47 import xml
.etree
.ElementTree
50 from .compat
import functools
# isort: split
52 compat_etree_fromstring
,
54 compat_HTMLParseError
,
58 from .dependencies
import brotli
, certifi
, websockets
, xattr
59 from .socks
import ProxyType
, sockssocket
62 def register_socks_protocols():
63 # "Register" SOCKS protocols
64 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
65 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
66 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
67 if scheme
not in urllib
.parse
.uses_netloc
:
68 urllib
.parse
.uses_netloc
.append(scheme
)
71 # This is not clearly defined otherwise
72 compiled_regex_type
= type(re
.compile(''))
75 def random_user_agent():
76 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
117 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
120 SUPPORTED_ENCODINGS
= [
124 SUPPORTED_ENCODINGS
.append('br')
127 'User-Agent': random_user_agent(),
128 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
129 'Accept-Language': 'en-us,en;q=0.5',
130 'Sec-Fetch-Mode': 'navigate',
135 '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',
139 NO_DEFAULT
= object()
140 IDENTITY
= lambda x
: x
142 ENGLISH_MONTH_NAMES
= [
143 'January', 'February', 'March', 'April', 'May', 'June',
144 'July', 'August', 'September', 'October', 'November', 'December']
147 'en': ENGLISH_MONTH_NAMES
,
149 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
150 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
151 # these follow the genitive grammatical case (dopełniacz)
152 # some websites might be using nominative, which will require another month list
153 # https://en.wikibooks.org/wiki/Polish/Noun_cases
154 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
155 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
158 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
160 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
161 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
162 'EST': -5, 'EDT': -4, # Eastern
163 'CST': -6, 'CDT': -5, # Central
164 'MST': -7, 'MDT': -6, # Mountain
165 'PST': -8, 'PDT': -7 # Pacific
168 # needed for sanitizing filenames in restricted mode
169 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
170 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
171 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
201 '%Y-%m-%d %H:%M:%S.%f',
202 '%Y-%m-%d %H:%M:%S:%f',
205 '%Y-%m-%dT%H:%M:%SZ',
206 '%Y-%m-%dT%H:%M:%S.%fZ',
207 '%Y-%m-%dT%H:%M:%S.%f0Z',
209 '%Y-%m-%dT%H:%M:%S.%f',
212 '%b %d %Y at %H:%M:%S',
214 '%B %d %Y at %H:%M:%S',
218 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
219 DATE_FORMATS_DAY_FIRST
.extend([
229 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
230 DATE_FORMATS_MONTH_FIRST
.extend([
238 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
239 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
241 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
245 def preferredencoding():
246 """Get preferred encoding.
248 Returns the best encoding scheme for the system, based on
249 locale.getpreferredencoding() and some further tweaks.
252 pref = locale.getpreferredencoding()
260 def write_json_file(obj, fn):
261 """ Encode obj as JSON and write it to fn, atomically if possible """
263 tf = tempfile.NamedTemporaryFile(
264 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
265 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
269 json.dump(obj, tf, ensure_ascii=False)
270 if sys.platform == 'win32
':
271 # Need to remove existing file on Windows, else os.rename raises
272 # WindowsError or FileExistsError.
273 with contextlib.suppress(OSError):
275 with contextlib.suppress(OSError):
278 os.chmod(tf.name, 0o666 & ~mask)
279 os.rename(tf.name, fn)
281 with contextlib.suppress(OSError):
286 def find_xpath_attr(node, xpath, key, val=None):
287 """ Find the xpath xpath[@key=val] """
288 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
289 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
290 return node.find(expr)
292 # On python2.6 the xml.etree.ElementTree.Element methods don't support
293 # the namespace parameter
296 def xpath_with_ns(path
, ns_map
):
297 components
= [c
.split(':') for c
in path
.split('/')]
301 replaced
.append(c
[0])
304 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
305 return '/'.join(replaced
)
308 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
309 def _find_xpath(xpath
):
310 return node
.find(xpath
)
312 if isinstance(xpath
, str):
313 n
= _find_xpath(xpath
)
321 if default
is not NO_DEFAULT
:
324 name
= xpath
if name
is None else name
325 raise ExtractorError('Could not find XML element %s' % name
)
331 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
332 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
333 if n
is None or n
== default
:
336 if default
is not NO_DEFAULT
:
339 name
= xpath
if name
is None else name
340 raise ExtractorError('Could not find XML element\'s text %s' % name
)
346 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
347 n
= find_xpath_attr(node
, xpath
, key
)
349 if default
is not NO_DEFAULT
:
352 name
= f
'{xpath}[@{key}]' if name
is None else name
353 raise ExtractorError('Could not find XML attribute %s' % name
)
359 def get_element_by_id(id, html
, **kwargs
):
360 """Return the content of the tag with the specified ID in the passed HTML document"""
361 return get_element_by_attribute('id', id, html
, **kwargs
)
364 def get_element_html_by_id(id, html
, **kwargs
):
365 """Return the html of the tag with the specified ID in the passed HTML document"""
366 return get_element_html_by_attribute('id', id, html
, **kwargs
)
369 def get_element_by_class(class_name
, html
):
370 """Return the content of the first tag with the specified class in the passed HTML document"""
371 retval
= get_elements_by_class(class_name
, html
)
372 return retval
[0] if retval
else None
375 def get_element_html_by_class(class_name
, html
):
376 """Return the html of the first tag with the specified class in the passed HTML document"""
377 retval
= get_elements_html_by_class(class_name
, html
)
378 return retval
[0] if retval
else None
381 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
382 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
383 return retval
[0] if retval
else None
386 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
387 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
388 return retval
[0] if retval
else None
391 def get_elements_by_class(class_name
, html
, **kargs
):
392 """Return the content of all tags with the specified class in the passed HTML document as a list"""
393 return get_elements_by_attribute(
394 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
395 html, escape_value=False)
398 def get_elements_html_by_class(class_name, html):
399 """Return the html of all tags with the specified class in the passed HTML document as a list"""
400 return get_elements_html_by_attribute(
401 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
402 html, escape_value=False)
405 def get_elements_by_attribute(*args, **kwargs):
406 """Return the content of the tag with the specified attribute in the passed HTML document"""
407 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
410 def get_elements_html_by_attribute(*args, **kwargs):
411 """Return the html of the tag with the specified attribute in the passed HTML document"""
412 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
415 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
417 Return the text (content) and the html (whole) of the tag with the specified
418 attribute in the passed HTML document
423 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
425 value = re.escape(value) if escape_value else value
427 partial_element_re = rf'''(?x
)
429 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
430 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
433 for m in re.finditer(partial_element_re, html):
434 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
437 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
442 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
444 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
445 closing tag for the first opening tag it has encountered, and can be used
449 class HTMLBreakOnClosingTagException(Exception):
453 self.tagstack = collections.deque()
454 html.parser.HTMLParser.__init__(self)
459 def __exit__(self, *_):
463 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
464 # so data remains buffered; we no longer have any interest in it, thus
465 # override this method to discard it
468 def handle_starttag(self, tag, _):
469 self.tagstack.append(tag)
471 def handle_endtag(self, tag):
472 if not self.tagstack:
473 raise compat_HTMLParseError('no tags
in the stack
')
475 inner_tag = self.tagstack.pop()
479 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
480 if not self.tagstack:
481 raise self.HTMLBreakOnClosingTagException()
484 # XXX: This should be far less strict
485 def get_element_text_and_html_by_tag(tag, html):
487 For the first element with the specified tag in the passed HTML document
488 return its' content (text
) and the whole
element (html
)
490 def find_or_raise(haystack, needle, exc):
492 return haystack.index(needle)
495 closing_tag = f'</{tag}>'
496 whole_start = find_or_raise(
497 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
498 content_start = find_or_raise(
499 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
500 content_start += whole_start + 1
501 with HTMLBreakOnClosingTagParser() as parser:
502 parser.feed(html[whole_start:content_start])
503 if not parser.tagstack or parser.tagstack[0] != tag:
504 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
505 offset = content_start
506 while offset < len(html):
507 next_closing_tag_start = find_or_raise(
508 html[offset:], closing_tag,
509 compat_HTMLParseError(f'closing {tag} tag not found'))
510 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
512 parser.feed(html[offset:offset + next_closing_tag_end])
513 offset += next_closing_tag_end
514 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
515 return html[content_start:offset + next_closing_tag_start], \
516 html[whole_start:offset + next_closing_tag_end]
517 raise compat_HTMLParseError('unexpected end of html')
520 class HTMLAttributeParser(html.parser.HTMLParser):
521 """Trivial HTML parser to gather the attributes
for a single element
"""
525 html.parser.HTMLParser.__init__(self)
527 def handle_starttag(self, tag, attrs):
528 self.attrs = dict(attrs)
529 raise compat_HTMLParseError('done')
532 class HTMLListAttrsParser(html.parser.HTMLParser):
533 """HTML parser to gather the attributes
for the elements of a
list"""
536 html.parser.HTMLParser.__init__(self)
540 def handle_starttag(self, tag, attrs):
541 if tag == 'li' and self._level == 0:
542 self.items.append(dict(attrs))
545 def handle_endtag(self, tag):
549 def extract_attributes(html_element):
550 """Given a string
for an HTML element such
as
552 a
="foo" B
="bar" c
="&98;az" d
=boz
553 empty
= noval entity
="&"
556 Decode
and return a dictionary of attributes
.
558 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
559 'empty': '', 'noval': None, 'entity': '&',
560 'sq': '"', 'dq': '\''
563 parser = HTMLAttributeParser()
564 with contextlib.suppress(compat_HTMLParseError):
565 parser.feed(html_element)
570 def parse_list(webpage):
571 """Given a string
for an series of HTML
<li
> elements
,
572 return a dictionary of their attributes
"""
573 parser = HTMLListAttrsParser()
579 def clean_html(html):
580 """Clean an HTML snippet into a readable string
"""
582 if html is None: # Convenience for sanitizing descriptions etc.
585 html = re.sub(r'\s+', ' ', html)
586 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
587 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
589 html = re.sub('<.*?>', '', html)
590 # Replace html entities
591 html = unescapeHTML(html)
595 class LenientJSONDecoder(json.JSONDecoder):
596 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
597 self.transform_source, self.ignore_extra = transform_source, ignore_extra
598 super().__init__(*args, **kwargs)
601 if self.transform_source:
602 s = self.transform_source(s)
604 if self.ignore_extra:
605 return self.raw_decode(s.lstrip())[0]
606 return super().decode(s)
607 except json.JSONDecodeError as e:
608 if e.pos is not None:
609 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
613 def sanitize_open(filename, open_mode):
614 """Try to
open the given filename
, and slightly tweak it
if this fails
.
616 Attempts to
open the given filename
. If this fails
, it tries to change
617 the filename slightly
, step by step
, until it
's either able to open it
618 or it fails and raises a final exception, like the standard open()
621 It returns the tuple (stream, definitive_file_name).
624 if sys.platform == 'win32
':
627 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
628 with contextlib.suppress(io.UnsupportedOperation):
629 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
630 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
632 for attempt in range(2):
635 if sys.platform == 'win32
':
636 # FIXME: An exclusive lock also locks the file from being read.
637 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
638 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
639 raise LockingUnsupportedError()
640 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
642 stream
= open(filename
, open_mode
)
643 return stream
, filename
644 except OSError as err
:
645 if attempt
or err
.errno
in (errno
.EACCES
,):
647 old_filename
, filename
= filename
, sanitize_path(filename
)
648 if old_filename
== filename
:
652 def timeconvert(timestr
):
653 """Convert RFC 2822 defined time string into system timestamp"""
655 timetuple
= email
.utils
.parsedate_tz(timestr
)
656 if timetuple
is not None:
657 timestamp
= email
.utils
.mktime_tz(timetuple
)
661 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
662 """Sanitizes a string so it could be used as part of a filename.
663 @param restricted Use a stricter subset of allowed characters
664 @param is_id Whether this is an ID that should be kept unchanged if possible.
665 If unset, yt-dlp's new sanitization rules are in effect
670 def replace_insane(char
):
671 if restricted
and char
in ACCENT_CHARS
:
672 return ACCENT_CHARS
[char
]
673 elif not restricted
and char
== '\n':
675 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
676 # Replace with their full-width unicode counterparts
677 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
678 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
681 return '' if restricted
else '\''
683 return '\0_\0-' if restricted
else '\0 \0-'
684 elif char
in '\\/|*<>':
686 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
690 # Replace look-alike Unicode glyphs
691 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
692 s
= unicodedata
.normalize('NFKC', s
)
693 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
694 result
= ''.join(map(replace_insane
, s
))
695 if is_id
is NO_DEFAULT
:
696 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
697 STRIP_RE
= r
'(?:\0.|[ _-])*'
698 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
699 result
= result
.replace('\0', '') or '_'
702 while '__' in result
:
703 result
= result
.replace('__', '_')
704 result
= result
.strip('_')
705 # Common case of "Foreign band name - English song title"
706 if restricted
and result
.startswith('-_'):
708 if result
.startswith('-'):
709 result
= '_' + result
[len('-'):]
710 result
= result
.lstrip('.')
716 def sanitize_path(s
, force
=False):
717 """Sanitizes and normalizes path on Windows"""
718 if sys
.platform
== 'win32':
720 drive_or_unc
, _
= os
.path
.splitdrive(s
)
726 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
730 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
731 for path_part
in norm_path
]
733 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
734 elif force
and s
and s
[0] == os
.path
.sep
:
735 sanitized_path
.insert(0, os
.path
.sep
)
736 return os
.path
.join(*sanitized_path
)
739 def sanitize_url(url
, *, scheme
='http'):
740 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
741 # the number of unwanted failures due to missing protocol
744 elif url
.startswith('//'):
745 return f
'{scheme}:{url}'
746 # Fix some common typos seen so far
748 # https://github.com/ytdl-org/youtube-dl/issues/15649
749 (r
'^httpss://', r
'https://'),
750 # https://bx1.be/lives/direct-tv/
751 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
753 for mistake
, fixup
in COMMON_TYPOS
:
754 if re
.match(mistake
, url
):
755 return re
.sub(mistake
, fixup
, url
)
759 def extract_basic_auth(url
):
760 parts
= urllib
.parse
.urlsplit(url
)
761 if parts
.username
is None:
763 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
764 parts
.hostname
if parts
.port
is None
765 else '%s:%d' % (parts
.hostname
, parts
.port
))))
766 auth_payload
= base64
.b64encode(
767 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
768 return url
, f
'Basic {auth_payload.decode()}'
771 def sanitized_Request(url
, *args
, **kwargs
):
772 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
773 if auth_header
is not None:
774 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
775 headers
['Authorization'] = auth_header
776 return urllib
.request
.Request(url
, *args
, **kwargs
)
780 """Expand shell variables and ~"""
781 return os
.path
.expandvars(compat_expanduser(s
))
784 def orderedSet(iterable
, *, lazy
=False):
785 """Remove all duplicates from the input iterable"""
787 seen
= [] # Do not use set since the items can be unhashable
793 return _iter() if lazy
else list(_iter())
796 def _htmlentity_transform(entity_with_semicolon
):
797 """Transforms an HTML entity to a character."""
798 entity
= entity_with_semicolon
[:-1]
800 # Known non-numeric HTML entity
801 if entity
in html
.entities
.name2codepoint
:
802 return chr(html
.entities
.name2codepoint
[entity
])
804 # TODO: HTML5 allows entities without a semicolon.
805 # E.g. 'Éric' should be decoded as 'Éric'.
806 if entity_with_semicolon
in html
.entities
.html5
:
807 return html
.entities
.html5
[entity_with_semicolon
]
809 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
811 numstr
= mobj
.group(1)
812 if numstr
.startswith('x'):
814 numstr
= '0%s' % numstr
817 # See https://github.com/ytdl-org/youtube-dl/issues/7518
818 with contextlib
.suppress(ValueError):
819 return chr(int(numstr
, base
))
821 # Unknown entity in name, return its literal representation
822 return '&%s;' % entity
828 assert isinstance(s
, str)
831 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
834 def escapeHTML(text
):
837 .replace('&', '&')
838 .replace('<', '<')
839 .replace('>', '>')
840 .replace('"', '"')
841 .replace("'", ''')
845 def process_communicate_or_kill(p
, *args
, **kwargs
):
846 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
847 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
848 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
851 class Popen(subprocess
.Popen
):
852 if sys
.platform
== 'win32':
853 _startupinfo
= subprocess
.STARTUPINFO()
854 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
859 def _fix_pyinstaller_ld_path(env
):
860 """Restore LD_LIBRARY_PATH when using PyInstaller
861 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
862 https://github.com/yt-dlp/yt-dlp/issues/4573
864 if not hasattr(sys
, '_MEIPASS'):
868 orig
= env
.get(f
'{key}_ORIG')
874 _fix('LD_LIBRARY_PATH') # Linux
875 _fix('DYLD_LIBRARY_PATH') # macOS
877 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
879 env
= os
.environ
.copy()
880 self
._fix
_pyinstaller
_ld
_path
(env
)
883 kwargs
['universal_newlines'] = True # For 3.6 compatibility
884 kwargs
.setdefault('encoding', 'utf-8')
885 kwargs
.setdefault('errors', 'replace')
886 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
888 def communicate_or_kill(self
, *args
, **kwargs
):
890 return self
.communicate(*args
, **kwargs
)
891 except BaseException
: # Including KeyboardInterrupt
892 self
.kill(timeout
=None)
895 def kill(self
, *, timeout
=0):
898 self
.wait(timeout
=timeout
)
901 def run(cls
, *args
, timeout
=None, **kwargs
):
902 with cls(*args
, **kwargs
) as proc
:
903 default
= '' if proc
.text_mode
else b
''
904 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
905 return stdout
or default
, stderr
or default
, proc
.returncode
908 def get_subprocess_encoding():
909 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
910 # For subprocess calls, encode with locale encoding
911 # Refer to http://stackoverflow.com/a/9951851/35070
912 encoding
= preferredencoding()
914 encoding
= sys
.getfilesystemencoding()
920 def encodeFilename(s
, for_subprocess
=False):
921 assert isinstance(s
, str)
925 def decodeFilename(b
, for_subprocess
=False):
929 def encodeArgument(s
):
930 # Legacy code that uses byte strings
931 # Uncomment the following line after fixing all post processors
932 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
933 return s
if isinstance(s
, str) else s
.decode('ascii')
936 def decodeArgument(b
):
940 def decodeOption(optval
):
943 if isinstance(optval
, bytes):
944 optval
= optval
.decode(preferredencoding())
946 assert isinstance(optval
, str)
950 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
953 def timetuple_from_msec(msec
):
954 secs
, msec
= divmod(msec
, 1000)
955 mins
, secs
= divmod(secs
, 60)
956 hrs
, mins
= divmod(mins
, 60)
957 return _timetuple(hrs
, mins
, secs
, msec
)
960 def formatSeconds(secs
, delim
=':', msec
=False):
961 time
= timetuple_from_msec(secs
* 1000)
963 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
965 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
967 ret
= '%d' % time
.seconds
968 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
971 def _ssl_load_windows_store_certs(ssl_context
, storename
):
972 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
974 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
975 if encoding
== 'x509_asn' and (
976 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
977 except PermissionError
:
980 with contextlib
.suppress(ssl
.SSLError
):
981 ssl_context
.load_verify_locations(cadata
=cert
)
984 def make_HTTPS_handler(params
, **kwargs
):
985 opts_check_certificate
= not params
.get('nocheckcertificate')
986 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
987 context
.check_hostname
= opts_check_certificate
988 if params
.get('legacyserverconnect'):
989 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
990 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
991 context
.set_ciphers('DEFAULT')
993 sys
.version_info
< (3, 10)
994 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
995 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
997 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
998 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
999 # in some situations [2][3].
1000 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
1001 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1002 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1003 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1004 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1005 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1006 # 4. https://peps.python.org/pep-0644/
1007 # 5. https://peps.python.org/pep-0644/#libressl-support
1008 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1009 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1010 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1012 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1013 if opts_check_certificate
:
1014 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1015 context
.load_verify_locations(cafile
=certifi
.where())
1018 context
.load_default_certs()
1019 # Work around the issue in load_default_certs when there are bad certificates. See:
1020 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1021 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1022 except ssl
.SSLError
:
1023 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1024 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1025 for storename
in ('CA', 'ROOT'):
1026 _ssl_load_windows_store_certs(context
, storename
)
1027 context
.set_default_verify_paths()
1029 client_certfile
= params
.get('client_certificate')
1032 context
.load_cert_chain(
1033 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1034 password
=params
.get('client_certificate_password'))
1035 except ssl
.SSLError
:
1036 raise YoutubeDLError('Unable to load client certificate')
1038 # Some servers may reject requests if ALPN extension is not sent. See:
1039 # https://github.com/python/cpython/issues/85140
1040 # https://github.com/yt-dlp/yt-dlp/issues/3878
1041 with contextlib
.suppress(NotImplementedError):
1042 context
.set_alpn_protocols(['http/1.1'])
1044 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1047 def bug_reports_message(before
=';'):
1048 from .update
import REPOSITORY
1050 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1051 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1053 before
= before
.rstrip()
1054 if not before
or before
.endswith(('.', '!', '?')):
1055 msg
= msg
[0].title() + msg
[1:]
1057 return (before
+ ' ' if before
else '') + msg
1060 class YoutubeDLError(Exception):
1061 """Base exception for YoutubeDL errors."""
1064 def __init__(self
, msg
=None):
1067 elif self
.msg
is None:
1068 self
.msg
= type(self
).__name
__
1069 super().__init
__(self
.msg
)
1072 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1073 if hasattr(ssl
, 'CertificateError'):
1074 network_exceptions
.append(ssl
.CertificateError
)
1075 network_exceptions
= tuple(network_exceptions
)
1078 class ExtractorError(YoutubeDLError
):
1079 """Error during info extraction."""
1081 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1082 """ tb, if given, is the original traceback (so that it can be printed out).
1083 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1085 if sys
.exc_info()[0] in network_exceptions
:
1088 self
.orig_msg
= str(msg
)
1090 self
.expected
= expected
1092 self
.video_id
= video_id
1094 self
.exc_info
= sys
.exc_info() # preserve original exception
1095 if isinstance(self
.exc_info
[1], ExtractorError
):
1096 self
.exc_info
= self
.exc_info
[1].exc_info
1097 super().__init
__(self
.__msg
)
1102 format_field(self
.ie
, None, '[%s] '),
1103 format_field(self
.video_id
, None, '%s: '),
1105 format_field(self
.cause
, None, ' (caused by %r)'),
1106 '' if self
.expected
else bug_reports_message()))
1108 def format_traceback(self
):
1109 return join_nonempty(
1110 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1111 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1114 def __setattr__(self
, name
, value
):
1115 super().__setattr
__(name
, value
)
1116 if getattr(self
, 'msg', None) and name
not in ('msg', 'args'):
1117 self
.msg
= self
.__msg
or type(self
).__name
__
1118 self
.args
= (self
.msg
, ) # Cannot be property
1121 class UnsupportedError(ExtractorError
):
1122 def __init__(self
, url
):
1124 'Unsupported URL: %s' % url
, expected
=True)
1128 class RegexNotFoundError(ExtractorError
):
1129 """Error when a regex didn't match"""
1133 class GeoRestrictedError(ExtractorError
):
1134 """Geographic restriction Error exception.
1136 This exception may be thrown when a video is not available from your
1137 geographic location due to geographic restrictions imposed by a website.
1140 def __init__(self
, msg
, countries
=None, **kwargs
):
1141 kwargs
['expected'] = True
1142 super().__init
__(msg
, **kwargs
)
1143 self
.countries
= countries
1146 class UserNotLive(ExtractorError
):
1147 """Error when a channel/user is not live"""
1149 def __init__(self
, msg
=None, **kwargs
):
1150 kwargs
['expected'] = True
1151 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1154 class DownloadError(YoutubeDLError
):
1155 """Download Error exception.
1157 This exception may be thrown by FileDownloader objects if they are not
1158 configured to continue on errors. They will contain the appropriate
1162 def __init__(self
, msg
, exc_info
=None):
1163 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1164 super().__init
__(msg
)
1165 self
.exc_info
= exc_info
1168 class EntryNotInPlaylist(YoutubeDLError
):
1169 """Entry not in playlist exception.
1171 This exception will be thrown by YoutubeDL when a requested entry
1172 is not found in the playlist info_dict
1174 msg
= 'Entry not found in info'
1177 class SameFileError(YoutubeDLError
):
1178 """Same File exception.
1180 This exception will be thrown by FileDownloader objects if they detect
1181 multiple files would have to be downloaded to the same file on disk.
1183 msg
= 'Fixed output name but more than one file to download'
1185 def __init__(self
, filename
=None):
1186 if filename
is not None:
1187 self
.msg
+= f
': {filename}'
1188 super().__init
__(self
.msg
)
1191 class PostProcessingError(YoutubeDLError
):
1192 """Post Processing exception.
1194 This exception may be raised by PostProcessor's .run() method to
1195 indicate an error in the postprocessing task.
1199 class DownloadCancelled(YoutubeDLError
):
1200 """ Exception raised when the download queue should be interrupted """
1201 msg
= 'The download was cancelled'
1204 class ExistingVideoReached(DownloadCancelled
):
1205 """ --break-on-existing triggered """
1206 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1209 class RejectedVideoReached(DownloadCancelled
):
1210 """ --break-on-reject triggered """
1211 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1214 class MaxDownloadsReached(DownloadCancelled
):
1215 """ --max-downloads limit has been reached. """
1216 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1219 class ReExtractInfo(YoutubeDLError
):
1220 """ Video info needs to be re-extracted. """
1222 def __init__(self
, msg
, expected
=False):
1223 super().__init
__(msg
)
1224 self
.expected
= expected
1227 class ThrottledDownload(ReExtractInfo
):
1228 """ Download speed below --throttled-rate. """
1229 msg
= 'The download speed is below throttle limit'
1232 super().__init
__(self
.msg
, expected
=False)
1235 class UnavailableVideoError(YoutubeDLError
):
1236 """Unavailable Format exception.
1238 This exception will be thrown when a video is requested
1239 in a format that is not available for that video.
1241 msg
= 'Unable to download video'
1243 def __init__(self
, err
=None):
1245 self
.msg
+= f
': {err}'
1246 super().__init
__(self
.msg
)
1249 class ContentTooShortError(YoutubeDLError
):
1250 """Content Too Short exception.
1252 This exception may be raised by FileDownloader objects when a file they
1253 download is too small for what the server announced first, indicating
1254 the connection was probably interrupted.
1257 def __init__(self
, downloaded
, expected
):
1258 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1260 self
.downloaded
= downloaded
1261 self
.expected
= expected
1264 class XAttrMetadataError(YoutubeDLError
):
1265 def __init__(self
, code
=None, msg
='Unknown error'):
1266 super().__init
__(msg
)
1270 # Parsing code and msg
1271 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1272 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1273 self
.reason
= 'NO_SPACE'
1274 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1275 self
.reason
= 'VALUE_TOO_LONG'
1277 self
.reason
= 'NOT_SUPPORTED'
1280 class XAttrUnavailableError(YoutubeDLError
):
1284 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1285 hc
= http_class(*args
, **kwargs
)
1286 source_address
= ydl_handler
._params
.get('source_address')
1288 if source_address
is not None:
1289 # This is to workaround _create_connection() from socket where it will try all
1290 # address data from getaddrinfo() including IPv6. This filters the result from
1291 # getaddrinfo() based on the source_address value.
1292 # This is based on the cpython socket.create_connection() function.
1293 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1294 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1295 host
, port
= address
1297 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1298 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1299 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1300 if addrs
and not ip_addrs
:
1301 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1303 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1304 % (ip_version
, source_address
[0]))
1305 for res
in ip_addrs
:
1306 af
, socktype
, proto
, canonname
, sa
= res
1309 sock
= socket
.socket(af
, socktype
, proto
)
1310 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1311 sock
.settimeout(timeout
)
1312 sock
.bind(source_address
)
1314 err
= None # Explicitly break reference cycle
1316 except OSError as _
:
1318 if sock
is not None:
1323 raise OSError('getaddrinfo returns an empty list')
1324 if hasattr(hc
, '_create_connection'):
1325 hc
._create
_connection
= _create_connection
1326 hc
.source_address
= (source_address
, 0)
1331 def handle_youtubedl_headers(headers
):
1332 filtered_headers
= headers
1334 if 'Youtubedl-no-compression' in filtered_headers
:
1335 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1336 del filtered_headers
['Youtubedl-no-compression']
1338 return filtered_headers
1341 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1342 """Handler for HTTP requests and responses.
1344 This class, when installed with an OpenerDirector, automatically adds
1345 the standard headers to every HTTP request and handles gzipped and
1346 deflated responses from web servers. If compression is to be avoided in
1347 a particular request, the original request in the program code only has
1348 to include the HTTP header "Youtubedl-no-compression", which will be
1349 removed before making the real request.
1351 Part of this code was copied from:
1353 http://techknack.net/python-urllib2-handlers/
1355 Andrew Rowls, the author of that code, agreed to release it to the
1359 def __init__(self
, params
, *args
, **kwargs
):
1360 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1361 self
._params
= params
1363 def http_open(self
, req
):
1364 conn_class
= http
.client
.HTTPConnection
1366 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1368 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1369 del req
.headers
['Ytdl-socks-proxy']
1371 return self
.do_open(functools
.partial(
1372 _create_http_connection
, self
, conn_class
, False),
1380 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1382 return zlib
.decompress(data
)
1388 return brotli
.decompress(data
)
1390 def http_request(self
, req
):
1391 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1392 # always respected by websites, some tend to give out URLs with non percent-encoded
1393 # non-ASCII characters (see telemb.py, ard.py [#3412])
1394 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1395 # To work around aforementioned issue we will replace request's original URL with
1396 # percent-encoded one
1397 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1398 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1399 url
= req
.get_full_url()
1400 url_escaped
= escape_url(url
)
1402 # Substitute URL if any change after escaping
1403 if url
!= url_escaped
:
1404 req
= update_Request(req
, url
=url_escaped
)
1406 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1407 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1408 # The dict keys are capitalized because of this bug by urllib
1409 if h
.capitalize() not in req
.headers
:
1410 req
.add_header(h
, v
)
1412 if 'Accept-encoding' not in req
.headers
:
1413 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1415 req
.headers
= handle_youtubedl_headers(req
.headers
)
1417 return super().do_request_(req
)
1419 def http_response(self
, req
, resp
):
1422 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1423 content
= resp
.read()
1424 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1426 uncompressed
= io
.BytesIO(gz
.read())
1427 except OSError as original_ioerror
:
1428 # There may be junk add the end of the file
1429 # See http://stackoverflow.com/q/4928560/35070 for details
1430 for i
in range(1, 1024):
1432 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1433 uncompressed
= io
.BytesIO(gz
.read())
1438 raise original_ioerror
1439 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1440 resp
.msg
= old_resp
.msg
1441 del resp
.headers
['Content-encoding']
1443 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1444 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1445 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1446 resp
.msg
= old_resp
.msg
1447 del resp
.headers
['Content-encoding']
1449 if resp
.headers
.get('Content-encoding', '') == 'br':
1450 resp
= urllib
.request
.addinfourl(
1451 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1452 resp
.msg
= old_resp
.msg
1453 del resp
.headers
['Content-encoding']
1454 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1455 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1456 if 300 <= resp
.code
< 400:
1457 location
= resp
.headers
.get('Location')
1459 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1460 location
= location
.encode('iso-8859-1').decode()
1461 location_escaped
= escape_url(location
)
1462 if location
!= location_escaped
:
1463 del resp
.headers
['Location']
1464 resp
.headers
['Location'] = location_escaped
1467 https_request
= http_request
1468 https_response
= http_response
1471 def make_socks_conn_class(base_class
, socks_proxy
):
1472 assert issubclass(base_class
, (
1473 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1475 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1476 if url_components
.scheme
.lower() == 'socks5':
1477 socks_type
= ProxyType
.SOCKS5
1478 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1479 socks_type
= ProxyType
.SOCKS4
1480 elif url_components
.scheme
.lower() == 'socks4a':
1481 socks_type
= ProxyType
.SOCKS4A
1483 def unquote_if_non_empty(s
):
1486 return urllib
.parse
.unquote_plus(s
)
1490 url_components
.hostname
, url_components
.port
or 1080,
1492 unquote_if_non_empty(url_components
.username
),
1493 unquote_if_non_empty(url_components
.password
),
1496 class SocksConnection(base_class
):
1498 self
.sock
= sockssocket()
1499 self
.sock
.setproxy(*proxy_args
)
1500 if isinstance(self
.timeout
, (int, float)):
1501 self
.sock
.settimeout(self
.timeout
)
1502 self
.sock
.connect((self
.host
, self
.port
))
1504 if isinstance(self
, http
.client
.HTTPSConnection
):
1505 if hasattr(self
, '_context'): # Python > 2.6
1506 self
.sock
= self
._context
.wrap_socket(
1507 self
.sock
, server_hostname
=self
.host
)
1509 self
.sock
= ssl
.wrap_socket(self
.sock
)
1511 return SocksConnection
1514 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1515 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1516 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1517 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1518 self
._params
= params
1520 def https_open(self
, req
):
1522 conn_class
= self
._https
_conn
_class
1524 if hasattr(self
, '_context'): # python > 2.6
1525 kwargs
['context'] = self
._context
1526 if hasattr(self
, '_check_hostname'): # python 3.x
1527 kwargs
['check_hostname'] = self
._check
_hostname
1529 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1531 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1532 del req
.headers
['Ytdl-socks-proxy']
1535 return self
.do_open(
1536 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1537 except urllib
.error
.URLError
as e
:
1538 if (isinstance(e
.reason
, ssl
.SSLError
)
1539 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1540 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1544 def is_path_like(f
):
1545 return isinstance(f
, (str, bytes, os
.PathLike
))
1548 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1550 See [1] for cookie file format.
1552 1. https://curl.haxx.se/docs/http-cookies.html
1554 _HTTPONLY_PREFIX
= '#HttpOnly_'
1556 _HEADER
= '''# Netscape HTTP Cookie File
1557 # This file is generated by yt-dlp. Do not edit.
1560 _CookieFileEntry
= collections
.namedtuple(
1562 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1564 def __init__(self
, filename
=None, *args
, **kwargs
):
1565 super().__init
__(None, *args
, **kwargs
)
1566 if is_path_like(filename
):
1567 filename
= os
.fspath(filename
)
1568 self
.filename
= filename
1571 def _true_or_false(cndn
):
1572 return 'TRUE' if cndn
else 'FALSE'
1574 @contextlib.contextmanager
1575 def open(self
, file, *, write
=False):
1576 if is_path_like(file):
1577 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1584 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1587 if (not ignore_discard
and cookie
.discard
1588 or not ignore_expires
and cookie
.is_expired(now
)):
1590 name
, value
= cookie
.name
, cookie
.value
1592 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1593 # with no name, whereas http.cookiejar regards it as a
1594 # cookie with no value.
1595 name
, value
= '', name
1596 f
.write('%s\n' % '\t'.join((
1598 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1600 self
._true
_or
_false
(cookie
.secure
),
1601 str_or_none(cookie
.expires
, default
=''),
1605 def save(self
, filename
=None, *args
, **kwargs
):
1607 Save cookies to a file.
1608 Code is taken from CPython 3.6
1609 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1611 if filename
is None:
1612 if self
.filename
is not None:
1613 filename
= self
.filename
1615 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1617 # Store session cookies with `expires` set to 0 instead of an empty string
1619 if cookie
.expires
is None:
1622 with self
.open(filename
, write
=True) as f
:
1623 f
.write(self
._HEADER
)
1624 self
._really
_save
(f
, *args
, **kwargs
)
1626 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1627 """Load cookies from a file."""
1628 if filename
is None:
1629 if self
.filename
is not None:
1630 filename
= self
.filename
1632 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1634 def prepare_line(line
):
1635 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1636 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1637 # comments and empty lines are fine
1638 if line
.startswith('#') or not line
.strip():
1640 cookie_list
= line
.split('\t')
1641 if len(cookie_list
) != self
._ENTRY
_LEN
:
1642 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1643 cookie
= self
._CookieFileEntry
(*cookie_list
)
1644 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1645 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1649 with self
.open(filename
) as f
:
1652 cf
.write(prepare_line(line
))
1653 except http
.cookiejar
.LoadError
as e
:
1654 if f
'{line.strip()} '[0] in '[{"':
1655 raise http
.cookiejar
.LoadError(
1656 'Cookies file must be Netscape formatted, not JSON. See '
1657 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1658 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1661 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1662 # Session cookies are denoted by either `expires` field set to
1663 # an empty string or 0. MozillaCookieJar only recognizes the former
1664 # (see [1]). So we need force the latter to be recognized as session
1665 # cookies on our own.
1666 # Session cookies may be important for cookies-based authentication,
1667 # e.g. usually, when user does not check 'Remember me' check box while
1668 # logging in on a site, some important cookies are stored as session
1669 # cookies so that not recognizing them will result in failed login.
1670 # 1. https://bugs.python.org/issue17164
1672 # Treat `expires=0` cookies as session cookies
1673 if cookie
.expires
== 0:
1674 cookie
.expires
= None
1675 cookie
.discard
= True
1678 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1679 def __init__(self
, cookiejar
=None):
1680 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1682 def http_response(self
, request
, response
):
1683 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1685 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1686 https_response
= http_response
1689 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1690 """YoutubeDL redirect handler
1692 The code is based on HTTPRedirectHandler implementation from CPython [1].
1694 This redirect handler solves two issues:
1695 - ensures redirect URL is always unicode under python 2
1696 - introduces support for experimental HTTP response status code
1697 308 Permanent Redirect [2] used by some sites [3]
1699 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1700 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1701 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1704 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1706 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1707 """Return a Request or None in response to a redirect.
1709 This is called by the http_error_30x methods when a
1710 redirection response is received. If a redirection should
1711 take place, return a new Request to allow http_error_30x to
1712 perform the redirect. Otherwise, raise HTTPError if no-one
1713 else should try to handle this url. Return None if you can't
1714 but another Handler might.
1716 m
= req
.get_method()
1717 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1718 or code
in (301, 302, 303) and m
== "POST")):
1719 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1720 # Strictly (according to RFC 2616), 301 or 302 in response to
1721 # a POST MUST NOT cause a redirection without confirmation
1722 # from the user (of urllib.request, in this case). In practice,
1723 # essentially all clients do redirect in this case, so we do
1726 # Be conciliant with URIs containing a space. This is mainly
1727 # redundant with the more complete encoding done in http_error_302(),
1728 # but it is kept for compatibility with other callers.
1729 newurl
= newurl
.replace(' ', '%20')
1731 CONTENT_HEADERS
= ("content-length", "content-type")
1732 # NB: don't use dict comprehension for python 2.6 compatibility
1733 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1735 # A 303 must either use GET or HEAD for subsequent request
1736 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1737 if code
== 303 and m
!= 'HEAD':
1739 # 301 and 302 redirects are commonly turned into a GET from a POST
1740 # for subsequent requests by browsers, so we'll do the same.
1741 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1742 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1743 if code
in (301, 302) and m
== 'POST':
1746 return urllib
.request
.Request(
1747 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1748 unverifiable
=True, method
=m
)
1751 def extract_timezone(date_str
):
1754 ^.{8,}? # >=8 char non-TZ prefix, if present
1755 (?P<tz>Z| # just the UTC Z, or
1756 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1757 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1758 [ ]? # optional space
1759 (?P<sign>\+|-) # +/-
1760 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1764 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1765 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1766 if timezone
is not None:
1767 date_str
= date_str
[:-len(m
.group('tz'))]
1768 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1770 date_str
= date_str
[:-len(m
.group('tz'))]
1771 if not m
.group('sign'):
1772 timezone
= datetime
.timedelta()
1774 sign
= 1 if m
.group('sign') == '+' else -1
1775 timezone
= datetime
.timedelta(
1776 hours
=sign
* int(m
.group('hours')),
1777 minutes
=sign
* int(m
.group('minutes')))
1778 return timezone
, date_str
1781 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1782 """ Return a UNIX timestamp from the given date """
1784 if date_str
is None:
1787 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1789 if timezone
is None:
1790 timezone
, date_str
= extract_timezone(date_str
)
1792 with contextlib
.suppress(ValueError):
1793 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1794 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1795 return calendar
.timegm(dt
.timetuple())
1798 def date_formats(day_first
=True):
1799 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1802 def unified_strdate(date_str
, day_first
=True):
1803 """Return a string with the date in the format YYYYMMDD"""
1805 if date_str
is None:
1809 date_str
= date_str
.replace(',', ' ')
1810 # Remove AM/PM + timezone
1811 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1812 _
, date_str
= extract_timezone(date_str
)
1814 for expression
in date_formats(day_first
):
1815 with contextlib
.suppress(ValueError):
1816 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1817 if upload_date
is None:
1818 timetuple
= email
.utils
.parsedate_tz(date_str
)
1820 with contextlib
.suppress(ValueError):
1821 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1822 if upload_date
is not None:
1823 return str(upload_date
)
1826 def unified_timestamp(date_str
, day_first
=True):
1827 if date_str
is None:
1830 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1831 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1833 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1834 timezone
, date_str
= extract_timezone(date_str
)
1836 # Remove AM/PM + timezone
1837 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1839 # Remove unrecognized timezones from ISO 8601 alike timestamps
1840 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1842 date_str
= date_str
[:-len(m
.group('tz'))]
1844 # Python only supports microseconds, so remove nanoseconds
1845 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1847 date_str
= m
.group(1)
1849 for expression
in date_formats(day_first
):
1850 with contextlib
.suppress(ValueError):
1851 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1852 return calendar
.timegm(dt
.timetuple())
1854 timetuple
= email
.utils
.parsedate_tz(date_str
)
1856 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1859 def determine_ext(url
, default_ext
='unknown_video'):
1860 if url
is None or '.' not in url
:
1862 guess
= url
.partition('?')[0].rpartition('.')[2]
1863 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1865 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1866 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1867 return guess
.rstrip('/')
1872 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1873 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1876 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1878 Return a datetime object from a string.
1880 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1882 @param format strftime format of DATE
1883 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1884 auto: round to the unit provided in date_str (if applicable).
1886 auto_precision
= False
1887 if precision
== 'auto':
1888 auto_precision
= True
1889 precision
= 'microsecond'
1890 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1891 if date_str
in ('now', 'today'):
1893 if date_str
== 'yesterday':
1894 return today
- datetime
.timedelta(days
=1)
1896 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1898 if match
is not None:
1899 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1900 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1901 unit
= match
.group('unit')
1902 if unit
== 'month' or unit
== 'year':
1903 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1909 delta
= datetime
.timedelta(**{unit + 's': time}
)
1910 new_date
= start_time
+ delta
1912 return datetime_round(new_date
, unit
)
1915 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1918 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1920 Return a date object from a string using datetime_from_str
1922 @param strict Restrict allowed patterns to "YYYYMMDD" and
1923 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1925 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1926 raise ValueError(f
'Invalid date format "{date_str}"')
1927 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1930 def datetime_add_months(dt
, months
):
1931 """Increment/Decrement a datetime object by months."""
1932 month
= dt
.month
+ months
- 1
1933 year
= dt
.year
+ month
// 12
1934 month
= month
% 12 + 1
1935 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1936 return dt
.replace(year
, month
, day
)
1939 def datetime_round(dt
, precision
='day'):
1941 Round a datetime object's time to a specific precision
1943 if precision
== 'microsecond':
1952 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1953 timestamp
= calendar
.timegm(dt
.timetuple())
1954 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1957 def hyphenate_date(date_str
):
1959 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1960 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1961 if match
is not None:
1962 return '-'.join(match
.groups())
1968 """Represents a time interval between two dates"""
1970 def __init__(self
, start
=None, end
=None):
1971 """start and end must be strings in the format accepted by date"""
1972 if start
is not None:
1973 self
.start
= date_from_str(start
, strict
=True)
1975 self
.start
= datetime
.datetime
.min.date()
1977 self
.end
= date_from_str(end
, strict
=True)
1979 self
.end
= datetime
.datetime
.max.date()
1980 if self
.start
> self
.end
:
1981 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1985 """Returns a range that only contains the given day"""
1986 return cls(day
, day
)
1988 def __contains__(self
, date
):
1989 """Check if the date is in the range"""
1990 if not isinstance(date
, datetime
.date
):
1991 date
= date_from_str(date
)
1992 return self
.start
<= date
<= self
.end
1995 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1997 def __eq__(self
, other
):
1998 return (isinstance(other
, DateRange
)
1999 and self
.start
== other
.start
and self
.end
== other
.end
)
2002 def platform_name():
2003 """ Returns the platform name as a str """
2004 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
2005 return platform
.platform()
2009 def system_identifier():
2010 python_implementation
= platform
.python_implementation()
2011 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2012 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2014 with contextlib
.suppress(OSError): # We may not have access to the executable
2015 libc_ver
= platform
.libc_ver()
2017 return 'Python %s (%s %s %s) - %s (%s%s)' % (
2018 platform
.python_version(),
2019 python_implementation
,
2021 platform
.architecture()[0],
2022 platform
.platform(),
2023 ssl
.OPENSSL_VERSION
,
2024 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2029 def get_windows_version():
2030 ''' Get Windows version. returns () if it's not running on Windows '''
2031 if compat_os_name
== 'nt':
2032 return version_tuple(platform
.win32_ver()[1])
2037 def write_string(s
, out
=None, encoding
=None):
2038 assert isinstance(s
, str)
2039 out
= out
or sys
.stderr
2041 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2042 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2044 enc
, buffer = None, out
2045 if 'b' in getattr(out
, 'mode', ''):
2046 enc
= encoding
or preferredencoding()
2047 elif hasattr(out
, 'buffer'):
2049 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2051 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2055 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2056 from . import _IN_CLI
2058 if msg
in deprecation_warning
._cache
:
2060 deprecation_warning
._cache
.add(msg
)
2062 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2063 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2066 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2069 deprecation_warning
._cache
= set()
2072 def bytes_to_intlist(bs
):
2075 if isinstance(bs
[0], int): # Python 3
2078 return [ord(c
) for c
in bs
]
2081 def intlist_to_bytes(xs
):
2084 return struct
.pack('%dB' % len(xs
), *xs
)
2087 class LockingUnsupportedError(OSError):
2088 msg
= 'File locking is not supported'
2091 super().__init
__(self
.msg
)
2094 # Cross-platform file locking
2095 if sys
.platform
== 'win32':
2097 import ctypes
.wintypes
2100 class OVERLAPPED(ctypes
.Structure
):
2102 ('Internal', ctypes
.wintypes
.LPVOID
),
2103 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2104 ('Offset', ctypes
.wintypes
.DWORD
),
2105 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2106 ('hEvent', ctypes
.wintypes
.HANDLE
),
2109 kernel32
= ctypes
.windll
.kernel32
2110 LockFileEx
= kernel32
.LockFileEx
2111 LockFileEx
.argtypes
= [
2112 ctypes
.wintypes
.HANDLE
, # hFile
2113 ctypes
.wintypes
.DWORD
, # dwFlags
2114 ctypes
.wintypes
.DWORD
, # dwReserved
2115 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2116 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2117 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2119 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2120 UnlockFileEx
= kernel32
.UnlockFileEx
2121 UnlockFileEx
.argtypes
= [
2122 ctypes
.wintypes
.HANDLE
, # hFile
2123 ctypes
.wintypes
.DWORD
, # dwReserved
2124 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2125 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2126 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2128 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2129 whole_low
= 0xffffffff
2130 whole_high
= 0x7fffffff
2132 def _lock_file(f
, exclusive
, block
):
2133 overlapped
= OVERLAPPED()
2134 overlapped
.Offset
= 0
2135 overlapped
.OffsetHigh
= 0
2136 overlapped
.hEvent
= 0
2137 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2139 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2140 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2141 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2142 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2143 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2145 def _unlock_file(f
):
2146 assert f
._lock
_file
_overlapped
_p
2147 handle
= msvcrt
.get_osfhandle(f
.fileno())
2148 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2149 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2155 def _lock_file(f
, exclusive
, block
):
2156 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2158 flags |
= fcntl
.LOCK_NB
2160 fcntl
.flock(f
, flags
)
2161 except BlockingIOError
:
2163 except OSError: # AOSP does not have flock()
2164 fcntl
.lockf(f
, flags
)
2166 def _unlock_file(f
):
2168 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2170 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2174 def _lock_file(f
, exclusive
, block
):
2175 raise LockingUnsupportedError()
2177 def _unlock_file(f
):
2178 raise LockingUnsupportedError()
2184 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2185 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2186 raise NotImplementedError(mode
)
2187 self
.mode
, self
.block
= mode
, block
2189 writable
= any(f
in mode
for f
in 'wax+')
2190 readable
= any(f
in mode
for f
in 'r+')
2191 flags
= functools
.reduce(operator
.ior
, (
2192 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2193 getattr(os
, 'O_BINARY', 0), # Windows only
2194 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2195 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2196 os
.O_APPEND
if 'a' in mode
else 0,
2197 os
.O_EXCL
if 'x' in mode
else 0,
2198 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2201 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2203 def __enter__(self
):
2204 exclusive
= 'r' not in self
.mode
2206 _lock_file(self
.f
, exclusive
, self
.block
)
2211 if 'w' in self
.mode
:
2214 except OSError as e
:
2216 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2217 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2226 _unlock_file(self
.f
)
2230 def __exit__(self
, *_
):
2239 def __getattr__(self
, attr
):
2240 return getattr(self
.f
, attr
)
2247 def get_filesystem_encoding():
2248 encoding
= sys
.getfilesystemencoding()
2249 return encoding
if encoding
is not None else 'utf-8'
2252 def shell_quote(args
):
2254 encoding
= get_filesystem_encoding()
2256 if isinstance(a
, bytes):
2257 # We may get a filename encoded with 'encodeFilename'
2258 a
= a
.decode(encoding
)
2259 quoted_args
.append(compat_shlex_quote(a
))
2260 return ' '.join(quoted_args
)
2263 def smuggle_url(url
, data
):
2264 """ Pass additional data in a URL for internal use. """
2266 url
, idata
= unsmuggle_url(url
, {})
2268 sdata
= urllib
.parse
.urlencode(
2269 {'__youtubedl_smuggle': json.dumps(data)}
)
2270 return url
+ '#' + sdata
2273 def unsmuggle_url(smug_url
, default
=None):
2274 if '#__youtubedl_smuggle' not in smug_url
:
2275 return smug_url
, default
2276 url
, _
, sdata
= smug_url
.rpartition('#')
2277 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2278 data
= json
.loads(jsond
)
2282 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2283 """ Formats numbers with decimal sufixes like K, M, etc """
2284 num
, factor
= float_or_none(num
), float(factor
)
2285 if num
is None or num
< 0:
2287 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2288 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2289 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2291 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2292 converted
= num
/ (factor
** exponent
)
2293 return fmt
% (converted
, suffix
)
2296 def format_bytes(bytes):
2297 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2300 def lookup_unit_table(unit_table
, s
, strict
=False):
2301 num_re
= NUMBER_RE
if strict
else NUMBER_RE
.replace(R
'\.', '[,.]')
2302 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2303 m
= (re
.fullmatch
if strict
else re
.match
)(
2304 rf
'(?P<num>{num_re})\s*(?P<unit>{units_re})\b', s
)
2308 num
= float(m
.group('num').replace(',', '.'))
2309 mult
= unit_table
[m
.group('unit')]
2310 return round(num
* mult
)
2314 """Parse a string indicating a byte quantity into an integer"""
2315 return lookup_unit_table(
2316 {u: 1024**i for i, u in enumerate(['', *'KMGTPEZY'])}
,
2317 s
.upper(), strict
=True)
2320 def parse_filesize(s
):
2324 # The lower-case forms are of course incorrect and unofficial,
2325 # but we support those too
2342 'megabytes': 1000 ** 2,
2343 'mebibytes': 1024 ** 2,
2349 'gigabytes': 1000 ** 3,
2350 'gibibytes': 1024 ** 3,
2356 'terabytes': 1000 ** 4,
2357 'tebibytes': 1024 ** 4,
2363 'petabytes': 1000 ** 5,
2364 'pebibytes': 1024 ** 5,
2370 'exabytes': 1000 ** 6,
2371 'exbibytes': 1024 ** 6,
2377 'zettabytes': 1000 ** 7,
2378 'zebibytes': 1024 ** 7,
2384 'yottabytes': 1000 ** 8,
2385 'yobibytes': 1024 ** 8,
2388 return lookup_unit_table(_UNIT_TABLE
, s
)
2395 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2397 if re
.match(r
'^[\d,.]+$', s
):
2398 return str_to_int(s
)
2411 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2415 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2417 return str_to_int(mobj
.group(1))
2420 def parse_resolution(s
, *, lenient
=False):
2425 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2427 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2430 'width': int(mobj
.group('w')),
2431 'height': int(mobj
.group('h')),
2434 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2436 return {'height': int(mobj.group(1))}
2438 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2440 return {'height': int(mobj.group(1)) * 540}
2445 def parse_bitrate(s
):
2446 if not isinstance(s
, str):
2448 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2450 return int(mobj
.group(1))
2453 def month_by_name(name
, lang
='en'):
2454 """ Return the number of a month by (locale-independently) English name """
2456 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2459 return month_names
.index(name
) + 1
2464 def month_by_abbreviation(abbrev
):
2465 """ Return the number of a month by (locale-independently) English
2469 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2474 def fix_xml_ampersands(xml_str
):
2475 """Replace all the '&' by '&' in XML"""
2477 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2482 def setproctitle(title
):
2483 assert isinstance(title
, str)
2485 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2492 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2496 # LoadLibrary in Windows Python 2.7.13 only expects
2497 # a bytestring, but since unicode_literals turns
2498 # every string into a unicode string, it fails.
2500 title_bytes
= title
.encode()
2501 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2502 buf
.value
= title_bytes
2504 libc
.prctl(15, buf
, 0, 0, 0)
2505 except AttributeError:
2506 return # Strange libc, just skip this
2509 def remove_start(s
, start
):
2510 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2513 def remove_end(s
, end
):
2514 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2517 def remove_quotes(s
):
2518 if s
is None or len(s
) < 2:
2520 for quote
in ('"', "'", ):
2521 if s
[0] == quote
and s
[-1] == quote
:
2526 def get_domain(url
):
2528 This implementation is inconsistent, but is kept for compatibility.
2529 Use this only for "webpage_url_domain"
2531 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2534 def url_basename(url
):
2535 path
= urllib
.parse
.urlparse(url
).path
2536 return path
.strip('/').split('/')[-1]
2540 return re
.match(r
'https?://[^?#]+/', url
).group()
2543 def urljoin(base
, path
):
2544 if isinstance(path
, bytes):
2545 path
= path
.decode()
2546 if not isinstance(path
, str) or not path
:
2548 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2550 if isinstance(base
, bytes):
2551 base
= base
.decode()
2552 if not isinstance(base
, str) or not re
.match(
2553 r
'^(?:https?:)?//', base
):
2555 return urllib
.parse
.urljoin(base
, path
)
2558 class HEADRequest(urllib
.request
.Request
):
2559 def get_method(self
):
2563 class PUTRequest(urllib
.request
.Request
):
2564 def get_method(self
):
2568 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2569 if get_attr
and v
is not None:
2570 v
= getattr(v
, get_attr
, None)
2572 return int(v
) * invscale
// scale
2573 except (ValueError, TypeError, OverflowError):
2577 def str_or_none(v
, default
=None):
2578 return default
if v
is None else str(v
)
2581 def str_to_int(int_str
):
2582 """ A more relaxed version of int_or_none """
2583 if isinstance(int_str
, int):
2585 elif isinstance(int_str
, str):
2586 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2587 return int_or_none(int_str
)
2590 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2594 return float(v
) * invscale
/ scale
2595 except (ValueError, TypeError):
2599 def bool_or_none(v
, default
=None):
2600 return v
if isinstance(v
, bool) else default
2603 def strip_or_none(v
, default
=None):
2604 return v
.strip() if isinstance(v
, str) else default
2607 def url_or_none(url
):
2608 if not url
or not isinstance(url
, str):
2611 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2614 def request_to_url(req
):
2615 if isinstance(req
, urllib
.request
.Request
):
2616 return req
.get_full_url()
2621 def strftime_or_none(timestamp
, date_format
, default
=None):
2622 datetime_object
= None
2624 if isinstance(timestamp
, (int, float)): # unix timestamp
2625 # Using naive datetime here can break timestamp() in Windows
2626 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2627 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2628 elif isinstance(timestamp
, str): # assume YYYYMMDD
2629 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2630 date_format
= re
.sub( # Support %s on windows
2631 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2632 return datetime_object
.strftime(date_format
)
2633 except (ValueError, TypeError, AttributeError):
2637 def parse_duration(s
):
2638 if not isinstance(s
, str):
2644 days
, hours
, mins
, secs
, ms
= [None] * 5
2645 m
= re
.match(r
'''(?x)
2647 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2648 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2649 (?P<ms>[.:][0-9]+)?Z?$
2652 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2657 [0-9]+\s*y(?:ears?)?,?\s*
2660 [0-9]+\s*m(?:onths?)?,?\s*
2663 [0-9]+\s*w(?:eeks?)?,?\s*
2666 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2670 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2673 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2676 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2679 days
, hours
, mins
, secs
, ms
= m
.groups()
2681 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2683 hours
, mins
= m
.groups()
2688 ms
= ms
.replace(':', '.')
2689 return sum(float(part
or 0) * mult
for part
, mult
in (
2690 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2693 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2694 name
, real_ext
= os
.path
.splitext(filename
)
2696 f
'{name}.{ext}{real_ext}'
2697 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2698 else f
'{filename}.{ext}')
2701 def replace_extension(filename
, ext
, expected_real_ext
=None):
2702 name
, real_ext
= os
.path
.splitext(filename
)
2703 return '{}.{}'.format(
2704 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2708 def check_executable(exe
, args
=[]):
2709 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2710 args can be a list of arguments for a short output (like -version) """
2712 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2718 def _get_exe_version_output(exe
, args
):
2720 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2721 # SIGTTOU if yt-dlp is run in the background.
2722 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2723 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2724 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2730 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2731 assert isinstance(output
, str)
2732 if version_re
is None:
2733 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2734 m
= re
.search(version_re
, output
)
2741 def get_exe_version(exe
, args
=['--version'],
2742 version_re
=None, unrecognized
='present'):
2743 """ Returns the version of the specified executable,
2744 or False if the executable is not present """
2745 out
= _get_exe_version_output(exe
, args
)
2746 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2749 def frange(start
=0, stop
=None, step
=1):
2752 start
, stop
= 0, start
2753 sign
= [-1, 1][step
> 0] if step
else 0
2754 while sign
* start
< sign
* stop
:
2759 class LazyList(collections
.abc
.Sequence
):
2760 """Lazy immutable list from an iterable
2761 Note that slices of a LazyList are lists and not LazyList"""
2763 class IndexError(IndexError):
2766 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2767 self
._iterable
= iter(iterable
)
2768 self
._cache
= [] if _cache
is None else _cache
2769 self
._reversed
= reverse
2773 # We need to consume the entire iterable to iterate in reverse
2774 yield from self
.exhaust()
2776 yield from self
._cache
2777 for item
in self
._iterable
:
2778 self
._cache
.append(item
)
2782 self
._cache
.extend(self
._iterable
)
2783 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2787 """Evaluate the entire iterable"""
2788 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2791 def _reverse_index(x
):
2792 return None if x
is None else ~x
2794 def __getitem__(self
, idx
):
2795 if isinstance(idx
, slice):
2797 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2798 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2799 elif isinstance(idx
, int):
2801 idx
= self
._reverse
_index
(idx
)
2802 start
, stop
, step
= idx
, idx
, 0
2804 raise TypeError('indices must be integers or slices')
2805 if ((start
or 0) < 0 or (stop
or 0) < 0
2806 or (start
is None and step
< 0)
2807 or (stop
is None and step
> 0)):
2808 # We need to consume the entire iterable to be able to slice from the end
2809 # Obviously, never use this with infinite iterables
2812 return self
._cache
[idx
]
2813 except IndexError as e
:
2814 raise self
.IndexError(e
) from e
2815 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2817 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2819 return self
._cache
[idx
]
2820 except IndexError as e
:
2821 raise self
.IndexError(e
) from e
2825 self
[-1] if self
._reversed
else self
[0]
2826 except self
.IndexError:
2832 return len(self
._cache
)
2834 def __reversed__(self
):
2835 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2838 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2841 # repr and str should mimic a list. So we exhaust the iterable
2842 return repr(self
.exhaust())
2845 return repr(self
.exhaust())
2850 class IndexError(IndexError):
2854 # This is only useful for tests
2855 return len(self
.getslice())
2857 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2858 self
._pagefunc
= pagefunc
2859 self
._pagesize
= pagesize
2860 self
._pagecount
= float('inf')
2861 self
._use
_cache
= use_cache
2864 def getpage(self
, pagenum
):
2865 page_results
= self
._cache
.get(pagenum
)
2866 if page_results
is None:
2867 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2869 self
._cache
[pagenum
] = page_results
2872 def getslice(self
, start
=0, end
=None):
2873 return list(self
._getslice
(start
, end
))
2875 def _getslice(self
, start
, end
):
2876 raise NotImplementedError('This method must be implemented by subclasses')
2878 def __getitem__(self
, idx
):
2879 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2880 if not isinstance(idx
, int) or idx
< 0:
2881 raise TypeError('indices must be non-negative integers')
2882 entries
= self
.getslice(idx
, idx
+ 1)
2884 raise self
.IndexError()
2888 class OnDemandPagedList(PagedList
):
2889 """Download pages until a page with less than maximum results"""
2891 def _getslice(self
, start
, end
):
2892 for pagenum
in itertools
.count(start
// self
._pagesize
):
2893 firstid
= pagenum
* self
._pagesize
2894 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2895 if start
>= nextfirstid
:
2899 start
% self
._pagesize
2900 if firstid
<= start
< nextfirstid
2903 ((end
- 1) % self
._pagesize
) + 1
2904 if (end
is not None and firstid
<= end
<= nextfirstid
)
2908 page_results
= self
.getpage(pagenum
)
2910 self
._pagecount
= pagenum
- 1
2912 if startv
!= 0 or endv
is not None:
2913 page_results
= page_results
[startv
:endv
]
2914 yield from page_results
2916 # A little optimization - if current page is not "full", ie. does
2917 # not contain page_size videos then we can assume that this page
2918 # is the last one - there are no more ids on further pages -
2919 # i.e. no need to query again.
2920 if len(page_results
) + startv
< self
._pagesize
:
2923 # If we got the whole page, but the next page is not interesting,
2924 # break out early as well
2925 if end
== nextfirstid
:
2929 class InAdvancePagedList(PagedList
):
2930 """PagedList with total number of pages known in advance"""
2932 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2933 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2934 self
._pagecount
= pagecount
2936 def _getslice(self
, start
, end
):
2937 start_page
= start
// self
._pagesize
2938 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2939 skip_elems
= start
- start_page
* self
._pagesize
2940 only_more
= None if end
is None else end
- start
2941 for pagenum
in range(start_page
, end_page
):
2942 page_results
= self
.getpage(pagenum
)
2944 page_results
= page_results
[skip_elems
:]
2946 if only_more
is not None:
2947 if len(page_results
) < only_more
:
2948 only_more
-= len(page_results
)
2950 yield from page_results
[:only_more
]
2952 yield from page_results
2955 class PlaylistEntries
:
2956 MissingEntry
= object()
2957 is_exhausted
= False
2959 def __init__(self
, ydl
, info_dict
):
2962 # _entries must be assigned now since infodict can change during iteration
2963 entries
= info_dict
.get('entries')
2965 raise EntryNotInPlaylist('There are no entries')
2966 elif isinstance(entries
, list):
2967 self
.is_exhausted
= True
2969 requested_entries
= info_dict
.get('requested_entries')
2970 self
.is_incomplete
= requested_entries
is not None
2971 if self
.is_incomplete
:
2972 assert self
.is_exhausted
2973 self
._entries
= [self
.MissingEntry
] * max(requested_entries
or [0])
2974 for i
, entry
in zip(requested_entries
, entries
):
2975 self
._entries
[i
- 1] = entry
2976 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2977 self
._entries
= entries
2979 self
._entries
= LazyList(entries
)
2981 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2982 (?P<start>[+-]?\d+)?
2984 (?P<end>[+-]?\d+|inf(?:inite)?)?
2985 (?::(?P<step>[+-]?\d+))?
2989 def parse_playlist_items(cls
, string
):
2990 for segment
in string
.split(','):
2992 raise ValueError('There is two or more consecutive commas')
2993 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2995 raise ValueError(f
'{segment!r} is not a valid specification')
2996 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2997 if int_or_none(step
) == 0:
2998 raise ValueError(f
'Step in {segment!r} cannot be zero')
2999 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
3001 def get_requested_items(self
):
3002 playlist_items
= self
.ydl
.params
.get('playlist_items')
3003 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
3004 playlist_end
= self
.ydl
.params
.get('playlistend')
3005 # For backwards compatibility, interpret -1 as whole list
3006 if playlist_end
in (-1, None):
3008 if not playlist_items
:
3009 playlist_items
= f
'{playlist_start}:{playlist_end}'
3010 elif playlist_start
!= 1 or playlist_end
:
3011 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
3013 for index
in self
.parse_playlist_items(playlist_items
):
3014 for i
, entry
in self
[index
]:
3019 # TODO: Add auto-generated fields
3020 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3021 except (ExistingVideoReached
, RejectedVideoReached
):
3024 def get_full_count(self
):
3025 if self
.is_exhausted
and not self
.is_incomplete
:
3027 elif isinstance(self
._entries
, InAdvancePagedList
):
3028 if self
._entries
._pagesize
== 1:
3029 return self
._entries
._pagecount
3031 @functools.cached_property
3033 if isinstance(self
._entries
, list):
3036 entry
= self
._entries
[i
]
3038 entry
= self
.MissingEntry
3039 if not self
.is_incomplete
:
3040 raise self
.IndexError()
3041 if entry
is self
.MissingEntry
:
3042 raise EntryNotInPlaylist(f
'Entry {i + 1} cannot be found')
3047 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3048 except (LazyList
.IndexError, PagedList
.IndexError):
3049 raise self
.IndexError()
3052 def __getitem__(self
, idx
):
3053 if isinstance(idx
, int):
3054 idx
= slice(idx
, idx
)
3056 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3057 step
= 1 if idx
.step
is None else idx
.step
3058 if idx
.start
is None:
3059 start
= 0 if step
> 0 else len(self
) - 1
3061 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3063 # NB: Do not call len(self) when idx == [:]
3064 if idx
.stop
is None:
3065 stop
= 0 if step
< 0 else float('inf')
3067 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3068 stop
+= [-1, 1][step
> 0]
3070 for i
in frange(start
, stop
, step
):
3074 entry
= self
._getter
(i
)
3075 except self
.IndexError:
3076 self
.is_exhausted
= True
3083 return len(tuple(self
[:]))
3085 class IndexError(IndexError):
3089 def uppercase_escape(s
):
3090 unicode_escape
= codecs
.getdecoder('unicode_escape')
3092 r
'\\U[0-9a-fA-F]{8}',
3093 lambda m
: unicode_escape(m
.group(0))[0],
3097 def lowercase_escape(s
):
3098 unicode_escape
= codecs
.getdecoder('unicode_escape')
3100 r
'\\u[0-9a-fA-F]{4}',
3101 lambda m
: unicode_escape(m
.group(0))[0],
3105 def escape_rfc3986(s
):
3106 """Escape non-ASCII characters as suggested by RFC 3986"""
3107 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3110 def escape_url(url
):
3111 """Escape URL as suggested by RFC 3986"""
3112 url_parsed
= urllib
.parse
.urlparse(url
)
3113 return url_parsed
._replace
(
3114 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3115 path
=escape_rfc3986(url_parsed
.path
),
3116 params
=escape_rfc3986(url_parsed
.params
),
3117 query
=escape_rfc3986(url_parsed
.query
),
3118 fragment
=escape_rfc3986(url_parsed
.fragment
)
3122 def parse_qs(url
, **kwargs
):
3123 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3126 def read_batch_urls(batch_fd
):
3128 if not isinstance(url
, str):
3129 url
= url
.decode('utf-8', 'replace')
3130 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3131 for bom
in BOM_UTF8
:
3132 if url
.startswith(bom
):
3133 url
= url
[len(bom
):]
3135 if not url
or url
.startswith(('#', ';', ']')):
3137 # "#" cannot be stripped out since it is part of the URI
3138 # However, it can be safely stripped out if following a whitespace
3139 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3141 with contextlib
.closing(batch_fd
) as fd
:
3142 return [url
for url
in map(fixup
, fd
) if url
]
3145 def urlencode_postdata(*args
, **kargs
):
3146 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3149 def update_url_query(url
, query
):
3152 parsed_url
= urllib
.parse
.urlparse(url
)
3153 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3155 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3156 query
=urllib
.parse
.urlencode(qs
, True)))
3159 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3160 req_headers
= req
.headers
.copy()
3161 req_headers
.update(headers
or {})
3162 req_data
= data
or req
.data
3163 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3164 req_get_method
= req
.get_method()
3165 if req_get_method
== 'HEAD':
3166 req_type
= HEADRequest
3167 elif req_get_method
== 'PUT':
3168 req_type
= PUTRequest
3170 req_type
= urllib
.request
.Request
3172 req_url
, data
=req_data
, headers
=req_headers
,
3173 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3174 if hasattr(req
, 'timeout'):
3175 new_req
.timeout
= req
.timeout
3179 def _multipart_encode_impl(data
, boundary
):
3180 content_type
= 'multipart/form-data; boundary=%s' % boundary
3183 for k
, v
in data
.items():
3184 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3185 if isinstance(k
, str):
3187 if isinstance(v
, str):
3189 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3190 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3191 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3192 if boundary
.encode('ascii') in content
:
3193 raise ValueError('Boundary overlaps with data')
3196 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3198 return out
, content_type
3201 def multipart_encode(data
, boundary
=None):
3203 Encode a dict to RFC 7578-compliant form-data
3206 A dict where keys and values can be either Unicode or bytes-like
3209 If specified a Unicode object, it's used as the boundary. Otherwise
3210 a random boundary is generated.
3212 Reference: https://tools.ietf.org/html/rfc7578
3214 has_specified_boundary
= boundary
is not None
3217 if boundary
is None:
3218 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3221 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3224 if has_specified_boundary
:
3228 return out
, content_type
3231 def variadic(x
, allowed_types
=(str, bytes, dict)):
3232 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3235 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3236 for val
in map(d
.get
, variadic(key_or_keys
)):
3237 if val
is not None and (val
or not skip_false_values
):
3242 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3245 val
= f(*args
, **kwargs
)
3246 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3249 if expected_type
is None or isinstance(val
, expected_type
):
3253 def try_get(src
, getter
, expected_type
=None):
3254 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3257 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3258 return {k: v for k, v in dct.items() if cndn(k, v)}
3261 def merge_dicts(*dicts
):
3263 for a_dict
in dicts
:
3264 for k
, v
in a_dict
.items():
3265 if (v
is not None and k
not in merged
3266 or isinstance(v
, str) and merged
[k
] == ''):
3271 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3272 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3284 TV_PARENTAL_GUIDELINES
= {
3294 def parse_age_limit(s
):
3295 # isinstance(False, int) is True. So type() must be used instead
3296 if type(s
) is int: # noqa: E721
3297 return s
if 0 <= s
<= 21 else None
3298 elif not isinstance(s
, str):
3300 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3302 return int(m
.group('age'))
3305 return US_RATINGS
[s
]
3306 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3308 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3312 def strip_jsonp(code
):
3315 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3316 (?:\s*&&\s*(?P=func_name))?
3317 \s*\(\s*(?P<callback_data>.*)\);?
3318 \s*?(?://[^\n]*)*$''',
3319 r
'\g<callback_data>', code
)
3322 def js_to_json(code
, vars={}, *, strict
=False):
3323 # vars is a dict of var, val pairs to substitute
3324 STRING_QUOTES
= '\'"'
3325 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3326 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3327 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3329 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3330 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3333 def process_escape(match
):
3334 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3335 escape
= match
.group(1) or match
.group(2)
3337 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3338 else R
'\u00' if escape
== 'x'
3339 else '' if escape
== '\n'
3344 if v
in ('true', 'false', 'null'):
3346 elif v
in ('undefined', 'void 0'):
3348 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3351 if v
[0] in STRING_QUOTES
:
3352 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3353 return f
'"{escaped}"'
3355 for regex
, base
in INTEGER_TABLE
:
3356 im
= re
.match(regex
, v
)
3358 i
= int(im
.group(1), base
)
3359 return f
'"{i}":' if v
.endswith(':') else str(i
)
3365 except json
.decoder
.JSONDecodeError
:
3366 return json
.dumps(vars[v
])
3373 raise ValueError(f
'Unknown value: {v}')
3375 def create_map(mobj
):
3376 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3378 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3380 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3381 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3383 return re
.sub(rf
'''(?sx)
3385 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3386 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3387 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3388 [0-9]+(?={SKIP_RE}:)|
3393 def qualities(quality_ids
):
3394 """ Get a numeric quality value out of a list of possible values """
3397 return quality_ids
.index(qid
)
3403 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'video', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3407 'default': '%(title)s [%(id)s].%(ext)s',
3408 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3414 'description': 'description',
3415 'annotation': 'annotations.xml',
3416 'infojson': 'info.json',
3419 'pl_thumbnail': None,
3420 'pl_description': 'description',
3421 'pl_infojson': 'info.json',
3424 # As of [1] format syntax is:
3425 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3426 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3427 STR_FORMAT_RE_TMPL
= r
'''(?x)
3428 (?<!%)(?P<prefix>(?:%%)*)
3430 (?P<has_key>\((?P<key>{0})\))?
3432 (?P<conversion>[#0\-+ ]+)?
3434 (?P<precision>\.\d+)?
3435 (?P<len_mod>[hlL])? # unused in python
3436 {1} # conversion type
3441 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3444 def limit_length(s
, length
):
3445 """ Add ellipses to overly long strings """
3450 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3454 def version_tuple(v
):
3455 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3458 def is_outdated_version(version
, limit
, assume_new
=True):
3460 return not assume_new
3462 return version_tuple(version
) < version_tuple(limit
)
3464 return not assume_new
3467 def ytdl_is_updateable():
3468 """ Returns if yt-dlp can be updated with -U """
3470 from .update
import is_non_updateable
3472 return not is_non_updateable()
3475 def args_to_str(args
):
3476 # Get a short string representation for a subprocess command
3477 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3480 def error_to_compat_str(err
):
3484 def error_to_str(err
):
3485 return f
'{type(err).__name__}: {err}'
3488 def mimetype2ext(mt
, default
=NO_DEFAULT
):
3489 if not isinstance(mt
, str):
3490 if default
is not NO_DEFAULT
:
3506 'x-matroska': 'mkv',
3508 'x-mp4-fragmented': 'mp4',
3513 # application (streaming playlists)
3517 'vnd.apple.mpegurl': 'm3u8',
3518 'vnd.ms-sstr+xml': 'ism',
3519 'x-mpegurl': 'm3u8',
3523 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3.
3524 # Using .mp3 as it's the most popular one
3525 'audio/mpeg': 'mp3',
3526 'audio/webm': 'weba',
3527 'audio/x-matroska': 'mka',
3528 'audio/x-mpegurl': 'm3u',
3536 'x-realaudio': 'ra',
3547 'vnd.wap.wbmp': 'wbmp',
3554 'filmstrip+json': 'fs',
3555 'smptett+xml': 'tt',
3558 'x-ms-sami': 'sami',
3567 mimetype
= mt
.partition(';')[0].strip().lower()
3568 _
, _
, subtype
= mimetype
.rpartition('/')
3570 ext
= traverse_obj(MAP
, mimetype
, subtype
, subtype
.rsplit('+')[-1])
3573 elif default
is not NO_DEFAULT
:
3575 return subtype
.replace('+', '.')
3578 def ext2mimetype(ext_or_url
):
3581 if '.' not in ext_or_url
:
3582 ext_or_url
= f
'file.{ext_or_url}'
3583 return mimetypes
.guess_type(ext_or_url
)[0]
3586 def parse_codecs(codecs_str
):
3587 # http://tools.ietf.org/html/rfc6381
3590 split_codecs
= list(filter(None, map(
3591 str.strip
, codecs_str
.strip().strip(',').split(','))))
3592 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3593 for full_codec
in split_codecs
:
3594 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3595 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3596 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3600 if parts
[0] in ('dvh1', 'dvhe'):
3602 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3604 elif parts
[:2] == ['vp9', '2']:
3606 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-4',
3607 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3608 acodec
= acodec
or full_codec
3609 elif parts
[0] in ('stpp', 'wvtt'):
3610 scodec
= scodec
or full_codec
3612 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3613 if vcodec
or acodec
or scodec
:
3615 'vcodec': vcodec
or 'none',
3616 'acodec': acodec
or 'none',
3617 'dynamic_range': hdr
,
3618 **({'scodec': scodec}
if scodec
is not None else {}),
3620 elif len(split_codecs
) == 2:
3622 'vcodec': split_codecs
[0],
3623 'acodec': split_codecs
[1],
3628 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3629 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3631 allow_mkv
= not preferences
or 'mkv' in preferences
3633 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3634 return 'mkv' # TODO: any other format allows this?
3636 # TODO: All codecs supported by parse_codecs isn't handled here
3637 COMPATIBLE_CODECS
= {
3639 'av1', 'hevc', 'avc1', 'mp4a', 'ac-4', # fourcc (m3u8, mpd)
3640 'h264', 'aacl', 'ec-3', # Set in ISM
3643 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3644 'vp9x', 'vp8x', # in the webm spec
3648 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3649 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3651 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3652 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3653 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3657 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3660 for ext
in preferences
or vexts
:
3661 current_exts
= {ext, *vexts, *aexts}
3662 if ext
== 'mkv' or current_exts
== {ext}
or any(
3663 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3665 return 'mkv' if allow_mkv
else preferences
[-1]
3668 def urlhandle_detect_ext(url_handle
, default
=NO_DEFAULT
):
3669 getheader
= url_handle
.headers
.get
3671 cd
= getheader('Content-Disposition')
3673 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3675 e
= determine_ext(m
.group('filename'), default_ext
=None)
3679 meta_ext
= getheader('x-amz-meta-name')
3681 e
= meta_ext
.rpartition('.')[2]
3685 return mimetype2ext(getheader('Content-Type'), default
=default
)
3688 def encode_data_uri(data
, mime_type
):
3689 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3692 def age_restricted(content_limit
, age_limit
):
3693 """ Returns True iff the content should be blocked """
3695 if age_limit
is None: # No limit set
3697 if content_limit
is None:
3698 return False # Content available for everyone
3699 return age_limit
< content_limit
3702 # List of known byte-order-marks (BOM)
3704 (b
'\xef\xbb\xbf', 'utf-8'),
3705 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3706 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3707 (b
'\xff\xfe', 'utf-16-le'),
3708 (b
'\xfe\xff', 'utf-16-be'),
3712 def is_html(first_bytes
):
3713 """ Detect whether a file contains HTML by examining its first bytes. """
3716 for bom
, enc
in BOMS
:
3717 while first_bytes
.startswith(bom
):
3718 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3720 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3723 def determine_protocol(info_dict
):
3724 protocol
= info_dict
.get('protocol')
3725 if protocol
is not None:
3728 url
= sanitize_url(info_dict
['url'])
3729 if url
.startswith('rtmp'):
3731 elif url
.startswith('mms'):
3733 elif url
.startswith('rtsp'):
3736 ext
= determine_ext(url
)
3738 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3742 return urllib
.parse
.urlparse(url
).scheme
3745 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3746 """ Render a list of rows, each as a list of values.
3747 Text after a \t will be right aligned """
3749 return len(remove_terminal_sequences(string
).replace('\t', ''))
3751 def get_max_lens(table
):
3752 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3754 def filter_using_list(row
, filterArray
):
3755 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3757 max_lens
= get_max_lens(data
) if hide_empty
else []
3758 header_row
= filter_using_list(header_row
, max_lens
)
3759 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3761 table
= [header_row
] + data
3762 max_lens
= get_max_lens(table
)
3765 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3766 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3768 for pos
, text
in enumerate(map(str, row
)):
3770 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3772 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3773 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3777 def _match_one(filter_part
, dct
, incomplete
):
3778 # TODO: Generalize code with YoutubeDL._build_format_filter
3779 STRING_OPERATORS
= {
3780 '*=': operator
.contains
,
3781 '^=': lambda attr
, value
: attr
.startswith(value
),
3782 '$=': lambda attr
, value
: attr
.endswith(value
),
3783 '~=': lambda attr
, value
: re
.search(value
, attr
),
3785 COMPARISON_OPERATORS
= {
3787 '<=': operator
.le
, # "<=" must be defined above "<"
3794 if isinstance(incomplete
, bool):
3795 is_incomplete
= lambda _
: incomplete
3797 is_incomplete
= lambda k
: k
in incomplete
3799 operator_rex
= re
.compile(r
'''(?x)
3801 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3803 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3806 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3807 m = operator_rex.fullmatch(filter_part.strip())
3810 unnegated_op = COMPARISON_OPERATORS[m['op']]
3812 op = lambda attr, value: not unnegated_op(attr, value)
3815 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3817 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3818 actual_value = dct.get(m['key'])
3819 numeric_comparison = None
3820 if isinstance(actual_value, (int, float)):
3821 # If the original field is a string and matching comparisonvalue is
3822 # a number we should respect the origin of the original field
3823 # and process comparison value as a string (see
3824 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3826 numeric_comparison = int(comparison_value)
3828 numeric_comparison = parse_filesize(comparison_value)
3829 if numeric_comparison is None:
3830 numeric_comparison = parse_filesize(f'{comparison_value}B')
3831 if numeric_comparison is None:
3832 numeric_comparison = parse_duration(comparison_value)
3833 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3834 raise ValueError('Operator %s only supports string values!' % m['op'])
3835 if actual_value is None:
3836 return is_incomplete(m['key']) or m['none_inclusive']
3837 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3840 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3841 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3843 operator_rex = re.compile(r'''(?x
)
3844 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3845 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3846 m = operator_rex.fullmatch(filter_part.strip())
3848 op = UNARY_OPERATORS[m.group('op')]
3849 actual_value = dct.get(m.group('key'))
3850 if is_incomplete(m.group('key')) and actual_value is None:
3852 return op(actual_value)
3854 raise ValueError('Invalid filter part %r' % filter_part)
3857 def match_str(filter_str, dct, incomplete=False):
3858 """ Filter a dictionary with a simple string syntax.
3859 @returns Whether the filter passes
3860 @param incomplete Set of keys that is expected to be missing from dct.
3861 Can be True/False to indicate all/none of the keys may be missing.
3862 All conditions on incomplete keys pass if the key is missing
3865 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3866 for filter_part in re.split(r'(?<!\\)&', filter_str))
3869 def match_filter_func(filters):
3872 filters = set(variadic(filters))
3874 interactive = '-' in filters
3878 def _match_func(info_dict, incomplete=False):
3879 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3880 return NO_DEFAULT if interactive and not incomplete else None
3882 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3883 filter_str = ') | ('.join(map(str.strip, filters))
3884 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3888 class download_range_func:
3889 def __init__(self, chapters, ranges):
3890 self.chapters, self.ranges = chapters, ranges
3892 def __call__(self, info_dict, ydl):
3893 if not self.ranges and not self.chapters:
3896 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3897 else 'Cannot match chapters since chapter information is unavailable')
3898 for regex in self.chapters or []:
3899 for i, chapter in enumerate(info_dict.get('chapters') or []):
3900 if re.search(regex, chapter['title']):
3902 yield {**chapter, 'index': i}
3903 if self.chapters and warning:
3904 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3906 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3908 def __eq__(self, other):
3909 return (isinstance(other, download_range_func)
3910 and self.chapters == other.chapters and self.ranges == other.ranges)
3913 return f'{type(self).__name__}({self.chapters}, {self.ranges})'
3916 def parse_dfxp_time_expr(time_expr):
3920 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3922 return float(mobj.group('time_offset'))
3924 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3926 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3929 def srt_subtitles_timecode(seconds):
3930 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3933 def ass_subtitles_timecode(seconds):
3934 time = timetuple_from_msec(seconds * 1000)
3935 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3938 def dfxp2srt(dfxp_data):
3940 @param dfxp_data A
bytes-like
object containing DFXP data
3941 @returns A
unicode object containing converted SRT data
3943 LEGACY_NAMESPACES = (
3944 (b'http://www.w3.org/ns/ttml', [
3945 b'http://www.w3.org/2004/11/ttaf1',
3946 b'http://www.w3.org/2006/04/ttaf1',
3947 b'http://www.w3.org/2006/10/ttaf1',
3949 (b'http://www.w3.org/ns/ttml#styling', [
3950 b'http://www.w3.org/ns/ttml#style',
3954 SUPPORTED_STYLING = [
3963 _x = functools.partial(xpath_with_ns, ns_map={
3964 'xml': 'http://www.w3.org/XML/1998/namespace',
3965 'ttml': 'http://www.w3.org/ns/ttml',
3966 'tts': 'http://www.w3.org/ns/ttml#styling',
3972 class TTMLPElementParser:
3974 _unclosed_elements = []
3975 _applied_styles = []
3977 def start(self, tag, attrib):
3978 if tag in (_x('ttml:br'), 'br'):
3981 unclosed_elements = []
3983 element_style_id = attrib.get('style')
3985 style.update(default_style)
3986 if element_style_id:
3987 style.update(styles.get(element_style_id, {}))
3988 for prop in SUPPORTED_STYLING:
3989 prop_val = attrib.get(_x('tts:' + prop))
3991 style[prop] = prop_val
3994 for k, v in sorted(style.items()):
3995 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3998 font += ' color="%s"' % v
3999 elif k == 'fontSize':
4000 font += ' size="%s"' % v
4001 elif k == 'fontFamily':
4002 font += ' face="%s"' % v
4003 elif k == 'fontWeight' and v == 'bold':
4005 unclosed_elements.append('b')
4006 elif k == 'fontStyle' and v == 'italic':
4008 unclosed_elements.append('i')
4009 elif k == 'textDecoration' and v == 'underline':
4011 unclosed_elements.append('u')
4013 self._out += '<font' + font + '>'
4014 unclosed_elements.append('font')
4016 if self._applied_styles:
4017 applied_style.update(self._applied_styles[-1])
4018 applied_style.update(style)
4019 self._applied_styles.append(applied_style)
4020 self._unclosed_elements.append(unclosed_elements)
4023 if tag not in (_x('ttml:br'), 'br'):
4024 unclosed_elements = self._unclosed_elements.pop()
4025 for element in reversed(unclosed_elements):
4026 self._out += '</%s>' % element
4027 if unclosed_elements and self._applied_styles:
4028 self._applied_styles.pop()
4030 def data(self, data):
4034 return self._out.strip()
4036 def parse_node(node):
4037 target = TTMLPElementParser()
4038 parser = xml.etree.ElementTree.XMLParser(target=target)
4039 parser.feed(xml.etree.ElementTree.tostring(node))
4040 return parser.close()
4042 for k, v in LEGACY_NAMESPACES:
4044 dfxp_data = dfxp_data.replace(ns, k)
4046 dfxp = compat_etree_fromstring(dfxp_data)
4048 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
4051 raise ValueError('Invalid dfxp/TTML subtitle')
4055 for style in dfxp.findall(_x('.//ttml:style')):
4056 style_id = style.get('id') or style.get(_x('xml:id'))
4059 parent_style_id = style.get('style')
4061 if parent_style_id not in styles:
4064 styles[style_id] = styles[parent_style_id].copy()
4065 for prop in SUPPORTED_STYLING:
4066 prop_val = style.get(_x('tts:' + prop))
4068 styles.setdefault(style_id, {})[prop] = prop_val
4074 for p in ('body', 'div'):
4075 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
4078 style = styles.get(ele.get('style'))
4081 default_style.update(style)
4083 for para, index in zip(paras, itertools.count(1)):
4084 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
4085 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
4086 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
4087 if begin_time is None:
4092 end_time = begin_time + dur
4093 out.append('%d\n%s --> %s\n%s\n\n' % (
4095 srt_subtitles_timecode(begin_time),
4096 srt_subtitles_timecode(end_time),
4102 def cli_option(params, command_option, param, separator=None):
4103 param = params.get(param)
4104 return ([] if param is None
4105 else [command_option, str(param)] if separator is None
4106 else [f'{command_option}{separator}{param}'])
4109 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4110 param = params.get(param)
4111 assert param in (True, False, None)
4112 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4115 def cli_valueless_option(params, command_option, param, expected_value=True):
4116 return [command_option] if params.get(param) == expected_value else []
4119 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4120 if isinstance(argdict, (list, tuple)): # for backward compatibility
4127 assert isinstance(argdict, dict)
4129 assert isinstance(keys, (list, tuple))
4130 for key_list in keys:
4131 arg_list = list(filter(
4132 lambda x: x is not None,
4133 [argdict.get(key.lower()) for key in variadic(key_list)]))
4135 return [arg for args in arg_list for arg in args]
4139 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4140 main_key, exe = main_key.lower(), exe.lower()
4141 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4142 keys = [f'{root_key}{k}' for k in (keys or [''])]
4143 if root_key in keys:
4145 keys.append((main_key, exe))
4146 keys.append('default')
4149 return cli_configuration_args(argdict, keys, default, use_compat)
4153 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4212 'iw': 'heb', # Replaced by he in 1989 revision
4222 'in': 'ind', # Replaced by id in 1989 revision
4337 'ji': 'yid', # Replaced by yi in 1989 revision
4345 def short2long(cls, code):
4346 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4347 return cls._lang_map.get(code[:2])
4350 def long2short(cls, code):
4351 """Convert language code from ISO 639-2/T to ISO 639-1"""
4352 for short_name, long_name in cls._lang_map.items():
4353 if long_name == code:
4358 # From http://data.okfn.org/data/core/country-list
4360 'AF': 'Afghanistan',
4361 'AX': 'Åland Islands',
4364 'AS': 'American Samoa',
4369 'AG': 'Antigua and Barbuda',
4386 'BO': 'Bolivia, Plurinational State of',
4387 'BQ': 'Bonaire, Sint Eustatius and Saba',
4388 'BA': 'Bosnia and Herzegovina',
4390 'BV': 'Bouvet Island',
4392 'IO': 'British Indian Ocean Territory',
4393 'BN': 'Brunei Darussalam',
4395 'BF': 'Burkina Faso',
4401 'KY': 'Cayman Islands',
4402 'CF': 'Central African Republic',
4406 'CX': 'Christmas Island',
4407 'CC': 'Cocos (Keeling) Islands',
4411 'CD': 'Congo, the Democratic Republic of the',
4412 'CK': 'Cook Islands',
4414 'CI': 'Côte d\'Ivoire',
4419 'CZ': 'Czech Republic',
4423 'DO': 'Dominican Republic',
4426 'SV': 'El Salvador',
4427 'GQ': 'Equatorial Guinea',
4431 'FK': 'Falkland Islands (Malvinas)',
4432 'FO': 'Faroe Islands',
4436 'GF': 'French Guiana',
4437 'PF': 'French Polynesia',
4438 'TF': 'French Southern Territories',
4453 'GW': 'Guinea-Bissau',
4456 'HM': 'Heard Island and McDonald Islands',
4457 'VA': 'Holy See (Vatican City State)',
4464 'IR': 'Iran, Islamic Republic of',
4467 'IM': 'Isle of Man',
4477 'KP': 'Korea, Democratic People\'s Republic of',
4478 'KR': 'Korea, Republic of',
4481 'LA': 'Lao People\'s Democratic Republic',
4487 'LI': 'Liechtenstein',
4491 'MK': 'Macedonia, the Former Yugoslav Republic of',
4498 'MH': 'Marshall Islands',
4504 'FM': 'Micronesia, Federated States of',
4505 'MD': 'Moldova, Republic of',
4516 'NL': 'Netherlands',
4517 'NC': 'New Caledonia',
4518 'NZ': 'New Zealand',
4523 'NF': 'Norfolk Island',
4524 'MP': 'Northern Mariana Islands',
4529 'PS': 'Palestine, State of',
4531 'PG': 'Papua New Guinea',
4534 'PH': 'Philippines',
4538 'PR': 'Puerto Rico',
4542 'RU': 'Russian Federation',
4544 'BL': 'Saint Barthélemy',
4545 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4546 'KN': 'Saint Kitts and Nevis',
4547 'LC': 'Saint Lucia',
4548 'MF': 'Saint Martin (French part)',
4549 'PM': 'Saint Pierre and Miquelon',
4550 'VC': 'Saint Vincent and the Grenadines',
4553 'ST': 'Sao Tome and Principe',
4554 'SA': 'Saudi Arabia',
4558 'SL': 'Sierra Leone',
4560 'SX': 'Sint Maarten (Dutch part)',
4563 'SB': 'Solomon Islands',
4565 'ZA': 'South Africa',
4566 'GS': 'South Georgia and the South Sandwich Islands',
4567 'SS': 'South Sudan',
4572 'SJ': 'Svalbard and Jan Mayen',
4575 'CH': 'Switzerland',
4576 'SY': 'Syrian Arab Republic',
4577 'TW': 'Taiwan, Province of China',
4579 'TZ': 'Tanzania, United Republic of',
4581 'TL': 'Timor-Leste',
4585 'TT': 'Trinidad and Tobago',
4588 'TM': 'Turkmenistan',
4589 'TC': 'Turks and Caicos Islands',
4593 'AE': 'United Arab Emirates',
4594 'GB': 'United Kingdom',
4595 'US': 'United States',
4596 'UM': 'United States Minor Outlying Islands',
4600 'VE': 'Venezuela, Bolivarian Republic of',
4602 'VG': 'Virgin Islands, British',
4603 'VI': 'Virgin Islands, U.S.',
4604 'WF': 'Wallis and Futuna',
4605 'EH': 'Western Sahara',
4609 # Not ISO 3166 codes, but used for IP blocks
4610 'AP': 'Asia/Pacific Region',
4615 def short2full(cls, code):
4616 """Convert an ISO 3166-2 country code to the corresponding full name"""
4617 return cls._country_map.get(code.upper())
4621 # Major IPv4 address blocks per country
4623 'AD': '46.172.224.0/19',
4624 'AE': '94.200.0.0/13',
4625 'AF': '149.54.0.0/17',
4626 'AG': '209.59.64.0/18',
4627 'AI': '204.14.248.0/21',
4628 'AL': '46.99.0.0/16',
4629 'AM': '46.70.0.0/15',
4630 'AO': '105.168.0.0/13',
4631 'AP': '182.50.184.0/21',
4632 'AQ': '23.154.160.0/24',
4633 'AR': '181.0.0.0/12',
4634 'AS': '202.70.112.0/20',
4635 'AT': '77.116.0.0/14',
4636 'AU': '1.128.0.0/11',
4637 'AW': '181.41.0.0/18',
4638 'AX': '185.217.4.0/22',
4639 'AZ': '5.197.0.0/16',
4640 'BA': '31.176.128.0/17',
4641 'BB': '65.48.128.0/17',
4642 'BD': '114.130.0.0/16',
4644 'BF': '102.178.0.0/15',
4645 'BG': '95.42.0.0/15',
4646 'BH': '37.131.0.0/17',
4647 'BI': '154.117.192.0/18',
4648 'BJ': '137.255.0.0/16',
4649 'BL': '185.212.72.0/23',
4650 'BM': '196.12.64.0/18',
4651 'BN': '156.31.0.0/16',
4652 'BO': '161.56.0.0/16',
4653 'BQ': '161.0.80.0/20',
4654 'BR': '191.128.0.0/12',
4655 'BS': '24.51.64.0/18',
4656 'BT': '119.2.96.0/19',
4657 'BW': '168.167.0.0/16',
4658 'BY': '178.120.0.0/13',
4659 'BZ': '179.42.192.0/18',
4660 'CA': '99.224.0.0/11',
4661 'CD': '41.243.0.0/16',
4662 'CF': '197.242.176.0/21',
4663 'CG': '160.113.0.0/16',
4664 'CH': '85.0.0.0/13',
4665 'CI': '102.136.0.0/14',
4666 'CK': '202.65.32.0/19',
4667 'CL': '152.172.0.0/14',
4668 'CM': '102.244.0.0/14',
4669 'CN': '36.128.0.0/10',
4670 'CO': '181.240.0.0/12',
4671 'CR': '201.192.0.0/12',
4672 'CU': '152.206.0.0/15',
4673 'CV': '165.90.96.0/19',
4674 'CW': '190.88.128.0/17',
4675 'CY': '31.153.0.0/16',
4676 'CZ': '88.100.0.0/14',
4678 'DJ': '197.241.0.0/17',
4679 'DK': '87.48.0.0/12',
4680 'DM': '192.243.48.0/20',
4681 'DO': '152.166.0.0/15',
4682 'DZ': '41.96.0.0/12',
4683 'EC': '186.68.0.0/15',
4684 'EE': '90.190.0.0/15',
4685 'EG': '156.160.0.0/11',
4686 'ER': '196.200.96.0/20',
4687 'ES': '88.0.0.0/11',
4688 'ET': '196.188.0.0/14',
4689 'EU': '2.16.0.0/13',
4690 'FI': '91.152.0.0/13',
4691 'FJ': '144.120.0.0/16',
4692 'FK': '80.73.208.0/21',
4693 'FM': '119.252.112.0/20',
4694 'FO': '88.85.32.0/19',
4696 'GA': '41.158.0.0/15',
4698 'GD': '74.122.88.0/21',
4699 'GE': '31.146.0.0/16',
4700 'GF': '161.22.64.0/18',
4701 'GG': '62.68.160.0/19',
4702 'GH': '154.160.0.0/12',
4703 'GI': '95.164.0.0/16',
4704 'GL': '88.83.0.0/19',
4705 'GM': '160.182.0.0/15',
4706 'GN': '197.149.192.0/18',
4707 'GP': '104.250.0.0/19',
4708 'GQ': '105.235.224.0/20',
4709 'GR': '94.64.0.0/13',
4710 'GT': '168.234.0.0/16',
4711 'GU': '168.123.0.0/16',
4712 'GW': '197.214.80.0/20',
4713 'GY': '181.41.64.0/18',
4714 'HK': '113.252.0.0/14',
4715 'HN': '181.210.0.0/16',
4716 'HR': '93.136.0.0/13',
4717 'HT': '148.102.128.0/17',
4718 'HU': '84.0.0.0/14',
4719 'ID': '39.192.0.0/10',
4720 'IE': '87.32.0.0/12',
4721 'IL': '79.176.0.0/13',
4722 'IM': '5.62.80.0/20',
4723 'IN': '117.192.0.0/10',
4724 'IO': '203.83.48.0/21',
4725 'IQ': '37.236.0.0/14',
4726 'IR': '2.176.0.0/12',
4727 'IS': '82.221.0.0/16',
4728 'IT': '79.0.0.0/10',
4729 'JE': '87.244.64.0/18',
4730 'JM': '72.27.0.0/17',
4731 'JO': '176.29.0.0/16',
4732 'JP': '133.0.0.0/8',
4733 'KE': '105.48.0.0/12',
4734 'KG': '158.181.128.0/17',
4735 'KH': '36.37.128.0/17',
4736 'KI': '103.25.140.0/22',
4737 'KM': '197.255.224.0/20',
4738 'KN': '198.167.192.0/19',
4739 'KP': '175.45.176.0/22',
4740 'KR': '175.192.0.0/10',
4741 'KW': '37.36.0.0/14',
4742 'KY': '64.96.0.0/15',
4743 'KZ': '2.72.0.0/13',
4744 'LA': '115.84.64.0/18',
4745 'LB': '178.135.0.0/16',
4746 'LC': '24.92.144.0/20',
4747 'LI': '82.117.0.0/19',
4748 'LK': '112.134.0.0/15',
4749 'LR': '102.183.0.0/16',
4750 'LS': '129.232.0.0/17',
4751 'LT': '78.56.0.0/13',
4752 'LU': '188.42.0.0/16',
4753 'LV': '46.109.0.0/16',
4754 'LY': '41.252.0.0/14',
4755 'MA': '105.128.0.0/11',
4756 'MC': '88.209.64.0/18',
4757 'MD': '37.246.0.0/16',
4758 'ME': '178.175.0.0/17',
4759 'MF': '74.112.232.0/21',
4760 'MG': '154.126.0.0/17',
4761 'MH': '117.103.88.0/21',
4762 'MK': '77.28.0.0/15',
4763 'ML': '154.118.128.0/18',
4764 'MM': '37.111.0.0/17',
4765 'MN': '49.0.128.0/17',
4766 'MO': '60.246.0.0/16',
4767 'MP': '202.88.64.0/20',
4768 'MQ': '109.203.224.0/19',
4769 'MR': '41.188.64.0/18',
4770 'MS': '208.90.112.0/22',
4771 'MT': '46.11.0.0/16',
4772 'MU': '105.16.0.0/12',
4773 'MV': '27.114.128.0/18',
4774 'MW': '102.70.0.0/15',
4775 'MX': '187.192.0.0/11',
4776 'MY': '175.136.0.0/13',
4777 'MZ': '197.218.0.0/15',
4778 'NA': '41.182.0.0/16',
4779 'NC': '101.101.0.0/18',
4780 'NE': '197.214.0.0/18',
4781 'NF': '203.17.240.0/22',
4782 'NG': '105.112.0.0/12',
4783 'NI': '186.76.0.0/15',
4784 'NL': '145.96.0.0/11',
4785 'NO': '84.208.0.0/13',
4786 'NP': '36.252.0.0/15',
4787 'NR': '203.98.224.0/19',
4788 'NU': '49.156.48.0/22',
4789 'NZ': '49.224.0.0/14',
4790 'OM': '5.36.0.0/15',
4791 'PA': '186.72.0.0/15',
4792 'PE': '186.160.0.0/14',
4793 'PF': '123.50.64.0/18',
4794 'PG': '124.240.192.0/19',
4795 'PH': '49.144.0.0/13',
4796 'PK': '39.32.0.0/11',
4797 'PL': '83.0.0.0/11',
4798 'PM': '70.36.0.0/20',
4799 'PR': '66.50.0.0/16',
4800 'PS': '188.161.0.0/16',
4801 'PT': '85.240.0.0/13',
4802 'PW': '202.124.224.0/20',
4803 'PY': '181.120.0.0/14',
4804 'QA': '37.210.0.0/15',
4805 'RE': '102.35.0.0/16',
4806 'RO': '79.112.0.0/13',
4807 'RS': '93.86.0.0/15',
4808 'RU': '5.136.0.0/13',
4809 'RW': '41.186.0.0/16',
4810 'SA': '188.48.0.0/13',
4811 'SB': '202.1.160.0/19',
4812 'SC': '154.192.0.0/11',
4813 'SD': '102.120.0.0/13',
4814 'SE': '78.64.0.0/12',
4815 'SG': '8.128.0.0/10',
4816 'SI': '188.196.0.0/14',
4817 'SK': '78.98.0.0/15',
4818 'SL': '102.143.0.0/17',
4819 'SM': '89.186.32.0/19',
4820 'SN': '41.82.0.0/15',
4821 'SO': '154.115.192.0/18',
4822 'SR': '186.179.128.0/17',
4823 'SS': '105.235.208.0/21',
4824 'ST': '197.159.160.0/19',
4825 'SV': '168.243.0.0/16',
4826 'SX': '190.102.0.0/20',
4828 'SZ': '41.84.224.0/19',
4829 'TC': '65.255.48.0/20',
4830 'TD': '154.68.128.0/19',
4831 'TG': '196.168.0.0/14',
4832 'TH': '171.96.0.0/13',
4833 'TJ': '85.9.128.0/18',
4834 'TK': '27.96.24.0/21',
4835 'TL': '180.189.160.0/20',
4836 'TM': '95.85.96.0/19',
4837 'TN': '197.0.0.0/11',
4838 'TO': '175.176.144.0/21',
4839 'TR': '78.160.0.0/11',
4840 'TT': '186.44.0.0/15',
4841 'TV': '202.2.96.0/19',
4842 'TW': '120.96.0.0/11',
4843 'TZ': '156.156.0.0/14',
4844 'UA': '37.52.0.0/14',
4845 'UG': '102.80.0.0/13',
4847 'UY': '167.56.0.0/13',
4848 'UZ': '84.54.64.0/18',
4849 'VA': '212.77.0.0/19',
4850 'VC': '207.191.240.0/21',
4851 'VE': '186.88.0.0/13',
4852 'VG': '66.81.192.0/20',
4853 'VI': '146.226.0.0/16',
4854 'VN': '14.160.0.0/11',
4855 'VU': '202.80.32.0/20',
4856 'WF': '117.20.32.0/21',
4857 'WS': '202.4.32.0/19',
4858 'YE': '134.35.0.0/16',
4859 'YT': '41.242.116.0/22',
4860 'ZA': '41.0.0.0/11',
4861 'ZM': '102.144.0.0/13',
4862 'ZW': '102.177.192.0/18',
4866 def random_ipv4(cls, code_or_block):
4867 if len(code_or_block) == 2:
4868 block = cls._country_ip_map.get(code_or_block.upper())
4872 block = code_or_block
4873 addr, preflen = block.split('/')
4874 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4875 addr_max = addr_min | (0xffffffff >> int(preflen))
4876 return str(socket.inet_ntoa(
4877 struct.pack('!L', random.randint(addr_min, addr_max))))
4880 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4881 def __init__(self, proxies=None):
4882 # Set default handlers
4883 for type in ('http', 'https'):
4884 setattr(self, '%s_open' % type,
4885 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4886 meth(r, proxy, type))
4887 urllib.request.ProxyHandler.__init__(self, proxies)
4889 def proxy_open(self, req, proxy, type):
4890 req_proxy = req.headers.get('Ytdl-request-proxy')
4891 if req_proxy is not None:
4893 del req.headers['Ytdl-request-proxy']
4895 if proxy == '__noproxy__':
4896 return None # No Proxy
4897 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4898 req.add_header('Ytdl-socks-proxy', proxy)
4899 # yt-dlp's http/https handlers do wrapping the socket with socks
4901 return urllib.request.ProxyHandler.proxy_open(
4902 self, req, proxy, type)
4905 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4906 # released into Public Domain
4907 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4909 def long_to_bytes(n, blocksize=0):
4910 """long_to_bytes(n:long, blocksize:int) : string
4911 Convert a long integer to a byte string.
4913 If optional blocksize is given and greater than zero, pad the front of the
4914 byte string with binary zeros so that the length is a multiple of
4917 # after much testing, this algorithm was deemed to be the fastest
4921 s = struct.pack('>I', n & 0xffffffff) + s
4923 # strip off leading zeros
4924 for i in range(len(s)):
4925 if s[i] != b'\000'[0]:
4928 # only happens when n == 0
4932 # add back some pad bytes. this could be done more efficiently w.r.t. the
4933 # de-padding being done above, but sigh...
4934 if blocksize > 0 and len(s) % blocksize:
4935 s = (blocksize - len(s) % blocksize) * b'\000' + s
4939 def bytes_to_long(s):
4940 """bytes_to_long(string) : long
4941 Convert a byte string to a long integer.
4943 This is (essentially) the inverse of long_to_bytes().
4948 extra = (4 - length % 4)
4949 s = b'\000' * extra + s
4950 length = length + extra
4951 for i in range(0, length, 4):
4952 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4956 def ohdave_rsa_encrypt(data, exponent, modulus):
4958 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4961 data: data to encrypt, bytes-like object
4962 exponent, modulus: parameter e and N of RSA algorithm, both integer
4963 Output: hex string of encrypted data
4965 Limitation: supports one block encryption only
4968 payload = int(binascii.hexlify(data[::-1]), 16)
4969 encrypted = pow(payload, exponent, modulus)
4970 return '%x' % encrypted
4973 def pkcs1pad(data, length):
4975 Padding input data with PKCS#1 scheme
4977 @param {int[]} data input data
4978 @param {int} length target length
4979 @returns {int[]} padded data
4981 if len(data) > length - 11:
4982 raise ValueError('Input data too
long for PKCS
#1 padding')
4984 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4985 return [0, 2] + pseudo_random
+ [0] + data
4988 def _base_n_table(n
, table
):
4989 if not table
and not n
:
4990 raise ValueError('Either table or n must be specified')
4991 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4993 if n
and n
!= len(table
):
4994 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4998 def encode_base_n(num
, n
=None, table
=None):
4999 """Convert given int to a base-n string"""
5000 table
= _base_n_table(n
, table
)
5004 result
, base
= '', len(table
)
5006 result
= table
[num
% base
] + result
5011 def decode_base_n(string
, n
=None, table
=None):
5012 """Convert given base-n string to int"""
5013 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
5014 result
, base
= 0, len(table
)
5016 result
= result
* base
+ table
[char
]
5020 def decode_base(value
, digits
):
5021 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
5022 f
'in a future version. Use {__name__}.decode_base_n instead')
5023 return decode_base_n(value
, table
=digits
)
5026 def decode_packed_codes(code
):
5027 mobj
= re
.search(PACKED_CODES_RE
, code
)
5028 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
5031 symbols
= symbols
.split('|')
5036 base_n_count
= encode_base_n(count
, base
)
5037 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
5040 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
5044 def caesar(s
, alphabet
, shift
):
5049 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
5054 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
5057 def parse_m3u8_attributes(attrib
):
5059 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5060 if val
.startswith('"'):
5066 def urshift(val
, n
):
5067 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5070 # Based on png2str() written by @gdkchan and improved by @yokrysty
5071 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5072 def decode_png(png_data
):
5073 # Reference: https://www.w3.org/TR/PNG/
5074 header
= png_data
[8:]
5076 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5077 raise OSError('Not a valid PNG file.')
5079 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5080 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5085 length
= unpack_integer(header
[:4])
5088 chunk_type
= header
[:4]
5091 chunk_data
= header
[:length
]
5092 header
= header
[length
:]
5094 header
= header
[4:] # Skip CRC
5102 ihdr
= chunks
[0]['data']
5104 width
= unpack_integer(ihdr
[:4])
5105 height
= unpack_integer(ihdr
[4:8])
5109 for chunk
in chunks
:
5110 if chunk
['type'] == b
'IDAT':
5111 idat
+= chunk
['data']
5114 raise OSError('Unable to read PNG data.')
5116 decompressed_data
= bytearray(zlib
.decompress(idat
))
5121 def _get_pixel(idx
):
5126 for y
in range(height
):
5127 basePos
= y
* (1 + stride
)
5128 filter_type
= decompressed_data
[basePos
]
5132 pixels
.append(current_row
)
5134 for x
in range(stride
):
5135 color
= decompressed_data
[1 + basePos
+ x
]
5136 basex
= y
* stride
+ x
5141 left
= _get_pixel(basex
- 3)
5143 up
= _get_pixel(basex
- stride
)
5145 if filter_type
== 1: # Sub
5146 color
= (color
+ left
) & 0xff
5147 elif filter_type
== 2: # Up
5148 color
= (color
+ up
) & 0xff
5149 elif filter_type
== 3: # Average
5150 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5151 elif filter_type
== 4: # Paeth
5157 c
= _get_pixel(basex
- stride
- 3)
5165 if pa
<= pb
and pa
<= pc
:
5166 color
= (color
+ a
) & 0xff
5168 color
= (color
+ b
) & 0xff
5170 color
= (color
+ c
) & 0xff
5172 current_row
.append(color
)
5174 return width
, height
, pixels
5177 def write_xattr(path
, key
, value
):
5178 # Windows: Write xattrs to NTFS Alternate Data Streams:
5179 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5180 if compat_os_name
== 'nt':
5181 assert ':' not in key
5182 assert os
.path
.exists(path
)
5185 with open(f
'{path}:{key}', 'wb') as f
:
5187 except OSError as e
:
5188 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5191 # UNIX Method 1. Use xattrs/pyxattrs modules
5194 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5195 # Unicode arguments are not supported in pyxattr until version 0.5.0
5196 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5197 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5198 setxattr
= xattr
.set
5200 setxattr
= xattr
.setxattr
5204 setxattr(path
, key
, value
)
5205 except OSError as e
:
5206 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5209 # UNIX Method 2. Use setfattr/xattr executables
5210 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5211 else 'xattr' if check_executable('xattr', ['-h']) else None)
5213 raise XAttrUnavailableError(
5214 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5215 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5217 value
= value
.decode()
5219 _
, stderr
, returncode
= Popen
.run(
5220 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5221 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5222 except OSError as e
:
5223 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5225 raise XAttrMetadataError(returncode
, stderr
)
5228 def random_birthday(year_field
, month_field
, day_field
):
5229 start_date
= datetime
.date(1950, 1, 1)
5230 end_date
= datetime
.date(1995, 12, 31)
5231 offset
= random
.randint(0, (end_date
- start_date
).days
)
5232 random_date
= start_date
+ datetime
.timedelta(offset
)
5234 year_field
: str(random_date
.year
),
5235 month_field
: str(random_date
.month
),
5236 day_field
: str(random_date
.day
),
5240 # Templates for internet shortcut files, which are plain text files.
5241 DOT_URL_LINK_TEMPLATE
= '''\
5246 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5247 <?xml version="1.0" encoding="UTF-8"?>
5248 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5249 <plist version="1.0">
5252 \t<string>%(url)s</string>
5257 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5267 'url': DOT_URL_LINK_TEMPLATE
,
5268 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5269 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5273 def iri_to_uri(iri
):
5275 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5277 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.
5280 iri_parts
= urllib
.parse
.urlparse(iri
)
5282 if '[' in iri_parts
.netloc
:
5283 raise ValueError('IPv6 URIs are not, yet, supported.')
5284 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5286 # 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.
5289 if iri_parts
.username
:
5290 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5291 if iri_parts
.password
is not None:
5292 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5295 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5296 # The 'idna' encoding produces ASCII text.
5297 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5298 net_location
+= ':' + str(iri_parts
.port
)
5300 return urllib
.parse
.urlunparse(
5304 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5306 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5307 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5309 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5310 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5312 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5314 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5317 def to_high_limit_path(path
):
5318 if sys
.platform
in ['win32', 'cygwin']:
5319 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5320 return '\\\\?\\' + os
.path
.abspath(path
)
5325 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5326 val
= traverse_obj(obj
, *variadic(field
))
5327 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5329 return template
% func(val
)
5332 def clean_podcast_url(url
):
5333 return re
.sub(r
'''(?x)
5337 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5340 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5343 cn\.co| # https://podcorn.com/analytics-prefix/
5344 st\.fm # https://podsights.com/docs/
5349 _HEX_TABLE
= '0123456789abcdef'
5352 def random_uuidv4():
5353 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5356 def make_dir(path
, to_screen
=None):
5358 dn
= os
.path
.dirname(path
)
5359 if dn
and not os
.path
.exists(dn
):
5362 except OSError as err
:
5363 if callable(to_screen
) is not None:
5364 to_screen('unable to create directory ' + error_to_compat_str(err
))
5368 def get_executable_path():
5369 from .update
import _get_variant_and_executable_path
5371 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5374 def get_user_config_dirs(package_name
):
5377 # .config (e.g. ~/.config/package_name)
5378 xdg_config_home
= os
.getenv('XDG_CONFIG_HOME') or compat_expanduser('~/.config')
5379 config_dir
= os
.path
.join(xdg_config_home
, package_name
)
5380 if os
.path
.isdir(config_dir
):
5381 locations
.add(config_dir
)
5383 # appdata (%APPDATA%/package_name)
5384 appdata_dir
= os
.getenv('appdata')
5386 config_dir
= os
.path
.join(appdata_dir
, package_name
)
5387 if os
.path
.isdir(config_dir
):
5388 locations
.add(config_dir
)
5390 # home (~/.package_name)
5391 user_config_directory
= os
.path
.join(compat_expanduser('~'), '.%s' % package_name
)
5392 if os
.path
.isdir(user_config_directory
):
5393 locations
.add(user_config_directory
)
5398 def get_system_config_dirs(package_name
):
5401 system_config_directory
= os
.path
.join('/etc', package_name
)
5402 if os
.path
.isdir(system_config_directory
):
5403 locations
.add(system_config_directory
)
5408 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5409 casesense
=True, is_user_input
=False, traverse_string
=False):
5411 Safely traverse nested `dict`s and `Sequence`s
5413 >>> obj = [{}, {"key": "value"}]
5414 >>> traverse_obj(obj, (1, "key"))
5417 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5418 The next path will also be tested if the path branched but no results could be found.
5419 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5420 A value of None is treated as the absence of a value.
5422 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5424 The keys in the path can be one of:
5425 - `None`: Return the current object.
5426 - `str`/`int`: Return `obj[key]`. For `re.Match, return `obj.group(key)`.
5427 - `slice`: Branch out and return all values in `obj[key]`.
5428 - `Ellipsis`: Branch out and return a list of all values.
5429 - `tuple`/`list`: Branch out and return a list of all matching values.
5430 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5431 - `function`: Branch out and return values filtered by the function.
5432 Read as: `[value for key, value in obj if function(key, value)]`.
5433 For `Sequence`s, `key` is the index of the value.
5434 - `dict` Transform the current object and return a matching dict.
5435 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5437 `tuple`, `list`, and `dict` all support nested paths and branches.
5439 @params paths Paths which to traverse by.
5440 @param default Value to return if the paths do not match.
5441 @param expected_type If a `type`, only accept final values of this type.
5442 If any other callable, try to call the function on each result.
5443 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5444 @param casesense If `False`, consider string dictionary keys as case insensitive.
5446 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5448 @param is_user_input Whether the keys are generated from user input.
5449 If `True` strings get converted to `int`/`slice` if needed.
5450 @param traverse_string Whether to traverse into objects as strings.
5451 If `True`, any non-compatible object will first be
5452 converted into a string and then traversed into.
5455 @returns The result of the object traversal.
5456 If successful, `get_all=True`, and the path branches at least once,
5457 then a list of results is returned instead.
5458 A list is always returned if the last path branches and no `default` is given.
5460 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5461 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5463 if isinstance(expected_type
, type):
5464 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5466 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5468 def apply_key(key
, obj
):
5475 elif isinstance(key
, (list, tuple)):
5477 _
, result
= apply_path(obj
, branch
)
5481 if isinstance(obj
, collections
.abc
.Mapping
):
5482 yield from obj
.values()
5483 elif is_sequence(obj
):
5485 elif isinstance(obj
, re
.Match
):
5486 yield from obj
.groups()
5487 elif traverse_string
:
5491 if is_sequence(obj
):
5492 iter_obj
= enumerate(obj
)
5493 elif isinstance(obj
, collections
.abc
.Mapping
):
5494 iter_obj
= obj
.items()
5495 elif isinstance(obj
, re
.Match
):
5496 iter_obj
= enumerate((obj
.group(), *obj
.groups()))
5497 elif traverse_string
:
5498 iter_obj
= enumerate(str(obj
))
5501 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5503 elif isinstance(key
, dict):
5504 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5505 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5506 if v
is not None or default
is not NO_DEFAULT
}
5508 elif isinstance(obj
, collections
.abc
.Mapping
):
5509 yield (obj
.get(key
) if casesense
or (key
in obj
)
5510 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5512 elif isinstance(obj
, re
.Match
):
5513 if isinstance(key
, int) or casesense
:
5514 with contextlib
.suppress(IndexError):
5515 yield obj
.group(key
)
5518 if not isinstance(key
, str):
5521 yield next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5525 key
= (int_or_none(key
) if ':' not in key
5526 else slice(*map(int_or_none
, key
.split(':'))))
5528 if not isinstance(key
, (int, slice)):
5531 if not is_sequence(obj
):
5532 if not traverse_string
:
5536 with contextlib
.suppress(IndexError):
5539 def apply_path(start_obj
, path
):
5541 has_branched
= False
5543 for key
in variadic(path
):
5544 if is_user_input
and key
== ':':
5547 if not casesense
and isinstance(key
, str):
5548 key
= key
.casefold()
5550 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5553 key_func
= functools
.partial(apply_key
, key
)
5554 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5556 return has_branched
, objs
5558 def _traverse_obj(obj
, path
, use_list
=True):
5559 has_branched
, results
= apply_path(obj
, path
)
5560 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5562 if get_all
and has_branched
:
5563 return results
.exhaust() if results
or use_list
else None
5565 return results
[0] if results
else None
5567 for index
, path
in enumerate(paths
, 1):
5568 use_list
= default
is NO_DEFAULT
and index
== len(paths
)
5569 result
= _traverse_obj(obj
, path
, use_list
)
5570 if result
is not None:
5573 return None if default
is NO_DEFAULT
else default
5576 def traverse_dict(dictn
, keys
, casesense
=True):
5577 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5578 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5579 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5582 def get_first(obj
, keys
, **kwargs
):
5583 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5586 def time_seconds(**kwargs
):
5587 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5588 return t
.timestamp()
5591 # create a JSON Web Signature (jws) with HS256 algorithm
5592 # the resulting format is in JWS Compact Serialization
5593 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5594 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5595 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5601 header_data
.update(headers
)
5602 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5603 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5604 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5605 signature_b64
= base64
.b64encode(h
.digest())
5606 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5610 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5611 def jwt_decode_hs256(jwt
):
5612 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5613 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5614 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5618 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5622 def supports_terminal_sequences(stream
):
5623 if compat_os_name
== 'nt':
5624 if not WINDOWS_VT_MODE
:
5626 elif not os
.getenv('TERM'):
5629 return stream
.isatty()
5630 except BaseException
:
5634 def windows_enable_vt_mode():
5635 """Ref: https://bugs.python.org/issue30075 """
5636 if get_windows_version() < (10, 0, 10586):
5640 import ctypes
.wintypes
5643 ENABLE_VIRTUAL_TERMINAL_PROCESSING
= 0x0004
5645 dll
= ctypes
.WinDLL('kernel32', use_last_error
=False)
5646 handle
= os
.open('CONOUT$', os
.O_RDWR
)
5649 h_out
= ctypes
.wintypes
.HANDLE(msvcrt
.get_osfhandle(handle
))
5650 dw_original_mode
= ctypes
.wintypes
.DWORD()
5651 success
= dll
.GetConsoleMode(h_out
, ctypes
.byref(dw_original_mode
))
5653 raise Exception('GetConsoleMode failed')
5655 success
= dll
.SetConsoleMode(h_out
, ctypes
.wintypes
.DWORD(
5656 dw_original_mode
.value | ENABLE_VIRTUAL_TERMINAL_PROCESSING
))
5658 raise Exception('SetConsoleMode failed')
5659 except Exception as e
:
5660 write_string(f
'WARNING: Cannot enable VT mode - {e}')
5662 global WINDOWS_VT_MODE
5663 WINDOWS_VT_MODE
= True
5664 supports_terminal_sequences
.cache_clear()
5669 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5672 def remove_terminal_sequences(string
):
5673 return _terminal_sequences_re
.sub('', string
)
5676 def number_of_digits(number
):
5677 return len('%d' % number
)
5680 def join_nonempty(*values
, delim
='-', from_dict
=None):
5681 if from_dict
is not None:
5682 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5683 return delim
.join(map(str, filter(None, values
)))
5686 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5688 Find the largest format dimensions in terms of video width and, for each thumbnail:
5689 * Modify the URL: Match the width with the provided regex and replace with the former width
5692 This function is useful with video services that scale the provided thumbnails on demand
5694 _keys
= ('width', 'height')
5695 max_dimensions
= max(
5696 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5698 if not max_dimensions
[0]:
5702 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5703 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5704 for thumbnail
in thumbnails
5708 def parse_http_range(range):
5709 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5711 return None, None, None
5712 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5714 return None, None, None
5715 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5718 def read_stdin(what
):
5719 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5720 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5724 def determine_file_encoding(data
):
5726 Detect the text encoding used
5727 @returns (encoding, bytes to skip)
5730 # BOM marks are given priority over declarations
5731 for bom
, enc
in BOMS
:
5732 if data
.startswith(bom
):
5733 return enc
, len(bom
)
5735 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5736 # We ignore the endianness to get a good enough match
5737 data
= data
.replace(b
'\0', b
'')
5738 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5739 return mobj
.group(1).decode() if mobj
else None, 0
5746 __initialized
= False
5748 def __init__(self
, parser
, label
=None):
5749 self
.parser
, self
.label
= parser
, label
5750 self
._loaded
_paths
, self
.configs
= set(), []
5752 def init(self
, args
=None, filename
=None):
5753 assert not self
.__initialized
5754 self
.own_args
, self
.filename
= args
, filename
5755 return self
.load_configs()
5757 def load_configs(self
):
5760 location
= os
.path
.realpath(self
.filename
)
5761 directory
= os
.path
.dirname(location
)
5762 if location
in self
._loaded
_paths
:
5764 self
._loaded
_paths
.add(location
)
5766 self
.__initialized
= True
5767 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5768 self
.parsed_args
= self
.own_args
5769 for location
in opts
.config_locations
or []:
5771 if location
in self
._loaded
_paths
:
5773 self
._loaded
_paths
.add(location
)
5774 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5776 location
= os
.path
.join(directory
, expand_path(location
))
5777 if os
.path
.isdir(location
):
5778 location
= os
.path
.join(location
, 'yt-dlp.conf')
5779 if not os
.path
.exists(location
):
5780 self
.parser
.error(f
'config location {location} does not exist')
5781 self
.append_config(self
.read_file(location
), location
)
5785 label
= join_nonempty(
5786 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5788 return join_nonempty(
5789 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5790 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5794 def read_file(filename
, default
=[]):
5796 optionf
= open(filename
, 'rb')
5798 return default
# silently skip if file is not present
5800 enc
, skip
= determine_file_encoding(optionf
.read(512))
5801 optionf
.seek(skip
, io
.SEEK_SET
)
5803 enc
= None # silently skip read errors
5805 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5806 contents
= optionf
.read().decode(enc
or preferredencoding())
5807 res
= shlex
.split(contents
, comments
=True)
5808 except Exception as err
:
5809 raise ValueError(f
'Unable to parse "{filename}": {err}')
5815 def hide_login_info(opts
):
5816 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5817 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5822 return m
.group('key') + '=PRIVATE'
5826 opts
= list(map(_scrub_eq
, opts
))
5827 for idx
, opt
in enumerate(opts
):
5828 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5829 opts
[idx
+ 1] = 'PRIVATE'
5832 def append_config(self
, *args
, label
=None):
5833 config
= type(self
)(self
.parser
, label
)
5834 config
._loaded
_paths
= self
._loaded
_paths
5835 if config
.init(*args
):
5836 self
.configs
.append(config
)
5840 for config
in reversed(self
.configs
):
5841 yield from config
.all_args
5842 yield from self
.parsed_args
or []
5844 def parse_known_args(self
, **kwargs
):
5845 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5847 def parse_args(self
):
5848 return self
.parser
.parse_args(self
.all_args
)
5851 class WebSocketsWrapper
:
5852 """Wraps websockets module to use in non-async scopes"""
5855 def __init__(self
, url
, headers
=None, connect
=True):
5856 self
.loop
= asyncio
.new_event_loop()
5857 # XXX: "loop" is deprecated
5858 self
.conn
= websockets
.connect(
5859 url
, extra_headers
=headers
, ping_interval
=None,
5860 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5863 atexit
.register(self
.__exit
__, None, None, None)
5865 def __enter__(self
):
5867 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5870 def send(self
, *args
):
5871 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5873 def recv(self
, *args
):
5874 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5876 def __exit__(self
, type, value
, traceback
):
5878 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5881 self
._cancel
_all
_tasks
(self
.loop
)
5883 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5884 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5886 def run_with_loop(main
, loop
):
5887 if not asyncio
.iscoroutine(main
):
5888 raise ValueError(f
'a coroutine was expected, got {main!r}')
5891 return loop
.run_until_complete(main
)
5893 loop
.run_until_complete(loop
.shutdown_asyncgens())
5894 if hasattr(loop
, 'shutdown_default_executor'):
5895 loop
.run_until_complete(loop
.shutdown_default_executor())
5898 def _cancel_all_tasks(loop
):
5899 to_cancel
= asyncio
.all_tasks(loop
)
5904 for task
in to_cancel
:
5907 # XXX: "loop" is removed in python 3.10+
5908 loop
.run_until_complete(
5909 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5911 for task
in to_cancel
:
5912 if task
.cancelled():
5914 if task
.exception() is not None:
5915 loop
.call_exception_handler({
5916 'message': 'unhandled exception during asyncio.run() shutdown',
5917 'exception': task
.exception(),
5922 def merge_headers(*dicts
):
5923 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5924 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5927 def cached_method(f
):
5928 """Cache a method"""
5929 signature
= inspect
.signature(f
)
5932 def wrapper(self
, *args
, **kwargs
):
5933 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5934 bound_args
.apply_defaults()
5935 key
= tuple(bound_args
.arguments
.values())[1:]
5937 cache
= vars(self
).setdefault('_cached_method__cache', {}).setdefault(f.__name__, {}
)
5938 if key
not in cache
:
5939 cache
[key
] = f(self
, *args
, **kwargs
)
5944 class classproperty
:
5945 """property access for class methods with optional caching"""
5946 def __new__(cls
, func
=None, *args
, **kwargs
):
5948 return functools
.partial(cls
, *args
, **kwargs
)
5949 return super().__new
__(cls
)
5951 def __init__(self
, func
, *, cache
=False):
5952 functools
.update_wrapper(self
, func
)
5954 self
._cache
= {} if cache
else None
5956 def __get__(self
, _
, cls
):
5957 if self
._cache
is None:
5958 return self
.func(cls
)
5959 elif cls
not in self
._cache
:
5960 self
._cache
[cls
] = self
.func(cls
)
5961 return self
._cache
[cls
]
5964 class Namespace(types
.SimpleNamespace
):
5965 """Immutable namespace"""
5968 return iter(self
.__dict
__.values())
5972 return self
.__dict
__.items()
5975 MEDIA_EXTENSIONS
= Namespace(
5976 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5977 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5978 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5979 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma', 'weba'),
5980 thumbnails
=('jpg', 'png', 'webp'),
5981 storyboards
=('mhtml', ),
5982 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5983 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5985 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5986 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5988 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5993 for retry in RetryManager(...):
5996 except SomeException as err:
6000 attempt
, _error
= 0, None
6002 def __init__(self
, _retries
, _error_callback
, **kwargs
):
6003 self
.retries
= _retries
or 0
6004 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
6006 def _should_retry(self
):
6007 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
6011 if self
._error
is NO_DEFAULT
:
6016 def error(self
, value
):
6020 while self
._should
_retry
():
6021 self
.error
= NO_DEFAULT
6025 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
6028 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
6029 """Utility function for reporting retries"""
6032 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
6037 elif isinstance(e
, ExtractorError
):
6038 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
6039 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
6041 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
6043 info(f
'Sleeping {delay:.2f} seconds ...')
6047 def make_archive_id(ie
, video_id
):
6048 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
6049 return f
'{ie_key.lower()} {video_id}'
6052 def truncate_string(s
, left
, right
=0):
6053 assert left
> 3 and right
>= 0
6054 if s
is None or len(s
) <= left
+ right
:
6056 return f
'{s[:left-3]}...{s[-right:] if right else ""}'
6059 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
6060 assert 'all' in alias_dict
, '"all" alias is required'
6061 requested
= list(start
or [])
6063 discard
= val
.startswith('-')
6067 if val
in alias_dict
:
6068 val
= alias_dict
[val
] if not discard
else [
6069 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
6070 # NB: Do not allow regex in aliases for performance
6071 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
6074 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
6075 else [val
] if val
in alias_dict
['all'] else None)
6077 raise ValueError(val
)
6080 for item
in current
:
6081 while item
in requested
:
6082 requested
.remove(item
)
6084 requested
.extend(current
)
6086 return orderedSet(requested
)
6090 regex
= r
' *((?P<reverse>\+)?(?P<field>[a-zA-Z0-9_]+)((?P<separator>[~:])(?P<limit>.*?))?)? *$'
6092 default
= ('hidden', 'aud_or_vid', 'hasvid', 'ie_pref', 'lang', 'quality',
6093 'res', 'fps', 'hdr:12', 'vcodec:vp9.2', 'channels', 'acodec',
6094 'size', 'br', 'asr', 'proto', 'ext', 'hasaud', 'source', 'id') # These must not be aliases
6095 ytdl_default
= ('hasaud', 'lang', 'quality', 'tbr', 'filesize', 'vbr',
6096 'height', 'width', 'proto', 'vext', 'abr', 'aext',
6097 'fps', 'fs_approx', 'source', 'id')
6100 'vcodec': {'type': 'ordered', 'regex': True,
6101 'order': ['av0?1', 'vp0?9.2', 'vp0?9', '[hx]265|he?vc?', '[hx]264|avc', 'vp0?8', 'mp4v|h263', 'theora', '', None, 'none']},
6102 'acodec': {'type': 'ordered', 'regex': True,
6103 'order': ['[af]lac', 'wav|aiff', 'opus', 'vorbis|ogg', 'aac', 'mp?4a?', 'mp3', 'ac-?4', 'e-?a?c-?3', 'ac-?3', 'dts', '', None, 'none']},
6104 'hdr': {'type': 'ordered', 'regex': True, 'field': 'dynamic_range',
6105 'order': ['dv', '(hdr)?12', r
'(hdr)?10\+', '(hdr)?10', 'hlg', '', 'sdr', None]},
6106 'proto': {'type': 'ordered', 'regex': True, 'field': 'protocol',
6107 'order': ['(ht|f)tps', '(ht|f)tp$', 'm3u8.*', '.*dash', 'websocket_frag', 'rtmpe?', '', 'mms|rtsp', 'ws|websocket', 'f4']},
6108 'vext': {'type': 'ordered', 'field': 'video_ext',
6109 'order': ('mp4', 'mov', 'webm', 'flv', '', 'none'),
6110 'order_free': ('webm', 'mp4', 'mov', 'flv', '', 'none')},
6111 'aext': {'type': 'ordered', 'regex': True, 'field': 'audio_ext',
6112 'order': ('m4a', 'aac', 'mp3', 'ogg', 'opus', 'web[am]', '', 'none'),
6113 'order_free': ('ogg', 'opus', 'web[am]', 'mp3', 'm4a', 'aac', '', 'none')},
6114 'hidden': {'visible': False, 'forced': True, 'type': 'extractor', 'max': -1000}
,
6115 'aud_or_vid': {'visible': False, 'forced': True, 'type': 'multiple',
6116 'field': ('vcodec', 'acodec'),
6117 'function': lambda it
: int(any(v
!= 'none' for v
in it
))},
6118 'ie_pref': {'priority': True, 'type': 'extractor'}
,
6119 'hasvid': {'priority': True, 'field': 'vcodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6120 'hasaud': {'field': 'acodec', 'type': 'boolean', 'not_in_list': ('none',)}
,
6121 'lang': {'convert': 'float', 'field': 'language_preference', 'default': -1}
,
6122 'quality': {'convert': 'float', 'default': -1}
,
6123 'filesize': {'convert': 'bytes'}
,
6124 'fs_approx': {'convert': 'bytes', 'field': 'filesize_approx'}
,
6125 'id': {'convert': 'string', 'field': 'format_id'}
,
6126 'height': {'convert': 'float_none'}
,
6127 'width': {'convert': 'float_none'}
,
6128 'fps': {'convert': 'float_none'}
,
6129 'channels': {'convert': 'float_none', 'field': 'audio_channels'}
,
6130 'tbr': {'convert': 'float_none'}
,
6131 'vbr': {'convert': 'float_none'}
,
6132 'abr': {'convert': 'float_none'}
,
6133 'asr': {'convert': 'float_none'}
,
6134 'source': {'convert': 'float', 'field': 'source_preference', 'default': -1}
,
6136 'codec': {'type': 'combined', 'field': ('vcodec', 'acodec')}
,
6137 'br': {'type': 'combined', 'field': ('tbr', 'vbr', 'abr'), 'same_limit': True}
,
6138 'size': {'type': 'combined', 'same_limit': True, 'field': ('filesize', 'fs_approx')}
,
6139 'ext': {'type': 'combined', 'field': ('vext', 'aext')}
,
6140 'res': {'type': 'multiple', 'field': ('height', 'width'),
6141 'function': lambda it
: (lambda l
: min(l
) if l
else 0)(tuple(filter(None, it
)))},
6143 # Actual field names
6144 'format_id': {'type': 'alias', 'field': 'id'}
,
6145 'preference': {'type': 'alias', 'field': 'ie_pref'}
,
6146 'language_preference': {'type': 'alias', 'field': 'lang'}
,
6147 'source_preference': {'type': 'alias', 'field': 'source'}
,
6148 'protocol': {'type': 'alias', 'field': 'proto'}
,
6149 'filesize_approx': {'type': 'alias', 'field': 'fs_approx'}
,
6150 'audio_channels': {'type': 'alias', 'field': 'channels'}
,
6153 'dimension': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6154 'resolution': {'type': 'alias', 'field': 'res', 'deprecated': True}
,
6155 'extension': {'type': 'alias', 'field': 'ext', 'deprecated': True}
,
6156 'bitrate': {'type': 'alias', 'field': 'br', 'deprecated': True}
,
6157 'total_bitrate': {'type': 'alias', 'field': 'tbr', 'deprecated': True}
,
6158 'video_bitrate': {'type': 'alias', 'field': 'vbr', 'deprecated': True}
,
6159 'audio_bitrate': {'type': 'alias', 'field': 'abr', 'deprecated': True}
,
6160 'framerate': {'type': 'alias', 'field': 'fps', 'deprecated': True}
,
6161 'filesize_estimate': {'type': 'alias', 'field': 'size', 'deprecated': True}
,
6162 'samplerate': {'type': 'alias', 'field': 'asr', 'deprecated': True}
,
6163 'video_ext': {'type': 'alias', 'field': 'vext', 'deprecated': True}
,
6164 'audio_ext': {'type': 'alias', 'field': 'aext', 'deprecated': True}
,
6165 'video_codec': {'type': 'alias', 'field': 'vcodec', 'deprecated': True}
,
6166 'audio_codec': {'type': 'alias', 'field': 'acodec', 'deprecated': True}
,
6167 'video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6168 'has_video': {'type': 'alias', 'field': 'hasvid', 'deprecated': True}
,
6169 'audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6170 'has_audio': {'type': 'alias', 'field': 'hasaud', 'deprecated': True}
,
6171 'extractor': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6172 'extractor_preference': {'type': 'alias', 'field': 'ie_pref', 'deprecated': True}
,
6175 def __init__(self
, ydl
, field_preference
):
6178 self
.evaluate_params(self
.ydl
.params
, field_preference
)
6179 if ydl
.params
.get('verbose'):
6180 self
.print_verbose_info(self
.ydl
.write_debug
)
6182 def _get_field_setting(self
, field
, key
):
6183 if field
not in self
.settings
:
6184 if key
in ('forced', 'priority'):
6186 self
.ydl
.deprecated_feature(f
'Using arbitrary fields ({field}) for format sorting is '
6187 'deprecated and may be removed in a future version')
6188 self
.settings
[field
] = {}
6189 propObj
= self
.settings
[field
]
6190 if key
not in propObj
:
6191 type = propObj
.get('type')
6193 default
= 'preference' if type == 'extractor' else (field
,) if type in ('combined', 'multiple') else field
6194 elif key
== 'convert':
6195 default
= 'order' if type == 'ordered' else 'float_string' if field
else 'ignore'
6197 default
= {'type': 'field', 'visible': True, 'order': [], 'not_in_list': (None,)}
.get(key
, None)
6198 propObj
[key
] = default
6201 def _resolve_field_value(self
, field
, value
, convertNone
=False):
6206 value
= value
.lower()
6207 conversion
= self
._get
_field
_setting
(field
, 'convert')
6208 if conversion
== 'ignore':
6210 if conversion
== 'string':
6212 elif conversion
== 'float_none':
6213 return float_or_none(value
)
6214 elif conversion
== 'bytes':
6215 return parse_bytes(value
)
6216 elif conversion
== 'order':
6217 order_list
= (self
._use
_free
_order
and self
._get
_field
_setting
(field
, 'order_free')) or self
._get
_field
_setting
(field
, 'order')
6218 use_regex
= self
._get
_field
_setting
(field
, 'regex')
6219 list_length
= len(order_list
)
6220 empty_pos
= order_list
.index('') if '' in order_list
else list_length
+ 1
6221 if use_regex
and value
is not None:
6222 for i
, regex
in enumerate(order_list
):
6223 if regex
and re
.match(regex
, value
):
6224 return list_length
- i
6225 return list_length
- empty_pos
# not in list
6226 else: # not regex or value = None
6227 return list_length
- (order_list
.index(value
) if value
in order_list
else empty_pos
)
6229 if value
.isnumeric():
6232 self
.settings
[field
]['convert'] = 'string'
6235 def evaluate_params(self
, params
, sort_extractor
):
6236 self
._use
_free
_order
= params
.get('prefer_free_formats', False)
6237 self
._sort
_user
= params
.get('format_sort', [])
6238 self
._sort
_extractor
= sort_extractor
6240 def add_item(field
, reverse
, closest
, limit_text
):
6241 field
= field
.lower()
6242 if field
in self
._order
:
6244 self
._order
.append(field
)
6245 limit
= self
._resolve
_field
_value
(field
, limit_text
)
6248 'closest': False if limit
is None else closest
,
6249 'limit_text': limit_text
,
6251 if field
in self
.settings
:
6252 self
.settings
[field
].update(data
)
6254 self
.settings
[field
] = data
6257 tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'forced'))
6258 + (tuple() if params
.get('format_sort_force', False)
6259 else tuple(field
for field
in self
.default
if self
._get
_field
_setting
(field
, 'priority')))
6260 + tuple(self
._sort
_user
) + tuple(sort_extractor
) + self
.default
)
6262 for item
in sort_list
:
6263 match
= re
.match(self
.regex
, item
)
6265 raise ExtractorError('Invalid format sort string "%s" given by extractor' % item
)
6266 field
= match
.group('field')
6269 if self
._get
_field
_setting
(field
, 'type') == 'alias':
6270 alias
, field
= field
, self
._get
_field
_setting
(field
, 'field')
6271 if self
._get
_field
_setting
(alias
, 'deprecated'):
6272 self
.ydl
.deprecated_feature(f
'Format sorting alias {alias} is deprecated and may '
6273 f
'be removed in a future version. Please use {field} instead')
6274 reverse
= match
.group('reverse') is not None
6275 closest
= match
.group('separator') == '~'
6276 limit_text
= match
.group('limit')
6278 has_limit
= limit_text
is not None
6279 has_multiple_fields
= self
._get
_field
_setting
(field
, 'type') == 'combined'
6280 has_multiple_limits
= has_limit
and has_multiple_fields
and not self
._get
_field
_setting
(field
, 'same_limit')
6282 fields
= self
._get
_field
_setting
(field
, 'field') if has_multiple_fields
else (field
,)
6283 limits
= limit_text
.split(':') if has_multiple_limits
else (limit_text
,) if has_limit
else tuple()
6284 limit_count
= len(limits
)
6285 for (i
, f
) in enumerate(fields
):
6286 add_item(f
, reverse
, closest
,
6287 limits
[i
] if i
< limit_count
6288 else limits
[0] if has_limit
and not has_multiple_limits
6291 def print_verbose_info(self
, write_debug
):
6293 write_debug('Sort order given by user: %s' % ', '.join(self
._sort
_user
))
6294 if self
._sort
_extractor
:
6295 write_debug('Sort order given by extractor: %s' % ', '.join(self
._sort
_extractor
))
6296 write_debug('Formats sorted by: %s' % ', '.join(['%s%s%s' % (
6297 '+' if self
._get
_field
_setting
(field
, 'reverse') else '', field
,
6298 '%s%s(%s)' % ('~' if self
._get
_field
_setting
(field
, 'closest') else ':',
6299 self
._get
_field
_setting
(field
, 'limit_text'),
6300 self
._get
_field
_setting
(field
, 'limit'))
6301 if self
._get
_field
_setting
(field
, 'limit_text') is not None else '')
6302 for field
in self
._order
if self
._get
_field
_setting
(field
, 'visible')]))
6304 def _calculate_field_preference_from_value(self
, format
, field
, type, value
):
6305 reverse
= self
._get
_field
_setting
(field
, 'reverse')
6306 closest
= self
._get
_field
_setting
(field
, 'closest')
6307 limit
= self
._get
_field
_setting
(field
, 'limit')
6309 if type == 'extractor':
6310 maximum
= self
._get
_field
_setting
(field
, 'max')
6311 if value
is None or (maximum
is not None and value
>= maximum
):
6313 elif type == 'boolean':
6314 in_list
= self
._get
_field
_setting
(field
, 'in_list')
6315 not_in_list
= self
._get
_field
_setting
(field
, 'not_in_list')
6316 value
= 0 if ((in_list
is None or value
in in_list
) and (not_in_list
is None or value
not in not_in_list
)) else -1
6317 elif type == 'ordered':
6318 value
= self
._resolve
_field
_value
(field
, value
, True)
6320 # try to convert to number
6321 val_num
= float_or_none(value
, default
=self
._get
_field
_setting
(field
, 'default'))
6322 is_num
= self
._get
_field
_setting
(field
, 'convert') != 'string' and val_num
is not None
6326 return ((-10, 0) if value
is None
6327 else (1, value
, 0) if not is_num
# if a field has mixed strings and numbers, strings are sorted higher
6328 else (0, -abs(value
- limit
), value
- limit
if reverse
else limit
- value
) if closest
6329 else (0, value
, 0) if not reverse
and (limit
is None or value
<= limit
)
6330 else (0, -value
, 0) if limit
is None or (reverse
and value
== limit
) or value
> limit
6331 else (-1, value
, 0))
6333 def _calculate_field_preference(self
, format
, field
):
6334 type = self
._get
_field
_setting
(field
, 'type') # extractor, boolean, ordered, field, multiple
6335 get_value
= lambda f
: format
.get(self
._get
_field
_setting
(f
, 'field'))
6336 if type == 'multiple':
6337 type = 'field' # Only 'field' is allowed in multiple for now
6338 actual_fields
= self
._get
_field
_setting
(field
, 'field')
6340 value
= self
._get
_field
_setting
(field
, 'function')(get_value(f
) for f
in actual_fields
)
6342 value
= get_value(field
)
6343 return self
._calculate
_field
_preference
_from
_value
(format
, field
, type, value
)
6345 def calculate_preference(self
, format
):
6346 # Determine missing protocol
6347 if not format
.get('protocol'):
6348 format
['protocol'] = determine_protocol(format
)
6350 # Determine missing ext
6351 if not format
.get('ext') and 'url' in format
:
6352 format
['ext'] = determine_ext(format
['url'])
6353 if format
.get('vcodec') == 'none':
6354 format
['audio_ext'] = format
['ext'] if format
.get('acodec') != 'none' else 'none'
6355 format
['video_ext'] = 'none'
6357 format
['video_ext'] = format
['ext']
6358 format
['audio_ext'] = 'none'
6359 # if format.get('preference') is None and format.get('ext') in ('f4f', 'f4m'): # Not supported?
6360 # format['preference'] = -1000
6362 if format
.get('preference') is None and format
.get('ext') == 'flv' and re
.match('[hx]265|he?vc?', format
.get('vcodec') or ''):
6363 # HEVC-over-FLV is out-of-spec by FLV's original spec
6364 # ref. https://trac.ffmpeg.org/ticket/6389
6365 # ref. https://github.com/yt-dlp/yt-dlp/pull/5821
6366 format
['preference'] = -100
6368 # Determine missing bitrates
6369 if format
.get('tbr') is None:
6370 if format
.get('vbr') is not None and format
.get('abr') is not None:
6371 format
['tbr'] = format
.get('vbr', 0) + format
.get('abr', 0)
6373 if format
.get('vcodec') != 'none' and format
.get('vbr') is None:
6374 format
['vbr'] = format
.get('tbr') - format
.get('abr', 0)
6375 if format
.get('acodec') != 'none' and format
.get('abr') is None:
6376 format
['abr'] = format
.get('tbr') - format
.get('vbr', 0)
6378 return tuple(self
._calculate
_field
_preference
(format
, field
) for field
in self
._order
)
6382 has_certifi
= bool(certifi
)
6383 has_websockets
= bool(websockets
)
6386 def load_plugins(name
, suffix
, namespace
):
6387 from .plugins
import load_plugins
6388 ret
= load_plugins(name
, suffix
)
6389 namespace
.update(ret
)