39 import xml
.etree
.ElementTree
42 from .compat
import asyncio
, functools
# isort: split
46 compat_etree_fromstring
,
49 compat_html_entities_html5
,
50 compat_HTMLParseError
,
61 compat_urllib_parse_unquote_plus
,
62 compat_urllib_parse_urlencode
,
63 compat_urllib_parse_urlparse
,
64 compat_urllib_request
,
67 from .dependencies
import brotli
, certifi
, websockets
68 from .socks
import ProxyType
, sockssocket
71 def register_socks_protocols():
72 # "Register" SOCKS protocols
73 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
74 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
75 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
76 if scheme
not in compat_urlparse
.uses_netloc
:
77 compat_urlparse
.uses_netloc
.append(scheme
)
80 # This is not clearly defined otherwise
81 compiled_regex_type
= type(re
.compile(''))
84 def random_user_agent():
85 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
126 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
129 SUPPORTED_ENCODINGS
= [
133 SUPPORTED_ENCODINGS
.append('br')
136 'User-Agent': random_user_agent(),
137 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
138 'Accept-Language': 'en-us,en;q=0.5',
139 'Sec-Fetch-Mode': 'navigate',
144 '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',
148 NO_DEFAULT
= object()
150 ENGLISH_MONTH_NAMES
= [
151 'January', 'February', 'March', 'April', 'May', 'June',
152 'July', 'August', 'September', 'October', 'November', 'December']
155 'en': ENGLISH_MONTH_NAMES
,
157 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
158 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
162 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
163 'flv', 'f4v', 'f4a', 'f4b',
164 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
165 'mkv', 'mka', 'mk3d',
174 'f4f', 'f4m', 'm3u8', 'smil')
176 # needed for sanitizing filenames in restricted mode
177 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
178 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
179 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
209 '%Y-%m-%d %H:%M:%S.%f',
210 '%Y-%m-%d %H:%M:%S:%f',
213 '%Y-%m-%dT%H:%M:%SZ',
214 '%Y-%m-%dT%H:%M:%S.%fZ',
215 '%Y-%m-%dT%H:%M:%S.%f0Z',
217 '%Y-%m-%dT%H:%M:%S.%f',
220 '%b %d %Y at %H:%M:%S',
222 '%B %d %Y at %H:%M:%S',
226 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
227 DATE_FORMATS_DAY_FIRST
.extend([
236 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
237 DATE_FORMATS_MONTH_FIRST
.extend([
245 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
246 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>(?P
<json_ld
>.+?
)</script
>'
248 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
252 def preferredencoding():
253 """Get preferred encoding.
255 Returns the best encoding scheme for the system, based on
256 locale.getpreferredencoding() and some further tweaks.
259 pref = locale.getpreferredencoding()
267 def write_json_file(obj, fn):
268 """ Encode obj as JSON and write it to fn, atomically if possible """
270 tf = tempfile.NamedTemporaryFile(
271 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
272 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
276 json.dump(obj, tf, ensure_ascii=False)
277 if sys.platform == 'win32
':
278 # Need to remove existing file on Windows, else os.rename raises
279 # WindowsError or FileExistsError.
280 with contextlib.suppress(OSError):
282 with contextlib.suppress(OSError):
285 os.chmod(tf.name, 0o666 & ~mask)
286 os.rename(tf.name, fn)
288 with contextlib.suppress(OSError):
293 def find_xpath_attr(node, xpath, key, val=None):
294 """ Find the xpath xpath[@key=val] """
295 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
296 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
297 return node.find(expr)
299 # On python2.6 the xml.etree.ElementTree.Element methods don't support
300 # the namespace parameter
303 def xpath_with_ns(path
, ns_map
):
304 components
= [c
.split(':') for c
in path
.split('/')]
308 replaced
.append(c
[0])
311 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
312 return '/'.join(replaced
)
315 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
316 def _find_xpath(xpath
):
317 return node
.find(xpath
)
319 if isinstance(xpath
, (str, compat_str
)):
320 n
= _find_xpath(xpath
)
328 if default
is not NO_DEFAULT
:
331 name
= xpath
if name
is None else name
332 raise ExtractorError('Could not find XML element %s' % name
)
338 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
339 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
340 if n
is None or n
== default
:
343 if default
is not NO_DEFAULT
:
346 name
= xpath
if name
is None else name
347 raise ExtractorError('Could not find XML element\'s text %s' % name
)
353 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
354 n
= find_xpath_attr(node
, xpath
, key
)
356 if default
is not NO_DEFAULT
:
359 name
= f
'{xpath}[@{key}]' if name
is None else name
360 raise ExtractorError('Could not find XML attribute %s' % name
)
366 def get_element_by_id(id, html
, **kwargs
):
367 """Return the content of the tag with the specified ID in the passed HTML document"""
368 return get_element_by_attribute('id', id, html
, **kwargs
)
371 def get_element_html_by_id(id, html
, **kwargs
):
372 """Return the html of the tag with the specified ID in the passed HTML document"""
373 return get_element_html_by_attribute('id', id, html
, **kwargs
)
376 def get_element_by_class(class_name
, html
):
377 """Return the content of the first tag with the specified class in the passed HTML document"""
378 retval
= get_elements_by_class(class_name
, html
)
379 return retval
[0] if retval
else None
382 def get_element_html_by_class(class_name
, html
):
383 """Return the html of the first tag with the specified class in the passed HTML document"""
384 retval
= get_elements_html_by_class(class_name
, html
)
385 return retval
[0] if retval
else None
388 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
389 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
390 return retval
[0] if retval
else None
393 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
394 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
395 return retval
[0] if retval
else None
398 def get_elements_by_class(class_name
, html
, **kargs
):
399 """Return the content of all tags with the specified class in the passed HTML document as a list"""
400 return get_elements_by_attribute(
401 'class', r
'[^\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
402 html, escape_value=False)
405 def get_elements_html_by_class(class_name, html):
406 """Return the html of all tags with the specified class in the passed HTML document as a list"""
407 return get_elements_html_by_attribute(
408 'class', r'[^
\'"]*(?<=[\'"\s
])%s(?
=[\'"\s])[^\'"]*' % re.escape(class_name),
409 html, escape_value=False)
412 def get_elements_by_attribute(*args, **kwargs):
413 """Return the content of the tag with the specified attribute in the passed HTML document"""
414 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
417 def get_elements_html_by_attribute(*args, **kwargs):
418 """Return the html of the tag with the specified attribute in the passed HTML document"""
419 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
422 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
424 Return the text (content) and the html (whole) of the tag with the specified
425 attribute in the passed HTML document
428 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
430 value = re.escape(value) if escape_value else value
432 partial_element_re = rf'''(?x
)
433 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
434 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
435 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
438 for m in re.finditer(partial_element_re, html):
439 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
442 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
447 class HTMLBreakOnClosingTagParser(compat_HTMLParser):
449 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
450 closing tag for the first opening tag it has encountered, and can be used
454 class HTMLBreakOnClosingTagException(Exception):
458 self.tagstack = collections.deque()
459 compat_HTMLParser.__init__(self)
464 def __exit__(self, *_):
468 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
469 # so data remains buffered; we no longer have any interest in it, thus
470 # override this method to discard it
473 def handle_starttag(self, tag, _):
474 self.tagstack.append(tag)
476 def handle_endtag(self, tag):
477 if not self.tagstack:
478 raise compat_HTMLParseError('no tags
in the stack
')
480 inner_tag = self.tagstack.pop()
484 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
485 if not self.tagstack:
486 raise self.HTMLBreakOnClosingTagException()
489 def get_element_text_and_html_by_tag(tag, html):
491 For the first element with the specified tag in the passed HTML document
492 return its' content (text
) and the whole
element (html
)
494 def find_or_raise(haystack, needle, exc):
496 return haystack.index(needle)
499 closing_tag = f'</{tag}>'
500 whole_start = find_or_raise(
501 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
502 content_start = find_or_raise(
503 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
504 content_start += whole_start + 1
505 with HTMLBreakOnClosingTagParser() as parser:
506 parser.feed(html[whole_start:content_start])
507 if not parser.tagstack or parser.tagstack[0] != tag:
508 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
509 offset = content_start
510 while offset < len(html):
511 next_closing_tag_start = find_or_raise(
512 html[offset:], closing_tag,
513 compat_HTMLParseError(f'closing {tag} tag not found'))
514 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
516 parser.feed(html[offset:offset + next_closing_tag_end])
517 offset += next_closing_tag_end
518 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
519 return html[content_start:offset + next_closing_tag_start], \
520 html[whole_start:offset + next_closing_tag_end]
521 raise compat_HTMLParseError('unexpected end of html')
524 class HTMLAttributeParser(compat_HTMLParser):
525 """Trivial HTML parser to gather the attributes
for a single element
"""
529 compat_HTMLParser.__init__(self)
531 def handle_starttag(self, tag, attrs):
532 self.attrs = dict(attrs)
535 class HTMLListAttrsParser(compat_HTMLParser):
536 """HTML parser to gather the attributes
for the elements of a
list"""
539 compat_HTMLParser.__init__(self)
543 def handle_starttag(self, tag, attrs):
544 if tag == 'li' and self._level == 0:
545 self.items.append(dict(attrs))
548 def handle_endtag(self, tag):
552 def extract_attributes(html_element):
553 """Given a string
for an HTML element such
as
555 a
="foo" B
="bar" c
="&98;az" d
=boz
556 empty
= noval entity
="&"
559 Decode
and return a dictionary of attributes
.
561 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
562 'empty': '', 'noval': None, 'entity': '&',
563 'sq': '"', 'dq': '\''
566 parser = HTMLAttributeParser()
567 with contextlib.suppress(compat_HTMLParseError):
568 parser.feed(html_element)
573 def parse_list(webpage):
574 """Given a string
for an series of HTML
<li
> elements
,
575 return a dictionary of their attributes
"""
576 parser = HTMLListAttrsParser()
582 def clean_html(html):
583 """Clean an HTML snippet into a readable string
"""
585 if html is None: # Convenience for sanitizing descriptions etc.
588 html = re.sub(r'\s+', ' ', html)
589 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
590 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
592 html = re.sub('<.*?>', '', html)
593 # Replace html entities
594 html = unescapeHTML(html)
598 class LenientJSONDecoder(json.JSONDecoder):
599 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
600 self.transform_source, self.ignore_extra = transform_source, ignore_extra
601 super().__init__(*args, **kwargs)
604 if self.transform_source:
605 s = self.transform_source(s)
606 if self.ignore_extra:
607 return self.raw_decode(s.lstrip())[0]
608 return super().decode(s)
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
':
624 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
625 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
627 for attempt in range(2):
630 if sys.platform == 'win32
':
631 # FIXME: An exclusive lock also locks the file from being read.
632 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
633 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
634 raise LockingUnsupportedError()
635 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
637 stream
= open(filename
, open_mode
)
638 return stream
, filename
639 except OSError as err
:
640 if attempt
or err
.errno
in (errno
.EACCES
,):
642 old_filename
, filename
= filename
, sanitize_path(filename
)
643 if old_filename
== filename
:
647 def timeconvert(timestr
):
648 """Convert RFC 2822 defined time string into system timestamp"""
650 timetuple
= email
.utils
.parsedate_tz(timestr
)
651 if timetuple
is not None:
652 timestamp
= email
.utils
.mktime_tz(timetuple
)
656 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
657 """Sanitizes a string so it could be used as part of a filename.
658 @param restricted Use a stricter subset of allowed characters
659 @param is_id Whether this is an ID that should be kept unchanged if possible.
660 If unset, yt-dlp's new sanitization rules are in effect
665 def replace_insane(char
):
666 if restricted
and char
in ACCENT_CHARS
:
667 return ACCENT_CHARS
[char
]
668 elif not restricted
and char
== '\n':
670 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
673 return '' if restricted
else '\''
675 return '\0_\0-' if restricted
else '\0 \0-'
676 elif char
in '\\/|*<>':
678 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
682 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
683 result
= ''.join(map(replace_insane
, s
))
684 if is_id
is NO_DEFAULT
:
685 result
= re
.sub('(\0.)(?:(?=\\1)..)+', r
'\1', result
) # Remove repeated substitute chars
686 STRIP_RE
= '(?:\0.|[ _-])*'
687 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
688 result
= result
.replace('\0', '') or '_'
691 while '__' in result
:
692 result
= result
.replace('__', '_')
693 result
= result
.strip('_')
694 # Common case of "Foreign band name - English song title"
695 if restricted
and result
.startswith('-_'):
697 if result
.startswith('-'):
698 result
= '_' + result
[len('-'):]
699 result
= result
.lstrip('.')
705 def sanitize_path(s
, force
=False):
706 """Sanitizes and normalizes path on Windows"""
707 if sys
.platform
== 'win32':
709 drive_or_unc
, _
= os
.path
.splitdrive(s
)
715 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
719 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
720 for path_part
in norm_path
]
722 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
723 elif force
and s
and s
[0] == os
.path
.sep
:
724 sanitized_path
.insert(0, os
.path
.sep
)
725 return os
.path
.join(*sanitized_path
)
728 def sanitize_url(url
):
729 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
730 # the number of unwanted failures due to missing protocol
733 elif url
.startswith('//'):
734 return 'http:%s' % url
735 # Fix some common typos seen so far
737 # https://github.com/ytdl-org/youtube-dl/issues/15649
738 (r
'^httpss://', r
'https://'),
739 # https://bx1.be/lives/direct-tv/
740 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
742 for mistake
, fixup
in COMMON_TYPOS
:
743 if re
.match(mistake
, url
):
744 return re
.sub(mistake
, fixup
, url
)
748 def extract_basic_auth(url
):
749 parts
= compat_urlparse
.urlsplit(url
)
750 if parts
.username
is None:
752 url
= compat_urlparse
.urlunsplit(parts
._replace
(netloc
=(
753 parts
.hostname
if parts
.port
is None
754 else '%s:%d' % (parts
.hostname
, parts
.port
))))
755 auth_payload
= base64
.b64encode(
756 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
757 return url
, f
'Basic {auth_payload.decode()}'
760 def sanitized_Request(url
, *args
, **kwargs
):
761 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
762 if auth_header
is not None:
763 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
764 headers
['Authorization'] = auth_header
765 return compat_urllib_request
.Request(url
, *args
, **kwargs
)
769 """Expand shell variables and ~"""
770 return os
.path
.expandvars(compat_expanduser(s
))
773 def orderedSet(iterable
):
774 """ Remove all duplicates from the input iterable """
782 def _htmlentity_transform(entity_with_semicolon
):
783 """Transforms an HTML entity to a character."""
784 entity
= entity_with_semicolon
[:-1]
786 # Known non-numeric HTML entity
787 if entity
in compat_html_entities
.name2codepoint
:
788 return compat_chr(compat_html_entities
.name2codepoint
[entity
])
790 # TODO: HTML5 allows entities without a semicolon. For example,
791 # 'Éric' should be decoded as 'Éric'.
792 if entity_with_semicolon
in compat_html_entities_html5
:
793 return compat_html_entities_html5
[entity_with_semicolon
]
795 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
797 numstr
= mobj
.group(1)
798 if numstr
.startswith('x'):
800 numstr
= '0%s' % numstr
803 # See https://github.com/ytdl-org/youtube-dl/issues/7518
804 with contextlib
.suppress(ValueError):
805 return compat_chr(int(numstr
, base
))
807 # Unknown entity in name, return its literal representation
808 return '&%s;' % entity
814 assert isinstance(s
, str)
817 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
820 def escapeHTML(text
):
823 .replace('&', '&')
824 .replace('<', '<')
825 .replace('>', '>')
826 .replace('"', '"')
827 .replace("'", ''')
831 def process_communicate_or_kill(p
, *args
, **kwargs
):
832 write_string('DeprecationWarning: yt_dlp.utils.process_communicate_or_kill is deprecated '
833 'and may be removed in a future version. Use yt_dlp.utils.Popen.communicate_or_kill instead')
834 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
837 class Popen(subprocess
.Popen
):
838 if sys
.platform
== 'win32':
839 _startupinfo
= subprocess
.STARTUPINFO()
840 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
844 def __init__(self
, *args
, **kwargs
):
845 super().__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
847 def communicate_or_kill(self
, *args
, **kwargs
):
849 return self
.communicate(*args
, **kwargs
)
850 except BaseException
: # Including KeyboardInterrupt
856 def get_subprocess_encoding():
857 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
858 # For subprocess calls, encode with locale encoding
859 # Refer to http://stackoverflow.com/a/9951851/35070
860 encoding
= preferredencoding()
862 encoding
= sys
.getfilesystemencoding()
868 def encodeFilename(s
, for_subprocess
=False):
869 assert isinstance(s
, str)
873 def decodeFilename(b
, for_subprocess
=False):
877 def encodeArgument(s
):
878 # Legacy code that uses byte strings
879 # Uncomment the following line after fixing all post processors
880 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
881 return s
if isinstance(s
, str) else s
.decode('ascii')
884 def decodeArgument(b
):
888 def decodeOption(optval
):
891 if isinstance(optval
, bytes):
892 optval
= optval
.decode(preferredencoding())
894 assert isinstance(optval
, compat_str
)
898 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
901 def timetuple_from_msec(msec
):
902 secs
, msec
= divmod(msec
, 1000)
903 mins
, secs
= divmod(secs
, 60)
904 hrs
, mins
= divmod(mins
, 60)
905 return _timetuple(hrs
, mins
, secs
, msec
)
908 def formatSeconds(secs
, delim
=':', msec
=False):
909 time
= timetuple_from_msec(secs
* 1000)
911 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
913 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
915 ret
= '%d' % time
.seconds
916 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
919 def _ssl_load_windows_store_certs(ssl_context
, storename
):
920 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
922 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
923 if encoding
== 'x509_asn' and (
924 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
925 except PermissionError
:
928 with contextlib
.suppress(ssl
.SSLError
):
929 ssl_context
.load_verify_locations(cadata
=cert
)
932 def make_HTTPS_handler(params
, **kwargs
):
933 opts_check_certificate
= not params
.get('nocheckcertificate')
934 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
935 context
.check_hostname
= opts_check_certificate
936 if params
.get('legacyserverconnect'):
937 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
938 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
939 context
.set_ciphers('DEFAULT')
941 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
942 if opts_check_certificate
:
943 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
944 context
.load_verify_locations(cafile
=certifi
.where())
946 context
.load_default_certs()
947 # Work around the issue in load_default_certs when there are bad certificates. See:
948 # https://github.com/yt-dlp/yt-dlp/issues/1060,
949 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
951 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
952 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
953 for storename
in ('CA', 'ROOT'):
954 _ssl_load_windows_store_certs(context
, storename
)
955 context
.set_default_verify_paths()
957 client_certfile
= params
.get('client_certificate')
960 context
.load_cert_chain(
961 client_certfile
, keyfile
=params
.get('client_certificate_key'),
962 password
=params
.get('client_certificate_password'))
964 raise YoutubeDLError('Unable to load client certificate')
966 # Some servers may reject requests if ALPN extension is not sent. See:
967 # https://github.com/python/cpython/issues/85140
968 # https://github.com/yt-dlp/yt-dlp/issues/3878
969 with contextlib
.suppress(NotImplementedError):
970 context
.set_alpn_protocols(['http/1.1'])
972 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
975 def bug_reports_message(before
=';'):
976 msg
= ('please report this issue on https://github.com/yt-dlp/yt-dlp/issues?q= , '
977 'filling out the appropriate issue template. '
978 'Confirm you are on the latest version using yt-dlp -U')
980 before
= before
.rstrip()
981 if not before
or before
.endswith(('.', '!', '?')):
982 msg
= msg
[0].title() + msg
[1:]
984 return (before
+ ' ' if before
else '') + msg
987 class YoutubeDLError(Exception):
988 """Base exception for YoutubeDL errors."""
991 def __init__(self
, msg
=None):
994 elif self
.msg
is None:
995 self
.msg
= type(self
).__name
__
996 super().__init
__(self
.msg
)
999 network_exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
1000 if hasattr(ssl
, 'CertificateError'):
1001 network_exceptions
.append(ssl
.CertificateError
)
1002 network_exceptions
= tuple(network_exceptions
)
1005 class ExtractorError(YoutubeDLError
):
1006 """Error during info extraction."""
1008 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1009 """ tb, if given, is the original traceback (so that it can be printed out).
1010 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1012 if sys
.exc_info()[0] in network_exceptions
:
1015 self
.orig_msg
= str(msg
)
1017 self
.expected
= expected
1019 self
.video_id
= video_id
1021 self
.exc_info
= sys
.exc_info() # preserve original exception
1023 super().__init
__(''.join((
1024 format_field(ie
, template
='[%s] '),
1025 format_field(video_id
, template
='%s: '),
1027 format_field(cause
, template
=' (caused by %r)'),
1028 '' if expected
else bug_reports_message())))
1030 def format_traceback(self
):
1031 return join_nonempty(
1032 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1033 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1037 class UnsupportedError(ExtractorError
):
1038 def __init__(self
, url
):
1040 'Unsupported URL: %s' % url
, expected
=True)
1044 class RegexNotFoundError(ExtractorError
):
1045 """Error when a regex didn't match"""
1049 class GeoRestrictedError(ExtractorError
):
1050 """Geographic restriction Error exception.
1052 This exception may be thrown when a video is not available from your
1053 geographic location due to geographic restrictions imposed by a website.
1056 def __init__(self
, msg
, countries
=None, **kwargs
):
1057 kwargs
['expected'] = True
1058 super().__init
__(msg
, **kwargs
)
1059 self
.countries
= countries
1062 class DownloadError(YoutubeDLError
):
1063 """Download Error exception.
1065 This exception may be thrown by FileDownloader objects if they are not
1066 configured to continue on errors. They will contain the appropriate
1070 def __init__(self
, msg
, exc_info
=None):
1071 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1072 super().__init
__(msg
)
1073 self
.exc_info
= exc_info
1076 class EntryNotInPlaylist(YoutubeDLError
):
1077 """Entry not in playlist exception.
1079 This exception will be thrown by YoutubeDL when a requested entry
1080 is not found in the playlist info_dict
1082 msg
= 'Entry not found in info'
1085 class SameFileError(YoutubeDLError
):
1086 """Same File exception.
1088 This exception will be thrown by FileDownloader objects if they detect
1089 multiple files would have to be downloaded to the same file on disk.
1091 msg
= 'Fixed output name but more than one file to download'
1093 def __init__(self
, filename
=None):
1094 if filename
is not None:
1095 self
.msg
+= f
': {filename}'
1096 super().__init
__(self
.msg
)
1099 class PostProcessingError(YoutubeDLError
):
1100 """Post Processing exception.
1102 This exception may be raised by PostProcessor's .run() method to
1103 indicate an error in the postprocessing task.
1107 class DownloadCancelled(YoutubeDLError
):
1108 """ Exception raised when the download queue should be interrupted """
1109 msg
= 'The download was cancelled'
1112 class ExistingVideoReached(DownloadCancelled
):
1113 """ --break-on-existing triggered """
1114 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1117 class RejectedVideoReached(DownloadCancelled
):
1118 """ --break-on-reject triggered """
1119 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1122 class MaxDownloadsReached(DownloadCancelled
):
1123 """ --max-downloads limit has been reached. """
1124 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1127 class ReExtractInfo(YoutubeDLError
):
1128 """ Video info needs to be re-extracted. """
1130 def __init__(self
, msg
, expected
=False):
1131 super().__init
__(msg
)
1132 self
.expected
= expected
1135 class ThrottledDownload(ReExtractInfo
):
1136 """ Download speed below --throttled-rate. """
1137 msg
= 'The download speed is below throttle limit'
1140 super().__init
__(self
.msg
, expected
=False)
1143 class UnavailableVideoError(YoutubeDLError
):
1144 """Unavailable Format exception.
1146 This exception will be thrown when a video is requested
1147 in a format that is not available for that video.
1149 msg
= 'Unable to download video'
1151 def __init__(self
, err
=None):
1153 self
.msg
+= f
': {err}'
1154 super().__init
__(self
.msg
)
1157 class ContentTooShortError(YoutubeDLError
):
1158 """Content Too Short exception.
1160 This exception may be raised by FileDownloader objects when a file they
1161 download is too small for what the server announced first, indicating
1162 the connection was probably interrupted.
1165 def __init__(self
, downloaded
, expected
):
1166 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1168 self
.downloaded
= downloaded
1169 self
.expected
= expected
1172 class XAttrMetadataError(YoutubeDLError
):
1173 def __init__(self
, code
=None, msg
='Unknown error'):
1174 super().__init
__(msg
)
1178 # Parsing code and msg
1179 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1180 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1181 self
.reason
= 'NO_SPACE'
1182 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1183 self
.reason
= 'VALUE_TOO_LONG'
1185 self
.reason
= 'NOT_SUPPORTED'
1188 class XAttrUnavailableError(YoutubeDLError
):
1192 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1193 hc
= http_class(*args
, **kwargs
)
1194 source_address
= ydl_handler
._params
.get('source_address')
1196 if source_address
is not None:
1197 # This is to workaround _create_connection() from socket where it will try all
1198 # address data from getaddrinfo() including IPv6. This filters the result from
1199 # getaddrinfo() based on the source_address value.
1200 # This is based on the cpython socket.create_connection() function.
1201 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1202 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1203 host
, port
= address
1205 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1206 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1207 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1208 if addrs
and not ip_addrs
:
1209 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1211 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1212 % (ip_version
, source_address
[0]))
1213 for res
in ip_addrs
:
1214 af
, socktype
, proto
, canonname
, sa
= res
1217 sock
= socket
.socket(af
, socktype
, proto
)
1218 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1219 sock
.settimeout(timeout
)
1220 sock
.bind(source_address
)
1222 err
= None # Explicitly break reference cycle
1224 except OSError as _
:
1226 if sock
is not None:
1231 raise OSError('getaddrinfo returns an empty list')
1232 if hasattr(hc
, '_create_connection'):
1233 hc
._create
_connection
= _create_connection
1234 hc
.source_address
= (source_address
, 0)
1239 def handle_youtubedl_headers(headers
):
1240 filtered_headers
= headers
1242 if 'Youtubedl-no-compression' in filtered_headers
:
1243 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1244 del filtered_headers
['Youtubedl-no-compression']
1246 return filtered_headers
1249 class YoutubeDLHandler(compat_urllib_request
.HTTPHandler
):
1250 """Handler for HTTP requests and responses.
1252 This class, when installed with an OpenerDirector, automatically adds
1253 the standard headers to every HTTP request and handles gzipped and
1254 deflated responses from web servers. If compression is to be avoided in
1255 a particular request, the original request in the program code only has
1256 to include the HTTP header "Youtubedl-no-compression", which will be
1257 removed before making the real request.
1259 Part of this code was copied from:
1261 http://techknack.net/python-urllib2-handlers/
1263 Andrew Rowls, the author of that code, agreed to release it to the
1267 def __init__(self
, params
, *args
, **kwargs
):
1268 compat_urllib_request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1269 self
._params
= params
1271 def http_open(self
, req
):
1272 conn_class
= compat_http_client
.HTTPConnection
1274 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1276 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1277 del req
.headers
['Ytdl-socks-proxy']
1279 return self
.do_open(functools
.partial(
1280 _create_http_connection
, self
, conn_class
, False),
1288 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1290 return zlib
.decompress(data
)
1296 return brotli
.decompress(data
)
1298 def http_request(self
, req
):
1299 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1300 # always respected by websites, some tend to give out URLs with non percent-encoded
1301 # non-ASCII characters (see telemb.py, ard.py [#3412])
1302 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1303 # To work around aforementioned issue we will replace request's original URL with
1304 # percent-encoded one
1305 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1306 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1307 url
= req
.get_full_url()
1308 url_escaped
= escape_url(url
)
1310 # Substitute URL if any change after escaping
1311 if url
!= url_escaped
:
1312 req
= update_Request(req
, url
=url_escaped
)
1314 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1315 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1316 # The dict keys are capitalized because of this bug by urllib
1317 if h
.capitalize() not in req
.headers
:
1318 req
.add_header(h
, v
)
1320 if 'Accept-encoding' not in req
.headers
:
1321 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1323 req
.headers
= handle_youtubedl_headers(req
.headers
)
1327 def http_response(self
, req
, resp
):
1330 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1331 content
= resp
.read()
1332 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1334 uncompressed
= io
.BytesIO(gz
.read())
1335 except OSError as original_ioerror
:
1336 # There may be junk add the end of the file
1337 # See http://stackoverflow.com/q/4928560/35070 for details
1338 for i
in range(1, 1024):
1340 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1341 uncompressed
= io
.BytesIO(gz
.read())
1346 raise original_ioerror
1347 resp
= compat_urllib_request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1348 resp
.msg
= old_resp
.msg
1349 del resp
.headers
['Content-encoding']
1351 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1352 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1353 resp
= compat_urllib_request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1354 resp
.msg
= old_resp
.msg
1355 del resp
.headers
['Content-encoding']
1357 if resp
.headers
.get('Content-encoding', '') == 'br':
1358 resp
= compat_urllib_request
.addinfourl(
1359 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1360 resp
.msg
= old_resp
.msg
1361 del resp
.headers
['Content-encoding']
1362 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1363 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1364 if 300 <= resp
.code
< 400:
1365 location
= resp
.headers
.get('Location')
1367 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1368 location
= location
.encode('iso-8859-1').decode()
1369 location_escaped
= escape_url(location
)
1370 if location
!= location_escaped
:
1371 del resp
.headers
['Location']
1372 resp
.headers
['Location'] = location_escaped
1375 https_request
= http_request
1376 https_response
= http_response
1379 def make_socks_conn_class(base_class
, socks_proxy
):
1380 assert issubclass(base_class
, (
1381 compat_http_client
.HTTPConnection
, compat_http_client
.HTTPSConnection
))
1383 url_components
= compat_urlparse
.urlparse(socks_proxy
)
1384 if url_components
.scheme
.lower() == 'socks5':
1385 socks_type
= ProxyType
.SOCKS5
1386 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1387 socks_type
= ProxyType
.SOCKS4
1388 elif url_components
.scheme
.lower() == 'socks4a':
1389 socks_type
= ProxyType
.SOCKS4A
1391 def unquote_if_non_empty(s
):
1394 return compat_urllib_parse_unquote_plus(s
)
1398 url_components
.hostname
, url_components
.port
or 1080,
1400 unquote_if_non_empty(url_components
.username
),
1401 unquote_if_non_empty(url_components
.password
),
1404 class SocksConnection(base_class
):
1406 self
.sock
= sockssocket()
1407 self
.sock
.setproxy(*proxy_args
)
1408 if isinstance(self
.timeout
, (int, float)):
1409 self
.sock
.settimeout(self
.timeout
)
1410 self
.sock
.connect((self
.host
, self
.port
))
1412 if isinstance(self
, compat_http_client
.HTTPSConnection
):
1413 if hasattr(self
, '_context'): # Python > 2.6
1414 self
.sock
= self
._context
.wrap_socket(
1415 self
.sock
, server_hostname
=self
.host
)
1417 self
.sock
= ssl
.wrap_socket(self
.sock
)
1419 return SocksConnection
1422 class YoutubeDLHTTPSHandler(compat_urllib_request
.HTTPSHandler
):
1423 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1424 compat_urllib_request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1425 self
._https
_conn
_class
= https_conn_class
or compat_http_client
.HTTPSConnection
1426 self
._params
= params
1428 def https_open(self
, req
):
1430 conn_class
= self
._https
_conn
_class
1432 if hasattr(self
, '_context'): # python > 2.6
1433 kwargs
['context'] = self
._context
1434 if hasattr(self
, '_check_hostname'): # python 3.x
1435 kwargs
['check_hostname'] = self
._check
_hostname
1437 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1439 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1440 del req
.headers
['Ytdl-socks-proxy']
1443 return self
.do_open(
1444 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1445 except urllib
.error
.URLError
as e
:
1446 if (isinstance(e
.reason
, ssl
.SSLError
)
1447 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1448 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1452 class YoutubeDLCookieJar(compat_cookiejar
.MozillaCookieJar
):
1454 See [1] for cookie file format.
1456 1. https://curl.haxx.se/docs/http-cookies.html
1458 _HTTPONLY_PREFIX
= '#HttpOnly_'
1460 _HEADER
= '''# Netscape HTTP Cookie File
1461 # This file is generated by yt-dlp. Do not edit.
1464 _CookieFileEntry
= collections
.namedtuple(
1466 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1468 def __init__(self
, filename
=None, *args
, **kwargs
):
1469 super().__init
__(None, *args
, **kwargs
)
1470 if self
.is_path(filename
):
1471 filename
= os
.fspath(filename
)
1472 self
.filename
= filename
1475 def _true_or_false(cndn
):
1476 return 'TRUE' if cndn
else 'FALSE'
1480 return isinstance(file, (str, bytes, os
.PathLike
))
1482 @contextlib.contextmanager
1483 def open(self
, file, *, write
=False):
1484 if self
.is_path(file):
1485 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1492 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1495 if (not ignore_discard
and cookie
.discard
1496 or not ignore_expires
and cookie
.is_expired(now
)):
1498 name
, value
= cookie
.name
, cookie
.value
1500 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1501 # with no name, whereas http.cookiejar regards it as a
1502 # cookie with no value.
1503 name
, value
= '', name
1504 f
.write('%s\n' % '\t'.join((
1506 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1508 self
._true
_or
_false
(cookie
.secure
),
1509 str_or_none(cookie
.expires
, default
=''),
1513 def save(self
, filename
=None, *args
, **kwargs
):
1515 Save cookies to a file.
1516 Code is taken from CPython 3.6
1517 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1519 if filename
is None:
1520 if self
.filename
is not None:
1521 filename
= self
.filename
1523 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1525 # Store session cookies with `expires` set to 0 instead of an empty string
1527 if cookie
.expires
is None:
1530 with self
.open(filename
, write
=True) as f
:
1531 f
.write(self
._HEADER
)
1532 self
._really
_save
(f
, *args
, **kwargs
)
1534 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1535 """Load cookies from a file."""
1536 if filename
is None:
1537 if self
.filename
is not None:
1538 filename
= self
.filename
1540 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1542 def prepare_line(line
):
1543 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1544 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1545 # comments and empty lines are fine
1546 if line
.startswith('#') or not line
.strip():
1548 cookie_list
= line
.split('\t')
1549 if len(cookie_list
) != self
._ENTRY
_LEN
:
1550 raise compat_cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1551 cookie
= self
._CookieFileEntry
(*cookie_list
)
1552 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1553 raise compat_cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1557 with self
.open(filename
) as f
:
1560 cf
.write(prepare_line(line
))
1561 except compat_cookiejar
.LoadError
as e
:
1562 if f
'{line.strip()} '[0] in '[{"':
1563 raise compat_cookiejar
.LoadError(
1564 'Cookies file must be Netscape formatted, not JSON. See '
1565 'https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl')
1566 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1569 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1570 # Session cookies are denoted by either `expires` field set to
1571 # an empty string or 0. MozillaCookieJar only recognizes the former
1572 # (see [1]). So we need force the latter to be recognized as session
1573 # cookies on our own.
1574 # Session cookies may be important for cookies-based authentication,
1575 # e.g. usually, when user does not check 'Remember me' check box while
1576 # logging in on a site, some important cookies are stored as session
1577 # cookies so that not recognizing them will result in failed login.
1578 # 1. https://bugs.python.org/issue17164
1580 # Treat `expires=0` cookies as session cookies
1581 if cookie
.expires
== 0:
1582 cookie
.expires
= None
1583 cookie
.discard
= True
1586 class YoutubeDLCookieProcessor(compat_urllib_request
.HTTPCookieProcessor
):
1587 def __init__(self
, cookiejar
=None):
1588 compat_urllib_request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1590 def http_response(self
, request
, response
):
1591 return compat_urllib_request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1593 https_request
= compat_urllib_request
.HTTPCookieProcessor
.http_request
1594 https_response
= http_response
1597 class YoutubeDLRedirectHandler(compat_urllib_request
.HTTPRedirectHandler
):
1598 """YoutubeDL redirect handler
1600 The code is based on HTTPRedirectHandler implementation from CPython [1].
1602 This redirect handler solves two issues:
1603 - ensures redirect URL is always unicode under python 2
1604 - introduces support for experimental HTTP response status code
1605 308 Permanent Redirect [2] used by some sites [3]
1607 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1608 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1609 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1612 http_error_301
= http_error_303
= http_error_307
= http_error_308
= compat_urllib_request
.HTTPRedirectHandler
.http_error_302
1614 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1615 """Return a Request or None in response to a redirect.
1617 This is called by the http_error_30x methods when a
1618 redirection response is received. If a redirection should
1619 take place, return a new Request to allow http_error_30x to
1620 perform the redirect. Otherwise, raise HTTPError if no-one
1621 else should try to handle this url. Return None if you can't
1622 but another Handler might.
1624 m
= req
.get_method()
1625 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1626 or code
in (301, 302, 303) and m
== "POST")):
1627 raise compat_HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1628 # Strictly (according to RFC 2616), 301 or 302 in response to
1629 # a POST MUST NOT cause a redirection without confirmation
1630 # from the user (of urllib.request, in this case). In practice,
1631 # essentially all clients do redirect in this case, so we do
1634 # Be conciliant with URIs containing a space. This is mainly
1635 # redundant with the more complete encoding done in http_error_302(),
1636 # but it is kept for compatibility with other callers.
1637 newurl
= newurl
.replace(' ', '%20')
1639 CONTENT_HEADERS
= ("content-length", "content-type")
1640 # NB: don't use dict comprehension for python 2.6 compatibility
1641 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1643 # A 303 must either use GET or HEAD for subsequent request
1644 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1645 if code
== 303 and m
!= 'HEAD':
1647 # 301 and 302 redirects are commonly turned into a GET from a POST
1648 # for subsequent requests by browsers, so we'll do the same.
1649 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1650 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1651 if code
in (301, 302) and m
== 'POST':
1654 return compat_urllib_request
.Request(
1655 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1656 unverifiable
=True, method
=m
)
1659 def extract_timezone(date_str
):
1662 ^.{8,}? # >=8 char non-TZ prefix, if present
1663 (?P<tz>Z| # just the UTC Z, or
1664 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1665 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1666 [ ]? # optional space
1667 (?P<sign>\+|-) # +/-
1668 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1672 timezone
= datetime
.timedelta()
1674 date_str
= date_str
[:-len(m
.group('tz'))]
1675 if not m
.group('sign'):
1676 timezone
= datetime
.timedelta()
1678 sign
= 1 if m
.group('sign') == '+' else -1
1679 timezone
= datetime
.timedelta(
1680 hours
=sign
* int(m
.group('hours')),
1681 minutes
=sign
* int(m
.group('minutes')))
1682 return timezone
, date_str
1685 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1686 """ Return a UNIX timestamp from the given date """
1688 if date_str
is None:
1691 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1693 if timezone
is None:
1694 timezone
, date_str
= extract_timezone(date_str
)
1696 with contextlib
.suppress(ValueError):
1697 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1698 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1699 return calendar
.timegm(dt
.timetuple())
1702 def date_formats(day_first
=True):
1703 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1706 def unified_strdate(date_str
, day_first
=True):
1707 """Return a string with the date in the format YYYYMMDD"""
1709 if date_str
is None:
1713 date_str
= date_str
.replace(',', ' ')
1714 # Remove AM/PM + timezone
1715 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1716 _
, date_str
= extract_timezone(date_str
)
1718 for expression
in date_formats(day_first
):
1719 with contextlib
.suppress(ValueError):
1720 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1721 if upload_date
is None:
1722 timetuple
= email
.utils
.parsedate_tz(date_str
)
1724 with contextlib
.suppress(ValueError):
1725 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1726 if upload_date
is not None:
1727 return compat_str(upload_date
)
1730 def unified_timestamp(date_str
, day_first
=True):
1731 if date_str
is None:
1734 date_str
= re
.sub(r
'[,|]', '', date_str
)
1736 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1737 timezone
, date_str
= extract_timezone(date_str
)
1739 # Remove AM/PM + timezone
1740 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1742 # Remove unrecognized timezones from ISO 8601 alike timestamps
1743 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1745 date_str
= date_str
[:-len(m
.group('tz'))]
1747 # Python only supports microseconds, so remove nanoseconds
1748 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]+$', date_str
)
1750 date_str
= m
.group(1)
1752 for expression
in date_formats(day_first
):
1753 with contextlib
.suppress(ValueError):
1754 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1755 return calendar
.timegm(dt
.timetuple())
1756 timetuple
= email
.utils
.parsedate_tz(date_str
)
1758 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1761 def determine_ext(url
, default_ext
='unknown_video'):
1762 if url
is None or '.' not in url
:
1764 guess
= url
.partition('?')[0].rpartition('.')[2]
1765 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1767 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1768 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1769 return guess
.rstrip('/')
1774 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1775 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1778 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1780 Return a datetime object from a string.
1782 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1784 @param format strftime format of DATE
1785 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1786 auto: round to the unit provided in date_str (if applicable).
1788 auto_precision
= False
1789 if precision
== 'auto':
1790 auto_precision
= True
1791 precision
= 'microsecond'
1792 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1793 if date_str
in ('now', 'today'):
1795 if date_str
== 'yesterday':
1796 return today
- datetime
.timedelta(days
=1)
1798 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1800 if match
is not None:
1801 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1802 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1803 unit
= match
.group('unit')
1804 if unit
== 'month' or unit
== 'year':
1805 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1811 delta
= datetime
.timedelta(**{unit + 's': time}
)
1812 new_date
= start_time
+ delta
1814 return datetime_round(new_date
, unit
)
1817 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1820 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1822 Return a date object from a string using datetime_from_str
1824 @param strict Restrict allowed patterns to "YYYYMMDD" and
1825 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1827 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1828 raise ValueError(f
'Invalid date format "{date_str}"')
1829 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1832 def datetime_add_months(dt
, months
):
1833 """Increment/Decrement a datetime object by months."""
1834 month
= dt
.month
+ months
- 1
1835 year
= dt
.year
+ month
// 12
1836 month
= month
% 12 + 1
1837 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1838 return dt
.replace(year
, month
, day
)
1841 def datetime_round(dt
, precision
='day'):
1843 Round a datetime object's time to a specific precision
1845 if precision
== 'microsecond':
1854 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1855 timestamp
= calendar
.timegm(dt
.timetuple())
1856 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1859 def hyphenate_date(date_str
):
1861 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1862 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1863 if match
is not None:
1864 return '-'.join(match
.groups())
1870 """Represents a time interval between two dates"""
1872 def __init__(self
, start
=None, end
=None):
1873 """start and end must be strings in the format accepted by date"""
1874 if start
is not None:
1875 self
.start
= date_from_str(start
, strict
=True)
1877 self
.start
= datetime
.datetime
.min.date()
1879 self
.end
= date_from_str(end
, strict
=True)
1881 self
.end
= datetime
.datetime
.max.date()
1882 if self
.start
> self
.end
:
1883 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1887 """Returns a range that only contains the given day"""
1888 return cls(day
, day
)
1890 def __contains__(self
, date
):
1891 """Check if the date is in the range"""
1892 if not isinstance(date
, datetime
.date
):
1893 date
= date_from_str(date
)
1894 return self
.start
<= date
<= self
.end
1897 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1900 def platform_name():
1901 """ Returns the platform name as a compat_str """
1902 res
= platform
.platform()
1903 if isinstance(res
, bytes):
1904 res
= res
.decode(preferredencoding())
1906 assert isinstance(res
, compat_str
)
1911 def get_windows_version():
1912 ''' Get Windows version. returns () if it's not running on Windows '''
1913 if compat_os_name
== 'nt':
1914 return version_tuple(platform
.win32_ver()[1])
1919 def write_string(s
, out
=None, encoding
=None):
1920 assert isinstance(s
, str)
1921 out
= out
or sys
.stderr
1923 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
1924 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1926 enc
, buffer = None, out
1927 if 'b' in getattr(out
, 'mode', ''):
1928 enc
= encoding
or preferredencoding()
1929 elif hasattr(out
, 'buffer'):
1931 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1933 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
1937 def bytes_to_intlist(bs
):
1940 if isinstance(bs
[0], int): # Python 3
1943 return [ord(c
) for c
in bs
]
1946 def intlist_to_bytes(xs
):
1949 return compat_struct_pack('%dB' % len(xs
), *xs
)
1952 class LockingUnsupportedError(OSError):
1953 msg
= 'File locking is not supported'
1956 super().__init
__(self
.msg
)
1959 # Cross-platform file locking
1960 if sys
.platform
== 'win32':
1961 import ctypes
.wintypes
1964 class OVERLAPPED(ctypes
.Structure
):
1966 ('Internal', ctypes
.wintypes
.LPVOID
),
1967 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1968 ('Offset', ctypes
.wintypes
.DWORD
),
1969 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1970 ('hEvent', ctypes
.wintypes
.HANDLE
),
1973 kernel32
= ctypes
.windll
.kernel32
1974 LockFileEx
= kernel32
.LockFileEx
1975 LockFileEx
.argtypes
= [
1976 ctypes
.wintypes
.HANDLE
, # hFile
1977 ctypes
.wintypes
.DWORD
, # dwFlags
1978 ctypes
.wintypes
.DWORD
, # dwReserved
1979 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1980 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1981 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1983 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
1984 UnlockFileEx
= kernel32
.UnlockFileEx
1985 UnlockFileEx
.argtypes
= [
1986 ctypes
.wintypes
.HANDLE
, # hFile
1987 ctypes
.wintypes
.DWORD
, # dwReserved
1988 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1989 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1990 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1992 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
1993 whole_low
= 0xffffffff
1994 whole_high
= 0x7fffffff
1996 def _lock_file(f
, exclusive
, block
):
1997 overlapped
= OVERLAPPED()
1998 overlapped
.Offset
= 0
1999 overlapped
.OffsetHigh
= 0
2000 overlapped
.hEvent
= 0
2001 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2003 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2004 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2005 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2006 raise BlockingIOError('Locking file failed: %r' % ctypes
.FormatError())
2008 def _unlock_file(f
):
2009 assert f
._lock
_file
_overlapped
_p
2010 handle
= msvcrt
.get_osfhandle(f
.fileno())
2011 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2012 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2018 def _lock_file(f
, exclusive
, block
):
2019 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2021 flags |
= fcntl
.LOCK_NB
2023 fcntl
.flock(f
, flags
)
2024 except BlockingIOError
:
2026 except OSError: # AOSP does not have flock()
2027 fcntl
.lockf(f
, flags
)
2029 def _unlock_file(f
):
2031 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2033 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2037 def _lock_file(f
, exclusive
, block
):
2038 raise LockingUnsupportedError()
2040 def _unlock_file(f
):
2041 raise LockingUnsupportedError()
2047 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2048 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2049 raise NotImplementedError(mode
)
2050 self
.mode
, self
.block
= mode
, block
2052 writable
= any(f
in mode
for f
in 'wax+')
2053 readable
= any(f
in mode
for f
in 'r+')
2054 flags
= functools
.reduce(operator
.ior
, (
2055 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2056 getattr(os
, 'O_BINARY', 0), # Windows only
2057 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2058 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2059 os
.O_APPEND
if 'a' in mode
else 0,
2060 os
.O_EXCL
if 'x' in mode
else 0,
2061 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2064 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2066 def __enter__(self
):
2067 exclusive
= 'r' not in self
.mode
2069 _lock_file(self
.f
, exclusive
, self
.block
)
2074 if 'w' in self
.mode
:
2077 except OSError as e
:
2079 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2080 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2089 _unlock_file(self
.f
)
2093 def __exit__(self
, *_
):
2102 def __getattr__(self
, attr
):
2103 return getattr(self
.f
, attr
)
2110 def get_filesystem_encoding():
2111 encoding
= sys
.getfilesystemencoding()
2112 return encoding
if encoding
is not None else 'utf-8'
2115 def shell_quote(args
):
2117 encoding
= get_filesystem_encoding()
2119 if isinstance(a
, bytes):
2120 # We may get a filename encoded with 'encodeFilename'
2121 a
= a
.decode(encoding
)
2122 quoted_args
.append(compat_shlex_quote(a
))
2123 return ' '.join(quoted_args
)
2126 def smuggle_url(url
, data
):
2127 """ Pass additional data in a URL for internal use. """
2129 url
, idata
= unsmuggle_url(url
, {})
2131 sdata
= compat_urllib_parse_urlencode(
2132 {'__youtubedl_smuggle': json.dumps(data)}
)
2133 return url
+ '#' + sdata
2136 def unsmuggle_url(smug_url
, default
=None):
2137 if '#__youtubedl_smuggle' not in smug_url
:
2138 return smug_url
, default
2139 url
, _
, sdata
= smug_url
.rpartition('#')
2140 jsond
= compat_parse_qs(sdata
)['__youtubedl_smuggle'][0]
2141 data
= json
.loads(jsond
)
2145 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2146 """ Formats numbers with decimal sufixes like K, M, etc """
2147 num
, factor
= float_or_none(num
), float(factor
)
2148 if num
is None or num
< 0:
2150 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2151 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2152 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2154 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2155 converted
= num
/ (factor
** exponent
)
2156 return fmt
% (converted
, suffix
)
2159 def format_bytes(bytes):
2160 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2163 def lookup_unit_table(unit_table
, s
):
2164 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2166 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2169 num_str
= m
.group('num').replace(',', '.')
2170 mult
= unit_table
[m
.group('unit')]
2171 return int(float(num_str
) * mult
)
2174 def parse_filesize(s
):
2178 # The lower-case forms are of course incorrect and unofficial,
2179 # but we support those too
2196 'megabytes': 1000 ** 2,
2197 'mebibytes': 1024 ** 2,
2203 'gigabytes': 1000 ** 3,
2204 'gibibytes': 1024 ** 3,
2210 'terabytes': 1000 ** 4,
2211 'tebibytes': 1024 ** 4,
2217 'petabytes': 1000 ** 5,
2218 'pebibytes': 1024 ** 5,
2224 'exabytes': 1000 ** 6,
2225 'exbibytes': 1024 ** 6,
2231 'zettabytes': 1000 ** 7,
2232 'zebibytes': 1024 ** 7,
2238 'yottabytes': 1000 ** 8,
2239 'yobibytes': 1024 ** 8,
2242 return lookup_unit_table(_UNIT_TABLE
, s
)
2249 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2251 if re
.match(r
'^[\d,.]+$', s
):
2252 return str_to_int(s
)
2265 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2269 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2271 return str_to_int(mobj
.group(1))
2274 def parse_resolution(s
, *, lenient
=False):
2279 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2281 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2284 'width': int(mobj
.group('w')),
2285 'height': int(mobj
.group('h')),
2288 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2290 return {'height': int(mobj.group(1))}
2292 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2294 return {'height': int(mobj.group(1)) * 540}
2299 def parse_bitrate(s
):
2300 if not isinstance(s
, compat_str
):
2302 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2304 return int(mobj
.group(1))
2307 def month_by_name(name
, lang
='en'):
2308 """ Return the number of a month by (locale-independently) English name """
2310 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2313 return month_names
.index(name
) + 1
2318 def month_by_abbreviation(abbrev
):
2319 """ Return the number of a month by (locale-independently) English
2323 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2328 def fix_xml_ampersands(xml_str
):
2329 """Replace all the '&' by '&' in XML"""
2331 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2336 def setproctitle(title
):
2337 assert isinstance(title
, compat_str
)
2339 # ctypes in Jython is not complete
2340 # http://bugs.jython.org/issue2148
2341 if sys
.platform
.startswith('java'):
2345 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2349 # LoadLibrary in Windows Python 2.7.13 only expects
2350 # a bytestring, but since unicode_literals turns
2351 # every string into a unicode string, it fails.
2353 title_bytes
= title
.encode()
2354 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2355 buf
.value
= title_bytes
2357 libc
.prctl(15, buf
, 0, 0, 0)
2358 except AttributeError:
2359 return # Strange libc, just skip this
2362 def remove_start(s
, start
):
2363 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2366 def remove_end(s
, end
):
2367 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2370 def remove_quotes(s
):
2371 if s
is None or len(s
) < 2:
2373 for quote
in ('"', "'", ):
2374 if s
[0] == quote
and s
[-1] == quote
:
2379 def get_domain(url
):
2380 domain
= re
.match(r
'(?:https?:\/\/)?(?:www\.)?(?P<domain>[^\n\/]+\.[^\n\/]+)(?:\/(.*))?', url
)
2381 return domain
.group('domain') if domain
else None
2384 def url_basename(url
):
2385 path
= compat_urlparse
.urlparse(url
).path
2386 return path
.strip('/').split('/')[-1]
2390 return re
.match(r
'https?://[^?#&]+/', url
).group()
2393 def urljoin(base
, path
):
2394 if isinstance(path
, bytes):
2395 path
= path
.decode()
2396 if not isinstance(path
, compat_str
) or not path
:
2398 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2400 if isinstance(base
, bytes):
2401 base
= base
.decode()
2402 if not isinstance(base
, compat_str
) or not re
.match(
2403 r
'^(?:https?:)?//', base
):
2405 return compat_urlparse
.urljoin(base
, path
)
2408 class HEADRequest(compat_urllib_request
.Request
):
2409 def get_method(self
):
2413 class PUTRequest(compat_urllib_request
.Request
):
2414 def get_method(self
):
2418 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2419 if get_attr
and v
is not None:
2420 v
= getattr(v
, get_attr
, None)
2422 return int(v
) * invscale
// scale
2423 except (ValueError, TypeError, OverflowError):
2427 def str_or_none(v
, default
=None):
2428 return default
if v
is None else compat_str(v
)
2431 def str_to_int(int_str
):
2432 """ A more relaxed version of int_or_none """
2433 if isinstance(int_str
, int):
2435 elif isinstance(int_str
, compat_str
):
2436 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2437 return int_or_none(int_str
)
2440 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2444 return float(v
) * invscale
/ scale
2445 except (ValueError, TypeError):
2449 def bool_or_none(v
, default
=None):
2450 return v
if isinstance(v
, bool) else default
2453 def strip_or_none(v
, default
=None):
2454 return v
.strip() if isinstance(v
, compat_str
) else default
2457 def url_or_none(url
):
2458 if not url
or not isinstance(url
, compat_str
):
2461 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2464 def request_to_url(req
):
2465 if isinstance(req
, compat_urllib_request
.Request
):
2466 return req
.get_full_url()
2471 def strftime_or_none(timestamp
, date_format
, default
=None):
2472 datetime_object
= None
2474 if isinstance(timestamp
, (int, float)): # unix timestamp
2475 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2476 elif isinstance(timestamp
, compat_str
): # assume YYYYMMDD
2477 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2478 return datetime_object
.strftime(date_format
)
2479 except (ValueError, TypeError, AttributeError):
2483 def parse_duration(s
):
2484 if not isinstance(s
, str):
2490 days
, hours
, mins
, secs
, ms
= [None] * 5
2491 m
= re
.match(r
'''(?x)
2493 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2494 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2495 (?P<ms>[.:][0-9]+)?Z?$
2498 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2503 [0-9]+\s*y(?:ears?)?,?\s*
2506 [0-9]+\s*m(?:onths?)?,?\s*
2509 [0-9]+\s*w(?:eeks?)?,?\s*
2512 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2516 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2519 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2522 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2525 days
, hours
, mins
, secs
, ms
= m
.groups()
2527 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2529 hours
, mins
= m
.groups()
2534 ms
= ms
.replace(':', '.')
2535 return sum(float(part
or 0) * mult
for part
, mult
in (
2536 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2539 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2540 name
, real_ext
= os
.path
.splitext(filename
)
2542 f
'{name}.{ext}{real_ext}'
2543 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2544 else f
'{filename}.{ext}')
2547 def replace_extension(filename
, ext
, expected_real_ext
=None):
2548 name
, real_ext
= os
.path
.splitext(filename
)
2549 return '{}.{}'.format(
2550 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2554 def check_executable(exe
, args
=[]):
2555 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2556 args can be a list of arguments for a short output (like -version) """
2558 Popen([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
).communicate_or_kill()
2564 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2566 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2568 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2569 # SIGTTOU if yt-dlp is run in the background.
2570 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2572 [encodeArgument(exe
)] + args
, stdin
=subprocess
.PIPE
,
2573 stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
).communicate_or_kill()
2576 if isinstance(out
, bytes): # Python 2.x
2577 out
= out
.decode('ascii', 'ignore')
2581 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2582 assert isinstance(output
, compat_str
)
2583 if version_re
is None:
2584 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2585 m
= re
.search(version_re
, output
)
2592 def get_exe_version(exe
, args
=['--version'],
2593 version_re
=None, unrecognized
='present'):
2594 """ Returns the version of the specified executable,
2595 or False if the executable is not present """
2596 out
= _get_exe_version_output(exe
, args
)
2597 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2600 class LazyList(collections
.abc
.Sequence
):
2601 """Lazy immutable list from an iterable
2602 Note that slices of a LazyList are lists and not LazyList"""
2604 class IndexError(IndexError):
2607 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2608 self
._iterable
= iter(iterable
)
2609 self
._cache
= [] if _cache
is None else _cache
2610 self
._reversed
= reverse
2614 # We need to consume the entire iterable to iterate in reverse
2615 yield from self
.exhaust()
2617 yield from self
._cache
2618 for item
in self
._iterable
:
2619 self
._cache
.append(item
)
2623 self
._cache
.extend(self
._iterable
)
2624 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2628 """Evaluate the entire iterable"""
2629 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2632 def _reverse_index(x
):
2633 return None if x
is None else -(x
+ 1)
2635 def __getitem__(self
, idx
):
2636 if isinstance(idx
, slice):
2638 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2639 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2640 elif isinstance(idx
, int):
2642 idx
= self
._reverse
_index
(idx
)
2643 start
, stop
, step
= idx
, idx
, 0
2645 raise TypeError('indices must be integers or slices')
2646 if ((start
or 0) < 0 or (stop
or 0) < 0
2647 or (start
is None and step
< 0)
2648 or (stop
is None and step
> 0)):
2649 # We need to consume the entire iterable to be able to slice from the end
2650 # Obviously, never use this with infinite iterables
2653 return self
._cache
[idx
]
2654 except IndexError as e
:
2655 raise self
.IndexError(e
) from e
2656 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2658 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2660 return self
._cache
[idx
]
2661 except IndexError as e
:
2662 raise self
.IndexError(e
) from e
2666 self
[-1] if self
._reversed
else self
[0]
2667 except self
.IndexError:
2673 return len(self
._cache
)
2675 def __reversed__(self
):
2676 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2679 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2682 # repr and str should mimic a list. So we exhaust the iterable
2683 return repr(self
.exhaust())
2686 return repr(self
.exhaust())
2691 class IndexError(IndexError):
2695 # This is only useful for tests
2696 return len(self
.getslice())
2698 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2699 self
._pagefunc
= pagefunc
2700 self
._pagesize
= pagesize
2701 self
._pagecount
= float('inf')
2702 self
._use
_cache
= use_cache
2705 def getpage(self
, pagenum
):
2706 page_results
= self
._cache
.get(pagenum
)
2707 if page_results
is None:
2708 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2710 self
._cache
[pagenum
] = page_results
2713 def getslice(self
, start
=0, end
=None):
2714 return list(self
._getslice
(start
, end
))
2716 def _getslice(self
, start
, end
):
2717 raise NotImplementedError('This method must be implemented by subclasses')
2719 def __getitem__(self
, idx
):
2720 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2721 if not isinstance(idx
, int) or idx
< 0:
2722 raise TypeError('indices must be non-negative integers')
2723 entries
= self
.getslice(idx
, idx
+ 1)
2725 raise self
.IndexError()
2729 class OnDemandPagedList(PagedList
):
2730 """Download pages until a page with less than maximum results"""
2732 def _getslice(self
, start
, end
):
2733 for pagenum
in itertools
.count(start
// self
._pagesize
):
2734 firstid
= pagenum
* self
._pagesize
2735 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2736 if start
>= nextfirstid
:
2740 start
% self
._pagesize
2741 if firstid
<= start
< nextfirstid
2744 ((end
- 1) % self
._pagesize
) + 1
2745 if (end
is not None and firstid
<= end
<= nextfirstid
)
2749 page_results
= self
.getpage(pagenum
)
2751 self
._pagecount
= pagenum
- 1
2753 if startv
!= 0 or endv
is not None:
2754 page_results
= page_results
[startv
:endv
]
2755 yield from page_results
2757 # A little optimization - if current page is not "full", ie. does
2758 # not contain page_size videos then we can assume that this page
2759 # is the last one - there are no more ids on further pages -
2760 # i.e. no need to query again.
2761 if len(page_results
) + startv
< self
._pagesize
:
2764 # If we got the whole page, but the next page is not interesting,
2765 # break out early as well
2766 if end
== nextfirstid
:
2770 class InAdvancePagedList(PagedList
):
2771 """PagedList with total number of pages known in advance"""
2773 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2774 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2775 self
._pagecount
= pagecount
2777 def _getslice(self
, start
, end
):
2778 start_page
= start
// self
._pagesize
2779 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2780 skip_elems
= start
- start_page
* self
._pagesize
2781 only_more
= None if end
is None else end
- start
2782 for pagenum
in range(start_page
, end_page
):
2783 page_results
= self
.getpage(pagenum
)
2785 page_results
= page_results
[skip_elems
:]
2787 if only_more
is not None:
2788 if len(page_results
) < only_more
:
2789 only_more
-= len(page_results
)
2791 yield from page_results
[:only_more
]
2793 yield from page_results
2796 def uppercase_escape(s
):
2797 unicode_escape
= codecs
.getdecoder('unicode_escape')
2799 r
'\\U[0-9a-fA-F]{8}',
2800 lambda m
: unicode_escape(m
.group(0))[0],
2804 def lowercase_escape(s
):
2805 unicode_escape
= codecs
.getdecoder('unicode_escape')
2807 r
'\\u[0-9a-fA-F]{4}',
2808 lambda m
: unicode_escape(m
.group(0))[0],
2812 def escape_rfc3986(s
):
2813 """Escape non-ASCII characters as suggested by RFC 3986"""
2814 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2817 def escape_url(url
):
2818 """Escape URL as suggested by RFC 3986"""
2819 url_parsed
= compat_urllib_parse_urlparse(url
)
2820 return url_parsed
._replace
(
2821 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2822 path
=escape_rfc3986(url_parsed
.path
),
2823 params
=escape_rfc3986(url_parsed
.params
),
2824 query
=escape_rfc3986(url_parsed
.query
),
2825 fragment
=escape_rfc3986(url_parsed
.fragment
)
2830 return compat_parse_qs(compat_urllib_parse_urlparse(url
).query
)
2833 def read_batch_urls(batch_fd
):
2835 if not isinstance(url
, compat_str
):
2836 url
= url
.decode('utf-8', 'replace')
2837 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
2838 for bom
in BOM_UTF8
:
2839 if url
.startswith(bom
):
2840 url
= url
[len(bom
):]
2842 if not url
or url
.startswith(('#', ';', ']')):
2844 # "#" cannot be stripped out since it is part of the URI
2845 # However, it can be safely stipped out if follwing a whitespace
2846 return re
.split(r
'\s#', url
, 1)[0].rstrip()
2848 with contextlib
.closing(batch_fd
) as fd
:
2849 return [url
for url
in map(fixup
, fd
) if url
]
2852 def urlencode_postdata(*args
, **kargs
):
2853 return compat_urllib_parse_urlencode(*args
, **kargs
).encode('ascii')
2856 def update_url_query(url
, query
):
2859 parsed_url
= compat_urlparse
.urlparse(url
)
2860 qs
= compat_parse_qs(parsed_url
.query
)
2862 return compat_urlparse
.urlunparse(parsed_url
._replace
(
2863 query
=compat_urllib_parse_urlencode(qs
, True)))
2866 def update_Request(req
, url
=None, data
=None, headers
={}, query={}
):
2867 req_headers
= req
.headers
.copy()
2868 req_headers
.update(headers
)
2869 req_data
= data
or req
.data
2870 req_url
= update_url_query(url
or req
.get_full_url(), query
)
2871 req_get_method
= req
.get_method()
2872 if req_get_method
== 'HEAD':
2873 req_type
= HEADRequest
2874 elif req_get_method
== 'PUT':
2875 req_type
= PUTRequest
2877 req_type
= compat_urllib_request
.Request
2879 req_url
, data
=req_data
, headers
=req_headers
,
2880 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
2881 if hasattr(req
, 'timeout'):
2882 new_req
.timeout
= req
.timeout
2886 def _multipart_encode_impl(data
, boundary
):
2887 content_type
= 'multipart/form-data; boundary=%s' % boundary
2890 for k
, v
in data
.items():
2891 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
2892 if isinstance(k
, compat_str
):
2894 if isinstance(v
, compat_str
):
2896 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
2897 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
2898 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
2899 if boundary
.encode('ascii') in content
:
2900 raise ValueError('Boundary overlaps with data')
2903 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
2905 return out
, content_type
2908 def multipart_encode(data
, boundary
=None):
2910 Encode a dict to RFC 7578-compliant form-data
2913 A dict where keys and values can be either Unicode or bytes-like
2916 If specified a Unicode object, it's used as the boundary. Otherwise
2917 a random boundary is generated.
2919 Reference: https://tools.ietf.org/html/rfc7578
2921 has_specified_boundary
= boundary
is not None
2924 if boundary
is None:
2925 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
2928 out
, content_type
= _multipart_encode_impl(data
, boundary
)
2931 if has_specified_boundary
:
2935 return out
, content_type
2938 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
2939 for val
in map(d
.get
, variadic(key_or_keys
)):
2940 if val
is not None and (val
or not skip_false_values
):
2945 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
2948 val
= f(*args
, **kwargs
)
2949 except (AttributeError, KeyError, TypeError, IndexError, ZeroDivisionError):
2952 if expected_type
is None or isinstance(val
, expected_type
):
2956 def try_get(src
, getter
, expected_type
=None):
2957 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
2960 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
2961 return {k: v for k, v in dct.items() if cndn(k, v)}
2964 def merge_dicts(*dicts
):
2966 for a_dict
in dicts
:
2967 for k
, v
in a_dict
.items():
2968 if (v
is not None and k
not in merged
2969 or isinstance(v
, str) and merged
[k
] == ''):
2974 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
2975 return string
if isinstance(string
, compat_str
) else compat_str(string
, encoding
, errors
)
2987 TV_PARENTAL_GUIDELINES
= {
2997 def parse_age_limit(s
):
2998 # isinstance(False, int) is True. So type() must be used instead
2999 if type(s
) is int: # noqa: E721
3000 return s
if 0 <= s
<= 21 else None
3001 elif not isinstance(s
, str):
3003 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3005 return int(m
.group('age'))
3008 return US_RATINGS
[s
]
3009 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3011 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3015 def strip_jsonp(code
):
3018 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3019 (?:\s*&&\s*(?P=func_name))?
3020 \s*\(\s*(?P<callback_data>.*)\);?
3021 \s*?(?://[^\n]*)*$''',
3022 r
'\g<callback_data>', code
)
3025 def js_to_json(code
, vars={}):
3026 # vars is a dict of var, val pairs to substitute
3027 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3028 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3030 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3031 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3036 if v
in ('true', 'false', 'null'):
3038 elif v
in ('undefined', 'void 0'):
3040 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3043 if v
[0] in ("'", '"'):
3044 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3049 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3051 for regex
, base
in INTEGER_TABLE
:
3052 im
= re
.match(regex
, v
)
3054 i
= int(im
.group(1), base
)
3055 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3062 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3064 return re
.sub(r
'''(?sx)
3065 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3066 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3067 {comment}|,(?={skip}[\]}}])|
3068 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3069 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3072 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3075 def qualities(quality_ids
):
3076 """ Get a numeric quality value out of a list of possible values """
3079 return quality_ids
.index(qid
)
3085 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'after_move', 'post_process', 'after_video', 'playlist')
3089 'default': '%(title)s [%(id)s].%(ext)s',
3090 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3096 'description': 'description',
3097 'annotation': 'annotations.xml',
3098 'infojson': 'info.json',
3101 'pl_thumbnail': None,
3102 'pl_description': 'description',
3103 'pl_infojson': 'info.json',
3106 # As of [1] format syntax is:
3107 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3108 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3109 STR_FORMAT_RE_TMPL
= r
'''(?x)
3110 (?<!%)(?P<prefix>(?:%%)*)
3112 (?P<has_key>\((?P<key>{0})\))?
3114 (?P<conversion>[#0\-+ ]+)?
3116 (?P<precision>\.\d+)?
3117 (?P<len_mod>[hlL])? # unused in python
3118 {1} # conversion type
3123 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3126 def limit_length(s
, length
):
3127 """ Add ellipses to overly long strings """
3132 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3136 def version_tuple(v
):
3137 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3140 def is_outdated_version(version
, limit
, assume_new
=True):
3142 return not assume_new
3144 return version_tuple(version
) < version_tuple(limit
)
3146 return not assume_new
3149 def ytdl_is_updateable():
3150 """ Returns if yt-dlp can be updated with -U """
3152 from .update
import is_non_updateable
3154 return not is_non_updateable()
3157 def args_to_str(args
):
3158 # Get a short string representation for a subprocess command
3159 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3162 def error_to_compat_str(err
):
3166 def error_to_str(err
):
3167 return f
'{type(err).__name__}: {err}'
3170 def mimetype2ext(mt
):
3174 mt
, _
, params
= mt
.partition(';')
3179 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3180 # it's the most popular one
3181 'audio/mpeg': 'mp3',
3182 'audio/x-wav': 'wav',
3184 'audio/wave': 'wav',
3187 ext
= FULL_MAP
.get(mt
)
3193 'smptett+xml': 'tt',
3197 'x-mp4-fragmented': 'mp4',
3198 'x-ms-sami': 'sami',
3201 'x-mpegurl': 'm3u8',
3202 'vnd.apple.mpegurl': 'm3u8',
3206 'vnd.ms-sstr+xml': 'ism',
3210 'filmstrip+json': 'fs',
3214 _
, _
, subtype
= mt
.rpartition('/')
3215 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3226 _
, _
, suffix
= subtype
.partition('+')
3227 ext
= SUFFIX_MAP
.get(suffix
)
3231 return subtype
.replace('+', '.')
3234 def ext2mimetype(ext_or_url
):
3237 if '.' not in ext_or_url
:
3238 ext_or_url
= f
'file.{ext_or_url}'
3239 return mimetypes
.guess_type(ext_or_url
)[0]
3242 def parse_codecs(codecs_str
):
3243 # http://tools.ietf.org/html/rfc6381
3246 split_codecs
= list(filter(None, map(
3247 str.strip
, codecs_str
.strip().strip(',').split(','))))
3248 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3249 for full_codec
in split_codecs
:
3250 parts
= full_codec
.split('.')
3251 codec
= parts
[0].replace('0', '')
3252 if codec
in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3253 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3255 vcodec
= '.'.join(parts
[:4]) if codec
in ('vp9', 'av1', 'hvc1') else full_codec
3256 if codec
in ('dvh1', 'dvhe'):
3258 elif codec
== 'av1' and len(parts
) > 3 and parts
[3] == '10':
3260 elif full_codec
.replace('0', '').startswith('vp9.2'):
3262 elif codec
in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3265 elif codec
in ('stpp', 'wvtt',):
3269 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3270 if vcodec
or acodec
or scodec
:
3272 'vcodec': vcodec
or 'none',
3273 'acodec': acodec
or 'none',
3274 'dynamic_range': hdr
,
3275 **({'scodec': scodec}
if scodec
is not None else {}),
3277 elif len(split_codecs
) == 2:
3279 'vcodec': split_codecs
[0],
3280 'acodec': split_codecs
[1],
3285 def urlhandle_detect_ext(url_handle
):
3286 getheader
= url_handle
.headers
.get
3288 cd
= getheader('Content-Disposition')
3290 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3292 e
= determine_ext(m
.group('filename'), default_ext
=None)
3296 return mimetype2ext(getheader('Content-Type'))
3299 def encode_data_uri(data
, mime_type
):
3300 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3303 def age_restricted(content_limit
, age_limit
):
3304 """ Returns True iff the content should be blocked """
3306 if age_limit
is None: # No limit set
3308 if content_limit
is None:
3309 return False # Content available for everyone
3310 return age_limit
< content_limit
3313 def is_html(first_bytes
):
3314 """ Detect whether a file contains HTML by examining its first bytes. """
3317 (b
'\xef\xbb\xbf', 'utf-8'),
3318 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3319 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3320 (b
'\xff\xfe', 'utf-16-le'),
3321 (b
'\xfe\xff', 'utf-16-be'),
3325 for bom
, enc
in BOMS
:
3326 while first_bytes
.startswith(bom
):
3327 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3329 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3332 def determine_protocol(info_dict
):
3333 protocol
= info_dict
.get('protocol')
3334 if protocol
is not None:
3337 url
= sanitize_url(info_dict
['url'])
3338 if url
.startswith('rtmp'):
3340 elif url
.startswith('mms'):
3342 elif url
.startswith('rtsp'):
3345 ext
= determine_ext(url
)
3351 return compat_urllib_parse_urlparse(url
).scheme
3354 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3355 """ Render a list of rows, each as a list of values.
3356 Text after a \t will be right aligned """
3358 return len(remove_terminal_sequences(string
).replace('\t', ''))
3360 def get_max_lens(table
):
3361 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3363 def filter_using_list(row
, filterArray
):
3364 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3366 max_lens
= get_max_lens(data
) if hide_empty
else []
3367 header_row
= filter_using_list(header_row
, max_lens
)
3368 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3370 table
= [header_row
] + data
3371 max_lens
= get_max_lens(table
)
3374 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3375 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3377 for pos
, text
in enumerate(map(str, row
)):
3379 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3381 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3382 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3386 def _match_one(filter_part
, dct
, incomplete
):
3387 # TODO: Generalize code with YoutubeDL._build_format_filter
3388 STRING_OPERATORS
= {
3389 '*=': operator
.contains
,
3390 '^=': lambda attr
, value
: attr
.startswith(value
),
3391 '$=': lambda attr
, value
: attr
.endswith(value
),
3392 '~=': lambda attr
, value
: re
.search(value
, attr
),
3394 COMPARISON_OPERATORS
= {
3396 '<=': operator
.le
, # "<=" must be defined above "<"
3403 if isinstance(incomplete
, bool):
3404 is_incomplete
= lambda _
: incomplete
3406 is_incomplete
= lambda k
: k
in incomplete
3408 operator_rex
= re
.compile(r
'''(?x)
3410 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3412 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3415 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3416 m = operator_rex.fullmatch(filter_part.strip())
3419 unnegated_op = COMPARISON_OPERATORS[m['op']]
3421 op = lambda attr, value: not unnegated_op(attr, value)
3424 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3426 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3427 actual_value = dct.get(m['key'])
3428 numeric_comparison = None
3429 if isinstance(actual_value, (int, float)):
3430 # If the original field is a string and matching comparisonvalue is
3431 # a number we should respect the origin of the original field
3432 # and process comparison value as a string (see
3433 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3435 numeric_comparison = int(comparison_value)
3437 numeric_comparison = parse_filesize(comparison_value)
3438 if numeric_comparison is None:
3439 numeric_comparison = parse_filesize(f'{comparison_value}B')
3440 if numeric_comparison is None:
3441 numeric_comparison = parse_duration(comparison_value)
3442 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3443 raise ValueError('Operator %s only supports string values!' % m['op'])
3444 if actual_value is None:
3445 return is_incomplete(m['key']) or m['none_inclusive']
3446 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3449 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3450 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3452 operator_rex = re.compile(r'''(?x
)
3453 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3454 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3455 m = operator_rex.fullmatch(filter_part.strip())
3457 op = UNARY_OPERATORS[m.group('op')]
3458 actual_value = dct.get(m.group('key'))
3459 if is_incomplete(m.group('key')) and actual_value is None:
3461 return op(actual_value)
3463 raise ValueError('Invalid filter part %r' % filter_part)
3466 def match_str(filter_str, dct, incomplete=False):
3467 """ Filter a dictionary with a simple string syntax.
3468 @returns Whether the filter passes
3469 @param incomplete Set of keys that is expected to be missing from dct.
3470 Can be True/False to indicate all/none of the keys may be missing.
3471 All conditions on incomplete keys pass if the key is missing
3474 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3475 for filter_part in re.split(r'(?<!\\)&', filter_str))
3478 def match_filter_func(filters):
3481 filters = set(variadic(filters))
3483 interactive = '-' in filters
3487 def _match_func(info_dict, incomplete=False):
3488 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3489 return NO_DEFAULT if interactive and not incomplete else None
3491 video_title = info_dict.get('title') or info_dict.get('id') or 'video'
3492 filter_str = ') | ('.join(map(str.strip, filters))
3493 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3497 def download_range_func(chapters, ranges):
3498 def inner(info_dict, ydl):
3499 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3500 else 'Cannot match chapters since chapter information is unavailable')
3501 for regex in chapters or []:
3502 for i, chapter in enumerate(info_dict.get('chapters') or []):
3503 if re.search(regex, chapter['title']):
3505 yield {**chapter, 'index': i}
3506 if chapters and warning:
3507 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3509 yield from ({'start_time': start, 'end_time': end} for start, end in ranges or [])
3514 def parse_dfxp_time_expr(time_expr):
3518 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3520 return float(mobj.group('time_offset'))
3522 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3524 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3527 def srt_subtitles_timecode(seconds):
3528 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3531 def ass_subtitles_timecode(seconds):
3532 time = timetuple_from_msec(seconds * 1000)
3533 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3536 def dfxp2srt(dfxp_data):
3538 @param dfxp_data A
bytes-like
object containing DFXP data
3539 @returns A
unicode object containing converted SRT data
3541 LEGACY_NAMESPACES = (
3542 (b'http://www.w3.org/ns/ttml', [
3543 b'http://www.w3.org/2004/11/ttaf1',
3544 b'http://www.w3.org/2006/04/ttaf1',
3545 b'http://www.w3.org/2006/10/ttaf1',
3547 (b'http://www.w3.org/ns/ttml#styling', [
3548 b'http://www.w3.org/ns/ttml#style',
3552 SUPPORTED_STYLING = [
3561 _x = functools.partial(xpath_with_ns, ns_map={
3562 'xml': 'http://www.w3.org/XML/1998/namespace',
3563 'ttml': 'http://www.w3.org/ns/ttml',
3564 'tts': 'http://www.w3.org/ns/ttml#styling',
3570 class TTMLPElementParser:
3572 _unclosed_elements = []
3573 _applied_styles = []
3575 def start(self, tag, attrib):
3576 if tag in (_x('ttml:br'), 'br'):
3579 unclosed_elements = []
3581 element_style_id = attrib.get('style')
3583 style.update(default_style)
3584 if element_style_id:
3585 style.update(styles.get(element_style_id, {}))
3586 for prop in SUPPORTED_STYLING:
3587 prop_val = attrib.get(_x('tts:' + prop))
3589 style[prop] = prop_val
3592 for k, v in sorted(style.items()):
3593 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3596 font += ' color="%s"' % v
3597 elif k == 'fontSize':
3598 font += ' size="%s"' % v
3599 elif k == 'fontFamily':
3600 font += ' face="%s"' % v
3601 elif k == 'fontWeight' and v == 'bold':
3603 unclosed_elements.append('b')
3604 elif k == 'fontStyle' and v == 'italic':
3606 unclosed_elements.append('i')
3607 elif k == 'textDecoration' and v == 'underline':
3609 unclosed_elements.append('u')
3611 self._out += '<font' + font + '>'
3612 unclosed_elements.append('font')
3614 if self._applied_styles:
3615 applied_style.update(self._applied_styles[-1])
3616 applied_style.update(style)
3617 self._applied_styles.append(applied_style)
3618 self._unclosed_elements.append(unclosed_elements)
3621 if tag not in (_x('ttml:br'), 'br'):
3622 unclosed_elements = self._unclosed_elements.pop()
3623 for element in reversed(unclosed_elements):
3624 self._out += '</%s>' % element
3625 if unclosed_elements and self._applied_styles:
3626 self._applied_styles.pop()
3628 def data(self, data):
3632 return self._out.strip()
3634 def parse_node(node):
3635 target = TTMLPElementParser()
3636 parser = xml.etree.ElementTree.XMLParser(target=target)
3637 parser.feed(xml.etree.ElementTree.tostring(node))
3638 return parser.close()
3640 for k, v in LEGACY_NAMESPACES:
3642 dfxp_data = dfxp_data.replace(ns, k)
3644 dfxp = compat_etree_fromstring(dfxp_data)
3646 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3649 raise ValueError('Invalid dfxp/TTML subtitle')
3653 for style in dfxp.findall(_x('.//ttml:style')):
3654 style_id = style.get('id') or style.get(_x('xml:id'))
3657 parent_style_id = style.get('style')
3659 if parent_style_id not in styles:
3662 styles[style_id] = styles[parent_style_id].copy()
3663 for prop in SUPPORTED_STYLING:
3664 prop_val = style.get(_x('tts:' + prop))
3666 styles.setdefault(style_id, {})[prop] = prop_val
3672 for p in ('body', 'div'):
3673 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3676 style = styles.get(ele.get('style'))
3679 default_style.update(style)
3681 for para, index in zip(paras, itertools.count(1)):
3682 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3683 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3684 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3685 if begin_time is None:
3690 end_time = begin_time + dur
3691 out.append('%d\n%s --> %s\n%s\n\n' % (
3693 srt_subtitles_timecode(begin_time),
3694 srt_subtitles_timecode(end_time),
3700 def cli_option(params, command_option, param, separator=None):
3701 param = params.get(param)
3702 return ([] if param is None
3703 else [command_option, str(param)] if separator is None
3704 else [f'{command_option}{separator}{param}'])
3707 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3708 param = params.get(param)
3709 assert param in (True, False, None)
3710 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
3713 def cli_valueless_option(params, command_option, param, expected_value=True):
3714 return [command_option] if params.get(param) == expected_value else []
3717 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3718 if isinstance(argdict, (list, tuple)): # for backward compatibility
3725 assert isinstance(argdict, dict)
3727 assert isinstance(keys, (list, tuple))
3728 for key_list in keys:
3729 arg_list = list(filter(
3730 lambda x: x is not None,
3731 [argdict.get(key.lower()) for key in variadic(key_list)]))
3733 return [arg for args in arg_list for arg in args]
3737 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3738 main_key, exe = main_key.lower(), exe.lower()
3739 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3740 keys = [f'{root_key}{k}' for k in (keys or [''])]
3741 if root_key in keys:
3743 keys.append((main_key, exe))
3744 keys.append('default')
3747 return cli_configuration_args(argdict, keys, default, use_compat)
3751 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
3810 'iw': 'heb', # Replaced by he in 1989 revision
3820 'in': 'ind', # Replaced by id in 1989 revision
3935 'ji': 'yid', # Replaced by yi in 1989 revision
3943 def short2long(cls, code):
3944 """Convert language code from ISO 639-1 to ISO 639-2/T"""
3945 return cls._lang_map.get(code[:2])
3948 def long2short(cls, code):
3949 """Convert language code from ISO 639-2/T to ISO 639-1"""
3950 for short_name, long_name in cls._lang_map.items():
3951 if long_name == code:
3956 # From http://data.okfn.org/data/core/country-list
3958 'AF': 'Afghanistan',
3959 'AX': 'Åland Islands',
3962 'AS': 'American Samoa',
3967 'AG': 'Antigua and Barbuda',
3984 'BO': 'Bolivia, Plurinational State of',
3985 'BQ': 'Bonaire, Sint Eustatius and Saba',
3986 'BA': 'Bosnia and Herzegovina',
3988 'BV': 'Bouvet Island',
3990 'IO': 'British Indian Ocean Territory',
3991 'BN': 'Brunei Darussalam',
3993 'BF': 'Burkina Faso',
3999 'KY': 'Cayman Islands',
4000 'CF': 'Central African Republic',
4004 'CX': 'Christmas Island',
4005 'CC': 'Cocos (Keeling) Islands',
4009 'CD': 'Congo, the Democratic Republic of the',
4010 'CK': 'Cook Islands',
4012 'CI': 'Côte d\'Ivoire',
4017 'CZ': 'Czech Republic',
4021 'DO': 'Dominican Republic',
4024 'SV': 'El Salvador',
4025 'GQ': 'Equatorial Guinea',
4029 'FK': 'Falkland Islands (Malvinas)',
4030 'FO': 'Faroe Islands',
4034 'GF': 'French Guiana',
4035 'PF': 'French Polynesia',
4036 'TF': 'French Southern Territories',
4051 'GW': 'Guinea-Bissau',
4054 'HM': 'Heard Island and McDonald Islands',
4055 'VA': 'Holy See (Vatican City State)',
4062 'IR': 'Iran, Islamic Republic of',
4065 'IM': 'Isle of Man',
4075 'KP': 'Korea, Democratic People\'s Republic of',
4076 'KR': 'Korea, Republic of',
4079 'LA': 'Lao People\'s Democratic Republic',
4085 'LI': 'Liechtenstein',
4089 'MK': 'Macedonia, the Former Yugoslav Republic of',
4096 'MH': 'Marshall Islands',
4102 'FM': 'Micronesia, Federated States of',
4103 'MD': 'Moldova, Republic of',
4114 'NL': 'Netherlands',
4115 'NC': 'New Caledonia',
4116 'NZ': 'New Zealand',
4121 'NF': 'Norfolk Island',
4122 'MP': 'Northern Mariana Islands',
4127 'PS': 'Palestine, State of',
4129 'PG': 'Papua New Guinea',
4132 'PH': 'Philippines',
4136 'PR': 'Puerto Rico',
4140 'RU': 'Russian Federation',
4142 'BL': 'Saint Barthélemy',
4143 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4144 'KN': 'Saint Kitts and Nevis',
4145 'LC': 'Saint Lucia',
4146 'MF': 'Saint Martin (French part)',
4147 'PM': 'Saint Pierre and Miquelon',
4148 'VC': 'Saint Vincent and the Grenadines',
4151 'ST': 'Sao Tome and Principe',
4152 'SA': 'Saudi Arabia',
4156 'SL': 'Sierra Leone',
4158 'SX': 'Sint Maarten (Dutch part)',
4161 'SB': 'Solomon Islands',
4163 'ZA': 'South Africa',
4164 'GS': 'South Georgia and the South Sandwich Islands',
4165 'SS': 'South Sudan',
4170 'SJ': 'Svalbard and Jan Mayen',
4173 'CH': 'Switzerland',
4174 'SY': 'Syrian Arab Republic',
4175 'TW': 'Taiwan, Province of China',
4177 'TZ': 'Tanzania, United Republic of',
4179 'TL': 'Timor-Leste',
4183 'TT': 'Trinidad and Tobago',
4186 'TM': 'Turkmenistan',
4187 'TC': 'Turks and Caicos Islands',
4191 'AE': 'United Arab Emirates',
4192 'GB': 'United Kingdom',
4193 'US': 'United States',
4194 'UM': 'United States Minor Outlying Islands',
4198 'VE': 'Venezuela, Bolivarian Republic of',
4200 'VG': 'Virgin Islands, British',
4201 'VI': 'Virgin Islands, U.S.',
4202 'WF': 'Wallis and Futuna',
4203 'EH': 'Western Sahara',
4207 # Not ISO 3166 codes, but used for IP blocks
4208 'AP': 'Asia/Pacific Region',
4213 def short2full(cls, code):
4214 """Convert an ISO 3166-2 country code to the corresponding full name"""
4215 return cls._country_map.get(code.upper())
4219 # Major IPv4 address blocks per country
4221 'AD': '46.172.224.0/19',
4222 'AE': '94.200.0.0/13',
4223 'AF': '149.54.0.0/17',
4224 'AG': '209.59.64.0/18',
4225 'AI': '204.14.248.0/21',
4226 'AL': '46.99.0.0/16',
4227 'AM': '46.70.0.0/15',
4228 'AO': '105.168.0.0/13',
4229 'AP': '182.50.184.0/21',
4230 'AQ': '23.154.160.0/24',
4231 'AR': '181.0.0.0/12',
4232 'AS': '202.70.112.0/20',
4233 'AT': '77.116.0.0/14',
4234 'AU': '1.128.0.0/11',
4235 'AW': '181.41.0.0/18',
4236 'AX': '185.217.4.0/22',
4237 'AZ': '5.197.0.0/16',
4238 'BA': '31.176.128.0/17',
4239 'BB': '65.48.128.0/17',
4240 'BD': '114.130.0.0/16',
4242 'BF': '102.178.0.0/15',
4243 'BG': '95.42.0.0/15',
4244 'BH': '37.131.0.0/17',
4245 'BI': '154.117.192.0/18',
4246 'BJ': '137.255.0.0/16',
4247 'BL': '185.212.72.0/23',
4248 'BM': '196.12.64.0/18',
4249 'BN': '156.31.0.0/16',
4250 'BO': '161.56.0.0/16',
4251 'BQ': '161.0.80.0/20',
4252 'BR': '191.128.0.0/12',
4253 'BS': '24.51.64.0/18',
4254 'BT': '119.2.96.0/19',
4255 'BW': '168.167.0.0/16',
4256 'BY': '178.120.0.0/13',
4257 'BZ': '179.42.192.0/18',
4258 'CA': '99.224.0.0/11',
4259 'CD': '41.243.0.0/16',
4260 'CF': '197.242.176.0/21',
4261 'CG': '160.113.0.0/16',
4262 'CH': '85.0.0.0/13',
4263 'CI': '102.136.0.0/14',
4264 'CK': '202.65.32.0/19',
4265 'CL': '152.172.0.0/14',
4266 'CM': '102.244.0.0/14',
4267 'CN': '36.128.0.0/10',
4268 'CO': '181.240.0.0/12',
4269 'CR': '201.192.0.0/12',
4270 'CU': '152.206.0.0/15',
4271 'CV': '165.90.96.0/19',
4272 'CW': '190.88.128.0/17',
4273 'CY': '31.153.0.0/16',
4274 'CZ': '88.100.0.0/14',
4276 'DJ': '197.241.0.0/17',
4277 'DK': '87.48.0.0/12',
4278 'DM': '192.243.48.0/20',
4279 'DO': '152.166.0.0/15',
4280 'DZ': '41.96.0.0/12',
4281 'EC': '186.68.0.0/15',
4282 'EE': '90.190.0.0/15',
4283 'EG': '156.160.0.0/11',
4284 'ER': '196.200.96.0/20',
4285 'ES': '88.0.0.0/11',
4286 'ET': '196.188.0.0/14',
4287 'EU': '2.16.0.0/13',
4288 'FI': '91.152.0.0/13',
4289 'FJ': '144.120.0.0/16',
4290 'FK': '80.73.208.0/21',
4291 'FM': '119.252.112.0/20',
4292 'FO': '88.85.32.0/19',
4294 'GA': '41.158.0.0/15',
4296 'GD': '74.122.88.0/21',
4297 'GE': '31.146.0.0/16',
4298 'GF': '161.22.64.0/18',
4299 'GG': '62.68.160.0/19',
4300 'GH': '154.160.0.0/12',
4301 'GI': '95.164.0.0/16',
4302 'GL': '88.83.0.0/19',
4303 'GM': '160.182.0.0/15',
4304 'GN': '197.149.192.0/18',
4305 'GP': '104.250.0.0/19',
4306 'GQ': '105.235.224.0/20',
4307 'GR': '94.64.0.0/13',
4308 'GT': '168.234.0.0/16',
4309 'GU': '168.123.0.0/16',
4310 'GW': '197.214.80.0/20',
4311 'GY': '181.41.64.0/18',
4312 'HK': '113.252.0.0/14',
4313 'HN': '181.210.0.0/16',
4314 'HR': '93.136.0.0/13',
4315 'HT': '148.102.128.0/17',
4316 'HU': '84.0.0.0/14',
4317 'ID': '39.192.0.0/10',
4318 'IE': '87.32.0.0/12',
4319 'IL': '79.176.0.0/13',
4320 'IM': '5.62.80.0/20',
4321 'IN': '117.192.0.0/10',
4322 'IO': '203.83.48.0/21',
4323 'IQ': '37.236.0.0/14',
4324 'IR': '2.176.0.0/12',
4325 'IS': '82.221.0.0/16',
4326 'IT': '79.0.0.0/10',
4327 'JE': '87.244.64.0/18',
4328 'JM': '72.27.0.0/17',
4329 'JO': '176.29.0.0/16',
4330 'JP': '133.0.0.0/8',
4331 'KE': '105.48.0.0/12',
4332 'KG': '158.181.128.0/17',
4333 'KH': '36.37.128.0/17',
4334 'KI': '103.25.140.0/22',
4335 'KM': '197.255.224.0/20',
4336 'KN': '198.167.192.0/19',
4337 'KP': '175.45.176.0/22',
4338 'KR': '175.192.0.0/10',
4339 'KW': '37.36.0.0/14',
4340 'KY': '64.96.0.0/15',
4341 'KZ': '2.72.0.0/13',
4342 'LA': '115.84.64.0/18',
4343 'LB': '178.135.0.0/16',
4344 'LC': '24.92.144.0/20',
4345 'LI': '82.117.0.0/19',
4346 'LK': '112.134.0.0/15',
4347 'LR': '102.183.0.0/16',
4348 'LS': '129.232.0.0/17',
4349 'LT': '78.56.0.0/13',
4350 'LU': '188.42.0.0/16',
4351 'LV': '46.109.0.0/16',
4352 'LY': '41.252.0.0/14',
4353 'MA': '105.128.0.0/11',
4354 'MC': '88.209.64.0/18',
4355 'MD': '37.246.0.0/16',
4356 'ME': '178.175.0.0/17',
4357 'MF': '74.112.232.0/21',
4358 'MG': '154.126.0.0/17',
4359 'MH': '117.103.88.0/21',
4360 'MK': '77.28.0.0/15',
4361 'ML': '154.118.128.0/18',
4362 'MM': '37.111.0.0/17',
4363 'MN': '49.0.128.0/17',
4364 'MO': '60.246.0.0/16',
4365 'MP': '202.88.64.0/20',
4366 'MQ': '109.203.224.0/19',
4367 'MR': '41.188.64.0/18',
4368 'MS': '208.90.112.0/22',
4369 'MT': '46.11.0.0/16',
4370 'MU': '105.16.0.0/12',
4371 'MV': '27.114.128.0/18',
4372 'MW': '102.70.0.0/15',
4373 'MX': '187.192.0.0/11',
4374 'MY': '175.136.0.0/13',
4375 'MZ': '197.218.0.0/15',
4376 'NA': '41.182.0.0/16',
4377 'NC': '101.101.0.0/18',
4378 'NE': '197.214.0.0/18',
4379 'NF': '203.17.240.0/22',
4380 'NG': '105.112.0.0/12',
4381 'NI': '186.76.0.0/15',
4382 'NL': '145.96.0.0/11',
4383 'NO': '84.208.0.0/13',
4384 'NP': '36.252.0.0/15',
4385 'NR': '203.98.224.0/19',
4386 'NU': '49.156.48.0/22',
4387 'NZ': '49.224.0.0/14',
4388 'OM': '5.36.0.0/15',
4389 'PA': '186.72.0.0/15',
4390 'PE': '186.160.0.0/14',
4391 'PF': '123.50.64.0/18',
4392 'PG': '124.240.192.0/19',
4393 'PH': '49.144.0.0/13',
4394 'PK': '39.32.0.0/11',
4395 'PL': '83.0.0.0/11',
4396 'PM': '70.36.0.0/20',
4397 'PR': '66.50.0.0/16',
4398 'PS': '188.161.0.0/16',
4399 'PT': '85.240.0.0/13',
4400 'PW': '202.124.224.0/20',
4401 'PY': '181.120.0.0/14',
4402 'QA': '37.210.0.0/15',
4403 'RE': '102.35.0.0/16',
4404 'RO': '79.112.0.0/13',
4405 'RS': '93.86.0.0/15',
4406 'RU': '5.136.0.0/13',
4407 'RW': '41.186.0.0/16',
4408 'SA': '188.48.0.0/13',
4409 'SB': '202.1.160.0/19',
4410 'SC': '154.192.0.0/11',
4411 'SD': '102.120.0.0/13',
4412 'SE': '78.64.0.0/12',
4413 'SG': '8.128.0.0/10',
4414 'SI': '188.196.0.0/14',
4415 'SK': '78.98.0.0/15',
4416 'SL': '102.143.0.0/17',
4417 'SM': '89.186.32.0/19',
4418 'SN': '41.82.0.0/15',
4419 'SO': '154.115.192.0/18',
4420 'SR': '186.179.128.0/17',
4421 'SS': '105.235.208.0/21',
4422 'ST': '197.159.160.0/19',
4423 'SV': '168.243.0.0/16',
4424 'SX': '190.102.0.0/20',
4426 'SZ': '41.84.224.0/19',
4427 'TC': '65.255.48.0/20',
4428 'TD': '154.68.128.0/19',
4429 'TG': '196.168.0.0/14',
4430 'TH': '171.96.0.0/13',
4431 'TJ': '85.9.128.0/18',
4432 'TK': '27.96.24.0/21',
4433 'TL': '180.189.160.0/20',
4434 'TM': '95.85.96.0/19',
4435 'TN': '197.0.0.0/11',
4436 'TO': '175.176.144.0/21',
4437 'TR': '78.160.0.0/11',
4438 'TT': '186.44.0.0/15',
4439 'TV': '202.2.96.0/19',
4440 'TW': '120.96.0.0/11',
4441 'TZ': '156.156.0.0/14',
4442 'UA': '37.52.0.0/14',
4443 'UG': '102.80.0.0/13',
4445 'UY': '167.56.0.0/13',
4446 'UZ': '84.54.64.0/18',
4447 'VA': '212.77.0.0/19',
4448 'VC': '207.191.240.0/21',
4449 'VE': '186.88.0.0/13',
4450 'VG': '66.81.192.0/20',
4451 'VI': '146.226.0.0/16',
4452 'VN': '14.160.0.0/11',
4453 'VU': '202.80.32.0/20',
4454 'WF': '117.20.32.0/21',
4455 'WS': '202.4.32.0/19',
4456 'YE': '134.35.0.0/16',
4457 'YT': '41.242.116.0/22',
4458 'ZA': '41.0.0.0/11',
4459 'ZM': '102.144.0.0/13',
4460 'ZW': '102.177.192.0/18',
4464 def random_ipv4(cls, code_or_block):
4465 if len(code_or_block) == 2:
4466 block = cls._country_ip_map.get(code_or_block.upper())
4470 block = code_or_block
4471 addr, preflen = block.split('/')
4472 addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
4473 addr_max = addr_min | (0xffffffff >> int(preflen))
4474 return compat_str(socket.inet_ntoa(
4475 compat_struct_pack('!L', random.randint(addr_min, addr_max))))
4478 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
4479 def __init__(self, proxies=None):
4480 # Set default handlers
4481 for type in ('http', 'https'):
4482 setattr(self, '%s_open' % type,
4483 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4484 meth(r, proxy, type))
4485 compat_urllib_request.ProxyHandler.__init__(self, proxies)
4487 def proxy_open(self, req, proxy, type):
4488 req_proxy = req.headers.get('Ytdl-request-proxy')
4489 if req_proxy is not None:
4491 del req.headers['Ytdl-request-proxy']
4493 if proxy == '__noproxy__':
4494 return None # No Proxy
4495 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4496 req.add_header('Ytdl-socks-proxy', proxy)
4497 # yt-dlp's http/https handlers do wrapping the socket with socks
4499 return compat_urllib_request.ProxyHandler.proxy_open(
4500 self, req, proxy, type)
4503 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4504 # released into Public Domain
4505 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4507 def long_to_bytes(n, blocksize=0):
4508 """long_to_bytes(n:long, blocksize:int) : string
4509 Convert a long integer to a byte string.
4511 If optional blocksize is given and greater than zero, pad the front of the
4512 byte string with binary zeros so that the length is a multiple of
4515 # after much testing, this algorithm was deemed to be the fastest
4519 s = compat_struct_pack('>I', n & 0xffffffff) + s
4521 # strip off leading zeros
4522 for i in range(len(s)):
4523 if s[i] != b'\000'[0]:
4526 # only happens when n == 0
4530 # add back some pad bytes. this could be done more efficiently w.r.t. the
4531 # de-padding being done above, but sigh...
4532 if blocksize > 0 and len(s) % blocksize:
4533 s = (blocksize - len(s) % blocksize) * b'\000' + s
4537 def bytes_to_long(s):
4538 """bytes_to_long(string) : long
4539 Convert a byte string to a long integer.
4541 This is (essentially) the inverse of long_to_bytes().
4546 extra = (4 - length % 4)
4547 s = b'\000' * extra + s
4548 length = length + extra
4549 for i in range(0, length, 4):
4550 acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
4554 def ohdave_rsa_encrypt(data, exponent, modulus):
4556 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4559 data: data to encrypt, bytes-like object
4560 exponent, modulus: parameter e and N of RSA algorithm, both integer
4561 Output: hex string of encrypted data
4563 Limitation: supports one block encryption only
4566 payload = int(binascii.hexlify(data[::-1]), 16)
4567 encrypted = pow(payload, exponent, modulus)
4568 return '%x' % encrypted
4571 def pkcs1pad(data, length):
4573 Padding input data with PKCS#1 scheme
4575 @param {int[]} data input data
4576 @param {int} length target length
4577 @returns {int[]} padded data
4579 if len(data) > length - 11:
4580 raise ValueError('Input data too
long for PKCS
#1 padding')
4582 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4583 return [0, 2] + pseudo_random
+ [0] + data
4586 def encode_base_n(num
, n
, table
=None):
4587 FULL_TABLE
= '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
4589 table
= FULL_TABLE
[:n
]
4592 raise ValueError('base %d exceeds table length %d' % (n
, len(table
)))
4599 ret
= table
[num
% n
] + ret
4604 def decode_packed_codes(code
):
4605 mobj
= re
.search(PACKED_CODES_RE
, code
)
4606 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4609 symbols
= symbols
.split('|')
4614 base_n_count
= encode_base_n(count
, base
)
4615 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4618 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4622 def caesar(s
, alphabet
, shift
):
4627 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4632 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4635 def parse_m3u8_attributes(attrib
):
4637 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4638 if val
.startswith('"'):
4644 def urshift(val
, n
):
4645 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4648 # Based on png2str() written by @gdkchan and improved by @yokrysty
4649 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4650 def decode_png(png_data
):
4651 # Reference: https://www.w3.org/TR/PNG/
4652 header
= png_data
[8:]
4654 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4655 raise OSError('Not a valid PNG file.')
4657 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4658 unpack_integer
= lambda x
: compat_struct_unpack(int_map
[len(x
)], x
)[0]
4663 length
= unpack_integer(header
[:4])
4666 chunk_type
= header
[:4]
4669 chunk_data
= header
[:length
]
4670 header
= header
[length
:]
4672 header
= header
[4:] # Skip CRC
4680 ihdr
= chunks
[0]['data']
4682 width
= unpack_integer(ihdr
[:4])
4683 height
= unpack_integer(ihdr
[4:8])
4687 for chunk
in chunks
:
4688 if chunk
['type'] == b
'IDAT':
4689 idat
+= chunk
['data']
4692 raise OSError('Unable to read PNG data.')
4694 decompressed_data
= bytearray(zlib
.decompress(idat
))
4699 def _get_pixel(idx
):
4704 for y
in range(height
):
4705 basePos
= y
* (1 + stride
)
4706 filter_type
= decompressed_data
[basePos
]
4710 pixels
.append(current_row
)
4712 for x
in range(stride
):
4713 color
= decompressed_data
[1 + basePos
+ x
]
4714 basex
= y
* stride
+ x
4719 left
= _get_pixel(basex
- 3)
4721 up
= _get_pixel(basex
- stride
)
4723 if filter_type
== 1: # Sub
4724 color
= (color
+ left
) & 0xff
4725 elif filter_type
== 2: # Up
4726 color
= (color
+ up
) & 0xff
4727 elif filter_type
== 3: # Average
4728 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4729 elif filter_type
== 4: # Paeth
4735 c
= _get_pixel(basex
- stride
- 3)
4743 if pa
<= pb
and pa
<= pc
:
4744 color
= (color
+ a
) & 0xff
4746 color
= (color
+ b
) & 0xff
4748 color
= (color
+ c
) & 0xff
4750 current_row
.append(color
)
4752 return width
, height
, pixels
4755 def write_xattr(path
, key
, value
):
4756 # Windows: Write xattrs to NTFS Alternate Data Streams:
4757 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4758 if compat_os_name
== 'nt':
4759 assert ':' not in key
4760 assert os
.path
.exists(path
)
4763 with open(f
'{path}:{key}', 'wb') as f
:
4765 except OSError as e
:
4766 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4769 # UNIX Method 1. Use xattrs/pyxattrs modules
4770 from .dependencies
import xattr
4773 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
4774 # Unicode arguments are not supported in pyxattr until version 0.5.0
4775 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4776 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
4777 setxattr
= xattr
.set
4779 setxattr
= xattr
.setxattr
4783 setxattr(path
, key
, value
)
4784 except OSError as e
:
4785 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4788 # UNIX Method 2. Use setfattr/xattr executables
4789 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
4790 else 'xattr' if check_executable('xattr', ['-h']) else None)
4792 raise XAttrUnavailableError(
4793 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
4794 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
4796 value
= value
.decode()
4799 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
4800 stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4801 except OSError as e
:
4802 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4803 stderr
= p
.communicate_or_kill()[1].decode('utf-8', 'replace')
4805 raise XAttrMetadataError(p
.returncode
, stderr
)
4808 def random_birthday(year_field
, month_field
, day_field
):
4809 start_date
= datetime
.date(1950, 1, 1)
4810 end_date
= datetime
.date(1995, 12, 31)
4811 offset
= random
.randint(0, (end_date
- start_date
).days
)
4812 random_date
= start_date
+ datetime
.timedelta(offset
)
4814 year_field
: str(random_date
.year
),
4815 month_field
: str(random_date
.month
),
4816 day_field
: str(random_date
.day
),
4820 # Templates for internet shortcut files, which are plain text files.
4821 DOT_URL_LINK_TEMPLATE
= '''\
4826 DOT_WEBLOC_LINK_TEMPLATE
= '''\
4827 <?xml version="1.0" encoding="UTF-8"?>
4828 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
4829 <plist version="1.0">
4832 \t<string>%(url)s</string>
4837 DOT_DESKTOP_LINK_TEMPLATE
= '''\
4847 'url': DOT_URL_LINK_TEMPLATE
,
4848 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
4849 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
4853 def iri_to_uri(iri
):
4855 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
4857 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.
4860 iri_parts
= compat_urllib_parse_urlparse(iri
)
4862 if '[' in iri_parts
.netloc
:
4863 raise ValueError('IPv6 URIs are not, yet, supported.')
4864 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
4866 # 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.
4869 if iri_parts
.username
:
4870 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
4871 if iri_parts
.password
is not None:
4872 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
4875 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
4876 # The 'idna' encoding produces ASCII text.
4877 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
4878 net_location
+= ':' + str(iri_parts
.port
)
4880 return urllib
.parse
.urlunparse(
4884 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
4886 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
4887 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
4889 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
4890 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
4892 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
4894 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
4897 def to_high_limit_path(path
):
4898 if sys
.platform
in ['win32', 'cygwin']:
4899 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
4900 return '\\\\?\\' + os
.path
.abspath(path
)
4905 def format_field(obj
, field
=None, template
='%s', ignore
=NO_DEFAULT
, default
='', func
=None):
4906 val
= traverse_obj(obj
, *variadic(field
))
4907 if (not val
and val
!= 0) if ignore
is NO_DEFAULT
else val
in ignore
:
4909 return template
% (func(val
) if func
else val
)
4912 def clean_podcast_url(url
):
4913 return re
.sub(r
'''(?x)
4917 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
4920 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
4923 cn\.co| # https://podcorn.com/analytics-prefix/
4924 st\.fm # https://podsights.com/docs/
4929 _HEX_TABLE
= '0123456789abcdef'
4932 def random_uuidv4():
4933 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
4936 def make_dir(path
, to_screen
=None):
4938 dn
= os
.path
.dirname(path
)
4939 if dn
and not os
.path
.exists(dn
):
4942 except OSError as err
:
4943 if callable(to_screen
) is not None:
4944 to_screen('unable to create directory ' + error_to_compat_str(err
))
4948 def get_executable_path():
4949 from .update
import _get_variant_and_executable_path
4951 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
4954 def load_plugins(name
, suffix
, namespace
):
4956 with contextlib
.suppress(FileNotFoundError
):
4957 plugins_spec
= importlib
.util
.spec_from_file_location(
4958 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
4959 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
4960 sys
.modules
[plugins_spec
.name
] = plugins
4961 plugins_spec
.loader
.exec_module(plugins
)
4962 for name
in dir(plugins
):
4963 if name
in namespace
:
4965 if not name
.endswith(suffix
):
4967 klass
= getattr(plugins
, name
)
4968 classes
[name
] = namespace
[name
] = klass
4973 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
4974 casesense
=True, is_user_input
=False, traverse_string
=False):
4975 ''' Traverse nested list/dict/tuple
4976 @param path_list A list of paths which are checked one by one.
4977 Each path is a list of keys where each key is a:
4979 - string: A dictionary key
4980 - int: An index into a list
4981 - tuple: A list of keys all of which will be traversed
4982 - Ellipsis: Fetch all values in the object
4983 - Function: Takes the key and value as arguments
4984 and returns whether the key matches or not
4985 @param default Default value to return
4986 @param expected_type Only accept final value of this type (Can also be any callable)
4987 @param get_all Return all the values obtained from a path or only the first one
4988 @param casesense Whether to consider dictionary keys as case sensitive
4989 @param is_user_input Whether the keys are generated from user input. If True,
4990 strings are converted to int/slice if necessary
4991 @param traverse_string Whether to traverse inside strings. If True, any
4992 non-compatible object will also be converted into a string
4996 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
4997 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
4999 def _traverse_obj(obj
, path
, _current_depth
=0):
5001 path
= tuple(variadic(path
))
5002 for i
, key
in enumerate(path
):
5003 if None in (key
, obj
):
5005 if isinstance(key
, (list, tuple)):
5006 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
5009 obj
= (obj
.values() if isinstance(obj
, dict)
5010 else obj
if isinstance(obj
, (list, tuple, LazyList
))
5011 else str(obj
) if traverse_string
else [])
5013 depth
= max(depth
, _current_depth
)
5014 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
5016 if isinstance(obj
, (list, tuple, LazyList
)):
5017 obj
= enumerate(obj
)
5018 elif isinstance(obj
, dict):
5021 if not traverse_string
:
5025 depth
= max(depth
, _current_depth
)
5026 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if try_call(key
, args
=(k
, v
))]
5027 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
5028 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
5029 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
5032 key
= (int_or_none(key
) if ':' not in key
5033 else slice(*map(int_or_none
, key
.split(':'))))
5034 if key
== slice(None):
5035 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
5036 if not isinstance(key
, (int, slice)):
5038 if not isinstance(obj
, (list, tuple, LazyList
)):
5039 if not traverse_string
:
5048 if isinstance(expected_type
, type):
5049 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5050 elif expected_type
is not None:
5051 type_test
= expected_type
5053 type_test
= lambda val
: val
5055 for path
in path_list
:
5057 val
= _traverse_obj(obj
, path
)
5060 for _
in range(depth
- 1):
5061 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5062 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5064 return val
if get_all
else val
[0]
5066 val
= type_test(val
)
5072 def traverse_dict(dictn
, keys
, casesense
=True):
5073 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5074 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5075 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5078 def get_first(obj
, keys
, **kwargs
):
5079 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5082 def variadic(x
, allowed_types
=(str, bytes, dict)):
5083 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5086 def decode_base(value
, digits
):
5087 # This will convert given base-x string to scalar (long or int)
5088 table
= {char: index for index, char in enumerate(digits)}
5093 result
+= table
[chr]
5097 def time_seconds(**kwargs
):
5098 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5099 return t
.timestamp()
5102 # create a JSON Web Signature (jws) with HS256 algorithm
5103 # the resulting format is in JWS Compact Serialization
5104 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5105 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5106 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5112 header_data
.update(headers
)
5113 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5114 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5115 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5116 signature_b64
= base64
.b64encode(h
.digest())
5117 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5121 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5122 def jwt_decode_hs256(jwt
):
5123 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5124 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5128 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5132 def supports_terminal_sequences(stream
):
5133 if compat_os_name
== 'nt':
5134 if not WINDOWS_VT_MODE
:
5136 elif not os
.getenv('TERM'):
5139 return stream
.isatty()
5140 except BaseException
:
5144 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5145 if get_windows_version() < (10, 0, 10586):
5147 global WINDOWS_VT_MODE
5148 startupinfo
= subprocess
.STARTUPINFO()
5149 startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
5151 subprocess
.Popen('', shell
=True, startupinfo
=startupinfo
).wait()
5155 WINDOWS_VT_MODE
= True
5156 supports_terminal_sequences
.cache_clear()
5159 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5162 def remove_terminal_sequences(string
):
5163 return _terminal_sequences_re
.sub('', string
)
5166 def number_of_digits(number
):
5167 return len('%d' % number
)
5170 def join_nonempty(*values
, delim
='-', from_dict
=None):
5171 if from_dict
is not None:
5172 values
= map(from_dict
.get
, values
)
5173 return delim
.join(map(str, filter(None, values
)))
5176 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5178 Find the largest format dimensions in terms of video width and, for each thumbnail:
5179 * Modify the URL: Match the width with the provided regex and replace with the former width
5182 This function is useful with video services that scale the provided thumbnails on demand
5184 _keys
= ('width', 'height')
5185 max_dimensions
= max(
5186 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5188 if not max_dimensions
[0]:
5192 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5193 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5194 for thumbnail
in thumbnails
5198 def parse_http_range(range):
5199 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5201 return None, None, None
5202 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5204 return None, None, None
5205 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5208 def read_stdin(what
):
5209 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5210 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5218 __initialized
= False
5220 def __init__(self
, parser
, label
=None):
5221 self
.parser
, self
.label
= parser
, label
5222 self
._loaded
_paths
, self
.configs
= set(), []
5224 def init(self
, args
=None, filename
=None):
5225 assert not self
.__initialized
5228 location
= os
.path
.realpath(filename
)
5229 directory
= os
.path
.dirname(location
)
5230 if location
in self
._loaded
_paths
:
5232 self
._loaded
_paths
.add(location
)
5234 self
.own_args
, self
.__initialized
= args
, True
5235 opts
, _
= self
.parser
.parse_known_args(args
)
5236 self
.parsed_args
, self
.filename
= args
, filename
5238 for location
in opts
.config_locations
or []:
5240 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5242 location
= os
.path
.join(directory
, expand_path(location
))
5243 if os
.path
.isdir(location
):
5244 location
= os
.path
.join(location
, 'yt-dlp.conf')
5245 if not os
.path
.exists(location
):
5246 self
.parser
.error(f
'config location {location} does not exist')
5247 self
.append_config(self
.read_file(location
), location
)
5251 label
= join_nonempty(
5252 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5254 return join_nonempty(
5255 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5256 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5260 def read_file(filename
, default
=[]):
5262 optionf
= open(filename
)
5264 return default
# silently skip if file is not present
5266 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5267 contents
= optionf
.read()
5268 res
= shlex
.split(contents
, comments
=True)
5274 def hide_login_info(opts
):
5275 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5276 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5281 return m
.group('key') + '=PRIVATE'
5285 opts
= list(map(_scrub_eq
, opts
))
5286 for idx
, opt
in enumerate(opts
):
5287 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5288 opts
[idx
+ 1] = 'PRIVATE'
5291 def append_config(self
, *args
, label
=None):
5292 config
= type(self
)(self
.parser
, label
)
5293 config
._loaded
_paths
= self
._loaded
_paths
5294 if config
.init(*args
):
5295 self
.configs
.append(config
)
5299 for config
in reversed(self
.configs
):
5300 yield from config
.all_args
5301 yield from self
.parsed_args
or []
5303 def parse_known_args(self
, **kwargs
):
5304 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5306 def parse_args(self
):
5307 return self
.parser
.parse_args(self
.all_args
)
5310 class WebSocketsWrapper():
5311 """Wraps websockets module to use in non-async scopes"""
5314 def __init__(self
, url
, headers
=None, connect
=True):
5315 self
.loop
= asyncio
.new_event_loop()
5316 # XXX: "loop" is deprecated
5317 self
.conn
= websockets
.connect(
5318 url
, extra_headers
=headers
, ping_interval
=None,
5319 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5322 atexit
.register(self
.__exit
__, None, None, None)
5324 def __enter__(self
):
5326 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5329 def send(self
, *args
):
5330 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5332 def recv(self
, *args
):
5333 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5335 def __exit__(self
, type, value
, traceback
):
5337 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5340 self
._cancel
_all
_tasks
(self
.loop
)
5342 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5343 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5345 def run_with_loop(main
, loop
):
5346 if not asyncio
.iscoroutine(main
):
5347 raise ValueError(f
'a coroutine was expected, got {main!r}')
5350 return loop
.run_until_complete(main
)
5352 loop
.run_until_complete(loop
.shutdown_asyncgens())
5353 if hasattr(loop
, 'shutdown_default_executor'):
5354 loop
.run_until_complete(loop
.shutdown_default_executor())
5357 def _cancel_all_tasks(loop
):
5358 to_cancel
= asyncio
.all_tasks(loop
)
5363 for task
in to_cancel
:
5366 # XXX: "loop" is removed in python 3.10+
5367 loop
.run_until_complete(
5368 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5370 for task
in to_cancel
:
5371 if task
.cancelled():
5373 if task
.exception() is not None:
5374 loop
.call_exception_handler({
5375 'message': 'unhandled exception during asyncio.run() shutdown',
5376 'exception': task
.exception(),
5381 def merge_headers(*dicts
):
5382 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5383 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5386 class classproperty
:
5387 """classmethod(property(func)) that works in py < 3.9"""
5389 def __init__(self
, func
):
5390 functools
.update_wrapper(self
, func
)
5393 def __get__(self
, _
, cls
):
5394 return self
.func(cls
)
5397 class Namespace(types
.SimpleNamespace
):
5398 """Immutable namespace"""
5401 return iter(self
.__dict
__.values())
5405 return self
.__dict
__.items()
5409 has_certifi
= bool(certifi
)
5410 has_websockets
= bool(websockets
)