39 import xml
.etree
.ElementTree
46 compat_etree_fromstring
,
49 compat_html_entities_html5
,
50 compat_HTMLParseError
,
61 compat_urllib_parse_unquote_plus
,
62 compat_urllib_parse_urlencode
,
63 compat_urllib_parse_urlparse
,
64 compat_urllib_request
,
67 from .dependencies
import brotli
, certifi
, websockets
68 from .socks
import ProxyType
, sockssocket
71 def register_socks_protocols():
72 # "Register" SOCKS protocols
73 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
74 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
75 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
76 if scheme
not in compat_urlparse
.uses_netloc
:
77 compat_urlparse
.uses_netloc
.append(scheme
)
80 # This is not clearly defined otherwise
81 compiled_regex_type
= type(re
.compile(''))
84 def random_user_agent():
85 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
126 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
129 SUPPORTED_ENCODINGS
= [
133 SUPPORTED_ENCODINGS
.append('br')
136 'User-Agent': random_user_agent(),
137 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
138 'Accept-Language': 'en-us,en;q=0.5',
139 'Sec-Fetch-Mode': 'navigate',
144 'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
148 NO_DEFAULT
= object()
150 ENGLISH_MONTH_NAMES
= [
151 'January', 'February', 'March', 'April', 'May', 'June',
152 'July', 'August', 'September', 'October', 'November', 'December']
155 'en': ENGLISH_MONTH_NAMES
,
157 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
158 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
162 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
163 'flv', 'f4v', 'f4a', 'f4b',
164 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
165 'mkv', 'mka', 'mk3d',
174 'f4f', 'f4m', 'm3u8', 'smil')
176 # needed for sanitizing filenames in restricted mode
177 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
178 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
179 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
209 '%Y-%m-%d %H:%M:%S.%f',
210 '%Y-%m-%d %H:%M:%S:%f',
213 '%Y-%m-%dT%H:%M:%SZ',
214 '%Y-%m-%dT%H:%M:%S.%fZ',
215 '%Y-%m-%dT%H:%M:%S.%f0Z',
217 '%Y-%m-%dT%H:%M:%S.%f',
220 '%b %d %Y at %H:%M:%S',
222 '%B %d %Y at %H:%M:%S',
226 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
227 DATE_FORMATS_DAY_FIRST
.extend([
236 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
237 DATE_FORMATS_MONTH_FIRST
.extend([
245 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
246 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>(?P
<json_ld
>.+?
)</script
>'
248 NUMBER_RE = r'\d
+(?
:\
.\d
+)?
'
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
):
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
)
370 def get_element_html_by_id(id, html
):
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
)
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
, escape_value
=True):
388 retval
= get_elements_by_attribute(attribute
, value
, html
, escape_value
)
389 return retval
[0] if retval
else None
392 def get_element_html_by_attribute(attribute
, value
, html
, escape_value
=True):
393 retval
= get_elements_html_by_attribute(attribute
, value
, html
, escape_value
)
394 return retval
[0] if retval
else None
397 def get_elements_by_class(class_name
, html
):
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 def sanitize_open(filename, open_mode):
598 """Try to
open the given filename
, and slightly tweak it
if this fails
.
600 Attempts to
open the given filename
. If this fails
, it tries to change
601 the filename slightly
, step by step
, until it
's either able to open it
602 or it fails and raises a final exception, like the standard open()
605 It returns the tuple (stream, definitive_file_name).
608 if sys.platform == 'win32
':
610 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
611 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
613 for attempt in range(2):
616 if sys.platform == 'win32
':
617 # FIXME: An exclusive lock also locks the file from being read.
618 # Since windows locks are mandatory, don't lock the
file on
windows (for now
).
619 # Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
620 raise LockingUnsupportedError()
621 stream
= locked_file(filename
, open_mode
, block
=False).__enter
__()
622 except LockingUnsupportedError
:
623 stream
= open(filename
, open_mode
)
624 return (stream
, filename
)
625 except OSError as err
:
626 if attempt
or err
.errno
in (errno
.EACCES
,):
628 old_filename
, filename
= filename
, sanitize_path(filename
)
629 if old_filename
== filename
:
633 def timeconvert(timestr
):
634 """Convert RFC 2822 defined time string into system timestamp"""
636 timetuple
= email
.utils
.parsedate_tz(timestr
)
637 if timetuple
is not None:
638 timestamp
= email
.utils
.mktime_tz(timetuple
)
642 def sanitize_filename(s
, restricted
=False, is_id
=NO_DEFAULT
):
643 """Sanitizes a string so it could be used as part of a filename.
644 @param restricted Use a stricter subset of allowed characters
645 @param is_id Whether this is an ID that should be kept unchanged if possible.
646 If unset, yt-dlp's new sanitization rules are in effect
651 def replace_insane(char
):
652 if restricted
and char
in ACCENT_CHARS
:
653 return ACCENT_CHARS
[char
]
654 elif not restricted
and char
== '\n':
656 elif char
== '?' or ord(char
) < 32 or ord(char
) == 127:
659 return '' if restricted
else '\''
661 return '\0_\0-' if restricted
else '\0 \0-'
662 elif char
in '\\/|*<>':
664 if restricted
and (char
in '!&\'()[]{}$;`^,#' or char
.isspace() or ord(char
) > 127):
668 s
= re
.sub(r
'[0-9]+(?::[0-9]+)+', lambda m
: m
.group(0).replace(':', '_'), s
) # Handle timestamps
669 result
= ''.join(map(replace_insane
, s
))
670 if is_id
is NO_DEFAULT
:
671 result
= re
.sub('(\0.)(?:(?=\\1)..)+', r
'\1', result
) # Remove repeated substitute chars
672 STRIP_RE
= '(?:\0.|[ _-])*'
673 result
= re
.sub(f
'^\0.{STRIP_RE}|{STRIP_RE}\0.$', '', result
) # Remove substitute chars from start/end
674 result
= result
.replace('\0', '') or '_'
677 while '__' in result
:
678 result
= result
.replace('__', '_')
679 result
= result
.strip('_')
680 # Common case of "Foreign band name - English song title"
681 if restricted
and result
.startswith('-_'):
683 if result
.startswith('-'):
684 result
= '_' + result
[len('-'):]
685 result
= result
.lstrip('.')
691 def sanitize_path(s
, force
=False):
692 """Sanitizes and normalizes path on Windows"""
693 if sys
.platform
== 'win32':
695 drive_or_unc
, _
= os
.path
.splitdrive(s
)
701 norm_path
= os
.path
.normpath(remove_start(s
, drive_or_unc
)).split(os
.path
.sep
)
705 path_part
if path_part
in ['.', '..'] else re
.sub(r
'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part
)
706 for path_part
in norm_path
]
708 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
709 elif force
and s
and s
[0] == os
.path
.sep
:
710 sanitized_path
.insert(0, os
.path
.sep
)
711 return os
.path
.join(*sanitized_path
)
714 def sanitize_url(url
):
715 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
716 # the number of unwanted failures due to missing protocol
717 if url
.startswith('//'):
718 return 'http:%s' % url
719 # Fix some common typos seen so far
721 # https://github.com/ytdl-org/youtube-dl/issues/15649
722 (r
'^httpss://', r
'https://'),
723 # https://bx1.be/lives/direct-tv/
724 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
726 for mistake
, fixup
in COMMON_TYPOS
:
727 if re
.match(mistake
, url
):
728 return re
.sub(mistake
, fixup
, url
)
732 def extract_basic_auth(url
):
733 parts
= compat_urlparse
.urlsplit(url
)
734 if parts
.username
is None:
736 url
= compat_urlparse
.urlunsplit(parts
._replace
(netloc
=(
737 parts
.hostname
if parts
.port
is None
738 else '%s:%d' % (parts
.hostname
, parts
.port
))))
739 auth_payload
= base64
.b64encode(
740 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode('utf-8'))
741 return url
, 'Basic ' + auth_payload
.decode('utf-8')
744 def sanitized_Request(url
, *args
, **kwargs
):
745 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
746 if auth_header
is not None:
747 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
748 headers
['Authorization'] = auth_header
749 return compat_urllib_request
.Request(url
, *args
, **kwargs
)
753 """Expand shell variables and ~"""
754 return os
.path
.expandvars(compat_expanduser(s
))
757 def orderedSet(iterable
):
758 """ Remove all duplicates from the input iterable """
766 def _htmlentity_transform(entity_with_semicolon
):
767 """Transforms an HTML entity to a character."""
768 entity
= entity_with_semicolon
[:-1]
770 # Known non-numeric HTML entity
771 if entity
in compat_html_entities
.name2codepoint
:
772 return compat_chr(compat_html_entities
.name2codepoint
[entity
])
774 # TODO: HTML5 allows entities without a semicolon. For example,
775 # 'Éric' should be decoded as 'Éric'.
776 if entity_with_semicolon
in compat_html_entities_html5
:
777 return compat_html_entities_html5
[entity_with_semicolon
]
779 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
781 numstr
= mobj
.group(1)
782 if numstr
.startswith('x'):
784 numstr
= '0%s' % numstr
787 # See https://github.com/ytdl-org/youtube-dl/issues/7518
788 with contextlib
.suppress(ValueError):
789 return compat_chr(int(numstr
, base
))
791 # Unknown entity in name, return its literal representation
792 return '&%s;' % entity
798 assert isinstance(s
, str)
801 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
804 def escapeHTML(text
):
807 .replace('&', '&')
808 .replace('<', '<')
809 .replace('>', '>')
810 .replace('"', '"')
811 .replace("'", ''')
815 def process_communicate_or_kill(p
, *args
, **kwargs
):
817 return p
.communicate(*args
, **kwargs
)
818 except BaseException
: # Including KeyboardInterrupt
824 class Popen(subprocess
.Popen
):
825 if sys
.platform
== 'win32':
826 _startupinfo
= subprocess
.STARTUPINFO()
827 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
831 def __init__(self
, *args
, **kwargs
):
832 super().__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
834 def communicate_or_kill(self
, *args
, **kwargs
):
835 return process_communicate_or_kill(self
, *args
, **kwargs
)
838 def get_subprocess_encoding():
839 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
840 # For subprocess calls, encode with locale encoding
841 # Refer to http://stackoverflow.com/a/9951851/35070
842 encoding
= preferredencoding()
844 encoding
= sys
.getfilesystemencoding()
850 def encodeFilename(s
, for_subprocess
=False):
851 assert isinstance(s
, str)
855 def decodeFilename(b
, for_subprocess
=False):
859 def encodeArgument(s
):
860 # Legacy code that uses byte strings
861 # Uncomment the following line after fixing all post processors
862 # assert isinstance(s, str), 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
863 return s
if isinstance(s
, str) else s
.decode('ascii')
866 def decodeArgument(b
):
870 def decodeOption(optval
):
873 if isinstance(optval
, bytes):
874 optval
= optval
.decode(preferredencoding())
876 assert isinstance(optval
, compat_str
)
880 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
883 def timetuple_from_msec(msec
):
884 secs
, msec
= divmod(msec
, 1000)
885 mins
, secs
= divmod(secs
, 60)
886 hrs
, mins
= divmod(mins
, 60)
887 return _timetuple(hrs
, mins
, secs
, msec
)
890 def formatSeconds(secs
, delim
=':', msec
=False):
891 time
= timetuple_from_msec(secs
* 1000)
893 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
895 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
897 ret
= '%d' % time
.seconds
898 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
901 def _ssl_load_windows_store_certs(ssl_context
, storename
):
902 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
904 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
905 if encoding
== 'x509_asn' and (
906 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
907 except PermissionError
:
910 with contextlib
.suppress(ssl
.SSLError
):
911 ssl_context
.load_verify_locations(cadata
=cert
)
914 def make_HTTPS_handler(params
, **kwargs
):
915 opts_check_certificate
= not params
.get('nocheckcertificate')
916 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
917 context
.check_hostname
= opts_check_certificate
918 if params
.get('legacyserverconnect'):
919 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
920 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
921 if opts_check_certificate
:
922 if has_certifi
and 'no-certifi' not in params
.get('compat_opts', []):
923 context
.load_verify_locations(cafile
=certifi
.where())
926 context
.load_default_certs()
927 # Work around the issue in load_default_certs when there are bad certificates. See:
928 # https://github.com/yt-dlp/yt-dlp/issues/1060,
929 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
931 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
932 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
933 # Create a new context to discard any certificates that were already loaded
934 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
935 context
.check_hostname
, context
.verify_mode
= True, ssl
.CERT_REQUIRED
936 for storename
in ('CA', 'ROOT'):
937 _ssl_load_windows_store_certs(context
, storename
)
938 context
.set_default_verify_paths()
939 client_certfile
= params
.get('client_certificate')
942 context
.load_cert_chain(
943 client_certfile
, keyfile
=params
.get('client_certificate_key'),
944 password
=params
.get('client_certificate_password'))
946 raise YoutubeDLError('Unable to load client certificate')
947 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
950 def bug_reports_message(before
=';'):
951 msg
= ('please report this issue on https://github.com/yt-dlp/yt-dlp/issues?q= , '
952 'filling out the appropriate issue template. '
953 'Confirm you are on the latest version using yt-dlp -U')
955 before
= before
.rstrip()
956 if not before
or before
.endswith(('.', '!', '?')):
957 msg
= msg
[0].title() + msg
[1:]
959 return (before
+ ' ' if before
else '') + msg
962 class YoutubeDLError(Exception):
963 """Base exception for YoutubeDL errors."""
966 def __init__(self
, msg
=None):
969 elif self
.msg
is None:
970 self
.msg
= type(self
).__name
__
971 super().__init
__(self
.msg
)
974 network_exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
975 if hasattr(ssl
, 'CertificateError'):
976 network_exceptions
.append(ssl
.CertificateError
)
977 network_exceptions
= tuple(network_exceptions
)
980 class ExtractorError(YoutubeDLError
):
981 """Error during info extraction."""
983 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
984 """ tb, if given, is the original traceback (so that it can be printed out).
985 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
987 if sys
.exc_info()[0] in network_exceptions
:
990 self
.orig_msg
= str(msg
)
992 self
.expected
= expected
994 self
.video_id
= video_id
996 self
.exc_info
= sys
.exc_info() # preserve original exception
998 super().__init
__(''.join((
999 format_field(ie
, template
='[%s] '),
1000 format_field(video_id
, template
='%s: '),
1002 format_field(cause
, template
=' (caused by %r)'),
1003 '' if expected
else bug_reports_message())))
1005 def format_traceback(self
):
1006 return join_nonempty(
1007 self
.traceback
and ''.join(traceback
.format_tb(self
.traceback
)),
1008 self
.cause
and ''.join(traceback
.format_exception(None, self
.cause
, self
.cause
.__traceback
__)[1:]),
1012 class UnsupportedError(ExtractorError
):
1013 def __init__(self
, url
):
1015 'Unsupported URL: %s' % url
, expected
=True)
1019 class RegexNotFoundError(ExtractorError
):
1020 """Error when a regex didn't match"""
1024 class GeoRestrictedError(ExtractorError
):
1025 """Geographic restriction Error exception.
1027 This exception may be thrown when a video is not available from your
1028 geographic location due to geographic restrictions imposed by a website.
1031 def __init__(self
, msg
, countries
=None, **kwargs
):
1032 kwargs
['expected'] = True
1033 super().__init
__(msg
, **kwargs
)
1034 self
.countries
= countries
1037 class DownloadError(YoutubeDLError
):
1038 """Download Error exception.
1040 This exception may be thrown by FileDownloader objects if they are not
1041 configured to continue on errors. They will contain the appropriate
1045 def __init__(self
, msg
, exc_info
=None):
1046 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1047 super().__init
__(msg
)
1048 self
.exc_info
= exc_info
1051 class EntryNotInPlaylist(YoutubeDLError
):
1052 """Entry not in playlist exception.
1054 This exception will be thrown by YoutubeDL when a requested entry
1055 is not found in the playlist info_dict
1057 msg
= 'Entry not found in info'
1060 class SameFileError(YoutubeDLError
):
1061 """Same File exception.
1063 This exception will be thrown by FileDownloader objects if they detect
1064 multiple files would have to be downloaded to the same file on disk.
1066 msg
= 'Fixed output name but more than one file to download'
1068 def __init__(self
, filename
=None):
1069 if filename
is not None:
1070 self
.msg
+= f
': {filename}'
1071 super().__init
__(self
.msg
)
1074 class PostProcessingError(YoutubeDLError
):
1075 """Post Processing exception.
1077 This exception may be raised by PostProcessor's .run() method to
1078 indicate an error in the postprocessing task.
1082 class DownloadCancelled(YoutubeDLError
):
1083 """ Exception raised when the download queue should be interrupted """
1084 msg
= 'The download was cancelled'
1087 class ExistingVideoReached(DownloadCancelled
):
1088 """ --break-on-existing triggered """
1089 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1092 class RejectedVideoReached(DownloadCancelled
):
1093 """ --break-on-reject triggered """
1094 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1097 class MaxDownloadsReached(DownloadCancelled
):
1098 """ --max-downloads limit has been reached. """
1099 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1102 class ReExtractInfo(YoutubeDLError
):
1103 """ Video info needs to be re-extracted. """
1105 def __init__(self
, msg
, expected
=False):
1106 super().__init
__(msg
)
1107 self
.expected
= expected
1110 class ThrottledDownload(ReExtractInfo
):
1111 """ Download speed below --throttled-rate. """
1112 msg
= 'The download speed is below throttle limit'
1115 super().__init
__(self
.msg
, expected
=False)
1118 class UnavailableVideoError(YoutubeDLError
):
1119 """Unavailable Format exception.
1121 This exception will be thrown when a video is requested
1122 in a format that is not available for that video.
1124 msg
= 'Unable to download video'
1126 def __init__(self
, err
=None):
1128 self
.msg
+= f
': {err}'
1129 super().__init
__(self
.msg
)
1132 class ContentTooShortError(YoutubeDLError
):
1133 """Content Too Short exception.
1135 This exception may be raised by FileDownloader objects when a file they
1136 download is too small for what the server announced first, indicating
1137 the connection was probably interrupted.
1140 def __init__(self
, downloaded
, expected
):
1141 super().__init
__(f
'Downloaded {downloaded} bytes, expected {expected} bytes')
1143 self
.downloaded
= downloaded
1144 self
.expected
= expected
1147 class XAttrMetadataError(YoutubeDLError
):
1148 def __init__(self
, code
=None, msg
='Unknown error'):
1149 super().__init
__(msg
)
1153 # Parsing code and msg
1154 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1155 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1156 self
.reason
= 'NO_SPACE'
1157 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1158 self
.reason
= 'VALUE_TOO_LONG'
1160 self
.reason
= 'NOT_SUPPORTED'
1163 class XAttrUnavailableError(YoutubeDLError
):
1167 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1168 hc
= http_class(*args
, **kwargs
)
1169 source_address
= ydl_handler
._params
.get('source_address')
1171 if source_address
is not None:
1172 # This is to workaround _create_connection() from socket where it will try all
1173 # address data from getaddrinfo() including IPv6. This filters the result from
1174 # getaddrinfo() based on the source_address value.
1175 # This is based on the cpython socket.create_connection() function.
1176 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1177 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1178 host
, port
= address
1180 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1181 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1182 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1183 if addrs
and not ip_addrs
:
1184 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1186 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1187 % (ip_version
, source_address
[0]))
1188 for res
in ip_addrs
:
1189 af
, socktype
, proto
, canonname
, sa
= res
1192 sock
= socket
.socket(af
, socktype
, proto
)
1193 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1194 sock
.settimeout(timeout
)
1195 sock
.bind(source_address
)
1197 err
= None # Explicitly break reference cycle
1199 except OSError as _
:
1201 if sock
is not None:
1206 raise OSError('getaddrinfo returns an empty list')
1207 if hasattr(hc
, '_create_connection'):
1208 hc
._create
_connection
= _create_connection
1209 hc
.source_address
= (source_address
, 0)
1214 def handle_youtubedl_headers(headers
):
1215 filtered_headers
= headers
1217 if 'Youtubedl-no-compression' in filtered_headers
:
1218 filtered_headers
= {k: v for k, v in filtered_headers.items() if k.lower() != 'accept-encoding'}
1219 del filtered_headers
['Youtubedl-no-compression']
1221 return filtered_headers
1224 class YoutubeDLHandler(compat_urllib_request
.HTTPHandler
):
1225 """Handler for HTTP requests and responses.
1227 This class, when installed with an OpenerDirector, automatically adds
1228 the standard headers to every HTTP request and handles gzipped and
1229 deflated responses from web servers. If compression is to be avoided in
1230 a particular request, the original request in the program code only has
1231 to include the HTTP header "Youtubedl-no-compression", which will be
1232 removed before making the real request.
1234 Part of this code was copied from:
1236 http://techknack.net/python-urllib2-handlers/
1238 Andrew Rowls, the author of that code, agreed to release it to the
1242 def __init__(self
, params
, *args
, **kwargs
):
1243 compat_urllib_request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1244 self
._params
= params
1246 def http_open(self
, req
):
1247 conn_class
= compat_http_client
.HTTPConnection
1249 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1251 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1252 del req
.headers
['Ytdl-socks-proxy']
1254 return self
.do_open(functools
.partial(
1255 _create_http_connection
, self
, conn_class
, False),
1263 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1265 return zlib
.decompress(data
)
1271 return brotli
.decompress(data
)
1273 def http_request(self
, req
):
1274 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1275 # always respected by websites, some tend to give out URLs with non percent-encoded
1276 # non-ASCII characters (see telemb.py, ard.py [#3412])
1277 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1278 # To work around aforementioned issue we will replace request's original URL with
1279 # percent-encoded one
1280 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1281 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1282 url
= req
.get_full_url()
1283 url_escaped
= escape_url(url
)
1285 # Substitute URL if any change after escaping
1286 if url
!= url_escaped
:
1287 req
= update_Request(req
, url
=url_escaped
)
1289 for h
, v
in self
._params
.get('http_headers', std_headers
).items():
1290 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1291 # The dict keys are capitalized because of this bug by urllib
1292 if h
.capitalize() not in req
.headers
:
1293 req
.add_header(h
, v
)
1295 if 'Accept-encoding' not in req
.headers
:
1296 req
.add_header('Accept-encoding', ', '.join(SUPPORTED_ENCODINGS
))
1298 req
.headers
= handle_youtubedl_headers(req
.headers
)
1302 def http_response(self
, req
, resp
):
1305 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1306 content
= resp
.read()
1307 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1309 uncompressed
= io
.BytesIO(gz
.read())
1310 except OSError as original_ioerror
:
1311 # There may be junk add the end of the file
1312 # See http://stackoverflow.com/q/4928560/35070 for details
1313 for i
in range(1, 1024):
1315 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1316 uncompressed
= io
.BytesIO(gz
.read())
1321 raise original_ioerror
1322 resp
= compat_urllib_request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1323 resp
.msg
= old_resp
.msg
1324 del resp
.headers
['Content-encoding']
1326 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1327 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1328 resp
= compat_urllib_request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1329 resp
.msg
= old_resp
.msg
1330 del resp
.headers
['Content-encoding']
1332 if resp
.headers
.get('Content-encoding', '') == 'br':
1333 resp
= compat_urllib_request
.addinfourl(
1334 io
.BytesIO(self
.brotli(resp
.read())), old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1335 resp
.msg
= old_resp
.msg
1336 del resp
.headers
['Content-encoding']
1337 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1338 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1339 if 300 <= resp
.code
< 400:
1340 location
= resp
.headers
.get('Location')
1342 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1343 location
= location
.encode('iso-8859-1').decode('utf-8')
1344 location_escaped
= escape_url(location
)
1345 if location
!= location_escaped
:
1346 del resp
.headers
['Location']
1347 resp
.headers
['Location'] = location_escaped
1350 https_request
= http_request
1351 https_response
= http_response
1354 def make_socks_conn_class(base_class
, socks_proxy
):
1355 assert issubclass(base_class
, (
1356 compat_http_client
.HTTPConnection
, compat_http_client
.HTTPSConnection
))
1358 url_components
= compat_urlparse
.urlparse(socks_proxy
)
1359 if url_components
.scheme
.lower() == 'socks5':
1360 socks_type
= ProxyType
.SOCKS5
1361 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1362 socks_type
= ProxyType
.SOCKS4
1363 elif url_components
.scheme
.lower() == 'socks4a':
1364 socks_type
= ProxyType
.SOCKS4A
1366 def unquote_if_non_empty(s
):
1369 return compat_urllib_parse_unquote_plus(s
)
1373 url_components
.hostname
, url_components
.port
or 1080,
1375 unquote_if_non_empty(url_components
.username
),
1376 unquote_if_non_empty(url_components
.password
),
1379 class SocksConnection(base_class
):
1381 self
.sock
= sockssocket()
1382 self
.sock
.setproxy(*proxy_args
)
1383 if isinstance(self
.timeout
, (int, float)):
1384 self
.sock
.settimeout(self
.timeout
)
1385 self
.sock
.connect((self
.host
, self
.port
))
1387 if isinstance(self
, compat_http_client
.HTTPSConnection
):
1388 if hasattr(self
, '_context'): # Python > 2.6
1389 self
.sock
= self
._context
.wrap_socket(
1390 self
.sock
, server_hostname
=self
.host
)
1392 self
.sock
= ssl
.wrap_socket(self
.sock
)
1394 return SocksConnection
1397 class YoutubeDLHTTPSHandler(compat_urllib_request
.HTTPSHandler
):
1398 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1399 compat_urllib_request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1400 self
._https
_conn
_class
= https_conn_class
or compat_http_client
.HTTPSConnection
1401 self
._params
= params
1403 def https_open(self
, req
):
1405 conn_class
= self
._https
_conn
_class
1407 if hasattr(self
, '_context'): # python > 2.6
1408 kwargs
['context'] = self
._context
1409 if hasattr(self
, '_check_hostname'): # python 3.x
1410 kwargs
['check_hostname'] = self
._check
_hostname
1412 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1414 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1415 del req
.headers
['Ytdl-socks-proxy']
1417 return self
.do_open(functools
.partial(
1418 _create_http_connection
, self
, conn_class
, True),
1422 class YoutubeDLCookieJar(compat_cookiejar
.MozillaCookieJar
):
1424 See [1] for cookie file format.
1426 1. https://curl.haxx.se/docs/http-cookies.html
1428 _HTTPONLY_PREFIX
= '#HttpOnly_'
1430 _HEADER
= '''# Netscape HTTP Cookie File
1431 # This file is generated by yt-dlp. Do not edit.
1434 _CookieFileEntry
= collections
.namedtuple(
1436 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1438 def save(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1440 Save cookies to a file.
1442 Most of the code is taken from CPython 3.8 and slightly adapted
1443 to support cookie files with UTF-8 in both python 2 and 3.
1445 if filename
is None:
1446 if self
.filename
is not None:
1447 filename
= self
.filename
1449 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1451 # Store session cookies with `expires` set to 0 instead of an empty
1454 if cookie
.expires
is None:
1457 with open(filename
, 'w', encoding
='utf-8') as f
:
1458 f
.write(self
._HEADER
)
1461 if not ignore_discard
and cookie
.discard
:
1463 if not ignore_expires
and cookie
.is_expired(now
):
1469 if cookie
.domain
.startswith('.'):
1470 initial_dot
= 'TRUE'
1472 initial_dot
= 'FALSE'
1473 if cookie
.expires
is not None:
1474 expires
= compat_str(cookie
.expires
)
1477 if cookie
.value
is None:
1478 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1479 # with no name, whereas http.cookiejar regards it as a
1480 # cookie with no value.
1485 value
= cookie
.value
1487 '\t'.join([cookie
.domain
, initial_dot
, cookie
.path
,
1488 secure
, expires
, name
, value
]) + '\n')
1490 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1491 """Load cookies from a file."""
1492 if filename
is None:
1493 if self
.filename
is not None:
1494 filename
= self
.filename
1496 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1498 def prepare_line(line
):
1499 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1500 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1501 # comments and empty lines are fine
1502 if line
.startswith('#') or not line
.strip():
1504 cookie_list
= line
.split('\t')
1505 if len(cookie_list
) != self
._ENTRY
_LEN
:
1506 raise compat_cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1507 cookie
= self
._CookieFileEntry
(*cookie_list
)
1508 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1509 raise compat_cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1513 with open(filename
, encoding
='utf-8') as f
:
1516 cf
.write(prepare_line(line
))
1517 except compat_cookiejar
.LoadError
as e
:
1518 if f
'{line.strip()} '[0] in '[{"':
1519 raise compat_cookiejar
.LoadError(
1520 'Cookies file must be Netscape formatted, not JSON. See '
1521 'https://github.com/ytdl-org/youtube-dl#how-do-i-pass-cookies-to-youtube-dl')
1522 write_string(f
'WARNING: skipping cookie file entry due to {e}: {line!r}\n')
1525 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1526 # Session cookies are denoted by either `expires` field set to
1527 # an empty string or 0. MozillaCookieJar only recognizes the former
1528 # (see [1]). So we need force the latter to be recognized as session
1529 # cookies on our own.
1530 # Session cookies may be important for cookies-based authentication,
1531 # e.g. usually, when user does not check 'Remember me' check box while
1532 # logging in on a site, some important cookies are stored as session
1533 # cookies so that not recognizing them will result in failed login.
1534 # 1. https://bugs.python.org/issue17164
1536 # Treat `expires=0` cookies as session cookies
1537 if cookie
.expires
== 0:
1538 cookie
.expires
= None
1539 cookie
.discard
= True
1542 class YoutubeDLCookieProcessor(compat_urllib_request
.HTTPCookieProcessor
):
1543 def __init__(self
, cookiejar
=None):
1544 compat_urllib_request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1546 def http_response(self
, request
, response
):
1547 return compat_urllib_request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1549 https_request
= compat_urllib_request
.HTTPCookieProcessor
.http_request
1550 https_response
= http_response
1553 class YoutubeDLRedirectHandler(compat_urllib_request
.HTTPRedirectHandler
):
1554 """YoutubeDL redirect handler
1556 The code is based on HTTPRedirectHandler implementation from CPython [1].
1558 This redirect handler solves two issues:
1559 - ensures redirect URL is always unicode under python 2
1560 - introduces support for experimental HTTP response status code
1561 308 Permanent Redirect [2] used by some sites [3]
1563 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1564 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1565 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1568 http_error_301
= http_error_303
= http_error_307
= http_error_308
= compat_urllib_request
.HTTPRedirectHandler
.http_error_302
1570 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1571 """Return a Request or None in response to a redirect.
1573 This is called by the http_error_30x methods when a
1574 redirection response is received. If a redirection should
1575 take place, return a new Request to allow http_error_30x to
1576 perform the redirect. Otherwise, raise HTTPError if no-one
1577 else should try to handle this url. Return None if you can't
1578 but another Handler might.
1580 m
= req
.get_method()
1581 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1582 or code
in (301, 302, 303) and m
== "POST")):
1583 raise compat_HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1584 # Strictly (according to RFC 2616), 301 or 302 in response to
1585 # a POST MUST NOT cause a redirection without confirmation
1586 # from the user (of urllib.request, in this case). In practice,
1587 # essentially all clients do redirect in this case, so we do
1590 # Be conciliant with URIs containing a space. This is mainly
1591 # redundant with the more complete encoding done in http_error_302(),
1592 # but it is kept for compatibility with other callers.
1593 newurl
= newurl
.replace(' ', '%20')
1595 CONTENT_HEADERS
= ("content-length", "content-type")
1596 # NB: don't use dict comprehension for python 2.6 compatibility
1597 newheaders
= {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
1599 # A 303 must either use GET or HEAD for subsequent request
1600 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
1601 if code
== 303 and m
!= 'HEAD':
1603 # 301 and 302 redirects are commonly turned into a GET from a POST
1604 # for subsequent requests by browsers, so we'll do the same.
1605 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
1606 # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
1607 if code
in (301, 302) and m
== 'POST':
1610 return compat_urllib_request
.Request(
1611 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1612 unverifiable
=True, method
=m
)
1615 def extract_timezone(date_str
):
1618 ^.{8,}? # >=8 char non-TZ prefix, if present
1619 (?P<tz>Z| # just the UTC Z, or
1620 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1621 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1622 [ ]? # optional space
1623 (?P<sign>\+|-) # +/-
1624 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1628 timezone
= datetime
.timedelta()
1630 date_str
= date_str
[:-len(m
.group('tz'))]
1631 if not m
.group('sign'):
1632 timezone
= datetime
.timedelta()
1634 sign
= 1 if m
.group('sign') == '+' else -1
1635 timezone
= datetime
.timedelta(
1636 hours
=sign
* int(m
.group('hours')),
1637 minutes
=sign
* int(m
.group('minutes')))
1638 return timezone
, date_str
1641 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1642 """ Return a UNIX timestamp from the given date """
1644 if date_str
is None:
1647 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1649 if timezone
is None:
1650 timezone
, date_str
= extract_timezone(date_str
)
1652 with contextlib
.suppress(ValueError):
1653 date_format
= f
'%Y-%m-%d{delimiter}%H:%M:%S'
1654 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1655 return calendar
.timegm(dt
.timetuple())
1658 def date_formats(day_first
=True):
1659 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1662 def unified_strdate(date_str
, day_first
=True):
1663 """Return a string with the date in the format YYYYMMDD"""
1665 if date_str
is None:
1669 date_str
= date_str
.replace(',', ' ')
1670 # Remove AM/PM + timezone
1671 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1672 _
, date_str
= extract_timezone(date_str
)
1674 for expression
in date_formats(day_first
):
1675 with contextlib
.suppress(ValueError):
1676 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1677 if upload_date
is None:
1678 timetuple
= email
.utils
.parsedate_tz(date_str
)
1680 with contextlib
.suppress(ValueError):
1681 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1682 if upload_date
is not None:
1683 return compat_str(upload_date
)
1686 def unified_timestamp(date_str
, day_first
=True):
1687 if date_str
is None:
1690 date_str
= re
.sub(r
'[,|]', '', date_str
)
1692 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1693 timezone
, date_str
= extract_timezone(date_str
)
1695 # Remove AM/PM + timezone
1696 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1698 # Remove unrecognized timezones from ISO 8601 alike timestamps
1699 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1701 date_str
= date_str
[:-len(m
.group('tz'))]
1703 # Python only supports microseconds, so remove nanoseconds
1704 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]+$', date_str
)
1706 date_str
= m
.group(1)
1708 for expression
in date_formats(day_first
):
1709 with contextlib
.suppress(ValueError):
1710 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1711 return calendar
.timegm(dt
.timetuple())
1712 timetuple
= email
.utils
.parsedate_tz(date_str
)
1714 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1717 def determine_ext(url
, default_ext
='unknown_video'):
1718 if url
is None or '.' not in url
:
1720 guess
= url
.partition('?')[0].rpartition('.')[2]
1721 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1723 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1724 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1725 return guess
.rstrip('/')
1730 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1731 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1734 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1736 Return a datetime object from a string in the format YYYYMMDD or
1737 (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
1739 format: string date format used to return datetime object from
1740 precision: round the time portion of a datetime object.
1741 auto|microsecond|second|minute|hour|day.
1742 auto: round to the unit provided in date_str (if applicable).
1744 auto_precision
= False
1745 if precision
== 'auto':
1746 auto_precision
= True
1747 precision
= 'microsecond'
1748 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1749 if date_str
in ('now', 'today'):
1751 if date_str
== 'yesterday':
1752 return today
- datetime
.timedelta(days
=1)
1754 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)(s)?',
1756 if match
is not None:
1757 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1758 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1759 unit
= match
.group('unit')
1760 if unit
== 'month' or unit
== 'year':
1761 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1767 delta
= datetime
.timedelta(**{unit + 's': time}
)
1768 new_date
= start_time
+ delta
1770 return datetime_round(new_date
, unit
)
1773 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1776 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1778 Return a datetime object from a string in the format YYYYMMDD or
1779 (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
1781 If "strict", only (now|today)[+-][0-9](day|week|month|year)(s)? is allowed
1783 format: string date format used to return datetime object from
1785 if strict
and not re
.fullmatch(r
'\d{8}|(now|today)[+-]\d+(day|week|month|year)(s)?', date_str
):
1786 raise ValueError(f
'Invalid date format {date_str}')
1787 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1790 def datetime_add_months(dt
, months
):
1791 """Increment/Decrement a datetime object by months."""
1792 month
= dt
.month
+ months
- 1
1793 year
= dt
.year
+ month
// 12
1794 month
= month
% 12 + 1
1795 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1796 return dt
.replace(year
, month
, day
)
1799 def datetime_round(dt
, precision
='day'):
1801 Round a datetime object's time to a specific precision
1803 if precision
== 'microsecond':
1812 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1813 timestamp
= calendar
.timegm(dt
.timetuple())
1814 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1817 def hyphenate_date(date_str
):
1819 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1820 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1821 if match
is not None:
1822 return '-'.join(match
.groups())
1828 """Represents a time interval between two dates"""
1830 def __init__(self
, start
=None, end
=None):
1831 """start and end must be strings in the format accepted by date"""
1832 if start
is not None:
1833 self
.start
= date_from_str(start
, strict
=True)
1835 self
.start
= datetime
.datetime
.min.date()
1837 self
.end
= date_from_str(end
, strict
=True)
1839 self
.end
= datetime
.datetime
.max.date()
1840 if self
.start
> self
.end
:
1841 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1845 """Returns a range that only contains the given day"""
1846 return cls(day
, day
)
1848 def __contains__(self
, date
):
1849 """Check if the date is in the range"""
1850 if not isinstance(date
, datetime
.date
):
1851 date
= date_from_str(date
)
1852 return self
.start
<= date
<= self
.end
1855 return f
'{self.start.isoformat()} - {self.end.isoformat()}'
1858 def platform_name():
1859 """ Returns the platform name as a compat_str """
1860 res
= platform
.platform()
1861 if isinstance(res
, bytes):
1862 res
= res
.decode(preferredencoding())
1864 assert isinstance(res
, compat_str
)
1868 def get_windows_version():
1869 ''' Get Windows version. None if it's not running on Windows '''
1870 if compat_os_name
== 'nt':
1871 return version_tuple(platform
.win32_ver()[1])
1876 def write_string(s
, out
=None, encoding
=None):
1877 assert isinstance(s
, str)
1878 out
= out
or sys
.stderr
1880 from .compat
import WINDOWS_VT_MODE
# Must be imported locally
1882 s
= re
.sub(r
'([\r\n]+)', r
' \1', s
)
1884 if 'b' in getattr(out
, 'mode', ''):
1885 byt
= s
.encode(encoding
or preferredencoding(), 'ignore')
1887 elif hasattr(out
, 'buffer'):
1888 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
1889 byt
= s
.encode(enc
, 'ignore')
1890 out
.buffer.write(byt
)
1896 def bytes_to_intlist(bs
):
1899 if isinstance(bs
[0], int): # Python 3
1902 return [ord(c
) for c
in bs
]
1905 def intlist_to_bytes(xs
):
1908 return compat_struct_pack('%dB' % len(xs
), *xs
)
1911 class LockingUnsupportedError(IOError):
1912 msg
= 'File locking is not supported on this platform'
1915 super().__init
__(self
.msg
)
1918 # Cross-platform file locking
1919 if sys
.platform
== 'win32':
1920 import ctypes
.wintypes
1923 class OVERLAPPED(ctypes
.Structure
):
1925 ('Internal', ctypes
.wintypes
.LPVOID
),
1926 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
1927 ('Offset', ctypes
.wintypes
.DWORD
),
1928 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
1929 ('hEvent', ctypes
.wintypes
.HANDLE
),
1932 kernel32
= ctypes
.windll
.kernel32
1933 LockFileEx
= kernel32
.LockFileEx
1934 LockFileEx
.argtypes
= [
1935 ctypes
.wintypes
.HANDLE
, # hFile
1936 ctypes
.wintypes
.DWORD
, # dwFlags
1937 ctypes
.wintypes
.DWORD
, # dwReserved
1938 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1939 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1940 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1942 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
1943 UnlockFileEx
= kernel32
.UnlockFileEx
1944 UnlockFileEx
.argtypes
= [
1945 ctypes
.wintypes
.HANDLE
, # hFile
1946 ctypes
.wintypes
.DWORD
, # dwReserved
1947 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
1948 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
1949 ctypes
.POINTER(OVERLAPPED
) # Overlapped
1951 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
1952 whole_low
= 0xffffffff
1953 whole_high
= 0x7fffffff
1955 def _lock_file(f
, exclusive
, block
):
1956 overlapped
= OVERLAPPED()
1957 overlapped
.Offset
= 0
1958 overlapped
.OffsetHigh
= 0
1959 overlapped
.hEvent
= 0
1960 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
1962 if not LockFileEx(msvcrt
.get_osfhandle(f
.fileno()),
1963 (0x2 if exclusive
else 0x0) |
(0x0 if block
else 0x1),
1964 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
1965 raise BlockingIOError('Locking file failed: %r' % ctypes
.FormatError())
1967 def _unlock_file(f
):
1968 assert f
._lock
_file
_overlapped
_p
1969 handle
= msvcrt
.get_osfhandle(f
.fileno())
1970 if not UnlockFileEx(handle
, 0, whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
1971 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
1977 def _lock_file(f
, exclusive
, block
):
1978 flags
= fcntl
.LOCK_EX
if exclusive
else fcntl
.LOCK_SH
1980 flags |
= fcntl
.LOCK_NB
1982 fcntl
.flock(f
, flags
)
1983 except BlockingIOError
:
1985 except OSError: # AOSP does not have flock()
1986 fcntl
.lockf(f
, flags
)
1988 def _unlock_file(f
):
1990 fcntl
.flock(f
, fcntl
.LOCK_UN
)
1992 fcntl
.lockf(f
, fcntl
.LOCK_UN
)
1996 def _lock_file(f
, exclusive
, block
):
1997 raise LockingUnsupportedError()
1999 def _unlock_file(f
):
2000 raise LockingUnsupportedError()
2006 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2007 if mode
not in {'r', 'rb', 'a', 'ab', 'w', 'wb'}
:
2008 raise NotImplementedError(mode
)
2009 self
.mode
, self
.block
= mode
, block
2011 writable
= any(f
in mode
for f
in 'wax+')
2012 readable
= any(f
in mode
for f
in 'r+')
2013 flags
= functools
.reduce(operator
.ior
, (
2014 getattr(os
, 'O_CLOEXEC', 0), # UNIX only
2015 getattr(os
, 'O_BINARY', 0), # Windows only
2016 getattr(os
, 'O_NOINHERIT', 0), # Windows only
2017 os
.O_CREAT
if writable
else 0, # O_TRUNC only after locking
2018 os
.O_APPEND
if 'a' in mode
else 0,
2019 os
.O_EXCL
if 'x' in mode
else 0,
2020 os
.O_RDONLY
if not writable
else os
.O_RDWR
if readable
else os
.O_WRONLY
,
2023 self
.f
= os
.fdopen(os
.open(filename
, flags
, 0o666), mode
, encoding
=encoding
)
2025 def __enter__(self
):
2026 exclusive
= 'r' not in self
.mode
2028 _lock_file(self
.f
, exclusive
, self
.block
)
2033 if 'w' in self
.mode
:
2036 except OSError as e
:
2037 if e
.errno
!= 29: # Illegal seek, expected when self.f is a FIFO
2045 _unlock_file(self
.f
)
2049 def __exit__(self
, *_
):
2058 def __getattr__(self
, attr
):
2059 return getattr(self
.f
, attr
)
2065 def get_filesystem_encoding():
2066 encoding
= sys
.getfilesystemencoding()
2067 return encoding
if encoding
is not None else 'utf-8'
2070 def shell_quote(args
):
2072 encoding
= get_filesystem_encoding()
2074 if isinstance(a
, bytes):
2075 # We may get a filename encoded with 'encodeFilename'
2076 a
= a
.decode(encoding
)
2077 quoted_args
.append(compat_shlex_quote(a
))
2078 return ' '.join(quoted_args
)
2081 def smuggle_url(url
, data
):
2082 """ Pass additional data in a URL for internal use. """
2084 url
, idata
= unsmuggle_url(url
, {})
2086 sdata
= compat_urllib_parse_urlencode(
2087 {'__youtubedl_smuggle': json.dumps(data)}
)
2088 return url
+ '#' + sdata
2091 def unsmuggle_url(smug_url
, default
=None):
2092 if '#__youtubedl_smuggle' not in smug_url
:
2093 return smug_url
, default
2094 url
, _
, sdata
= smug_url
.rpartition('#')
2095 jsond
= compat_parse_qs(sdata
)['__youtubedl_smuggle'][0]
2096 data
= json
.loads(jsond
)
2100 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2101 """ Formats numbers with decimal sufixes like K, M, etc """
2102 num
, factor
= float_or_none(num
), float(factor
)
2103 if num
is None or num
< 0:
2105 POSSIBLE_SUFFIXES
= 'kMGTPEZY'
2106 exponent
= 0 if num
== 0 else min(int(math
.log(num
, factor
)), len(POSSIBLE_SUFFIXES
))
2107 suffix
= ['', *POSSIBLE_SUFFIXES
][exponent
]
2109 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2110 converted
= num
/ (factor
** exponent
)
2111 return fmt
% (converted
, suffix
)
2114 def format_bytes(bytes):
2115 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2118 def lookup_unit_table(unit_table
, s
):
2119 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2121 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2124 num_str
= m
.group('num').replace(',', '.')
2125 mult
= unit_table
[m
.group('unit')]
2126 return int(float(num_str
) * mult
)
2129 def parse_filesize(s
):
2133 # The lower-case forms are of course incorrect and unofficial,
2134 # but we support those too
2151 'megabytes': 1000 ** 2,
2152 'mebibytes': 1024 ** 2,
2158 'gigabytes': 1000 ** 3,
2159 'gibibytes': 1024 ** 3,
2165 'terabytes': 1000 ** 4,
2166 'tebibytes': 1024 ** 4,
2172 'petabytes': 1000 ** 5,
2173 'pebibytes': 1024 ** 5,
2179 'exabytes': 1000 ** 6,
2180 'exbibytes': 1024 ** 6,
2186 'zettabytes': 1000 ** 7,
2187 'zebibytes': 1024 ** 7,
2193 'yottabytes': 1000 ** 8,
2194 'yobibytes': 1024 ** 8,
2197 return lookup_unit_table(_UNIT_TABLE
, s
)
2204 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2206 if re
.match(r
'^[\d,.]+$', s
):
2207 return str_to_int(s
)
2220 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2224 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2226 return str_to_int(mobj
.group(1))
2229 def parse_resolution(s
, *, lenient
=False):
2234 mobj
= re
.search(r
'(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)', s
)
2236 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2239 'width': int(mobj
.group('w')),
2240 'height': int(mobj
.group('h')),
2243 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2245 return {'height': int(mobj.group(1))}
2247 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2249 return {'height': int(mobj.group(1)) * 540}
2254 def parse_bitrate(s
):
2255 if not isinstance(s
, compat_str
):
2257 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2259 return int(mobj
.group(1))
2262 def month_by_name(name
, lang
='en'):
2263 """ Return the number of a month by (locale-independently) English name """
2265 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2268 return month_names
.index(name
) + 1
2273 def month_by_abbreviation(abbrev
):
2274 """ Return the number of a month by (locale-independently) English
2278 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2283 def fix_xml_ampersands(xml_str
):
2284 """Replace all the '&' by '&' in XML"""
2286 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2291 def setproctitle(title
):
2292 assert isinstance(title
, compat_str
)
2294 # ctypes in Jython is not complete
2295 # http://bugs.jython.org/issue2148
2296 if sys
.platform
.startswith('java'):
2300 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2304 # LoadLibrary in Windows Python 2.7.13 only expects
2305 # a bytestring, but since unicode_literals turns
2306 # every string into a unicode string, it fails.
2308 title_bytes
= title
.encode('utf-8')
2309 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2310 buf
.value
= title_bytes
2312 libc
.prctl(15, buf
, 0, 0, 0)
2313 except AttributeError:
2314 return # Strange libc, just skip this
2317 def remove_start(s
, start
):
2318 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2321 def remove_end(s
, end
):
2322 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2325 def remove_quotes(s
):
2326 if s
is None or len(s
) < 2:
2328 for quote
in ('"', "'", ):
2329 if s
[0] == quote
and s
[-1] == quote
:
2334 def get_domain(url
):
2335 domain
= re
.match(r
'(?:https?:\/\/)?(?:www\.)?(?P<domain>[^\n\/]+\.[^\n\/]+)(?:\/(.*))?', url
)
2336 return domain
.group('domain') if domain
else None
2339 def url_basename(url
):
2340 path
= compat_urlparse
.urlparse(url
).path
2341 return path
.strip('/').split('/')[-1]
2345 return re
.match(r
'https?://[^?#&]+/', url
).group()
2348 def urljoin(base
, path
):
2349 if isinstance(path
, bytes):
2350 path
= path
.decode('utf-8')
2351 if not isinstance(path
, compat_str
) or not path
:
2353 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2355 if isinstance(base
, bytes):
2356 base
= base
.decode('utf-8')
2357 if not isinstance(base
, compat_str
) or not re
.match(
2358 r
'^(?:https?:)?//', base
):
2360 return compat_urlparse
.urljoin(base
, path
)
2363 class HEADRequest(compat_urllib_request
.Request
):
2364 def get_method(self
):
2368 class PUTRequest(compat_urllib_request
.Request
):
2369 def get_method(self
):
2373 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2374 if get_attr
and v
is not None:
2375 v
= getattr(v
, get_attr
, None)
2377 return int(v
) * invscale
// scale
2378 except (ValueError, TypeError, OverflowError):
2382 def str_or_none(v
, default
=None):
2383 return default
if v
is None else compat_str(v
)
2386 def str_to_int(int_str
):
2387 """ A more relaxed version of int_or_none """
2388 if isinstance(int_str
, int):
2390 elif isinstance(int_str
, compat_str
):
2391 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2392 return int_or_none(int_str
)
2395 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2399 return float(v
) * invscale
/ scale
2400 except (ValueError, TypeError):
2404 def bool_or_none(v
, default
=None):
2405 return v
if isinstance(v
, bool) else default
2408 def strip_or_none(v
, default
=None):
2409 return v
.strip() if isinstance(v
, compat_str
) else default
2412 def url_or_none(url
):
2413 if not url
or not isinstance(url
, compat_str
):
2416 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2419 def request_to_url(req
):
2420 if isinstance(req
, compat_urllib_request
.Request
):
2421 return req
.get_full_url()
2426 def strftime_or_none(timestamp
, date_format
, default
=None):
2427 datetime_object
= None
2429 if isinstance(timestamp
, (int, float)): # unix timestamp
2430 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2431 elif isinstance(timestamp
, compat_str
): # assume YYYYMMDD
2432 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2433 return datetime_object
.strftime(date_format
)
2434 except (ValueError, TypeError, AttributeError):
2438 def parse_duration(s
):
2439 if not isinstance(s
, str):
2445 days
, hours
, mins
, secs
, ms
= [None] * 5
2446 m
= re
.match(r
'''(?x)
2448 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2449 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2450 (?P<ms>[.:][0-9]+)?Z?$
2453 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2458 [0-9]+\s*y(?:ears?)?,?\s*
2461 [0-9]+\s*m(?:onths?)?,?\s*
2464 [0-9]+\s*w(?:eeks?)?,?\s*
2467 (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
2471 (?P<hours>[0-9]+)\s*h(?:ours?)?,?\s*
2474 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
2477 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2480 days
, hours
, mins
, secs
, ms
= m
.groups()
2482 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2484 hours
, mins
= m
.groups()
2489 ms
= ms
.replace(':', '.')
2490 return sum(float(part
or 0) * mult
for part
, mult
in (
2491 (days
, 86400), (hours
, 3600), (mins
, 60), (secs
, 1), (ms
, 1)))
2494 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2495 name
, real_ext
= os
.path
.splitext(filename
)
2497 f
'{name}.{ext}{real_ext}'
2498 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2499 else f
'{filename}.{ext}')
2502 def replace_extension(filename
, ext
, expected_real_ext
=None):
2503 name
, real_ext
= os
.path
.splitext(filename
)
2504 return '{}.{}'.format(
2505 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2509 def check_executable(exe
, args
=[]):
2510 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2511 args can be a list of arguments for a short output (like -version) """
2513 Popen([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
).communicate_or_kill()
2519 def _get_exe_version_output(exe
, args
, *, to_screen
=None):
2521 to_screen(f
'Checking exe version: {shell_quote([exe] + args)}')
2523 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2524 # SIGTTOU if yt-dlp is run in the background.
2525 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2527 [encodeArgument(exe
)] + args
, stdin
=subprocess
.PIPE
,
2528 stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
).communicate_or_kill()
2531 if isinstance(out
, bytes): # Python 2.x
2532 out
= out
.decode('ascii', 'ignore')
2536 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2537 assert isinstance(output
, compat_str
)
2538 if version_re
is None:
2539 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2540 m
= re
.search(version_re
, output
)
2547 def get_exe_version(exe
, args
=['--version'],
2548 version_re
=None, unrecognized
='present'):
2549 """ Returns the version of the specified executable,
2550 or False if the executable is not present """
2551 out
= _get_exe_version_output(exe
, args
)
2552 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2555 class LazyList(collections
.abc
.Sequence
):
2556 ''' Lazy immutable list from an iterable
2557 Note that slices of a LazyList are lists and not LazyList'''
2559 class IndexError(IndexError):
2562 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2563 self
.__iterable
= iter(iterable
)
2564 self
.__cache
= [] if _cache
is None else _cache
2565 self
.__reversed
= reverse
2569 # We need to consume the entire iterable to iterate in reverse
2570 yield from self
.exhaust()
2572 yield from self
.__cache
2573 for item
in self
.__iterable
:
2574 self
.__cache
.append(item
)
2577 def __exhaust(self
):
2578 self
.__cache
.extend(self
.__iterable
)
2579 # Discard the emptied iterable to make it pickle-able
2580 self
.__iterable
= []
2584 ''' Evaluate the entire iterable '''
2585 return self
.__exhaust
()[::-1 if self
.__reversed
else 1]
2588 def __reverse_index(x
):
2589 return None if x
is None else -(x
+ 1)
2591 def __getitem__(self
, idx
):
2592 if isinstance(idx
, slice):
2594 idx
= slice(self
.__reverse
_index
(idx
.start
), self
.__reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2595 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2596 elif isinstance(idx
, int):
2598 idx
= self
.__reverse
_index
(idx
)
2599 start
, stop
, step
= idx
, idx
, 0
2601 raise TypeError('indices must be integers or slices')
2602 if ((start
or 0) < 0 or (stop
or 0) < 0
2603 or (start
is None and step
< 0)
2604 or (stop
is None and step
> 0)):
2605 # We need to consume the entire iterable to be able to slice from the end
2606 # Obviously, never use this with infinite iterables
2609 return self
.__cache
[idx
]
2610 except IndexError as e
:
2611 raise self
.IndexError(e
) from e
2612 n
= max(start
or 0, stop
or 0) - len(self
.__cache
) + 1
2614 self
.__cache
.extend(itertools
.islice(self
.__iterable
, n
))
2616 return self
.__cache
[idx
]
2617 except IndexError as e
:
2618 raise self
.IndexError(e
) from e
2622 self
[-1] if self
.__reversed
else self
[0]
2623 except self
.IndexError:
2629 return len(self
.__cache
)
2631 def __reversed__(self
):
2632 return type(self
)(self
.__iterable
, reverse
=not self
.__reversed
, _cache
=self
.__cache
)
2635 return type(self
)(self
.__iterable
, reverse
=self
.__reversed
, _cache
=self
.__cache
)
2638 # repr and str should mimic a list. So we exhaust the iterable
2639 return repr(self
.exhaust())
2642 return repr(self
.exhaust())
2647 class IndexError(IndexError):
2651 # This is only useful for tests
2652 return len(self
.getslice())
2654 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2655 self
._pagefunc
= pagefunc
2656 self
._pagesize
= pagesize
2657 self
._pagecount
= float('inf')
2658 self
._use
_cache
= use_cache
2661 def getpage(self
, pagenum
):
2662 page_results
= self
._cache
.get(pagenum
)
2663 if page_results
is None:
2664 page_results
= [] if pagenum
> self
._pagecount
else list(self
._pagefunc
(pagenum
))
2666 self
._cache
[pagenum
] = page_results
2669 def getslice(self
, start
=0, end
=None):
2670 return list(self
._getslice
(start
, end
))
2672 def _getslice(self
, start
, end
):
2673 raise NotImplementedError('This method must be implemented by subclasses')
2675 def __getitem__(self
, idx
):
2676 assert self
._use
_cache
, 'Indexing PagedList requires cache'
2677 if not isinstance(idx
, int) or idx
< 0:
2678 raise TypeError('indices must be non-negative integers')
2679 entries
= self
.getslice(idx
, idx
+ 1)
2681 raise self
.IndexError()
2685 class OnDemandPagedList(PagedList
):
2686 """Download pages until a page with less than maximum results"""
2688 def _getslice(self
, start
, end
):
2689 for pagenum
in itertools
.count(start
// self
._pagesize
):
2690 firstid
= pagenum
* self
._pagesize
2691 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2692 if start
>= nextfirstid
:
2696 start
% self
._pagesize
2697 if firstid
<= start
< nextfirstid
2700 ((end
- 1) % self
._pagesize
) + 1
2701 if (end
is not None and firstid
<= end
<= nextfirstid
)
2705 page_results
= self
.getpage(pagenum
)
2707 self
._pagecount
= pagenum
- 1
2709 if startv
!= 0 or endv
is not None:
2710 page_results
= page_results
[startv
:endv
]
2711 yield from page_results
2713 # A little optimization - if current page is not "full", ie. does
2714 # not contain page_size videos then we can assume that this page
2715 # is the last one - there are no more ids on further pages -
2716 # i.e. no need to query again.
2717 if len(page_results
) + startv
< self
._pagesize
:
2720 # If we got the whole page, but the next page is not interesting,
2721 # break out early as well
2722 if end
== nextfirstid
:
2726 class InAdvancePagedList(PagedList
):
2727 """PagedList with total number of pages known in advance"""
2729 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2730 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2731 self
._pagecount
= pagecount
2733 def _getslice(self
, start
, end
):
2734 start_page
= start
// self
._pagesize
2735 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2736 skip_elems
= start
- start_page
* self
._pagesize
2737 only_more
= None if end
is None else end
- start
2738 for pagenum
in range(start_page
, end_page
):
2739 page_results
= self
.getpage(pagenum
)
2741 page_results
= page_results
[skip_elems
:]
2743 if only_more
is not None:
2744 if len(page_results
) < only_more
:
2745 only_more
-= len(page_results
)
2747 yield from page_results
[:only_more
]
2749 yield from page_results
2752 def uppercase_escape(s
):
2753 unicode_escape
= codecs
.getdecoder('unicode_escape')
2755 r
'\\U[0-9a-fA-F]{8}',
2756 lambda m
: unicode_escape(m
.group(0))[0],
2760 def lowercase_escape(s
):
2761 unicode_escape
= codecs
.getdecoder('unicode_escape')
2763 r
'\\u[0-9a-fA-F]{4}',
2764 lambda m
: unicode_escape(m
.group(0))[0],
2768 def escape_rfc3986(s
):
2769 """Escape non-ASCII characters as suggested by RFC 3986"""
2770 return urllib
.parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2773 def escape_url(url
):
2774 """Escape URL as suggested by RFC 3986"""
2775 url_parsed
= compat_urllib_parse_urlparse(url
)
2776 return url_parsed
._replace
(
2777 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2778 path
=escape_rfc3986(url_parsed
.path
),
2779 params
=escape_rfc3986(url_parsed
.params
),
2780 query
=escape_rfc3986(url_parsed
.query
),
2781 fragment
=escape_rfc3986(url_parsed
.fragment
)
2786 return compat_parse_qs(compat_urllib_parse_urlparse(url
).query
)
2789 def read_batch_urls(batch_fd
):
2791 if not isinstance(url
, compat_str
):
2792 url
= url
.decode('utf-8', 'replace')
2793 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
2794 for bom
in BOM_UTF8
:
2795 if url
.startswith(bom
):
2796 url
= url
[len(bom
):]
2798 if not url
or url
.startswith(('#', ';', ']')):
2800 # "#" cannot be stripped out since it is part of the URI
2801 # However, it can be safely stipped out if follwing a whitespace
2802 return re
.split(r
'\s#', url
, 1)[0].rstrip()
2804 with contextlib
.closing(batch_fd
) as fd
:
2805 return [url
for url
in map(fixup
, fd
) if url
]
2808 def urlencode_postdata(*args
, **kargs
):
2809 return compat_urllib_parse_urlencode(*args
, **kargs
).encode('ascii')
2812 def update_url_query(url
, query
):
2815 parsed_url
= compat_urlparse
.urlparse(url
)
2816 qs
= compat_parse_qs(parsed_url
.query
)
2818 return compat_urlparse
.urlunparse(parsed_url
._replace
(
2819 query
=compat_urllib_parse_urlencode(qs
, True)))
2822 def update_Request(req
, url
=None, data
=None, headers
={}, query={}
):
2823 req_headers
= req
.headers
.copy()
2824 req_headers
.update(headers
)
2825 req_data
= data
or req
.data
2826 req_url
= update_url_query(url
or req
.get_full_url(), query
)
2827 req_get_method
= req
.get_method()
2828 if req_get_method
== 'HEAD':
2829 req_type
= HEADRequest
2830 elif req_get_method
== 'PUT':
2831 req_type
= PUTRequest
2833 req_type
= compat_urllib_request
.Request
2835 req_url
, data
=req_data
, headers
=req_headers
,
2836 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
2837 if hasattr(req
, 'timeout'):
2838 new_req
.timeout
= req
.timeout
2842 def _multipart_encode_impl(data
, boundary
):
2843 content_type
= 'multipart/form-data; boundary=%s' % boundary
2846 for k
, v
in data
.items():
2847 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
2848 if isinstance(k
, compat_str
):
2849 k
= k
.encode('utf-8')
2850 if isinstance(v
, compat_str
):
2851 v
= v
.encode('utf-8')
2852 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
2853 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
2854 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
2855 if boundary
.encode('ascii') in content
:
2856 raise ValueError('Boundary overlaps with data')
2859 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
2861 return out
, content_type
2864 def multipart_encode(data
, boundary
=None):
2866 Encode a dict to RFC 7578-compliant form-data
2869 A dict where keys and values can be either Unicode or bytes-like
2872 If specified a Unicode object, it's used as the boundary. Otherwise
2873 a random boundary is generated.
2875 Reference: https://tools.ietf.org/html/rfc7578
2877 has_specified_boundary
= boundary
is not None
2880 if boundary
is None:
2881 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
2884 out
, content_type
= _multipart_encode_impl(data
, boundary
)
2887 if has_specified_boundary
:
2891 return out
, content_type
2894 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
2895 for val
in map(d
.get
, variadic(key_or_keys
)):
2896 if val
is not None and (val
or not skip_false_values
):
2901 def try_call(*funcs
, expected_type
=None, args
=[], kwargs
={}):
2904 val
= f(*args
, **kwargs
)
2905 except (AttributeError, KeyError, TypeError, IndexError, ZeroDivisionError):
2908 if expected_type
is None or isinstance(val
, expected_type
):
2912 def try_get(src
, getter
, expected_type
=None):
2913 return try_call(*variadic(getter
), args
=(src
,), expected_type
=expected_type
)
2916 def filter_dict(dct
, cndn
=lambda _
, v
: v
is not None):
2917 return {k: v for k, v in dct.items() if cndn(k, v)}
2920 def merge_dicts(*dicts
):
2922 for a_dict
in dicts
:
2923 for k
, v
in a_dict
.items():
2924 if (v
is not None and k
not in merged
2925 or isinstance(v
, str) and merged
[k
] == ''):
2930 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
2931 return string
if isinstance(string
, compat_str
) else compat_str(string
, encoding
, errors
)
2943 TV_PARENTAL_GUIDELINES
= {
2953 def parse_age_limit(s
):
2954 # isinstance(False, int) is True. So type() must be used instead
2956 return s
if 0 <= s
<= 21 else None
2957 elif not isinstance(s
, str):
2959 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
2961 return int(m
.group('age'))
2964 return US_RATINGS
[s
]
2965 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
2967 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
2971 def strip_jsonp(code
):
2974 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
2975 (?:\s*&&\s*(?P=func_name))?
2976 \s*\(\s*(?P<callback_data>.*)\);?
2977 \s*?(?://[^\n]*)*$''',
2978 r
'\g<callback_data>', code
)
2981 def js_to_json(code
, vars={}):
2982 # vars is a dict of var, val pairs to substitute
2983 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
2984 SKIP_RE
= fr
'\s*(?:{COMMENT_RE})?\s*'
2986 (fr
'(?s)^(0[xX][0-9a-fA-F]+){SKIP_RE}:?$', 16),
2987 (fr
'(?s)^(0+[0-7]+){SKIP_RE}:?$', 8),
2992 if v
in ('true', 'false', 'null'):
2994 elif v
in ('undefined', 'void 0'):
2996 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
2999 if v
[0] in ("'", '"'):
3000 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3005 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3007 for regex
, base
in INTEGER_TABLE
:
3008 im
= re
.match(regex
, v
)
3010 i
= int(im
.group(1), base
)
3011 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3018 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3020 return re
.sub(r
'''(?sx)
3021 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3022 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3023 {comment}|,(?={skip}[\]}}])|
3024 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3025 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3028 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3031 def qualities(quality_ids
):
3032 """ Get a numeric quality value out of a list of possible values """
3035 return quality_ids
.index(qid
)
3041 POSTPROCESS_WHEN
= ('pre_process', 'after_filter', 'before_dl', 'after_move', 'post_process', 'after_video', 'playlist')
3045 'default': '%(title)s [%(id)s].%(ext)s',
3046 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3052 'description': 'description',
3053 'annotation': 'annotations.xml',
3054 'infojson': 'info.json',
3057 'pl_thumbnail': None,
3058 'pl_description': 'description',
3059 'pl_infojson': 'info.json',
3062 # As of [1] format syntax is:
3063 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3064 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3065 STR_FORMAT_RE_TMPL
= r
'''(?x)
3066 (?<!%)(?P<prefix>(?:%%)*)
3068 (?P<has_key>\((?P<key>{0})\))?
3070 (?P<conversion>[#0\-+ ]+)?
3072 (?P<precision>\.\d+)?
3073 (?P<len_mod>[hlL])? # unused in python
3074 {1} # conversion type
3079 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3082 def limit_length(s
, length
):
3083 """ Add ellipses to overly long strings """
3088 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3092 def version_tuple(v
):
3093 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3096 def is_outdated_version(version
, limit
, assume_new
=True):
3098 return not assume_new
3100 return version_tuple(version
) < version_tuple(limit
)
3102 return not assume_new
3105 def ytdl_is_updateable():
3106 """ Returns if yt-dlp can be updated with -U """
3108 from .update
import is_non_updateable
3110 return not is_non_updateable()
3113 def args_to_str(args
):
3114 # Get a short string representation for a subprocess command
3115 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3118 def error_to_compat_str(err
):
3122 def error_to_str(err
):
3123 return f
'{type(err).__name__}: {err}'
3126 def mimetype2ext(mt
):
3130 mt
, _
, params
= mt
.partition(';')
3135 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3136 # it's the most popular one
3137 'audio/mpeg': 'mp3',
3138 'audio/x-wav': 'wav',
3140 'audio/wave': 'wav',
3143 ext
= FULL_MAP
.get(mt
)
3149 'smptett+xml': 'tt',
3153 'x-mp4-fragmented': 'mp4',
3154 'x-ms-sami': 'sami',
3157 'x-mpegurl': 'm3u8',
3158 'vnd.apple.mpegurl': 'm3u8',
3162 'vnd.ms-sstr+xml': 'ism',
3166 'filmstrip+json': 'fs',
3170 _
, _
, subtype
= mt
.rpartition('/')
3171 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3182 _
, _
, suffix
= subtype
.partition('+')
3183 ext
= SUFFIX_MAP
.get(suffix
)
3187 return subtype
.replace('+', '.')
3190 def ext2mimetype(ext_or_url
):
3193 if '.' not in ext_or_url
:
3194 ext_or_url
= f
'file.{ext_or_url}'
3195 return mimetypes
.guess_type(ext_or_url
)[0]
3198 def parse_codecs(codecs_str
):
3199 # http://tools.ietf.org/html/rfc6381
3202 split_codecs
= list(filter(None, map(
3203 str.strip
, codecs_str
.strip().strip(',').split(','))))
3204 vcodec
, acodec
, scodec
, hdr
= None, None, None, None
3205 for full_codec
in split_codecs
:
3206 parts
= full_codec
.split('.')
3207 codec
= parts
[0].replace('0', '')
3208 if codec
in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3209 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3211 vcodec
= '.'.join(parts
[:4]) if codec
in ('vp9', 'av1', 'hvc1') else full_codec
3212 if codec
in ('dvh1', 'dvhe'):
3214 elif codec
== 'av1' and len(parts
) > 3 and parts
[3] == '10':
3216 elif full_codec
.replace('0', '').startswith('vp9.2'):
3218 elif codec
in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3221 elif codec
in ('stpp', 'wvtt',):
3225 write_string(f
'WARNING: Unknown codec {full_codec}\n')
3226 if vcodec
or acodec
or scodec
:
3228 'vcodec': vcodec
or 'none',
3229 'acodec': acodec
or 'none',
3230 'dynamic_range': hdr
,
3231 **({'scodec': scodec}
if scodec
is not None else {}),
3233 elif len(split_codecs
) == 2:
3235 'vcodec': split_codecs
[0],
3236 'acodec': split_codecs
[1],
3241 def urlhandle_detect_ext(url_handle
):
3242 getheader
= url_handle
.headers
.get
3244 cd
= getheader('Content-Disposition')
3246 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3248 e
= determine_ext(m
.group('filename'), default_ext
=None)
3252 return mimetype2ext(getheader('Content-Type'))
3255 def encode_data_uri(data
, mime_type
):
3256 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3259 def age_restricted(content_limit
, age_limit
):
3260 """ Returns True iff the content should be blocked """
3262 if age_limit
is None: # No limit set
3264 if content_limit
is None:
3265 return False # Content available for everyone
3266 return age_limit
< content_limit
3269 def is_html(first_bytes
):
3270 """ Detect whether a file contains HTML by examining its first bytes. """
3273 (b
'\xef\xbb\xbf', 'utf-8'),
3274 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3275 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3276 (b
'\xff\xfe', 'utf-16-le'),
3277 (b
'\xfe\xff', 'utf-16-be'),
3279 for bom
, enc
in BOMS
:
3280 if first_bytes
.startswith(bom
):
3281 s
= first_bytes
[len(bom
):].decode(enc
, 'replace')
3284 s
= first_bytes
.decode('utf-8', 'replace')
3286 return re
.match(r
'^\s*<', s
)
3289 def determine_protocol(info_dict
):
3290 protocol
= info_dict
.get('protocol')
3291 if protocol
is not None:
3294 url
= sanitize_url(info_dict
['url'])
3295 if url
.startswith('rtmp'):
3297 elif url
.startswith('mms'):
3299 elif url
.startswith('rtsp'):
3302 ext
= determine_ext(url
)
3308 return compat_urllib_parse_urlparse(url
).scheme
3311 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3312 """ Render a list of rows, each as a list of values.
3313 Text after a \t will be right aligned """
3315 return len(remove_terminal_sequences(string
).replace('\t', ''))
3317 def get_max_lens(table
):
3318 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3320 def filter_using_list(row
, filterArray
):
3321 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3323 max_lens
= get_max_lens(data
) if hide_empty
else []
3324 header_row
= filter_using_list(header_row
, max_lens
)
3325 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3327 table
= [header_row
] + data
3328 max_lens
= get_max_lens(table
)
3331 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3332 table
[1][-1] = table
[1][-1][:-extra_gap
* len(delim
)] # Remove extra_gap from end of delimiter
3334 for pos
, text
in enumerate(map(str, row
)):
3336 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3338 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3339 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3343 def _match_one(filter_part
, dct
, incomplete
):
3344 # TODO: Generalize code with YoutubeDL._build_format_filter
3345 STRING_OPERATORS
= {
3346 '*=': operator
.contains
,
3347 '^=': lambda attr
, value
: attr
.startswith(value
),
3348 '$=': lambda attr
, value
: attr
.endswith(value
),
3349 '~=': lambda attr
, value
: re
.search(value
, attr
),
3351 COMPARISON_OPERATORS
= {
3353 '<=': operator
.le
, # "<=" must be defined above "<"
3360 if isinstance(incomplete
, bool):
3361 is_incomplete
= lambda _
: incomplete
3363 is_incomplete
= lambda k
: k
in incomplete
3365 operator_rex
= re
.compile(r
'''(?x)\s*
3367 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3369 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3373 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3374 m = operator_rex.search(filter_part)
3377 unnegated_op = COMPARISON_OPERATORS[m['op']]
3379 op = lambda attr, value: not unnegated_op(attr, value)
3382 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3384 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3385 actual_value = dct.get(m['key'])
3386 numeric_comparison = None
3387 if isinstance(actual_value, (int, float)):
3388 # If the original field is a string and matching comparisonvalue is
3389 # a number we should respect the origin of the original field
3390 # and process comparison value as a string (see
3391 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3393 numeric_comparison = int(comparison_value)
3395 numeric_comparison = parse_filesize(comparison_value)
3396 if numeric_comparison is None:
3397 numeric_comparison = parse_filesize(f'{comparison_value}B')
3398 if numeric_comparison is None:
3399 numeric_comparison = parse_duration(comparison_value)
3400 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3401 raise ValueError('Operator %s only supports string values!' % m['op'])
3402 if actual_value is None:
3403 return is_incomplete(m['key']) or m['none_inclusive']
3404 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3407 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3408 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3410 operator_rex = re.compile(r'''(?x
)\s
*
3411 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3413 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3414 m = operator_rex.search(filter_part)
3416 op = UNARY_OPERATORS[m.group('op')]
3417 actual_value = dct.get(m.group('key'))
3418 if is_incomplete(m.group('key')) and actual_value is None:
3420 return op(actual_value)
3422 raise ValueError('Invalid filter part %r' % filter_part)
3425 def match_str(filter_str, dct, incomplete=False):
3426 """ Filter a dictionary with a simple string syntax.
3427 @returns Whether the filter passes
3428 @param incomplete Set of keys that is expected to be missing from dct.
3429 Can be True/False to indicate all/none of the keys may be missing.
3430 All conditions on incomplete keys pass if the key is missing
3433 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3434 for filter_part in re.split(r'(?<!\\)&', filter_str))
3437 def match_filter_func(filters):
3440 filters = set(variadic(filters))
3442 interactive = '-' in filters
3446 def _match_func(info_dict, incomplete=False):
3447 if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
3448 return NO_DEFAULT if interactive and not incomplete else None
3450 video_title = info_dict.get('title') or info_dict.get('id') or 'video'
3451 filter_str = ') | ('.join(map(str.strip, filters))
3452 return f'{video_title} does not pass filter ({filter_str}), skipping ..'
3456 def parse_dfxp_time_expr(time_expr):
3460 mobj = re.match(rf'^(?P<time_offset>{NUMBER_RE})s?$', time_expr)
3462 return float(mobj.group('time_offset'))
3464 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3466 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3469 def srt_subtitles_timecode(seconds):
3470 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3473 def ass_subtitles_timecode(seconds):
3474 time = timetuple_from_msec(seconds * 1000)
3475 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3478 def dfxp2srt(dfxp_data):
3480 @param dfxp_data A
bytes-like
object containing DFXP data
3481 @returns A
unicode object containing converted SRT data
3483 LEGACY_NAMESPACES = (
3484 (b'http://www.w3.org/ns/ttml', [
3485 b'http://www.w3.org/2004/11/ttaf1',
3486 b'http://www.w3.org/2006/04/ttaf1',
3487 b'http://www.w3.org/2006/10/ttaf1',
3489 (b'http://www.w3.org/ns/ttml#styling', [
3490 b'http://www.w3.org/ns/ttml#style',
3494 SUPPORTED_STYLING = [
3503 _x = functools.partial(xpath_with_ns, ns_map={
3504 'xml': 'http://www.w3.org/XML/1998/namespace',
3505 'ttml': 'http://www.w3.org/ns/ttml',
3506 'tts': 'http://www.w3.org/ns/ttml#styling',
3512 class TTMLPElementParser:
3514 _unclosed_elements = []
3515 _applied_styles = []
3517 def start(self, tag, attrib):
3518 if tag in (_x('ttml:br'), 'br'):
3521 unclosed_elements = []
3523 element_style_id = attrib.get('style')
3525 style.update(default_style)
3526 if element_style_id:
3527 style.update(styles.get(element_style_id, {}))
3528 for prop in SUPPORTED_STYLING:
3529 prop_val = attrib.get(_x('tts:' + prop))
3531 style[prop] = prop_val
3534 for k, v in sorted(style.items()):
3535 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3538 font += ' color="%s"' % v
3539 elif k == 'fontSize':
3540 font += ' size="%s"' % v
3541 elif k == 'fontFamily':
3542 font += ' face="%s"' % v
3543 elif k == 'fontWeight' and v == 'bold':
3545 unclosed_elements.append('b')
3546 elif k == 'fontStyle' and v == 'italic':
3548 unclosed_elements.append('i')
3549 elif k == 'textDecoration' and v == 'underline':
3551 unclosed_elements.append('u')
3553 self._out += '<font' + font + '>'
3554 unclosed_elements.append('font')
3556 if self._applied_styles:
3557 applied_style.update(self._applied_styles[-1])
3558 applied_style.update(style)
3559 self._applied_styles.append(applied_style)
3560 self._unclosed_elements.append(unclosed_elements)
3563 if tag not in (_x('ttml:br'), 'br'):
3564 unclosed_elements = self._unclosed_elements.pop()
3565 for element in reversed(unclosed_elements):
3566 self._out += '</%s>' % element
3567 if unclosed_elements and self._applied_styles:
3568 self._applied_styles.pop()
3570 def data(self, data):
3574 return self._out.strip()
3576 def parse_node(node):
3577 target = TTMLPElementParser()
3578 parser = xml.etree.ElementTree.XMLParser(target=target)
3579 parser.feed(xml.etree.ElementTree.tostring(node))
3580 return parser.close()
3582 for k, v in LEGACY_NAMESPACES:
3584 dfxp_data = dfxp_data.replace(ns, k)
3586 dfxp = compat_etree_fromstring(dfxp_data)
3588 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3591 raise ValueError('Invalid dfxp/TTML subtitle')
3595 for style in dfxp.findall(_x('.//ttml:style')):
3596 style_id = style.get('id') or style.get(_x('xml:id'))
3599 parent_style_id = style.get('style')
3601 if parent_style_id not in styles:
3604 styles[style_id] = styles[parent_style_id].copy()
3605 for prop in SUPPORTED_STYLING:
3606 prop_val = style.get(_x('tts:' + prop))
3608 styles.setdefault(style_id, {})[prop] = prop_val
3614 for p in ('body', 'div'):
3615 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3618 style = styles.get(ele.get('style'))
3621 default_style.update(style)
3623 for para, index in zip(paras, itertools.count(1)):
3624 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3625 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3626 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3627 if begin_time is None:
3632 end_time = begin_time + dur
3633 out.append('%d\n%s --> %s\n%s\n\n' % (
3635 srt_subtitles_timecode(begin_time),
3636 srt_subtitles_timecode(end_time),
3642 def cli_option(params, command_option, param):
3643 param = params.get(param)
3645 param = compat_str(param)
3646 return [command_option, param] if param is not None else []
3649 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3650 param = params.get(param)
3653 assert isinstance(param, bool)
3655 return [command_option + separator + (true_value if param else false_value)]
3656 return [command_option, true_value if param else false_value]
3659 def cli_valueless_option(params, command_option, param, expected_value=True):
3660 param = params.get(param)
3661 return [command_option] if param == expected_value else []
3664 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3665 if isinstance(argdict, (list, tuple)): # for backward compatibility
3672 assert isinstance(argdict, dict)
3674 assert isinstance(keys, (list, tuple))
3675 for key_list in keys:
3676 arg_list = list(filter(
3677 lambda x: x is not None,
3678 [argdict.get(key.lower()) for key in variadic(key_list)]))
3680 return [arg for args in arg_list for arg in args]
3684 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3685 main_key, exe = main_key.lower(), exe.lower()
3686 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3687 keys = [f'{root_key}{k}' for k in (keys or [''])]
3688 if root_key in keys:
3690 keys.append((main_key, exe))
3691 keys.append('default')
3694 return cli_configuration_args(argdict, keys, default, use_compat)
3698 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
3757 'iw': 'heb', # Replaced by he in 1989 revision
3767 'in': 'ind', # Replaced by id in 1989 revision
3882 'ji': 'yid', # Replaced by yi in 1989 revision
3890 def short2long(cls, code):
3891 """Convert language code from ISO 639-1 to ISO 639-2/T"""
3892 return cls._lang_map.get(code[:2])
3895 def long2short(cls, code):
3896 """Convert language code from ISO 639-2/T to ISO 639-1"""
3897 for short_name, long_name in cls._lang_map.items():
3898 if long_name == code:
3903 # From http://data.okfn.org/data/core/country-list
3905 'AF': 'Afghanistan',
3906 'AX': 'Åland Islands',
3909 'AS': 'American Samoa',
3914 'AG': 'Antigua and Barbuda',
3931 'BO': 'Bolivia, Plurinational State of',
3932 'BQ': 'Bonaire, Sint Eustatius and Saba',
3933 'BA': 'Bosnia and Herzegovina',
3935 'BV': 'Bouvet Island',
3937 'IO': 'British Indian Ocean Territory',
3938 'BN': 'Brunei Darussalam',
3940 'BF': 'Burkina Faso',
3946 'KY': 'Cayman Islands',
3947 'CF': 'Central African Republic',
3951 'CX': 'Christmas Island',
3952 'CC': 'Cocos (Keeling) Islands',
3956 'CD': 'Congo, the Democratic Republic of the',
3957 'CK': 'Cook Islands',
3959 'CI': 'Côte d\'Ivoire',
3964 'CZ': 'Czech Republic',
3968 'DO': 'Dominican Republic',
3971 'SV': 'El Salvador',
3972 'GQ': 'Equatorial Guinea',
3976 'FK': 'Falkland Islands (Malvinas)',
3977 'FO': 'Faroe Islands',
3981 'GF': 'French Guiana',
3982 'PF': 'French Polynesia',
3983 'TF': 'French Southern Territories',
3998 'GW': 'Guinea-Bissau',
4001 'HM': 'Heard Island and McDonald Islands',
4002 'VA': 'Holy See (Vatican City State)',
4009 'IR': 'Iran, Islamic Republic of',
4012 'IM': 'Isle of Man',
4022 'KP': 'Korea, Democratic People\'s Republic of',
4023 'KR': 'Korea, Republic of',
4026 'LA': 'Lao People\'s Democratic Republic',
4032 'LI': 'Liechtenstein',
4036 'MK': 'Macedonia, the Former Yugoslav Republic of',
4043 'MH': 'Marshall Islands',
4049 'FM': 'Micronesia, Federated States of',
4050 'MD': 'Moldova, Republic of',
4061 'NL': 'Netherlands',
4062 'NC': 'New Caledonia',
4063 'NZ': 'New Zealand',
4068 'NF': 'Norfolk Island',
4069 'MP': 'Northern Mariana Islands',
4074 'PS': 'Palestine, State of',
4076 'PG': 'Papua New Guinea',
4079 'PH': 'Philippines',
4083 'PR': 'Puerto Rico',
4087 'RU': 'Russian Federation',
4089 'BL': 'Saint Barthélemy',
4090 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4091 'KN': 'Saint Kitts and Nevis',
4092 'LC': 'Saint Lucia',
4093 'MF': 'Saint Martin (French part)',
4094 'PM': 'Saint Pierre and Miquelon',
4095 'VC': 'Saint Vincent and the Grenadines',
4098 'ST': 'Sao Tome and Principe',
4099 'SA': 'Saudi Arabia',
4103 'SL': 'Sierra Leone',
4105 'SX': 'Sint Maarten (Dutch part)',
4108 'SB': 'Solomon Islands',
4110 'ZA': 'South Africa',
4111 'GS': 'South Georgia and the South Sandwich Islands',
4112 'SS': 'South Sudan',
4117 'SJ': 'Svalbard and Jan Mayen',
4120 'CH': 'Switzerland',
4121 'SY': 'Syrian Arab Republic',
4122 'TW': 'Taiwan, Province of China',
4124 'TZ': 'Tanzania, United Republic of',
4126 'TL': 'Timor-Leste',
4130 'TT': 'Trinidad and Tobago',
4133 'TM': 'Turkmenistan',
4134 'TC': 'Turks and Caicos Islands',
4138 'AE': 'United Arab Emirates',
4139 'GB': 'United Kingdom',
4140 'US': 'United States',
4141 'UM': 'United States Minor Outlying Islands',
4145 'VE': 'Venezuela, Bolivarian Republic of',
4147 'VG': 'Virgin Islands, British',
4148 'VI': 'Virgin Islands, U.S.',
4149 'WF': 'Wallis and Futuna',
4150 'EH': 'Western Sahara',
4157 def short2full(cls, code):
4158 """Convert an ISO 3166-2 country code to the corresponding full name"""
4159 return cls._country_map.get(code.upper())
4163 # Major IPv4 address blocks per country
4165 'AD': '46.172.224.0/19',
4166 'AE': '94.200.0.0/13',
4167 'AF': '149.54.0.0/17',
4168 'AG': '209.59.64.0/18',
4169 'AI': '204.14.248.0/21',
4170 'AL': '46.99.0.0/16',
4171 'AM': '46.70.0.0/15',
4172 'AO': '105.168.0.0/13',
4173 'AP': '182.50.184.0/21',
4174 'AQ': '23.154.160.0/24',
4175 'AR': '181.0.0.0/12',
4176 'AS': '202.70.112.0/20',
4177 'AT': '77.116.0.0/14',
4178 'AU': '1.128.0.0/11',
4179 'AW': '181.41.0.0/18',
4180 'AX': '185.217.4.0/22',
4181 'AZ': '5.197.0.0/16',
4182 'BA': '31.176.128.0/17',
4183 'BB': '65.48.128.0/17',
4184 'BD': '114.130.0.0/16',
4186 'BF': '102.178.0.0/15',
4187 'BG': '95.42.0.0/15',
4188 'BH': '37.131.0.0/17',
4189 'BI': '154.117.192.0/18',
4190 'BJ': '137.255.0.0/16',
4191 'BL': '185.212.72.0/23',
4192 'BM': '196.12.64.0/18',
4193 'BN': '156.31.0.0/16',
4194 'BO': '161.56.0.0/16',
4195 'BQ': '161.0.80.0/20',
4196 'BR': '191.128.0.0/12',
4197 'BS': '24.51.64.0/18',
4198 'BT': '119.2.96.0/19',
4199 'BW': '168.167.0.0/16',
4200 'BY': '178.120.0.0/13',
4201 'BZ': '179.42.192.0/18',
4202 'CA': '99.224.0.0/11',
4203 'CD': '41.243.0.0/16',
4204 'CF': '197.242.176.0/21',
4205 'CG': '160.113.0.0/16',
4206 'CH': '85.0.0.0/13',
4207 'CI': '102.136.0.0/14',
4208 'CK': '202.65.32.0/19',
4209 'CL': '152.172.0.0/14',
4210 'CM': '102.244.0.0/14',
4211 'CN': '36.128.0.0/10',
4212 'CO': '181.240.0.0/12',
4213 'CR': '201.192.0.0/12',
4214 'CU': '152.206.0.0/15',
4215 'CV': '165.90.96.0/19',
4216 'CW': '190.88.128.0/17',
4217 'CY': '31.153.0.0/16',
4218 'CZ': '88.100.0.0/14',
4220 'DJ': '197.241.0.0/17',
4221 'DK': '87.48.0.0/12',
4222 'DM': '192.243.48.0/20',
4223 'DO': '152.166.0.0/15',
4224 'DZ': '41.96.0.0/12',
4225 'EC': '186.68.0.0/15',
4226 'EE': '90.190.0.0/15',
4227 'EG': '156.160.0.0/11',
4228 'ER': '196.200.96.0/20',
4229 'ES': '88.0.0.0/11',
4230 'ET': '196.188.0.0/14',
4231 'EU': '2.16.0.0/13',
4232 'FI': '91.152.0.0/13',
4233 'FJ': '144.120.0.0/16',
4234 'FK': '80.73.208.0/21',
4235 'FM': '119.252.112.0/20',
4236 'FO': '88.85.32.0/19',
4238 'GA': '41.158.0.0/15',
4240 'GD': '74.122.88.0/21',
4241 'GE': '31.146.0.0/16',
4242 'GF': '161.22.64.0/18',
4243 'GG': '62.68.160.0/19',
4244 'GH': '154.160.0.0/12',
4245 'GI': '95.164.0.0/16',
4246 'GL': '88.83.0.0/19',
4247 'GM': '160.182.0.0/15',
4248 'GN': '197.149.192.0/18',
4249 'GP': '104.250.0.0/19',
4250 'GQ': '105.235.224.0/20',
4251 'GR': '94.64.0.0/13',
4252 'GT': '168.234.0.0/16',
4253 'GU': '168.123.0.0/16',
4254 'GW': '197.214.80.0/20',
4255 'GY': '181.41.64.0/18',
4256 'HK': '113.252.0.0/14',
4257 'HN': '181.210.0.0/16',
4258 'HR': '93.136.0.0/13',
4259 'HT': '148.102.128.0/17',
4260 'HU': '84.0.0.0/14',
4261 'ID': '39.192.0.0/10',
4262 'IE': '87.32.0.0/12',
4263 'IL': '79.176.0.0/13',
4264 'IM': '5.62.80.0/20',
4265 'IN': '117.192.0.0/10',
4266 'IO': '203.83.48.0/21',
4267 'IQ': '37.236.0.0/14',
4268 'IR': '2.176.0.0/12',
4269 'IS': '82.221.0.0/16',
4270 'IT': '79.0.0.0/10',
4271 'JE': '87.244.64.0/18',
4272 'JM': '72.27.0.0/17',
4273 'JO': '176.29.0.0/16',
4274 'JP': '133.0.0.0/8',
4275 'KE': '105.48.0.0/12',
4276 'KG': '158.181.128.0/17',
4277 'KH': '36.37.128.0/17',
4278 'KI': '103.25.140.0/22',
4279 'KM': '197.255.224.0/20',
4280 'KN': '198.167.192.0/19',
4281 'KP': '175.45.176.0/22',
4282 'KR': '175.192.0.0/10',
4283 'KW': '37.36.0.0/14',
4284 'KY': '64.96.0.0/15',
4285 'KZ': '2.72.0.0/13',
4286 'LA': '115.84.64.0/18',
4287 'LB': '178.135.0.0/16',
4288 'LC': '24.92.144.0/20',
4289 'LI': '82.117.0.0/19',
4290 'LK': '112.134.0.0/15',
4291 'LR': '102.183.0.0/16',
4292 'LS': '129.232.0.0/17',
4293 'LT': '78.56.0.0/13',
4294 'LU': '188.42.0.0/16',
4295 'LV': '46.109.0.0/16',
4296 'LY': '41.252.0.0/14',
4297 'MA': '105.128.0.0/11',
4298 'MC': '88.209.64.0/18',
4299 'MD': '37.246.0.0/16',
4300 'ME': '178.175.0.0/17',
4301 'MF': '74.112.232.0/21',
4302 'MG': '154.126.0.0/17',
4303 'MH': '117.103.88.0/21',
4304 'MK': '77.28.0.0/15',
4305 'ML': '154.118.128.0/18',
4306 'MM': '37.111.0.0/17',
4307 'MN': '49.0.128.0/17',
4308 'MO': '60.246.0.0/16',
4309 'MP': '202.88.64.0/20',
4310 'MQ': '109.203.224.0/19',
4311 'MR': '41.188.64.0/18',
4312 'MS': '208.90.112.0/22',
4313 'MT': '46.11.0.0/16',
4314 'MU': '105.16.0.0/12',
4315 'MV': '27.114.128.0/18',
4316 'MW': '102.70.0.0/15',
4317 'MX': '187.192.0.0/11',
4318 'MY': '175.136.0.0/13',
4319 'MZ': '197.218.0.0/15',
4320 'NA': '41.182.0.0/16',
4321 'NC': '101.101.0.0/18',
4322 'NE': '197.214.0.0/18',
4323 'NF': '203.17.240.0/22',
4324 'NG': '105.112.0.0/12',
4325 'NI': '186.76.0.0/15',
4326 'NL': '145.96.0.0/11',
4327 'NO': '84.208.0.0/13',
4328 'NP': '36.252.0.0/15',
4329 'NR': '203.98.224.0/19',
4330 'NU': '49.156.48.0/22',
4331 'NZ': '49.224.0.0/14',
4332 'OM': '5.36.0.0/15',
4333 'PA': '186.72.0.0/15',
4334 'PE': '186.160.0.0/14',
4335 'PF': '123.50.64.0/18',
4336 'PG': '124.240.192.0/19',
4337 'PH': '49.144.0.0/13',
4338 'PK': '39.32.0.0/11',
4339 'PL': '83.0.0.0/11',
4340 'PM': '70.36.0.0/20',
4341 'PR': '66.50.0.0/16',
4342 'PS': '188.161.0.0/16',
4343 'PT': '85.240.0.0/13',
4344 'PW': '202.124.224.0/20',
4345 'PY': '181.120.0.0/14',
4346 'QA': '37.210.0.0/15',
4347 'RE': '102.35.0.0/16',
4348 'RO': '79.112.0.0/13',
4349 'RS': '93.86.0.0/15',
4350 'RU': '5.136.0.0/13',
4351 'RW': '41.186.0.0/16',
4352 'SA': '188.48.0.0/13',
4353 'SB': '202.1.160.0/19',
4354 'SC': '154.192.0.0/11',
4355 'SD': '102.120.0.0/13',
4356 'SE': '78.64.0.0/12',
4357 'SG': '8.128.0.0/10',
4358 'SI': '188.196.0.0/14',
4359 'SK': '78.98.0.0/15',
4360 'SL': '102.143.0.0/17',
4361 'SM': '89.186.32.0/19',
4362 'SN': '41.82.0.0/15',
4363 'SO': '154.115.192.0/18',
4364 'SR': '186.179.128.0/17',
4365 'SS': '105.235.208.0/21',
4366 'ST': '197.159.160.0/19',
4367 'SV': '168.243.0.0/16',
4368 'SX': '190.102.0.0/20',
4370 'SZ': '41.84.224.0/19',
4371 'TC': '65.255.48.0/20',
4372 'TD': '154.68.128.0/19',
4373 'TG': '196.168.0.0/14',
4374 'TH': '171.96.0.0/13',
4375 'TJ': '85.9.128.0/18',
4376 'TK': '27.96.24.0/21',
4377 'TL': '180.189.160.0/20',
4378 'TM': '95.85.96.0/19',
4379 'TN': '197.0.0.0/11',
4380 'TO': '175.176.144.0/21',
4381 'TR': '78.160.0.0/11',
4382 'TT': '186.44.0.0/15',
4383 'TV': '202.2.96.0/19',
4384 'TW': '120.96.0.0/11',
4385 'TZ': '156.156.0.0/14',
4386 'UA': '37.52.0.0/14',
4387 'UG': '102.80.0.0/13',
4389 'UY': '167.56.0.0/13',
4390 'UZ': '84.54.64.0/18',
4391 'VA': '212.77.0.0/19',
4392 'VC': '207.191.240.0/21',
4393 'VE': '186.88.0.0/13',
4394 'VG': '66.81.192.0/20',
4395 'VI': '146.226.0.0/16',
4396 'VN': '14.160.0.0/11',
4397 'VU': '202.80.32.0/20',
4398 'WF': '117.20.32.0/21',
4399 'WS': '202.4.32.0/19',
4400 'YE': '134.35.0.0/16',
4401 'YT': '41.242.116.0/22',
4402 'ZA': '41.0.0.0/11',
4403 'ZM': '102.144.0.0/13',
4404 'ZW': '102.177.192.0/18',
4408 def random_ipv4(cls, code_or_block):
4409 if len(code_or_block) == 2:
4410 block = cls._country_ip_map.get(code_or_block.upper())
4414 block = code_or_block
4415 addr, preflen = block.split('/')
4416 addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
4417 addr_max = addr_min | (0xffffffff >> int(preflen))
4418 return compat_str(socket.inet_ntoa(
4419 compat_struct_pack('!L', random.randint(addr_min, addr_max))))
4422 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
4423 def __init__(self, proxies=None):
4424 # Set default handlers
4425 for type in ('http', 'https'):
4426 setattr(self, '%s_open' % type,
4427 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4428 meth(r, proxy, type))
4429 compat_urllib_request.ProxyHandler.__init__(self, proxies)
4431 def proxy_open(self, req, proxy, type):
4432 req_proxy = req.headers.get('Ytdl-request-proxy')
4433 if req_proxy is not None:
4435 del req.headers['Ytdl-request-proxy']
4437 if proxy == '__noproxy__':
4438 return None # No Proxy
4439 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4440 req.add_header('Ytdl-socks-proxy', proxy)
4441 # yt-dlp's http/https handlers do wrapping the socket with socks
4443 return compat_urllib_request.ProxyHandler.proxy_open(
4444 self, req, proxy, type)
4447 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4448 # released into Public Domain
4449 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4451 def long_to_bytes(n, blocksize=0):
4452 """long_to_bytes(n:long, blocksize:int) : string
4453 Convert a long integer to a byte string.
4455 If optional blocksize is given and greater than zero, pad the front of the
4456 byte string with binary zeros so that the length is a multiple of
4459 # after much testing, this algorithm was deemed to be the fastest
4463 s = compat_struct_pack('>I', n & 0xffffffff) + s
4465 # strip off leading zeros
4466 for i in range(len(s)):
4467 if s[i] != b'\000'[0]:
4470 # only happens when n == 0
4474 # add back some pad bytes. this could be done more efficiently w.r.t. the
4475 # de-padding being done above, but sigh...
4476 if blocksize > 0 and len(s) % blocksize:
4477 s = (blocksize - len(s) % blocksize) * b'\000' + s
4481 def bytes_to_long(s):
4482 """bytes_to_long(string) : long
4483 Convert a byte string to a long integer.
4485 This is (essentially) the inverse of long_to_bytes().
4490 extra = (4 - length % 4)
4491 s = b'\000' * extra + s
4492 length = length + extra
4493 for i in range(0, length, 4):
4494 acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
4498 def ohdave_rsa_encrypt(data, exponent, modulus):
4500 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4503 data: data to encrypt, bytes-like object
4504 exponent, modulus: parameter e and N of RSA algorithm, both integer
4505 Output: hex string of encrypted data
4507 Limitation: supports one block encryption only
4510 payload = int(binascii.hexlify(data[::-1]), 16)
4511 encrypted = pow(payload, exponent, modulus)
4512 return '%x' % encrypted
4515 def pkcs1pad(data, length):
4517 Padding input data with PKCS#1 scheme
4519 @param {int[]} data input data
4520 @param {int} length target length
4521 @returns {int[]} padded data
4523 if len(data) > length - 11:
4524 raise ValueError('Input data too
long for PKCS
#1 padding')
4526 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4527 return [0, 2] + pseudo_random
+ [0] + data
4530 def encode_base_n(num
, n
, table
=None):
4531 FULL_TABLE
= '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
4533 table
= FULL_TABLE
[:n
]
4536 raise ValueError('base %d exceeds table length %d' % (n
, len(table
)))
4543 ret
= table
[num
% n
] + ret
4548 def decode_packed_codes(code
):
4549 mobj
= re
.search(PACKED_CODES_RE
, code
)
4550 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4553 symbols
= symbols
.split('|')
4558 base_n_count
= encode_base_n(count
, base
)
4559 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4562 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4566 def caesar(s
, alphabet
, shift
):
4571 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4576 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4579 def parse_m3u8_attributes(attrib
):
4581 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4582 if val
.startswith('"'):
4588 def urshift(val
, n
):
4589 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4592 # Based on png2str() written by @gdkchan and improved by @yokrysty
4593 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4594 def decode_png(png_data
):
4595 # Reference: https://www.w3.org/TR/PNG/
4596 header
= png_data
[8:]
4598 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4599 raise OSError('Not a valid PNG file.')
4601 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4602 unpack_integer
= lambda x
: compat_struct_unpack(int_map
[len(x
)], x
)[0]
4607 length
= unpack_integer(header
[:4])
4610 chunk_type
= header
[:4]
4613 chunk_data
= header
[:length
]
4614 header
= header
[length
:]
4616 header
= header
[4:] # Skip CRC
4624 ihdr
= chunks
[0]['data']
4626 width
= unpack_integer(ihdr
[:4])
4627 height
= unpack_integer(ihdr
[4:8])
4631 for chunk
in chunks
:
4632 if chunk
['type'] == b
'IDAT':
4633 idat
+= chunk
['data']
4636 raise OSError('Unable to read PNG data.')
4638 decompressed_data
= bytearray(zlib
.decompress(idat
))
4643 def _get_pixel(idx
):
4648 for y
in range(height
):
4649 basePos
= y
* (1 + stride
)
4650 filter_type
= decompressed_data
[basePos
]
4654 pixels
.append(current_row
)
4656 for x
in range(stride
):
4657 color
= decompressed_data
[1 + basePos
+ x
]
4658 basex
= y
* stride
+ x
4663 left
= _get_pixel(basex
- 3)
4665 up
= _get_pixel(basex
- stride
)
4667 if filter_type
== 1: # Sub
4668 color
= (color
+ left
) & 0xff
4669 elif filter_type
== 2: # Up
4670 color
= (color
+ up
) & 0xff
4671 elif filter_type
== 3: # Average
4672 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4673 elif filter_type
== 4: # Paeth
4679 c
= _get_pixel(basex
- stride
- 3)
4687 if pa
<= pb
and pa
<= pc
:
4688 color
= (color
+ a
) & 0xff
4690 color
= (color
+ b
) & 0xff
4692 color
= (color
+ c
) & 0xff
4694 current_row
.append(color
)
4696 return width
, height
, pixels
4699 def write_xattr(path
, key
, value
):
4700 # Windows: Write xattrs to NTFS Alternate Data Streams:
4701 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4702 if compat_os_name
== 'nt':
4703 assert ':' not in key
4704 assert os
.path
.exists(path
)
4707 with open(f
'{path}:{key}', 'wb') as f
:
4709 except OSError as e
:
4710 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4713 # UNIX Method 1. Use xattrs/pyxattrs modules
4714 from .dependencies
import xattr
4717 if getattr(xattr
, '_yt_dlp__identifier', None) == 'pyxattr':
4718 # Unicode arguments are not supported in pyxattr until version 0.5.0
4719 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4720 if version_tuple(xattr
.__version
__) >= (0, 5, 0):
4721 setxattr
= xattr
.set
4723 setxattr
= xattr
.setxattr
4727 setxattr(path
, key
, value
)
4728 except OSError as e
:
4729 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4732 # UNIX Method 2. Use setfattr/xattr executables
4733 exe
= ('setfattr' if check_executable('setfattr', ['--version'])
4734 else 'xattr' if check_executable('xattr', ['-h']) else None)
4736 raise XAttrUnavailableError(
4737 'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
4738 + ('"xattr" binary' if sys
.platform
!= 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
4740 value
= value
.decode('utf-8')
4743 [exe
, '-w', key
, value
, path
] if exe
== 'xattr' else [exe
, '-n', key
, '-v', value
, path
],
4744 stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4745 except OSError as e
:
4746 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4747 stderr
= p
.communicate_or_kill()[1].decode('utf-8', 'replace')
4749 raise XAttrMetadataError(p
.returncode
, stderr
)
4752 def random_birthday(year_field
, month_field
, day_field
):
4753 start_date
= datetime
.date(1950, 1, 1)
4754 end_date
= datetime
.date(1995, 12, 31)
4755 offset
= random
.randint(0, (end_date
- start_date
).days
)
4756 random_date
= start_date
+ datetime
.timedelta(offset
)
4758 year_field
: str(random_date
.year
),
4759 month_field
: str(random_date
.month
),
4760 day_field
: str(random_date
.day
),
4764 # Templates for internet shortcut files, which are plain text files.
4765 DOT_URL_LINK_TEMPLATE
= '''\
4770 DOT_WEBLOC_LINK_TEMPLATE
= '''\
4771 <?xml version="1.0" encoding="UTF-8"?>
4772 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
4773 <plist version="1.0">
4776 \t<string>%(url)s</string>
4781 DOT_DESKTOP_LINK_TEMPLATE
= '''\
4791 'url': DOT_URL_LINK_TEMPLATE
,
4792 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
4793 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
4797 def iri_to_uri(iri
):
4799 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
4801 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.
4804 iri_parts
= compat_urllib_parse_urlparse(iri
)
4806 if '[' in iri_parts
.netloc
:
4807 raise ValueError('IPv6 URIs are not, yet, supported.')
4808 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
4810 # 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.
4813 if iri_parts
.username
:
4814 net_location
+= urllib
.parse
.quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
4815 if iri_parts
.password
is not None:
4816 net_location
+= ':' + urllib
.parse
.quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
4819 net_location
+= iri_parts
.hostname
.encode('idna').decode('utf-8') # Punycode for Unicode hostnames.
4820 # The 'idna' encoding produces ASCII text.
4821 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
4822 net_location
+= ':' + str(iri_parts
.port
)
4824 return urllib
.parse
.urlunparse(
4828 urllib
.parse
.quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
4830 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
4831 urllib
.parse
.quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
4833 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
4834 urllib
.parse
.quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
4836 urllib
.parse
.quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
4838 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
4841 def to_high_limit_path(path
):
4842 if sys
.platform
in ['win32', 'cygwin']:
4843 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
4844 return '\\\\?\\' + os
.path
.abspath(path
)
4849 def format_field(obj
, field
=None, template
='%s', ignore
=(None, ''), default
='', func
=None):
4850 val
= traverse_obj(obj
, *variadic(field
))
4853 return template
% (func(val
) if func
else val
)
4856 def clean_podcast_url(url
):
4857 return re
.sub(r
'''(?x)
4861 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
4864 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
4867 cn\.co| # https://podcorn.com/analytics-prefix/
4868 st\.fm # https://podsights.com/docs/
4873 _HEX_TABLE
= '0123456789abcdef'
4876 def random_uuidv4():
4877 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
4880 def make_dir(path
, to_screen
=None):
4882 dn
= os
.path
.dirname(path
)
4883 if dn
and not os
.path
.exists(dn
):
4886 except OSError as err
:
4887 if callable(to_screen
) is not None:
4888 to_screen('unable to create directory ' + error_to_compat_str(err
))
4892 def get_executable_path():
4893 from zipimport
import zipimporter
4894 if hasattr(sys
, 'frozen'): # Running from PyInstaller
4895 path
= os
.path
.dirname(sys
.executable
)
4896 elif isinstance(__loader__
, zipimporter
): # Running from ZIP
4897 path
= os
.path
.join(os
.path
.dirname(__file__
), '../..')
4899 path
= os
.path
.join(os
.path
.dirname(__file__
), '..')
4900 return os
.path
.abspath(path
)
4903 def load_plugins(name
, suffix
, namespace
):
4905 with contextlib
.suppress(FileNotFoundError
):
4906 plugins_spec
= importlib
.util
.spec_from_file_location(
4907 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
4908 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
4909 sys
.modules
[plugins_spec
.name
] = plugins
4910 plugins_spec
.loader
.exec_module(plugins
)
4911 for name
in dir(plugins
):
4912 if name
in namespace
:
4914 if not name
.endswith(suffix
):
4916 klass
= getattr(plugins
, name
)
4917 classes
[name
] = namespace
[name
] = klass
4922 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
4923 casesense
=True, is_user_input
=False, traverse_string
=False):
4924 ''' Traverse nested list/dict/tuple
4925 @param path_list A list of paths which are checked one by one.
4926 Each path is a list of keys where each key is a:
4928 - string: A dictionary key
4929 - int: An index into a list
4930 - tuple: A list of keys all of which will be traversed
4931 - Ellipsis: Fetch all values in the object
4932 - Function: Takes the key and value as arguments
4933 and returns whether the key matches or not
4934 @param default Default value to return
4935 @param expected_type Only accept final value of this type (Can also be any callable)
4936 @param get_all Return all the values obtained from a path or only the first one
4937 @param casesense Whether to consider dictionary keys as case sensitive
4938 @param is_user_input Whether the keys are generated from user input. If True,
4939 strings are converted to int/slice if necessary
4940 @param traverse_string Whether to traverse inside strings. If True, any
4941 non-compatible object will also be converted into a string
4945 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
4946 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
4948 def _traverse_obj(obj
, path
, _current_depth
=0):
4950 path
= tuple(variadic(path
))
4951 for i
, key
in enumerate(path
):
4952 if None in (key
, obj
):
4954 if isinstance(key
, (list, tuple)):
4955 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
4958 obj
= (obj
.values() if isinstance(obj
, dict)
4959 else obj
if isinstance(obj
, (list, tuple, LazyList
))
4960 else str(obj
) if traverse_string
else [])
4962 depth
= max(depth
, _current_depth
)
4963 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
4965 if isinstance(obj
, (list, tuple, LazyList
)):
4966 obj
= enumerate(obj
)
4967 elif isinstance(obj
, dict):
4970 if not traverse_string
:
4974 depth
= max(depth
, _current_depth
)
4975 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if try_call(key
, args
=(k
, v
))]
4976 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
4977 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
4978 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
4981 key
= (int_or_none(key
) if ':' not in key
4982 else slice(*map(int_or_none
, key
.split(':'))))
4983 if key
== slice(None):
4984 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
4985 if not isinstance(key
, (int, slice)):
4987 if not isinstance(obj
, (list, tuple, LazyList
)):
4988 if not traverse_string
:
4997 if isinstance(expected_type
, type):
4998 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
4999 elif expected_type
is not None:
5000 type_test
= expected_type
5002 type_test
= lambda val
: val
5004 for path
in path_list
:
5006 val
= _traverse_obj(obj
, path
)
5009 for _
in range(depth
- 1):
5010 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5011 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5013 return val
if get_all
else val
[0]
5015 val
= type_test(val
)
5021 def traverse_dict(dictn
, keys
, casesense
=True):
5022 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5023 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5024 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5027 def get_first(obj
, keys
, **kwargs
):
5028 return traverse_obj(obj
, (..., *variadic(keys
)), **kwargs
, get_all
=False)
5031 def variadic(x
, allowed_types
=(str, bytes, dict)):
5032 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5035 def decode_base(value
, digits
):
5036 # This will convert given base-x string to scalar (long or int)
5037 table
= {char: index for index, char in enumerate(digits)}
5042 result
+= table
[chr]
5046 def time_seconds(**kwargs
):
5047 t
= datetime
.datetime
.now(datetime
.timezone(datetime
.timedelta(**kwargs
)))
5048 return t
.timestamp()
5051 # create a JSON Web Signature (jws) with HS256 algorithm
5052 # the resulting format is in JWS Compact Serialization
5053 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5054 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5055 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5061 header_data
.update(headers
)
5062 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode('utf-8'))
5063 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode('utf-8'))
5064 h
= hmac
.new(key
.encode('utf-8'), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5065 signature_b64
= base64
.b64encode(h
.digest())
5066 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5070 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5071 def jwt_decode_hs256(jwt
):
5072 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5073 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5077 def supports_terminal_sequences(stream
):
5078 if compat_os_name
== 'nt':
5079 from .compat
import WINDOWS_VT_MODE
# Must be imported locally
5080 if not WINDOWS_VT_MODE
or get_windows_version() < (10, 0, 10586):
5082 elif not os
.getenv('TERM'):
5085 return stream
.isatty()
5086 except BaseException
:
5090 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5093 def remove_terminal_sequences(string
):
5094 return _terminal_sequences_re
.sub('', string
)
5097 def number_of_digits(number
):
5098 return len('%d' % number
)
5101 def join_nonempty(*values
, delim
='-', from_dict
=None):
5102 if from_dict
is not None:
5103 values
= map(from_dict
.get
, values
)
5104 return delim
.join(map(str, filter(None, values
)))
5107 def scale_thumbnails_to_max_format_width(formats
, thumbnails
, url_width_re
):
5109 Find the largest format dimensions in terms of video width and, for each thumbnail:
5110 * Modify the URL: Match the width with the provided regex and replace with the former width
5113 This function is useful with video services that scale the provided thumbnails on demand
5115 _keys
= ('width', 'height')
5116 max_dimensions
= max(
5117 (tuple(format
.get(k
) or 0 for k
in _keys
) for format
in formats
),
5119 if not max_dimensions
[0]:
5123 {'url': re.sub(url_width_re, str(max_dimensions[0]), thumbnail['url'])}
,
5124 dict(zip(_keys
, max_dimensions
)), thumbnail
)
5125 for thumbnail
in thumbnails
5129 def parse_http_range(range):
5130 """ Parse value of "Range" or "Content-Range" HTTP header into tuple. """
5132 return None, None, None
5133 crg
= re
.search(r
'bytes[ =](\d+)-(\d+)?(?:/(\d+))?', range)
5135 return None, None, None
5136 return int(crg
.group(1)), int_or_none(crg
.group(2)), int_or_none(crg
.group(3))
5142 __initialized
= False
5144 def __init__(self
, parser
, label
=None):
5145 self
._parser
, self
.label
= parser
, label
5146 self
._loaded
_paths
, self
.configs
= set(), []
5148 def init(self
, args
=None, filename
=None):
5149 assert not self
.__initialized
5152 location
= os
.path
.realpath(filename
)
5153 directory
= os
.path
.dirname(location
)
5154 if location
in self
._loaded
_paths
:
5156 self
._loaded
_paths
.add(location
)
5158 self
.__initialized
= True
5159 self
.own_args
, self
.filename
= args
, filename
5160 for location
in self
._parser
.parse_args(args
)[0].config_locations
or []:
5161 location
= os
.path
.join(directory
, expand_path(location
))
5162 if os
.path
.isdir(location
):
5163 location
= os
.path
.join(location
, 'yt-dlp.conf')
5164 if not os
.path
.exists(location
):
5165 self
._parser
.error(f
'config location {location} does not exist')
5166 self
.append_config(self
.read_file(location
), location
)
5170 label
= join_nonempty(
5171 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5173 return join_nonempty(
5174 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5175 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5179 def read_file(filename
, default
=[]):
5181 optionf
= open(filename
)
5183 return default
# silently skip if file is not present
5185 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5186 contents
= optionf
.read()
5187 res
= shlex
.split(contents
, comments
=True)
5193 def hide_login_info(opts
):
5194 PRIVATE_OPTS
= {'-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'}
5195 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5200 return m
.group('key') + '=PRIVATE'
5204 opts
= list(map(_scrub_eq
, opts
))
5205 for idx
, opt
in enumerate(opts
):
5206 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5207 opts
[idx
+ 1] = 'PRIVATE'
5210 def append_config(self
, *args
, label
=None):
5211 config
= type(self
)(self
._parser
, label
)
5212 config
._loaded
_paths
= self
._loaded
_paths
5213 if config
.init(*args
):
5214 self
.configs
.append(config
)
5218 for config
in reversed(self
.configs
):
5219 yield from config
.all_args
5220 yield from self
.own_args
or []
5222 def parse_args(self
):
5223 return self
._parser
.parse_args(self
.all_args
)
5226 class WebSocketsWrapper():
5227 """Wraps websockets module to use in non-async scopes"""
5230 def __init__(self
, url
, headers
=None, connect
=True):
5231 self
.loop
= asyncio
.new_event_loop()
5232 # XXX: "loop" is deprecated
5233 self
.conn
= websockets
.connect(
5234 url
, extra_headers
=headers
, ping_interval
=None,
5235 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5238 atexit
.register(self
.__exit
__, None, None, None)
5240 def __enter__(self
):
5242 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5245 def send(self
, *args
):
5246 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5248 def recv(self
, *args
):
5249 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5251 def __exit__(self
, type, value
, traceback
):
5253 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5256 self
._cancel
_all
_tasks
(self
.loop
)
5258 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5259 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5261 def run_with_loop(main
, loop
):
5262 if not asyncio
.iscoroutine(main
):
5263 raise ValueError(f
'a coroutine was expected, got {main!r}')
5266 return loop
.run_until_complete(main
)
5268 loop
.run_until_complete(loop
.shutdown_asyncgens())
5269 if hasattr(loop
, 'shutdown_default_executor'):
5270 loop
.run_until_complete(loop
.shutdown_default_executor())
5273 def _cancel_all_tasks(loop
):
5274 to_cancel
= asyncio
.all_tasks(loop
)
5279 for task
in to_cancel
:
5282 # XXX: "loop" is removed in python 3.10+
5283 loop
.run_until_complete(
5284 asyncio
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5286 for task
in to_cancel
:
5287 if task
.cancelled():
5289 if task
.exception() is not None:
5290 loop
.call_exception_handler({
5291 'message': 'unhandled exception during asyncio.run() shutdown',
5292 'exception': task
.exception(),
5297 def merge_headers(*dicts
):
5298 """Merge dicts of http headers case insensitively, prioritizing the latter ones"""
5299 return {k.title(): v for k, v in itertools.chain.from_iterable(map(dict.items, dicts))}
5302 class classproperty
:
5303 def __init__(self
, f
):
5306 def __get__(self
, _
, cls
):
5310 def Namespace(**kwargs
):
5311 return collections
.namedtuple('Namespace', kwargs
)(**kwargs
)
5315 has_certifi
= bool(certifi
)
5316 has_websockets
= bool(websockets
)