38 import xml
.etree
.ElementTree
41 from .compat
import asyncio
, functools
# isort: split
45 compat_etree_fromstring
,
48 compat_html_entities_html5
,
49 compat_HTMLParseError
,
60 compat_urllib_parse_unquote_plus
,
61 compat_urllib_parse_urlencode
,
62 compat_urllib_parse_urlparse
,
63 compat_urllib_request
,
66 from .dependencies
import brotli
, certifi
, websockets
67 from .socks
import ProxyType
, sockssocket
70 def register_socks_protocols():
71 # "Register" SOCKS protocols
72 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
73 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
74 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
75 if scheme
not in compat_urlparse
.uses_netloc
:
76 compat_urlparse
.uses_netloc
.append(scheme
)
79 # This is not clearly defined otherwise
80 compiled_regex_type
= type(re
.compile(''))
83 def random_user_agent():
84 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
125 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
128 SUPPORTED_ENCODINGS
= [
132 SUPPORTED_ENCODINGS
.append('br')
135 'User-Agent': random_user_agent(),
136 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
137 'Accept-Language': 'en-us,en;q=0.5',
138 'Sec-Fetch-Mode': 'navigate',
143 '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',
147 NO_DEFAULT
= object()
149 ENGLISH_MONTH_NAMES
= [
150 'January', 'February', 'March', 'April', 'May', 'June',
151 'July', 'August', 'September', 'October', 'November', 'December']
154 'en': ENGLISH_MONTH_NAMES
,
156 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
157 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
161 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
162 'flv', 'f4v', 'f4a', 'f4b',
163 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
164 'mkv', 'mka', 'mk3d',
173 'f4f', 'f4m', 'm3u8', 'smil')
175 # needed for sanitizing filenames in restricted mode
176 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
177 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
178 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
208 '%Y-%m-%d %H:%M:%S.%f',
209 '%Y-%m-%d %H:%M:%S:%f',
212 '%Y-%m-%dT%H:%M:%SZ',
213 '%Y-%m-%dT%H:%M:%S.%fZ',
214 '%Y-%m-%dT%H:%M:%S.%f0Z',
216 '%Y-%m-%dT%H:%M:%S.%f',
219 '%b %d %Y at %H:%M:%S',
221 '%B %d %Y at %H:%M:%S',
225 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
226 DATE_FORMATS_DAY_FIRST
.extend([
235 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
236 DATE_FORMATS_MONTH_FIRST
.extend([
244 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
245 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>(?P
<json_ld
>.+?
)</script
>'
247 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
251 def preferredencoding():
252 """Get preferred encoding.
254 Returns the best encoding scheme for the system, based on
255 locale.getpreferredencoding() and some further tweaks.
258 pref = locale.getpreferredencoding()
266 def write_json_file(obj, fn):
267 """ Encode obj as JSON and write it to fn, atomically if possible """
269 tf = tempfile.NamedTemporaryFile(
270 prefix=f'{os.path.basename(fn)}
.', dir=os.path.dirname(fn),
271 suffix='.tmp
', delete=False, mode='w
', encoding='utf
-8')
275 json.dump(obj, tf, ensure_ascii=False)
276 if sys.platform == 'win32
':
277 # Need to remove existing file on Windows, else os.rename raises
278 # WindowsError or FileExistsError.
279 with contextlib.suppress(OSError):
281 with contextlib.suppress(OSError):
284 os.chmod(tf.name, 0o666 & ~mask)
285 os.rename(tf.name, fn)
287 with contextlib.suppress(OSError):
292 def find_xpath_attr(node, xpath, key, val=None):
293 """ Find the xpath xpath[@key=val] """
294 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
295 expr = xpath + ('[@%s]' % key if val is None else f"[@{key}='{val}
']")
296 return node.find(expr)
298 # On python2.6 the xml.etree.ElementTree.Element methods don't support
299 # the namespace parameter
302 def xpath_with_ns(path
, ns_map
):
303 components
= [c
.split(':') for c
in path
.split('/')]
307 replaced
.append(c
[0])
310 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
311 return '/'.join(replaced
)
314 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
315 def _find_xpath(xpath
):
316 return node
.find(xpath
)
318 if isinstance(xpath
, (str, compat_str
)):
319 n
= _find_xpath(xpath
)
327 if default
is not NO_DEFAULT
:
330 name
= xpath
if name
is None else name
331 raise ExtractorError('Could not find XML element %s' % name
)
337 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
338 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
339 if n
is None or n
== default
:
342 if default
is not NO_DEFAULT
:
345 name
= xpath
if name
is None else name
346 raise ExtractorError('Could not find XML element\'s text %s' % name
)
352 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
353 n
= find_xpath_attr(node
, xpath
, key
)
355 if default
is not NO_DEFAULT
:
358 name
= f
'{xpath}[@{key}]' if name
is None else name
359 raise ExtractorError('Could not find XML attribute %s' % name
)
365 def get_element_by_id(id, html
, **kwargs
):
366 """Return the content of the tag with the specified ID in the passed HTML document"""
367 return get_element_by_attribute('id', id, html
, **kwargs
)
370 def get_element_html_by_id(id, html
, **kwargs
):
371 """Return the html of the tag with the specified ID in the passed HTML document"""
372 return get_element_html_by_attribute('id', id, html
, **kwargs
)
375 def get_element_by_class(class_name
, html
):
376 """Return the content of the first tag with the specified class in the passed HTML document"""
377 retval
= get_elements_by_class(class_name
, html
)
378 return retval
[0] if retval
else None
381 def get_element_html_by_class(class_name
, html
):
382 """Return the html of the first tag with the specified class in the passed HTML document"""
383 retval
= get_elements_html_by_class(class_name
, html
)
384 return retval
[0] if retval
else None
387 def get_element_by_attribute(attribute
, value
, html
, **kwargs
):
388 retval
= get_elements_by_attribute(attribute
, value
, html
, **kwargs
)
389 return retval
[0] if retval
else None
392 def get_element_html_by_attribute(attribute
, value
, html
, **kargs
):
393 retval
= get_elements_html_by_attribute(attribute
, value
, html
, **kargs
)
394 return retval
[0] if retval
else None
397 def get_elements_by_class(class_name
, html
, **kargs
):
398 """Return the content of all tags with the specified class in the passed HTML document as a list"""
399 return get_elements_by_attribute(
400 'class', r
'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
401 html, escape_value=False)
404 def get_elements_html_by_class(class_name, html):
405 """Return the html of all tags with the specified class in the passed HTML document as a list"""
406 return get_elements_html_by_attribute(
407 'class', r'[^
\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
408 html, escape_value=False)
411 def get_elements_by_attribute(*args, **kwargs):
412 """Return the content of the tag with the specified attribute in the passed HTML document"""
413 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
416 def get_elements_html_by_attribute(*args, **kwargs):
417 """Return the html of the tag with the specified attribute in the passed HTML document"""
418 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
421 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
423 Return the text (content) and the html (whole) of the tag with the specified
424 attribute in the passed HTML document
427 quote = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
429 value = re.escape(value) if escape_value else value
431 partial_element_re = rf'''(?x
)
432 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
433 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
434 \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
437 for m in re.finditer(partial_element_re, html):
438 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
441 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
446 class HTMLBreakOnClosingTagParser(compat_HTMLParser):
448 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
449 closing tag for the first opening tag it has encountered, and can be used
453 class HTMLBreakOnClosingTagException(Exception):
457 self.tagstack = collections.deque()
458 compat_HTMLParser.__init__(self)
463 def __exit__(self, *_):
467 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
468 # so data remains buffered; we no longer have any interest in it, thus
469 # override this method to discard it
472 def handle_starttag(self, tag, _):
473 self.tagstack.append(tag)
475 def handle_endtag(self, tag):
476 if not self.tagstack:
477 raise compat_HTMLParseError('no tags
in the stack
')
479 inner_tag = self.tagstack.pop()
483 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
484 if not self.tagstack:
485 raise self.HTMLBreakOnClosingTagException()
488 def get_element_text_and_html_by_tag(tag, html):
490 For the first element with the specified tag in the passed HTML document
491 return its' content (text
) and the whole
element (html
)
493 def find_or_raise(haystack, needle, exc):
495 return haystack.index(needle)
498 closing_tag = f'</{tag}>'
499 whole_start = find_or_raise(
500 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
501 content_start = find_or_raise(
502 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
503 content_start += whole_start + 1
504 with HTMLBreakOnClosingTagParser() as parser:
505 parser.feed(html[whole_start:content_start])
506 if not parser.tagstack or parser.tagstack[0] != tag:
507 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
508 offset = content_start
509 while offset < len(html):
510 next_closing_tag_start = find_or_raise(
511 html[offset:], closing_tag,
512 compat_HTMLParseError(f'closing {tag} tag not found'))
513 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
515 parser.feed(html[offset:offset + next_closing_tag_end])
516 offset += next_closing_tag_end
517 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
518 return html[content_start:offset + next_closing_tag_start], \
519 html[whole_start:offset + next_closing_tag_end]
520 raise compat_HTMLParseError('unexpected end of html')
523 class HTMLAttributeParser(compat_HTMLParser):
524 """Trivial HTML parser to gather the attributes
for a single element
"""
528 compat_HTMLParser.__init__(self)
530 def handle_starttag(self, tag, attrs):
531 self.attrs = dict(attrs)
534 class HTMLListAttrsParser(compat_HTMLParser):
535 """HTML parser to gather the attributes
for the elements of a
list"""
538 compat_HTMLParser.__init__(self)
542 def handle_starttag(self, tag, attrs):
543 if tag == 'li' and self._level == 0:
544 self.items.append(dict(attrs))
547 def handle_endtag(self, tag):
551 def extract_attributes(html_element):
552 """Given a string
for an HTML element such
as
554 a
="foo" B
="bar" c
="&98;az" d
=boz
555 empty
= noval entity
="&"
558 Decode
and return a dictionary of attributes
.
560 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
561 'empty': '', 'noval': None, 'entity': '&',
562 'sq': '"', 'dq': '\''
565 parser = HTMLAttributeParser()
566 with contextlib.suppress(compat_HTMLParseError):
567 parser.feed(html_element)
572 def parse_list(webpage):
573 """Given a string
for an series of HTML
<li
> elements
,
574 return a dictionary of their attributes
"""
575 parser = HTMLListAttrsParser()
581 def clean_html(html):
582 """Clean an HTML snippet into a readable string
"""
584 if html is None: # Convenience for sanitizing descriptions etc.
587 html = re.sub(r'\s+', ' ', html)
588 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
589 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
591 html = re.sub('<.*?>', '', html)
592 # Replace html entities
593 html = unescapeHTML(html)
597 class LenientJSONDecoder(json.JSONDecoder):
598 def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
599 self.transform_source, self.ignore_extra = transform_source, ignore_extra
600 super().__init__(*args, **kwargs)
603 if self.transform_source:
604 s = self.transform_source(s)
605 if self.ignore_extra:
606 return self.raw_decode(s.lstrip())[0]
607 return super().decode(s)
610 def sanitize_open(filename, open_mode):
611 """Try to
open the given filename
, and slightly tweak it
if this fails
.
613 Attempts to
open the given filename
. If this fails
, it tries to change
614 the filename slightly
, step by step
, until it
's either able to open it
615 or it fails and raises a final exception, like the standard open()
618 It returns the tuple (stream, definitive_file_name).
621 if sys.platform == 'win32
':
623 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
624 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
626 for attempt in range(2):
629 if sys.platform == 'win32
':
630 # FIXME: An exclusive lock also locks the file from being read.
631 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
632 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
633 raise LockingUnsupportedError()
634 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
636 stream
= open(filename
, open_mode
)
637 return stream
, filename
638 except OSError as err
:
639 if attempt
or err
.errno
in (errno
.EACCES
,):
641 old_filename
, filename
= filename
, sanitize_path(filename
)
642 if old_filename
== filename
:
646 def timeconvert(timestr
):
647 """Convert RFC 2822 defined time string into system timestamp"""
649 timetuple
= email
.utils
.parsedate_tz(timestr
)
650 if timetuple
is not None:
651 timestamp
= email
.utils
.mktime_tz(timetuple
)
655 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
656 """Sanitizes a string so it could be used as part of a filename.
657 @param restricted Use a stricter subset of allowed characters
658 @param is_id Whether this is an ID that should be kept unchanged if possible.
659 If unset, yt-dlp's new sanitization rules are in effect
664 def replace_insane(char
):
665 if restricted
and char
in ACCENT_CHARS
:
666 return ACCENT_CHARS
[char
]
667 elif not restricted
and char
== '\n':
669 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
672 return '' if restricted
else '\''
674 return '\0_\0-' if restricted
else '\0 \0-'
675 elif char
in '\\/|*<>':
677 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
681 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
682 result
= ''.join(map(replace_insane
, s
))
683 if is_id
is NO_DEFAULT
:
684 result
= re
.sub('(\0.)(?:(?=\\1)..)+', r
'\1', result
) # Remove repeated substitute chars
685 STRIP_RE
= '(?:\0.|[ _-])*'
686 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
687 result
= result
.replace('\0', '') or '_'
690 while '__' in result
:
691 result
= result
.replace('__', '_')
692 result
= result
.strip('_')
693 # Common case of "Foreign band name - English song title"
694 if restricted
and result
.startswith('-_'):
696 if result
.startswith('-'):
697 result
= '_' + result
[len('-'):]
698 result
= result
.lstrip('.')
704 def sanitize_path(s
, force
=False):
705 """Sanitizes and normalizes path on Windows"""
706 if sys
.platform
== 'win32':
708 drive_or_unc
, _
= os
.path
.splitdrive(s
)
714 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
718 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
719 for path_part
in norm_path
]
721 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
722 elif force
and s
and s
[0] == os
.path
.sep
:
723 sanitized_path
.insert(0, os
.path
.sep
)
724 return os
.path
.join(*sanitized_path
)
727 def sanitize_url(url
):
728 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
729 # the number of unwanted failures due to missing protocol
732 elif url
.startswith('//'):
733 return 'http:%s' % url
734 # Fix some common typos seen so far
736 # https://github.com/ytdl-org/youtube-dl/issues/15649
737 (r
'^httpss://', r
'https://'),
738 # https://bx1.be/lives/direct-tv/
739 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
741 for mistake
, fixup
in COMMON_TYPOS
:
742 if re
.match(mistake
, url
):
743 return re
.sub(mistake
, fixup
, url
)
747 def extract_basic_auth(url
):
748 parts
= compat_urlparse
.urlsplit(url
)
749 if parts
.username
is None:
751 url
= compat_urlparse
.urlunsplit(parts
._replace
(netloc
=(
752 parts
.hostname
if parts
.port
is None
753 else '%s:%d' % (parts
.hostname
, parts
.port
))))
754 auth_payload
= base64
.b64encode(
755 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode())
756 return url
, f
'Basic {auth_payload.decode()}'
759 def sanitized_Request(url
, *args
, **kwargs
):
760 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
761 if auth_header
is not None:
762 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
763 headers
['Authorization'] = auth_header
764 return compat_urllib_request
.Request(url
, *args
, **kwargs
)
768 """Expand shell variables and ~"""
769 return os
.path
.expandvars(compat_expanduser(s
))
772 def orderedSet(iterable
):
773 """ Remove all duplicates from the input iterable """
781 def _htmlentity_transform(entity_with_semicolon
):
782 """Transforms an HTML entity to a character."""
783 entity
= entity_with_semicolon
[:-1]
785 # Known non-numeric HTML entity
786 if entity
in compat_html_entities
.name2codepoint
:
787 return compat_chr(compat_html_entities
.name2codepoint
[entity
])
789 # TODO: HTML5 allows entities without a semicolon. For example,
790 # 'Éric' should be decoded as 'Éric'.
791 if entity_with_semicolon
in compat_html_entities_html5
:
792 return compat_html_entities_html5
[entity_with_semicolon
]
794 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
796 numstr
= mobj
.group(1)
797 if numstr
.startswith('x'):
799 numstr
= '0%s' % numstr
802 # See https://github.com/ytdl-org/youtube-dl/issues/7518
803 with contextlib
.suppress(ValueError):
804 return compat_chr(int(numstr
, base
))
806 # Unknown entity in name, return its literal representation
807 return '&%s;' % entity
813 assert isinstance(s
, str)
816 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
819 def escapeHTML(text
):
822 .replace('&', '&')
823 .replace('<', '<')
824 .replace('>', '>')
825 .replace('"', '"')
826 .replace("'", ''')
830 def process_communicate_or_kill(p
, *args
, **kwargs
):
831 write_string('DeprecationWarning: yt_dlp.utils.process_communicate_or_kill is deprecated '
832 'and may be removed in a future version. Use yt_dlp.utils.Popen.communicate_or_kill instead')
833 return Popen
.communicate_or_kill(p
, *args
, **kwargs
)
836 class Popen(subprocess
.Popen
):
837 if sys
.platform
== 'win32':
838 _startupinfo
= subprocess
.STARTUPINFO()
839 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
843 def __init__(self
, *args
, **kwargs
):
844 super().__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
846 def communicate_or_kill(self
, *args
, **kwargs
):
848 return self
.communicate(*args
, **kwargs
)
849 except BaseException
: # Including KeyboardInterrupt
855 def get_subprocess_encoding():
856 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
857 # For subprocess calls, encode with locale encoding
858 # Refer to http://stackoverflow.com/a/9951851/35070
859 encoding
= preferredencoding()
861 encoding
= sys
.getfilesystemencoding()
867 def encodeFilename(s
, for_subprocess
=False):
868 assert isinstance(s
, str)
872 def decodeFilename(b
, for_subprocess
=False):
876 def encodeArgument(s
):
877 # Legacy code that uses byte strings
878 # Uncomment the following line after fixing all post processors
879 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
880 return s
if isinstance(s
, str) else s
.decode('ascii')
883 def decodeArgument(b
):
887 def decodeOption(optval
):
890 if isinstance(optval
, bytes):
891 optval
= optval
.decode(preferredencoding())
893 assert isinstance(optval
, compat_str
)
897 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
900 def timetuple_from_msec(msec
):
901 secs
, msec
= divmod(msec
, 1000)
902 mins
, secs
= divmod(secs
, 60)
903 hrs
, mins
= divmod(mins
, 60)
904 return _timetuple(hrs
, mins
, secs
, msec
)
907 def formatSeconds(secs
, delim
=':', msec
=False):
908 time
= timetuple_from_msec(secs
* 1000)
910 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
912 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
914 ret
= '%d' % time
.seconds
915 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
918 def _ssl_load_windows_store_certs(ssl_context
, storename
):
919 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
921 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
922 if encoding
== 'x509_asn' and (
923 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
924 except PermissionError
:
927 with contextlib
.suppress(ssl
.SSLError
):
928 ssl_context
.load_verify_locations(cadata
=cert
)
931 def make_HTTPS_handler(params
, **kwargs
):
932 opts_check_certificate
= not params
.get('nocheckcertificate')
933 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
934 context
.check_hostname
= opts_check_certificate
935 if params
.get('legacyserverconnect'):
936 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
937 # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
938 context
.set_ciphers('DEFAULT')
940 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
941 if opts_check_certificate
:
942 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
943 context
.load_verify_locations(cafile
=certifi
.where())
945 context
.load_default_certs()
946 # Work around the issue in load_default_certs when there are bad certificates. See:
947 # https://github.com/yt-dlp/yt-dlp/issues/1060,
948 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
950 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
951 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
952 for storename
in ('CA', 'ROOT'):
953 _ssl_load_windows_store_certs(context
, storename
)
954 context
.set_default_verify_paths()
956 client_certfile
= params
.get('client_certificate')
959 context
.load_cert_chain(
960 client_certfile
, keyfile
=params
.get('client_certificate_key'),
961 password
=params
.get('client_certificate_password'))
963 raise YoutubeDLError('Unable to load client certificate')
965 # Some servers may reject requests if ALPN extension is not sent. See:
966 # https://github.com/python/cpython/issues/85140
967 # https://github.com/yt-dlp/yt-dlp/issues/3878
968 with contextlib
.suppress(NotImplementedError):
969 context
.set_alpn_protocols(['http/1.1'])
971 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
974 def bug_reports_message(before
=';'):
975 msg
= ('please report this issue on https://github.com/yt-dlp/yt-dlp/issues?q= , '
976 'filling out the appropriate issue template. '
977 'Confirm you are on the latest version using yt-dlp -U')
979 before
= before
.rstrip()
980 if not before
or before
.endswith(('.', '!', '?')):
981 msg
= msg
[0].title() + msg
[1:]
983 return (before
+ ' ' if before
else '') + msg
986 class YoutubeDLError(Exception):
987 """Base exception for YoutubeDL errors."""
990 def __init__(self
, msg
=None):
993 elif self
.msg
is None:
994 self
.msg
= type(self
).__name
__
995 super().__init
__(self
.msg
)
998 network_exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
999 if hasattr(ssl
, 'CertificateError'):
1000 network_exceptions
.append(ssl
.CertificateError
)
1001 network_exceptions
= tuple(network_exceptions
)
1004 class ExtractorError(YoutubeDLError
):
1005 """Error during info extraction."""
1007 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1008 """ tb, if given, is the original traceback (so that it can be printed out).
1009 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1011 if sys
.exc_info()[0] in network_exceptions
:
1014 self
.orig_msg
= str(msg
)
1016 self
.expected
= expected
1018 self
.video_id
= video_id
1020 self
.exc_info
= sys
.exc_info() # preserve original exception
1022 super().__init
__(''.join((
1023 format_field(ie
, template
='[%s] '),
1024 format_field(video_id
, template
='%s: '),
1026 format_field(cause
, template
=' (caused by %r)'),
1027 '' if expected
else bug_reports_message())))
1029 def format_traceback(self
):
1030 return join_nonempty(
1031 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1032 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1036 class UnsupportedError(ExtractorError
):
1037 def __init__(self
, url
):
1039 'Unsupported URL: %s' % url
, expected
=True)
1043 class RegexNotFoundError(ExtractorError
):
1044 """Error when a regex didn't match"""
1048 class GeoRestrictedError(ExtractorError
):
1049 """Geographic restriction Error exception.
1051 This exception may be thrown when a video is not available from your
1052 geographic location due to geographic restrictions imposed by a website.
1055 def __init__(self
, msg
, countries
=None, **kwargs
):
1056 kwargs
['expected'] = True
1057 super().__init
__(msg
, **kwargs
)
1058 self
.countries
= countries
1061 class DownloadError(YoutubeDLError
):
1062 """Download Error exception.
1064 This exception may be thrown by FileDownloader objects if they are not
1065 configured to continue on errors. They will contain the appropriate
1069 def __init__(self
, msg
, exc_info
=None):
1070 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1071 super().__init
__(msg
)
1072 self
.exc_info
= exc_info
1075 class EntryNotInPlaylist(YoutubeDLError
):
1076 """Entry not in playlist exception.
1078 This exception will be thrown by YoutubeDL when a requested entry
1079 is not found in the playlist info_dict
1081 msg
= 'Entry not found in info'
1084 class SameFileError(YoutubeDLError
):
1085 """Same File exception.
1087 This exception will be thrown by FileDownloader objects if they detect
1088 multiple files would have to be downloaded to the same file on disk.
1090 msg
= 'Fixed output name but more than one file to download'
1092 def __init__(self
, filename
=None):
1093 if filename
is not None:
1094 self
.msg
+= f
': {filename}'
1095 super().__init
__(self
.msg
)
1098 class PostProcessingError(YoutubeDLError
):
1099 """Post Processing exception.
1101 This exception may be raised by PostProcessor's .run() method to
1102 indicate an error in the postprocessing task.
1106 class DownloadCancelled(YoutubeDLError
):
1107 """ Exception raised when the download queue should be interrupted """
1108 msg
= 'The download was cancelled'
1111 class ExistingVideoReached(DownloadCancelled
):
1112 """ --break-on-existing triggered """
1113 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1116 class RejectedVideoReached(DownloadCancelled
):
1117 """ --break-on-reject triggered """
1118 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1121 class MaxDownloadsReached(DownloadCancelled
):
1122 """ --max-downloads limit has been reached. """
1123 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1126 class ReExtractInfo(YoutubeDLError
):
1127 """ Video info needs to be re-extracted. """
1129 def __init__(self
, msg
, expected
=False):
1130 super().__init
__(msg
)
1131 self
.expected
= expected
1134 class ThrottledDownload(ReExtractInfo
):
1135 """ Download speed below --throttled-rate. """
1136 msg
= 'The download speed is below throttle limit'
1139 super().__init
__(self
.msg
, expected
=False)
1142 class UnavailableVideoError(YoutubeDLError
):
1143 """Unavailable Format exception.
1145 This exception will be thrown when a video is requested
1146 in a format that is not available for that video.
1148 msg
= 'Unable to download video'
1150 def __init__(self
, err
=None):
1152 self
.msg
+= f
': {err}'
1153 super().__init
__(self
.msg
)
1156 class ContentTooShortError(YoutubeDLError
):
1157 """Content Too Short exception.
1159 This exception may be raised by FileDownloader objects when a file they
1160 download is too small for what the server announced first, indicating
1161 the connection was probably interrupted.
1164 def __init__(self
, downloaded
, expected
):
1165 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1167 self
.downloaded
= downloaded
1168 self
.expected
= expected
1171 class XAttrMetadataError(YoutubeDLError
):
1172 def __init__(self
, code
=None, msg
='Unknown error'):
1173 super().__init
__(msg
)
1177 # Parsing code and msg
1178 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1179 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1180 self
.reason
= 'NO_SPACE'
1181 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1182 self
.reason
= 'VALUE_TOO_LONG'
1184 self
.reason
= 'NOT_SUPPORTED'
1187 class XAttrUnavailableError(YoutubeDLError
):
1191 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1192 hc
= http_class(*args
, **kwargs
)
1193 source_address
= ydl_handler
._params
.get('source_address')
1195 if source_address
is not None:
1196 # This is to workaround _create_connection() from socket where it will try all
1197 # address data from getaddrinfo() including IPv6. This filters the result from
1198 # getaddrinfo() based on the source_address value.
1199 # This is based on the cpython socket.create_connection() function.
1200 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1201 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1202 host
, port
= address
1204 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1205 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1206 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1207 if addrs
and not ip_addrs
:
1208 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1210 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1211 % (ip_version
, source_address
[0]))
1212 for res
in ip_addrs
:
1213 af
, socktype
, proto
, canonname
, sa
= res
1216 sock
= socket
.socket(af
, socktype
, proto
)
1217 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1218 sock
.settimeout(timeout
)
1219 sock
.bind(source_address
)
1221 err
= None # Explicitly break reference cycle
1223 except OSError as _
:
1225 if sock
is not None:
1230 raise OSError('getaddrinfo returns an empty list')
1231 if hasattr(hc
, '_create_connection'):
1232 hc
._create
_connection
= _create_connection
1233 hc
.source_address
= (source_address
, 0)
1238 def handle_youtubedl_headers(headers
):
1239 filtered_headers
= headers
1241 if 'Youtubedl-no-compression' in filtered_headers
:
1242 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1243 del filtered_headers
['Youtubedl-no-compression']
1245 return filtered_headers
1248 class YoutubeDLHandler(compat_urllib_request
.HTTPHandler
):
1249 """Handler for HTTP requests and responses.
1251 This class, when installed with an OpenerDirector, automatically adds
1252 the standard headers to every HTTP request and handles gzipped and
1253 deflated responses from web servers. If compression is to be avoided in
1254 a particular request, the original request in the program code only has
1255 to include the HTTP header "Youtubedl-no-compression", which will be
1256 removed before making the real request.
1258 Part of this code was copied from:
1260 http://techknack.net/python-urllib2-handlers/
1262 Andrew Rowls, the author of that code, agreed to release it to the
1266 def __init__(self
, params
, *args
, **kwargs
):
1267 compat_urllib_request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1268 self
._params
= params
1270 def http_open(self
, req
):
1271 conn_class
= compat_http_client
.HTTPConnection
1273 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1275 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1276 del req
.headers
['Ytdl-socks-proxy']
1278 return self
.do_open(functools
.partial(
1279 _create_http_connection
, self
, conn_class
, False),
1287 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1289 return zlib
.decompress(data
)
1295 return brotli
.decompress(data
)
1297 def http_request(self
, req
):
1298 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1299 # always respected by websites, some tend to give out URLs with non percent-encoded
1300 # non-ASCII characters (see telemb.py, ard.py [#3412])
1301 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1302 # To work around aforementioned issue we will replace request's original URL with
1303 # percent-encoded one
1304 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1305 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1306 url
= req
.get_full_url()
1307 url_escaped
= escape_url(url
)
1309 # Substitute URL if any change after escaping
1310 if url
!= url_escaped
:
1311 req
= update_Request(req
, url
=url_escaped
)
1313 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1314 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1315 # The dict keys are capitalized because of this bug by urllib
1316 if h
.capitalize() not in req
.headers
:
1317 req
.add_header(h
, v
)
1319 if 'Accept-encoding' not in req
.headers
:
1320 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1322 req
.headers
= handle_youtubedl_headers(req
.headers
)
1326 def http_response(self
, req
, resp
):
1329 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1330 content
= resp
.read()
1331 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1333 uncompressed
= io
.BytesIO(gz
.read())
1334 except OSError as original_ioerror
:
1335 # There may be junk add the end of the file
1336 # See http://stackoverflow.com/q/4928560/35070 for details
1337 for i
in range(1, 1024):
1339 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1340 uncompressed
= io
.BytesIO(gz
.read())
1345 raise original_ioerror
1346 resp
= compat_urllib_request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1347 resp
.msg
= old_resp
.msg
1348 del resp
.headers
['Content-encoding']
1350 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1351 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1352 resp
= compat_urllib_request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1353 resp
.msg
= old_resp
.msg
1354 del resp
.headers
['Content-encoding']
1356 if resp
.headers
.get('Content-encoding', '') == 'br':
1357 resp
= compat_urllib_request
.addinfourl(
1358 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1359 resp
.msg
= old_resp
.msg
1360 del resp
.headers
['Content-encoding']
1361 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1362 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1363 if 300 <= resp
.code
< 400:
1364 location
= resp
.headers
.get('Location')
1366 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1367 location
= location
.encode('iso-8859-1').decode()
1368 location_escaped
= escape_url(location
)
1369 if location
!= location_escaped
:
1370 del resp
.headers
['Location']
1371 resp
.headers
['Location'] = location_escaped
1374 https_request
= http_request
1375 https_response
= http_response
1378 def make_socks_conn_class(base_class
, socks_proxy
):
1379 assert issubclass(base_class
, (
1380 compat_http_client
.HTTPConnection
, compat_http_client
.HTTPSConnection
))
1382 url_components
= compat_urlparse
.urlparse(socks_proxy
)
1383 if url_components
.scheme
.lower() == 'socks5':
1384 socks_type
= ProxyType
.SOCKS5
1385 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1386 socks_type
= ProxyType
.SOCKS4
1387 elif url_components
.scheme
.lower() == 'socks4a':
1388 socks_type
= ProxyType
.SOCKS4A
1390 def unquote_if_non_empty(s
):
1393 return compat_urllib_parse_unquote_plus(s
)
1397 url_components
.hostname
, url_components
.port
or 1080,
1399 unquote_if_non_empty(url_components
.username
),
1400 unquote_if_non_empty(url_components
.password
),
1403 class SocksConnection(base_class
):
1405 self
.sock
= sockssocket()
1406 self
.sock
.setproxy(*proxy_args
)
1407 if isinstance(self
.timeout
, (int, float)):
1408 self
.sock
.settimeout(self
.timeout
)
1409 self
.sock
.connect((self
.host
, self
.port
))
1411 if isinstance(self
, compat_http_client
.HTTPSConnection
):
1412 if hasattr(self
, '_context'): # Python > 2.6
1413 self
.sock
= self
._context
.wrap_socket(
1414 self
.sock
, server_hostname
=self
.host
)
1416 self
.sock
= ssl
.wrap_socket(self
.sock
)
1418 return SocksConnection
1421 class YoutubeDLHTTPSHandler(compat_urllib_request
.HTTPSHandler
):
1422 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1423 compat_urllib_request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1424 self
._https
_conn
_class
= https_conn_class
or compat_http_client
.HTTPSConnection
1425 self
._params
= params
1427 def https_open(self
, req
):
1429 conn_class
= self
._https
_conn
_class
1431 if hasattr(self
, '_context'): # python > 2.6
1432 kwargs
['context'] = self
._context
1433 if hasattr(self
, '_check_hostname'): # python 3.x
1434 kwargs
['check_hostname'] = self
._check
_hostname
1436 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1438 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1439 del req
.headers
['Ytdl-socks-proxy']
1442 return self
.do_open(
1443 functools
.partial(_create_http_connection
, self
, conn_class
, True), req
, **kwargs
)
1444 except urllib
.error
.URLError
as e
:
1445 if (isinstance(e
.reason
, ssl
.SSLError
)
1446 and getattr(e
.reason
, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
1447 raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
1451 class YoutubeDLCookieJar(compat_cookiejar
.MozillaCookieJar
):
1453 See [1] for cookie file format.
1455 1. https://curl.haxx.se/docs/http-cookies.html
1457 _HTTPONLY_PREFIX
= '#HttpOnly_'
1459 _HEADER
= '''# Netscape HTTP Cookie File
1460 # This file is generated by yt-dlp. Do not edit.
1463 _CookieFileEntry
= collections
.namedtuple(
1465 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1467 def __init__(self
, filename
=None, *args
, **kwargs
):
1468 super().__init
__(None, *args
, **kwargs
)
1469 if self
.is_path(filename
):
1470 filename
= os
.fspath(filename
)
1471 self
.filename
= filename
1474 def _true_or_false(cndn
):
1475 return 'TRUE' if cndn
else 'FALSE'
1479 return isinstance(file, (str, bytes, os
.PathLike
))
1481 @contextlib.contextmanager
1482 def open(self
, file, *, write
=False):
1483 if self
.is_path(file):
1484 with open(file, 'w' if write
else 'r', encoding
='utf-8') as f
:
1491 def _really_save(self
, f
, ignore_discard
=False, ignore_expires
=False):
1494 if (not ignore_discard
and cookie
.discard
1495 or not ignore_expires
and cookie
.is_expired(now
)):
1497 name
, value
= cookie
.name
, cookie
.value
1499 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1500 # with no name, whereas http.cookiejar regards it as a
1501 # cookie with no value.
1502 name
, value
= '', name
1503 f
.write('%s\n' % '\t'.join((
1505 self
._true
_or
_false
(cookie
.domain
.startswith('.')),
1507 self
._true
_or
_false
(cookie
.secure
),
1508 str_or_none(cookie
.expires
, default
=''),
1512 def save(self
, filename
=None, *args
, **kwargs
):
1514 Save cookies to a file.
1515 Code is taken from CPython 3.6
1516 https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
1518 if filename
is None:
1519 if self
.filename
is not None:
1520 filename
= self
.filename
1522 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1524 # Store session cookies with `expires` set to 0 instead of an empty string
1526 if cookie
.expires
is None:
1529 with self
.open(filename
, write
=True) as f
:
1530 f
.write(self
._HEADER
)
1531 self
._really
_save
(f
, *args
, **kwargs
)
1533 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1534 """Load cookies from a file."""
1535 if filename
is None:
1536 if self
.filename
is not None:
1537 filename
= self
.filename
1539 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1541 def prepare_line(line
):
1542 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1543 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1544 # comments and empty lines are fine
1545 if line
.startswith('#') or not line
.strip():
1547 cookie_list
= line
.split('\t')
1548 if len(cookie_list
) != self
._ENTRY
_LEN
:
1549 raise compat_cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1550 cookie
= self
._CookieFileEntry
(*cookie_list
)
1551 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1552 raise compat_cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1556 with self
.open(filename
) as f
:
1559 cf
.write(prepare_line(line
))
1560 except compat_cookiejar
.LoadError
as e
:
1561 if f
'{line.strip()} '[0] in '[{"':
1562 raise compat_cookiejar
.LoadError(
1563 'Cookies file must be Netscape formatted, not JSON. See '
1564 'https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl')
1565 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1568 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1569 # Session cookies are denoted by either `expires` field set to
1570 # an empty string or 0. MozillaCookieJar only recognizes the former
1571 # (see [1]). So we need force the latter to be recognized as session
1572 # cookies on our own.
1573 # Session cookies may be important for cookies-based authentication,
1574 # e.g. usually, when user does not check 'Remember me' check box while
1575 # logging in on a site, some important cookies are stored as session
1576 # cookies so that not recognizing them will result in failed login.
1577 # 1. https://bugs.python.org/issue17164
1579 # Treat `expires=0` cookies as session cookies
1580 if cookie
.expires
== 0:
1581 cookie
.expires
= None
1582 cookie
.discard
= True
1585 class YoutubeDLCookieProcessor(compat_urllib_request
.HTTPCookieProcessor
):
1586 def __init__(self
, cookiejar
=None):
1587 compat_urllib_request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1589 def http_response(self
, request
, response
):
1590 return compat_urllib_request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1592 https_request
= compat_urllib_request
.HTTPCookieProcessor
.http_request
1593 https_response
= http_response
1596 class YoutubeDLRedirectHandler(compat_urllib_request
.HTTPRedirectHandler
):
1597 """YoutubeDL redirect handler
1599 The code is based on HTTPRedirectHandler implementation from CPython [1].
1601 This redirect handler solves two issues:
1602 - ensures redirect URL is always unicode under python 2
1603 - introduces support for experimental HTTP response status code
1604 308 Permanent Redirect [2] used by some sites [3]
1606 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1607 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1608 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1611 http_error_301
= http_error_303
= http_error_307
= http_error_308
= compat_urllib_request
.HTTPRedirectHandler
.http_error_302
1613 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1614 """Return a Request or None in response to a redirect.
1616 This is called by the http_error_30x methods when a
1617 redirection response is received. If a redirection should
1618 take place, return a new Request to allow http_error_30x to
1619 perform the redirect. Otherwise, raise HTTPError if no-one
1620 else should try to handle this url. Return None if you can't
1621 but another Handler might.
1623 m
= req
.get_method()
1624 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1625 or code
in (301, 302, 303) and m
== "POST")):
1626 raise compat_HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1627 # Strictly (according to RFC 2616), 301 or 302 in response to
1628 # a POST MUST NOT cause a redirection without confirmation
1629 # from the user (of urllib.request, in this case). In practice,
1630 # essentially all clients do redirect in this case, so we do
1633 # Be conciliant with URIs containing a space. This is mainly
1634 # redundant with the more complete encoding done in http_error_302(),
1635 # but it is kept for compatibility with other callers.
1636 newurl
= newurl
.replace(' ', '%20')
1638 CONTENT_HEADERS
= ("content-length", "content-type")
1639 # NB: don't use dict comprehension for python 2.6 compatibility
1640 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1642 # A 303 must either use GET or HEAD for subsequent request
1643 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1644 if code
== 303 and m
!= 'HEAD':
1646 # 301 and 302 redirects are commonly turned into a GET from a POST
1647 # for subsequent requests by browsers, so we'll do the same.
1648 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1649 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1650 if code
in (301, 302) and m
== 'POST':
1653 return compat_urllib_request
.Request(
1654 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1655 unverifiable
=True, method
=m
)
1658 def extract_timezone(date_str
):
1661 ^.{8,}? # >=8 char non-TZ prefix, if present
1662 (?P<tz>Z| # just the UTC Z, or
1663 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1664 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1665 [ ]? # optional space
1666 (?P<sign>\+|-) # +/-
1667 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1671 timezone
= datetime
.timedelta()
1673 date_str
= date_str
[:-len(m
.group('tz'))]
1674 if not m
.group('sign'):
1675 timezone
= datetime
.timedelta()
1677 sign
= 1 if m
.group('sign') == '+' else -1
1678 timezone
= datetime
.timedelta(
1679 hours
=sign
* int(m
.group('hours')),
1680 minutes
=sign
* int(m
.group('minutes')))
1681 return timezone
, date_str
1684 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1685 """ Return a UNIX timestamp from the given date """
1687 if date_str
is None:
1690 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1692 if timezone
is None:
1693 timezone
, date_str
= extract_timezone(date_str
)
1695 with contextlib
.suppress(ValueError):
1696 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1697 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1698 return calendar
.timegm(dt
.timetuple())
1701 def date_formats(day_first
=True):
1702 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1705 def unified_strdate(date_str
, day_first
=True):
1706 """Return a string with the date in the format YYYYMMDD"""
1708 if date_str
is None:
1712 date_str
= date_str
.replace(',', ' ')
1713 # Remove AM/PM + timezone
1714 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1715 _
, date_str
= extract_timezone(date_str
)
1717 for expression
in date_formats(day_first
):
1718 with contextlib
.suppress(ValueError):
1719 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1720 if upload_date
is None:
1721 timetuple
= email
.utils
.parsedate_tz(date_str
)
1723 with contextlib
.suppress(ValueError):
1724 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1725 if upload_date
is not None:
1726 return compat_str(upload_date
)
1729 def unified_timestamp(date_str
, day_first
=True):
1730 if date_str
is None:
1733 date_str
= re
.sub(r
'[,|]', '', date_str
)
1735 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1736 timezone
, date_str
= extract_timezone(date_str
)
1738 # Remove AM/PM + timezone
1739 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1741 # Remove unrecognized timezones from ISO 8601 alike timestamps
1742 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1744 date_str
= date_str
[:-len(m
.group('tz'))]
1746 # Python only supports microseconds, so remove nanoseconds
1747 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]+$', date_str
)
1749 date_str
= m
.group(1)
1751 for expression
in date_formats(day_first
):
1752 with contextlib
.suppress(ValueError):
1753 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1754 return calendar
.timegm(dt
.timetuple())
1755 timetuple
= email
.utils
.parsedate_tz(date_str
)
1757 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1760 def determine_ext(url
, default_ext
='unknown_video'):
1761 if url
is None or '.' not in url
:
1763 guess
= url
.partition('?')[0].rpartition('.')[2]
1764 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1766 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1767 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1768 return guess
.rstrip('/')
1773 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1774 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1777 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1779 Return a datetime object from a string.
1781 (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
1783 @param format strftime format of DATE
1784 @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
1785 auto: round to the unit provided in date_str (if applicable).
1787 auto_precision
= False
1788 if precision
== 'auto':
1789 auto_precision
= True
1790 precision
= 'microsecond'
1791 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1792 if date_str
in ('now', 'today'):
1794 if date_str
== 'yesterday':
1795 return today
- datetime
.timedelta(days
=1)
1797 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
1799 if match
is not None:
1800 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1801 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1802 unit
= match
.group('unit')
1803 if unit
== 'month' or unit
== 'year':
1804 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1810 delta
= datetime
.timedelta(**{unit + 's': time}
)
1811 new_date
= start_time
+ delta
1813 return datetime_round(new_date
, unit
)
1816 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1819 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1821 Return a date object from a string using datetime_from_str
1823 @param strict Restrict allowed patterns to "YYYYMMDD" and
1824 (now|today|yesterday)(-\d+(day|week|month|year)s?)?
1826 if strict
and not re
.fullmatch(r
'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str
):
1827 raise ValueError(f
'Invalid date format "{date_str}"')
1828 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1831 def datetime_add_months(dt
, months
):
1832 """Increment/Decrement a datetime object by months."""
1833 month
= dt
.month
+ months
- 1
1834 year
= dt
.year
+ month
// 12
1835 month
= month
% 12 + 1
1836 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1837 return dt
.replace(year
, month
, day
)
1840 def datetime_round(dt
, precision
='day'):
1842 Round a datetime object's time to a specific precision
1844 if precision
== 'microsecond':
1853 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1854 timestamp
= calendar
.timegm(dt
.timetuple())
1855 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1858 def hyphenate_date(date_str
):
1860 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1861 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1862 if match
is not None:
1863 return '-'.join(match
.groups())
1869 """Represents a time interval between two dates"""
1871 def __init__(self
, start
=None, end
=None):
1872 """start and end must be strings in the format accepted by date"""
1873 if start
is not None:
1874 self
.start
= date_from_str(start
, strict
=True)
1876 self
.start
= datetime
.datetime
.min.date()
1878 self
.end
= date_from_str(end
, strict
=True)
1880 self
.end
= datetime
.datetime
.max.date()
1881 if self
.start
> self
.end
:
1882 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1886 """Returns a range that only contains the given day"""
1887 return cls(day
, day
)
1889 def __contains__(self
, date
):
1890 """Check if the date is in the range"""
1891 if not isinstance(date
, datetime
.date
):
1892 date
= date_from_str(date
)
1893 return self
.start
<= date
<= self
.end
1896 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1899 def platform_name():
1900 """ Returns the platform name as a compat_str """
1901 res
= platform
.platform()
1902 if isinstance(res
, bytes):
1903 res
= res
.decode(preferredencoding())
1905 assert isinstance(res
, compat_str
)
1910 def get_windows_version():
1911 ''' Get Windows version. returns () if it's not running on Windows '''
1912 if compat_os_name
== 'nt':
1913 return version_tuple(platform
.win32_ver()[1])
1918 def write_string(s
, out
=None, encoding
=None):
1919 assert isinstance(s
, str)
1920 out
= out
or sys
.stderr
1922 if compat_os_name
== 'nt' and supports_terminal_sequences(out
):
1923 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1925 enc
, buffer = None, out
1926 if 'b' in getattr(out
, 'mode', ''):
1927 enc
= encoding
or preferredencoding()
1928 elif hasattr(out
, 'buffer'):
1930 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1932 buffer.write(s
.encode(enc
, 'ignore') if enc
else s
)
1936 def bytes_to_intlist(bs
):
1939 if isinstance(bs
[0], int): # Python 3
1942 return [ord(c
) for c
in bs
]
1945 def intlist_to_bytes(xs
):
1948 return compat_struct_pack('%dB' % len(xs
), *xs
)
1951 class LockingUnsupportedError(OSError):
1952 msg
= 'File locking is not supported'
1955 super().__init
__(self
.msg
)
1958 # Cross-platform file locking
1959 if sys
.platform
== 'win32':
1960 import ctypes
.wintypes
1963 class OVERLAPPED(ctypes
.Structure
):
1965 ('Internal', ctypes
.wintypes
.LPVOID
),
1966 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1967 ('Offset', ctypes
.wintypes
.DWORD
),
1968 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1969 ('hEvent', ctypes
.wintypes
.HANDLE
),
1972 kernel32
= ctypes
.windll
.kernel32
1973 LockFileEx
= kernel32
.LockFileEx
1974 LockFileEx
.argtypes
= [
1975 ctypes
.wintypes
.HANDLE
, # hFile
1976 ctypes
.wintypes
.DWORD
, # dwFlags
1977 ctypes
.wintypes
.DWORD
, # dwReserved
1978 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1979 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1980 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1982 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
1983 UnlockFileEx
= kernel32
.UnlockFileEx
1984 UnlockFileEx
.argtypes
= [
1985 ctypes
.wintypes
.HANDLE
, # hFile
1986 ctypes
.wintypes
.DWORD
, # dwReserved
1987 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1988 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1989 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1991 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
1992 whole_low
= 0xffffffff
1993 whole_high
= 0x7fffffff
1995 def _lock_file(f
, exclusive
, block
):
1996 overlapped
= OVERLAPPED()
1997 overlapped
.Offset
= 0
1998 overlapped
.OffsetHigh
= 0
1999 overlapped
.hEvent
= 0
2000 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2002 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
2003 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
2004 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2005 raise BlockingIOError('Locking file failed: %r' % ctypes
.FormatError())
2007 def _unlock_file(f
):
2008 assert f
._lock
_file
_overlapped
_p
2009 handle
= msvcrt
.get_osfhandle(f
.fileno())
2010 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2011 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2017 def _lock_file(f
, exclusive
, block
):
2018 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
2020 flags |
= fcntl
.LOCK_NB
2022 fcntl
.flock(f
, flags
)
2023 except BlockingIOError
:
2025 except OSError: # AOSP does not have flock()
2026 fcntl
.lockf(f
, flags
)
2028 def _unlock_file(f
):
2030 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2032 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
2036 def _lock_file(f
, exclusive
, block
):
2037 raise LockingUnsupportedError()
2039 def _unlock_file(f
):
2040 raise LockingUnsupportedError()
2046 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2047 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2048 raise NotImplementedError(mode
)
2049 self
.mode
, self
.block
= mode
, block
2051 writable
= any(f
in mode
for f
in 'wax+')
2052 readable
= any(f
in mode
for f
in 'r+')
2053 flags
= functools
.reduce(operator
.ior
, (
2054 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2055 getattr(os
, 'O_BINARY', 0), # Windows only
2056 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2057 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2058 os
.O_APPEND
if 'a' in mode
else 0,
2059 os
.O_EXCL
if 'x' in mode
else 0,
2060 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2063 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2065 def __enter__(self
):
2066 exclusive
= 'r' not in self
.mode
2068 _lock_file(self
.f
, exclusive
, self
.block
)
2073 if 'w' in self
.mode
:
2076 except OSError as e
:
2078 errno
.ESPIPE
, # Illegal seek - expected for FIFO
2079 errno
.EINVAL
, # Invalid argument - expected for /dev/null
2088 _unlock_file(self
.f
)
2092 def __exit__(self
, *_
):
2101 def __getattr__(self
, attr
):
2102 return getattr(self
.f
, attr
)
2109 def get_filesystem_encoding():
2110 encoding
= sys
.getfilesystemencoding()
2111 return encoding
if encoding
is not None else 'utf-8'
2114 def shell_quote(args
):
2116 encoding
= get_filesystem_encoding()
2118 if isinstance(a
, bytes):
2119 # We may get a filename encoded with 'encodeFilename'
2120 a
= a
.decode(encoding
)
2121 quoted_args
.append(compat_shlex_quote(a
))
2122 return ' '.join(quoted_args
)
2125 def smuggle_url(url
, data
):
2126 """ Pass additional data in a URL for internal use. """
2128 url
, idata
= unsmuggle_url(url
, {})
2130 sdata
= compat_urllib_parse_urlencode(
2131 {'__youtubedl_smuggle': json.dumps(data)}
)
2132 return url
+ '#' + sdata
2135 def unsmuggle_url(smug_url
, default
=None):
2136 if '#__youtubedl_smuggle' not in smug_url
:
2137 return smug_url
, default
2138 url
, _
, sdata
= smug_url
.rpartition('#')
2139 jsond
= compat_parse_qs(sdata
)['__youtubedl_smuggle'][0]
2140 data
= json
.loads(jsond
)
2144 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2145 """ Formats numbers with decimal sufixes like K, M, etc """
2146 num
, factor
= float_or_none(num
), float(factor
)
2147 if num
is None or num
< 0:
2149 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2150 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2151 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2153 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2154 converted
= num
/ (factor
** exponent
)
2155 return fmt
% (converted
, suffix
)
2158 def format_bytes(bytes):
2159 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2162 def lookup_unit_table(unit_table
, s
):
2163 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2165 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2168 num_str
= m
.group('num').replace(',', '.')
2169 mult
= unit_table
[m
.group('unit')]
2170 return int(float(num_str
) * mult
)
2173 def parse_filesize(s
):
2177 # The lower-case forms are of course incorrect and unofficial,
2178 # but we support those too
2195 'megabytes': 1000 ** 2,
2196 'mebibytes': 1024 ** 2,
2202 'gigabytes': 1000 ** 3,
2203 'gibibytes': 1024 ** 3,
2209 'terabytes': 1000 ** 4,
2210 'tebibytes': 1024 ** 4,
2216 'petabytes': 1000 ** 5,
2217 'pebibytes': 1024 ** 5,
2223 'exabytes': 1000 ** 6,
2224 'exbibytes': 1024 ** 6,
2230 'zettabytes': 1000 ** 7,
2231 'zebibytes': 1024 ** 7,
2237 'yottabytes': 1000 ** 8,
2238 'yobibytes': 1024 ** 8,
2241 return lookup_unit_table(_UNIT_TABLE
, s
)
2248 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2250 if re
.match(r
'^[\d,.]+$', s
):
2251 return str_to_int(s
)
2264 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2268 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2270 return str_to_int(mobj
.group(1))
2273 def parse_resolution(s
, *, lenient
=False):
2278 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2280 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2283 'width': int(mobj
.group('w')),
2284 'height': int(mobj
.group('h')),
2287 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2289 return {'height': int(mobj.group(1))}
2291 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2293 return {'height': int(mobj.group(1)) * 540}
2298 def parse_bitrate(s
):
2299 if not isinstance(s
, compat_str
):
2301 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2303 return int(mobj
.group(1))
2306 def month_by_name(name
, lang
='en'):
2307 """ Return the number of a month by (locale-independently) English name """
2309 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2312 return month_names
.index(name
) + 1
2317 def month_by_abbreviation(abbrev
):
2318 """ Return the number of a month by (locale-independently) English
2322 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2327 def fix_xml_ampersands(xml_str
):
2328 """Replace all the '&' by '&' in XML"""
2330 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2335 def setproctitle(title
):
2336 assert isinstance(title
, compat_str
)
2338 # ctypes in Jython is not complete
2339 # http://bugs.jython.org/issue2148
2340 if sys
.platform
.startswith('java'):
2344 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2348 # LoadLibrary in Windows Python 2.7.13 only expects
2349 # a bytestring, but since unicode_literals turns
2350 # every string into a unicode string, it fails.
2352 title_bytes
= title
.encode()
2353 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2354 buf
.value
= title_bytes
2356 libc
.prctl(15, buf
, 0, 0, 0)
2357 except AttributeError:
2358 return # Strange libc, just skip this
2361 def remove_start(s
, start
):
2362 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2365 def remove_end(s
, end
):
2366 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2369 def remove_quotes(s
):
2370 if s
is None or len(s
) < 2:
2372 for quote
in ('"', "'", ):
2373 if s
[0] == quote
and s
[-1] == quote
:
2378 def get_domain(url
):
2379 domain
= re
.match(r
'(?:https?:\/\/)?(?:www\.)?(?P<domain>[^\n\/]+\.[^\n\/]+)(?:\/(.*))?', url
)
2380 return domain
.group('domain') if domain
else None
2383 def url_basename(url
):
2384 path
= compat_urlparse
.urlparse(url
).path
2385 return path
.strip('/').split('/')[-1]
2389 return re
.match(r
'https?://[^?#&]+/', url
).group()
2392 def urljoin(base
, path
):
2393 if isinstance(path
, bytes):
2394 path
= path
.decode()
2395 if not isinstance(path
, compat_str
) or not path
:
2397 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2399 if isinstance(base
, bytes):
2400 base
= base
.decode()
2401 if not isinstance(base
, compat_str
) or not re
.match(
2402 r
'^(?:https?:)?//', base
):
2404 return compat_urlparse
.urljoin(base
, path
)
2407 class HEADRequest(compat_urllib_request
.Request
):
2408 def get_method(self
):
2412 class PUTRequest(compat_urllib_request
.Request
):
2413 def get_method(self
):
2417 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2418 if get_attr
and v
is not None:
2419 v
= getattr(v
, get_attr
, None)
2421 return int(v
) * invscale
// scale
2422 except (ValueError, TypeError, OverflowError):
2426 def str_or_none(v
, default
=None):
2427 return default
if v
is None else compat_str(v
)
2430 def str_to_int(int_str
):
2431 """ A more relaxed version of int_or_none """
2432 if isinstance(int_str
, int):
2434 elif isinstance(int_str
, compat_str
):
2435 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2436 return int_or_none(int_str
)
2439 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2443 return float(v
) * invscale
/ scale
2444 except (ValueError, TypeError):
2448 def bool_or_none(v
, default
=None):
2449 return v
if isinstance(v
, bool) else default
2452 def strip_or_none(v
, default
=None):
2453 return v
.strip() if isinstance(v
, compat_str
) else default
2456 def url_or_none(url
):
2457 if not url
or not isinstance(url
, compat_str
):
2460 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2463 def request_to_url(req
):
2464 if isinstance(req
, compat_urllib_request
.Request
):
2465 return req
.get_full_url()
2470 def strftime_or_none(timestamp
, date_format
, default
=None):
2471 datetime_object
= None
2473 if isinstance(timestamp
, (int, float)): # unix timestamp
2474 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2475 elif isinstance(timestamp
, compat_str
): # assume YYYYMMDD
2476 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2477 return datetime_object
.strftime(date_format
)
2478 except (ValueError, TypeError, AttributeError):
2482 def parse_duration(s
):
2483 if not isinstance(s
, str):
2489 days
, hours
, mins
, secs
, ms
= [None] * 5
2490 m
= re
.match(r
'''(?x)
2492 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2493 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2494 (?P<ms>[.:][0-9]+)?Z?$
2497 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2502 [0-9]+\s*y(?:ears?)?,?\s*
2505 [0-9]+\s*m(?:onths?)?,?\s*
2508 [0-9]+\s*w(?:eeks?)?,?\s*
2511 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2515 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2518 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2521 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2524 days
, hours
, mins
, secs
, ms
= m
.groups()
2526 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2528 hours
, mins
= m
.groups()
2533 ms
= ms
.replace(':', '.')
2534 return sum(float(part
or 0) * mult
for part
, mult
in (
2535 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2538 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2539 name
, real_ext
= os
.path
.splitext(filename
)
2541 f
'{name}.{ext}{real_ext}'
2542 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2543 else f
'{filename}.{ext}')
2546 def replace_extension(filename
, ext
, expected_real_ext
=None):
2547 name
, real_ext
= os
.path
.splitext(filename
)
2548 return '{}.{}'.format(
2549 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2553 def check_executable(exe
, args
=[]):
2554 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2555 args can be a list of arguments for a short output (like -version) """
2557 Popen([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
).communicate_or_kill()
2563 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2565 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2567 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2568 # SIGTTOU if yt-dlp is run in the background.
2569 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2571 [encodeArgument(exe
)] + args
, stdin
=subprocess
.PIPE
,
2572 stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
).communicate_or_kill()
2575 if isinstance(out
, bytes): # Python 2.x
2576 out
= out
.decode('ascii', 'ignore')
2580 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2581 assert isinstance(output
, compat_str
)
2582 if version_re
is None:
2583 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2584 m
= re
.search(version_re
, output
)
2591 def get_exe_version(exe
, args
=['--version'],
2592 version_re
=None, unrecognized
='present'):
2593 """ Returns the version of the specified executable,
2594 or False if the executable is not present """
2595 out
= _get_exe_version_output(exe
, args
)
2596 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2599 class LazyList(collections
.abc
.Sequence
):
2600 """Lazy immutable list from an iterable
2601 Note that slices of a LazyList are lists and not LazyList"""
2603 class IndexError(IndexError):
2606 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2607 self
._iterable
= iter(iterable
)
2608 self
._cache
= [] if _cache
is None else _cache
2609 self
._reversed
= reverse
2613 # We need to consume the entire iterable to iterate in reverse
2614 yield from self
.exhaust()
2616 yield from self
._cache
2617 for item
in self
._iterable
:
2618 self
._cache
.append(item
)
2622 self
._cache
.extend(self
._iterable
)
2623 self
._iterable
= [] # Discard the emptied iterable to make it pickle-able
2627 """Evaluate the entire iterable"""
2628 return self
._exhaust
()[::-1 if self
._reversed
else 1]
2631 def _reverse_index(x
):
2632 return None if x
is None else -(x
+ 1)
2634 def __getitem__(self
, idx
):
2635 if isinstance(idx
, slice):
2637 idx
= slice(self
._reverse
_index
(idx
.start
), self
._reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2638 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2639 elif isinstance(idx
, int):
2641 idx
= self
._reverse
_index
(idx
)
2642 start
, stop
, step
= idx
, idx
, 0
2644 raise TypeError('indices must be integers or slices')
2645 if ((start
or 0) < 0 or (stop
or 0) < 0
2646 or (start
is None and step
< 0)
2647 or (stop
is None and step
> 0)):
2648 # We need to consume the entire iterable to be able to slice from the end
2649 # Obviously, never use this with infinite iterables
2652 return self
._cache
[idx
]
2653 except IndexError as e
:
2654 raise self
.IndexError(e
) from e
2655 n
= max(start
or 0, stop
or 0) - len(self
._cache
) + 1
2657 self
._cache
.extend(itertools
.islice(self
._iterable
, n
))
2659 return self
._cache
[idx
]
2660 except IndexError as e
:
2661 raise self
.IndexError(e
) from e
2665 self
[-1] if self
._reversed
else self
[0]
2666 except self
.IndexError:
2672 return len(self
._cache
)
2674 def __reversed__(self
):
2675 return type(self
)(self
._iterable
, reverse
=not self
._reversed
, _cache
=self
._cache
)
2678 return type(self
)(self
._iterable
, reverse
=self
._reversed
, _cache
=self
._cache
)
2681 # repr and str should mimic a list. So we exhaust the iterable
2682 return repr(self
.exhaust())
2685 return repr(self
.exhaust())
2690 class IndexError(IndexError):
2694 # This is only useful for tests
2695 return len(self
.getslice())
2697 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2698 self
._pagefunc
= pagefunc
2699 self
._pagesize
= pagesize
2700 self
._pagecount
= float('inf')
2701 self
._use
_cache
= use_cache
2704 def getpage(self
, pagenum
):
2705 page_results
= self
._cache
.get(pagenum
)
2706 if page_results
is None:
2707 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2709 self
._cache
[pagenum
] = page_results
2712 def getslice(self
, start
=0, end
=None):
2713 return list(self
._getslice
(start
, end
))
2715 def _getslice(self
, start
, end
):
2716 raise NotImplementedError('This method must be implemented by subclasses')
2718 def __getitem__(self
, idx
):
2719 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2720 if not isinstance(idx
, int) or idx
< 0:
2721 raise TypeError('indices must be non-negative integers')
2722 entries
= self
.getslice(idx
, idx
+ 1)
2724 raise self
.IndexError()
2728 class OnDemandPagedList(PagedList
):
2729 """Download pages until a page with less than maximum results"""
2731 def _getslice(self
, start
, end
):
2732 for pagenum
in itertools
.count(start
// self
._pagesize
):
2733 firstid
= pagenum
* self
._pagesize
2734 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2735 if start
>= nextfirstid
:
2739 start
% self
._pagesize
2740 if firstid
<= start
< nextfirstid
2743 ((end
- 1) % self
._pagesize
) + 1
2744 if (end
is not None and firstid
<= end
<= nextfirstid
)
2748 page_results
= self
.getpage(pagenum
)
2750 self
._pagecount
= pagenum
- 1
2752 if startv
!= 0 or endv
is not None:
2753 page_results
= page_results
[startv
:endv
]
2754 yield from page_results
2756 # A little optimization - if current page is not "full", ie. does
2757 # not contain page_size videos then we can assume that this page
2758 # is the last one - there are no more ids on further pages -
2759 # i.e. no need to query again.
2760 if len(page_results
) + startv
< self
._pagesize
:
2763 # If we got the whole page, but the next page is not interesting,
2764 # break out early as well
2765 if end
== nextfirstid
:
2769 class InAdvancePagedList(PagedList
):
2770 """PagedList with total number of pages known in advance"""
2772 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2773 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2774 self
._pagecount
= pagecount
2776 def _getslice(self
, start
, end
):
2777 start_page
= start
// self
._pagesize
2778 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2779 skip_elems
= start
- start_page
* self
._pagesize
2780 only_more
= None if end
is None else end
- start
2781 for pagenum
in range(start_page
, end_page
):
2782 page_results
= self
.getpage(pagenum
)
2784 page_results
= page_results
[skip_elems
:]
2786 if only_more
is not None:
2787 if len(page_results
) < only_more
:
2788 only_more
-= len(page_results
)
2790 yield from page_results
[:only_more
]
2792 yield from page_results
2795 def uppercase_escape(s
):
2796 unicode_escape
= codecs
.getdecoder('unicode_escape')
2798 r
'\\U[0-9a-fA-F]{8}',
2799 lambda m
: unicode_escape(m
.group(0))[0],
2803 def lowercase_escape(s
):
2804 unicode_escape
= codecs
.getdecoder('unicode_escape')
2806 r
'\\u[0-9a-fA-F]{4}',
2807 lambda m
: unicode_escape(m
.group(0))[0],
2811 def escape_rfc3986(s
):
2812 """Escape non-ASCII characters as suggested by RFC 3986"""
2813 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2816 def escape_url(url
):
2817 """Escape URL as suggested by RFC 3986"""
2818 url_parsed
= compat_urllib_parse_urlparse(url
)
2819 return url_parsed
._replace
(
2820 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2821 path
=escape_rfc3986(url_parsed
.path
),
2822 params
=escape_rfc3986(url_parsed
.params
),
2823 query
=escape_rfc3986(url_parsed
.query
),
2824 fragment
=escape_rfc3986(url_parsed
.fragment
)
2829 return compat_parse_qs(compat_urllib_parse_urlparse(url
).query
)
2832 def read_batch_urls(batch_fd
):
2834 if not isinstance(url
, compat_str
):
2835 url
= url
.decode('utf-8', 'replace')
2836 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
2837 for bom
in BOM_UTF8
:
2838 if url
.startswith(bom
):
2839 url
= url
[len(bom
):]
2841 if not url
or url
.startswith(('#', ';', ']')):
2843 # "#" cannot be stripped out since it is part of the URI
2844 # However, it can be safely stipped out if follwing a whitespace
2845 return re
.split(r
'\s#', url
, 1)[0].rstrip()
2847 with contextlib
.closing(batch_fd
) as fd
:
2848 return [url
for url
in map(fixup
, fd
) if url
]
2851 def urlencode_postdata(*args
, **kargs
):
2852 return compat_urllib_parse_urlencode(*args
, **kargs
).encode('ascii')
2855 def update_url_query(url
, query
):
2858 parsed_url
= compat_urlparse
.urlparse(url
)
2859 qs
= compat_parse_qs(parsed_url
.query
)
2861 return compat_urlparse
.urlunparse(parsed_url
._replace
(
2862 query
=compat_urllib_parse_urlencode(qs
, True)))
2865 def update_Request(req
, url
=None, data
=None, headers
={}, query={}
):
2866 req_headers
= req
.headers
.copy()
2867 req_headers
.update(headers
)
2868 req_data
= data
or req
.data
2869 req_url
= update_url_query(url
or req
.get_full_url(), query
)
2870 req_get_method
= req
.get_method()
2871 if req_get_method
== 'HEAD':
2872 req_type
= HEADRequest
2873 elif req_get_method
== 'PUT':
2874 req_type
= PUTRequest
2876 req_type
= compat_urllib_request
.Request
2878 req_url
, data
=req_data
, headers
=req_headers
,
2879 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
2880 if hasattr(req
, 'timeout'):
2881 new_req
.timeout
= req
.timeout
2885 def _multipart_encode_impl(data
, boundary
):
2886 content_type
= 'multipart/form-data; boundary=%s' % boundary
2889 for k
, v
in data
.items():
2890 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
2891 if isinstance(k
, compat_str
):
2893 if isinstance(v
, compat_str
):
2895 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
2896 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
2897 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
2898 if boundary
.encode('ascii') in content
:
2899 raise ValueError('Boundary overlaps with data')
2902 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
2904 return out
, content_type
2907 def multipart_encode(data
, boundary
=None):
2909 Encode a dict to RFC 7578-compliant form-data
2912 A dict where keys and values can be either Unicode or bytes-like
2915 If specified a Unicode object, it's used as the boundary. Otherwise
2916 a random boundary is generated.
2918 Reference: https://tools.ietf.org/html/rfc7578
2920 has_specified_boundary
= boundary
is not None
2923 if boundary
is None:
2924 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
2927 out
, content_type
= _multipart_encode_impl(data
, boundary
)
2930 if has_specified_boundary
:
2934 return out
, content_type
2937 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
2938 for val
in map(d
.get
, variadic(key_or_keys
)):
2939 if val
is not None and (val
or not skip_false_values
):
2944 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
2947 val
= f(*args
, **kwargs
)
2948 except (AttributeError, KeyError, TypeError, IndexError, ZeroDivisionError):
2951 if expected_type
is None or isinstance(val
, expected_type
):
2955 def try_get(src
, getter
, expected_type
=None):
2956 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
2959 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
2960 return {k: v for k, v in dct.items() if cndn(k, v)}
2963 def merge_dicts(*dicts
):
2965 for a_dict
in dicts
:
2966 for k
, v
in a_dict
.items():
2967 if (v
is not None and k
not in merged
2968 or isinstance(v
, str) and merged
[k
] == ''):
2973 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
2974 return string
if isinstance(string
, compat_str
) else compat_str(string
, encoding
, errors
)
2986 TV_PARENTAL_GUIDELINES
= {
2996 def parse_age_limit(s
):
2997 # isinstance(False, int) is True. So type() must be used instead
2998 if type(s
) is int: # noqa: E721
2999 return s
if 0 <= s
<= 21 else None
3000 elif not isinstance(s
, str):
3002 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3004 return int(m
.group('age'))
3007 return US_RATINGS
[s
]
3008 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3010 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3014 def strip_jsonp(code
):
3017 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3018 (?:\s*&&\s*(?P=func_name))?
3019 \s*\(\s*(?P<callback_data>.*)\);?
3020 \s*?(?://[^\n]*)*$''',
3021 r
'\g<callback_data>', code
)
3024 def js_to_json(code
, vars={}):
3025 # vars is a dict of var, val pairs to substitute
3026 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3027 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
3029 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
3030 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
3035 if v
in ('true', 'false', 'null'):
3037 elif v
in ('undefined', 'void 0'):
3039 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3042 if v
[0] in ("'", '"'):
3043 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3048 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3050 for regex
, base
in INTEGER_TABLE
:
3051 im
= re
.match(regex
, v
)
3053 i
= int(im
.group(1), base
)
3054 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3061 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3063 return re
.sub(r
'''(?sx)
3064 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3065 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3066 {comment}|,(?={skip}[\]}}])|
3067 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3068 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3071 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3074 def qualities(quality_ids
):
3075 """ Get a numeric quality value out of a list of possible values """
3078 return quality_ids
.index(qid
)
3084 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'after_move', 'post_process', 'after_video', 'playlist')
3088 'default': '%(title)s [%(id)s].%(ext)s',
3089 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3095 'description': 'description',
3096 'annotation': 'annotations.xml',
3097 'infojson': 'info.json',
3100 'pl_thumbnail': None,
3101 'pl_description': 'description',
3102 'pl_infojson': 'info.json',
3105 # As of [1] format syntax is:
3106 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3107 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3108 STR_FORMAT_RE_TMPL
= r
'''(?x)
3109 (?<!%)(?P<prefix>(?:%%)*)
3111 (?P<has_key>\((?P<key>{0})\))?
3113 (?P<conversion>[#0\-+ ]+)?
3115 (?P<precision>\.\d+)?
3116 (?P<len_mod>[hlL])? # unused in python
3117 {1} # conversion type
3122 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3125 def limit_length(s
, length
):
3126 """ Add ellipses to overly long strings """
3131 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3135 def version_tuple(v
):
3136 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3139 def is_outdated_version(version
, limit
, assume_new
=True):
3141 return not assume_new
3143 return version_tuple(version
) < version_tuple(limit
)
3145 return not assume_new
3148 def ytdl_is_updateable():
3149 """ Returns if yt-dlp can be updated with -U """
3151 from .update
import is_non_updateable
3153 return not is_non_updateable()
3156 def args_to_str(args
):
3157 # Get a short string representation for a subprocess command
3158 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3161 def error_to_compat_str(err
):
3165 def error_to_str(err
):
3166 return f
'{type(err).__name__}: {err}'
3169 def mimetype2ext(mt
):
3173 mt
, _
, params
= mt
.partition(';')
3178 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3179 # it's the most popular one
3180 'audio/mpeg': 'mp3',
3181 'audio/x-wav': 'wav',
3183 'audio/wave': 'wav',
3186 ext
= FULL_MAP
.get(mt
)
3192 'smptett+xml': 'tt',
3196 'x-mp4-fragmented': 'mp4',
3197 'x-ms-sami': 'sami',
3200 'x-mpegurl': 'm3u8',
3201 'vnd.apple.mpegurl': 'm3u8',
3205 'vnd.ms-sstr+xml': 'ism',
3209 'filmstrip+json': 'fs',
3213 _
, _
, subtype
= mt
.rpartition('/')
3214 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3225 _
, _
, suffix
= subtype
.partition('+')
3226 ext
= SUFFIX_MAP
.get(suffix
)
3230 return subtype
.replace('+', '.')
3233 def ext2mimetype(ext_or_url
):
3236 if '.' not in ext_or_url
:
3237 ext_or_url
= f
'file.{ext_or_url}'
3238 return mimetypes
.guess_type(ext_or_url
)[0]
3241 def parse_codecs(codecs_str
):
3242 # http://tools.ietf.org/html/rfc6381
3245 split_codecs
= list(filter(None, map(
3246 str.strip
, codecs_str
.strip().strip(',').split(','))))
3247 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3248 for full_codec
in split_codecs
:
3249 parts
= full_codec
.split('.')
3250 codec
= parts
[0].replace('0', '')
3251 if codec
in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3252 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3254 vcodec
= '.'.join(parts
[:4]) if codec
in ('vp9', 'av1', 'hvc1') else full_codec
3255 if codec
in ('dvh1', 'dvhe'):
3257 elif codec
== 'av1' and len(parts
) > 3 and parts
[3] == '10':
3259 elif full_codec
.replace('0', '').startswith('vp9.2'):
3261 elif codec
in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3264 elif codec
in ('stpp', 'wvtt',):
3268 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3269 if vcodec
or acodec
or scodec
:
3271 'vcodec': vcodec
or 'none',
3272 'acodec': acodec
or 'none',
3273 'dynamic_range': hdr
,
3274 **({'scodec': scodec}
if scodec
is not None else {}),
3276 elif len(split_codecs
) == 2:
3278 'vcodec': split_codecs
[0],
3279 'acodec': split_codecs
[1],
3284 def urlhandle_detect_ext(url_handle
):
3285 getheader
= url_handle
.headers
.get
3287 cd
= getheader('Content-Disposition')
3289 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3291 e
= determine_ext(m
.group('filename'), default_ext
=None)
3295 return mimetype2ext(getheader('Content-Type'))
3298 def encode_data_uri(data
, mime_type
):
3299 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3302 def age_restricted(content_limit
, age_limit
):
3303 """ Returns True iff the content should be blocked """
3305 if age_limit
is None: # No limit set
3307 if content_limit
is None:
3308 return False # Content available for everyone
3309 return age_limit
< content_limit
3312 def is_html(first_bytes
):
3313 """ Detect whether a file contains HTML by examining its first bytes. """
3316 (b
'\xef\xbb\xbf', 'utf-8'),
3317 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3318 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3319 (b
'\xff\xfe', 'utf-16-le'),
3320 (b
'\xfe\xff', 'utf-16-be'),
3324 for bom
, enc
in BOMS
:
3325 while first_bytes
.startswith(bom
):
3326 encoding
, first_bytes
= enc
, first_bytes
[len(bom
):]
3328 return re
.match(r
'^\s*<', first_bytes
.decode(encoding
, 'replace'))
3331 def determine_protocol(info_dict
):
3332 protocol
= info_dict
.get('protocol')
3333 if protocol
is not None:
3336 url
= sanitize_url(info_dict
['url'])
3337 if url
.startswith('rtmp'):
3339 elif url
.startswith('mms'):
3341 elif url
.startswith('rtsp'):
3344 ext
= determine_ext(url
)
3350 return compat_urllib_parse_urlparse(url
).scheme
3353 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3354 """ Render a list of rows, each as a list of values.
3355 Text after a \t will be right aligned """
3357 return len(remove_terminal_sequences(string
).replace('\t', ''))
3359 def get_max_lens(table
):
3360 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3362 def filter_using_list(row
, filterArray
):
3363 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3365 max_lens
= get_max_lens(data
) if hide_empty
else []
3366 header_row
= filter_using_list(header_row
, max_lens
)
3367 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3369 table
= [header_row
] + data
3370 max_lens
= get_max_lens(table
)
3373 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3374 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3376 for pos
, text
in enumerate(map(str, row
)):
3378 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3380 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3381 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3385 def _match_one(filter_part
, dct
, incomplete
):
3386 # TODO: Generalize code with YoutubeDL._build_format_filter
3387 STRING_OPERATORS
= {
3388 '*=': operator
.contains
,
3389 '^=': lambda attr
, value
: attr
.startswith(value
),
3390 '$=': lambda attr
, value
: attr
.endswith(value
),
3391 '~=': lambda attr
, value
: re
.search(value
, attr
),
3393 COMPARISON_OPERATORS
= {
3395 '<=': operator
.le
, # "<=" must be defined above "<"
3402 if isinstance(incomplete
, bool):
3403 is_incomplete
= lambda _
: incomplete
3405 is_incomplete
= lambda k
: k
in incomplete
3407 operator_rex
= re
.compile(r
'''(?x)\s*
3409 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3411 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3415 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3416 m = operator_rex.search(filter_part)
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
)\s
*
3453 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3455 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3456 m = operator_rex.search(filter_part)
3458 op = UNARY_OPERATORS[m.group('op')]
3459 actual_value = dct.get(m.group('key'))
3460 if is_incomplete(m.group('key')) and actual_value is None:
3462 return op(actual_value)
3464 raise ValueError('Invalid filter part %r' % filter_part)
3467 def match_str(filter_str, dct, incomplete=False):
3468 """ Filter a dictionary with a simple string syntax.
3469 @returns Whether the filter passes
3470 @param incomplete Set of keys that is expected to be missing from dct.
3471 Can be True/False to indicate all/none of the keys may be missing.
3472 All conditions on incomplete keys pass if the key is missing
3475 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3476 for filter_part in re.split(r'(?<!\\)&', filter_str))
3479 def match_filter_func(filters):
3482 filters = set(variadic(filters))
3484 interactive = '-' in filters
3488 def _match_func(info_dict, incomplete=False):
3489 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3490 return NO_DEFAULT if interactive and not incomplete else None
3492 video_title = info_dict.get('title') or info_dict.get('id') or 'video'
3493 filter_str = ') | ('.join(map(str.strip, filters))
3494 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3498 def download_range_func(chapters, ranges):
3499 def inner(info_dict, ydl):
3500 warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
3501 else 'Chapter information is unavailable')
3502 for regex in chapters or []:
3503 for i, chapter in enumerate(info_dict.get('chapters') or []):
3504 if re.search(regex, chapter['title']):
3506 yield {**chapter, 'index': i}
3508 ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
3510 yield from ({'start_time': start, 'end_time': end} for start, end in ranges or [])
3515 def parse_dfxp_time_expr(time_expr):
3519 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3521 return float(mobj.group('time_offset'))
3523 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3525 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3528 def srt_subtitles_timecode(seconds):
3529 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3532 def ass_subtitles_timecode(seconds):
3533 time = timetuple_from_msec(seconds * 1000)
3534 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3537 def dfxp2srt(dfxp_data):
3539 @param dfxp_data A
bytes-like
object containing DFXP data
3540 @returns A
unicode object containing converted SRT data
3542 LEGACY_NAMESPACES = (
3543 (b'http://www.w3.org/ns/ttml', [
3544 b'http://www.w3.org/2004/11/ttaf1',
3545 b'http://www.w3.org/2006/04/ttaf1',
3546 b'http://www.w3.org/2006/10/ttaf1',
3548 (b'http://www.w3.org/ns/ttml#styling', [
3549 b'http://www.w3.org/ns/ttml#style',
3553 SUPPORTED_STYLING = [
3562 _x = functools.partial(xpath_with_ns, ns_map={
3563 'xml': 'http://www.w3.org/XML/1998/namespace',
3564 'ttml': 'http://www.w3.org/ns/ttml',
3565 'tts': 'http://www.w3.org/ns/ttml#styling',
3571 class TTMLPElementParser:
3573 _unclosed_elements = []
3574 _applied_styles = []
3576 def start(self, tag, attrib):
3577 if tag in (_x('ttml:br'), 'br'):
3580 unclosed_elements = []
3582 element_style_id = attrib.get('style')
3584 style.update(default_style)
3585 if element_style_id:
3586 style.update(styles.get(element_style_id, {}))
3587 for prop in SUPPORTED_STYLING:
3588 prop_val = attrib.get(_x('tts:' + prop))
3590 style[prop] = prop_val
3593 for k, v in sorted(style.items()):
3594 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3597 font += ' color="%s"' % v
3598 elif k == 'fontSize':
3599 font += ' size="%s"' % v
3600 elif k == 'fontFamily':
3601 font += ' face="%s"' % v
3602 elif k == 'fontWeight' and v == 'bold':
3604 unclosed_elements.append('b')
3605 elif k == 'fontStyle' and v == 'italic':
3607 unclosed_elements.append('i')
3608 elif k == 'textDecoration' and v == 'underline':
3610 unclosed_elements.append('u')
3612 self._out += '<font' + font + '>'
3613 unclosed_elements.append('font')
3615 if self._applied_styles:
3616 applied_style.update(self._applied_styles[-1])
3617 applied_style.update(style)
3618 self._applied_styles.append(applied_style)
3619 self._unclosed_elements.append(unclosed_elements)
3622 if tag not in (_x('ttml:br'), 'br'):
3623 unclosed_elements = self._unclosed_elements.pop()
3624 for element in reversed(unclosed_elements):
3625 self._out += '</%s>' % element
3626 if unclosed_elements and self._applied_styles:
3627 self._applied_styles.pop()
3629 def data(self, data):
3633 return self._out.strip()
3635 def parse_node(node):
3636 target = TTMLPElementParser()
3637 parser = xml.etree.ElementTree.XMLParser(target=target)
3638 parser.feed(xml.etree.ElementTree.tostring(node))
3639 return parser.close()
3641 for k, v in LEGACY_NAMESPACES:
3643 dfxp_data = dfxp_data.replace(ns, k)
3645 dfxp = compat_etree_fromstring(dfxp_data)
3647 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3650 raise ValueError('Invalid dfxp/TTML subtitle')
3654 for style in dfxp.findall(_x('.//ttml:style')):
3655 style_id = style.get('id') or style.get(_x('xml:id'))
3658 parent_style_id = style.get('style')
3660 if parent_style_id not in styles:
3663 styles[style_id] = styles[parent_style_id].copy()
3664 for prop in SUPPORTED_STYLING:
3665 prop_val = style.get(_x('tts:' + prop))
3667 styles.setdefault(style_id, {})[prop] = prop_val
3673 for p in ('body', 'div'):
3674 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3677 style = styles.get(ele.get('style'))
3680 default_style.update(style)
3682 for para, index in zip(paras, itertools.count(1)):
3683 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3684 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3685 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3686 if begin_time is None:
3691 end_time = begin_time + dur
3692 out.append('%d\n%s --> %s\n%s\n\n' % (
3694 srt_subtitles_timecode(begin_time),
3695 srt_subtitles_timecode(end_time),
3701 def cli_option(params, command_option, param, separator=None):
3702 param = params.get(param)
3703 return ([] if param is None
3704 else [command_option, str(param)] if separator is None
3705 else [f'{command_option}{separator}{param}'])
3708 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3709 param = params.get(param)
3710 assert param in (True, False, None)
3711 return cli_option({True: true_value, False: false_value}, command_option, param, separator)
3714 def cli_valueless_option(params, command_option, param, expected_value=True):
3715 return [command_option] if params.get(param) == expected_value else []
3718 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3719 if isinstance(argdict, (list, tuple)): # for backward compatibility
3726 assert isinstance(argdict, dict)
3728 assert isinstance(keys, (list, tuple))
3729 for key_list in keys:
3730 arg_list = list(filter(
3731 lambda x: x is not None,
3732 [argdict.get(key.lower()) for key in variadic(key_list)]))
3734 return [arg for args in arg_list for arg in args]
3738 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3739 main_key, exe = main_key.lower(), exe.lower()
3740 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3741 keys = [f'{root_key}{k}' for k in (keys or [''])]
3742 if root_key in keys:
3744 keys.append((main_key, exe))
3745 keys.append('default')
3748 return cli_configuration_args(argdict, keys, default, use_compat)
3752 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
3811 'iw': 'heb', # Replaced by he in 1989 revision
3821 'in': 'ind', # Replaced by id in 1989 revision
3936 'ji': 'yid', # Replaced by yi in 1989 revision
3944 def short2long(cls, code):
3945 """Convert language code from ISO 639-1 to ISO 639-2/T"""
3946 return cls._lang_map.get(code[:2])
3949 def long2short(cls, code):
3950 """Convert language code from ISO 639-2/T to ISO 639-1"""
3951 for short_name, long_name in cls._lang_map.items():
3952 if long_name == code:
3957 # From http://data.okfn.org/data/core/country-list
3959 'AF': 'Afghanistan',
3960 'AX': 'Åland Islands',
3963 'AS': 'American Samoa',
3968 'AG': 'Antigua and Barbuda',
3985 'BO': 'Bolivia, Plurinational State of',
3986 'BQ': 'Bonaire, Sint Eustatius and Saba',
3987 'BA': 'Bosnia and Herzegovina',
3989 'BV': 'Bouvet Island',
3991 'IO': 'British Indian Ocean Territory',
3992 'BN': 'Brunei Darussalam',
3994 'BF': 'Burkina Faso',
4000 'KY': 'Cayman Islands',
4001 'CF': 'Central African Republic',
4005 'CX': 'Christmas Island',
4006 'CC': 'Cocos (Keeling) Islands',
4010 'CD': 'Congo, the Democratic Republic of the',
4011 'CK': 'Cook Islands',
4013 'CI': 'Côte d\'Ivoire',
4018 'CZ': 'Czech Republic',
4022 'DO': 'Dominican Republic',
4025 'SV': 'El Salvador',
4026 'GQ': 'Equatorial Guinea',
4030 'FK': 'Falkland Islands (Malvinas)',
4031 'FO': 'Faroe Islands',
4035 'GF': 'French Guiana',
4036 'PF': 'French Polynesia',
4037 'TF': 'French Southern Territories',
4052 'GW': 'Guinea-Bissau',
4055 'HM': 'Heard Island and McDonald Islands',
4056 'VA': 'Holy See (Vatican City State)',
4063 'IR': 'Iran, Islamic Republic of',
4066 'IM': 'Isle of Man',
4076 'KP': 'Korea, Democratic People\'s Republic of',
4077 'KR': 'Korea, Republic of',
4080 'LA': 'Lao People\'s Democratic Republic',
4086 'LI': 'Liechtenstein',
4090 'MK': 'Macedonia, the Former Yugoslav Republic of',
4097 'MH': 'Marshall Islands',
4103 'FM': 'Micronesia, Federated States of',
4104 'MD': 'Moldova, Republic of',
4115 'NL': 'Netherlands',
4116 'NC': 'New Caledonia',
4117 'NZ': 'New Zealand',
4122 'NF': 'Norfolk Island',
4123 'MP': 'Northern Mariana Islands',
4128 'PS': 'Palestine, State of',
4130 'PG': 'Papua New Guinea',
4133 'PH': 'Philippines',
4137 'PR': 'Puerto Rico',
4141 'RU': 'Russian Federation',
4143 'BL': 'Saint Barthélemy',
4144 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4145 'KN': 'Saint Kitts and Nevis',
4146 'LC': 'Saint Lucia',
4147 'MF': 'Saint Martin (French part)',
4148 'PM': 'Saint Pierre and Miquelon',
4149 'VC': 'Saint Vincent and the Grenadines',
4152 'ST': 'Sao Tome and Principe',
4153 'SA': 'Saudi Arabia',
4157 'SL': 'Sierra Leone',
4159 'SX': 'Sint Maarten (Dutch part)',
4162 'SB': 'Solomon Islands',
4164 'ZA': 'South Africa',
4165 'GS': 'South Georgia and the South Sandwich Islands',
4166 'SS': 'South Sudan',
4171 'SJ': 'Svalbard and Jan Mayen',
4174 'CH': 'Switzerland',
4175 'SY': 'Syrian Arab Republic',
4176 'TW': 'Taiwan, Province of China',
4178 'TZ': 'Tanzania, United Republic of',
4180 'TL': 'Timor-Leste',
4184 'TT': 'Trinidad and Tobago',
4187 'TM': 'Turkmenistan',
4188 'TC': 'Turks and Caicos Islands',
4192 'AE': 'United Arab Emirates',
4193 'GB': 'United Kingdom',
4194 'US': 'United States',
4195 'UM': 'United States Minor Outlying Islands',
4199 'VE': 'Venezuela, Bolivarian Republic of',
4201 'VG': 'Virgin Islands, British',
4202 'VI': 'Virgin Islands, U.S.',
4203 'WF': 'Wallis and Futuna',
4204 'EH': 'Western Sahara',
4208 # Not ISO 3166 codes, but used for IP blocks
4209 'AP': 'Asia/Pacific Region',
4214 def short2full(cls, code):
4215 """Convert an ISO 3166-2 country code to the corresponding full name"""
4216 return cls._country_map.get(code.upper())
4220 # Major IPv4 address blocks per country
4222 'AD': '46.172.224.0/19',
4223 'AE': '94.200.0.0/13',
4224 'AF': '149.54.0.0/17',
4225 'AG': '209.59.64.0/18',
4226 'AI': '204.14.248.0/21',
4227 'AL': '46.99.0.0/16',
4228 'AM': '46.70.0.0/15',
4229 'AO': '105.168.0.0/13',
4230 'AP': '182.50.184.0/21',
4231 'AQ': '23.154.160.0/24',
4232 'AR': '181.0.0.0/12',
4233 'AS': '202.70.112.0/20',
4234 'AT': '77.116.0.0/14',
4235 'AU': '1.128.0.0/11',
4236 'AW': '181.41.0.0/18',
4237 'AX': '185.217.4.0/22',
4238 'AZ': '5.197.0.0/16',
4239 'BA': '31.176.128.0/17',
4240 'BB': '65.48.128.0/17',
4241 'BD': '114.130.0.0/16',
4243 'BF': '102.178.0.0/15',
4244 'BG': '95.42.0.0/15',
4245 'BH': '37.131.0.0/17',
4246 'BI': '154.117.192.0/18',
4247 'BJ': '137.255.0.0/16',
4248 'BL': '185.212.72.0/23',
4249 'BM': '196.12.64.0/18',
4250 'BN': '156.31.0.0/16',
4251 'BO': '161.56.0.0/16',
4252 'BQ': '161.0.80.0/20',
4253 'BR': '191.128.0.0/12',
4254 'BS': '24.51.64.0/18',
4255 'BT': '119.2.96.0/19',
4256 'BW': '168.167.0.0/16',
4257 'BY': '178.120.0.0/13',
4258 'BZ': '179.42.192.0/18',
4259 'CA': '99.224.0.0/11',
4260 'CD': '41.243.0.0/16',
4261 'CF': '197.242.176.0/21',
4262 'CG': '160.113.0.0/16',
4263 'CH': '85.0.0.0/13',
4264 'CI': '102.136.0.0/14',
4265 'CK': '202.65.32.0/19',
4266 'CL': '152.172.0.0/14',
4267 'CM': '102.244.0.0/14',
4268 'CN': '36.128.0.0/10',
4269 'CO': '181.240.0.0/12',
4270 'CR': '201.192.0.0/12',
4271 'CU': '152.206.0.0/15',
4272 'CV': '165.90.96.0/19',
4273 'CW': '190.88.128.0/17',
4274 'CY': '31.153.0.0/16',
4275 'CZ': '88.100.0.0/14',
4277 'DJ': '197.241.0.0/17',
4278 'DK': '87.48.0.0/12',
4279 'DM': '192.243.48.0/20',
4280 'DO': '152.166.0.0/15',
4281 'DZ': '41.96.0.0/12',
4282 'EC': '186.68.0.0/15',
4283 'EE': '90.190.0.0/15',
4284 'EG': '156.160.0.0/11',
4285 'ER': '196.200.96.0/20',
4286 'ES': '88.0.0.0/11',
4287 'ET': '196.188.0.0/14',
4288 'EU': '2.16.0.0/13',
4289 'FI': '91.152.0.0/13',
4290 'FJ': '144.120.0.0/16',
4291 'FK': '80.73.208.0/21',
4292 'FM': '119.252.112.0/20',
4293 'FO': '88.85.32.0/19',
4295 'GA': '41.158.0.0/15',
4297 'GD': '74.122.88.0/21',
4298 'GE': '31.146.0.0/16',
4299 'GF': '161.22.64.0/18',
4300 'GG': '62.68.160.0/19',
4301 'GH': '154.160.0.0/12',
4302 'GI': '95.164.0.0/16',
4303 'GL': '88.83.0.0/19',
4304 'GM': '160.182.0.0/15',
4305 'GN': '197.149.192.0/18',
4306 'GP': '104.250.0.0/19',
4307 'GQ': '105.235.224.0/20',
4308 'GR': '94.64.0.0/13',
4309 'GT': '168.234.0.0/16',
4310 'GU': '168.123.0.0/16',
4311 'GW': '197.214.80.0/20',
4312 'GY': '181.41.64.0/18',
4313 'HK': '113.252.0.0/14',
4314 'HN': '181.210.0.0/16',
4315 'HR': '93.136.0.0/13',
4316 'HT': '148.102.128.0/17',
4317 'HU': '84.0.0.0/14',
4318 'ID': '39.192.0.0/10',
4319 'IE': '87.32.0.0/12',
4320 'IL': '79.176.0.0/13',
4321 'IM': '5.62.80.0/20',
4322 'IN': '117.192.0.0/10',
4323 'IO': '203.83.48.0/21',
4324 'IQ': '37.236.0.0/14',
4325 'IR': '2.176.0.0/12',
4326 'IS': '82.221.0.0/16',
4327 'IT': '79.0.0.0/10',
4328 'JE': '87.244.64.0/18',
4329 'JM': '72.27.0.0/17',
4330 'JO': '176.29.0.0/16',
4331 'JP': '133.0.0.0/8',
4332 'KE': '105.48.0.0/12',
4333 'KG': '158.181.128.0/17',
4334 'KH': '36.37.128.0/17',
4335 'KI': '103.25.140.0/22',
4336 'KM': '197.255.224.0/20',
4337 'KN': '198.167.192.0/19',
4338 'KP': '175.45.176.0/22',
4339 'KR': '175.192.0.0/10',
4340 'KW': '37.36.0.0/14',
4341 'KY': '64.96.0.0/15',
4342 'KZ': '2.72.0.0/13',
4343 'LA': '115.84.64.0/18',
4344 'LB': '178.135.0.0/16',
4345 'LC': '24.92.144.0/20',
4346 'LI': '82.117.0.0/19',
4347 'LK': '112.134.0.0/15',
4348 'LR': '102.183.0.0/16',
4349 'LS': '129.232.0.0/17',
4350 'LT': '78.56.0.0/13',
4351 'LU': '188.42.0.0/16',
4352 'LV': '46.109.0.0/16',
4353 'LY': '41.252.0.0/14',
4354 'MA': '105.128.0.0/11',
4355 'MC': '88.209.64.0/18',
4356 'MD': '37.246.0.0/16',
4357 'ME': '178.175.0.0/17',
4358 'MF': '74.112.232.0/21',
4359 'MG': '154.126.0.0/17',
4360 'MH': '117.103.88.0/21',
4361 'MK': '77.28.0.0/15',
4362 'ML': '154.118.128.0/18',
4363 'MM': '37.111.0.0/17',
4364 'MN': '49.0.128.0/17',
4365 'MO': '60.246.0.0/16',
4366 'MP': '202.88.64.0/20',
4367 'MQ': '109.203.224.0/19',
4368 'MR': '41.188.64.0/18',
4369 'MS': '208.90.112.0/22',
4370 'MT': '46.11.0.0/16',
4371 'MU': '105.16.0.0/12',
4372 'MV': '27.114.128.0/18',
4373 'MW': '102.70.0.0/15',
4374 'MX': '187.192.0.0/11',
4375 'MY': '175.136.0.0/13',
4376 'MZ': '197.218.0.0/15',
4377 'NA': '41.182.0.0/16',
4378 'NC': '101.101.0.0/18',
4379 'NE': '197.214.0.0/18',
4380 'NF': '203.17.240.0/22',
4381 'NG': '105.112.0.0/12',
4382 'NI': '186.76.0.0/15',
4383 'NL': '145.96.0.0/11',
4384 'NO': '84.208.0.0/13',
4385 'NP': '36.252.0.0/15',
4386 'NR': '203.98.224.0/19',
4387 'NU': '49.156.48.0/22',
4388 'NZ': '49.224.0.0/14',
4389 'OM': '5.36.0.0/15',
4390 'PA': '186.72.0.0/15',
4391 'PE': '186.160.0.0/14',
4392 'PF': '123.50.64.0/18',
4393 'PG': '124.240.192.0/19',
4394 'PH': '49.144.0.0/13',
4395 'PK': '39.32.0.0/11',
4396 'PL': '83.0.0.0/11',
4397 'PM': '70.36.0.0/20',
4398 'PR': '66.50.0.0/16',
4399 'PS': '188.161.0.0/16',
4400 'PT': '85.240.0.0/13',
4401 'PW': '202.124.224.0/20',
4402 'PY': '181.120.0.0/14',
4403 'QA': '37.210.0.0/15',
4404 'RE': '102.35.0.0/16',
4405 'RO': '79.112.0.0/13',
4406 'RS': '93.86.0.0/15',
4407 'RU': '5.136.0.0/13',
4408 'RW': '41.186.0.0/16',
4409 'SA': '188.48.0.0/13',
4410 'SB': '202.1.160.0/19',
4411 'SC': '154.192.0.0/11',
4412 'SD': '102.120.0.0/13',
4413 'SE': '78.64.0.0/12',
4414 'SG': '8.128.0.0/10',
4415 'SI': '188.196.0.0/14',
4416 'SK': '78.98.0.0/15',
4417 'SL': '102.143.0.0/17',
4418 'SM': '89.186.32.0/19',
4419 'SN': '41.82.0.0/15',
4420 'SO': '154.115.192.0/18',
4421 'SR': '186.179.128.0/17',
4422 'SS': '105.235.208.0/21',
4423 'ST': '197.159.160.0/19',
4424 'SV': '168.243.0.0/16',
4425 'SX': '190.102.0.0/20',
4427 'SZ': '41.84.224.0/19',
4428 'TC': '65.255.48.0/20',
4429 'TD': '154.68.128.0/19',
4430 'TG': '196.168.0.0/14',
4431 'TH': '171.96.0.0/13',
4432 'TJ': '85.9.128.0/18',
4433 'TK': '27.96.24.0/21',
4434 'TL': '180.189.160.0/20',
4435 'TM': '95.85.96.0/19',
4436 'TN': '197.0.0.0/11',
4437 'TO': '175.176.144.0/21',
4438 'TR': '78.160.0.0/11',
4439 'TT': '186.44.0.0/15',
4440 'TV': '202.2.96.0/19',
4441 'TW': '120.96.0.0/11',
4442 'TZ': '156.156.0.0/14',
4443 'UA': '37.52.0.0/14',
4444 'UG': '102.80.0.0/13',
4446 'UY': '167.56.0.0/13',
4447 'UZ': '84.54.64.0/18',
4448 'VA': '212.77.0.0/19',
4449 'VC': '207.191.240.0/21',
4450 'VE': '186.88.0.0/13',
4451 'VG': '66.81.192.0/20',
4452 'VI': '146.226.0.0/16',
4453 'VN': '14.160.0.0/11',
4454 'VU': '202.80.32.0/20',
4455 'WF': '117.20.32.0/21',
4456 'WS': '202.4.32.0/19',
4457 'YE': '134.35.0.0/16',
4458 'YT': '41.242.116.0/22',
4459 'ZA': '41.0.0.0/11',
4460 'ZM': '102.144.0.0/13',
4461 'ZW': '102.177.192.0/18',
4465 def random_ipv4(cls, code_or_block):
4466 if len(code_or_block) == 2:
4467 block = cls._country_ip_map.get(code_or_block.upper())
4471 block = code_or_block
4472 addr, preflen = block.split('/')
4473 addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
4474 addr_max = addr_min | (0xffffffff >> int(preflen))
4475 return compat_str(socket.inet_ntoa(
4476 compat_struct_pack('!L', random.randint(addr_min, addr_max))))
4479 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
4480 def __init__(self, proxies=None):
4481 # Set default handlers
4482 for type in ('http', 'https'):
4483 setattr(self, '%s_open' % type,
4484 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4485 meth(r, proxy, type))
4486 compat_urllib_request.ProxyHandler.__init__(self, proxies)
4488 def proxy_open(self, req, proxy, type):
4489 req_proxy = req.headers.get('Ytdl-request-proxy')
4490 if req_proxy is not None:
4492 del req.headers['Ytdl-request-proxy']
4494 if proxy == '__noproxy__':
4495 return None # No Proxy
4496 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4497 req.add_header('Ytdl-socks-proxy', proxy)
4498 # yt-dlp's http/https handlers do wrapping the socket with socks
4500 return compat_urllib_request.ProxyHandler.proxy_open(
4501 self, req, proxy, type)
4504 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4505 # released into Public Domain
4506 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4508 def long_to_bytes(n, blocksize=0):
4509 """long_to_bytes(n:long, blocksize:int) : string
4510 Convert a long integer to a byte string.
4512 If optional blocksize is given and greater than zero, pad the front of the
4513 byte string with binary zeros so that the length is a multiple of
4516 # after much testing, this algorithm was deemed to be the fastest
4520 s = compat_struct_pack('>I', n & 0xffffffff) + s
4522 # strip off leading zeros
4523 for i in range(len(s)):
4524 if s[i] != b'\000'[0]:
4527 # only happens when n == 0
4531 # add back some pad bytes. this could be done more efficiently w.r.t. the
4532 # de-padding being done above, but sigh...
4533 if blocksize > 0 and len(s) % blocksize:
4534 s = (blocksize - len(s) % blocksize) * b'\000' + s
4538 def bytes_to_long(s):
4539 """bytes_to_long(string) : long
4540 Convert a byte string to a long integer.
4542 This is (essentially) the inverse of long_to_bytes().
4547 extra = (4 - length % 4)
4548 s = b'\000' * extra + s
4549 length = length + extra
4550 for i in range(0, length, 4):
4551 acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
4555 def ohdave_rsa_encrypt(data, exponent, modulus):
4557 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4560 data: data to encrypt, bytes-like object
4561 exponent, modulus: parameter e and N of RSA algorithm, both integer
4562 Output: hex string of encrypted data
4564 Limitation: supports one block encryption only
4567 payload = int(binascii.hexlify(data[::-1]), 16)
4568 encrypted = pow(payload, exponent, modulus)
4569 return '%x' % encrypted
4572 def pkcs1pad(data, length):
4574 Padding input data with PKCS#1 scheme
4576 @param {int[]} data input data
4577 @param {int} length target length
4578 @returns {int[]} padded data
4580 if len(data) > length - 11:
4581 raise ValueError('Input data too
long for PKCS
#1 padding')
4583 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4584 return [0, 2] + pseudo_random
+ [0] + data
4587 def encode_base_n(num
, n
, table
=None):
4588 FULL_TABLE
= '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
4590 table
= FULL_TABLE
[:n
]
4593 raise ValueError('base %d exceeds table length %d' % (n
, len(table
)))
4600 ret
= table
[num
% n
] + ret
4605 def decode_packed_codes(code
):
4606 mobj
= re
.search(PACKED_CODES_RE
, code
)
4607 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4610 symbols
= symbols
.split('|')
4615 base_n_count
= encode_base_n(count
, base
)
4616 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4619 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4623 def caesar(s
, alphabet
, shift
):
4628 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4633 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4636 def parse_m3u8_attributes(attrib
):
4638 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4639 if val
.startswith('"'):
4645 def urshift(val
, n
):
4646 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4649 # Based on png2str() written by @gdkchan and improved by @yokrysty
4650 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4651 def decode_png(png_data
):
4652 # Reference: https://www.w3.org/TR/PNG/
4653 header
= png_data
[8:]
4655 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4656 raise OSError('Not a valid PNG file.')
4658 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4659 unpack_integer
= lambda x
: compat_struct_unpack(int_map
[len(x
)], x
)[0]
4664 length
= unpack_integer(header
[:4])
4667 chunk_type
= header
[:4]
4670 chunk_data
= header
[:length
]
4671 header
= header
[length
:]
4673 header
= header
[4:] # Skip CRC
4681 ihdr
= chunks
[0]['data']
4683 width
= unpack_integer(ihdr
[:4])
4684 height
= unpack_integer(ihdr
[4:8])
4688 for chunk
in chunks
:
4689 if chunk
['type'] == b
'IDAT':
4690 idat
+= chunk
['data']
4693 raise OSError('Unable to read PNG data.')
4695 decompressed_data
= bytearray(zlib
.decompress(idat
))
4700 def _get_pixel(idx
):
4705 for y
in range(height
):
4706 basePos
= y
* (1 + stride
)
4707 filter_type
= decompressed_data
[basePos
]
4711 pixels
.append(current_row
)
4713 for x
in range(stride
):
4714 color
= decompressed_data
[1 + basePos
+ x
]
4715 basex
= y
* stride
+ x
4720 left
= _get_pixel(basex
- 3)
4722 up
= _get_pixel(basex
- stride
)
4724 if filter_type
== 1: # Sub
4725 color
= (color
+ left
) & 0xff
4726 elif filter_type
== 2: # Up
4727 color
= (color
+ up
) & 0xff
4728 elif filter_type
== 3: # Average
4729 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4730 elif filter_type
== 4: # Paeth
4736 c
= _get_pixel(basex
- stride
- 3)
4744 if pa
<= pb
and pa
<= pc
:
4745 color
= (color
+ a
) & 0xff
4747 color
= (color
+ b
) & 0xff
4749 color
= (color
+ c
) & 0xff
4751 current_row
.append(color
)
4753 return width
, height
, pixels
4756 def write_xattr(path
, key
, value
):
4757 # Windows: Write xattrs to NTFS Alternate Data Streams:
4758 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4759 if compat_os_name
== 'nt':
4760 assert ':' not in key
4761 assert os
.path
.exists(path
)
4764 with open(f
'{path}:{key}', 'wb') as f
:
4766 except OSError as e
:
4767 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4770 # UNIX Method 1. Use xattrs/pyxattrs modules
4771 from .dependencies
import xattr
4774 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
4775 # Unicode arguments are not supported in pyxattr until version 0.5.0
4776 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4777 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
4778 setxattr
= xattr
.set
4780 setxattr
= xattr
.setxattr
4784 setxattr(path
, key
, value
)
4785 except OSError as e
:
4786 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4789 # UNIX Method 2. Use setfattr/xattr executables
4790 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
4791 else 'xattr' if check_executable('xattr', ['-h']) else None)
4793 raise XAttrUnavailableError(
4794 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
4795 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
4797 value
= value
.decode()
4800 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
4801 stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4802 except OSError as e
:
4803 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4804 stderr
= p
.communicate_or_kill()[1].decode('utf-8', 'replace')
4806 raise XAttrMetadataError(p
.returncode
, stderr
)
4809 def random_birthday(year_field
, month_field
, day_field
):
4810 start_date
= datetime
.date(1950, 1, 1)
4811 end_date
= datetime
.date(1995, 12, 31)
4812 offset
= random
.randint(0, (end_date
- start_date
).days
)
4813 random_date
= start_date
+ datetime
.timedelta(offset
)
4815 year_field
: str(random_date
.year
),
4816 month_field
: str(random_date
.month
),
4817 day_field
: str(random_date
.day
),
4821 # Templates for internet shortcut files, which are plain text files.
4822 DOT_URL_LINK_TEMPLATE
= '''\
4827 DOT_WEBLOC_LINK_TEMPLATE
= '''\
4828 <?xml version="1.0" encoding="UTF-8"?>
4829 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
4830 <plist version="1.0">
4833 \t<string>%(url)s</string>
4838 DOT_DESKTOP_LINK_TEMPLATE
= '''\
4848 'url': DOT_URL_LINK_TEMPLATE
,
4849 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
4850 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
4854 def iri_to_uri(iri
):
4856 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
4858 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.
4861 iri_parts
= compat_urllib_parse_urlparse(iri
)
4863 if '[' in iri_parts
.netloc
:
4864 raise ValueError('IPv6 URIs are not, yet, supported.')
4865 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
4867 # 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.
4870 if iri_parts
.username
:
4871 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
4872 if iri_parts
.password
is not None:
4873 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
4876 net_location
+= iri_parts
.hostname
.encode('idna').decode() # Punycode for Unicode hostnames.
4877 # The 'idna' encoding produces ASCII text.
4878 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
4879 net_location
+= ':' + str(iri_parts
.port
)
4881 return urllib
.parse
.urlunparse(
4885 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
4887 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
4888 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
4890 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
4891 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
4893 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
4895 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
4898 def to_high_limit_path(path
):
4899 if sys
.platform
in ['win32', 'cygwin']:
4900 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
4901 return '\\\\?\\' + os
.path
.abspath(path
)
4906 def format_field(obj
, field
=None, template
='%s', ignore
=(None, ''), default
='', func
=None):
4907 val
= traverse_obj(obj
, *variadic(field
))
4910 return template
% (func(val
) if func
else val
)
4913 def clean_podcast_url(url
):
4914 return re
.sub(r
'''(?x)
4918 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
4921 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
4924 cn\.co| # https://podcorn.com/analytics-prefix/
4925 st\.fm # https://podsights.com/docs/
4930 _HEX_TABLE
= '0123456789abcdef'
4933 def random_uuidv4():
4934 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
4937 def make_dir(path
, to_screen
=None):
4939 dn
= os
.path
.dirname(path
)
4940 if dn
and not os
.path
.exists(dn
):
4943 except OSError as err
:
4944 if callable(to_screen
) is not None:
4945 to_screen('unable to create directory ' + error_to_compat_str(err
))
4949 def get_executable_path():
4950 from .update
import _get_variant_and_executable_path
4952 return os
.path
.dirname(os
.path
.abspath(_get_variant_and_executable_path()[1]))
4955 def load_plugins(name
, suffix
, namespace
):
4957 with contextlib
.suppress(FileNotFoundError
):
4958 plugins_spec
= importlib
.util
.spec_from_file_location(
4959 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
4960 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
4961 sys
.modules
[plugins_spec
.name
] = plugins
4962 plugins_spec
.loader
.exec_module(plugins
)
4963 for name
in dir(plugins
):
4964 if name
in namespace
:
4966 if not name
.endswith(suffix
):
4968 klass
= getattr(plugins
, name
)
4969 classes
[name
] = namespace
[name
] = klass
4974 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
4975 casesense
=True, is_user_input
=False, traverse_string
=False):
4976 ''' Traverse nested list/dict/tuple
4977 @param path_list A list of paths which are checked one by one.
4978 Each path is a list of keys where each key is a:
4980 - string: A dictionary key
4981 - int: An index into a list
4982 - tuple: A list of keys all of which will be traversed
4983 - Ellipsis: Fetch all values in the object
4984 - Function: Takes the key and value as arguments
4985 and returns whether the key matches or not
4986 @param default Default value to return
4987 @param expected_type Only accept final value of this type (Can also be any callable)
4988 @param get_all Return all the values obtained from a path or only the first one
4989 @param casesense Whether to consider dictionary keys as case sensitive
4990 @param is_user_input Whether the keys are generated from user input. If True,
4991 strings are converted to int/slice if necessary
4992 @param traverse_string Whether to traverse inside strings. If True, any
4993 non-compatible object will also be converted into a string
4997 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
4998 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
5000 def _traverse_obj(obj
, path
, _current_depth
=0):
5002 path
= tuple(variadic(path
))
5003 for i
, key
in enumerate(path
):
5004 if None in (key
, obj
):
5006 if isinstance(key
, (list, tuple)):
5007 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
5010 obj
= (obj
.values() if isinstance(obj
, dict)
5011 else obj
if isinstance(obj
, (list, tuple, LazyList
))
5012 else str(obj
) if traverse_string
else [])
5014 depth
= max(depth
, _current_depth
)
5015 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
5017 if isinstance(obj
, (list, tuple, LazyList
)):
5018 obj
= enumerate(obj
)
5019 elif isinstance(obj
, dict):
5022 if not traverse_string
:
5026 depth
= max(depth
, _current_depth
)
5027 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if try_call(key
, args
=(k
, v
))]
5028 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
5029 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
5030 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
5033 key
= (int_or_none(key
) if ':' not in key
5034 else slice(*map(int_or_none
, key
.split(':'))))
5035 if key
== slice(None):
5036 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
5037 if not isinstance(key
, (int, slice)):
5039 if not isinstance(obj
, (list, tuple, LazyList
)):
5040 if not traverse_string
:
5049 if isinstance(expected_type
, type):
5050 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5051 elif expected_type
is not None:
5052 type_test
= expected_type
5054 type_test
= lambda val
: val
5056 for path
in path_list
:
5058 val
= _traverse_obj(obj
, path
)
5061 for _
in range(depth
- 1):
5062 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5063 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5065 return val
if get_all
else val
[0]
5067 val
= type_test(val
)
5073 def traverse_dict(dictn
, keys
, casesense
=True):
5074 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5075 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5076 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5079 def get_first(obj
, keys
, **kwargs
):
5080 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5083 def variadic(x
, allowed_types
=(str, bytes, dict)):
5084 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5087 def decode_base(value
, digits
):
5088 # This will convert given base-x string to scalar (long or int)
5089 table
= {char: index for index, char in enumerate(digits)}
5094 result
+= table
[chr]
5098 def time_seconds(**kwargs
):
5099 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5100 return t
.timestamp()
5103 # create a JSON Web Signature (jws) with HS256 algorithm
5104 # the resulting format is in JWS Compact Serialization
5105 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5106 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5107 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5113 header_data
.update(headers
)
5114 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode())
5115 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode())
5116 h
= hmac
.new(key
.encode(), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5117 signature_b64
= base64
.b64encode(h
.digest())
5118 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5122 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5123 def jwt_decode_hs256(jwt
):
5124 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5125 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5129 WINDOWS_VT_MODE
= False if compat_os_name
== 'nt' else None
5133 def supports_terminal_sequences(stream
):
5134 if compat_os_name
== 'nt':
5135 if not WINDOWS_VT_MODE
:
5137 elif not os
.getenv('TERM'):
5140 return stream
.isatty()
5141 except BaseException
:
5145 def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
5146 if get_windows_version() < (10, 0, 10586):
5148 global WINDOWS_VT_MODE
5149 startupinfo
= subprocess
.STARTUPINFO()
5150 startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
5152 subprocess
.Popen('', shell
=True, startupinfo
=startupinfo
).wait()
5156 WINDOWS_VT_MODE
= True
5157 supports_terminal_sequences
.cache_clear()
5160 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5163 def remove_terminal_sequences(string
):
5164 return _terminal_sequences_re
.sub('', string
)
5167 def number_of_digits(number
):
5168 return len('%d' % number
)
5171 def join_nonempty(*values
, delim
='-', from_dict
=None):
5172 if from_dict
is not None:
5173 values
= map(from_dict
.get
, values
)
5174 return delim
.join(map(str, filter(None, values
)))
5177 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5179 Find the largest format dimensions in terms of video width and, for each thumbnail:
5180 * Modify the URL: Match the width with the provided regex and replace with the former width
5183 This function is useful with video services that scale the provided thumbnails on demand
5185 _keys
= ('width', 'height')
5186 max_dimensions
= max(
5187 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5189 if not max_dimensions
[0]:
5193 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5194 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5195 for thumbnail
in thumbnails
5199 def parse_http_range(range):
5200 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5202 return None, None, None
5203 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5205 return None, None, None
5206 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5209 def read_stdin(what
):
5210 eof
= 'Ctrl+Z' if compat_os_name
== 'nt' else 'Ctrl+D'
5211 write_string(f
'Reading {what} from STDIN - EOF ({eof}) to end:\n')
5219 __initialized
= False
5221 def __init__(self
, parser
, label
=None):
5222 self
.parser
, self
.label
= parser
, label
5223 self
._loaded
_paths
, self
.configs
= set(), []
5225 def init(self
, args
=None, filename
=None):
5226 assert not self
.__initialized
5229 location
= os
.path
.realpath(filename
)
5230 directory
= os
.path
.dirname(location
)
5231 if location
in self
._loaded
_paths
:
5233 self
._loaded
_paths
.add(location
)
5235 self
.own_args
, self
.__initialized
= args
, True
5236 opts
, _
= self
.parser
.parse_known_args(args
)
5237 self
.parsed_args
, self
.filename
= args
, filename
5239 for location
in opts
.config_locations
or []:
5241 self
.append_config(shlex
.split(read_stdin('options'), comments
=True), label
='stdin')
5243 location
= os
.path
.join(directory
, expand_path(location
))
5244 if os
.path
.isdir(location
):
5245 location
= os
.path
.join(location
, 'yt-dlp.conf')
5246 if not os
.path
.exists(location
):
5247 self
.parser
.error(f
'config location {location} does not exist')
5248 self
.append_config(self
.read_file(location
), location
)
5252 label
= join_nonempty(
5253 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5255 return join_nonempty(
5256 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5257 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5261 def read_file(filename
, default
=[]):
5263 optionf
= open(filename
)
5265 return default
# silently skip if file is not present
5267 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5268 contents
= optionf
.read()
5269 res
= shlex
.split(contents
, comments
=True)
5275 def hide_login_info(opts
):
5276 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5277 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5282 return m
.group('key') + '=PRIVATE'
5286 opts
= list(map(_scrub_eq
, opts
))
5287 for idx
, opt
in enumerate(opts
):
5288 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5289 opts
[idx
+ 1] = 'PRIVATE'
5292 def append_config(self
, *args
, label
=None):
5293 config
= type(self
)(self
.parser
, label
)
5294 config
._loaded
_paths
= self
._loaded
_paths
5295 if config
.init(*args
):
5296 self
.configs
.append(config
)
5300 for config
in reversed(self
.configs
):
5301 yield from config
.all_args
5302 yield from self
.parsed_args
or []
5304 def parse_known_args(self
, **kwargs
):
5305 return self
.parser
.parse_known_args(self
.all_args
, **kwargs
)
5307 def parse_args(self
):
5308 return self
.parser
.parse_args(self
.all_args
)
5311 class WebSocketsWrapper():
5312 """Wraps websockets module to use in non-async scopes"""
5315 def __init__(self
, url
, headers
=None, connect
=True):
5316 self
.loop
= asyncio
.new_event_loop()
5317 # XXX: "loop" is deprecated
5318 self
.conn
= websockets
.connect(
5319 url
, extra_headers
=headers
, ping_interval
=None,
5320 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5323 atexit
.register(self
.__exit
__, None, None, None)
5325 def __enter__(self
):
5327 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5330 def send(self
, *args
):
5331 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5333 def recv(self
, *args
):
5334 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5336 def __exit__(self
, type, value
, traceback
):
5338 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5341 self
._cancel
_all
_tasks
(self
.loop
)
5343 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5344 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5346 def run_with_loop(main
, loop
):
5347 if not asyncio
.iscoroutine(main
):
5348 raise ValueError(f
'a coroutine was expected, got {main!r}')
5351 return loop
.run_until_complete(main
)
5353 loop
.run_until_complete(loop
.shutdown_asyncgens())
5354 if hasattr(loop
, 'shutdown_default_executor'):
5355 loop
.run_until_complete(loop
.shutdown_default_executor())
5358 def _cancel_all_tasks(loop
):
5359 to_cancel
= asyncio
.all_tasks(loop
)
5364 for task
in to_cancel
:
5367 # XXX: "loop" is removed in python 3.10+
5368 loop
.run_until_complete(
5369 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5371 for task
in to_cancel
:
5372 if task
.cancelled():
5374 if task
.exception() is not None:
5375 loop
.call_exception_handler({
5376 'message': 'unhandled exception during asyncio.run() shutdown',
5377 'exception': task
.exception(),
5382 def merge_headers(*dicts
):
5383 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5384 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5387 class classproperty
:
5388 """classmethod(property(func)) that works in py < 3.9"""
5390 def __init__(self
, func
):
5391 functools
.update_wrapper(self
, func
)
5394 def __get__(self
, _
, cls
):
5395 return self
.func(cls
)
5399 """Immutable namespace"""
5401 def __init__(self
, **kwargs
):
5404 def __getattr__(self
, attr
):
5405 return self
._dict
[attr
]
5407 def __contains__(self
, item
):
5408 return item
in self
._dict
.values()
5411 return iter(self
._dict
.items())
5414 return f
'{type(self).__name__}({", ".join(f"{k}={v}" for k, v in self)})'
5418 has_certifi
= bool(certifi
)
5419 has_websockets
= bool(websockets
)