48 import xml
.etree
.ElementTree
51 from .compat
import functools
# isort: split
53 compat_etree_fromstring
,
55 compat_HTMLParseError
,
59 from .dependencies
import brotli
, certifi
, websockets
, xattr
60 from .socks
import ProxyType
, sockssocket
63 def register_socks_protocols():
64 # "Register" SOCKS protocols
65 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
66 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
67 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
68 if scheme
not in urllib
.parse
.uses_netloc
:
69 urllib
.parse
.uses_netloc
.append(scheme
)
72 # This is not clearly defined otherwise
73 compiled_regex_type
= type(re
.compile(''))
76 def random_user_agent():
77 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
118 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
121 SUPPORTED_ENCODINGS
= [
125 SUPPORTED_ENCODINGS
.append('br')
128 'User-Agent': random_user_agent(),
129 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
130 'Accept-Language': 'en-us,en;q=0.5',
131 'Sec-Fetch-Mode': 'navigate',
136 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
140 NO_DEFAULT
= object()
141 IDENTITY
= lambda x
: x
143 ENGLISH_MONTH_NAMES
= [
144 'January', 'February', 'March', 'April', 'May', 'June',
145 'July', 'August', 'September', 'October', 'November', 'December']
148 'en': ENGLISH_MONTH_NAMES
,
150 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
151 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
152 # these follow the genitive grammatical case (dopełniacz)
153 # some websites might be using nominative, which will require another month list
154 # https://en.wikibooks.org/wiki/Polish/Noun_cases
155 'pl': ['stycznia', 'lutego', 'marca', 'kwietnia', 'maja', 'czerwca',
156 'lipca', 'sierpnia', 'września', 'października', 'listopada', 'grudnia'],
159 # From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42
161 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0,
162 'AST': -4, 'ADT': -3, # Atlantic (used in Canada)
163 'EST': -5, 'EDT': -4, # Eastern
164 'CST': -6, 'CDT': -5, # Central
165 'MST': -7, 'MDT': -6, # Mountain
166 'PST': -8, 'PDT': -7 # Pacific
169 # needed for sanitizing filenames in restricted mode
170 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
171 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
172 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
202 '%Y-%m-%d %H:%M:%S.%f',
203 '%Y-%m-%d %H:%M:%S:%f',
206 '%Y-%m-%dT%H:%M:%SZ',
207 '%Y-%m-%dT%H:%M:%S.%fZ',
208 '%Y-%m-%dT%H:%M:%S.%f0Z',
210 '%Y-%m-%dT%H:%M:%S.%f',
213 '%b %d %Y at %H:%M:%S',
215 '%B %d %Y at %H:%M:%S',
219 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
220 DATE_FORMATS_DAY_FIRST
.extend([
230 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
231 DATE_FORMATS_MONTH_FIRST
.extend([
239 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
240 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>\s
*(?P
<json_ld
>{.+?}|\
[.+?\
])\s
*</script
>'
242 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
246 def preferredencoding():
247 """Get preferred encoding.
249 Returns the best encoding scheme for the system, based on
250 locale.getpreferredencoding() and some further tweaks.
253 pref = locale.getpreferredencoding()
261 def write_json_file(obj, fn):
262 """ Encode obj as JSON and write it to fn, atomically if possible """
264 tf = tempfile.NamedTemporaryFile(
265 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
266 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
270 json.dump(obj, tf, ensure_ascii=False)
271 if sys.platform == 'win32
':
272 # Need to remove existing file on Windows, else os.rename raises
273 # WindowsError or FileExistsError.
274 with contextlib.suppress(OSError):
276 with contextlib.suppress(OSError):
279 os.chmod(tf.name, 0o666 & ~mask)
280 os.rename(tf.name, fn)
282 with contextlib.suppress(OSError):
287 def find_xpath_attr(node, xpath, key, val=None):
288 """ Find the xpath xpath[@key=val] """
289 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
290 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
291 return node.find(expr)
293 # On python2.6 the xml.etree.ElementTree.Element methods don't support
294 # the namespace parameter
297 def xpath_with_ns(path
, ns_map
):
298 components
= [c
.split(':') for c
in path
.split('/')]
302 replaced
.append(c
[0])
305 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
306 return '/'.join(replaced
)
309 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
310 def _find_xpath(xpath
):
311 return node
.find(xpath
)
313 if isinstance(xpath
, str):
314 n
= _find_xpath(xpath
)
322 if default
is not NO_DEFAULT
:
325 name
= xpath
if name
is None else name
326 raise ExtractorError('Could not find XML element %s' % name
)
332 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
333 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
334 if n
is None or n
== default
:
337 if default
is not NO_DEFAULT
:
340 name
= xpath
if name
is None else name
341 raise ExtractorError('Could not find XML element\'s text %s' % name
)
347 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
348 n
= find_xpath_attr(node
, xpath
, key
)
350 if default
is not NO_DEFAULT
:
353 name
= f
'{xpath}[@{key}]' if name
is None else name
354 raise ExtractorError('Could not find XML attribute %s' % name
)
360 def get_element_by_id(id, html
, **kwargs
):
361 """Return the content of the tag with the specified ID in the passed HTML document"""
362 return get_element_by_attribute('id', id, html
, **kwargs
)
365 def get_element_html_by_id(id, html
, **kwargs
):
366 """Return the html of the tag with the specified ID in the passed HTML document"""
367 return get_element_html_by_attribute('id', id, html
, **kwargs
)
370 def get_element_by_class(class_name
, html
):
371 """Return the content of the first tag with the specified class in the passed HTML document"""
372 retval
= get_elements_by_class(class_name
, html
)
373 return retval
[0] if retval
else None
376 def get_element_html_by_class(class_name
, html
):
377 """Return the html of the first tag with the specified class in the passed HTML document"""
378 retval
= get_elements_html_by_class(class_name
, html
)
379 return retval
[0] if retval
else None
382 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
383 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
384 return retval
[0] if retval
else None
387 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
388 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
389 return retval
[0] if retval
else None
392 def get_elements_by_class(class_name
, html
, **kargs
):
393 """Return the content of all tags with the specified class in the passed HTML document as a list"""
394 return get_elements_by_attribute(
395 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
396 html, escape_value=False)
399 def get_elements_html_by_class(class_name, html):
400 """Return the html of all tags with the specified class in the passed HTML document as a list"""
401 return get_elements_html_by_attribute(
402 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
403 html, escape_value=False)
406 def get_elements_by_attribute(*args, **kwargs):
407 """Return the content of the tag with the specified attribute in the passed HTML document"""
408 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
411 def get_elements_html_by_attribute(*args, **kwargs):
412 """Return the html of the tag with the specified attribute in the passed HTML document"""
413 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
416 def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
:.-]+', escape_value=True):
418 Return the text (content) and the html (whole) of the tag with the specified
419 attribute in the passed HTML document
422 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
424 value = re.escape(value) if escape_value else value
426 partial_element_re = rf'''(?x
)
428 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
429 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
432 for m in re.finditer(partial_element_re, html):
433 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
436 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
441 class HTMLBreakOnClosingTagParser(html.parser.HTMLParser):
443 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
444 closing tag for the first opening tag it has encountered, and can be used
448 class HTMLBreakOnClosingTagException(Exception):
452 self.tagstack = collections.deque()
453 html.parser.HTMLParser.__init__(self)
458 def __exit__(self, *_):
462 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
463 # so data remains buffered; we no longer have any interest in it, thus
464 # override this method to discard it
467 def handle_starttag(self, tag, _):
468 self.tagstack.append(tag)
470 def handle_endtag(self, tag):
471 if not self.tagstack:
472 raise compat_HTMLParseError('no tags
in the stack
')
474 inner_tag = self.tagstack.pop()
478 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
479 if not self.tagstack:
480 raise self.HTMLBreakOnClosingTagException()
483 def get_element_text_and_html_by_tag(tag, html):
485 For the first element with the specified tag in the passed HTML document
486 return its' content (text
) and the whole
element (html
)
488 def find_or_raise(haystack, needle, exc):
490 return haystack.index(needle)
493 closing_tag = f'</{tag}>'
494 whole_start = find_or_raise(
495 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
496 content_start = find_or_raise(
497 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
498 content_start += whole_start + 1
499 with HTMLBreakOnClosingTagParser() as parser:
500 parser.feed(html[whole_start:content_start])
501 if not parser.tagstack or parser.tagstack[0] != tag:
502 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
503 offset = content_start
504 while offset < len(html):
505 next_closing_tag_start = find_or_raise(
506 html[offset:], closing_tag,
507 compat_HTMLParseError(f'closing {tag} tag not found'))
508 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
510 parser.feed(html[offset:offset + next_closing_tag_end])
511 offset += next_closing_tag_end
512 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
513 return html[content_start:offset + next_closing_tag_start], \
514 html[whole_start:offset + next_closing_tag_end]
515 raise compat_HTMLParseError('unexpected end of html')
518 class HTMLAttributeParser(html.parser.HTMLParser):
519 """Trivial HTML parser to gather the attributes
for a single element
"""
523 html.parser.HTMLParser.__init__(self)
525 def handle_starttag(self, tag, attrs):
526 self.attrs = dict(attrs)
527 raise compat_HTMLParseError('done')
530 class HTMLListAttrsParser(html.parser.HTMLParser):
531 """HTML parser to gather the attributes
for the elements of a
list"""
534 html.parser.HTMLParser.__init__(self)
538 def handle_starttag(self, tag, attrs):
539 if tag == 'li' and self._level == 0:
540 self.items.append(dict(attrs))
543 def handle_endtag(self, tag):
547 def extract_attributes(html_element):
548 """Given a string
for an HTML element such
as
550 a
="foo" B
="bar" c
="&98;az" d
=boz
551 empty
= noval entity
="&"
554 Decode
and return a dictionary of attributes
.
556 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
557 'empty': '', 'noval': None, 'entity': '&',
558 'sq': '"', 'dq': '\''
561 parser = HTMLAttributeParser()
562 with contextlib.suppress(compat_HTMLParseError):
563 parser.feed(html_element)
568 def parse_list(webpage):
569 """Given a string
for an series of HTML
<li
> elements
,
570 return a dictionary of their attributes
"""
571 parser = HTMLListAttrsParser()
577 def clean_html(html):
578 """Clean an HTML snippet into a readable string
"""
580 if html is None: # Convenience for sanitizing descriptions etc.
583 html = re.sub(r'\s+', ' ', html)
584 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
585 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
587 html = re.sub('<.*?>', '', html)
588 # Replace html entities
589 html = unescapeHTML(html)
593 class LenientJSONDecoder(json.JSONDecoder):
594 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
595 self.transform_source, self.ignore_extra = transform_source, ignore_extra
596 super().__init__(*args, **kwargs)
599 if self.transform_source:
600 s = self.transform_source(s)
602 if self.ignore_extra:
603 return self.raw_decode(s.lstrip())[0]
604 return super().decode(s)
605 except json.JSONDecodeError as e:
606 if e.pos is not None:
607 raise type(e)(f'{e.msg} in {s[e.pos-10:e.pos+10]!r}', s, e.pos)
611 def sanitize_open(filename, open_mode):
612 """Try to
open the given filename
, and slightly tweak it
if this fails
.
614 Attempts to
open the given filename
. If this fails
, it tries to change
615 the filename slightly
, step by step
, until it
's either able to open it
616 or it fails and raises a final exception, like the standard open()
619 It returns the tuple (stream, definitive_file_name).
622 if sys.platform == 'win32
':
625 # stdout may be any IO stream, e.g. when using contextlib.redirect_stdout
626 with contextlib.suppress(io.UnsupportedOperation):
627 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
628 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
630 for attempt in range(2):
633 if sys.platform == 'win32
':
634 # FIXME: An exclusive lock also locks the file from being read.
635 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
636 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
637 raise LockingUnsupportedError()
638 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
640 stream
= open(filename
, open_mode
)
641 return stream
, filename
642 except OSError as err
:
643 if attempt
or err
.errno
in (errno
.EACCES
,):
645 old_filename
, filename
= filename
, sanitize_path(filename
)
646 if old_filename
== filename
:
650 def timeconvert(timestr
):
651 """Convert RFC 2822 defined time string into system timestamp"""
653 timetuple
= email
.utils
.parsedate_tz(timestr
)
654 if timetuple
is not None:
655 timestamp
= email
.utils
.mktime_tz(timetuple
)
659 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
660 """Sanitizes a string so it could be used as part of a filename.
661 @param restricted Use a stricter subset of allowed characters
662 @param is_id Whether this is an ID that should be kept unchanged if possible.
663 If unset, yt-dlp's new sanitization rules are in effect
668 def replace_insane(char
):
669 if restricted
and char
in ACCENT_CHARS
:
670 return ACCENT_CHARS
[char
]
671 elif not restricted
and char
== '\n':
673 elif is_id
is NO_DEFAULT
and not restricted
and char
in '"*:<>?|/\\':
674 # Replace with their full-width unicode counterparts
675 return {'/': '\u29F8', '\\': '\u29f9'}
.get(char
, chr(ord(char
) + 0xfee0))
676 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
679 return '' if restricted
else '\''
681 return '\0_\0-' if restricted
else '\0 \0-'
682 elif char
in '\\/|*<>':
684 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
688 # Replace look-alike Unicode glyphs
689 if restricted
and (is_id
is NO_DEFAULT
or not is_id
):
690 s
= unicodedata
.normalize('NFKC', s
)
691 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
692 result
= ''.join(map(replace_insane
, s
))
693 if is_id
is NO_DEFAULT
:
694 result
= re
.sub(r
'(\0.)(?:(?=\1)..)+', r
'\1', result
) # Remove repeated substitute chars
695 STRIP_RE
= r
'(?:\0.|[ _-])*'
696 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
697 result
= result
.replace('\0', '') or '_'
700 while '__' in result
:
701 result
= result
.replace('__', '_')
702 result
= result
.strip('_')
703 # Common case of "Foreign band name - English song title"
704 if restricted
and result
.startswith('-_'):
706 if result
.startswith('-'):
707 result
= '_' + result
[len('-'):]
708 result
= result
.lstrip('.')
714 def sanitize_path(s
, force
=False):
715 """Sanitizes and normalizes path on Windows"""
716 if sys
.platform
== 'win32':
718 drive_or_unc
, _
= os
.path
.splitdrive(s
)
724 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
728 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
729 for path_part
in norm_path
]
731 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
732 elif force
and s
and s
[0] == os
.path
.sep
:
733 sanitized_path
.insert(0, os
.path
.sep
)
734 return os
.path
.join(*sanitized_path
)
737 def sanitize_url(url
, *, scheme
='http'):
738 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
739 # the number of unwanted failures due to missing protocol
742 elif url
.startswith('//'):
743 return f
'{scheme}:{url}'
744 # Fix some common typos seen so far
746 # https://github.com/ytdl-org/youtube-dl/issues/15649
747 (r
'^httpss://', r
'https://'),
748 # https://bx1.be/lives/direct-tv/
749 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
751 for mistake
, fixup
in COMMON_TYPOS
:
752 if re
.match(mistake
, url
):
753 return re
.sub(mistake
, fixup
, url
)
757 def extract_basic_auth(url
):
758 parts
= urllib
.parse
.urlsplit(url
)
759 if parts
.username
is None:
761 url
= urllib
.parse
.urlunsplit(parts
._replace
(netloc
=(
762 parts
.hostname
if parts
.port
is None
763 else '%s:%d' % (parts
.hostname
, parts
.port
))))
764 auth_payload
= base64
.b64encode(
765 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
766 return url
, f
'Basic {auth_payload.decode()}'
769 def sanitized_Request(url
, *args
, **kwargs
):
770 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
771 if auth_header
is not None:
772 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
773 headers
['Authorization'] = auth_header
774 return urllib
.request
.Request(url
, *args
, **kwargs
)
778 """Expand shell variables and ~"""
779 return os
.path
.expandvars(compat_expanduser(s
))
782 def orderedSet(iterable
, *, lazy
=False):
783 """Remove all duplicates from the input iterable"""
785 seen
= [] # Do not use set since the items can be unhashable
791 return _iter() if lazy
else list(_iter())
794 def _htmlentity_transform(entity_with_semicolon
):
795 """Transforms an HTML entity to a character."""
796 entity
= entity_with_semicolon
[:-1]
798 # Known non-numeric HTML entity
799 if entity
in html
.entities
.name2codepoint
:
800 return chr(html
.entities
.name2codepoint
[entity
])
802 # TODO: HTML5 allows entities without a semicolon.
803 # E.g. 'Éric' should be decoded as 'Éric'.
804 if entity_with_semicolon
in html
.entities
.html5
:
805 return html
.entities
.html5
[entity_with_semicolon
]
807 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
809 numstr
= mobj
.group(1)
810 if numstr
.startswith('x'):
812 numstr
= '0%s' % numstr
815 # See https://github.com/ytdl-org/youtube-dl/issues/7518
816 with contextlib
.suppress(ValueError):
817 return chr(int(numstr
, base
))
819 # Unknown entity in name, return its literal representation
820 return '&%s;' % entity
826 assert isinstance(s
, str)
829 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
832 def escapeHTML(text
):
835 .replace('&', '&')
836 .replace('<', '<')
837 .replace('>', '>')
838 .replace('"', '"')
839 .replace("'", ''')
843 def process_communicate_or_kill(p
, *args
, **kwargs
):
844 deprecation_warning(f
'"{__name__}.process_communicate_or_kill" is deprecated and may be removed '
845 f
'in a future version. Use "{__name__}.Popen.communicate_or_kill" instead')
846 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
849 class Popen(subprocess
.Popen
):
850 if sys
.platform
== 'win32':
851 _startupinfo
= subprocess
.STARTUPINFO()
852 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
857 def _fix_pyinstaller_ld_path(env
):
858 """Restore LD_LIBRARY_PATH when using PyInstaller
859 Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
860 https://github.com/yt-dlp/yt-dlp/issues/4573
862 if not hasattr(sys
, '_MEIPASS'):
866 orig
= env
.get(f
'{key}_ORIG')
872 _fix('LD_LIBRARY_PATH') # Linux
873 _fix('DYLD_LIBRARY_PATH') # macOS
875 def __init__(self
, *args
, env
=None, text
=False, **kwargs
):
877 env
= os
.environ
.copy()
878 self
._fix
_pyinstaller
_ld
_path
(env
)
881 kwargs
['universal_newlines'] = True # For 3.6 compatibility
882 kwargs
.setdefault('encoding', 'utf-8')
883 kwargs
.setdefault('errors', 'replace')
884 super().__init
__(*args
, env
=env
, **kwargs
, startupinfo
=self
._startupinfo
)
886 def communicate_or_kill(self
, *args
, **kwargs
):
888 return self
.communicate(*args
, **kwargs
)
889 except BaseException
: # Including KeyboardInterrupt
890 self
.kill(timeout
=None)
893 def kill(self
, *, timeout
=0):
896 self
.wait(timeout
=timeout
)
899 def run(cls
, *args
, timeout
=None, **kwargs
):
900 with cls(*args
, **kwargs
) as proc
:
901 default
= '' if proc
.text_mode
else b
''
902 stdout
, stderr
= proc
.communicate_or_kill(timeout
=timeout
)
903 return stdout
or default
, stderr
or default
, proc
.returncode
906 def get_subprocess_encoding():
907 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
908 # For subprocess calls, encode with locale encoding
909 # Refer to http://stackoverflow.com/a/9951851/35070
910 encoding
= preferredencoding()
912 encoding
= sys
.getfilesystemencoding()
918 def encodeFilename(s
, for_subprocess
=False):
919 assert isinstance(s
, str)
923 def decodeFilename(b
, for_subprocess
=False):
927 def encodeArgument(s
):
928 # Legacy code that uses byte strings
929 # Uncomment the following line after fixing all post processors
930 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, str, type(s))
931 return s
if isinstance(s
, str) else s
.decode('ascii')
934 def decodeArgument(b
):
938 def decodeOption(optval
):
941 if isinstance(optval
, bytes):
942 optval
= optval
.decode(preferredencoding())
944 assert isinstance(optval
, str)
948 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
951 def timetuple_from_msec(msec
):
952 secs
, msec
= divmod(msec
, 1000)
953 mins
, secs
= divmod(secs
, 60)
954 hrs
, mins
= divmod(mins
, 60)
955 return _timetuple(hrs
, mins
, secs
, msec
)
958 def formatSeconds(secs
, delim
=':', msec
=False):
959 time
= timetuple_from_msec(secs
* 1000)
961 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
963 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
965 ret
= '%d' % time
.seconds
966 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
969 def _ssl_load_windows_store_certs(ssl_context
, storename
):
970 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
972 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
973 if encoding
== 'x509_asn' and (
974 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
975 except PermissionError
:
978 with contextlib
.suppress(ssl
.SSLError
):
979 ssl_context
.load_verify_locations(cadata
=cert
)
982 def make_HTTPS_handler(params
, **kwargs
):
983 opts_check_certificate
= not params
.get('nocheckcertificate')
984 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
985 context
.check_hostname
= opts_check_certificate
986 if params
.get('legacyserverconnect'):
987 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
988 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
989 context
.set_ciphers('DEFAULT')
991 sys
.version_info
< (3, 10)
992 and ssl
.OPENSSL_VERSION_INFO
>= (1, 1, 1)
993 and not ssl
.OPENSSL_VERSION
.startswith('LibreSSL')
995 # Backport the default SSL ciphers and minimum TLS version settings from Python 3.10 [1].
996 # This is to ensure consistent behavior across Python versions, and help avoid fingerprinting
997 # in some situations [2][3].
998 # Python 3.10 only supports OpenSSL 1.1.1+ [4]. Because this change is likely
999 # untested on older versions, we only apply this to OpenSSL 1.1.1+ to be safe.
1000 # LibreSSL is excluded until further investigation due to cipher support issues [5][6].
1001 # 1. https://github.com/python/cpython/commit/e983252b516edb15d4338b0a47631b59ef1e2536
1002 # 2. https://github.com/yt-dlp/yt-dlp/issues/4627
1003 # 3. https://github.com/yt-dlp/yt-dlp/pull/5294
1004 # 4. https://peps.python.org/pep-0644/
1005 # 5. https://peps.python.org/pep-0644/#libressl-support
1006 # 6. https://github.com/yt-dlp/yt-dlp/commit/5b9f253fa0aee996cf1ed30185d4b502e00609c4#commitcomment-89054368
1007 context
.set_ciphers('@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM')
1008 context
.minimum_version
= ssl
.TLSVersion
.TLSv1_2
1010 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1011 if opts_check_certificate
:
1012 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
1013 context
.load_verify_locations(cafile
=certifi
.where())
1016 context
.load_default_certs()
1017 # Work around the issue in load_default_certs when there are bad certificates. See:
1018 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1019 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1020 except ssl
.SSLError
:
1021 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1022 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1023 for storename
in ('CA', 'ROOT'):
1024 _ssl_load_windows_store_certs(context
, storename
)
1025 context
.set_default_verify_paths()
1027 client_certfile
= params
.get('client_certificate')
1030 context
.load_cert_chain(
1031 client_certfile
, keyfile
=params
.get('client_certificate_key'),
1032 password
=params
.get('client_certificate_password'))
1033 except ssl
.SSLError
:
1034 raise YoutubeDLError('Unable to load client certificate')
1036 # Some servers may reject requests if ALPN extension is not sent. See:
1037 # https://github.com/python/cpython/issues/85140
1038 # https://github.com/yt-dlp/yt-dlp/issues/3878
1039 with contextlib
.suppress(NotImplementedError):
1040 context
.set_alpn_protocols(['http/1.1'])
1042 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1045 def bug_reports_message(before
=';'):
1046 from .update
import REPOSITORY
1048 msg
= (f
'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
1049 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
1051 before
= before
.rstrip()
1052 if not before
or before
.endswith(('.', '!', '?')):
1053 msg
= msg
[0].title() + msg
[1:]
1055 return (before
+ ' ' if before
else '') + msg
1058 class YoutubeDLError(Exception):
1059 """Base exception for YoutubeDL errors."""
1062 def __init__(self
, msg
=None):
1065 elif self
.msg
is None:
1066 self
.msg
= type(self
).__name
__
1067 super().__init
__(self
.msg
)
1070 network_exceptions
= [urllib
.error
.URLError
, http
.client
.HTTPException
, socket
.error
]
1071 if hasattr(ssl
, 'CertificateError'):
1072 network_exceptions
.append(ssl
.CertificateError
)
1073 network_exceptions
= tuple(network_exceptions
)
1076 class ExtractorError(YoutubeDLError
):
1077 """Error during info extraction."""
1079 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1080 """ tb, if given, is the original traceback (so that it can be printed out).
1081 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1083 if sys
.exc_info()[0] in network_exceptions
:
1086 self
.orig_msg
= str(msg
)
1088 self
.expected
= expected
1090 self
.video_id
= video_id
1092 self
.exc_info
= sys
.exc_info() # preserve original exception
1093 if isinstance(self
.exc_info
[1], ExtractorError
):
1094 self
.exc_info
= self
.exc_info
[1].exc_info
1096 super().__init
__(''.join((
1097 format_field(ie
, None, '[%s] '),
1098 format_field(video_id
, None, '%s: '),
1100 format_field(cause
, None, ' (caused by %r)'),
1101 '' if expected
else bug_reports_message())))
1103 def format_traceback(self
):
1104 return join_nonempty(
1105 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1106 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1110 class UnsupportedError(ExtractorError
):
1111 def __init__(self
, url
):
1113 'Unsupported URL: %s' % url
, expected
=True)
1117 class RegexNotFoundError(ExtractorError
):
1118 """Error when a regex didn't match"""
1122 class GeoRestrictedError(ExtractorError
):
1123 """Geographic restriction Error exception.
1125 This exception may be thrown when a video is not available from your
1126 geographic location due to geographic restrictions imposed by a website.
1129 def __init__(self
, msg
, countries
=None, **kwargs
):
1130 kwargs
['expected'] = True
1131 super().__init
__(msg
, **kwargs
)
1132 self
.countries
= countries
1135 class UserNotLive(ExtractorError
):
1136 """Error when a channel/user is not live"""
1138 def __init__(self
, msg
=None, **kwargs
):
1139 kwargs
['expected'] = True
1140 super().__init
__(msg
or 'The channel is not currently live', **kwargs
)
1143 class DownloadError(YoutubeDLError
):
1144 """Download Error exception.
1146 This exception may be thrown by FileDownloader objects if they are not
1147 configured to continue on errors. They will contain the appropriate
1151 def __init__(self
, msg
, exc_info
=None):
1152 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1153 super().__init
__(msg
)
1154 self
.exc_info
= exc_info
1157 class EntryNotInPlaylist(YoutubeDLError
):
1158 """Entry not in playlist exception.
1160 This exception will be thrown by YoutubeDL when a requested entry
1161 is not found in the playlist info_dict
1163 msg
= 'Entry not found in info'
1166 class SameFileError(YoutubeDLError
):
1167 """Same File exception.
1169 This exception will be thrown by FileDownloader objects if they detect
1170 multiple files would have to be downloaded to the same file on disk.
1172 msg
= 'Fixed output name but more than one file to download'
1174 def __init__(self
, filename
=None):
1175 if filename
is not None:
1176 self
.msg
+= f
': {filename}'
1177 super().__init
__(self
.msg
)
1180 class PostProcessingError(YoutubeDLError
):
1181 """Post Processing exception.
1183 This exception may be raised by PostProcessor's .run() method to
1184 indicate an error in the postprocessing task.
1188 class DownloadCancelled(YoutubeDLError
):
1189 """ Exception raised when the download queue should be interrupted """
1190 msg
= 'The download was cancelled'
1193 class ExistingVideoReached(DownloadCancelled
):
1194 """ --break-on-existing triggered """
1195 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1198 class RejectedVideoReached(DownloadCancelled
):
1199 """ --break-on-reject triggered """
1200 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1203 class MaxDownloadsReached(DownloadCancelled
):
1204 """ --max-downloads limit has been reached. """
1205 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1208 class ReExtractInfo(YoutubeDLError
):
1209 """ Video info needs to be re-extracted. """
1211 def __init__(self
, msg
, expected
=False):
1212 super().__init
__(msg
)
1213 self
.expected
= expected
1216 class ThrottledDownload(ReExtractInfo
):
1217 """ Download speed below --throttled-rate. """
1218 msg
= 'The download speed is below throttle limit'
1221 super().__init
__(self
.msg
, expected
=False)
1224 class UnavailableVideoError(YoutubeDLError
):
1225 """Unavailable Format exception.
1227 This exception will be thrown when a video is requested
1228 in a format that is not available for that video.
1230 msg
= 'Unable to download video'
1232 def __init__(self
, err
=None):
1234 self
.msg
+= f
': {err}'
1235 super().__init
__(self
.msg
)
1238 class ContentTooShortError(YoutubeDLError
):
1239 """Content Too Short exception.
1241 This exception may be raised by FileDownloader objects when a file they
1242 download is too small for what the server announced first, indicating
1243 the connection was probably interrupted.
1246 def __init__(self
, downloaded
, expected
):
1247 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1249 self
.downloaded
= downloaded
1250 self
.expected
= expected
1253 class XAttrMetadataError(YoutubeDLError
):
1254 def __init__(self
, code
=None, msg
='Unknown error'):
1255 super().__init
__(msg
)
1259 # Parsing code and msg
1260 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1261 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1262 self
.reason
= 'NO_SPACE'
1263 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1264 self
.reason
= 'VALUE_TOO_LONG'
1266 self
.reason
= 'NOT_SUPPORTED'
1269 class XAttrUnavailableError(YoutubeDLError
):
1273 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1274 hc
= http_class(*args
, **kwargs
)
1275 source_address
= ydl_handler
._params
.get('source_address')
1277 if source_address
is not None:
1278 # This is to workaround _create_connection() from socket where it will try all
1279 # address data from getaddrinfo() including IPv6. This filters the result from
1280 # getaddrinfo() based on the source_address value.
1281 # This is based on the cpython socket.create_connection() function.
1282 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1283 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1284 host
, port
= address
1286 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1287 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1288 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1289 if addrs
and not ip_addrs
:
1290 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1292 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1293 % (ip_version
, source_address
[0]))
1294 for res
in ip_addrs
:
1295 af
, socktype
, proto
, canonname
, sa
= res
1298 sock
= socket
.socket(af
, socktype
, proto
)
1299 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1300 sock
.settimeout(timeout
)
1301 sock
.bind(source_address
)
1303 err
= None # Explicitly break reference cycle
1305 except OSError as _
:
1307 if sock
is not None:
1312 raise OSError('getaddrinfo returns an empty list')
1313 if hasattr(hc
, '_create_connection'):
1314 hc
._create
_connection
= _create_connection
1315 hc
.source_address
= (source_address
, 0)
1320 def handle_youtubedl_headers(headers
):
1321 filtered_headers
= headers
1323 if 'Youtubedl-no-compression' in filtered_headers
:
1324 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1325 del filtered_headers
['Youtubedl-no-compression']
1327 return filtered_headers
1330 class YoutubeDLHandler(urllib
.request
.HTTPHandler
):
1331 """Handler for HTTP requests and responses.
1333 This class, when installed with an OpenerDirector, automatically adds
1334 the standard headers to every HTTP request and handles gzipped and
1335 deflated responses from web servers. If compression is to be avoided in
1336 a particular request, the original request in the program code only has
1337 to include the HTTP header "Youtubedl-no-compression", which will be
1338 removed before making the real request.
1340 Part of this code was copied from:
1342 http://techknack.net/python-urllib2-handlers/
1344 Andrew Rowls, the author of that code, agreed to release it to the
1348 def __init__(self
, params
, *args
, **kwargs
):
1349 urllib
.request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1350 self
._params
= params
1352 def http_open(self
, req
):
1353 conn_class
= http
.client
.HTTPConnection
1355 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1357 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1358 del req
.headers
['Ytdl-socks-proxy']
1360 return self
.do_open(functools
.partial(
1361 _create_http_connection
, self
, conn_class
, False),
1369 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1371 return zlib
.decompress(data
)
1377 return brotli
.decompress(data
)
1379 def http_request(self
, req
):
1380 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1381 # always respected by websites, some tend to give out URLs with non percent-encoded
1382 # non-ASCII characters (see telemb.py, ard.py [#3412])
1383 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1384 # To work around aforementioned issue we will replace request's original URL with
1385 # percent-encoded one
1386 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1387 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1388 url
= req
.get_full_url()
1389 url_escaped
= escape_url(url
)
1391 # Substitute URL if any change after escaping
1392 if url
!= url_escaped
:
1393 req
= update_Request(req
, url
=url_escaped
)
1395 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1396 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1397 # The dict keys are capitalized because of this bug by urllib
1398 if h
.capitalize() not in req
.headers
:
1399 req
.add_header(h
, v
)
1401 if 'Accept-encoding' not in req
.headers
:
1402 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1404 req
.headers
= handle_youtubedl_headers(req
.headers
)
1406 return super().do_request_(req
)
1408 def http_response(self
, req
, resp
):
1411 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1412 content
= resp
.read()
1413 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1415 uncompressed
= io
.BytesIO(gz
.read())
1416 except OSError as original_ioerror
:
1417 # There may be junk add the end of the file
1418 # See http://stackoverflow.com/q/4928560/35070 for details
1419 for i
in range(1, 1024):
1421 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1422 uncompressed
= io
.BytesIO(gz
.read())
1427 raise original_ioerror
1428 resp
= urllib
.request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1429 resp
.msg
= old_resp
.msg
1430 del resp
.headers
['Content-encoding']
1432 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1433 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1434 resp
= urllib
.request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1435 resp
.msg
= old_resp
.msg
1436 del resp
.headers
['Content-encoding']
1438 if resp
.headers
.get('Content-encoding', '') == 'br':
1439 resp
= urllib
.request
.addinfourl(
1440 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1441 resp
.msg
= old_resp
.msg
1442 del resp
.headers
['Content-encoding']
1443 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1444 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1445 if 300 <= resp
.code
< 400:
1446 location
= resp
.headers
.get('Location')
1448 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1449 location
= location
.encode('iso-8859-1').decode()
1450 location_escaped
= escape_url(location
)
1451 if location
!= location_escaped
:
1452 del resp
.headers
['Location']
1453 resp
.headers
['Location'] = location_escaped
1456 https_request
= http_request
1457 https_response
= http_response
1460 def make_socks_conn_class(base_class
, socks_proxy
):
1461 assert issubclass(base_class
, (
1462 http
.client
.HTTPConnection
, http
.client
.HTTPSConnection
))
1464 url_components
= urllib
.parse
.urlparse(socks_proxy
)
1465 if url_components
.scheme
.lower() == 'socks5':
1466 socks_type
= ProxyType
.SOCKS5
1467 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1468 socks_type
= ProxyType
.SOCKS4
1469 elif url_components
.scheme
.lower() == 'socks4a':
1470 socks_type
= ProxyType
.SOCKS4A
1472 def unquote_if_non_empty(s
):
1475 return urllib
.parse
.unquote_plus(s
)
1479 url_components
.hostname
, url_components
.port
or 1080,
1481 unquote_if_non_empty(url_components
.username
),
1482 unquote_if_non_empty(url_components
.password
),
1485 class SocksConnection(base_class
):
1487 self
.sock
= sockssocket()
1488 self
.sock
.setproxy(*proxy_args
)
1489 if isinstance(self
.timeout
, (int, float)):
1490 self
.sock
.settimeout(self
.timeout
)
1491 self
.sock
.connect((self
.host
, self
.port
))
1493 if isinstance(self
, http
.client
.HTTPSConnection
):
1494 if hasattr(self
, '_context'): # Python > 2.6
1495 self
.sock
= self
._context
.wrap_socket(
1496 self
.sock
, server_hostname
=self
.host
)
1498 self
.sock
= ssl
.wrap_socket(self
.sock
)
1500 return SocksConnection
1503 class YoutubeDLHTTPSHandler(urllib
.request
.HTTPSHandler
):
1504 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1505 urllib
.request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1506 self
._https
_conn
_class
= https_conn_class
or http
.client
.HTTPSConnection
1507 self
._params
= params
1509 def https_open(self
, req
):
1511 conn_class
= self
._https
_conn
_class
1513 if hasattr(self
, '_context'): # python > 2.6
1514 kwargs
['context'] = self
._context
1515 if hasattr(self
, '_check_hostname'): # python 3.x
1516 kwargs
['check_hostname'] = self
._check
_hostname
1518 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1520 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1521 del req
.headers
['Ytdl-socks-proxy']
1524 return self
.do_open(
1525 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1526 except urllib
.error
.URLError
as e
:
1527 if (isinstance(e
.reason
, ssl
.SSLError
)
1528 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1529 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1533 def is_path_like(f
):
1534 return isinstance(f
, (str, bytes, os
.PathLike
))
1537 class YoutubeDLCookieJar(http
.cookiejar
.MozillaCookieJar
):
1539 See [1] for cookie file format.
1541 1. https://curl.haxx.se/docs/http-cookies.html
1543 _HTTPONLY_PREFIX
= '#HttpOnly_'
1545 _HEADER
= '''# Netscape HTTP Cookie File
1546 # This file is generated by yt-dlp. Do not edit.
1549 _CookieFileEntry
= collections
.namedtuple(
1551 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1553 def __init__(self
, filename
=None, *args
, **kwargs
):
1554 super().__init
__(None, *args
, **kwargs
)
1555 if is_path_like(filename
):
1556 filename
= os
.fspath(filename
)
1557 self
.filename
= filename
1560 def _true_or_false(cndn
):
1561 return 'TRUE' if cndn
else 'FALSE'
1563 @contextlib.contextmanager
1564 def open(self
, file, *, write
=False):
1565 if is_path_like(file):
1566 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1573 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1576 if (not ignore_discard
and cookie
.discard
1577 or not ignore_expires
and cookie
.is_expired(now
)):
1579 name
, value
= cookie
.name
, cookie
.value
1581 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1582 # with no name, whereas http.cookiejar regards it as a
1583 # cookie with no value.
1584 name
, value
= '', name
1585 f
.write('%s\n' % '\t'.join((
1587 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1589 self
._true
_or
_false
(cookie
.secure
),
1590 str_or_none(cookie
.expires
, default
=''),
1594 def save(self
, filename
=None, *args
, **kwargs
):
1596 Save cookies to a file.
1597 Code is taken from CPython 3.6
1598 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1600 if filename
is None:
1601 if self
.filename
is not None:
1602 filename
= self
.filename
1604 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1606 # Store session cookies with `expires` set to 0 instead of an empty string
1608 if cookie
.expires
is None:
1611 with self
.open(filename
, write
=True) as f
:
1612 f
.write(self
._HEADER
)
1613 self
._really
_save
(f
, *args
, **kwargs
)
1615 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1616 """Load cookies from a file."""
1617 if filename
is None:
1618 if self
.filename
is not None:
1619 filename
= self
.filename
1621 raise ValueError(http
.cookiejar
.MISSING_FILENAME_TEXT
)
1623 def prepare_line(line
):
1624 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1625 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1626 # comments and empty lines are fine
1627 if line
.startswith('#') or not line
.strip():
1629 cookie_list
= line
.split('\t')
1630 if len(cookie_list
) != self
._ENTRY
_LEN
:
1631 raise http
.cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1632 cookie
= self
._CookieFileEntry
(*cookie_list
)
1633 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1634 raise http
.cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1638 with self
.open(filename
) as f
:
1641 cf
.write(prepare_line(line
))
1642 except http
.cookiejar
.LoadError
as e
:
1643 if f
'{line.strip()} '[0] in '[{"':
1644 raise http
.cookiejar
.LoadError(
1645 'Cookies file must be Netscape formatted, not JSON. See '
1646 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp')
1647 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1650 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1651 # Session cookies are denoted by either `expires` field set to
1652 # an empty string or 0. MozillaCookieJar only recognizes the former
1653 # (see [1]). So we need force the latter to be recognized as session
1654 # cookies on our own.
1655 # Session cookies may be important for cookies-based authentication,
1656 # e.g. usually, when user does not check 'Remember me' check box while
1657 # logging in on a site, some important cookies are stored as session
1658 # cookies so that not recognizing them will result in failed login.
1659 # 1. https://bugs.python.org/issue17164
1661 # Treat `expires=0` cookies as session cookies
1662 if cookie
.expires
== 0:
1663 cookie
.expires
= None
1664 cookie
.discard
= True
1667 class YoutubeDLCookieProcessor(urllib
.request
.HTTPCookieProcessor
):
1668 def __init__(self
, cookiejar
=None):
1669 urllib
.request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1671 def http_response(self
, request
, response
):
1672 return urllib
.request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1674 https_request
= urllib
.request
.HTTPCookieProcessor
.http_request
1675 https_response
= http_response
1678 class YoutubeDLRedirectHandler(urllib
.request
.HTTPRedirectHandler
):
1679 """YoutubeDL redirect handler
1681 The code is based on HTTPRedirectHandler implementation from CPython [1].
1683 This redirect handler solves two issues:
1684 - ensures redirect URL is always unicode under python 2
1685 - introduces support for experimental HTTP response status code
1686 308 Permanent Redirect [2] used by some sites [3]
1688 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1689 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1690 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1693 http_error_301
= http_error_303
= http_error_307
= http_error_308
= urllib
.request
.HTTPRedirectHandler
.http_error_302
1695 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1696 """Return a Request or None in response to a redirect.
1698 This is called by the http_error_30x methods when a
1699 redirection response is received. If a redirection should
1700 take place, return a new Request to allow http_error_30x to
1701 perform the redirect. Otherwise, raise HTTPError if no-one
1702 else should try to handle this url. Return None if you can't
1703 but another Handler might.
1705 m
= req
.get_method()
1706 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1707 or code
in (301, 302, 303) and m
== "POST")):
1708 raise urllib
.error
.HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1709 # Strictly (according to RFC 2616), 301 or 302 in response to
1710 # a POST MUST NOT cause a redirection without confirmation
1711 # from the user (of urllib.request, in this case). In practice,
1712 # essentially all clients do redirect in this case, so we do
1715 # Be conciliant with URIs containing a space. This is mainly
1716 # redundant with the more complete encoding done in http_error_302(),
1717 # but it is kept for compatibility with other callers.
1718 newurl
= newurl
.replace(' ', '%20')
1720 CONTENT_HEADERS
= ("content-length", "content-type")
1721 # NB: don't use dict comprehension for python 2.6 compatibility
1722 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1724 # A 303 must either use GET or HEAD for subsequent request
1725 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1726 if code
== 303 and m
!= 'HEAD':
1728 # 301 and 302 redirects are commonly turned into a GET from a POST
1729 # for subsequent requests by browsers, so we'll do the same.
1730 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1731 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1732 if code
in (301, 302) and m
== 'POST':
1735 return urllib
.request
.Request(
1736 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1737 unverifiable
=True, method
=m
)
1740 def extract_timezone(date_str
):
1743 ^.{8,}? # >=8 char non-TZ prefix, if present
1744 (?P<tz>Z| # just the UTC Z, or
1745 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1746 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1747 [ ]? # optional space
1748 (?P<sign>\+|-) # +/-
1749 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1753 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1754 timezone
= TIMEZONE_NAMES
.get(m
and m
.group('tz').strip())
1755 if timezone
is not None:
1756 date_str
= date_str
[:-len(m
.group('tz'))]
1757 timezone
= datetime
.timedelta(hours
=timezone
or 0)
1759 date_str
= date_str
[:-len(m
.group('tz'))]
1760 if not m
.group('sign'):
1761 timezone
= datetime
.timedelta()
1763 sign
= 1 if m
.group('sign') == '+' else -1
1764 timezone
= datetime
.timedelta(
1765 hours
=sign
* int(m
.group('hours')),
1766 minutes
=sign
* int(m
.group('minutes')))
1767 return timezone
, date_str
1770 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1771 """ Return a UNIX timestamp from the given date """
1773 if date_str
is None:
1776 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1778 if timezone
is None:
1779 timezone
, date_str
= extract_timezone(date_str
)
1781 with contextlib
.suppress(ValueError):
1782 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1783 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1784 return calendar
.timegm(dt
.timetuple())
1787 def date_formats(day_first
=True):
1788 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1791 def unified_strdate(date_str
, day_first
=True):
1792 """Return a string with the date in the format YYYYMMDD"""
1794 if date_str
is None:
1798 date_str
= date_str
.replace(',', ' ')
1799 # Remove AM/PM + timezone
1800 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1801 _
, date_str
= extract_timezone(date_str
)
1803 for expression
in date_formats(day_first
):
1804 with contextlib
.suppress(ValueError):
1805 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1806 if upload_date
is None:
1807 timetuple
= email
.utils
.parsedate_tz(date_str
)
1809 with contextlib
.suppress(ValueError):
1810 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1811 if upload_date
is not None:
1812 return str(upload_date
)
1815 def unified_timestamp(date_str
, day_first
=True):
1816 if date_str
is None:
1819 date_str
= re
.sub(r
'\s+', ' ', re
.sub(
1820 r
'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str
))
1822 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1823 timezone
, date_str
= extract_timezone(date_str
)
1825 # Remove AM/PM + timezone
1826 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1828 # Remove unrecognized timezones from ISO 8601 alike timestamps
1829 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1831 date_str
= date_str
[:-len(m
.group('tz'))]
1833 # Python only supports microseconds, so remove nanoseconds
1834 m
= re
.search(r
'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str
)
1836 date_str
= m
.group(1)
1838 for expression
in date_formats(day_first
):
1839 with contextlib
.suppress(ValueError):
1840 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1841 return calendar
.timegm(dt
.timetuple())
1843 timetuple
= email
.utils
.parsedate_tz(date_str
)
1845 return calendar
.timegm(timetuple
) + pm_delta
* 3600 - timezone
.total_seconds()
1848 def determine_ext(url
, default_ext
='unknown_video'):
1849 if url
is None or '.' not in url
:
1851 guess
= url
.partition('?')[0].rpartition('.')[2]
1852 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1854 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1855 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1856 return guess
.rstrip('/')
1861 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1862 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1865 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1867 Return a datetime object from a string.
1869 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1871 @param format strftime format of DATE
1872 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1873 auto: round to the unit provided in date_str (if applicable).
1875 auto_precision
= False
1876 if precision
== 'auto':
1877 auto_precision
= True
1878 precision
= 'microsecond'
1879 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1880 if date_str
in ('now', 'today'):
1882 if date_str
== 'yesterday':
1883 return today
- datetime
.timedelta(days
=1)
1885 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1887 if match
is not None:
1888 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1889 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1890 unit
= match
.group('unit')
1891 if unit
== 'month' or unit
== 'year':
1892 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1898 delta
= datetime
.timedelta(**{unit + 's': time}
)
1899 new_date
= start_time
+ delta
1901 return datetime_round(new_date
, unit
)
1904 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1907 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1909 Return a date object from a string using datetime_from_str
1911 @param strict Restrict allowed patterns to "YYYYMMDD" and
1912 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1914 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1915 raise ValueError(f
'Invalid date format "{date_str}"')
1916 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1919 def datetime_add_months(dt
, months
):
1920 """Increment/Decrement a datetime object by months."""
1921 month
= dt
.month
+ months
- 1
1922 year
= dt
.year
+ month
// 12
1923 month
= month
% 12 + 1
1924 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1925 return dt
.replace(year
, month
, day
)
1928 def datetime_round(dt
, precision
='day'):
1930 Round a datetime object's time to a specific precision
1932 if precision
== 'microsecond':
1941 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1942 timestamp
= calendar
.timegm(dt
.timetuple())
1943 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1946 def hyphenate_date(date_str
):
1948 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1949 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1950 if match
is not None:
1951 return '-'.join(match
.groups())
1957 """Represents a time interval between two dates"""
1959 def __init__(self
, start
=None, end
=None):
1960 """start and end must be strings in the format accepted by date"""
1961 if start
is not None:
1962 self
.start
= date_from_str(start
, strict
=True)
1964 self
.start
= datetime
.datetime
.min.date()
1966 self
.end
= date_from_str(end
, strict
=True)
1968 self
.end
= datetime
.datetime
.max.date()
1969 if self
.start
> self
.end
:
1970 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1974 """Returns a range that only contains the given day"""
1975 return cls(day
, day
)
1977 def __contains__(self
, date
):
1978 """Check if the date is in the range"""
1979 if not isinstance(date
, datetime
.date
):
1980 date
= date_from_str(date
)
1981 return self
.start
<= date
<= self
.end
1984 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1986 def __eq__(self
, other
):
1987 return (isinstance(other
, DateRange
)
1988 and self
.start
== other
.start
and self
.end
== other
.end
)
1991 def platform_name():
1992 """ Returns the platform name as a str """
1993 deprecation_warning(f
'"{__name__}.platform_name" is deprecated, use "platform.platform" instead')
1994 return platform
.platform()
1998 def system_identifier():
1999 python_implementation
= platform
.python_implementation()
2000 if python_implementation
== 'PyPy' and hasattr(sys
, 'pypy_version_info'):
2001 python_implementation
+= ' version %d.%d.%d' % sys
.pypy_version_info
[:3]
2003 with contextlib
.suppress(OSError): # We may not have access to the executable
2004 libc_ver
= platform
.libc_ver()
2006 return 'Python %s (%s %s) - %s (%s%s)' % (
2007 platform
.python_version(),
2008 python_implementation
,
2009 platform
.architecture()[0],
2010 platform
.platform(),
2011 ssl
.OPENSSL_VERSION
,
2012 format_field(join_nonempty(*libc_ver
, delim
=' '), None, ', %s'),
2017 def get_windows_version():
2018 ''' Get Windows version. returns () if it's not running on Windows '''
2019 if compat_os_name
== 'nt':
2020 return version_tuple(platform
.win32_ver()[1])
2025 def write_string(s
, out
=None, encoding
=None):
2026 assert isinstance(s
, str)
2027 out
= out
or sys
.stderr
2029 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
2030 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
2032 enc
, buffer = None, out
2033 if 'b' in getattr(out
, 'mode', ''):
2034 enc
= encoding
or preferredencoding()
2035 elif hasattr(out
, 'buffer'):
2037 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2039 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
2043 def deprecation_warning(msg
, *, printer
=None, stacklevel
=0, **kwargs
):
2044 from . import _IN_CLI
2046 if msg
in deprecation_warning
._cache
:
2048 deprecation_warning
._cache
.add(msg
)
2050 return printer(f
'{msg}{bug_reports_message()}', **kwargs
)
2051 return write_string(f
'ERROR: {msg}{bug_reports_message()}\n', **kwargs
)
2054 warnings
.warn(DeprecationWarning(msg
), stacklevel
=stacklevel
+ 3)
2057 deprecation_warning
._cache
= set()
2060 def bytes_to_intlist(bs
):
2063 if isinstance(bs
[0], int): # Python 3
2066 return [ord(c
) for c
in bs
]
2069 def intlist_to_bytes(xs
):
2072 return struct
.pack('%dB' % len(xs
), *xs
)
2075 class LockingUnsupportedError(OSError):
2076 msg
= 'File locking is not supported'
2079 super().__init
__(self
.msg
)
2082 # Cross-platform file locking
2083 if sys
.platform
== 'win32':
2085 import ctypes
.wintypes
2088 class OVERLAPPED(ctypes
.Structure
):
2090 ('Internal', ctypes
.wintypes
.LPVOID
),
2091 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2092 ('Offset', ctypes
.wintypes
.DWORD
),
2093 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2094 ('hEvent', ctypes
.wintypes
.HANDLE
),
2097 kernel32
= ctypes
.windll
.kernel32
2098 LockFileEx
= kernel32
.LockFileEx
2099 LockFileEx
.argtypes
= [
2100 ctypes
.wintypes
.HANDLE
, # hFile
2101 ctypes
.wintypes
.DWORD
, # dwFlags
2102 ctypes
.wintypes
.DWORD
, # dwReserved
2103 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2104 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2105 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2107 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2108 UnlockFileEx
= kernel32
.UnlockFileEx
2109 UnlockFileEx
.argtypes
= [
2110 ctypes
.wintypes
.HANDLE
, # hFile
2111 ctypes
.wintypes
.DWORD
, # dwReserved
2112 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2113 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2114 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2116 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2117 whole_low
= 0xffffffff
2118 whole_high
= 0x7fffffff
2120 def _lock_file(f
, exclusive
, block
):
2121 overlapped
= OVERLAPPED()
2122 overlapped
.Offset
= 0
2123 overlapped
.OffsetHigh
= 0
2124 overlapped
.hEvent
= 0
2125 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2127 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2128 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2129 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2130 # NB: No argument form of "ctypes.FormatError" does not work on PyPy
2131 raise BlockingIOError(f
'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
2133 def _unlock_file(f
):
2134 assert f
._lock
_file
_overlapped
_p
2135 handle
= msvcrt
.get_osfhandle(f
.fileno())
2136 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2137 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2143 def _lock_file(f
, exclusive
, block
):
2144 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2146 flags |
= fcntl
.LOCK_NB
2148 fcntl
.flock(f
, flags
)
2149 except BlockingIOError
:
2151 except OSError: # AOSP does not have flock()
2152 fcntl
.lockf(f
, flags
)
2154 def _unlock_file(f
):
2156 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2158 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2162 def _lock_file(f
, exclusive
, block
):
2163 raise LockingUnsupportedError()
2165 def _unlock_file(f
):
2166 raise LockingUnsupportedError()
2172 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2173 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2174 raise NotImplementedError(mode
)
2175 self
.mode
, self
.block
= mode
, block
2177 writable
= any(f
in mode
for f
in 'wax+')
2178 readable
= any(f
in mode
for f
in 'r+')
2179 flags
= functools
.reduce(operator
.ior
, (
2180 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2181 getattr(os
, 'O_BINARY', 0), # Windows only
2182 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2183 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2184 os
.O_APPEND
if 'a' in mode
else 0,
2185 os
.O_EXCL
if 'x' in mode
else 0,
2186 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2189 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2191 def __enter__(self
):
2192 exclusive
= 'r' not in self
.mode
2194 _lock_file(self
.f
, exclusive
, self
.block
)
2199 if 'w' in self
.mode
:
2202 except OSError as e
:
2204 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2205 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2214 _unlock_file(self
.f
)
2218 def __exit__(self
, *_
):
2227 def __getattr__(self
, attr
):
2228 return getattr(self
.f
, attr
)
2235 def get_filesystem_encoding():
2236 encoding
= sys
.getfilesystemencoding()
2237 return encoding
if encoding
is not None else 'utf-8'
2240 def shell_quote(args
):
2242 encoding
= get_filesystem_encoding()
2244 if isinstance(a
, bytes):
2245 # We may get a filename encoded with 'encodeFilename'
2246 a
= a
.decode(encoding
)
2247 quoted_args
.append(compat_shlex_quote(a
))
2248 return ' '.join(quoted_args
)
2251 def smuggle_url(url
, data
):
2252 """ Pass additional data in a URL for internal use. """
2254 url
, idata
= unsmuggle_url(url
, {})
2256 sdata
= urllib
.parse
.urlencode(
2257 {'__youtubedl_smuggle': json.dumps(data)}
)
2258 return url
+ '#' + sdata
2261 def unsmuggle_url(smug_url
, default
=None):
2262 if '#__youtubedl_smuggle' not in smug_url
:
2263 return smug_url
, default
2264 url
, _
, sdata
= smug_url
.rpartition('#')
2265 jsond
= urllib
.parse
.parse_qs(sdata
)['__youtubedl_smuggle'][0]
2266 data
= json
.loads(jsond
)
2270 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2271 """ Formats numbers with decimal sufixes like K, M, etc """
2272 num
, factor
= float_or_none(num
), float(factor
)
2273 if num
is None or num
< 0:
2275 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2276 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2277 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2279 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2280 converted
= num
/ (factor
** exponent
)
2281 return fmt
% (converted
, suffix
)
2284 def format_bytes(bytes):
2285 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2288 def lookup_unit_table(unit_table
, s
):
2289 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2291 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2294 num_str
= m
.group('num').replace(',', '.')
2295 mult
= unit_table
[m
.group('unit')]
2296 return int(float(num_str
) * mult
)
2299 def parse_filesize(s
):
2303 # The lower-case forms are of course incorrect and unofficial,
2304 # but we support those too
2321 'megabytes': 1000 ** 2,
2322 'mebibytes': 1024 ** 2,
2328 'gigabytes': 1000 ** 3,
2329 'gibibytes': 1024 ** 3,
2335 'terabytes': 1000 ** 4,
2336 'tebibytes': 1024 ** 4,
2342 'petabytes': 1000 ** 5,
2343 'pebibytes': 1024 ** 5,
2349 'exabytes': 1000 ** 6,
2350 'exbibytes': 1024 ** 6,
2356 'zettabytes': 1000 ** 7,
2357 'zebibytes': 1024 ** 7,
2363 'yottabytes': 1000 ** 8,
2364 'yobibytes': 1024 ** 8,
2367 return lookup_unit_table(_UNIT_TABLE
, s
)
2374 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2376 if re
.match(r
'^[\d,.]+$', s
):
2377 return str_to_int(s
)
2390 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2394 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2396 return str_to_int(mobj
.group(1))
2399 def parse_resolution(s
, *, lenient
=False):
2404 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2406 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2409 'width': int(mobj
.group('w')),
2410 'height': int(mobj
.group('h')),
2413 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2415 return {'height': int(mobj.group(1))}
2417 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2419 return {'height': int(mobj.group(1)) * 540}
2424 def parse_bitrate(s
):
2425 if not isinstance(s
, str):
2427 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2429 return int(mobj
.group(1))
2432 def month_by_name(name
, lang
='en'):
2433 """ Return the number of a month by (locale-independently) English name """
2435 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2438 return month_names
.index(name
) + 1
2443 def month_by_abbreviation(abbrev
):
2444 """ Return the number of a month by (locale-independently) English
2448 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2453 def fix_xml_ampersands(xml_str
):
2454 """Replace all the '&' by '&' in XML"""
2456 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2461 def setproctitle(title
):
2462 assert isinstance(title
, str)
2464 # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4541
2471 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2475 # LoadLibrary in Windows Python 2.7.13 only expects
2476 # a bytestring, but since unicode_literals turns
2477 # every string into a unicode string, it fails.
2479 title_bytes
= title
.encode()
2480 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2481 buf
.value
= title_bytes
2483 libc
.prctl(15, buf
, 0, 0, 0)
2484 except AttributeError:
2485 return # Strange libc, just skip this
2488 def remove_start(s
, start
):
2489 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2492 def remove_end(s
, end
):
2493 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2496 def remove_quotes(s
):
2497 if s
is None or len(s
) < 2:
2499 for quote
in ('"', "'", ):
2500 if s
[0] == quote
and s
[-1] == quote
:
2505 def get_domain(url
):
2507 This implementation is inconsistent, but is kept for compatibility.
2508 Use this only for "webpage_url_domain"
2510 return remove_start(urllib
.parse
.urlparse(url
).netloc
, 'www.') or None
2513 def url_basename(url
):
2514 path
= urllib
.parse
.urlparse(url
).path
2515 return path
.strip('/').split('/')[-1]
2519 return re
.match(r
'https?://[^?#]+/', url
).group()
2522 def urljoin(base
, path
):
2523 if isinstance(path
, bytes):
2524 path
= path
.decode()
2525 if not isinstance(path
, str) or not path
:
2527 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2529 if isinstance(base
, bytes):
2530 base
= base
.decode()
2531 if not isinstance(base
, str) or not re
.match(
2532 r
'^(?:https?:)?//', base
):
2534 return urllib
.parse
.urljoin(base
, path
)
2537 class HEADRequest(urllib
.request
.Request
):
2538 def get_method(self
):
2542 class PUTRequest(urllib
.request
.Request
):
2543 def get_method(self
):
2547 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2548 if get_attr
and v
is not None:
2549 v
= getattr(v
, get_attr
, None)
2551 return int(v
) * invscale
// scale
2552 except (ValueError, TypeError, OverflowError):
2556 def str_or_none(v
, default
=None):
2557 return default
if v
is None else str(v
)
2560 def str_to_int(int_str
):
2561 """ A more relaxed version of int_or_none """
2562 if isinstance(int_str
, int):
2564 elif isinstance(int_str
, str):
2565 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2566 return int_or_none(int_str
)
2569 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2573 return float(v
) * invscale
/ scale
2574 except (ValueError, TypeError):
2578 def bool_or_none(v
, default
=None):
2579 return v
if isinstance(v
, bool) else default
2582 def strip_or_none(v
, default
=None):
2583 return v
.strip() if isinstance(v
, str) else default
2586 def url_or_none(url
):
2587 if not url
or not isinstance(url
, str):
2590 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2593 def request_to_url(req
):
2594 if isinstance(req
, urllib
.request
.Request
):
2595 return req
.get_full_url()
2600 def strftime_or_none(timestamp
, date_format
, default
=None):
2601 datetime_object
= None
2603 if isinstance(timestamp
, (int, float)): # unix timestamp
2604 # Using naive datetime here can break timestamp() in Windows
2605 # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
2606 datetime_object
= datetime
.datetime
.fromtimestamp(timestamp
, datetime
.timezone
.utc
)
2607 elif isinstance(timestamp
, str): # assume YYYYMMDD
2608 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2609 date_format
= re
.sub( # Support %s on windows
2610 r
'(?<!%)(%%)*%s', rf
'\g<1>{int(datetime_object.timestamp())}', date_format
)
2611 return datetime_object
.strftime(date_format
)
2612 except (ValueError, TypeError, AttributeError):
2616 def parse_duration(s
):
2617 if not isinstance(s
, str):
2623 days
, hours
, mins
, secs
, ms
= [None] * 5
2624 m
= re
.match(r
'''(?x)
2626 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2627 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2628 (?P<ms>[.:][0-9]+)?Z?$
2631 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2636 [0-9]+\s*y(?:ears?)?,?\s*
2639 [0-9]+\s*m(?:onths?)?,?\s*
2642 [0-9]+\s*w(?:eeks?)?,?\s*
2645 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2649 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2652 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2655 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2658 days
, hours
, mins
, secs
, ms
= m
.groups()
2660 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2662 hours
, mins
= m
.groups()
2667 ms
= ms
.replace(':', '.')
2668 return sum(float(part
or 0) * mult
for part
, mult
in (
2669 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2672 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2673 name
, real_ext
= os
.path
.splitext(filename
)
2675 f
'{name}.{ext}{real_ext}'
2676 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2677 else f
'{filename}.{ext}')
2680 def replace_extension(filename
, ext
, expected_real_ext
=None):
2681 name
, real_ext
= os
.path
.splitext(filename
)
2682 return '{}.{}'.format(
2683 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2687 def check_executable(exe
, args
=[]):
2688 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2689 args can be a list of arguments for a short output (like -version) """
2691 Popen
.run([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
)
2697 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2699 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2701 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2702 # SIGTTOU if yt-dlp is run in the background.
2703 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2704 stdout
, _
, _
= Popen
.run([encodeArgument(exe
)] + args
, text
=True,
2705 stdin
=subprocess
.PIPE
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
)
2711 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2712 assert isinstance(output
, str)
2713 if version_re
is None:
2714 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2715 m
= re
.search(version_re
, output
)
2722 def get_exe_version(exe
, args
=['--version'],
2723 version_re
=None, unrecognized
='present'):
2724 """ Returns the version of the specified executable,
2725 or False if the executable is not present """
2726 out
= _get_exe_version_output(exe
, args
)
2727 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2730 def frange(start
=0, stop
=None, step
=1):
2733 start
, stop
= 0, start
2734 sign
= [-1, 1][step
> 0] if step
else 0
2735 while sign
* start
< sign
* stop
:
2740 class LazyList(collections
.abc
.Sequence
):
2741 """Lazy immutable list from an iterable
2742 Note that slices of a LazyList are lists and not LazyList"""
2744 class IndexError(IndexError):
2747 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2748 self
._iterable
= iter(iterable
)
2749 self
._cache
= [] if _cache
is None else _cache
2750 self
._reversed
= reverse
2754 # We need to consume the entire iterable to iterate in reverse
2755 yield from self
.exhaust()
2757 yield from self
._cache
2758 for item
in self
._iterable
:
2759 self
._cache
.append(item
)
2763 self
._cache
.extend(self
._iterable
)
2764 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2768 """Evaluate the entire iterable"""
2769 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2772 def _reverse_index(x
):
2773 return None if x
is None else ~x
2775 def __getitem__(self
, idx
):
2776 if isinstance(idx
, slice):
2778 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2779 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2780 elif isinstance(idx
, int):
2782 idx
= self
._reverse
_index
(idx
)
2783 start
, stop
, step
= idx
, idx
, 0
2785 raise TypeError('indices must be integers or slices')
2786 if ((start
or 0) < 0 or (stop
or 0) < 0
2787 or (start
is None and step
< 0)
2788 or (stop
is None and step
> 0)):
2789 # We need to consume the entire iterable to be able to slice from the end
2790 # Obviously, never use this with infinite iterables
2793 return self
._cache
[idx
]
2794 except IndexError as e
:
2795 raise self
.IndexError(e
) from e
2796 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2798 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2800 return self
._cache
[idx
]
2801 except IndexError as e
:
2802 raise self
.IndexError(e
) from e
2806 self
[-1] if self
._reversed
else self
[0]
2807 except self
.IndexError:
2813 return len(self
._cache
)
2815 def __reversed__(self
):
2816 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2819 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2822 # repr and str should mimic a list. So we exhaust the iterable
2823 return repr(self
.exhaust())
2826 return repr(self
.exhaust())
2831 class IndexError(IndexError):
2835 # This is only useful for tests
2836 return len(self
.getslice())
2838 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2839 self
._pagefunc
= pagefunc
2840 self
._pagesize
= pagesize
2841 self
._pagecount
= float('inf')
2842 self
._use
_cache
= use_cache
2845 def getpage(self
, pagenum
):
2846 page_results
= self
._cache
.get(pagenum
)
2847 if page_results
is None:
2848 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2850 self
._cache
[pagenum
] = page_results
2853 def getslice(self
, start
=0, end
=None):
2854 return list(self
._getslice
(start
, end
))
2856 def _getslice(self
, start
, end
):
2857 raise NotImplementedError('This method must be implemented by subclasses')
2859 def __getitem__(self
, idx
):
2860 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2861 if not isinstance(idx
, int) or idx
< 0:
2862 raise TypeError('indices must be non-negative integers')
2863 entries
= self
.getslice(idx
, idx
+ 1)
2865 raise self
.IndexError()
2869 class OnDemandPagedList(PagedList
):
2870 """Download pages until a page with less than maximum results"""
2872 def _getslice(self
, start
, end
):
2873 for pagenum
in itertools
.count(start
// self
._pagesize
):
2874 firstid
= pagenum
* self
._pagesize
2875 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2876 if start
>= nextfirstid
:
2880 start
% self
._pagesize
2881 if firstid
<= start
< nextfirstid
2884 ((end
- 1) % self
._pagesize
) + 1
2885 if (end
is not None and firstid
<= end
<= nextfirstid
)
2889 page_results
= self
.getpage(pagenum
)
2891 self
._pagecount
= pagenum
- 1
2893 if startv
!= 0 or endv
is not None:
2894 page_results
= page_results
[startv
:endv
]
2895 yield from page_results
2897 # A little optimization - if current page is not "full", ie. does
2898 # not contain page_size videos then we can assume that this page
2899 # is the last one - there are no more ids on further pages -
2900 # i.e. no need to query again.
2901 if len(page_results
) + startv
< self
._pagesize
:
2904 # If we got the whole page, but the next page is not interesting,
2905 # break out early as well
2906 if end
== nextfirstid
:
2910 class InAdvancePagedList(PagedList
):
2911 """PagedList with total number of pages known in advance"""
2913 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2914 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2915 self
._pagecount
= pagecount
2917 def _getslice(self
, start
, end
):
2918 start_page
= start
// self
._pagesize
2919 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2920 skip_elems
= start
- start_page
* self
._pagesize
2921 only_more
= None if end
is None else end
- start
2922 for pagenum
in range(start_page
, end_page
):
2923 page_results
= self
.getpage(pagenum
)
2925 page_results
= page_results
[skip_elems
:]
2927 if only_more
is not None:
2928 if len(page_results
) < only_more
:
2929 only_more
-= len(page_results
)
2931 yield from page_results
[:only_more
]
2933 yield from page_results
2936 class PlaylistEntries
:
2937 MissingEntry
= object()
2938 is_exhausted
= False
2940 def __init__(self
, ydl
, info_dict
):
2943 # _entries must be assigned now since infodict can change during iteration
2944 entries
= info_dict
.get('entries')
2946 raise EntryNotInPlaylist('There are no entries')
2947 elif isinstance(entries
, list):
2948 self
.is_exhausted
= True
2950 requested_entries
= info_dict
.get('requested_entries')
2951 self
.is_incomplete
= bool(requested_entries
)
2952 if self
.is_incomplete
:
2953 assert self
.is_exhausted
2954 self
._entries
= [self
.MissingEntry
] * max(requested_entries
)
2955 for i
, entry
in zip(requested_entries
, entries
):
2956 self
._entries
[i
- 1] = entry
2957 elif isinstance(entries
, (list, PagedList
, LazyList
)):
2958 self
._entries
= entries
2960 self
._entries
= LazyList(entries
)
2962 PLAYLIST_ITEMS_RE
= re
.compile(r
'''(?x)
2963 (?P<start>[+-]?\d+)?
2965 (?P<end>[+-]?\d+|inf(?:inite)?)?
2966 (?::(?P<step>[+-]?\d+))?
2970 def parse_playlist_items(cls
, string
):
2971 for segment
in string
.split(','):
2973 raise ValueError('There is two or more consecutive commas')
2974 mobj
= cls
.PLAYLIST_ITEMS_RE
.fullmatch(segment
)
2976 raise ValueError(f
'{segment!r} is not a valid specification')
2977 start
, end
, step
, has_range
= mobj
.group('start', 'end', 'step', 'range')
2978 if int_or_none(step
) == 0:
2979 raise ValueError(f
'Step in {segment!r} cannot be zero')
2980 yield slice(int_or_none(start
), float_or_none(end
), int_or_none(step
)) if has_range
else int(start
)
2982 def get_requested_items(self
):
2983 playlist_items
= self
.ydl
.params
.get('playlist_items')
2984 playlist_start
= self
.ydl
.params
.get('playliststart', 1)
2985 playlist_end
= self
.ydl
.params
.get('playlistend')
2986 # For backwards compatibility, interpret -1 as whole list
2987 if playlist_end
in (-1, None):
2989 if not playlist_items
:
2990 playlist_items
= f
'{playlist_start}:{playlist_end}'
2991 elif playlist_start
!= 1 or playlist_end
:
2992 self
.ydl
.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once
=True)
2994 for index
in self
.parse_playlist_items(playlist_items
):
2995 for i
, entry
in self
[index
]:
3000 # TODO: Add auto-generated fields
3001 self
.ydl
._match
_entry
(entry
, incomplete
=True, silent
=True)
3002 except (ExistingVideoReached
, RejectedVideoReached
):
3005 def get_full_count(self
):
3006 if self
.is_exhausted
and not self
.is_incomplete
:
3008 elif isinstance(self
._entries
, InAdvancePagedList
):
3009 if self
._entries
._pagesize
== 1:
3010 return self
._entries
._pagecount
3012 @functools.cached_property
3014 if isinstance(self
._entries
, list):
3017 entry
= self
._entries
[i
]
3019 entry
= self
.MissingEntry
3020 if not self
.is_incomplete
:
3021 raise self
.IndexError()
3022 if entry
is self
.MissingEntry
:
3023 raise EntryNotInPlaylist(f
'Entry {i} cannot be found')
3028 return type(self
.ydl
)._handle
_extraction
_exceptions
(lambda _
, i
: self
._entries
[i
])(self
.ydl
, i
)
3029 except (LazyList
.IndexError, PagedList
.IndexError):
3030 raise self
.IndexError()
3033 def __getitem__(self
, idx
):
3034 if isinstance(idx
, int):
3035 idx
= slice(idx
, idx
)
3037 # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
3038 step
= 1 if idx
.step
is None else idx
.step
3039 if idx
.start
is None:
3040 start
= 0 if step
> 0 else len(self
) - 1
3042 start
= idx
.start
- 1 if idx
.start
>= 0 else len(self
) + idx
.start
3044 # NB: Do not call len(self) when idx == [:]
3045 if idx
.stop
is None:
3046 stop
= 0 if step
< 0 else float('inf')
3048 stop
= idx
.stop
- 1 if idx
.stop
>= 0 else len(self
) + idx
.stop
3049 stop
+= [-1, 1][step
> 0]
3051 for i
in frange(start
, stop
, step
):
3055 entry
= self
._getter
(i
)
3056 except self
.IndexError:
3057 self
.is_exhausted
= True
3064 return len(tuple(self
[:]))
3066 class IndexError(IndexError):
3070 def uppercase_escape(s
):
3071 unicode_escape
= codecs
.getdecoder('unicode_escape')
3073 r
'\\U[0-9a-fA-F]{8}',
3074 lambda m
: unicode_escape(m
.group(0))[0],
3078 def lowercase_escape(s
):
3079 unicode_escape
= codecs
.getdecoder('unicode_escape')
3081 r
'\\u[0-9a-fA-F]{4}',
3082 lambda m
: unicode_escape(m
.group(0))[0],
3086 def escape_rfc3986(s
):
3087 """Escape non-ASCII characters as suggested by RFC 3986"""
3088 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
3091 def escape_url(url
):
3092 """Escape URL as suggested by RFC 3986"""
3093 url_parsed
= urllib
.parse
.urlparse(url
)
3094 return url_parsed
._replace
(
3095 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
3096 path
=escape_rfc3986(url_parsed
.path
),
3097 params
=escape_rfc3986(url_parsed
.params
),
3098 query
=escape_rfc3986(url_parsed
.query
),
3099 fragment
=escape_rfc3986(url_parsed
.fragment
)
3103 def parse_qs(url
, **kwargs
):
3104 return urllib
.parse
.parse_qs(urllib
.parse
.urlparse(url
).query
, **kwargs
)
3107 def read_batch_urls(batch_fd
):
3109 if not isinstance(url
, str):
3110 url
= url
.decode('utf-8', 'replace')
3111 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
3112 for bom
in BOM_UTF8
:
3113 if url
.startswith(bom
):
3114 url
= url
[len(bom
):]
3116 if not url
or url
.startswith(('#', ';', ']')):
3118 # "#" cannot be stripped out since it is part of the URI
3119 # However, it can be safely stripped out if following a whitespace
3120 return re
.split(r
'\s#', url
, 1)[0].rstrip()
3122 with contextlib
.closing(batch_fd
) as fd
:
3123 return [url
for url
in map(fixup
, fd
) if url
]
3126 def urlencode_postdata(*args
, **kargs
):
3127 return urllib
.parse
.urlencode(*args
, **kargs
).encode('ascii')
3130 def update_url_query(url
, query
):
3133 parsed_url
= urllib
.parse
.urlparse(url
)
3134 qs
= urllib
.parse
.parse_qs(parsed_url
.query
)
3136 return urllib
.parse
.urlunparse(parsed_url
._replace
(
3137 query
=urllib
.parse
.urlencode(qs
, True)))
3140 def update_Request(req
, url
=None, data
=None, headers
=None, query
=None):
3141 req_headers
= req
.headers
.copy()
3142 req_headers
.update(headers
or {})
3143 req_data
= data
or req
.data
3144 req_url
= update_url_query(url
or req
.get_full_url(), query
)
3145 req_get_method
= req
.get_method()
3146 if req_get_method
== 'HEAD':
3147 req_type
= HEADRequest
3148 elif req_get_method
== 'PUT':
3149 req_type
= PUTRequest
3151 req_type
= urllib
.request
.Request
3153 req_url
, data
=req_data
, headers
=req_headers
,
3154 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
3155 if hasattr(req
, 'timeout'):
3156 new_req
.timeout
= req
.timeout
3160 def _multipart_encode_impl(data
, boundary
):
3161 content_type
= 'multipart/form-data; boundary=%s' % boundary
3164 for k
, v
in data
.items():
3165 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
3166 if isinstance(k
, str):
3168 if isinstance(v
, str):
3170 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
3171 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
3172 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
3173 if boundary
.encode('ascii') in content
:
3174 raise ValueError('Boundary overlaps with data')
3177 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
3179 return out
, content_type
3182 def multipart_encode(data
, boundary
=None):
3184 Encode a dict to RFC 7578-compliant form-data
3187 A dict where keys and values can be either Unicode or bytes-like
3190 If specified a Unicode object, it's used as the boundary. Otherwise
3191 a random boundary is generated.
3193 Reference: https://tools.ietf.org/html/rfc7578
3195 has_specified_boundary
= boundary
is not None
3198 if boundary
is None:
3199 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3202 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3205 if has_specified_boundary
:
3209 return out
, content_type
3212 def variadic(x
, allowed_types
=(str, bytes, dict)):
3213 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
3216 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3217 for val
in map(d
.get
, variadic(key_or_keys
)):
3218 if val
is not None and (val
or not skip_false_values
):
3223 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
3226 val
= f(*args
, **kwargs
)
3227 except (AttributeError, KeyError, TypeError, IndexError, ValueError, ZeroDivisionError):
3230 if expected_type
is None or isinstance(val
, expected_type
):
3234 def try_get(src
, getter
, expected_type
=None):
3235 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
3238 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
3239 return {k: v for k, v in dct.items() if cndn(k, v)}
3242 def merge_dicts(*dicts
):
3244 for a_dict
in dicts
:
3245 for k
, v
in a_dict
.items():
3246 if (v
is not None and k
not in merged
3247 or isinstance(v
, str) and merged
[k
] == ''):
3252 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3253 return string
if isinstance(string
, str) else str(string
, encoding
, errors
)
3265 TV_PARENTAL_GUIDELINES
= {
3275 def parse_age_limit(s
):
3276 # isinstance(False, int) is True. So type() must be used instead
3277 if type(s
) is int: # noqa: E721
3278 return s
if 0 <= s
<= 21 else None
3279 elif not isinstance(s
, str):
3281 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3283 return int(m
.group('age'))
3286 return US_RATINGS
[s
]
3287 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3289 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3293 def strip_jsonp(code
):
3296 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3297 (?:\s*&&\s*(?P=func_name))?
3298 \s*\(\s*(?P<callback_data>.*)\);?
3299 \s*?(?://[^\n]*)*$''',
3300 r
'\g<callback_data>', code
)
3303 def js_to_json(code
, vars={}, *, strict
=False):
3304 # vars is a dict of var, val pairs to substitute
3305 STRING_QUOTES
= '\'"'
3306 STRING_RE
= '|'.join(rf
'{q}(?:\\.|[^\\{q}])*{q}' for q
in STRING_QUOTES
)
3307 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3308 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3310 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3311 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3314 def process_escape(match
):
3315 JSON_PASSTHROUGH_ESCAPES
= R
'"\bfnrtu'
3316 escape
= match
.group(1) or match
.group(2)
3318 return (Rf
'\{escape}' if escape
in JSON_PASSTHROUGH_ESCAPES
3319 else R
'\u00' if escape
== 'x'
3320 else '' if escape
== '\n'
3325 if v
in ('true', 'false', 'null'):
3327 elif v
in ('undefined', 'void 0'):
3329 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3332 if v
[0] in STRING_QUOTES
:
3333 escaped
= re
.sub(r
'(?s)(")|\\(.)', process_escape
, v
[1:-1])
3334 return f
'"{escaped}"'
3336 for regex
, base
in INTEGER_TABLE
:
3337 im
= re
.match(regex
, v
)
3339 i
= int(im
.group(1), base
)
3340 return f
'"{i}":' if v
.endswith(':') else str(i
)
3343 return json
.dumps(vars[v
])
3348 raise ValueError(f
'Unknown value: {v}')
3350 def create_map(mobj
):
3351 return json
.dumps(dict(json
.loads(js_to_json(mobj
.group(1) or '[]', vars=vars))))
3353 code
= re
.sub(r
'new Map\((\[.*?\])?\)', create_map
, code
)
3355 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3356 code
= re
.sub(r
'new \w+\((.*?)\)', lambda m
: json
.dumps(m
.group(0)), code
)
3358 return re
.sub(rf
'''(?sx)
3360 {COMMENT_RE}|,(?={SKIP_RE}[\]}}])|
3361 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3362 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{SKIP_RE}:)?|
3363 [0-9]+(?={SKIP_RE}:)|
3368 def qualities(quality_ids
):
3369 """ Get a numeric quality value out of a list of possible values """
3372 return quality_ids
.index(qid
)
3378 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
3382 'default': '%(title)s [%(id)s].%(ext)s',
3383 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3389 'description': 'description',
3390 'annotation': 'annotations.xml',
3391 'infojson': 'info.json',
3394 'pl_thumbnail': None,
3395 'pl_description': 'description',
3396 'pl_infojson': 'info.json',
3399 # As of [1] format syntax is:
3400 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3401 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3402 STR_FORMAT_RE_TMPL
= r
'''(?x)
3403 (?<!%)(?P<prefix>(?:%%)*)
3405 (?P<has_key>\((?P<key>{0})\))?
3407 (?P<conversion>[#0\-+ ]+)?
3409 (?P<precision>\.\d+)?
3410 (?P<len_mod>[hlL])? # unused in python
3411 {1} # conversion type
3416 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3419 def limit_length(s
, length
):
3420 """ Add ellipses to overly long strings """
3425 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3429 def version_tuple(v
):
3430 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3433 def is_outdated_version(version
, limit
, assume_new
=True):
3435 return not assume_new
3437 return version_tuple(version
) < version_tuple(limit
)
3439 return not assume_new
3442 def ytdl_is_updateable():
3443 """ Returns if yt-dlp can be updated with -U """
3445 from .update
import is_non_updateable
3447 return not is_non_updateable()
3450 def args_to_str(args
):
3451 # Get a short string representation for a subprocess command
3452 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3455 def error_to_compat_str(err
):
3459 def error_to_str(err
):
3460 return f
'{type(err).__name__}: {err}'
3463 def mimetype2ext(mt
):
3467 mt
, _
, params
= mt
.partition(';')
3472 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3473 # it's the most popular one
3474 'audio/mpeg': 'mp3',
3475 'audio/x-wav': 'wav',
3477 'audio/wave': 'wav',
3480 ext
= FULL_MAP
.get(mt
)
3486 'smptett+xml': 'tt',
3490 'x-mp4-fragmented': 'mp4',
3491 'x-ms-sami': 'sami',
3494 'x-mpegurl': 'm3u8',
3495 'vnd.apple.mpegurl': 'm3u8',
3499 'vnd.ms-sstr+xml': 'ism',
3503 'filmstrip+json': 'fs',
3507 _
, _
, subtype
= mt
.rpartition('/')
3508 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3519 _
, _
, suffix
= subtype
.partition('+')
3520 ext
= SUFFIX_MAP
.get(suffix
)
3524 return subtype
.replace('+', '.')
3527 def ext2mimetype(ext_or_url
):
3530 if '.' not in ext_or_url
:
3531 ext_or_url
= f
'file.{ext_or_url}'
3532 return mimetypes
.guess_type(ext_or_url
)[0]
3535 def parse_codecs(codecs_str
):
3536 # http://tools.ietf.org/html/rfc6381
3539 split_codecs
= list(filter(None, map(
3540 str.strip
, codecs_str
.strip().strip(',').split(','))))
3541 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3542 for full_codec
in split_codecs
:
3543 parts
= re
.sub(r
'0+(?=\d)', '', full_codec
).split('.')
3544 if parts
[0] in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3545 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3549 if parts
[0] in ('dvh1', 'dvhe'):
3551 elif parts
[0] == 'av1' and traverse_obj(parts
, 3) == '10':
3553 elif parts
[:2] == ['vp9', '2']:
3555 elif parts
[0] in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac',
3556 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3557 acodec
= acodec
or full_codec
3558 elif parts
[0] in ('stpp', 'wvtt'):
3559 scodec
= scodec
or full_codec
3561 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3562 if vcodec
or acodec
or scodec
:
3564 'vcodec': vcodec
or 'none',
3565 'acodec': acodec
or 'none',
3566 'dynamic_range': hdr
,
3567 **({'scodec': scodec}
if scodec
is not None else {}),
3569 elif len(split_codecs
) == 2:
3571 'vcodec': split_codecs
[0],
3572 'acodec': split_codecs
[1],
3577 def get_compatible_ext(*, vcodecs
, acodecs
, vexts
, aexts
, preferences
=None):
3578 assert len(vcodecs
) == len(vexts
) and len(acodecs
) == len(aexts
)
3580 allow_mkv
= not preferences
or 'mkv' in preferences
3582 if allow_mkv
and max(len(acodecs
), len(vcodecs
)) > 1:
3583 return 'mkv' # TODO: any other format allows this?
3585 # TODO: All codecs supported by parse_codecs isn't handled here
3586 COMPATIBLE_CODECS
= {
3588 'av1', 'hevc', 'avc1', 'mp4a', # fourcc (m3u8, mpd)
3589 'h264', 'aacl', 'ec-3', # Set in ISM
3592 'av1', 'vp9', 'vp8', 'opus', 'vrbs',
3593 'vp9x', 'vp8x', # in the webm spec
3597 sanitize_codec
= functools
.partial(try_get
, getter
=lambda x
: x
[0].split('.')[0].replace('0', ''))
3598 vcodec
, acodec
= sanitize_codec(vcodecs
), sanitize_codec(acodecs
)
3600 for ext
in preferences
or COMPATIBLE_CODECS
.keys():
3601 codec_set
= COMPATIBLE_CODECS
.get(ext
, set())
3602 if ext
== 'mkv' or codec_set
.issuperset((vcodec
, acodec
)):
3606 {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma', 'mov'}
,
3609 for ext
in preferences
or vexts
:
3610 current_exts
= {ext, *vexts, *aexts}
3611 if ext
== 'mkv' or current_exts
== {ext}
or any(
3612 ext_sets
.issuperset(current_exts
) for ext_sets
in COMPATIBLE_EXTS
):
3614 return 'mkv' if allow_mkv
else preferences
[-1]
3617 def urlhandle_detect_ext(url_handle
):
3618 getheader
= url_handle
.headers
.get
3620 cd
= getheader('Content-Disposition')
3622 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3624 e
= determine_ext(m
.group('filename'), default_ext
=None)
3628 return mimetype2ext(getheader('Content-Type'))
3631 def encode_data_uri(data
, mime_type
):
3632 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3635 def age_restricted(content_limit
, age_limit
):
3636 """ Returns True iff the content should be blocked """
3638 if age_limit
is None: # No limit set
3640 if content_limit
is None:
3641 return False # Content available for everyone
3642 return age_limit
< content_limit
3645 # List of known byte-order-marks (BOM)
3647 (b
'\xef\xbb\xbf', 'utf-8'),
3648 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3649 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3650 (b
'\xff\xfe', 'utf-16-le'),
3651 (b
'\xfe\xff', 'utf-16-be'),
3655 def is_html(first_bytes
):
3656 """ Detect whether a file contains HTML by examining its first bytes. """
3659 for bom
, enc
in BOMS
:
3660 while first_bytes
.startswith(bom
):
3661 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3663 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3666 def determine_protocol(info_dict
):
3667 protocol
= info_dict
.get('protocol')
3668 if protocol
is not None:
3671 url
= sanitize_url(info_dict
['url'])
3672 if url
.startswith('rtmp'):
3674 elif url
.startswith('mms'):
3676 elif url
.startswith('rtsp'):
3679 ext
= determine_ext(url
)
3681 return 'm3u8' if info_dict
.get('is_live') else 'm3u8_native'
3685 return urllib
.parse
.urlparse(url
).scheme
3688 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3689 """ Render a list of rows, each as a list of values.
3690 Text after a \t will be right aligned """
3692 return len(remove_terminal_sequences(string
).replace('\t', ''))
3694 def get_max_lens(table
):
3695 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3697 def filter_using_list(row
, filterArray
):
3698 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3700 max_lens
= get_max_lens(data
) if hide_empty
else []
3701 header_row
= filter_using_list(header_row
, max_lens
)
3702 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3704 table
= [header_row
] + data
3705 max_lens
= get_max_lens(table
)
3708 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3709 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3711 for pos
, text
in enumerate(map(str, row
)):
3713 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3715 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3716 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3720 def _match_one(filter_part
, dct
, incomplete
):
3721 # TODO: Generalize code with YoutubeDL._build_format_filter
3722 STRING_OPERATORS
= {
3723 '*=': operator
.contains
,
3724 '^=': lambda attr
, value
: attr
.startswith(value
),
3725 '$=': lambda attr
, value
: attr
.endswith(value
),
3726 '~=': lambda attr
, value
: re
.search(value
, attr
),
3728 COMPARISON_OPERATORS
= {
3730 '<=': operator
.le
, # "<=" must be defined above "<"
3737 if isinstance(incomplete
, bool):
3738 is_incomplete
= lambda _
: incomplete
3740 is_incomplete
= lambda k
: k
in incomplete
3742 operator_rex
= re
.compile(r
'''(?x)
3744 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3746 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3749 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3750 m = operator_rex.fullmatch(filter_part.strip())
3753 unnegated_op = COMPARISON_OPERATORS[m['op']]
3755 op = lambda attr, value: not unnegated_op(attr, value)
3758 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3760 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3761 actual_value = dct.get(m['key'])
3762 numeric_comparison = None
3763 if isinstance(actual_value, (int, float)):
3764 # If the original field is a string and matching comparisonvalue is
3765 # a number we should respect the origin of the original field
3766 # and process comparison value as a string (see
3767 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3769 numeric_comparison = int(comparison_value)
3771 numeric_comparison = parse_filesize(comparison_value)
3772 if numeric_comparison is None:
3773 numeric_comparison = parse_filesize(f'{comparison_value}B')
3774 if numeric_comparison is None:
3775 numeric_comparison = parse_duration(comparison_value)
3776 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3777 raise ValueError('Operator %s only supports string values!' % m['op'])
3778 if actual_value is None:
3779 return is_incomplete(m['key']) or m['none_inclusive']
3780 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3783 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3784 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3786 operator_rex = re.compile(r'''(?x
)
3787 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3788 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3789 m = operator_rex.fullmatch(filter_part.strip())
3791 op = UNARY_OPERATORS[m.group('op')]
3792 actual_value = dct.get(m.group('key'))
3793 if is_incomplete(m.group('key')) and actual_value is None:
3795 return op(actual_value)
3797 raise ValueError('Invalid filter part %r' % filter_part)
3800 def match_str(filter_str, dct, incomplete=False):
3801 """ Filter a dictionary with a simple string syntax.
3802 @returns Whether the filter passes
3803 @param incomplete Set of keys that is expected to be missing from dct.
3804 Can be True/False to indicate all/none of the keys may be missing.
3805 All conditions on incomplete keys pass if the key is missing
3808 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3809 for filter_part in re.split(r'(?<!\\)&', filter_str))
3812 def match_filter_func(filters):
3815 filters = set(variadic(filters))
3817 interactive = '-' in filters
3821 def _match_func(info_dict, incomplete=False):
3822 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3823 return NO_DEFAULT if interactive and not incomplete else None
3825 video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
3826 filter_str = ') | ('.join(map(str.strip, filters))
3827 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3831 class download_range_func:
3832 def __init__(self, chapters, ranges):
3833 self.chapters, self.ranges = chapters, ranges
3835 def __call__(self, info_dict, ydl):
3836 if not self.ranges and not self.chapters:
3839 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3840 else 'Cannot match chapters since chapter information is unavailable')
3841 for regex in self.chapters or []:
3842 for i, chapter in enumerate(info_dict.get('chapters') or []):
3843 if re.search(regex, chapter['title']):
3845 yield {**chapter, 'index': i}
3846 if self.chapters and warning:
3847 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3849 yield from ({'start_time': start, 'end_time': end} for start, end in self.ranges or [])
3851 def __eq__(self, other):
3852 return (isinstance(other, download_range_func)
3853 and self.chapters == other.chapters and self.ranges == other.ranges)
3856 def parse_dfxp_time_expr(time_expr):
3860 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3862 return float(mobj.group('time_offset'))
3864 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3866 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3869 def srt_subtitles_timecode(seconds):
3870 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3873 def ass_subtitles_timecode(seconds):
3874 time = timetuple_from_msec(seconds * 1000)
3875 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3878 def dfxp2srt(dfxp_data):
3880 @param dfxp_data A
bytes-like
object containing DFXP data
3881 @returns A
unicode object containing converted SRT data
3883 LEGACY_NAMESPACES = (
3884 (b'http://www.w3.org/ns/ttml', [
3885 b'http://www.w3.org/2004/11/ttaf1',
3886 b'http://www.w3.org/2006/04/ttaf1',
3887 b'http://www.w3.org/2006/10/ttaf1',
3889 (b'http://www.w3.org/ns/ttml#styling', [
3890 b'http://www.w3.org/ns/ttml#style',
3894 SUPPORTED_STYLING = [
3903 _x = functools.partial(xpath_with_ns, ns_map={
3904 'xml': 'http://www.w3.org/XML/1998/namespace',
3905 'ttml': 'http://www.w3.org/ns/ttml',
3906 'tts': 'http://www.w3.org/ns/ttml#styling',
3912 class TTMLPElementParser:
3914 _unclosed_elements = []
3915 _applied_styles = []
3917 def start(self, tag, attrib):
3918 if tag in (_x('ttml:br'), 'br'):
3921 unclosed_elements = []
3923 element_style_id = attrib.get('style')
3925 style.update(default_style)
3926 if element_style_id:
3927 style.update(styles.get(element_style_id, {}))
3928 for prop in SUPPORTED_STYLING:
3929 prop_val = attrib.get(_x('tts:' + prop))
3931 style[prop] = prop_val
3934 for k, v in sorted(style.items()):
3935 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3938 font += ' color="%s"' % v
3939 elif k == 'fontSize':
3940 font += ' size="%s"' % v
3941 elif k == 'fontFamily':
3942 font += ' face="%s"' % v
3943 elif k == 'fontWeight' and v == 'bold':
3945 unclosed_elements.append('b')
3946 elif k == 'fontStyle' and v == 'italic':
3948 unclosed_elements.append('i')
3949 elif k == 'textDecoration' and v == 'underline':
3951 unclosed_elements.append('u')
3953 self._out += '<font' + font + '>'
3954 unclosed_elements.append('font')
3956 if self._applied_styles:
3957 applied_style.update(self._applied_styles[-1])
3958 applied_style.update(style)
3959 self._applied_styles.append(applied_style)
3960 self._unclosed_elements.append(unclosed_elements)
3963 if tag not in (_x('ttml:br'), 'br'):
3964 unclosed_elements = self._unclosed_elements.pop()
3965 for element in reversed(unclosed_elements):
3966 self._out += '</%s>' % element
3967 if unclosed_elements and self._applied_styles:
3968 self._applied_styles.pop()
3970 def data(self, data):
3974 return self._out.strip()
3976 def parse_node(node):
3977 target = TTMLPElementParser()
3978 parser = xml.etree.ElementTree.XMLParser(target=target)
3979 parser.feed(xml.etree.ElementTree.tostring(node))
3980 return parser.close()
3982 for k, v in LEGACY_NAMESPACES:
3984 dfxp_data = dfxp_data.replace(ns, k)
3986 dfxp = compat_etree_fromstring(dfxp_data)
3988 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3991 raise ValueError('Invalid dfxp/TTML subtitle')
3995 for style in dfxp.findall(_x('.//ttml:style')):
3996 style_id = style.get('id') or style.get(_x('xml:id'))
3999 parent_style_id = style.get('style')
4001 if parent_style_id not in styles:
4004 styles[style_id] = styles[parent_style_id].copy()
4005 for prop in SUPPORTED_STYLING:
4006 prop_val = style.get(_x('tts:' + prop))
4008 styles.setdefault(style_id, {})[prop] = prop_val
4014 for p in ('body', 'div'):
4015 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
4018 style = styles.get(ele.get('style'))
4021 default_style.update(style)
4023 for para, index in zip(paras, itertools.count(1)):
4024 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
4025 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
4026 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
4027 if begin_time is None:
4032 end_time = begin_time + dur
4033 out.append('%d\n%s --> %s\n%s\n\n' % (
4035 srt_subtitles_timecode(begin_time),
4036 srt_subtitles_timecode(end_time),
4042 def cli_option(params, command_option, param, separator=None):
4043 param = params.get(param)
4044 return ([] if param is None
4045 else [command_option, str(param)] if separator is None
4046 else [f'{command_option}{separator}{param}'])
4049 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
4050 param = params.get(param)
4051 assert param in (True, False, None)
4052 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
4055 def cli_valueless_option(params, command_option, param, expected_value=True):
4056 return [command_option] if params.get(param) == expected_value else []
4059 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
4060 if isinstance(argdict, (list, tuple)): # for backward compatibility
4067 assert isinstance(argdict, dict)
4069 assert isinstance(keys, (list, tuple))
4070 for key_list in keys:
4071 arg_list = list(filter(
4072 lambda x: x is not None,
4073 [argdict.get(key.lower()) for key in variadic(key_list)]))
4075 return [arg for args in arg_list for arg in args]
4079 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
4080 main_key, exe = main_key.lower(), exe.lower()
4081 root_key = exe if main_key == exe else f'{main_key}+{exe}'
4082 keys = [f'{root_key}{k}' for k in (keys or [''])]
4083 if root_key in keys:
4085 keys.append((main_key, exe))
4086 keys.append('default')
4089 return cli_configuration_args(argdict, keys, default, use_compat)
4093 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
4152 'iw': 'heb', # Replaced by he in 1989 revision
4162 'in': 'ind', # Replaced by id in 1989 revision
4277 'ji': 'yid', # Replaced by yi in 1989 revision
4285 def short2long(cls, code):
4286 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4287 return cls._lang_map.get(code[:2])
4290 def long2short(cls, code):
4291 """Convert language code from ISO 639-2/T to ISO 639-1"""
4292 for short_name, long_name in cls._lang_map.items():
4293 if long_name == code:
4298 # From http://data.okfn.org/data/core/country-list
4300 'AF': 'Afghanistan',
4301 'AX': 'Åland Islands',
4304 'AS': 'American Samoa',
4309 'AG': 'Antigua and Barbuda',
4326 'BO': 'Bolivia, Plurinational State of',
4327 'BQ': 'Bonaire, Sint Eustatius and Saba',
4328 'BA': 'Bosnia and Herzegovina',
4330 'BV': 'Bouvet Island',
4332 'IO': 'British Indian Ocean Territory',
4333 'BN': 'Brunei Darussalam',
4335 'BF': 'Burkina Faso',
4341 'KY': 'Cayman Islands',
4342 'CF': 'Central African Republic',
4346 'CX': 'Christmas Island',
4347 'CC': 'Cocos (Keeling) Islands',
4351 'CD': 'Congo, the Democratic Republic of the',
4352 'CK': 'Cook Islands',
4354 'CI': 'Côte d\'Ivoire',
4359 'CZ': 'Czech Republic',
4363 'DO': 'Dominican Republic',
4366 'SV': 'El Salvador',
4367 'GQ': 'Equatorial Guinea',
4371 'FK': 'Falkland Islands (Malvinas)',
4372 'FO': 'Faroe Islands',
4376 'GF': 'French Guiana',
4377 'PF': 'French Polynesia',
4378 'TF': 'French Southern Territories',
4393 'GW': 'Guinea-Bissau',
4396 'HM': 'Heard Island and McDonald Islands',
4397 'VA': 'Holy See (Vatican City State)',
4404 'IR': 'Iran, Islamic Republic of',
4407 'IM': 'Isle of Man',
4417 'KP': 'Korea, Democratic People\'s Republic of',
4418 'KR': 'Korea, Republic of',
4421 'LA': 'Lao People\'s Democratic Republic',
4427 'LI': 'Liechtenstein',
4431 'MK': 'Macedonia, the Former Yugoslav Republic of',
4438 'MH': 'Marshall Islands',
4444 'FM': 'Micronesia, Federated States of',
4445 'MD': 'Moldova, Republic of',
4456 'NL': 'Netherlands',
4457 'NC': 'New Caledonia',
4458 'NZ': 'New Zealand',
4463 'NF': 'Norfolk Island',
4464 'MP': 'Northern Mariana Islands',
4469 'PS': 'Palestine, State of',
4471 'PG': 'Papua New Guinea',
4474 'PH': 'Philippines',
4478 'PR': 'Puerto Rico',
4482 'RU': 'Russian Federation',
4484 'BL': 'Saint Barthélemy',
4485 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4486 'KN': 'Saint Kitts and Nevis',
4487 'LC': 'Saint Lucia',
4488 'MF': 'Saint Martin (French part)',
4489 'PM': 'Saint Pierre and Miquelon',
4490 'VC': 'Saint Vincent and the Grenadines',
4493 'ST': 'Sao Tome and Principe',
4494 'SA': 'Saudi Arabia',
4498 'SL': 'Sierra Leone',
4500 'SX': 'Sint Maarten (Dutch part)',
4503 'SB': 'Solomon Islands',
4505 'ZA': 'South Africa',
4506 'GS': 'South Georgia and the South Sandwich Islands',
4507 'SS': 'South Sudan',
4512 'SJ': 'Svalbard and Jan Mayen',
4515 'CH': 'Switzerland',
4516 'SY': 'Syrian Arab Republic',
4517 'TW': 'Taiwan, Province of China',
4519 'TZ': 'Tanzania, United Republic of',
4521 'TL': 'Timor-Leste',
4525 'TT': 'Trinidad and Tobago',
4528 'TM': 'Turkmenistan',
4529 'TC': 'Turks and Caicos Islands',
4533 'AE': 'United Arab Emirates',
4534 'GB': 'United Kingdom',
4535 'US': 'United States',
4536 'UM': 'United States Minor Outlying Islands',
4540 'VE': 'Venezuela, Bolivarian Republic of',
4542 'VG': 'Virgin Islands, British',
4543 'VI': 'Virgin Islands, U.S.',
4544 'WF': 'Wallis and Futuna',
4545 'EH': 'Western Sahara',
4549 # Not ISO 3166 codes, but used for IP blocks
4550 'AP': 'Asia/Pacific Region',
4555 def short2full(cls, code):
4556 """Convert an ISO 3166-2 country code to the corresponding full name"""
4557 return cls._country_map.get(code.upper())
4561 # Major IPv4 address blocks per country
4563 'AD': '46.172.224.0/19',
4564 'AE': '94.200.0.0/13',
4565 'AF': '149.54.0.0/17',
4566 'AG': '209.59.64.0/18',
4567 'AI': '204.14.248.0/21',
4568 'AL': '46.99.0.0/16',
4569 'AM': '46.70.0.0/15',
4570 'AO': '105.168.0.0/13',
4571 'AP': '182.50.184.0/21',
4572 'AQ': '23.154.160.0/24',
4573 'AR': '181.0.0.0/12',
4574 'AS': '202.70.112.0/20',
4575 'AT': '77.116.0.0/14',
4576 'AU': '1.128.0.0/11',
4577 'AW': '181.41.0.0/18',
4578 'AX': '185.217.4.0/22',
4579 'AZ': '5.197.0.0/16',
4580 'BA': '31.176.128.0/17',
4581 'BB': '65.48.128.0/17',
4582 'BD': '114.130.0.0/16',
4584 'BF': '102.178.0.0/15',
4585 'BG': '95.42.0.0/15',
4586 'BH': '37.131.0.0/17',
4587 'BI': '154.117.192.0/18',
4588 'BJ': '137.255.0.0/16',
4589 'BL': '185.212.72.0/23',
4590 'BM': '196.12.64.0/18',
4591 'BN': '156.31.0.0/16',
4592 'BO': '161.56.0.0/16',
4593 'BQ': '161.0.80.0/20',
4594 'BR': '191.128.0.0/12',
4595 'BS': '24.51.64.0/18',
4596 'BT': '119.2.96.0/19',
4597 'BW': '168.167.0.0/16',
4598 'BY': '178.120.0.0/13',
4599 'BZ': '179.42.192.0/18',
4600 'CA': '99.224.0.0/11',
4601 'CD': '41.243.0.0/16',
4602 'CF': '197.242.176.0/21',
4603 'CG': '160.113.0.0/16',
4604 'CH': '85.0.0.0/13',
4605 'CI': '102.136.0.0/14',
4606 'CK': '202.65.32.0/19',
4607 'CL': '152.172.0.0/14',
4608 'CM': '102.244.0.0/14',
4609 'CN': '36.128.0.0/10',
4610 'CO': '181.240.0.0/12',
4611 'CR': '201.192.0.0/12',
4612 'CU': '152.206.0.0/15',
4613 'CV': '165.90.96.0/19',
4614 'CW': '190.88.128.0/17',
4615 'CY': '31.153.0.0/16',
4616 'CZ': '88.100.0.0/14',
4618 'DJ': '197.241.0.0/17',
4619 'DK': '87.48.0.0/12',
4620 'DM': '192.243.48.0/20',
4621 'DO': '152.166.0.0/15',
4622 'DZ': '41.96.0.0/12',
4623 'EC': '186.68.0.0/15',
4624 'EE': '90.190.0.0/15',
4625 'EG': '156.160.0.0/11',
4626 'ER': '196.200.96.0/20',
4627 'ES': '88.0.0.0/11',
4628 'ET': '196.188.0.0/14',
4629 'EU': '2.16.0.0/13',
4630 'FI': '91.152.0.0/13',
4631 'FJ': '144.120.0.0/16',
4632 'FK': '80.73.208.0/21',
4633 'FM': '119.252.112.0/20',
4634 'FO': '88.85.32.0/19',
4636 'GA': '41.158.0.0/15',
4638 'GD': '74.122.88.0/21',
4639 'GE': '31.146.0.0/16',
4640 'GF': '161.22.64.0/18',
4641 'GG': '62.68.160.0/19',
4642 'GH': '154.160.0.0/12',
4643 'GI': '95.164.0.0/16',
4644 'GL': '88.83.0.0/19',
4645 'GM': '160.182.0.0/15',
4646 'GN': '197.149.192.0/18',
4647 'GP': '104.250.0.0/19',
4648 'GQ': '105.235.224.0/20',
4649 'GR': '94.64.0.0/13',
4650 'GT': '168.234.0.0/16',
4651 'GU': '168.123.0.0/16',
4652 'GW': '197.214.80.0/20',
4653 'GY': '181.41.64.0/18',
4654 'HK': '113.252.0.0/14',
4655 'HN': '181.210.0.0/16',
4656 'HR': '93.136.0.0/13',
4657 'HT': '148.102.128.0/17',
4658 'HU': '84.0.0.0/14',
4659 'ID': '39.192.0.0/10',
4660 'IE': '87.32.0.0/12',
4661 'IL': '79.176.0.0/13',
4662 'IM': '5.62.80.0/20',
4663 'IN': '117.192.0.0/10',
4664 'IO': '203.83.48.0/21',
4665 'IQ': '37.236.0.0/14',
4666 'IR': '2.176.0.0/12',
4667 'IS': '82.221.0.0/16',
4668 'IT': '79.0.0.0/10',
4669 'JE': '87.244.64.0/18',
4670 'JM': '72.27.0.0/17',
4671 'JO': '176.29.0.0/16',
4672 'JP': '133.0.0.0/8',
4673 'KE': '105.48.0.0/12',
4674 'KG': '158.181.128.0/17',
4675 'KH': '36.37.128.0/17',
4676 'KI': '103.25.140.0/22',
4677 'KM': '197.255.224.0/20',
4678 'KN': '198.167.192.0/19',
4679 'KP': '175.45.176.0/22',
4680 'KR': '175.192.0.0/10',
4681 'KW': '37.36.0.0/14',
4682 'KY': '64.96.0.0/15',
4683 'KZ': '2.72.0.0/13',
4684 'LA': '115.84.64.0/18',
4685 'LB': '178.135.0.0/16',
4686 'LC': '24.92.144.0/20',
4687 'LI': '82.117.0.0/19',
4688 'LK': '112.134.0.0/15',
4689 'LR': '102.183.0.0/16',
4690 'LS': '129.232.0.0/17',
4691 'LT': '78.56.0.0/13',
4692 'LU': '188.42.0.0/16',
4693 'LV': '46.109.0.0/16',
4694 'LY': '41.252.0.0/14',
4695 'MA': '105.128.0.0/11',
4696 'MC': '88.209.64.0/18',
4697 'MD': '37.246.0.0/16',
4698 'ME': '178.175.0.0/17',
4699 'MF': '74.112.232.0/21',
4700 'MG': '154.126.0.0/17',
4701 'MH': '117.103.88.0/21',
4702 'MK': '77.28.0.0/15',
4703 'ML': '154.118.128.0/18',
4704 'MM': '37.111.0.0/17',
4705 'MN': '49.0.128.0/17',
4706 'MO': '60.246.0.0/16',
4707 'MP': '202.88.64.0/20',
4708 'MQ': '109.203.224.0/19',
4709 'MR': '41.188.64.0/18',
4710 'MS': '208.90.112.0/22',
4711 'MT': '46.11.0.0/16',
4712 'MU': '105.16.0.0/12',
4713 'MV': '27.114.128.0/18',
4714 'MW': '102.70.0.0/15',
4715 'MX': '187.192.0.0/11',
4716 'MY': '175.136.0.0/13',
4717 'MZ': '197.218.0.0/15',
4718 'NA': '41.182.0.0/16',
4719 'NC': '101.101.0.0/18',
4720 'NE': '197.214.0.0/18',
4721 'NF': '203.17.240.0/22',
4722 'NG': '105.112.0.0/12',
4723 'NI': '186.76.0.0/15',
4724 'NL': '145.96.0.0/11',
4725 'NO': '84.208.0.0/13',
4726 'NP': '36.252.0.0/15',
4727 'NR': '203.98.224.0/19',
4728 'NU': '49.156.48.0/22',
4729 'NZ': '49.224.0.0/14',
4730 'OM': '5.36.0.0/15',
4731 'PA': '186.72.0.0/15',
4732 'PE': '186.160.0.0/14',
4733 'PF': '123.50.64.0/18',
4734 'PG': '124.240.192.0/19',
4735 'PH': '49.144.0.0/13',
4736 'PK': '39.32.0.0/11',
4737 'PL': '83.0.0.0/11',
4738 'PM': '70.36.0.0/20',
4739 'PR': '66.50.0.0/16',
4740 'PS': '188.161.0.0/16',
4741 'PT': '85.240.0.0/13',
4742 'PW': '202.124.224.0/20',
4743 'PY': '181.120.0.0/14',
4744 'QA': '37.210.0.0/15',
4745 'RE': '102.35.0.0/16',
4746 'RO': '79.112.0.0/13',
4747 'RS': '93.86.0.0/15',
4748 'RU': '5.136.0.0/13',
4749 'RW': '41.186.0.0/16',
4750 'SA': '188.48.0.0/13',
4751 'SB': '202.1.160.0/19',
4752 'SC': '154.192.0.0/11',
4753 'SD': '102.120.0.0/13',
4754 'SE': '78.64.0.0/12',
4755 'SG': '8.128.0.0/10',
4756 'SI': '188.196.0.0/14',
4757 'SK': '78.98.0.0/15',
4758 'SL': '102.143.0.0/17',
4759 'SM': '89.186.32.0/19',
4760 'SN': '41.82.0.0/15',
4761 'SO': '154.115.192.0/18',
4762 'SR': '186.179.128.0/17',
4763 'SS': '105.235.208.0/21',
4764 'ST': '197.159.160.0/19',
4765 'SV': '168.243.0.0/16',
4766 'SX': '190.102.0.0/20',
4768 'SZ': '41.84.224.0/19',
4769 'TC': '65.255.48.0/20',
4770 'TD': '154.68.128.0/19',
4771 'TG': '196.168.0.0/14',
4772 'TH': '171.96.0.0/13',
4773 'TJ': '85.9.128.0/18',
4774 'TK': '27.96.24.0/21',
4775 'TL': '180.189.160.0/20',
4776 'TM': '95.85.96.0/19',
4777 'TN': '197.0.0.0/11',
4778 'TO': '175.176.144.0/21',
4779 'TR': '78.160.0.0/11',
4780 'TT': '186.44.0.0/15',
4781 'TV': '202.2.96.0/19',
4782 'TW': '120.96.0.0/11',
4783 'TZ': '156.156.0.0/14',
4784 'UA': '37.52.0.0/14',
4785 'UG': '102.80.0.0/13',
4787 'UY': '167.56.0.0/13',
4788 'UZ': '84.54.64.0/18',
4789 'VA': '212.77.0.0/19',
4790 'VC': '207.191.240.0/21',
4791 'VE': '186.88.0.0/13',
4792 'VG': '66.81.192.0/20',
4793 'VI': '146.226.0.0/16',
4794 'VN': '14.160.0.0/11',
4795 'VU': '202.80.32.0/20',
4796 'WF': '117.20.32.0/21',
4797 'WS': '202.4.32.0/19',
4798 'YE': '134.35.0.0/16',
4799 'YT': '41.242.116.0/22',
4800 'ZA': '41.0.0.0/11',
4801 'ZM': '102.144.0.0/13',
4802 'ZW': '102.177.192.0/18',
4806 def random_ipv4(cls, code_or_block):
4807 if len(code_or_block) == 2:
4808 block = cls._country_ip_map.get(code_or_block.upper())
4812 block = code_or_block
4813 addr, preflen = block.split('/')
4814 addr_min = struct.unpack('!L', socket.inet_aton(addr))[0]
4815 addr_max = addr_min | (0xffffffff >> int(preflen))
4816 return str(socket.inet_ntoa(
4817 struct.pack('!L', random.randint(addr_min, addr_max))))
4820 class PerRequestProxyHandler(urllib.request.ProxyHandler):
4821 def __init__(self, proxies=None):
4822 # Set default handlers
4823 for type in ('http', 'https'):
4824 setattr(self, '%s_open' % type,
4825 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4826 meth(r, proxy, type))
4827 urllib.request.ProxyHandler.__init__(self, proxies)
4829 def proxy_open(self, req, proxy, type):
4830 req_proxy = req.headers.get('Ytdl-request-proxy')
4831 if req_proxy is not None:
4833 del req.headers['Ytdl-request-proxy']
4835 if proxy == '__noproxy__':
4836 return None # No Proxy
4837 if urllib.parse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4838 req.add_header('Ytdl-socks-proxy', proxy)
4839 # yt-dlp's http/https handlers do wrapping the socket with socks
4841 return urllib.request.ProxyHandler.proxy_open(
4842 self, req, proxy, type)
4845 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4846 # released into Public Domain
4847 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4849 def long_to_bytes(n, blocksize=0):
4850 """long_to_bytes(n:long, blocksize:int) : string
4851 Convert a long integer to a byte string.
4853 If optional blocksize is given and greater than zero, pad the front of the
4854 byte string with binary zeros so that the length is a multiple of
4857 # after much testing, this algorithm was deemed to be the fastest
4861 s = struct.pack('>I', n & 0xffffffff) + s
4863 # strip off leading zeros
4864 for i in range(len(s)):
4865 if s[i] != b'\000'[0]:
4868 # only happens when n == 0
4872 # add back some pad bytes. this could be done more efficiently w.r.t. the
4873 # de-padding being done above, but sigh...
4874 if blocksize > 0 and len(s) % blocksize:
4875 s = (blocksize - len(s) % blocksize) * b'\000' + s
4879 def bytes_to_long(s):
4880 """bytes_to_long(string) : long
4881 Convert a byte string to a long integer.
4883 This is (essentially) the inverse of long_to_bytes().
4888 extra = (4 - length % 4)
4889 s = b'\000' * extra + s
4890 length = length + extra
4891 for i in range(0, length, 4):
4892 acc = (acc << 32) + struct.unpack('>I', s[i:i + 4])[0]
4896 def ohdave_rsa_encrypt(data, exponent, modulus):
4898 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4901 data: data to encrypt, bytes-like object
4902 exponent, modulus: parameter e and N of RSA algorithm, both integer
4903 Output: hex string of encrypted data
4905 Limitation: supports one block encryption only
4908 payload = int(binascii.hexlify(data[::-1]), 16)
4909 encrypted = pow(payload, exponent, modulus)
4910 return '%x' % encrypted
4913 def pkcs1pad(data, length):
4915 Padding input data with PKCS#1 scheme
4917 @param {int[]} data input data
4918 @param {int} length target length
4919 @returns {int[]} padded data
4921 if len(data) > length - 11:
4922 raise ValueError('Input data too
long for PKCS
#1 padding')
4924 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4925 return [0, 2] + pseudo_random
+ [0] + data
4928 def _base_n_table(n
, table
):
4929 if not table
and not n
:
4930 raise ValueError('Either table or n must be specified')
4931 table
= (table
or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n
]
4933 if n
and n
!= len(table
):
4934 raise ValueError(f
'base {n} exceeds table length {len(table)}')
4938 def encode_base_n(num
, n
=None, table
=None):
4939 """Convert given int to a base-n string"""
4940 table
= _base_n_table(n
, table
)
4944 result
, base
= '', len(table
)
4946 result
= table
[num
% base
] + result
4951 def decode_base_n(string
, n
=None, table
=None):
4952 """Convert given base-n string to int"""
4953 table
= {char: index for index, char in enumerate(_base_n_table(n, table))}
4954 result
, base
= 0, len(table
)
4956 result
= result
* base
+ table
[char
]
4960 def decode_base(value
, digits
):
4961 deprecation_warning(f
'{__name__}.decode_base is deprecated and may be removed '
4962 f
'in a future version. Use {__name__}.decode_base_n instead')
4963 return decode_base_n(value
, table
=digits
)
4966 def decode_packed_codes(code
):
4967 mobj
= re
.search(PACKED_CODES_RE
, code
)
4968 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4971 symbols
= symbols
.split('|')
4976 base_n_count
= encode_base_n(count
, base
)
4977 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4980 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4984 def caesar(s
, alphabet
, shift
):
4989 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4994 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4997 def parse_m3u8_attributes(attrib
):
4999 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
5000 if val
.startswith('"'):
5006 def urshift(val
, n
):
5007 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
5010 # Based on png2str() written by @gdkchan and improved by @yokrysty
5011 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
5012 def decode_png(png_data
):
5013 # Reference: https://www.w3.org/TR/PNG/
5014 header
= png_data
[8:]
5016 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
5017 raise OSError('Not a valid PNG file.')
5019 int_map
= {1: '>B', 2: '>H', 4: '>I'}
5020 unpack_integer
= lambda x
: struct
.unpack(int_map
[len(x
)], x
)[0]
5025 length
= unpack_integer(header
[:4])
5028 chunk_type
= header
[:4]
5031 chunk_data
= header
[:length
]
5032 header
= header
[length
:]
5034 header
= header
[4:] # Skip CRC
5042 ihdr
= chunks
[0]['data']
5044 width
= unpack_integer(ihdr
[:4])
5045 height
= unpack_integer(ihdr
[4:8])
5049 for chunk
in chunks
:
5050 if chunk
['type'] == b
'IDAT':
5051 idat
+= chunk
['data']
5054 raise OSError('Unable to read PNG data.')
5056 decompressed_data
= bytearray(zlib
.decompress(idat
))
5061 def _get_pixel(idx
):
5066 for y
in range(height
):
5067 basePos
= y
* (1 + stride
)
5068 filter_type
= decompressed_data
[basePos
]
5072 pixels
.append(current_row
)
5074 for x
in range(stride
):
5075 color
= decompressed_data
[1 + basePos
+ x
]
5076 basex
= y
* stride
+ x
5081 left
= _get_pixel(basex
- 3)
5083 up
= _get_pixel(basex
- stride
)
5085 if filter_type
== 1: # Sub
5086 color
= (color
+ left
) & 0xff
5087 elif filter_type
== 2: # Up
5088 color
= (color
+ up
) & 0xff
5089 elif filter_type
== 3: # Average
5090 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
5091 elif filter_type
== 4: # Paeth
5097 c
= _get_pixel(basex
- stride
- 3)
5105 if pa
<= pb
and pa
<= pc
:
5106 color
= (color
+ a
) & 0xff
5108 color
= (color
+ b
) & 0xff
5110 color
= (color
+ c
) & 0xff
5112 current_row
.append(color
)
5114 return width
, height
, pixels
5117 def write_xattr(path
, key
, value
):
5118 # Windows: Write xattrs to NTFS Alternate Data Streams:
5119 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
5120 if compat_os_name
== 'nt':
5121 assert ':' not in key
5122 assert os
.path
.exists(path
)
5125 with open(f
'{path}:{key}', 'wb') as f
:
5127 except OSError as e
:
5128 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5131 # UNIX Method 1. Use xattrs/pyxattrs modules
5134 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
5135 # Unicode arguments are not supported in pyxattr until version 0.5.0
5136 # See https://github.com/ytdl-org/youtube-dl/issues/5498
5137 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
5138 setxattr
= xattr
.set
5140 setxattr
= xattr
.setxattr
5144 setxattr(path
, key
, value
)
5145 except OSError as e
:
5146 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5149 # UNIX Method 2. Use setfattr/xattr executables
5150 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
5151 else 'xattr' if check_executable('xattr', ['-h']) else None)
5153 raise XAttrUnavailableError(
5154 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
5155 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
5157 value
= value
.decode()
5159 _
, stderr
, returncode
= Popen
.run(
5160 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
5161 text
=True, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
5162 except OSError as e
:
5163 raise XAttrMetadataError(e
.errno
, e
.strerror
)
5165 raise XAttrMetadataError(returncode
, stderr
)
5168 def random_birthday(year_field
, month_field
, day_field
):
5169 start_date
= datetime
.date(1950, 1, 1)
5170 end_date
= datetime
.date(1995, 12, 31)
5171 offset
= random
.randint(0, (end_date
- start_date
).days
)
5172 random_date
= start_date
+ datetime
.timedelta(offset
)
5174 year_field
: str(random_date
.year
),
5175 month_field
: str(random_date
.month
),
5176 day_field
: str(random_date
.day
),
5180 # Templates for internet shortcut files, which are plain text files.
5181 DOT_URL_LINK_TEMPLATE
= '''\
5186 DOT_WEBLOC_LINK_TEMPLATE
= '''\
5187 <?xml version="1.0" encoding="UTF-8"?>
5188 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
5189 <plist version="1.0">
5192 \t<string>%(url)s</string>
5197 DOT_DESKTOP_LINK_TEMPLATE
= '''\
5207 'url': DOT_URL_LINK_TEMPLATE
,
5208 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
5209 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
5213 def iri_to_uri(iri
):
5215 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
5217 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.
5220 iri_parts
= urllib
.parse
.urlparse(iri
)
5222 if '[' in iri_parts
.netloc
:
5223 raise ValueError('IPv6 URIs are not, yet, supported.')
5224 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
5226 # 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.
5229 if iri_parts
.username
:
5230 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
5231 if iri_parts
.password
is not None:
5232 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
5235 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
5236 # The 'idna' encoding produces ASCII text.
5237 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
5238 net_location
+= ':' + str(iri_parts
.port
)
5240 return urllib
.parse
.urlunparse(
5244 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
5246 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
5247 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
5249 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
5250 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
5252 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
5254 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
5257 def to_high_limit_path(path
):
5258 if sys
.platform
in ['win32', 'cygwin']:
5259 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
5260 return '\\\\?\\' + os
.path
.abspath(path
)
5265 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=IDENTITY
):
5266 val
= traverse_obj(obj
, *variadic(field
))
5267 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in variadic(ignore
):
5269 return template
% func(val
)
5272 def clean_podcast_url(url
):
5273 return re
.sub(r
'''(?x)
5277 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5280 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5283 cn\.co| # https://podcorn.com/analytics-prefix/
5284 st\.fm # https://podsights.com/docs/
5289 _HEX_TABLE
= '0123456789abcdef'
5292 def random_uuidv4():
5293 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5296 def make_dir(path
, to_screen
=None):
5298 dn
= os
.path
.dirname(path
)
5299 if dn
and not os
.path
.exists(dn
):
5302 except OSError as err
:
5303 if callable(to_screen
) is not None:
5304 to_screen('unable to create directory ' + error_to_compat_str(err
))
5308 def get_executable_path():
5309 from .update
import _get_variant_and_executable_path
5311 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
5314 def load_plugins(name
, suffix
, namespace
):
5316 with contextlib
.suppress(FileNotFoundError
):
5317 plugins_spec
= importlib
.util
.spec_from_file_location(
5318 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5319 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5320 sys
.modules
[plugins_spec
.name
] = plugins
5321 plugins_spec
.loader
.exec_module(plugins
)
5322 for name
in dir(plugins
):
5323 if name
in namespace
:
5325 if not name
.endswith(suffix
):
5327 klass
= getattr(plugins
, name
)
5328 classes
[name
] = namespace
[name
] = klass
5333 obj
, *paths
, default
=NO_DEFAULT
, expected_type
=None, get_all
=True,
5334 casesense
=True, is_user_input
=False, traverse_string
=False):
5336 Safely traverse nested `dict`s and `Sequence`s
5338 >>> obj = [{}, {"key": "value"}]
5339 >>> traverse_obj(obj, (1, "key"))
5342 Each of the provided `paths` is tested and the first producing a valid result will be returned.
5343 The next path will also be tested if the path branched but no results could be found.
5344 Supported values for traversal are `Mapping`, `Sequence` and `re.Match`.
5345 A value of None is treated as the absence of a value.
5347 The paths will be wrapped in `variadic`, so that `'key'` is conveniently the same as `('key', )`.
5349 The keys in the path can be one of:
5350 - `None`: Return the current object.
5351 - `str`/`int`: Return `obj[key]`. For `re.Match, return `obj.group(key)`.
5352 - `slice`: Branch out and return all values in `obj[key]`.
5353 - `Ellipsis`: Branch out and return a list of all values.
5354 - `tuple`/`list`: Branch out and return a list of all matching values.
5355 Read as: `[traverse_obj(obj, branch) for branch in branches]`.
5356 - `function`: Branch out and return values filtered by the function.
5357 Read as: `[value for key, value in obj if function(key, value)]`.
5358 For `Sequence`s, `key` is the index of the value.
5359 - `dict` Transform the current object and return a matching dict.
5360 Read as: `{key: traverse_obj(obj, path) for key, path in dct.items()}`.
5362 `tuple`, `list`, and `dict` all support nested paths and branches.
5364 @params paths Paths which to traverse by.
5365 @param default Value to return if the paths do not match.
5366 @param expected_type If a `type`, only accept final values of this type.
5367 If any other callable, try to call the function on each result.
5368 @param get_all If `False`, return the first matching result, otherwise all matching ones.
5369 @param casesense If `False`, consider string dictionary keys as case insensitive.
5371 The following are only meant to be used by YoutubeDL.prepare_outtmpl and are not part of the API
5373 @param is_user_input Whether the keys are generated from user input.
5374 If `True` strings get converted to `int`/`slice` if needed.
5375 @param traverse_string Whether to traverse into objects as strings.
5376 If `True`, any non-compatible object will first be
5377 converted into a string and then traversed into.
5380 @returns The result of the object traversal.
5381 If successful, `get_all=True`, and the path branches at least once,
5382 then a list of results is returned instead.
5383 A list is always returned if the last path branches and no `default` is given.
5385 is_sequence
= lambda x
: isinstance(x
, collections
.abc
.Sequence
) and not isinstance(x
, (str, bytes))
5386 casefold
= lambda k
: k
.casefold() if isinstance(k
, str) else k
5388 if isinstance(expected_type
, type):
5389 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5391 type_test
= lambda val
: try_call(expected_type
or IDENTITY
, args
=(val
,))
5393 def apply_key(key
, obj
):
5400 elif isinstance(key
, (list, tuple)):
5402 _
, result
= apply_path(obj
, branch
)
5406 if isinstance(obj
, collections
.abc
.Mapping
):
5407 yield from obj
.values()
5408 elif is_sequence(obj
):
5410 elif isinstance(obj
, re
.Match
):
5411 yield from obj
.groups()
5412 elif traverse_string
:
5416 if is_sequence(obj
):
5417 iter_obj
= enumerate(obj
)
5418 elif isinstance(obj
, collections
.abc
.Mapping
):
5419 iter_obj
= obj
.items()
5420 elif isinstance(obj
, re
.Match
):
5421 iter_obj
= enumerate((obj
.group(), *obj
.groups()))
5422 elif traverse_string
:
5423 iter_obj
= enumerate(str(obj
))
5426 yield from (v
for k
, v
in iter_obj
if try_call(key
, args
=(k
, v
)))
5428 elif isinstance(key
, dict):
5429 iter_obj
= ((k
, _traverse_obj(obj
, v
)) for k
, v
in key
.items())
5430 yield {k
: v
if v
is not None else default
for k
, v
in iter_obj
5431 if v
is not None or default
is not NO_DEFAULT
}
5433 elif isinstance(obj
, collections
.abc
.Mapping
):
5434 yield (obj
.get(key
) if casesense
or (key
in obj
)
5435 else next((v
for k
, v
in obj
.items() if casefold(k
) == key
), None))
5437 elif isinstance(obj
, re
.Match
):
5438 if isinstance(key
, int) or casesense
:
5439 with contextlib
.suppress(IndexError):
5440 yield obj
.group(key
)
5443 if not isinstance(key
, str):
5446 yield next((v
for k
, v
in obj
.groupdict().items() if casefold(k
) == key
), None)
5450 key
= (int_or_none(key
) if ':' not in key
5451 else slice(*map(int_or_none
, key
.split(':'))))
5453 if not isinstance(key
, (int, slice)):
5456 if not is_sequence(obj
):
5457 if not traverse_string
:
5461 with contextlib
.suppress(IndexError):
5464 def apply_path(start_obj
, path
):
5466 has_branched
= False
5468 for key
in variadic(path
):
5469 if is_user_input
and key
== ':':
5472 if not casesense
and isinstance(key
, str):
5473 key
= key
.casefold()
5475 if key
is ... or isinstance(key
, (list, tuple)) or callable(key
):
5478 key_func
= functools
.partial(apply_key
, key
)
5479 objs
= itertools
.chain
.from_iterable(map(key_func
, objs
))
5481 return has_branched
, objs
5483 def _traverse_obj(obj
, path
, use_list
=True):
5484 has_branched
, results
= apply_path(obj
, path
)
5485 results
= LazyList(x
for x
in map(type_test
, results
) if x
is not None)
5487 if get_all
and has_branched
:
5488 return results
.exhaust() if results
or use_list
else None
5490 return results
[0] if results
else None
5492 for index
, path
in enumerate(paths
, 1):
5493 use_list
= default
is NO_DEFAULT
and index
== len(paths
)
5494 result
= _traverse_obj(obj
, path
, use_list
)
5495 if result
is not None:
5498 return None if default
is NO_DEFAULT
else default
5501 def traverse_dict(dictn
, keys
, casesense
=True):
5502 deprecation_warning(f
'"{__name__}.traverse_dict" is deprecated and may be removed '
5503 f
'in a future version. Use "{__name__}.traverse_obj" instead')
5504 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5507 def get_first(obj
, keys
, **kwargs
):
5508 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5511 def time_seconds(**kwargs
):
5512 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5513 return t
.timestamp()
5516 # create a JSON Web Signature (jws) with HS256 algorithm
5517 # the resulting format is in JWS Compact Serialization
5518 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5519 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5520 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5526 header_data
.update(headers
)
5527 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5528 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5529 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5530 signature_b64
= base64
.b64encode(h
.digest())
5531 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5535 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5536 def jwt_decode_hs256(jwt
):
5537 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5538 # add trailing ='s that may have been stripped, superfluous ='s are ignored
5539 payload_data
= json
.loads(base64
.urlsafe_b64decode(f
'{payload_b64}==='))
5543 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5547 def supports_terminal_sequences(stream
):
5548 if compat_os_name
== 'nt':
5549 if not WINDOWS_VT_MODE
:
5551 elif not os
.getenv('TERM'):
5554 return stream
.isatty()
5555 except BaseException
:
5559 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5560 if get_windows_version() < (10, 0, 10586):
5562 global WINDOWS_VT_MODE
5564 Popen
.run('', shell
=True)
5568 WINDOWS_VT_MODE
= True
5569 supports_terminal_sequences
.cache_clear()
5572 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5575 def remove_terminal_sequences(string
):
5576 return _terminal_sequences_re
.sub('', string
)
5579 def number_of_digits(number
):
5580 return len('%d' % number
)
5583 def join_nonempty(*values
, delim
='-', from_dict
=None):
5584 if from_dict
is not None:
5585 values
= (traverse_obj(from_dict
, variadic(v
)) for v
in values
)
5586 return delim
.join(map(str, filter(None, values
)))
5589 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5591 Find the largest format dimensions in terms of video width and, for each thumbnail:
5592 * Modify the URL: Match the width with the provided regex and replace with the former width
5595 This function is useful with video services that scale the provided thumbnails on demand
5597 _keys
= ('width', 'height')
5598 max_dimensions
= max(
5599 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5601 if not max_dimensions
[0]:
5605 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5606 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5607 for thumbnail
in thumbnails
5611 def parse_http_range(range):
5612 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5614 return None, None, None
5615 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5617 return None, None, None
5618 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5621 def read_stdin(what
):
5622 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5623 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5627 def determine_file_encoding(data
):
5629 Detect the text encoding used
5630 @returns (encoding, bytes to skip)
5633 # BOM marks are given priority over declarations
5634 for bom
, enc
in BOMS
:
5635 if data
.startswith(bom
):
5636 return enc
, len(bom
)
5638 # Strip off all null bytes to match even when UTF-16 or UTF-32 is used.
5639 # We ignore the endianness to get a good enough match
5640 data
= data
.replace(b
'\0', b
'')
5641 mobj
= re
.match(rb
'(?m)^#\s*coding\s*:\s*(\S+)\s*$', data
)
5642 return mobj
.group(1).decode() if mobj
else None, 0
5649 __initialized
= False
5651 def __init__(self
, parser
, label
=None):
5652 self
.parser
, self
.label
= parser
, label
5653 self
._loaded
_paths
, self
.configs
= set(), []
5655 def init(self
, args
=None, filename
=None):
5656 assert not self
.__initialized
5657 self
.own_args
, self
.filename
= args
, filename
5658 return self
.load_configs()
5660 def load_configs(self
):
5663 location
= os
.path
.realpath(self
.filename
)
5664 directory
= os
.path
.dirname(location
)
5665 if location
in self
._loaded
_paths
:
5667 self
._loaded
_paths
.add(location
)
5669 self
.__initialized
= True
5670 opts
, _
= self
.parser
.parse_known_args(self
.own_args
)
5671 self
.parsed_args
= self
.own_args
5672 for location
in opts
.config_locations
or []:
5674 if location
in self
._loaded
_paths
:
5676 self
._loaded
_paths
.add(location
)
5677 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5679 location
= os
.path
.join(directory
, expand_path(location
))
5680 if os
.path
.isdir(location
):
5681 location
= os
.path
.join(location
, 'yt-dlp.conf')
5682 if not os
.path
.exists(location
):
5683 self
.parser
.error(f
'config location {location} does not exist')
5684 self
.append_config(self
.read_file(location
), location
)
5688 label
= join_nonempty(
5689 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5691 return join_nonempty(
5692 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5693 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5697 def read_file(filename
, default
=[]):
5699 optionf
= open(filename
, 'rb')
5701 return default
# silently skip if file is not present
5703 enc
, skip
= determine_file_encoding(optionf
.read(512))
5704 optionf
.seek(skip
, io
.SEEK_SET
)
5706 enc
= None # silently skip read errors
5708 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5709 contents
= optionf
.read().decode(enc
or preferredencoding())
5710 res
= shlex
.split(contents
, comments
=True)
5711 except Exception as err
:
5712 raise ValueError(f
'Unable to parse "{filename}": {err}')
5718 def hide_login_info(opts
):
5719 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5720 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5725 return m
.group('key') + '=PRIVATE'
5729 opts
= list(map(_scrub_eq
, opts
))
5730 for idx
, opt
in enumerate(opts
):
5731 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5732 opts
[idx
+ 1] = 'PRIVATE'
5735 def append_config(self
, *args
, label
=None):
5736 config
= type(self
)(self
.parser
, label
)
5737 config
._loaded
_paths
= self
._loaded
_paths
5738 if config
.init(*args
):
5739 self
.configs
.append(config
)
5743 for config
in reversed(self
.configs
):
5744 yield from config
.all_args
5745 yield from self
.parsed_args
or []
5747 def parse_known_args(self
, **kwargs
):
5748 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5750 def parse_args(self
):
5751 return self
.parser
.parse_args(self
.all_args
)
5754 class WebSocketsWrapper
:
5755 """Wraps websockets module to use in non-async scopes"""
5758 def __init__(self
, url
, headers
=None, connect
=True):
5759 self
.loop
= asyncio
.new_event_loop()
5760 # XXX: "loop" is deprecated
5761 self
.conn
= websockets
.connect(
5762 url
, extra_headers
=headers
, ping_interval
=None,
5763 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5766 atexit
.register(self
.__exit
__, None, None, None)
5768 def __enter__(self
):
5770 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5773 def send(self
, *args
):
5774 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5776 def recv(self
, *args
):
5777 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5779 def __exit__(self
, type, value
, traceback
):
5781 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5784 self
._cancel
_all
_tasks
(self
.loop
)
5786 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5787 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5789 def run_with_loop(main
, loop
):
5790 if not asyncio
.iscoroutine(main
):
5791 raise ValueError(f
'a coroutine was expected, got {main!r}')
5794 return loop
.run_until_complete(main
)
5796 loop
.run_until_complete(loop
.shutdown_asyncgens())
5797 if hasattr(loop
, 'shutdown_default_executor'):
5798 loop
.run_until_complete(loop
.shutdown_default_executor())
5801 def _cancel_all_tasks(loop
):
5802 to_cancel
= asyncio
.all_tasks(loop
)
5807 for task
in to_cancel
:
5810 # XXX: "loop" is removed in python 3.10+
5811 loop
.run_until_complete(
5812 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5814 for task
in to_cancel
:
5815 if task
.cancelled():
5817 if task
.exception() is not None:
5818 loop
.call_exception_handler({
5819 'message': 'unhandled exception during asyncio.run() shutdown',
5820 'exception': task
.exception(),
5825 def merge_headers(*dicts
):
5826 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5827 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5830 def cached_method(f
):
5831 """Cache a method"""
5832 signature
= inspect
.signature(f
)
5835 def wrapper(self
, *args
, **kwargs
):
5836 bound_args
= signature
.bind(self
, *args
, **kwargs
)
5837 bound_args
.apply_defaults()
5838 key
= tuple(bound_args
.arguments
.values())[1:]
5840 cache
= vars(self
).setdefault('__cached_method__cache', {}).setdefault(f.__name__, {}
)
5841 if key
not in cache
:
5842 cache
[key
] = f(self
, *args
, **kwargs
)
5847 class classproperty
:
5848 """property access for class methods"""
5850 def __init__(self
, func
):
5851 functools
.update_wrapper(self
, func
)
5854 def __get__(self
, _
, cls
):
5855 return self
.func(cls
)
5858 class Namespace(types
.SimpleNamespace
):
5859 """Immutable namespace"""
5862 return iter(self
.__dict
__.values())
5866 return self
.__dict
__.items()
5869 MEDIA_EXTENSIONS
= Namespace(
5870 common_video
=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
5871 video
=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
5872 common_audio
=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
5873 audio
=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
5874 thumbnails
=('jpg', 'png', 'webp'),
5875 storyboards
=('mhtml', ),
5876 subtitles
=('srt', 'vtt', 'ass', 'lrc'),
5877 manifests
=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
5879 MEDIA_EXTENSIONS
.video
+= MEDIA_EXTENSIONS
.common_video
5880 MEDIA_EXTENSIONS
.audio
+= MEDIA_EXTENSIONS
.common_audio
5882 KNOWN_EXTENSIONS
= (*MEDIA_EXTENSIONS
.video
, *MEDIA_EXTENSIONS
.audio
, *MEDIA_EXTENSIONS
.manifests
)
5887 for retry in RetryManager(...):
5890 except SomeException as err:
5894 attempt
, _error
= 0, None
5896 def __init__(self
, _retries
, _error_callback
, **kwargs
):
5897 self
.retries
= _retries
or 0
5898 self
.error_callback
= functools
.partial(_error_callback
, **kwargs
)
5900 def _should_retry(self
):
5901 return self
._error
is not NO_DEFAULT
and self
.attempt
<= self
.retries
5905 if self
._error
is NO_DEFAULT
:
5910 def error(self
, value
):
5914 while self
._should
_retry
():
5915 self
.error
= NO_DEFAULT
5919 self
.error_callback(self
.error
, self
.attempt
, self
.retries
)
5922 def report_retry(e
, count
, retries
, *, sleep_func
, info
, warn
, error
=None, suffix
=None):
5923 """Utility function for reporting retries"""
5926 return error(f
'{e}. Giving up after {count - 1} retries') if count
> 1 else error(str(e
))
5931 elif isinstance(e
, ExtractorError
):
5932 e
= remove_end(str_or_none(e
.cause
) or e
.orig_msg
, '.')
5933 warn(f
'{e}. Retrying{format_field(suffix, None, " %s")} ({count}/{retries})...')
5935 delay
= float_or_none(sleep_func(n
=count
- 1)) if callable(sleep_func
) else sleep_func
5937 info(f
'Sleeping {delay:.2f} seconds ...')
5941 def make_archive_id(ie
, video_id
):
5942 ie_key
= ie
if isinstance(ie
, str) else ie
.ie_key()
5943 return f
'{ie_key.lower()} {video_id}'
5946 def truncate_string(s
, left
, right
=0):
5947 assert left
> 3 and right
>= 0
5948 if s
is None or len(s
) <= left
+ right
:
5950 return f
'{s[:left-3]}...{s[-right:]}'
5953 def orderedSet_from_options(options
, alias_dict
, *, use_regex
=False, start
=None):
5954 assert 'all' in alias_dict
, '"all" alias is required'
5955 requested
= list(start
or [])
5957 discard
= val
.startswith('-')
5961 if val
in alias_dict
:
5962 val
= alias_dict
[val
] if not discard
else [
5963 i
[1:] if i
.startswith('-') else f
'-{i}' for i
in alias_dict
[val
]]
5964 # NB: Do not allow regex in aliases for performance
5965 requested
= orderedSet_from_options(val
, alias_dict
, start
=requested
)
5968 current
= (filter(re
.compile(val
, re
.I
).fullmatch
, alias_dict
['all']) if use_regex
5969 else [val
] if val
in alias_dict
['all'] else None)
5971 raise ValueError(val
)
5974 for item
in current
:
5975 while item
in requested
:
5976 requested
.remove(item
)
5978 requested
.extend(current
)
5980 return orderedSet(requested
)
5984 has_certifi
= bool(certifi
)
5985 has_websockets
= bool(websockets
)