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3357 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3358 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3360 return re
.sub(rf
'''(?sx)
3362 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3363 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3364 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3365 [0-9]+(?={SKIP_RE}:)|
3370 def qualities(quality_ids
):
3371 """ Get a numeric quality value out of a list of possible values """
3374 return quality_ids
.index(qid
)
3380 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3384 'default': '%(title)s [%(id)s].%(ext)s',
3385 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3391 'description': 'description',
3392 'annotation': 'annotations.xml',
3393 'infojson': 'info.json',
3396 'pl_thumbnail': None,
3397 'pl_description': 'description',
3398 'pl_infojson': 'info.json',
3401 # As of [1] format syntax is:
3402 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3403 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3404 STR_FORMAT_RE_TMPL
= r
'''(?x)
3405 (?<!%)(?P<prefix>(?:%%)*)
3407 (?P<has_key>\((?P<key>{0})\))?
3409 (?P<conversion>[#0\-+ ]+)?
3411 (?P<precision>\.\d+)?
3412 (?P<len_mod>[hlL])? # unused in python
3413 {1} # conversion type
3418 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3421 def limit_length(s
, length
):
3422 """ Add ellipses to overly long strings """
3427 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3431 def version_tuple(v
):
3432 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3435 def is_outdated_version(version
, limit
, assume_new
=True):
3437 return not assume_new
3439 return version_tuple(version
) < version_tuple(limit
)
3441 return not assume_new
3444 def ytdl_is_updateable():
3445 """ Returns if yt-dlp can be updated with -U """
3447 from .update
import is_non_updateable
3449 return not is_non_updateable()
3452 def args_to_str(args
):
3453 # Get a short string representation for a subprocess command
3454 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3457 def error_to_compat_str(err
):
3461 def error_to_str(err
):
3462 return f
'{type(err).__name__}: {err}'
3465 def mimetype2ext(mt
):
3469 mt
, _
, params
= mt
.partition(';')
3474 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3475 # it's the most popular one
3476 'audio/mpeg': 'mp3',
3477 'audio/x-wav': 'wav',
3479 'audio/wave': 'wav',
3482 ext
= FULL_MAP
.get(mt
)
3488 'smptett+xml': 'tt',
3492 'x-mp4-fragmented': 'mp4',
3493 'x-ms-sami': 'sami',
3496 'x-mpegurl': 'm3u8',
3497 'vnd.apple.mpegurl': 'm3u8',
3501 'vnd.ms-sstr+xml': 'ism',
3505 'filmstrip+json': 'fs',
3509 _
, _
, subtype
= mt
.rpartition('/')
3510 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3521 _
, _
, suffix
= subtype
.partition('+')
3522 ext
= SUFFIX_MAP
.get(suffix
)
3526 return subtype
.replace('+', '.')
3529 def ext2mimetype(ext_or_url
):
3532 if '.' not in ext_or_url
:
3533 ext_or_url
= f
'file.{ext_or_url}'
3534 return mimetypes
.guess_type(ext_or_url
)[0]
3537 def parse_codecs(codecs_str
):
3538 # http://tools.ietf.org/html/rfc6381
3541 split_codecs
= list(filter(None, map(
3542 str.strip
, codecs_str
.strip().strip(',').split(','))))
3543 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3544 for full_codec
in split_codecs
:
3545 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3546 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3547 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3551 if parts
[0] in ('dvh1', 'dvhe'):
3553 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3555 elif parts
[:2] == ['vp9', '2']:
3557 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3558 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3559 acodec
= acodec
or full_codec
3560 elif parts
[0] in ('stpp', 'wvtt'):
3561 scodec
= scodec
or full_codec
3563 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3564 if vcodec
or acodec
or scodec
:
3566 'vcodec': vcodec
or 'none',
3567 'acodec': acodec
or 'none',
3568 'dynamic_range': hdr
,
3569 **({'scodec': scodec}
if scodec
is not None else {}),
3571 elif len(split_codecs
) == 2:
3573 'vcodec': split_codecs
[0],
3574 'acodec': split_codecs
[1],
3579 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3580 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3582 allow_mkv
= not preferences
or 'mkv' in preferences
3584 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3585 return 'mkv' # TODO: any other format allows this?
3587 # TODO: All codecs supported by parse_codecs isn't handled here
3588 COMPATIBLE_CODECS
= {
3590 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3591 'h264', 'aacl', 'ec-3', # Set in ISM
3594 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3595 'vp9x', 'vp8x', # in the webm spec
3599 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3600 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3602 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3603 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3604 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3608 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3611 for ext
in preferences
or vexts
:
3612 current_exts
= {ext, *vexts, *aexts}
3613 if ext
== 'mkv' or current_exts
== {ext}
or any(
3614 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3616 return 'mkv' if allow_mkv
else preferences
[-1]
3619 def urlhandle_detect_ext(url_handle
):
3620 getheader
= url_handle
.headers
.get
3622 cd
= getheader('Content-Disposition')
3624 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3626 e
= determine_ext(m
.group('filename'), default_ext
=None)
3630 return mimetype2ext(getheader('Content-Type'))
3633 def encode_data_uri(data
, mime_type
):
3634 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3637 def age_restricted(content_limit
, age_limit
):
3638 """ Returns True iff the content should be blocked """
3640 if age_limit
is None: # No limit set
3642 if content_limit
is None:
3643 return False # Content available for everyone
3644 return age_limit
< content_limit
3647 # List of known byte-order-marks (BOM)
3649 (b
'\xef\xbb\xbf', 'utf-8'),
3650 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3651 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3652 (b
'\xff\xfe', 'utf-16-le'),
3653 (b
'\xfe\xff', 'utf-16-be'),
3657 def is_html(first_bytes
):
3658 """ Detect whether a file contains HTML by examining its first bytes. """
3661 for bom
, enc
in BOMS
:
3662 while first_bytes
.startswith(bom
):
3663 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3665 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3668 def determine_protocol(info_dict
):
3669 protocol
= info_dict
.get('protocol')
3670 if protocol
is not None:
3673 url
= sanitize_url(info_dict
['url'])
3674 if url
.startswith('rtmp'):
3676 elif url
.startswith('mms'):
3678 elif url
.startswith('rtsp'):
3681 ext
= determine_ext(url
)
3683 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3687 return urllib
.parse
.urlparse(url
).scheme
3690 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3691 """ Render a list of rows, each as a list of values.
3692 Text after a \t will be right aligned """
3694 return len(remove_terminal_sequences(string
).replace('\t', ''))
3696 def get_max_lens(table
):
3697 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3699 def filter_using_list(row
, filterArray
):
3700 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3702 max_lens
= get_max_lens(data
) if hide_empty
else []
3703 header_row
= filter_using_list(header_row
, max_lens
)
3704 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3706 table
= [header_row
] + data
3707 max_lens
= get_max_lens(table
)
3710 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3711 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3713 for pos
, text
in enumerate(map(str, row
)):
3715 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3717 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3718 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3722 def _match_one(filter_part
, dct
, incomplete
):
3723 # TODO: Generalize code with YoutubeDL._build_format_filter
3724 STRING_OPERATORS
= {
3725 '*=': operator
.contains
,
3726 '^=': lambda attr
, value
: attr
.startswith(value
),
3727 '$=': lambda attr
, value
: attr
.endswith(value
),
3728 '~=': lambda attr
, value
: re
.search(value
, attr
),
3730 COMPARISON_OPERATORS
= {
3732 '<=': operator
.le
, # "<=" must be defined above "<"
3739 if isinstance(incomplete
, bool):
3740 is_incomplete
= lambda _
: incomplete
3742 is_incomplete
= lambda k
: k
in incomplete
3744 operator_rex
= re
.compile(r
'''(?x)
3746 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3748 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3751 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3752 m = operator_rex.fullmatch(filter_part.strip())
3755 unnegated_op = COMPARISON_OPERATORS[m['op']]
3757 op = lambda attr, value: not unnegated_op(attr, value)
3760 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3762 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3763 actual_value = dct.get(m['key'])
3764 numeric_comparison = None
3765 if isinstance(actual_value, (int, float)):
3766 # If the original field is a string and matching comparisonvalue is
3767 # a number we should respect the origin of the original field
3768 # and process comparison value as a string (see
3769 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3771 numeric_comparison = int(comparison_value)
3773 numeric_comparison = parse_filesize(comparison_value)
3774 if numeric_comparison is None:
3775 numeric_comparison = parse_filesize(f'{comparison_value}B')
3776 if numeric_comparison is None:
3777 numeric_comparison = parse_duration(comparison_value)
3778 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3779 raise ValueError('Operator %s only supports string values!' % m['op'])
3780 if actual_value is None:
3781 return is_incomplete(m['key']) or m['none_inclusive']
3782 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3785 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3786 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3788 operator_rex = re.compile(r'''(?x
)
3789 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3790 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3791 m = operator_rex.fullmatch(filter_part.strip())
3793 op = UNARY_OPERATORS[m.group('op')]
3794 actual_value = dct.get(m.group('key'))
3795 if is_incomplete(m.group('key')) and actual_value is None:
3797 return op(actual_value)
3799 raise ValueError('Invalid filter part %r' % filter_part)
3802 def match_str(filter_str, dct, incomplete=False):
3803 """ Filter a dictionary with a simple string syntax.
3804 @returns Whether the filter passes
3805 @param incomplete Set of keys that is expected to be missing from dct.
3806 Can be True/False to indicate all/none of the keys may be missing.
3807 All conditions on incomplete keys pass if the key is missing
3810 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3811 for filter_part in re.split(r'(?<!\\)&', filter_str))
3814 def match_filter_func(filters):
3817 filters = set(variadic(filters))
3819 interactive = '-' in filters
3823 def _match_func(info_dict, incomplete=False):
3824 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3825 return NO_DEFAULT if interactive and not incomplete else None
3827 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3828 filter_str = ') | ('.join(map(str.strip, filters))
3829 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3833 class download_range_func:
3834 def __init__(self, chapters, ranges):
3835 self.chapters, self.ranges = chapters, ranges
3837 def __call__(self, info_dict, ydl):
3838 if not self.ranges and not self.chapters:
3841 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3842 else 'Cannot match chapters since chapter information is unavailable')
3843 for regex in self.chapters or []:
3844 for i, chapter in enumerate(info_dict.get('chapters') or []):
3845 if re.search(regex, chapter['title']):
3847 yield {**chapter, 'index': i}
3848 if self.chapters and warning:
3849 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3851 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3853 def __eq__(self, other):
3854 return (isinstance(other, download_range_func)
3855 and self.chapters == other.chapters and self.ranges == other.ranges)
3858 def parse_dfxp_time_expr(time_expr):
3862 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3864 return float(mobj.group('time_offset'))
3866 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3868 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3871 def srt_subtitles_timecode(seconds):
3872 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3875 def ass_subtitles_timecode(seconds):
3876 time = timetuple_from_msec(seconds * 1000)
3877 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3880 def dfxp2srt(dfxp_data):
3882 @param dfxp_data A
bytes-like
object containing DFXP data
3883 @returns A
unicode object containing converted SRT data
3885 LEGACY_NAMESPACES = (
3886 (b'http://www.w3.org/ns/ttml', [
3887 b'http://www.w3.org/2004/11/ttaf1',
3888 b'http://www.w3.org/2006/04/ttaf1',
3889 b'http://www.w3.org/2006/10/ttaf1',
3891 (b'http://www.w3.org/ns/ttml#styling', [
3892 b'http://www.w3.org/ns/ttml#style',
3896 SUPPORTED_STYLING = [
3905 _x = functools.partial(xpath_with_ns, ns_map={
3906 'xml': 'http://www.w3.org/XML/1998/namespace',
3907 'ttml': 'http://www.w3.org/ns/ttml',
3908 'tts': 'http://www.w3.org/ns/ttml#styling',
3914 class TTMLPElementParser:
3916 _unclosed_elements = []
3917 _applied_styles = []
3919 def start(self, tag, attrib):
3920 if tag in (_x('ttml:br'), 'br'):
3923 unclosed_elements = []
3925 element_style_id = attrib.get('style')
3927 style.update(default_style)
3928 if element_style_id:
3929 style.update(styles.get(element_style_id, {}))
3930 for prop in SUPPORTED_STYLING:
3931 prop_val = attrib.get(_x('tts:' + prop))
3933 style[prop] = prop_val
3936 for k, v in sorted(style.items()):
3937 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3940 font += ' color="%s"' % v
3941 elif k == 'fontSize':
3942 font += ' size="%s"' % v
3943 elif k == 'fontFamily':
3944 font += ' face="%s"' % v
3945 elif k == 'fontWeight' and v == 'bold':
3947 unclosed_elements.append('b')
3948 elif k == 'fontStyle' and v == 'italic':
3950 unclosed_elements.append('i')
3951 elif k == 'textDecoration' and v == 'underline':
3953 unclosed_elements.append('u')
3955 self._out += '<font' + font + '>'
3956 unclosed_elements.append('font')
3958 if self._applied_styles:
3959 applied_style.update(self._applied_styles[-1])
3960 applied_style.update(style)
3961 self._applied_styles.append(applied_style)
3962 self._unclosed_elements.append(unclosed_elements)
3965 if tag not in (_x('ttml:br'), 'br'):
3966 unclosed_elements = self._unclosed_elements.pop()
3967 for element in reversed(unclosed_elements):
3968 self._out += '</%s>' % element
3969 if unclosed_elements and self._applied_styles:
3970 self._applied_styles.pop()
3972 def data(self, data):
3976 return self._out.strip()
3978 def parse_node(node):
3979 target = TTMLPElementParser()
3980 parser = xml.etree.ElementTree.XMLParser(target=target)
3981 parser.feed(xml.etree.ElementTree.tostring(node))
3982 return parser.close()
3984 for k, v in LEGACY_NAMESPACES:
3986 dfxp_data = dfxp_data.replace(ns, k)
3988 dfxp = compat_etree_fromstring(dfxp_data)
3990 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3993 raise ValueError('Invalid dfxp/TTML subtitle')
3997 for style in dfxp.findall(_x('.//ttml:style')):
3998 style_id = style.get('id') or style.get(_x('xml:id'))
4001 parent_style_id = style.get('style')
4003 if parent_style_id not in styles:
4006 styles[style_id] = styles[parent_style_id].copy()
4007 for prop in SUPPORTED_STYLING:
4008 prop_val = style.get(_x('tts:' + prop))
4010 styles.setdefault(style_id, {})[prop] = prop_val
4016 for p in ('body', 'div'):
4017 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
4020 style = styles.get(ele.get('style'))
4023 default_style.update(style)
4025 for para, index in zip(paras, itertools.count(1)):
4026 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
4027 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
4028 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
4029 if begin_time is None:
4034 end_time = begin_time + dur
4035 out.append('%d\n%s --> %s\n%s\n\n' % (
4037 srt_subtitles_timecode(begin_time),
4038 srt_subtitles_timecode(end_time),
4044 def cli_option(params, command_option, param, separator=None):
4045 param = params.get(param)
4046 return ([] if param is None
4047 else [command_option, str(param)] if separator is None
4048 else [f'{command_option}{separator}{param}'])
4051 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4052 param = params.get(param)
4053 assert param in (True, False, None)
4054 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4057 def cli_valueless_option(params, command_option, param, expected_value=True):
4058 return [command_option] if params.get(param) == expected_value else []
4061 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4062 if isinstance(argdict, (list, tuple)): # for backward compatibility
4069 assert isinstance(argdict, dict)
4071 assert isinstance(keys, (list, tuple))
4072 for key_list in keys:
4073 arg_list = list(filter(
4074 lambda x: x is not None,
4075 [argdict.get(key.lower()) for key in variadic(key_list)]))
4077 return [arg for args in arg_list for arg in args]
4081 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4082 main_key, exe = main_key.lower(), exe.lower()
4083 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4084 keys = [f'{root_key}{k}' for k in (keys or [''])]
4085 if root_key in keys:
4087 keys.append((main_key, exe))
4088 keys.append('default')
4091 return cli_configuration_args(argdict, keys, default, use_compat)
4095 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4154 'iw': 'heb', # Replaced by he in 1989 revision
4164 'in': 'ind', # Replaced by id in 1989 revision
4279 'ji': 'yid', # Replaced by yi in 1989 revision
4287 def short2long(cls, code):
4288 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4289 return cls._lang_map.get(code[:2])
4292 def long2short(cls, code):
4293 """Convert language code from ISO 639-2/T to ISO 639-1"""
4294 for short_name, long_name in cls._lang_map.items():
4295 if long_name == code:
4300 # From http://data.okfn.org/data/core/country-list
4302 'AF': 'Afghanistan',
4303 'AX': 'Åland Islands',
4306 'AS': 'American Samoa',
4311 'AG': 'Antigua and Barbuda',
4328 'BO': 'Bolivia, Plurinational State of',
4329 'BQ': 'Bonaire, Sint Eustatius and Saba',
4330 'BA': 'Bosnia and Herzegovina',
4332 'BV': 'Bouvet Island',
4334 'IO': 'British Indian Ocean Territory',
4335 'BN': 'Brunei Darussalam',
4337 'BF': 'Burkina Faso',
4343 'KY': 'Cayman Islands',
4344 'CF': 'Central African Republic',
4348 'CX': 'Christmas Island',
4349 'CC': 'Cocos (Keeling) Islands',
4353 'CD': 'Congo, the Democratic Republic of the',
4354 'CK': 'Cook Islands',
4356 'CI': 'Côte d\'Ivoire',
4361 'CZ': 'Czech Republic',
4365 'DO': 'Dominican Republic',
4368 'SV': 'El Salvador',
4369 'GQ': 'Equatorial Guinea',
4373 'FK': 'Falkland Islands (Malvinas)',
4374 'FO': 'Faroe Islands',
4378 'GF': 'French Guiana',
4379 'PF': 'French Polynesia',
4380 'TF': 'French Southern Territories',
4395 'GW': 'Guinea-Bissau',
4398 'HM': 'Heard Island and McDonald Islands',
4399 'VA': 'Holy See (Vatican City State)',
4406 'IR': 'Iran, Islamic Republic of',
4409 'IM': 'Isle of Man',
4419 'KP': 'Korea, Democratic People\'s Republic of',
4420 'KR': 'Korea, Republic of',
4423 'LA': 'Lao People\'s Democratic Republic',
4429 'LI': 'Liechtenstein',
4433 'MK': 'Macedonia, the Former Yugoslav Republic of',
4440 'MH': 'Marshall Islands',
4446 'FM': 'Micronesia, Federated States of',
4447 'MD': 'Moldova, Republic of',
4458 'NL': 'Netherlands',
4459 'NC': 'New Caledonia',
4460 'NZ': 'New Zealand',
4465 'NF': 'Norfolk Island',
4466 'MP': 'Northern Mariana Islands',
4471 'PS': 'Palestine, State of',
4473 'PG': 'Papua New Guinea',
4476 'PH': 'Philippines',
4480 'PR': 'Puerto Rico',
4484 'RU': 'Russian Federation',
4486 'BL': 'Saint Barthélemy',
4487 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4488 'KN': 'Saint Kitts and Nevis',
4489 'LC': 'Saint Lucia',
4490 'MF': 'Saint Martin (French part)',
4491 'PM': 'Saint Pierre and Miquelon',
4492 'VC': 'Saint Vincent and the Grenadines',
4495 'ST': 'Sao Tome and Principe',
4496 'SA': 'Saudi Arabia',
4500 'SL': 'Sierra Leone',
4502 'SX': 'Sint Maarten (Dutch part)',
4505 'SB': 'Solomon Islands',
4507 'ZA': 'South Africa',
4508 'GS': 'South Georgia and the South Sandwich Islands',
4509 'SS': 'South Sudan',
4514 'SJ': 'Svalbard and Jan Mayen',
4517 'CH': 'Switzerland',
4518 'SY': 'Syrian Arab Republic',
4519 'TW': 'Taiwan, Province of China',
4521 'TZ': 'Tanzania, United Republic of',
4523 'TL': 'Timor-Leste',
4527 'TT': 'Trinidad and Tobago',
4530 'TM': 'Turkmenistan',
4531 'TC': 'Turks and Caicos Islands',
4535 'AE': 'United Arab Emirates',
4536 'GB': 'United Kingdom',
4537 'US': 'United States',
4538 'UM': 'United States Minor Outlying Islands',
4542 'VE': 'Venezuela, Bolivarian Republic of',
4544 'VG': 'Virgin Islands, British',
4545 'VI': 'Virgin Islands, U.S.',
4546 'WF': 'Wallis and Futuna',
4547 'EH': 'Western Sahara',
4551 # Not ISO 3166 codes, but used for IP blocks
4552 'AP': 'Asia/Pacific Region',
4557 def short2full(cls, code):
4558 """Convert an ISO 3166-2 country code to the corresponding full name"""
4559 return cls._country_map.get(code.upper())
4563 # Major IPv4 address blocks per country
4565 'AD': '46.172.224.0/19',
4566 'AE': '94.200.0.0/13',
4567 'AF': '149.54.0.0/17',
4568 'AG': '209.59.64.0/18',
4569 'AI': '204.14.248.0/21',
4570 'AL': '46.99.0.0/16',
4571 'AM': '46.70.0.0/15',
4572 'AO': '105.168.0.0/13',
4573 'AP': '182.50.184.0/21',
4574 'AQ': '23.154.160.0/24',
4575 'AR': '181.0.0.0/12',
4576 'AS': '202.70.112.0/20',
4577 'AT': '77.116.0.0/14',
4578 'AU': '1.128.0.0/11',
4579 'AW': '181.41.0.0/18',
4580 'AX': '185.217.4.0/22',
4581 'AZ': '5.197.0.0/16',
4582 'BA': '31.176.128.0/17',
4583 'BB': '65.48.128.0/17',
4584 'BD': '114.130.0.0/16',
4586 'BF': '102.178.0.0/15',
4587 'BG': '95.42.0.0/15',
4588 'BH': '37.131.0.0/17',
4589 'BI': '154.117.192.0/18',
4590 'BJ': '137.255.0.0/16',
4591 'BL': '185.212.72.0/23',
4592 'BM': '196.12.64.0/18',
4593 'BN': '156.31.0.0/16',
4594 'BO': '161.56.0.0/16',
4595 'BQ': '161.0.80.0/20',
4596 'BR': '191.128.0.0/12',
4597 'BS': '24.51.64.0/18',
4598 'BT': '119.2.96.0/19',
4599 'BW': '168.167.0.0/16',
4600 'BY': '178.120.0.0/13',
4601 'BZ': '179.42.192.0/18',
4602 'CA': '99.224.0.0/11',
4603 'CD': '41.243.0.0/16',
4604 'CF': '197.242.176.0/21',
4605 'CG': '160.113.0.0/16',
4606 'CH': '85.0.0.0/13',
4607 'CI': '102.136.0.0/14',
4608 'CK': '202.65.32.0/19',
4609 'CL': '152.172.0.0/14',
4610 'CM': '102.244.0.0/14',
4611 'CN': '36.128.0.0/10',
4612 'CO': '181.240.0.0/12',
4613 'CR': '201.192.0.0/12',
4614 'CU': '152.206.0.0/15',
4615 'CV': '165.90.96.0/19',
4616 'CW': '190.88.128.0/17',
4617 'CY': '31.153.0.0/16',
4618 'CZ': '88.100.0.0/14',
4620 'DJ': '197.241.0.0/17',
4621 'DK': '87.48.0.0/12',
4622 'DM': '192.243.48.0/20',
4623 'DO': '152.166.0.0/15',
4624 'DZ': '41.96.0.0/12',
4625 'EC': '186.68.0.0/15',
4626 'EE': '90.190.0.0/15',
4627 'EG': '156.160.0.0/11',
4628 'ER': '196.200.96.0/20',
4629 'ES': '88.0.0.0/11',
4630 'ET': '196.188.0.0/14',
4631 'EU': '2.16.0.0/13',
4632 'FI': '91.152.0.0/13',
4633 'FJ': '144.120.0.0/16',
4634 'FK': '80.73.208.0/21',
4635 'FM': '119.252.112.0/20',
4636 'FO': '88.85.32.0/19',
4638 'GA': '41.158.0.0/15',
4640 'GD': '74.122.88.0/21',
4641 'GE': '31.146.0.0/16',
4642 'GF': '161.22.64.0/18',
4643 'GG': '62.68.160.0/19',
4644 'GH': '154.160.0.0/12',
4645 'GI': '95.164.0.0/16',
4646 'GL': '88.83.0.0/19',
4647 'GM': '160.182.0.0/15',
4648 'GN': '197.149.192.0/18',
4649 'GP': '104.250.0.0/19',
4650 'GQ': '105.235.224.0/20',
4651 'GR': '94.64.0.0/13',
4652 'GT': '168.234.0.0/16',
4653 'GU': '168.123.0.0/16',
4654 'GW': '197.214.80.0/20',
4655 'GY': '181.41.64.0/18',
4656 'HK': '113.252.0.0/14',
4657 'HN': '181.210.0.0/16',
4658 'HR': '93.136.0.0/13',
4659 'HT': '148.102.128.0/17',
4660 'HU': '84.0.0.0/14',
4661 'ID': '39.192.0.0/10',
4662 'IE': '87.32.0.0/12',
4663 'IL': '79.176.0.0/13',
4664 'IM': '5.62.80.0/20',
4665 'IN': '117.192.0.0/10',
4666 'IO': '203.83.48.0/21',
4667 'IQ': '37.236.0.0/14',
4668 'IR': '2.176.0.0/12',
4669 'IS': '82.221.0.0/16',
4670 'IT': '79.0.0.0/10',
4671 'JE': '87.244.64.0/18',
4672 'JM': '72.27.0.0/17',
4673 'JO': '176.29.0.0/16',
4674 'JP': '133.0.0.0/8',
4675 'KE': '105.48.0.0/12',
4676 'KG': '158.181.128.0/17',
4677 'KH': '36.37.128.0/17',
4678 'KI': '103.25.140.0/22',
4679 'KM': '197.255.224.0/20',
4680 'KN': '198.167.192.0/19',
4681 'KP': '175.45.176.0/22',
4682 'KR': '175.192.0.0/10',
4683 'KW': '37.36.0.0/14',
4684 'KY': '64.96.0.0/15',
4685 'KZ': '2.72.0.0/13',
4686 'LA': '115.84.64.0/18',
4687 'LB': '178.135.0.0/16',
4688 'LC': '24.92.144.0/20',
4689 'LI': '82.117.0.0/19',
4690 'LK': '112.134.0.0/15',
4691 'LR': '102.183.0.0/16',
4692 'LS': '129.232.0.0/17',
4693 'LT': '78.56.0.0/13',
4694 'LU': '188.42.0.0/16',
4695 'LV': '46.109.0.0/16',
4696 'LY': '41.252.0.0/14',
4697 'MA': '105.128.0.0/11',
4698 'MC': '88.209.64.0/18',
4699 'MD': '37.246.0.0/16',
4700 'ME': '178.175.0.0/17',
4701 'MF': '74.112.232.0/21',
4702 'MG': '154.126.0.0/17',
4703 'MH': '117.103.88.0/21',
4704 'MK': '77.28.0.0/15',
4705 'ML': '154.118.128.0/18',
4706 'MM': '37.111.0.0/17',
4707 'MN': '49.0.128.0/17',
4708 'MO': '60.246.0.0/16',
4709 'MP': '202.88.64.0/20',
4710 'MQ': '109.203.224.0/19',
4711 'MR': '41.188.64.0/18',
4712 'MS': '208.90.112.0/22',
4713 'MT': '46.11.0.0/16',
4714 'MU': '105.16.0.0/12',
4715 'MV': '27.114.128.0/18',
4716 'MW': '102.70.0.0/15',
4717 'MX': '187.192.0.0/11',
4718 'MY': '175.136.0.0/13',
4719 'MZ': '197.218.0.0/15',
4720 'NA': '41.182.0.0/16',
4721 'NC': '101.101.0.0/18',
4722 'NE': '197.214.0.0/18',
4723 'NF': '203.17.240.0/22',
4724 'NG': '105.112.0.0/12',
4725 'NI': '186.76.0.0/15',
4726 'NL': '145.96.0.0/11',
4727 'NO': '84.208.0.0/13',
4728 'NP': '36.252.0.0/15',
4729 'NR': '203.98.224.0/19',
4730 'NU': '49.156.48.0/22',
4731 'NZ': '49.224.0.0/14',
4732 'OM': '5.36.0.0/15',
4733 'PA': '186.72.0.0/15',
4734 'PE': '186.160.0.0/14',
4735 'PF': '123.50.64.0/18',
4736 'PG': '124.240.192.0/19',
4737 'PH': '49.144.0.0/13',
4738 'PK': '39.32.0.0/11',
4739 'PL': '83.0.0.0/11',
4740 'PM': '70.36.0.0/20',
4741 'PR': '66.50.0.0/16',
4742 'PS': '188.161.0.0/16',
4743 'PT': '85.240.0.0/13',
4744 'PW': '202.124.224.0/20',
4745 'PY': '181.120.0.0/14',
4746 'QA': '37.210.0.0/15',
4747 'RE': '102.35.0.0/16',
4748 'RO': '79.112.0.0/13',
4749 'RS': '93.86.0.0/15',
4750 'RU': '5.136.0.0/13',
4751 'RW': '41.186.0.0/16',
4752 'SA': '188.48.0.0/13',
4753 'SB': '202.1.160.0/19',
4754 'SC': '154.192.0.0/11',
4755 'SD': '102.120.0.0/13',
4756 'SE': '78.64.0.0/12',
4757 'SG': '8.128.0.0/10',
4758 'SI': '188.196.0.0/14',
4759 'SK': '78.98.0.0/15',
4760 'SL': '102.143.0.0/17',
4761 'SM': '89.186.32.0/19',
4762 'SN': '41.82.0.0/15',
4763 'SO': '154.115.192.0/18',
4764 'SR': '186.179.128.0/17',
4765 'SS': '105.235.208.0/21',
4766 'ST': '197.159.160.0/19',
4767 'SV': '168.243.0.0/16',
4768 'SX': '190.102.0.0/20',
4770 'SZ': '41.84.224.0/19',
4771 'TC': '65.255.48.0/20',
4772 'TD': '154.68.128.0/19',
4773 'TG': '196.168.0.0/14',
4774 'TH': '171.96.0.0/13',
4775 'TJ': '85.9.128.0/18',
4776 'TK': '27.96.24.0/21',
4777 'TL': '180.189.160.0/20',
4778 'TM': '95.85.96.0/19',
4779 'TN': '197.0.0.0/11',
4780 'TO': '175.176.144.0/21',
4781 'TR': '78.160.0.0/11',
4782 'TT': '186.44.0.0/15',
4783 'TV': '202.2.96.0/19',
4784 'TW': '120.96.0.0/11',
4785 'TZ': '156.156.0.0/14',
4786 'UA': '37.52.0.0/14',
4787 'UG': '102.80.0.0/13',
4789 'UY': '167.56.0.0/13',
4790 'UZ': '84.54.64.0/18',
4791 'VA': '212.77.0.0/19',
4792 'VC': '207.191.240.0/21',
4793 'VE': '186.88.0.0/13',
4794 'VG': '66.81.192.0/20',
4795 'VI': '146.226.0.0/16',
4796 'VN': '14.160.0.0/11',
4797 'VU': '202.80.32.0/20',
4798 'WF': '117.20.32.0/21',
4799 'WS': '202.4.32.0/19',
4800 'YE': '134.35.0.0/16',
4801 'YT': '41.242.116.0/22',
4802 'ZA': '41.0.0.0/11',
4803 'ZM': '102.144.0.0/13',
4804 'ZW': '102.177.192.0/18',
4808 def random_ipv4(cls, code_or_block):
4809 if len(code_or_block) == 2:
4810 block = cls._country_ip_map.get(code_or_block.upper())
4814 block = code_or_block
4815 addr, preflen = block.split('/')
4816 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4817 addr_max = addr_min | (0xffffffff >> int(preflen))
4818 return str(socket.inet_ntoa(
4819 struct.pack('!L', random.randint(addr_min, addr_max))))
4822 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4823 def __init__(self, proxies=None):
4824 # Set default handlers
4825 for type in ('http', 'https'):
4826 setattr(self, '%s_open' % type,
4827 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4828 meth(r, proxy, type))
4829 urllib.request.ProxyHandler.__init__(self, proxies)
4831 def proxy_open(self, req, proxy, type):
4832 req_proxy = req.headers.get('Ytdl-request-proxy')
4833 if req_proxy is not None:
4835 del req.headers['Ytdl-request-proxy']
4837 if proxy == '__noproxy__':
4838 return None # No Proxy
4839 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4840 req.add_header('Ytdl-socks-proxy', proxy)
4841 # yt-dlp's http/https handlers do wrapping the socket with socks
4843 return urllib.request.ProxyHandler.proxy_open(
4844 self, req, proxy, type)
4847 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4848 # released into Public Domain
4849 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4851 def long_to_bytes(n, blocksize=0):
4852 """long_to_bytes(n:long, blocksize:int) : string
4853 Convert a long integer to a byte string.
4855 If optional blocksize is given and greater than zero, pad the front of the
4856 byte string with binary zeros so that the length is a multiple of
4859 # after much testing, this algorithm was deemed to be the fastest
4863 s = struct.pack('>I', n & 0xffffffff) + s
4865 # strip off leading zeros
4866 for i in range(len(s)):
4867 if s[i] != b'\000'[0]:
4870 # only happens when n == 0
4874 # add back some pad bytes. this could be done more efficiently w.r.t. the
4875 # de-padding being done above, but sigh...
4876 if blocksize > 0 and len(s) % blocksize:
4877 s = (blocksize - len(s) % blocksize) * b'\000' + s
4881 def bytes_to_long(s):
4882 """bytes_to_long(string) : long
4883 Convert a byte string to a long integer.
4885 This is (essentially) the inverse of long_to_bytes().
4890 extra = (4 - length % 4)
4891 s = b'\000' * extra + s
4892 length = length + extra
4893 for i in range(0, length, 4):
4894 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4898 def ohdave_rsa_encrypt(data, exponent, modulus):
4900 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4903 data: data to encrypt, bytes-like object
4904 exponent, modulus: parameter e and N of RSA algorithm, both integer
4905 Output: hex string of encrypted data
4907 Limitation: supports one block encryption only
4910 payload = int(binascii.hexlify(data[::-1]), 16)
4911 encrypted = pow(payload, exponent, modulus)
4912 return '%x' % encrypted
4915 def pkcs1pad(data, length):
4917 Padding input data with PKCS#1 scheme
4919 @param {int[]} data input data
4920 @param {int} length target length
4921 @returns {int[]} padded data
4923 if len(data) > length - 11:
4924 raise ValueError('Input data too
long for PKCS
#1 padding')
4926 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4927 return [0, 2] + pseudo_random
+ [0] + data
4930 def _base_n_table(n
, table
):
4931 if not table
and not n
:
4932 raise ValueError('Either table or n must be specified')
4933 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4935 if n
and n
!= len(table
):
4936 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4940 def encode_base_n(num
, n
=None, table
=None):
4941 """Convert given int to a base-n string"""
4942 table
= _base_n_table(n
, table
)
4946 result
, base
= '', len(table
)
4948 result
= table
[num
% base
] + result
4953 def decode_base_n(string
, n
=None, table
=None):
4954 """Convert given base-n string to int"""
4955 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4956 result
, base
= 0, len(table
)
4958 result
= result
* base
+ table
[char
]
4962 def decode_base(value
, digits
):
4963 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
4964 f
'in a future version. Use {__name__}.decode_base_n instead')
4965 return decode_base_n(value
, table
=digits
)
4968 def decode_packed_codes(code
):
4969 mobj
= re
.search(PACKED_CODES_RE
, code
)
4970 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4973 symbols
= symbols
.split('|')
4978 base_n_count
= encode_base_n(count
, base
)
4979 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4982 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4986 def caesar(s
, alphabet
, shift
):
4991 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4996 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4999 def parse_m3u8_attributes(attrib
):
5001 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5002 if val
.startswith('"'):
5008 def urshift(val
, n
):
5009 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5012 # Based on png2str() written by @gdkchan and improved by @yokrysty
5013 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5014 def decode_png(png_data
):
5015 # Reference: https://www.w3.org/TR/PNG/
5016 header
= png_data
[8:]
5018 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5019 raise OSError('Not a valid PNG file.')
5021 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5022 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5027 length
= unpack_integer(header
[:4])
5030 chunk_type
= header
[:4]
5033 chunk_data
= header
[:length
]
5034 header
= header
[length
:]
5036 header
= header
[4:] # Skip CRC
5044 ihdr
= chunks
[0]['data']
5046 width
= unpack_integer(ihdr
[:4])
5047 height
= unpack_integer(ihdr
[4:8])
5051 for chunk
in chunks
:
5052 if chunk
['type'] == b
'IDAT':
5053 idat
+= chunk
['data']
5056 raise OSError('Unable to read PNG data.')
5058 decompressed_data
= bytearray(zlib
.decompress(idat
))
5063 def _get_pixel(idx
):
5068 for y
in range(height
):
5069 basePos
= y
* (1 + stride
)
5070 filter_type
= decompressed_data
[basePos
]
5074 pixels
.append(current_row
)
5076 for x
in range(stride
):
5077 color
= decompressed_data
[1 + basePos
+ x
]
5078 basex
= y
* stride
+ x
5083 left
= _get_pixel(basex
- 3)
5085 up
= _get_pixel(basex
- stride
)
5087 if filter_type
== 1: # Sub
5088 color
= (color
+ left
) & 0xff
5089 elif filter_type
== 2: # Up
5090 color
= (color
+ up
) & 0xff
5091 elif filter_type
== 3: # Average
5092 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5093 elif filter_type
== 4: # Paeth
5099 c
= _get_pixel(basex
- stride
- 3)
5107 if pa
<= pb
and pa
<= pc
:
5108 color
= (color
+ a
) & 0xff
5110 color
= (color
+ b
) & 0xff
5112 color
= (color
+ c
) & 0xff
5114 current_row
.append(color
)
5116 return width
, height
, pixels
5119 def write_xattr(path
, key
, value
):
5120 # Windows: Write xattrs to NTFS Alternate Data Streams:
5121 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5122 if compat_os_name
== 'nt':
5123 assert ':' not in key
5124 assert os
.path
.exists(path
)
5127 with open(f
'{path}:{key}', 'wb') as f
:
5129 except OSError as e
:
5130 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5133 # UNIX Method 1. Use xattrs/pyxattrs modules
5136 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5137 # Unicode arguments are not supported in pyxattr until version 0.5.0
5138 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5139 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5140 setxattr
= xattr
.set
5142 setxattr
= xattr
.setxattr
5146 setxattr(path
, key
, value
)
5147 except OSError as e
:
5148 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5151 # UNIX Method 2. Use setfattr/xattr executables
5152 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5153 else 'xattr' if check_executable('xattr', ['-h']) else None)
5155 raise XAttrUnavailableError(
5156 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5157 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5159 value
= value
.decode()
5161 _
, stderr
, returncode
= Popen
.run(
5162 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5163 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5164 except OSError as e
:
5165 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5167 raise XAttrMetadataError(returncode
, stderr
)
5170 def random_birthday(year_field
, month_field
, day_field
):
5171 start_date
= datetime
.date(1950, 1, 1)
5172 end_date
= datetime
.date(1995, 12, 31)
5173 offset
= random
.randint(0, (end_date
- start_date
).days
)
5174 random_date
= start_date
+ datetime
.timedelta(offset
)
5176 year_field
: str(random_date
.year
),
5177 month_field
: str(random_date
.month
),
5178 day_field
: str(random_date
.day
),
5182 # Templates for internet shortcut files, which are plain text files.
5183 DOT_URL_LINK_TEMPLATE
= '''\
5188 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5189 <?xml version="1.0" encoding="UTF-8"?>
5190 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5191 <plist version="1.0">
5194 \t<string>%(url)s</string>
5199 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5209 'url': DOT_URL_LINK_TEMPLATE
,
5210 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5211 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5215 def iri_to_uri(iri
):
5217 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5219 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.
5222 iri_parts
= urllib
.parse
.urlparse(iri
)
5224 if '[' in iri_parts
.netloc
:
5225 raise ValueError('IPv6 URIs are not, yet, supported.')
5226 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5228 # 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.
5231 if iri_parts
.username
:
5232 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5233 if iri_parts
.password
is not None:
5234 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5237 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5238 # The 'idna' encoding produces ASCII text.
5239 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5240 net_location
+= ':' + str(iri_parts
.port
)
5242 return urllib
.parse
.urlunparse(
5246 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5248 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5249 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5251 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5252 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5254 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5256 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5259 def to_high_limit_path(path
):
5260 if sys
.platform
in ['win32', 'cygwin']:
5261 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5262 return '\\\\?\\' + os
.path
.abspath(path
)
5267 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5268 val
= traverse_obj(obj
, *variadic(field
))
5269 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5271 return template
% func(val
)
5274 def clean_podcast_url(url
):
5275 return re
.sub(r
'''(?x)
5279 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5282 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5285 cn\.co| # https://podcorn.com/analytics-prefix/
5286 st\.fm # https://podsights.com/docs/
5291 _HEX_TABLE
= '0123456789abcdef'
5294 def random_uuidv4():
5295 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5298 def make_dir(path
, to_screen
=None):
5300 dn
= os
.path
.dirname(path
)
5301 if dn
and not os
.path
.exists(dn
):
5304 except OSError as err
:
5305 if callable(to_screen
) is not None:
5306 to_screen('unable to create directory ' + error_to_compat_str(err
))
5310 def get_executable_path():
5311 from .update
import _get_variant_and_executable_path
5313 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5316 def load_plugins(name
, suffix
, namespace
):
5318 with contextlib
.suppress(FileNotFoundError
):
5319 plugins_spec
= importlib
.util
.spec_from_file_location(
5320 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5321 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5322 sys
.modules
[plugins_spec
.name
] = plugins
5323 plugins_spec
.loader
.exec_module(plugins
)
5324 for name
in dir(plugins
):
5325 if name
in namespace
:
5327 if not name
.endswith(suffix
):
5329 klass
= getattr(plugins
, name
)
5330 classes
[name
] = namespace
[name
] = klass
5335 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5336 casesense
=True, is_user_input
=False, traverse_string
=False):
5338 Safely traverse nested `dict`s and `Sequence`s
5340 >>> obj = [{}, {"key": "value"}]
5341 >>> traverse_obj(obj, (1, "key"))
5344 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5345 The next path will also be tested if the path branched but no results could be found.
5346 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5347 A value of None is treated as the absence of a value.
5349 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5351 The keys in the path can be one of:
5352 - `None`: Return the current object.
5353 - `str`/`int`: Return `obj[key]`. For `re.Match, return `obj.group(key)`.
5354 - `slice`: Branch out and return all values in `obj[key]`.
5355 - `Ellipsis`: Branch out and return a list of all values.
5356 - `tuple`/`list`: Branch out and return a list of all matching values.
5357 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5358 - `function`: Branch out and return values filtered by the function.
5359 Read as: `[value for key, value in obj if function(key, value)]`.
5360 For `Sequence`s, `key` is the index of the value.
5361 - `dict` Transform the current object and return a matching dict.
5362 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5364 `tuple`, `list`, and `dict` all support nested paths and branches.
5366 @params paths Paths which to traverse by.
5367 @param default Value to return if the paths do not match.
5368 @param expected_type If a `type`, only accept final values of this type.
5369 If any other callable, try to call the function on each result.
5370 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5371 @param casesense If `False`, consider string dictionary keys as case insensitive.
5373 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5375 @param is_user_input Whether the keys are generated from user input.
5376 If `True` strings get converted to `int`/`slice` if needed.
5377 @param traverse_string Whether to traverse into objects as strings.
5378 If `True`, any non-compatible object will first be
5379 converted into a string and then traversed into.
5382 @returns The result of the object traversal.
5383 If successful, `get_all=True`, and the path branches at least once,
5384 then a list of results is returned instead.
5385 A list is always returned if the last path branches and no `default` is given.
5387 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5388 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5390 if isinstance(expected_type
, type):
5391 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5393 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5395 def apply_key(key
, obj
):
5402 elif isinstance(key
, (list, tuple)):
5404 _
, result
= apply_path(obj
, branch
)
5408 if isinstance(obj
, collections
.abc
.Mapping
):
5409 yield from obj
.values()
5410 elif is_sequence(obj
):
5412 elif isinstance(obj
, re
.Match
):
5413 yield from obj
.groups()
5414 elif traverse_string
:
5418 if is_sequence(obj
):
5419 iter_obj
= enumerate(obj
)
5420 elif isinstance(obj
, collections
.abc
.Mapping
):
5421 iter_obj
= obj
.items()
5422 elif isinstance(obj
, re
.Match
):
5423 iter_obj
= enumerate((obj
.group(), *obj
.groups()))
5424 elif traverse_string
:
5425 iter_obj
= enumerate(str(obj
))
5428 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5430 elif isinstance(key
, dict):
5431 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5432 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5433 if v
is not None or default
is not NO_DEFAULT
}
5435 elif isinstance(obj
, collections
.abc
.Mapping
):
5436 yield (obj
.get(key
) if casesense
or (key
in obj
)
5437 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5439 elif isinstance(obj
, re
.Match
):
5440 if isinstance(key
, int) or casesense
:
5441 with contextlib
.suppress(IndexError):
5442 yield obj
.group(key
)
5445 if not isinstance(key
, str):
5448 yield next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5452 key
= (int_or_none(key
) if ':' not in key
5453 else slice(*map(int_or_none
, key
.split(':'))))
5455 if not isinstance(key
, (int, slice)):
5458 if not is_sequence(obj
):
5459 if not traverse_string
:
5463 with contextlib
.suppress(IndexError):
5466 def apply_path(start_obj
, path
):
5468 has_branched
= False
5470 for key
in variadic(path
):
5471 if is_user_input
and key
== ':':
5474 if not casesense
and isinstance(key
, str):
5475 key
= key
.casefold()
5477 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5480 key_func
= functools
.partial(apply_key
, key
)
5481 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5483 return has_branched
, objs
5485 def _traverse_obj(obj
, path
, use_list
=True):
5486 has_branched
, results
= apply_path(obj
, path
)
5487 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5489 if get_all
and has_branched
:
5490 return results
.exhaust() if results
or use_list
else None
5492 return results
[0] if results
else None
5494 for index
, path
in enumerate(paths
, 1):
5495 use_list
= default
is NO_DEFAULT
and index
== len(paths
)
5496 result
= _traverse_obj(obj
, path
, use_list
)
5497 if result
is not None:
5500 return None if default
is NO_DEFAULT
else default
5503 def traverse_dict(dictn
, keys
, casesense
=True):
5504 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5505 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5506 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5509 def get_first(obj
, keys
, **kwargs
):
5510 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5513 def time_seconds(**kwargs
):
5514 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5515 return t
.timestamp()
5518 # create a JSON Web Signature (jws) with HS256 algorithm
5519 # the resulting format is in JWS Compact Serialization
5520 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5521 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5522 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5528 header_data
.update(headers
)
5529 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5530 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5531 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5532 signature_b64
= base64
.b64encode(h
.digest())
5533 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5537 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5538 def jwt_decode_hs256(jwt
):
5539 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5540 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5541 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5545 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5549 def supports_terminal_sequences(stream
):
5550 if compat_os_name
== 'nt':
5551 if not WINDOWS_VT_MODE
:
5553 elif not os
.getenv('TERM'):
5556 return stream
.isatty()
5557 except BaseException
:
5561 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5562 if get_windows_version() < (10, 0, 10586):
5564 global WINDOWS_VT_MODE
5566 Popen
.run('', shell
=True)
5570 WINDOWS_VT_MODE
= True
5571 supports_terminal_sequences
.cache_clear()
5574 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5577 def remove_terminal_sequences(string
):
5578 return _terminal_sequences_re
.sub('', string
)
5581 def number_of_digits(number
):
5582 return len('%d' % number
)
5585 def join_nonempty(*values
, delim
='-', from_dict
=None):
5586 if from_dict
is not None:
5587 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5588 return delim
.join(map(str, filter(None, values
)))
5591 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5593 Find the largest format dimensions in terms of video width and, for each thumbnail:
5594 * Modify the URL: Match the width with the provided regex and replace with the former width
5597 This function is useful with video services that scale the provided thumbnails on demand
5599 _keys
= ('width', 'height')
5600 max_dimensions
= max(
5601 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5603 if not max_dimensions
[0]:
5607 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5608 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5609 for thumbnail
in thumbnails
5613 def parse_http_range(range):
5614 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5616 return None, None, None
5617 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5619 return None, None, None
5620 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5623 def read_stdin(what
):
5624 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5625 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5629 def determine_file_encoding(data
):
5631 Detect the text encoding used
5632 @returns (encoding, bytes to skip)
5635 # BOM marks are given priority over declarations
5636 for bom
, enc
in BOMS
:
5637 if data
.startswith(bom
):
5638 return enc
, len(bom
)
5640 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5641 # We ignore the endianness to get a good enough match
5642 data
= data
.replace(b
'\0', b
'')
5643 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5644 return mobj
.group(1).decode() if mobj
else None, 0
5651 __initialized
= False
5653 def __init__(self
, parser
, label
=None):
5654 self
.parser
, self
.label
= parser
, label
5655 self
._loaded
_paths
, self
.configs
= set(), []
5657 def init(self
, args
=None, filename
=None):
5658 assert not self
.__initialized
5659 self
.own_args
, self
.filename
= args
, filename
5660 return self
.load_configs()
5662 def load_configs(self
):
5665 location
= os
.path
.realpath(self
.filename
)
5666 directory
= os
.path
.dirname(location
)
5667 if location
in self
._loaded
_paths
:
5669 self
._loaded
_paths
.add(location
)
5671 self
.__initialized
= True
5672 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5673 self
.parsed_args
= self
.own_args
5674 for location
in opts
.config_locations
or []:
5676 if location
in self
._loaded
_paths
:
5678 self
._loaded
_paths
.add(location
)
5679 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5681 location
= os
.path
.join(directory
, expand_path(location
))
5682 if os
.path
.isdir(location
):
5683 location
= os
.path
.join(location
, 'yt-dlp.conf')
5684 if not os
.path
.exists(location
):
5685 self
.parser
.error(f
'config location {location} does not exist')
5686 self
.append_config(self
.read_file(location
), location
)
5690 label
= join_nonempty(
5691 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5693 return join_nonempty(
5694 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5695 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5699 def read_file(filename
, default
=[]):
5701 optionf
= open(filename
, 'rb')
5703 return default
# silently skip if file is not present
5705 enc
, skip
= determine_file_encoding(optionf
.read(512))
5706 optionf
.seek(skip
, io
.SEEK_SET
)
5708 enc
= None # silently skip read errors
5710 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5711 contents
= optionf
.read().decode(enc
or preferredencoding())
5712 res
= shlex
.split(contents
, comments
=True)
5713 except Exception as err
:
5714 raise ValueError(f
'Unable to parse "{filename}": {err}')
5720 def hide_login_info(opts
):
5721 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5722 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5727 return m
.group('key') + '=PRIVATE'
5731 opts
= list(map(_scrub_eq
, opts
))
5732 for idx
, opt
in enumerate(opts
):
5733 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5734 opts
[idx
+ 1] = 'PRIVATE'
5737 def append_config(self
, *args
, label
=None):
5738 config
= type(self
)(self
.parser
, label
)
5739 config
._loaded
_paths
= self
._loaded
_paths
5740 if config
.init(*args
):
5741 self
.configs
.append(config
)
5745 for config
in reversed(self
.configs
):
5746 yield from config
.all_args
5747 yield from self
.parsed_args
or []
5749 def parse_known_args(self
, **kwargs
):
5750 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5752 def parse_args(self
):
5753 return self
.parser
.parse_args(self
.all_args
)
5756 class WebSocketsWrapper
:
5757 """Wraps websockets module to use in non-async scopes"""
5760 def __init__(self
, url
, headers
=None, connect
=True):
5761 self
.loop
= asyncio
.new_event_loop()
5762 # XXX: "loop" is deprecated
5763 self
.conn
= websockets
.connect(
5764 url
, extra_headers
=headers
, ping_interval
=None,
5765 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5768 atexit
.register(self
.__exit
__, None, None, None)
5770 def __enter__(self
):
5772 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5775 def send(self
, *args
):
5776 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5778 def recv(self
, *args
):
5779 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5781 def __exit__(self
, type, value
, traceback
):
5783 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5786 self
._cancel
_all
_tasks
(self
.loop
)
5788 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5789 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5791 def run_with_loop(main
, loop
):
5792 if not asyncio
.iscoroutine(main
):
5793 raise ValueError(f
'a coroutine was expected, got {main!r}')
5796 return loop
.run_until_complete(main
)
5798 loop
.run_until_complete(loop
.shutdown_asyncgens())
5799 if hasattr(loop
, 'shutdown_default_executor'):
5800 loop
.run_until_complete(loop
.shutdown_default_executor())
5803 def _cancel_all_tasks(loop
):
5804 to_cancel
= asyncio
.all_tasks(loop
)
5809 for task
in to_cancel
:
5812 # XXX: "loop" is removed in python 3.10+
5813 loop
.run_until_complete(
5814 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5816 for task
in to_cancel
:
5817 if task
.cancelled():
5819 if task
.exception() is not None:
5820 loop
.call_exception_handler({
5821 'message': 'unhandled exception during asyncio.run() shutdown',
5822 'exception': task
.exception(),
5827 def merge_headers(*dicts
):
5828 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5829 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5832 def cached_method(f
):
5833 """Cache a method"""
5834 signature
= inspect
.signature(f
)
5837 def wrapper(self
, *args
, **kwargs
):
5838 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5839 bound_args
.apply_defaults()
5840 key
= tuple(bound_args
.arguments
.values())[1:]
5842 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
5843 if key
not in cache
:
5844 cache
[key
] = f(self
, *args
, **kwargs
)
5849 class classproperty
:
5850 """property access for class methods with optional caching"""
5851 def __new__(cls
, func
=None, *args
, **kwargs
):
5853 return functools
.partial(cls
, *args
, **kwargs
)
5854 return super().__new
__(cls
)
5856 def __init__(self
, func
, *, cache
=False):
5857 functools
.update_wrapper(self
, func
)
5859 self
._cache
= {} if cache
else None
5861 def __get__(self
, _
, cls
):
5862 if self
._cache
is None:
5863 return self
.func(cls
)
5864 elif cls
not in self
._cache
:
5865 self
._cache
[cls
] = self
.func(cls
)
5866 return self
._cache
[cls
]
5869 class Namespace(types
.SimpleNamespace
):
5870 """Immutable namespace"""
5873 return iter(self
.__dict
__.values())
5877 return self
.__dict
__.items()
5880 MEDIA_EXTENSIONS
= Namespace(
5881 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5882 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5883 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5884 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5885 thumbnails
=('jpg', 'png', 'webp'),
5886 storyboards
=('mhtml', ),
5887 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5888 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5890 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5891 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5893 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5898 for retry in RetryManager(...):
5901 except SomeException as err:
5905 attempt
, _error
= 0, None
5907 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5908 self
.retries
= _retries
or 0
5909 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5911 def _should_retry(self
):
5912 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5916 if self
._error
is NO_DEFAULT
:
5921 def error(self
, value
):
5925 while self
._should
_retry
():
5926 self
.error
= NO_DEFAULT
5930 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5933 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5934 """Utility function for reporting retries"""
5937 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5942 elif isinstance(e
, ExtractorError
):
5943 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5944 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5946 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5948 info(f
'Sleeping {delay:.2f} seconds ...')
5952 def make_archive_id(ie
, video_id
):
5953 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5954 return f
'{ie_key.lower()} {video_id}'
5957 def truncate_string(s
, left
, right
=0):
5958 assert left
> 3 and right
>= 0
5959 if s
is None or len(s
) <= left
+ right
:
5961 return f
'{s[:left-3]}...{s[-right:]}'
5964 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5965 assert 'all' in alias_dict
, '"all" alias is required'
5966 requested
= list(start
or [])
5968 discard
= val
.startswith('-')
5972 if val
in alias_dict
:
5973 val
= alias_dict
[val
] if not discard
else [
5974 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5975 # NB: Do not allow regex in aliases for performance
5976 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5979 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5980 else [val
] if val
in alias_dict
['all'] else None)
5982 raise ValueError(val
)
5985 for item
in current
:
5986 while item
in requested
:
5987 requested
.remove(item
)
5989 requested
.extend(current
)
5991 return orderedSet(requested
)
5995 has_certifi
= bool(certifi
)
5996 has_websockets
= bool(websockets
)