4 from __future__
import unicode_literals
40 import xml
.etree
.ElementTree
45 compat_HTMLParseError
,
51 compat_ctypes_WINFUNCTYPE
,
52 compat_etree_fromstring
,
55 compat_html_entities_html5
,
69 compat_urllib_parse_urlencode
,
70 compat_urllib_parse_urlparse
,
71 compat_urllib_parse_urlunparse
,
72 compat_urllib_parse_quote
,
73 compat_urllib_parse_quote_plus
,
74 compat_urllib_parse_unquote_plus
,
75 compat_urllib_request
,
87 def register_socks_protocols():
88 # "Register" SOCKS protocols
89 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
90 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
91 for scheme
in ('socks', 'socks4', 'socks4a', 'socks5'):
92 if scheme
not in compat_urlparse
.uses_netloc
:
93 compat_urlparse
.uses_netloc
.append(scheme
)
96 # This is not clearly defined otherwise
97 compiled_regex_type
= type(re
.compile(''))
100 def random_user_agent():
101 _USER_AGENT_TPL
= 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/%s Safari/537.36'
142 return _USER_AGENT_TPL
% random
.choice(_CHROME_VERSIONS
)
146 'User-Agent': random_user_agent(),
147 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
148 'Accept-Encoding': 'gzip, deflate',
149 'Accept-Language': 'en-us,en;q=0.5',
150 'Sec-Fetch-Mode': 'navigate',
155 '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',
159 NO_DEFAULT
= object()
161 ENGLISH_MONTH_NAMES
= [
162 'January', 'February', 'March', 'April', 'May', 'June',
163 'July', 'August', 'September', 'October', 'November', 'December']
166 'en': ENGLISH_MONTH_NAMES
,
168 'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
169 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
173 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
174 'flv', 'f4v', 'f4a', 'f4b',
175 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
176 'mkv', 'mka', 'mk3d',
185 'f4f', 'f4m', 'm3u8', 'smil')
187 # needed for sanitizing filenames in restricted mode
188 ACCENT_CHARS
= dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
189 itertools
.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
190 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
220 '%Y-%m-%d %H:%M:%S.%f',
221 '%Y-%m-%d %H:%M:%S:%f',
224 '%Y-%m-%dT%H:%M:%SZ',
225 '%Y-%m-%dT%H:%M:%S.%fZ',
226 '%Y-%m-%dT%H:%M:%S.%f0Z',
228 '%Y-%m-%dT%H:%M:%S.%f',
231 '%b %d %Y at %H:%M:%S',
233 '%B %d %Y at %H:%M:%S',
237 DATE_FORMATS_DAY_FIRST
= list(DATE_FORMATS
)
238 DATE_FORMATS_DAY_FIRST
.extend([
247 DATE_FORMATS_MONTH_FIRST
= list(DATE_FORMATS
)
248 DATE_FORMATS_MONTH_FIRST
.extend([
256 PACKED_CODES_RE
= r
"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
257 JSON_LD_RE
= r
'(?is)<script[^>]+type=(["\']?
)application
/ld\
+json\
1[^
>]*>(?P
<json_ld
>.+?
)</script
>'
260 def preferredencoding():
261 """Get preferred encoding.
263 Returns the best encoding scheme for the system, based on
264 locale.getpreferredencoding() and some further tweaks.
267 pref = locale.getpreferredencoding()
275 def write_json_file(obj, fn):
276 """ Encode obj as JSON and write it to fn, atomically if possible """
278 fn = encodeFilename(fn)
279 if sys.version_info < (3, 0) and sys.platform != 'win32
':
280 encoding = get_filesystem_encoding()
281 # os.path.basename returns a bytes object, but NamedTemporaryFile
282 # will fail if the filename contains non ascii characters unless we
283 # use a unicode object
284 path_basename = lambda f: os.path.basename(fn).decode(encoding)
285 # the same for os.path.dirname
286 path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
288 path_basename = os.path.basename
289 path_dirname = os.path.dirname
293 'prefix
': path_basename(fn) + '.',
294 'dir': path_dirname(fn),
298 # In Python 2.x, json.dump expects a bytestream.
299 # In Python 3.x, it writes to a character stream
300 if sys.version_info < (3, 0):
308 tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
312 json.dump(obj, tf, ensure_ascii=False)
313 if sys.platform == 'win32
':
314 # Need to remove existing file on Windows, else os.rename raises
315 # WindowsError or FileExistsError.
323 os.chmod(tf.name, 0o666 & ~mask)
326 os.rename(tf.name, fn)
335 if sys.version_info >= (2, 7):
336 def find_xpath_attr(node, xpath, key, val=None):
337 """ Find the xpath xpath[@key=val] """
338 assert re.match(r'^
[a
-zA
-Z_
-]+$
', key)
339 expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
340 return node.find(expr)
342 def find_xpath_attr(node, xpath, key, val=None):
343 for f in node.findall(compat_xpath(xpath)):
344 if key not in f.attrib:
346 if val is None or f.attrib.get(key) == val:
350 # On python2.6 the xml.etree.ElementTree.Element methods don't support
351 # the namespace parameter
354 def xpath_with_ns(path
, ns_map
):
355 components
= [c
.split(':') for c
in path
.split('/')]
359 replaced
.append(c
[0])
362 replaced
.append('{%s}%s' % (ns_map
[ns
], tag
))
363 return '/'.join(replaced
)
366 def xpath_element(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
367 def _find_xpath(xpath
):
368 return node
.find(compat_xpath(xpath
))
370 if isinstance(xpath
, (str, compat_str
)):
371 n
= _find_xpath(xpath
)
379 if default
is not NO_DEFAULT
:
382 name
= xpath
if name
is None else name
383 raise ExtractorError('Could not find XML element %s' % name
)
389 def xpath_text(node
, xpath
, name
=None, fatal
=False, default
=NO_DEFAULT
):
390 n
= xpath_element(node
, xpath
, name
, fatal
=fatal
, default
=default
)
391 if n
is None or n
== default
:
394 if default
is not NO_DEFAULT
:
397 name
= xpath
if name
is None else name
398 raise ExtractorError('Could not find XML element\'s text %s' % name
)
404 def xpath_attr(node
, xpath
, key
, name
=None, fatal
=False, default
=NO_DEFAULT
):
405 n
= find_xpath_attr(node
, xpath
, key
)
407 if default
is not NO_DEFAULT
:
410 name
= '%s[@%s]' % (xpath
, key
) if name
is None else name
411 raise ExtractorError('Could not find XML attribute %s' % name
)
417 def get_element_by_id(id, html
):
418 """Return the content of the tag with the specified ID in the passed HTML document"""
419 return get_element_by_attribute('id', id, html
)
422 def get_element_html_by_id(id, html
):
423 """Return the html of the tag with the specified ID in the passed HTML document"""
424 return get_element_html_by_attribute('id', id, html
)
427 def get_element_by_class(class_name
, html
):
428 """Return the content of the first tag with the specified class in the passed HTML document"""
429 retval
= get_elements_by_class(class_name
, html
)
430 return retval
[0] if retval
else None
433 def get_element_html_by_class(class_name
, html
):
434 """Return the html of the first tag with the specified class in the passed HTML document"""
435 retval
= get_elements_html_by_class(class_name
, html
)
436 return retval
[0] if retval
else None
439 def get_element_by_attribute(attribute
, value
, html
, escape_value
=True):
440 retval
= get_elements_by_attribute(attribute
, value
, html
, escape_value
)
441 return retval
[0] if retval
else None
444 def get_element_html_by_attribute(attribute
, value
, html
, escape_value
=True):
445 retval
= get_elements_html_by_attribute(attribute
, value
, html
, escape_value
)
446 return retval
[0] if retval
else None
449 def get_elements_by_class(class_name
, html
):
450 """Return the content of all tags with the specified class in the passed HTML document as a list"""
451 return get_elements_by_attribute(
452 'class', r
'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
453 html, escape_value=False)
456 def get_elements_html_by_class(class_name, html):
457 """Return the html of all tags with the specified class in the passed HTML document as a list"""
458 return get_elements_html_by_attribute(
459 'class', r'[^
\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
460 html, escape_value=False)
463 def get_elements_by_attribute(*args, **kwargs):
464 """Return the content of the tag with the specified attribute in the passed HTML document"""
465 return [content for content, _ in get_elements_text_and_html_by_attribute(*args, **kwargs)]
468 def get_elements_html_by_attribute(*args, **kwargs):
469 """Return the html of the tag with the specified attribute in the passed HTML document"""
470 return [whole for _, whole in get_elements_text_and_html_by_attribute(*args, **kwargs)]
473 def get_elements_text_and_html_by_attribute(attribute, value, html, escape_value=True):
475 Return the text (content) and the html (whole) of the tag with the specified
476 attribute in the passed HTML document
479 value_quote_optional = '' if re.match(r'''[\s"'`
=<>]''', value) else '?'
481 value = re.escape(value) if escape_value else value
483 partial_element_re = r'''(?x
)
484 <(?P
<tag
>[a
-zA
-Z0
-9:._-]+)
485 (?
:\
s(?
:[^
>"']|"[^
"]*"|
'[^']*')*)?
486 \s%(attribute)s\s*=\s*(?P<_q>['"]%(vqo)s)(?-x:%(value)s)(?P=_q)
487 ''' % {'attribute': re.escape(attribute), 'value': value, 'vqo': value_quote_optional}
489 for m in re.finditer(partial_element_re, html):
490 content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
493 unescapeHTML(re.sub(r'^(?P<q>["\'])(?P
<content
>.*)(?P
=q
)$
', r'\g
<content
>', content, flags=re.DOTALL)),
498 class HTMLBreakOnClosingTagParser(compat_HTMLParser):
500 HTML parser which raises HTMLBreakOnClosingTagException upon reaching the
501 closing tag for the first opening tag it has encountered, and can be used
505 class HTMLBreakOnClosingTagException(Exception):
509 self.tagstack = collections.deque()
510 compat_HTMLParser.__init__(self)
515 def __exit__(self, *_):
519 # handle_endtag does not return upon raising HTMLBreakOnClosingTagException,
520 # so data remains buffered; we no longer have any interest in it, thus
521 # override this method to discard it
524 def handle_starttag(self, tag, _):
525 self.tagstack.append(tag)
527 def handle_endtag(self, tag):
528 if not self.tagstack:
529 raise compat_HTMLParseError('no tags
in the stack
')
531 inner_tag = self.tagstack.pop()
535 raise compat_HTMLParseError(f'matching opening tag
for closing {tag} tag
not found
')
536 if not self.tagstack:
537 raise self.HTMLBreakOnClosingTagException()
540 def get_element_text_and_html_by_tag(tag, html):
542 For the first element with the specified tag in the passed HTML document
543 return its' content (text
) and the whole
element (html
)
545 def find_or_raise(haystack, needle, exc):
547 return haystack.index(needle)
550 closing_tag = f'</{tag}>'
551 whole_start = find_or_raise(
552 html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
553 content_start = find_or_raise(
554 html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
555 content_start += whole_start + 1
556 with HTMLBreakOnClosingTagParser() as parser:
557 parser.feed(html[whole_start:content_start])
558 if not parser.tagstack or parser.tagstack[0] != tag:
559 raise compat_HTMLParseError(f'parser did not match opening {tag} tag')
560 offset = content_start
561 while offset < len(html):
562 next_closing_tag_start = find_or_raise(
563 html[offset:], closing_tag,
564 compat_HTMLParseError(f'closing {tag} tag not found'))
565 next_closing_tag_end = next_closing_tag_start + len(closing_tag)
567 parser.feed(html[offset:offset + next_closing_tag_end])
568 offset += next_closing_tag_end
569 except HTMLBreakOnClosingTagParser.HTMLBreakOnClosingTagException:
570 return html[content_start:offset + next_closing_tag_start], \
571 html[whole_start:offset + next_closing_tag_end]
572 raise compat_HTMLParseError('unexpected end of html')
575 class HTMLAttributeParser(compat_HTMLParser):
576 """Trivial HTML parser to gather the attributes
for a single element
"""
580 compat_HTMLParser.__init__(self)
582 def handle_starttag(self, tag, attrs):
583 self.attrs = dict(attrs)
586 class HTMLListAttrsParser(compat_HTMLParser):
587 """HTML parser to gather the attributes
for the elements of a
list"""
590 compat_HTMLParser.__init__(self)
594 def handle_starttag(self, tag, attrs):
595 if tag == 'li' and self._level == 0:
596 self.items.append(dict(attrs))
599 def handle_endtag(self, tag):
603 def extract_attributes(html_element):
604 """Given a string
for an HTML element such
as
606 a
="foo" B
="bar" c
="&98;az" d
=boz
607 empty
= noval entity
="&"
610 Decode
and return a dictionary of attributes
.
612 'a': 'foo', 'b': 'bar', c
: 'baz', d
: 'boz',
613 'empty': '', 'noval': None, 'entity': '&',
614 'sq': '"', 'dq': '\''
616 NB HTMLParser
is stricter
in Python
2.6 & 3.2 than
in later versions
,
617 but the cases
in the unit test will work
for all of
2.6, 2.7, 3.2-3.5.
619 parser = HTMLAttributeParser()
621 parser.feed(html_element)
623 # Older Python may throw HTMLParseError in case of malformed HTML
624 except compat_HTMLParseError:
629 def parse_list(webpage):
630 """Given a string
for an series of HTML
<li
> elements
,
631 return a dictionary of their attributes
"""
632 parser = HTMLListAttrsParser()
638 def clean_html(html):
639 """Clean an HTML snippet into a readable string
"""
641 if html is None: # Convenience for sanitizing descriptions etc.
644 html = re.sub(r'\s+', ' ', html)
645 html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
646 html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
648 html = re.sub('<.*?>', '', html)
649 # Replace html entities
650 html = unescapeHTML(html)
654 def sanitize_open(filename, open_mode):
655 """Try to
open the given filename
, and slightly tweak it
if this fails
.
657 Attempts to
open the given filename
. If this fails
, it tries to change
658 the filename slightly
, step by step
, until it
's either able to open it
659 or it fails and raises a final exception, like the standard open()
662 It returns the tuple (stream, definitive_file_name).
666 if sys.platform == 'win32
':
668 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
669 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
670 stream = locked_file(filename, open_mode, block=False).open()
671 return (stream, filename)
672 except (IOError, OSError) as err:
673 if err.errno in (errno.EACCES,):
676 # In case of error, try to remove win32 forbidden chars
677 alt_filename = sanitize_path(filename)
678 if alt_filename == filename:
681 # An exception here should be caught in the caller
682 stream = locked_file(filename, open_mode, block=False).open()
683 return (stream, alt_filename)
686 def timeconvert(timestr):
687 """Convert RFC 2822 defined time string into system timestamp"""
689 timetuple = email.utils.parsedate_tz(timestr)
690 if timetuple is not None:
691 timestamp = email.utils.mktime_tz(timetuple)
695 def sanitize_filename(s, restricted=False, is_id=False):
696 """Sanitizes a string so it could be used as part of a filename.
697 If restricted is set, use a stricter subset of allowed characters.
698 Set is_id if this is not an arbitrary string, but an ID that should be kept
701 def replace_insane(char):
702 if restricted and char in ACCENT_CHARS:
703 return ACCENT_CHARS[char]
704 elif not restricted and char == '\n':
706 elif char == '?
' or ord(char) < 32 or ord(char) == 127:
709 return '' if restricted else '\''
711 return '_-' if restricted else ' -'
712 elif char in '\\/|*<>':
714 if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
716 if restricted and ord(char) > 127:
723 s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
724 result = ''.join(map(replace_insane, s))
726 while '__' in result:
727 result = result.replace('__', '_')
728 result = result.strip('_')
729 # Common case of "Foreign band name
- English song title
"
730 if restricted and result.startswith('-_'):
732 if result.startswith('-'):
733 result = '_' + result[len('-'):]
734 result = result.lstrip('.')
740 def sanitize_path(s, force=False):
741 """Sanitizes and normalizes path on Windows"""
742 if sys.platform == 'win32':
744 drive_or_unc, _ = os.path.splitdrive(s)
745 if sys.version_info < (2, 7) and not drive_or_unc:
746 drive_or_unc, _ = os.path.splitunc(s)
752 norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
756 path_part if path_part in ['.', '..'] else re.sub(r'(?:[/<>:"\|
\\?\
*]|
[\s
.]$
)', '#', path_part)
757 for path_part
in norm_path
]
759 sanitized_path
.insert(0, drive_or_unc
+ os
.path
.sep
)
760 elif force
and s
[0] == os
.path
.sep
:
761 sanitized_path
.insert(0, os
.path
.sep
)
762 return os
.path
.join(*sanitized_path
)
765 def sanitize_url(url
):
766 # Prepend protocol-less URLs with `http:` scheme in order to mitigate
767 # the number of unwanted failures due to missing protocol
768 if url
.startswith('//'):
769 return 'http:%s' % url
770 # Fix some common typos seen so far
772 # https://github.com/ytdl-org/youtube-dl/issues/15649
773 (r
'^httpss://', r
'https://'),
774 # https://bx1.be/lives/direct-tv/
775 (r
'^rmtp([es]?)://', r
'rtmp\1://'),
777 for mistake
, fixup
in COMMON_TYPOS
:
778 if re
.match(mistake
, url
):
779 return re
.sub(mistake
, fixup
, url
)
783 def extract_basic_auth(url
):
784 parts
= compat_urlparse
.urlsplit(url
)
785 if parts
.username
is None:
787 url
= compat_urlparse
.urlunsplit(parts
._replace
(netloc
=(
788 parts
.hostname
if parts
.port
is None
789 else '%s:%d' % (parts
.hostname
, parts
.port
))))
790 auth_payload
= base64
.b64encode(
791 ('%s:%s' % (parts
.username
, parts
.password
or '')).encode('utf-8'))
792 return url
, 'Basic ' + auth_payload
.decode('utf-8')
795 def sanitized_Request(url
, *args
, **kwargs
):
796 url
, auth_header
= extract_basic_auth(escape_url(sanitize_url(url
)))
797 if auth_header
is not None:
798 headers
= args
[1] if len(args
) >= 2 else kwargs
.setdefault('headers', {})
799 headers
['Authorization'] = auth_header
800 return compat_urllib_request
.Request(url
, *args
, **kwargs
)
804 """Expand shell variables and ~"""
805 return os
.path
.expandvars(compat_expanduser(s
))
808 def orderedSet(iterable
):
809 """ Remove all duplicates from the input iterable """
817 def _htmlentity_transform(entity_with_semicolon
):
818 """Transforms an HTML entity to a character."""
819 entity
= entity_with_semicolon
[:-1]
821 # Known non-numeric HTML entity
822 if entity
in compat_html_entities
.name2codepoint
:
823 return compat_chr(compat_html_entities
.name2codepoint
[entity
])
825 # TODO: HTML5 allows entities without a semicolon. For example,
826 # 'Éric' should be decoded as 'Éric'.
827 if entity_with_semicolon
in compat_html_entities_html5
:
828 return compat_html_entities_html5
[entity_with_semicolon
]
830 mobj
= re
.match(r
'#(x[0-9a-fA-F]+|[0-9]+)', entity
)
832 numstr
= mobj
.group(1)
833 if numstr
.startswith('x'):
835 numstr
= '0%s' % numstr
838 # See https://github.com/ytdl-org/youtube-dl/issues/7518
840 return compat_chr(int(numstr
, base
))
844 # Unknown entity in name, return its literal representation
845 return '&%s;' % entity
851 assert type(s
) == compat_str
854 r
'&([^&;]+;)', lambda m
: _htmlentity_transform(m
.group(1)), s
)
857 def escapeHTML(text
):
860 .replace('&', '&')
861 .replace('<', '<')
862 .replace('>', '>')
863 .replace('"', '"')
864 .replace("'", ''')
868 def process_communicate_or_kill(p
, *args
, **kwargs
):
870 return p
.communicate(*args
, **kwargs
)
871 except BaseException
: # Including KeyboardInterrupt
877 class Popen(subprocess
.Popen
):
878 if sys
.platform
== 'win32':
879 _startupinfo
= subprocess
.STARTUPINFO()
880 _startupinfo
.dwFlags |
= subprocess
.STARTF_USESHOWWINDOW
884 def __init__(self
, *args
, **kwargs
):
885 super(Popen
, self
).__init
__(*args
, **kwargs
, startupinfo
=self
._startupinfo
)
887 def communicate_or_kill(self
, *args
, **kwargs
):
888 return process_communicate_or_kill(self
, *args
, **kwargs
)
891 def get_subprocess_encoding():
892 if sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
893 # For subprocess calls, encode with locale encoding
894 # Refer to http://stackoverflow.com/a/9951851/35070
895 encoding
= preferredencoding()
897 encoding
= sys
.getfilesystemencoding()
903 def encodeFilename(s
, for_subprocess
=False):
905 @param s The name of the file
908 assert type(s
) == compat_str
910 # Python 3 has a Unicode API
911 if sys
.version_info
>= (3, 0):
914 # Pass '' directly to use Unicode APIs on Windows 2000 and up
915 # (Detecting Windows NT 4 is tricky because 'major >= 4' would
916 # match Windows 9x series as well. Besides, NT 4 is obsolete.)
917 if not for_subprocess
and sys
.platform
== 'win32' and sys
.getwindowsversion()[0] >= 5:
920 # Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
921 if sys
.platform
.startswith('java'):
924 return s
.encode(get_subprocess_encoding(), 'ignore')
927 def decodeFilename(b
, for_subprocess
=False):
929 if sys
.version_info
>= (3, 0):
932 if not isinstance(b
, bytes):
935 return b
.decode(get_subprocess_encoding(), 'ignore')
938 def encodeArgument(s
):
939 if not isinstance(s
, compat_str
):
940 # Legacy code that uses byte strings
941 # Uncomment the following line after fixing all post processors
942 # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
943 s
= s
.decode('ascii')
944 return encodeFilename(s
, True)
947 def decodeArgument(b
):
948 return decodeFilename(b
, True)
951 def decodeOption(optval
):
954 if isinstance(optval
, bytes):
955 optval
= optval
.decode(preferredencoding())
957 assert isinstance(optval
, compat_str
)
961 _timetuple
= collections
.namedtuple('Time', ('hours', 'minutes', 'seconds', 'milliseconds'))
964 def timetuple_from_msec(msec
):
965 secs
, msec
= divmod(msec
, 1000)
966 mins
, secs
= divmod(secs
, 60)
967 hrs
, mins
= divmod(mins
, 60)
968 return _timetuple(hrs
, mins
, secs
, msec
)
971 def formatSeconds(secs
, delim
=':', msec
=False):
972 time
= timetuple_from_msec(secs
* 1000)
974 ret
= '%d%s%02d%s%02d' % (time
.hours
, delim
, time
.minutes
, delim
, time
.seconds
)
976 ret
= '%d%s%02d' % (time
.minutes
, delim
, time
.seconds
)
978 ret
= '%d' % time
.seconds
979 return '%s.%03d' % (ret
, time
.milliseconds
) if msec
else ret
982 def _ssl_load_windows_store_certs(ssl_context
, storename
):
983 # Code adapted from _load_windows_store_certs in https://github.com/python/cpython/blob/main/Lib/ssl.py
985 certs
= [cert
for cert
, encoding
, trust
in ssl
.enum_certificates(storename
)
986 if encoding
== 'x509_asn' and (
987 trust
is True or ssl
.Purpose
.SERVER_AUTH
.oid
in trust
)]
988 except PermissionError
:
992 ssl_context
.load_verify_locations(cadata
=cert
)
997 def make_HTTPS_handler(params
, **kwargs
):
998 opts_check_certificate
= not params
.get('nocheckcertificate')
999 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
1000 context
.check_hostname
= opts_check_certificate
1001 if params
.get('legacyserverconnect'):
1002 context
.options |
= 4 # SSL_OP_LEGACY_SERVER_CONNECT
1003 context
.verify_mode
= ssl
.CERT_REQUIRED
if opts_check_certificate
else ssl
.CERT_NONE
1004 if opts_check_certificate
:
1006 context
.load_default_certs()
1007 # Work around the issue in load_default_certs when there are bad certificates. See:
1008 # https://github.com/yt-dlp/yt-dlp/issues/1060,
1009 # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
1010 except ssl
.SSLError
:
1011 # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
1012 if sys
.platform
== 'win32' and hasattr(ssl
, 'enum_certificates'):
1013 # Create a new context to discard any certificates that were already loaded
1014 context
= ssl
.SSLContext(ssl
.PROTOCOL_TLS_CLIENT
)
1015 context
.check_hostname
, context
.verify_mode
= True, ssl
.CERT_REQUIRED
1016 for storename
in ('CA', 'ROOT'):
1017 _ssl_load_windows_store_certs(context
, storename
)
1018 context
.set_default_verify_paths()
1019 return YoutubeDLHTTPSHandler(params
, context
=context
, **kwargs
)
1022 def bug_reports_message(before
=';'):
1023 msg
= ('please report this issue on https://github.com/yt-dlp/yt-dlp , '
1024 'filling out the "Broken site" issue template properly. '
1025 'Confirm you are on the latest version using -U')
1027 before
= before
.rstrip()
1028 if not before
or before
.endswith(('.', '!', '?')):
1029 msg
= msg
[0].title() + msg
[1:]
1031 return (before
+ ' ' if before
else '') + msg
1034 class YoutubeDLError(Exception):
1035 """Base exception for YoutubeDL errors."""
1038 def __init__(self
, msg
=None):
1041 elif self
.msg
is None:
1042 self
.msg
= type(self
).__name
__
1043 super().__init
__(self
.msg
)
1046 network_exceptions
= [compat_urllib_error
.URLError
, compat_http_client
.HTTPException
, socket
.error
]
1047 if hasattr(ssl
, 'CertificateError'):
1048 network_exceptions
.append(ssl
.CertificateError
)
1049 network_exceptions
= tuple(network_exceptions
)
1052 class ExtractorError(YoutubeDLError
):
1053 """Error during info extraction."""
1055 def __init__(self
, msg
, tb
=None, expected
=False, cause
=None, video_id
=None, ie
=None):
1056 """ tb, if given, is the original traceback (so that it can be printed out).
1057 If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
1059 if sys
.exc_info()[0] in network_exceptions
:
1064 self
.expected
= expected
1066 self
.video_id
= video_id
1068 self
.exc_info
= sys
.exc_info() # preserve original exception
1070 super(ExtractorError
, self
).__init
__(''.join((
1071 format_field(ie
, template
='[%s] '),
1072 format_field(video_id
, template
='%s: '),
1074 format_field(cause
, template
=' (caused by %r)'),
1075 '' if expected
else bug_reports_message())))
1077 def format_traceback(self
):
1078 if self
.traceback
is None:
1080 return ''.join(traceback
.format_tb(self
.traceback
))
1083 class UnsupportedError(ExtractorError
):
1084 def __init__(self
, url
):
1085 super(UnsupportedError
, self
).__init
__(
1086 'Unsupported URL: %s' % url
, expected
=True)
1090 class RegexNotFoundError(ExtractorError
):
1091 """Error when a regex didn't match"""
1095 class GeoRestrictedError(ExtractorError
):
1096 """Geographic restriction Error exception.
1098 This exception may be thrown when a video is not available from your
1099 geographic location due to geographic restrictions imposed by a website.
1102 def __init__(self
, msg
, countries
=None, **kwargs
):
1103 kwargs
['expected'] = True
1104 super(GeoRestrictedError
, self
).__init
__(msg
, **kwargs
)
1105 self
.countries
= countries
1108 class DownloadError(YoutubeDLError
):
1109 """Download Error exception.
1111 This exception may be thrown by FileDownloader objects if they are not
1112 configured to continue on errors. They will contain the appropriate
1116 def __init__(self
, msg
, exc_info
=None):
1117 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
1118 super(DownloadError
, self
).__init
__(msg
)
1119 self
.exc_info
= exc_info
1122 class EntryNotInPlaylist(YoutubeDLError
):
1123 """Entry not in playlist exception.
1125 This exception will be thrown by YoutubeDL when a requested entry
1126 is not found in the playlist info_dict
1128 msg
= 'Entry not found in info'
1131 class SameFileError(YoutubeDLError
):
1132 """Same File exception.
1134 This exception will be thrown by FileDownloader objects if they detect
1135 multiple files would have to be downloaded to the same file on disk.
1137 msg
= 'Fixed output name but more than one file to download'
1139 def __init__(self
, filename
=None):
1140 if filename
is not None:
1141 self
.msg
+= f
': {filename}'
1142 super().__init
__(self
.msg
)
1145 class PostProcessingError(YoutubeDLError
):
1146 """Post Processing exception.
1148 This exception may be raised by PostProcessor's .run() method to
1149 indicate an error in the postprocessing task.
1153 class DownloadCancelled(YoutubeDLError
):
1154 """ Exception raised when the download queue should be interrupted """
1155 msg
= 'The download was cancelled'
1158 class ExistingVideoReached(DownloadCancelled
):
1159 """ --break-on-existing triggered """
1160 msg
= 'Encountered a video that is already in the archive, stopping due to --break-on-existing'
1163 class RejectedVideoReached(DownloadCancelled
):
1164 """ --break-on-reject triggered """
1165 msg
= 'Encountered a video that did not match filter, stopping due to --break-on-reject'
1168 class MaxDownloadsReached(DownloadCancelled
):
1169 """ --max-downloads limit has been reached. """
1170 msg
= 'Maximum number of downloads reached, stopping due to --max-downloads'
1173 class ReExtractInfo(YoutubeDLError
):
1174 """ Video info needs to be re-extracted. """
1176 def __init__(self
, msg
, expected
=False):
1177 super().__init
__(msg
)
1178 self
.expected
= expected
1181 class ThrottledDownload(ReExtractInfo
):
1182 """ Download speed below --throttled-rate. """
1183 msg
= 'The download speed is below throttle limit'
1186 super().__init
__(self
.msg
, expected
=False)
1189 class UnavailableVideoError(YoutubeDLError
):
1190 """Unavailable Format exception.
1192 This exception will be thrown when a video is requested
1193 in a format that is not available for that video.
1195 msg
= 'Unable to download video'
1197 def __init__(self
, err
=None):
1199 self
.msg
+= f
': {err}'
1200 super().__init
__(self
.msg
)
1203 class ContentTooShortError(YoutubeDLError
):
1204 """Content Too Short exception.
1206 This exception may be raised by FileDownloader objects when a file they
1207 download is too small for what the server announced first, indicating
1208 the connection was probably interrupted.
1211 def __init__(self
, downloaded
, expected
):
1212 super(ContentTooShortError
, self
).__init
__(
1213 'Downloaded {0} bytes, expected {1} bytes'.format(downloaded
, expected
)
1216 self
.downloaded
= downloaded
1217 self
.expected
= expected
1220 class XAttrMetadataError(YoutubeDLError
):
1221 def __init__(self
, code
=None, msg
='Unknown error'):
1222 super(XAttrMetadataError
, self
).__init
__(msg
)
1226 # Parsing code and msg
1227 if (self
.code
in (errno
.ENOSPC
, errno
.EDQUOT
)
1228 or 'No space left' in self
.msg
or 'Disk quota exceeded' in self
.msg
):
1229 self
.reason
= 'NO_SPACE'
1230 elif self
.code
== errno
.E2BIG
or 'Argument list too long' in self
.msg
:
1231 self
.reason
= 'VALUE_TOO_LONG'
1233 self
.reason
= 'NOT_SUPPORTED'
1236 class XAttrUnavailableError(YoutubeDLError
):
1240 def _create_http_connection(ydl_handler
, http_class
, is_https
, *args
, **kwargs
):
1241 # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
1242 # expected HTTP responses to meet HTTP/1.0 or later (see also
1243 # https://github.com/ytdl-org/youtube-dl/issues/6727)
1244 if sys
.version_info
< (3, 0):
1245 kwargs
['strict'] = True
1246 hc
= http_class(*args
, **compat_kwargs(kwargs
))
1247 source_address
= ydl_handler
._params
.get('source_address')
1249 if source_address
is not None:
1250 # This is to workaround _create_connection() from socket where it will try all
1251 # address data from getaddrinfo() including IPv6. This filters the result from
1252 # getaddrinfo() based on the source_address value.
1253 # This is based on the cpython socket.create_connection() function.
1254 # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
1255 def _create_connection(address
, timeout
=socket
._GLOBAL
_DEFAULT
_TIMEOUT
, source_address
=None):
1256 host
, port
= address
1258 addrs
= socket
.getaddrinfo(host
, port
, 0, socket
.SOCK_STREAM
)
1259 af
= socket
.AF_INET
if '.' in source_address
[0] else socket
.AF_INET6
1260 ip_addrs
= [addr
for addr
in addrs
if addr
[0] == af
]
1261 if addrs
and not ip_addrs
:
1262 ip_version
= 'v4' if af
== socket
.AF_INET
else 'v6'
1264 "No remote IP%s addresses available for connect, can't use '%s' as source address"
1265 % (ip_version
, source_address
[0]))
1266 for res
in ip_addrs
:
1267 af
, socktype
, proto
, canonname
, sa
= res
1270 sock
= socket
.socket(af
, socktype
, proto
)
1271 if timeout
is not socket
._GLOBAL
_DEFAULT
_TIMEOUT
:
1272 sock
.settimeout(timeout
)
1273 sock
.bind(source_address
)
1275 err
= None # Explicitly break reference cycle
1277 except socket
.error
as _
:
1279 if sock
is not None:
1284 raise socket
.error('getaddrinfo returns an empty list')
1285 if hasattr(hc
, '_create_connection'):
1286 hc
._create
_connection
= _create_connection
1287 sa
= (source_address
, 0)
1288 if hasattr(hc
, 'source_address'): # Python 2.7+
1289 hc
.source_address
= sa
1291 def _hc_connect(self
, *args
, **kwargs
):
1292 sock
= _create_connection(
1293 (self
.host
, self
.port
), self
.timeout
, sa
)
1295 self
.sock
= ssl
.wrap_socket(
1296 sock
, self
.key_file
, self
.cert_file
,
1297 ssl_version
=ssl
.PROTOCOL_TLSv1
)
1300 hc
.connect
= functools
.partial(_hc_connect
, hc
)
1305 def handle_youtubedl_headers(headers
):
1306 filtered_headers
= headers
1308 if 'Youtubedl-no-compression' in filtered_headers
:
1309 filtered_headers
= dict((k
, v
) for k
, v
in filtered_headers
.items() if k
.lower() != 'accept-encoding')
1310 del filtered_headers
['Youtubedl-no-compression']
1312 return filtered_headers
1315 class YoutubeDLHandler(compat_urllib_request
.HTTPHandler
):
1316 """Handler for HTTP requests and responses.
1318 This class, when installed with an OpenerDirector, automatically adds
1319 the standard headers to every HTTP request and handles gzipped and
1320 deflated responses from web servers. If compression is to be avoided in
1321 a particular request, the original request in the program code only has
1322 to include the HTTP header "Youtubedl-no-compression", which will be
1323 removed before making the real request.
1325 Part of this code was copied from:
1327 http://techknack.net/python-urllib2-handlers/
1329 Andrew Rowls, the author of that code, agreed to release it to the
1333 def __init__(self
, params
, *args
, **kwargs
):
1334 compat_urllib_request
.HTTPHandler
.__init
__(self
, *args
, **kwargs
)
1335 self
._params
= params
1337 def http_open(self
, req
):
1338 conn_class
= compat_http_client
.HTTPConnection
1340 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1342 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1343 del req
.headers
['Ytdl-socks-proxy']
1345 return self
.do_open(functools
.partial(
1346 _create_http_connection
, self
, conn_class
, False),
1354 return zlib
.decompress(data
, -zlib
.MAX_WBITS
)
1356 return zlib
.decompress(data
)
1358 def http_request(self
, req
):
1359 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
1360 # always respected by websites, some tend to give out URLs with non percent-encoded
1361 # non-ASCII characters (see telemb.py, ard.py [#3412])
1362 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
1363 # To work around aforementioned issue we will replace request's original URL with
1364 # percent-encoded one
1365 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
1366 # the code of this workaround has been moved here from YoutubeDL.urlopen()
1367 url
= req
.get_full_url()
1368 url_escaped
= escape_url(url
)
1370 # Substitute URL if any change after escaping
1371 if url
!= url_escaped
:
1372 req
= update_Request(req
, url
=url_escaped
)
1374 for h
, v
in std_headers
.items():
1375 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
1376 # The dict keys are capitalized because of this bug by urllib
1377 if h
.capitalize() not in req
.headers
:
1378 req
.add_header(h
, v
)
1380 req
.headers
= handle_youtubedl_headers(req
.headers
)
1382 if sys
.version_info
< (2, 7) and '#' in req
.get_full_url():
1383 # Python 2.6 is brain-dead when it comes to fragments
1384 req
._Request
__original
= req
._Request
__original
.partition('#')[0]
1385 req
._Request
__r
_type
= req
._Request
__r
_type
.partition('#')[0]
1389 def http_response(self
, req
, resp
):
1392 if resp
.headers
.get('Content-encoding', '') == 'gzip':
1393 content
= resp
.read()
1394 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
), mode
='rb')
1396 uncompressed
= io
.BytesIO(gz
.read())
1397 except IOError as original_ioerror
:
1398 # There may be junk add the end of the file
1399 # See http://stackoverflow.com/q/4928560/35070 for details
1400 for i
in range(1, 1024):
1402 gz
= gzip
.GzipFile(fileobj
=io
.BytesIO(content
[:-i
]), mode
='rb')
1403 uncompressed
= io
.BytesIO(gz
.read())
1408 raise original_ioerror
1409 resp
= compat_urllib_request
.addinfourl(uncompressed
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1410 resp
.msg
= old_resp
.msg
1411 del resp
.headers
['Content-encoding']
1413 if resp
.headers
.get('Content-encoding', '') == 'deflate':
1414 gz
= io
.BytesIO(self
.deflate(resp
.read()))
1415 resp
= compat_urllib_request
.addinfourl(gz
, old_resp
.headers
, old_resp
.url
, old_resp
.code
)
1416 resp
.msg
= old_resp
.msg
1417 del resp
.headers
['Content-encoding']
1418 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
1419 # https://github.com/ytdl-org/youtube-dl/issues/6457).
1420 if 300 <= resp
.code
< 400:
1421 location
= resp
.headers
.get('Location')
1423 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
1424 if sys
.version_info
>= (3, 0):
1425 location
= location
.encode('iso-8859-1').decode('utf-8')
1427 location
= location
.decode('utf-8')
1428 location_escaped
= escape_url(location
)
1429 if location
!= location_escaped
:
1430 del resp
.headers
['Location']
1431 if sys
.version_info
< (3, 0):
1432 location_escaped
= location_escaped
.encode('utf-8')
1433 resp
.headers
['Location'] = location_escaped
1436 https_request
= http_request
1437 https_response
= http_response
1440 def make_socks_conn_class(base_class
, socks_proxy
):
1441 assert issubclass(base_class
, (
1442 compat_http_client
.HTTPConnection
, compat_http_client
.HTTPSConnection
))
1444 url_components
= compat_urlparse
.urlparse(socks_proxy
)
1445 if url_components
.scheme
.lower() == 'socks5':
1446 socks_type
= ProxyType
.SOCKS5
1447 elif url_components
.scheme
.lower() in ('socks', 'socks4'):
1448 socks_type
= ProxyType
.SOCKS4
1449 elif url_components
.scheme
.lower() == 'socks4a':
1450 socks_type
= ProxyType
.SOCKS4A
1452 def unquote_if_non_empty(s
):
1455 return compat_urllib_parse_unquote_plus(s
)
1459 url_components
.hostname
, url_components
.port
or 1080,
1461 unquote_if_non_empty(url_components
.username
),
1462 unquote_if_non_empty(url_components
.password
),
1465 class SocksConnection(base_class
):
1467 self
.sock
= sockssocket()
1468 self
.sock
.setproxy(*proxy_args
)
1469 if type(self
.timeout
) in (int, float):
1470 self
.sock
.settimeout(self
.timeout
)
1471 self
.sock
.connect((self
.host
, self
.port
))
1473 if isinstance(self
, compat_http_client
.HTTPSConnection
):
1474 if hasattr(self
, '_context'): # Python > 2.6
1475 self
.sock
= self
._context
.wrap_socket(
1476 self
.sock
, server_hostname
=self
.host
)
1478 self
.sock
= ssl
.wrap_socket(self
.sock
)
1480 return SocksConnection
1483 class YoutubeDLHTTPSHandler(compat_urllib_request
.HTTPSHandler
):
1484 def __init__(self
, params
, https_conn_class
=None, *args
, **kwargs
):
1485 compat_urllib_request
.HTTPSHandler
.__init
__(self
, *args
, **kwargs
)
1486 self
._https
_conn
_class
= https_conn_class
or compat_http_client
.HTTPSConnection
1487 self
._params
= params
1489 def https_open(self
, req
):
1491 conn_class
= self
._https
_conn
_class
1493 if hasattr(self
, '_context'): # python > 2.6
1494 kwargs
['context'] = self
._context
1495 if hasattr(self
, '_check_hostname'): # python 3.x
1496 kwargs
['check_hostname'] = self
._check
_hostname
1498 socks_proxy
= req
.headers
.get('Ytdl-socks-proxy')
1500 conn_class
= make_socks_conn_class(conn_class
, socks_proxy
)
1501 del req
.headers
['Ytdl-socks-proxy']
1503 return self
.do_open(functools
.partial(
1504 _create_http_connection
, self
, conn_class
, True),
1508 class YoutubeDLCookieJar(compat_cookiejar
.MozillaCookieJar
):
1510 See [1] for cookie file format.
1512 1. https://curl.haxx.se/docs/http-cookies.html
1514 _HTTPONLY_PREFIX
= '#HttpOnly_'
1516 _HEADER
= '''# Netscape HTTP Cookie File
1517 # This file is generated by yt-dlp. Do not edit.
1520 _CookieFileEntry
= collections
.namedtuple(
1522 ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
1524 def save(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1526 Save cookies to a file.
1528 Most of the code is taken from CPython 3.8 and slightly adapted
1529 to support cookie files with UTF-8 in both python 2 and 3.
1531 if filename
is None:
1532 if self
.filename
is not None:
1533 filename
= self
.filename
1535 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1537 # Store session cookies with `expires` set to 0 instead of an empty
1540 if cookie
.expires
is None:
1543 with io
.open(filename
, 'w', encoding
='utf-8') as f
:
1544 f
.write(self
._HEADER
)
1547 if not ignore_discard
and cookie
.discard
:
1549 if not ignore_expires
and cookie
.is_expired(now
):
1555 if cookie
.domain
.startswith('.'):
1556 initial_dot
= 'TRUE'
1558 initial_dot
= 'FALSE'
1559 if cookie
.expires
is not None:
1560 expires
= compat_str(cookie
.expires
)
1563 if cookie
.value
is None:
1564 # cookies.txt regards 'Set-Cookie: foo' as a cookie
1565 # with no name, whereas http.cookiejar regards it as a
1566 # cookie with no value.
1571 value
= cookie
.value
1573 '\t'.join([cookie
.domain
, initial_dot
, cookie
.path
,
1574 secure
, expires
, name
, value
]) + '\n')
1576 def load(self
, filename
=None, ignore_discard
=False, ignore_expires
=False):
1577 """Load cookies from a file."""
1578 if filename
is None:
1579 if self
.filename
is not None:
1580 filename
= self
.filename
1582 raise ValueError(compat_cookiejar
.MISSING_FILENAME_TEXT
)
1584 def prepare_line(line
):
1585 if line
.startswith(self
._HTTPONLY
_PREFIX
):
1586 line
= line
[len(self
._HTTPONLY
_PREFIX
):]
1587 # comments and empty lines are fine
1588 if line
.startswith('#') or not line
.strip():
1590 cookie_list
= line
.split('\t')
1591 if len(cookie_list
) != self
._ENTRY
_LEN
:
1592 raise compat_cookiejar
.LoadError('invalid length %d' % len(cookie_list
))
1593 cookie
= self
._CookieFileEntry
(*cookie_list
)
1594 if cookie
.expires_at
and not cookie
.expires_at
.isdigit():
1595 raise compat_cookiejar
.LoadError('invalid expires at %s' % cookie
.expires_at
)
1599 with io
.open(filename
, encoding
='utf-8') as f
:
1602 cf
.write(prepare_line(line
))
1603 except compat_cookiejar
.LoadError
as e
:
1605 'WARNING: skipping cookie file entry due to %s: %r\n'
1606 % (e
, line
), sys
.stderr
)
1609 self
._really
_load
(cf
, filename
, ignore_discard
, ignore_expires
)
1610 # Session cookies are denoted by either `expires` field set to
1611 # an empty string or 0. MozillaCookieJar only recognizes the former
1612 # (see [1]). So we need force the latter to be recognized as session
1613 # cookies on our own.
1614 # Session cookies may be important for cookies-based authentication,
1615 # e.g. usually, when user does not check 'Remember me' check box while
1616 # logging in on a site, some important cookies are stored as session
1617 # cookies so that not recognizing them will result in failed login.
1618 # 1. https://bugs.python.org/issue17164
1620 # Treat `expires=0` cookies as session cookies
1621 if cookie
.expires
== 0:
1622 cookie
.expires
= None
1623 cookie
.discard
= True
1626 class YoutubeDLCookieProcessor(compat_urllib_request
.HTTPCookieProcessor
):
1627 def __init__(self
, cookiejar
=None):
1628 compat_urllib_request
.HTTPCookieProcessor
.__init
__(self
, cookiejar
)
1630 def http_response(self
, request
, response
):
1631 # Python 2 will choke on next HTTP request in row if there are non-ASCII
1632 # characters in Set-Cookie HTTP header of last response (see
1633 # https://github.com/ytdl-org/youtube-dl/issues/6769).
1634 # In order to at least prevent crashing we will percent encode Set-Cookie
1635 # header before HTTPCookieProcessor starts processing it.
1636 # if sys.version_info < (3, 0) and response.headers:
1637 # for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
1638 # set_cookie = response.headers.get(set_cookie_header)
1640 # set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
1641 # if set_cookie != set_cookie_escaped:
1642 # del response.headers[set_cookie_header]
1643 # response.headers[set_cookie_header] = set_cookie_escaped
1644 return compat_urllib_request
.HTTPCookieProcessor
.http_response(self
, request
, response
)
1646 https_request
= compat_urllib_request
.HTTPCookieProcessor
.http_request
1647 https_response
= http_response
1650 class YoutubeDLRedirectHandler(compat_urllib_request
.HTTPRedirectHandler
):
1651 """YoutubeDL redirect handler
1653 The code is based on HTTPRedirectHandler implementation from CPython [1].
1655 This redirect handler solves two issues:
1656 - ensures redirect URL is always unicode under python 2
1657 - introduces support for experimental HTTP response status code
1658 308 Permanent Redirect [2] used by some sites [3]
1660 1. https://github.com/python/cpython/blob/master/Lib/urllib/request.py
1661 2. https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/308
1662 3. https://github.com/ytdl-org/youtube-dl/issues/28768
1665 http_error_301
= http_error_303
= http_error_307
= http_error_308
= compat_urllib_request
.HTTPRedirectHandler
.http_error_302
1667 def redirect_request(self
, req
, fp
, code
, msg
, headers
, newurl
):
1668 """Return a Request or None in response to a redirect.
1670 This is called by the http_error_30x methods when a
1671 redirection response is received. If a redirection should
1672 take place, return a new Request to allow http_error_30x to
1673 perform the redirect. Otherwise, raise HTTPError if no-one
1674 else should try to handle this url. Return None if you can't
1675 but another Handler might.
1677 m
= req
.get_method()
1678 if (not (code
in (301, 302, 303, 307, 308) and m
in ("GET", "HEAD")
1679 or code
in (301, 302, 303) and m
== "POST")):
1680 raise compat_HTTPError(req
.full_url
, code
, msg
, headers
, fp
)
1681 # Strictly (according to RFC 2616), 301 or 302 in response to
1682 # a POST MUST NOT cause a redirection without confirmation
1683 # from the user (of urllib.request, in this case). In practice,
1684 # essentially all clients do redirect in this case, so we do
1687 # On python 2 urlh.geturl() may sometimes return redirect URL
1688 # as byte string instead of unicode. This workaround allows
1689 # to force it always return unicode.
1690 if sys
.version_info
[0] < 3:
1691 newurl
= compat_str(newurl
)
1693 # Be conciliant with URIs containing a space. This is mainly
1694 # redundant with the more complete encoding done in http_error_302(),
1695 # but it is kept for compatibility with other callers.
1696 newurl
= newurl
.replace(' ', '%20')
1698 CONTENT_HEADERS
= ("content-length", "content-type")
1699 # NB: don't use dict comprehension for python 2.6 compatibility
1700 newheaders
= dict((k
, v
) for k
, v
in req
.headers
.items()
1701 if k
.lower() not in CONTENT_HEADERS
)
1702 return compat_urllib_request
.Request(
1703 newurl
, headers
=newheaders
, origin_req_host
=req
.origin_req_host
,
1707 def extract_timezone(date_str
):
1710 ^.{8,}? # >=8 char non-TZ prefix, if present
1711 (?P<tz>Z| # just the UTC Z, or
1712 (?:(?<=.\b\d{4}|\b\d{2}:\d\d)| # preceded by 4 digits or hh:mm or
1713 (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d)) # not preceded by 3 alpha word or >= 4 alpha or 2 digits
1714 [ ]? # optional space
1715 (?P<sign>\+|-) # +/-
1716 (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2}) # hh[:]mm
1720 timezone
= datetime
.timedelta()
1722 date_str
= date_str
[:-len(m
.group('tz'))]
1723 if not m
.group('sign'):
1724 timezone
= datetime
.timedelta()
1726 sign
= 1 if m
.group('sign') == '+' else -1
1727 timezone
= datetime
.timedelta(
1728 hours
=sign
* int(m
.group('hours')),
1729 minutes
=sign
* int(m
.group('minutes')))
1730 return timezone
, date_str
1733 def parse_iso8601(date_str
, delimiter
='T', timezone
=None):
1734 """ Return a UNIX timestamp from the given date """
1736 if date_str
is None:
1739 date_str
= re
.sub(r
'\.[0-9]+', '', date_str
)
1741 if timezone
is None:
1742 timezone
, date_str
= extract_timezone(date_str
)
1745 date_format
= '%Y-%m-%d{0}%H:%M:%S'.format(delimiter
)
1746 dt
= datetime
.datetime
.strptime(date_str
, date_format
) - timezone
1747 return calendar
.timegm(dt
.timetuple())
1752 def date_formats(day_first
=True):
1753 return DATE_FORMATS_DAY_FIRST
if day_first
else DATE_FORMATS_MONTH_FIRST
1756 def unified_strdate(date_str
, day_first
=True):
1757 """Return a string with the date in the format YYYYMMDD"""
1759 if date_str
is None:
1763 date_str
= date_str
.replace(',', ' ')
1764 # Remove AM/PM + timezone
1765 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1766 _
, date_str
= extract_timezone(date_str
)
1768 for expression
in date_formats(day_first
):
1770 upload_date
= datetime
.datetime
.strptime(date_str
, expression
).strftime('%Y%m%d')
1773 if upload_date
is None:
1774 timetuple
= email
.utils
.parsedate_tz(date_str
)
1777 upload_date
= datetime
.datetime(*timetuple
[:6]).strftime('%Y%m%d')
1780 if upload_date
is not None:
1781 return compat_str(upload_date
)
1784 def unified_timestamp(date_str
, day_first
=True):
1785 if date_str
is None:
1788 date_str
= re
.sub(r
'[,|]', '', date_str
)
1790 pm_delta
= 12 if re
.search(r
'(?i)PM', date_str
) else 0
1791 timezone
, date_str
= extract_timezone(date_str
)
1793 # Remove AM/PM + timezone
1794 date_str
= re
.sub(r
'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str
)
1796 # Remove unrecognized timezones from ISO 8601 alike timestamps
1797 m
= re
.search(r
'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str
)
1799 date_str
= date_str
[:-len(m
.group('tz'))]
1801 # Python only supports microseconds, so remove nanoseconds
1802 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]+$', date_str
)
1804 date_str
= m
.group(1)
1806 for expression
in date_formats(day_first
):
1808 dt
= datetime
.datetime
.strptime(date_str
, expression
) - timezone
+ datetime
.timedelta(hours
=pm_delta
)
1809 return calendar
.timegm(dt
.timetuple())
1812 timetuple
= email
.utils
.parsedate_tz(date_str
)
1814 return calendar
.timegm(timetuple
) + pm_delta
* 3600
1817 def determine_ext(url
, default_ext
='unknown_video'):
1818 if url
is None or '.' not in url
:
1820 guess
= url
.partition('?')[0].rpartition('.')[2]
1821 if re
.match(r
'^[A-Za-z0-9]+$', guess
):
1823 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1824 elif guess
.rstrip('/') in KNOWN_EXTENSIONS
:
1825 return guess
.rstrip('/')
1830 def subtitles_filename(filename
, sub_lang
, sub_format
, expected_real_ext
=None):
1831 return replace_extension(filename
, sub_lang
+ '.' + sub_format
, expected_real_ext
)
1834 def datetime_from_str(date_str
, precision
='auto', format
='%Y%m%d'):
1836 Return a datetime object from a string in the format YYYYMMDD or
1837 (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
1839 format: string date format used to return datetime object from
1840 precision: round the time portion of a datetime object.
1841 auto|microsecond|second|minute|hour|day.
1842 auto: round to the unit provided in date_str (if applicable).
1844 auto_precision
= False
1845 if precision
== 'auto':
1846 auto_precision
= True
1847 precision
= 'microsecond'
1848 today
= datetime_round(datetime
.datetime
.utcnow(), precision
)
1849 if date_str
in ('now', 'today'):
1851 if date_str
== 'yesterday':
1852 return today
- datetime
.timedelta(days
=1)
1854 r
'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)(s)?',
1856 if match
is not None:
1857 start_time
= datetime_from_str(match
.group('start'), precision
, format
)
1858 time
= int(match
.group('time')) * (-1 if match
.group('sign') == '-' else 1)
1859 unit
= match
.group('unit')
1860 if unit
== 'month' or unit
== 'year':
1861 new_date
= datetime_add_months(start_time
, time
* 12 if unit
== 'year' else time
)
1867 delta
= datetime
.timedelta(**{unit + 's': time}
)
1868 new_date
= start_time
+ delta
1870 return datetime_round(new_date
, unit
)
1873 return datetime_round(datetime
.datetime
.strptime(date_str
, format
), precision
)
1876 def date_from_str(date_str
, format
='%Y%m%d', strict
=False):
1878 Return a datetime object from a string in the format YYYYMMDD or
1879 (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
1881 If "strict", only (now|today)[+-][0-9](day|week|month|year)(s)? is allowed
1883 format: string date format used to return datetime object from
1885 if strict
and not re
.fullmatch(r
'\d{8}|(now|today)[+-]\d+(day|week|month|year)(s)?', date_str
):
1886 raise ValueError(f
'Invalid date format {date_str}')
1887 return datetime_from_str(date_str
, precision
='microsecond', format
=format
).date()
1890 def datetime_add_months(dt
, months
):
1891 """Increment/Decrement a datetime object by months."""
1892 month
= dt
.month
+ months
- 1
1893 year
= dt
.year
+ month
// 12
1894 month
= month
% 12 + 1
1895 day
= min(dt
.day
, calendar
.monthrange(year
, month
)[1])
1896 return dt
.replace(year
, month
, day
)
1899 def datetime_round(dt
, precision
='day'):
1901 Round a datetime object's time to a specific precision
1903 if precision
== 'microsecond':
1912 roundto
= lambda x
, n
: ((x
+ n
/ 2) // n
) * n
1913 timestamp
= calendar
.timegm(dt
.timetuple())
1914 return datetime
.datetime
.utcfromtimestamp(roundto(timestamp
, unit_seconds
[precision
]))
1917 def hyphenate_date(date_str
):
1919 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1920 match
= re
.match(r
'^(\d\d\d\d)(\d\d)(\d\d)$', date_str
)
1921 if match
is not None:
1922 return '-'.join(match
.groups())
1927 class DateRange(object):
1928 """Represents a time interval between two dates"""
1930 def __init__(self
, start
=None, end
=None):
1931 """start and end must be strings in the format accepted by date"""
1932 if start
is not None:
1933 self
.start
= date_from_str(start
, strict
=True)
1935 self
.start
= datetime
.datetime
.min.date()
1937 self
.end
= date_from_str(end
, strict
=True)
1939 self
.end
= datetime
.datetime
.max.date()
1940 if self
.start
> self
.end
:
1941 raise ValueError('Date range: "%s" , the start date must be before the end date' % self
)
1945 """Returns a range that only contains the given day"""
1946 return cls(day
, day
)
1948 def __contains__(self
, date
):
1949 """Check if the date is in the range"""
1950 if not isinstance(date
, datetime
.date
):
1951 date
= date_from_str(date
)
1952 return self
.start
<= date
<= self
.end
1955 return '%s - %s' % (self
.start
.isoformat(), self
.end
.isoformat())
1958 def platform_name():
1959 """ Returns the platform name as a compat_str """
1960 res
= platform
.platform()
1961 if isinstance(res
, bytes):
1962 res
= res
.decode(preferredencoding())
1964 assert isinstance(res
, compat_str
)
1968 def get_windows_version():
1969 ''' Get Windows version. None if it's not running on Windows '''
1970 if compat_os_name
== 'nt':
1971 return version_tuple(platform
.win32_ver()[1])
1976 def _windows_write_string(s
, out
):
1977 """ Returns True if the string was written using special methods,
1978 False if it has yet to be written out."""
1979 # Adapted from http://stackoverflow.com/a/3259271/35070
1981 import ctypes
.wintypes
1989 fileno
= out
.fileno()
1990 except AttributeError:
1991 # If the output stream doesn't have a fileno, it's virtual
1993 except io
.UnsupportedOperation
:
1994 # Some strange Windows pseudo files?
1996 if fileno
not in WIN_OUTPUT_IDS
:
1999 GetStdHandle
= compat_ctypes_WINFUNCTYPE(
2000 ctypes
.wintypes
.HANDLE
, ctypes
.wintypes
.DWORD
)(
2001 ('GetStdHandle', ctypes
.windll
.kernel32
))
2002 h
= GetStdHandle(WIN_OUTPUT_IDS
[fileno
])
2004 WriteConsoleW
= compat_ctypes_WINFUNCTYPE(
2005 ctypes
.wintypes
.BOOL
, ctypes
.wintypes
.HANDLE
, ctypes
.wintypes
.LPWSTR
,
2006 ctypes
.wintypes
.DWORD
, ctypes
.POINTER(ctypes
.wintypes
.DWORD
),
2007 ctypes
.wintypes
.LPVOID
)(('WriteConsoleW', ctypes
.windll
.kernel32
))
2008 written
= ctypes
.wintypes
.DWORD(0)
2010 GetFileType
= compat_ctypes_WINFUNCTYPE(ctypes
.wintypes
.DWORD
, ctypes
.wintypes
.DWORD
)(('GetFileType', ctypes
.windll
.kernel32
))
2011 FILE_TYPE_CHAR
= 0x0002
2012 FILE_TYPE_REMOTE
= 0x8000
2013 GetConsoleMode
= compat_ctypes_WINFUNCTYPE(
2014 ctypes
.wintypes
.BOOL
, ctypes
.wintypes
.HANDLE
,
2015 ctypes
.POINTER(ctypes
.wintypes
.DWORD
))(
2016 ('GetConsoleMode', ctypes
.windll
.kernel32
))
2017 INVALID_HANDLE_VALUE
= ctypes
.wintypes
.DWORD(-1).value
2019 def not_a_console(handle
):
2020 if handle
== INVALID_HANDLE_VALUE
or handle
is None:
2022 return ((GetFileType(handle
) & ~FILE_TYPE_REMOTE
) != FILE_TYPE_CHAR
2023 or GetConsoleMode(handle
, ctypes
.byref(ctypes
.wintypes
.DWORD())) == 0)
2025 if not_a_console(h
):
2028 def next_nonbmp_pos(s
):
2030 return next(i
for i
, c
in enumerate(s
) if ord(c
) > 0xffff)
2031 except StopIteration:
2035 count
= min(next_nonbmp_pos(s
), 1024)
2037 ret
= WriteConsoleW(
2038 h
, s
, count
if count
else 2, ctypes
.byref(written
), None)
2040 raise OSError('Failed to write string')
2041 if not count
: # We just wrote a non-BMP character
2042 assert written
.value
== 2
2045 assert written
.value
> 0
2046 s
= s
[written
.value
:]
2050 def write_string(s
, out
=None, encoding
=None):
2053 assert type(s
) == compat_str
2055 if sys
.platform
== 'win32' and encoding
is None and hasattr(out
, 'fileno'):
2056 if _windows_write_string(s
, out
):
2059 if ('b' in getattr(out
, 'mode', '')
2060 or sys
.version_info
[0] < 3): # Python 2 lies about mode of sys.stderr
2061 byt
= s
.encode(encoding
or preferredencoding(), 'ignore')
2063 elif hasattr(out
, 'buffer'):
2064 enc
= encoding
or getattr(out
, 'encoding', None) or preferredencoding()
2065 byt
= s
.encode(enc
, 'ignore')
2066 out
.buffer.write(byt
)
2072 def bytes_to_intlist(bs
):
2075 if isinstance(bs
[0], int): # Python 3
2078 return [ord(c
) for c
in bs
]
2081 def intlist_to_bytes(xs
):
2084 return compat_struct_pack('%dB' % len(xs
), *xs
)
2087 # Cross-platform file locking
2088 if sys
.platform
== 'win32':
2089 import ctypes
.wintypes
2092 class OVERLAPPED(ctypes
.Structure
):
2094 ('Internal', ctypes
.wintypes
.LPVOID
),
2095 ('InternalHigh', ctypes
.wintypes
.LPVOID
),
2096 ('Offset', ctypes
.wintypes
.DWORD
),
2097 ('OffsetHigh', ctypes
.wintypes
.DWORD
),
2098 ('hEvent', ctypes
.wintypes
.HANDLE
),
2101 kernel32
= ctypes
.windll
.kernel32
2102 LockFileEx
= kernel32
.LockFileEx
2103 LockFileEx
.argtypes
= [
2104 ctypes
.wintypes
.HANDLE
, # hFile
2105 ctypes
.wintypes
.DWORD
, # dwFlags
2106 ctypes
.wintypes
.DWORD
, # dwReserved
2107 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2108 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2109 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2111 LockFileEx
.restype
= ctypes
.wintypes
.BOOL
2112 UnlockFileEx
= kernel32
.UnlockFileEx
2113 UnlockFileEx
.argtypes
= [
2114 ctypes
.wintypes
.HANDLE
, # hFile
2115 ctypes
.wintypes
.DWORD
, # dwReserved
2116 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockLow
2117 ctypes
.wintypes
.DWORD
, # nNumberOfBytesToLockHigh
2118 ctypes
.POINTER(OVERLAPPED
) # Overlapped
2120 UnlockFileEx
.restype
= ctypes
.wintypes
.BOOL
2121 whole_low
= 0xffffffff
2122 whole_high
= 0x7fffffff
2124 def _lock_file(f
, exclusive
, block
): # todo: block unused on win32
2125 overlapped
= OVERLAPPED()
2126 overlapped
.Offset
= 0
2127 overlapped
.OffsetHigh
= 0
2128 overlapped
.hEvent
= 0
2129 f
._lock
_file
_overlapped
_p
= ctypes
.pointer(overlapped
)
2130 handle
= msvcrt
.get_osfhandle(f
.fileno())
2131 if not LockFileEx(handle
, 0x2 if exclusive
else 0x0, 0,
2132 whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2133 raise OSError('Locking file failed: %r' % ctypes
.FormatError())
2135 def _unlock_file(f
):
2136 assert f
._lock
_file
_overlapped
_p
2137 handle
= msvcrt
.get_osfhandle(f
.fileno())
2138 if not UnlockFileEx(handle
, 0,
2139 whole_low
, whole_high
, f
._lock
_file
_overlapped
_p
):
2140 raise OSError('Unlocking file failed: %r' % ctypes
.FormatError())
2143 # Some platforms, such as Jython, is missing fcntl
2147 def _lock_file(f
, exclusive
, block
):
2149 fcntl
.LOCK_SH
if not exclusive
2150 else fcntl
.LOCK_EX
if block
2151 else fcntl
.LOCK_EX | fcntl
.LOCK_NB
)
2153 def _unlock_file(f
):
2154 fcntl
.flock(f
, fcntl
.LOCK_UN
)
2157 UNSUPPORTED_MSG
= 'file locking is not supported on this platform'
2159 def _lock_file(f
, exclusive
, block
):
2160 raise IOError(UNSUPPORTED_MSG
)
2162 def _unlock_file(f
):
2163 raise IOError(UNSUPPORTED_MSG
)
2166 class locked_file(object):
2167 def __init__(self
, filename
, mode
, block
=True, encoding
=None):
2168 assert mode
in ['r', 'rb', 'a', 'ab', 'w', 'wb']
2169 self
.f
= io
.open(filename
, mode
, encoding
=encoding
)
2173 def __enter__(self
):
2174 exclusive
= 'r' not in self
.mode
2176 _lock_file(self
.f
, exclusive
, self
.block
)
2182 def __exit__(self
, etype
, value
, traceback
):
2184 _unlock_file(self
.f
)
2191 def write(self
, *args
):
2192 return self
.f
.write(*args
)
2194 def read(self
, *args
):
2195 return self
.f
.read(*args
)
2201 return self
.__enter
__()
2203 def close(self
, *args
):
2204 self
.__exit
__(self
, *args
, value
=False, traceback
=False)
2207 def get_filesystem_encoding():
2208 encoding
= sys
.getfilesystemencoding()
2209 return encoding
if encoding
is not None else 'utf-8'
2212 def shell_quote(args
):
2214 encoding
= get_filesystem_encoding()
2216 if isinstance(a
, bytes):
2217 # We may get a filename encoded with 'encodeFilename'
2218 a
= a
.decode(encoding
)
2219 quoted_args
.append(compat_shlex_quote(a
))
2220 return ' '.join(quoted_args
)
2223 def smuggle_url(url
, data
):
2224 """ Pass additional data in a URL for internal use. """
2226 url
, idata
= unsmuggle_url(url
, {})
2228 sdata
= compat_urllib_parse_urlencode(
2229 {'__youtubedl_smuggle': json.dumps(data)}
)
2230 return url
+ '#' + sdata
2233 def unsmuggle_url(smug_url
, default
=None):
2234 if '#__youtubedl_smuggle' not in smug_url
:
2235 return smug_url
, default
2236 url
, _
, sdata
= smug_url
.rpartition('#')
2237 jsond
= compat_parse_qs(sdata
)['__youtubedl_smuggle'][0]
2238 data
= json
.loads(jsond
)
2242 def format_decimal_suffix(num
, fmt
='%d%s', *, factor
=1000):
2243 """ Formats numbers with decimal sufixes like K, M, etc """
2244 num
, factor
= float_or_none(num
), float(factor
)
2247 exponent
= 0 if num
== 0 else int(math
.log(num
, factor
))
2248 suffix
= ['', *'kMGTPEZY'][exponent
]
2250 suffix
= {'k': 'Ki', '': ''}
.get(suffix
, f
'{suffix}i')
2251 converted
= num
/ (factor
** exponent
)
2252 return fmt
% (converted
, suffix
)
2255 def format_bytes(bytes):
2256 return format_decimal_suffix(bytes, '%.2f%sB', factor
=1024) or 'N/A'
2259 def lookup_unit_table(unit_table
, s
):
2260 units_re
= '|'.join(re
.escape(u
) for u
in unit_table
)
2262 r
'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re
, s
)
2265 num_str
= m
.group('num').replace(',', '.')
2266 mult
= unit_table
[m
.group('unit')]
2267 return int(float(num_str
) * mult
)
2270 def parse_filesize(s
):
2274 # The lower-case forms are of course incorrect and unofficial,
2275 # but we support those too
2292 'megabytes': 1000 ** 2,
2293 'mebibytes': 1024 ** 2,
2299 'gigabytes': 1000 ** 3,
2300 'gibibytes': 1024 ** 3,
2306 'terabytes': 1000 ** 4,
2307 'tebibytes': 1024 ** 4,
2313 'petabytes': 1000 ** 5,
2314 'pebibytes': 1024 ** 5,
2320 'exabytes': 1000 ** 6,
2321 'exbibytes': 1024 ** 6,
2327 'zettabytes': 1000 ** 7,
2328 'zebibytes': 1024 ** 7,
2334 'yottabytes': 1000 ** 8,
2335 'yobibytes': 1024 ** 8,
2338 return lookup_unit_table(_UNIT_TABLE
, s
)
2345 s
= re
.sub(r
'^[^\d]+\s', '', s
).strip()
2347 if re
.match(r
'^[\d,.]+$', s
):
2348 return str_to_int(s
)
2361 ret
= lookup_unit_table(_UNIT_TABLE
, s
)
2365 mobj
= re
.match(r
'([\d,.]+)(?:$|\s)', s
)
2367 return str_to_int(mobj
.group(1))
2370 def parse_resolution(s
):
2374 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(?P<w>\d+)\s*[xX×,]\s*(?P<h>\d+)(?![a-zA-Z0-9])', s
)
2377 'width': int(mobj
.group('w')),
2378 'height': int(mobj
.group('h')),
2381 mobj
= re
.search(r
'(?<![a-zA-Z0-9])(\d+)[pPiI](?![a-zA-Z0-9])', s
)
2383 return {'height': int(mobj.group(1))}
2385 mobj
= re
.search(r
'\b([48])[kK]\b', s
)
2387 return {'height': int(mobj.group(1)) * 540}
2392 def parse_bitrate(s
):
2393 if not isinstance(s
, compat_str
):
2395 mobj
= re
.search(r
'\b(\d+)\s*kbps', s
)
2397 return int(mobj
.group(1))
2400 def month_by_name(name
, lang
='en'):
2401 """ Return the number of a month by (locale-independently) English name """
2403 month_names
= MONTH_NAMES
.get(lang
, MONTH_NAMES
['en'])
2406 return month_names
.index(name
) + 1
2411 def month_by_abbreviation(abbrev
):
2412 """ Return the number of a month by (locale-independently) English
2416 return [s
[:3] for s
in ENGLISH_MONTH_NAMES
].index(abbrev
) + 1
2421 def fix_xml_ampersands(xml_str
):
2422 """Replace all the '&' by '&' in XML"""
2424 r
'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
2429 def setproctitle(title
):
2430 assert isinstance(title
, compat_str
)
2432 # ctypes in Jython is not complete
2433 # http://bugs.jython.org/issue2148
2434 if sys
.platform
.startswith('java'):
2438 libc
= ctypes
.cdll
.LoadLibrary('libc.so.6')
2442 # LoadLibrary in Windows Python 2.7.13 only expects
2443 # a bytestring, but since unicode_literals turns
2444 # every string into a unicode string, it fails.
2446 title_bytes
= title
.encode('utf-8')
2447 buf
= ctypes
.create_string_buffer(len(title_bytes
))
2448 buf
.value
= title_bytes
2450 libc
.prctl(15, buf
, 0, 0, 0)
2451 except AttributeError:
2452 return # Strange libc, just skip this
2455 def remove_start(s
, start
):
2456 return s
[len(start
):] if s
is not None and s
.startswith(start
) else s
2459 def remove_end(s
, end
):
2460 return s
[:-len(end
)] if s
is not None and s
.endswith(end
) else s
2463 def remove_quotes(s
):
2464 if s
is None or len(s
) < 2:
2466 for quote
in ('"', "'", ):
2467 if s
[0] == quote
and s
[-1] == quote
:
2472 def get_domain(url
):
2473 domain
= re
.match(r
'(?:https?:\/\/)?(?:www\.)?(?P<domain>[^\n\/]+\.[^\n\/]+)(?:\/(.*))?', url
)
2474 return domain
.group('domain') if domain
else None
2477 def url_basename(url
):
2478 path
= compat_urlparse
.urlparse(url
).path
2479 return path
.strip('/').split('/')[-1]
2483 return re
.match(r
'https?://[^?#&]+/', url
).group()
2486 def urljoin(base
, path
):
2487 if isinstance(path
, bytes):
2488 path
= path
.decode('utf-8')
2489 if not isinstance(path
, compat_str
) or not path
:
2491 if re
.match(r
'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path
):
2493 if isinstance(base
, bytes):
2494 base
= base
.decode('utf-8')
2495 if not isinstance(base
, compat_str
) or not re
.match(
2496 r
'^(?:https?:)?//', base
):
2498 return compat_urlparse
.urljoin(base
, path
)
2501 class HEADRequest(compat_urllib_request
.Request
):
2502 def get_method(self
):
2506 class PUTRequest(compat_urllib_request
.Request
):
2507 def get_method(self
):
2511 def int_or_none(v
, scale
=1, default
=None, get_attr
=None, invscale
=1):
2512 if get_attr
and v
is not None:
2513 v
= getattr(v
, get_attr
, None)
2515 return int(v
) * invscale
// scale
2516 except (ValueError, TypeError, OverflowError):
2520 def str_or_none(v
, default
=None):
2521 return default
if v
is None else compat_str(v
)
2524 def str_to_int(int_str
):
2525 """ A more relaxed version of int_or_none """
2526 if isinstance(int_str
, compat_integer_types
):
2528 elif isinstance(int_str
, compat_str
):
2529 int_str
= re
.sub(r
'[,\.\+]', '', int_str
)
2530 return int_or_none(int_str
)
2533 def float_or_none(v
, scale
=1, invscale
=1, default
=None):
2537 return float(v
) * invscale
/ scale
2538 except (ValueError, TypeError):
2542 def bool_or_none(v
, default
=None):
2543 return v
if isinstance(v
, bool) else default
2546 def strip_or_none(v
, default
=None):
2547 return v
.strip() if isinstance(v
, compat_str
) else default
2550 def url_or_none(url
):
2551 if not url
or not isinstance(url
, compat_str
):
2554 return url
if re
.match(r
'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url
) else None
2557 def strftime_or_none(timestamp
, date_format
, default
=None):
2558 datetime_object
= None
2560 if isinstance(timestamp
, compat_numeric_types
): # unix timestamp
2561 datetime_object
= datetime
.datetime
.utcfromtimestamp(timestamp
)
2562 elif isinstance(timestamp
, compat_str
): # assume YYYYMMDD
2563 datetime_object
= datetime
.datetime
.strptime(timestamp
, '%Y%m%d')
2564 return datetime_object
.strftime(date_format
)
2565 except (ValueError, TypeError, AttributeError):
2569 def parse_duration(s
):
2570 if not isinstance(s
, compat_basestring
):
2576 days
, hours
, mins
, secs
, ms
= [None] * 5
2577 m
= re
.match(r
'''(?x)
2579 (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?
2580 (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
2581 (?P<ms>[.:][0-9]+)?Z?$
2584 days
, hours
, mins
, secs
, ms
= m
.group('days', 'hours', 'mins', 'secs', 'ms')
2589 [0-9]+\s*y(?:ears?)?\s*
2592 [0-9]+\s*m(?:onths?)?\s*
2595 [0-9]+\s*w(?:eeks?)?\s*
2598 (?P<days>[0-9]+)\s*d(?:ays?)?\s*
2602 (?P<hours>[0-9]+)\s*h(?:ours?)?\s*
2605 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?\s*
2608 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
2611 days
, hours
, mins
, secs
, ms
= m
.groups()
2613 m
= re
.match(r
'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s
)
2615 hours
, mins
= m
.groups()
2621 duration
+= float(secs
)
2623 duration
+= float(mins
) * 60
2625 duration
+= float(hours
) * 60 * 60
2627 duration
+= float(days
) * 24 * 60 * 60
2629 duration
+= float(ms
.replace(':', '.'))
2633 def prepend_extension(filename
, ext
, expected_real_ext
=None):
2634 name
, real_ext
= os
.path
.splitext(filename
)
2636 '{0}.{1}{2}'.format(name
, ext
, real_ext
)
2637 if not expected_real_ext
or real_ext
[1:] == expected_real_ext
2638 else '{0}.{1}'.format(filename
, ext
))
2641 def replace_extension(filename
, ext
, expected_real_ext
=None):
2642 name
, real_ext
= os
.path
.splitext(filename
)
2643 return '{0}.{1}'.format(
2644 name
if not expected_real_ext
or real_ext
[1:] == expected_real_ext
else filename
,
2648 def check_executable(exe
, args
=[]):
2649 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
2650 args can be a list of arguments for a short output (like -version) """
2652 Popen([exe
] + args
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
).communicate_or_kill()
2658 def _get_exe_version_output(exe
, args
):
2660 # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
2661 # SIGTTOU if yt-dlp is run in the background.
2662 # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
2664 [encodeArgument(exe
)] + args
, stdin
=subprocess
.PIPE
,
2665 stdout
=subprocess
.PIPE
, stderr
=subprocess
.STDOUT
).communicate_or_kill()
2668 if isinstance(out
, bytes): # Python 2.x
2669 out
= out
.decode('ascii', 'ignore')
2673 def detect_exe_version(output
, version_re
=None, unrecognized
='present'):
2674 assert isinstance(output
, compat_str
)
2675 if version_re
is None:
2676 version_re
= r
'version\s+([-0-9._a-zA-Z]+)'
2677 m
= re
.search(version_re
, output
)
2684 def get_exe_version(exe
, args
=['--version'],
2685 version_re
=None, unrecognized
='present'):
2686 """ Returns the version of the specified executable,
2687 or False if the executable is not present """
2688 out
= _get_exe_version_output(exe
, args
)
2689 return detect_exe_version(out
, version_re
, unrecognized
) if out
else False
2692 class LazyList(collections
.abc
.Sequence
):
2693 ''' Lazy immutable list from an iterable
2694 Note that slices of a LazyList are lists and not LazyList'''
2696 class IndexError(IndexError):
2699 def __init__(self
, iterable
, *, reverse
=False, _cache
=None):
2700 self
.__iterable
= iter(iterable
)
2701 self
.__cache
= [] if _cache
is None else _cache
2702 self
.__reversed
= reverse
2706 # We need to consume the entire iterable to iterate in reverse
2707 yield from self
.exhaust()
2709 yield from self
.__cache
2710 for item
in self
.__iterable
:
2711 self
.__cache
.append(item
)
2714 def __exhaust(self
):
2715 self
.__cache
.extend(self
.__iterable
)
2716 # Discard the emptied iterable to make it pickle-able
2717 self
.__iterable
= []
2721 ''' Evaluate the entire iterable '''
2722 return self
.__exhaust
()[::-1 if self
.__reversed
else 1]
2725 def __reverse_index(x
):
2726 return None if x
is None else -(x
+ 1)
2728 def __getitem__(self
, idx
):
2729 if isinstance(idx
, slice):
2731 idx
= slice(self
.__reverse
_index
(idx
.start
), self
.__reverse
_index
(idx
.stop
), -(idx
.step
or 1))
2732 start
, stop
, step
= idx
.start
, idx
.stop
, idx
.step
or 1
2733 elif isinstance(idx
, int):
2735 idx
= self
.__reverse
_index
(idx
)
2736 start
, stop
, step
= idx
, idx
, 0
2738 raise TypeError('indices must be integers or slices')
2739 if ((start
or 0) < 0 or (stop
or 0) < 0
2740 or (start
is None and step
< 0)
2741 or (stop
is None and step
> 0)):
2742 # We need to consume the entire iterable to be able to slice from the end
2743 # Obviously, never use this with infinite iterables
2746 return self
.__cache
[idx
]
2747 except IndexError as e
:
2748 raise self
.IndexError(e
) from e
2749 n
= max(start
or 0, stop
or 0) - len(self
.__cache
) + 1
2751 self
.__cache
.extend(itertools
.islice(self
.__iterable
, n
))
2753 return self
.__cache
[idx
]
2754 except IndexError as e
:
2755 raise self
.IndexError(e
) from e
2759 self
[-1] if self
.__reversed
else self
[0]
2760 except self
.IndexError:
2766 return len(self
.__cache
)
2768 def __reversed__(self
):
2769 return type(self
)(self
.__iterable
, reverse
=not self
.__reversed
, _cache
=self
.__cache
)
2772 return type(self
)(self
.__iterable
, reverse
=self
.__reversed
, _cache
=self
.__cache
)
2775 # repr and str should mimic a list. So we exhaust the iterable
2776 return repr(self
.exhaust())
2779 return repr(self
.exhaust())
2784 class IndexError(IndexError):
2788 # This is only useful for tests
2789 return len(self
.getslice())
2791 def __init__(self
, pagefunc
, pagesize
, use_cache
=True):
2792 self
._pagefunc
= pagefunc
2793 self
._pagesize
= pagesize
2794 self
._use
_cache
= use_cache
2797 def getpage(self
, pagenum
):
2798 page_results
= self
._cache
.get(pagenum
)
2799 if page_results
is None:
2800 page_results
= list(self
._pagefunc
(pagenum
))
2802 self
._cache
[pagenum
] = page_results
2805 def getslice(self
, start
=0, end
=None):
2806 return list(self
._getslice
(start
, end
))
2808 def _getslice(self
, start
, end
):
2809 raise NotImplementedError('This method must be implemented by subclasses')
2811 def __getitem__(self
, idx
):
2812 # NOTE: cache must be enabled if this is used
2813 if not isinstance(idx
, int) or idx
< 0:
2814 raise TypeError('indices must be non-negative integers')
2815 entries
= self
.getslice(idx
, idx
+ 1)
2817 raise self
.IndexError()
2821 class OnDemandPagedList(PagedList
):
2822 def _getslice(self
, start
, end
):
2823 for pagenum
in itertools
.count(start
// self
._pagesize
):
2824 firstid
= pagenum
* self
._pagesize
2825 nextfirstid
= pagenum
* self
._pagesize
+ self
._pagesize
2826 if start
>= nextfirstid
:
2830 start
% self
._pagesize
2831 if firstid
<= start
< nextfirstid
2834 ((end
- 1) % self
._pagesize
) + 1
2835 if (end
is not None and firstid
<= end
<= nextfirstid
)
2838 page_results
= self
.getpage(pagenum
)
2839 if startv
!= 0 or endv
is not None:
2840 page_results
= page_results
[startv
:endv
]
2841 yield from page_results
2843 # A little optimization - if current page is not "full", ie. does
2844 # not contain page_size videos then we can assume that this page
2845 # is the last one - there are no more ids on further pages -
2846 # i.e. no need to query again.
2847 if len(page_results
) + startv
< self
._pagesize
:
2850 # If we got the whole page, but the next page is not interesting,
2851 # break out early as well
2852 if end
== nextfirstid
:
2856 class InAdvancePagedList(PagedList
):
2857 def __init__(self
, pagefunc
, pagecount
, pagesize
):
2858 self
._pagecount
= pagecount
2859 PagedList
.__init
__(self
, pagefunc
, pagesize
, True)
2861 def _getslice(self
, start
, end
):
2862 start_page
= start
// self
._pagesize
2863 end_page
= self
._pagecount
if end
is None else min(self
._pagecount
, end
// self
._pagesize
+ 1)
2864 skip_elems
= start
- start_page
* self
._pagesize
2865 only_more
= None if end
is None else end
- start
2866 for pagenum
in range(start_page
, end_page
):
2867 page_results
= self
.getpage(pagenum
)
2869 page_results
= page_results
[skip_elems
:]
2871 if only_more
is not None:
2872 if len(page_results
) < only_more
:
2873 only_more
-= len(page_results
)
2875 yield from page_results
[:only_more
]
2877 yield from page_results
2880 def uppercase_escape(s
):
2881 unicode_escape
= codecs
.getdecoder('unicode_escape')
2883 r
'\\U[0-9a-fA-F]{8}',
2884 lambda m
: unicode_escape(m
.group(0))[0],
2888 def lowercase_escape(s
):
2889 unicode_escape
= codecs
.getdecoder('unicode_escape')
2891 r
'\\u[0-9a-fA-F]{4}',
2892 lambda m
: unicode_escape(m
.group(0))[0],
2896 def escape_rfc3986(s
):
2897 """Escape non-ASCII characters as suggested by RFC 3986"""
2898 if sys
.version_info
< (3, 0) and isinstance(s
, compat_str
):
2899 s
= s
.encode('utf-8')
2900 return compat_urllib_parse
.quote(s
, b
"%/;:@&=+$,!~*'()?#[]")
2903 def escape_url(url
):
2904 """Escape URL as suggested by RFC 3986"""
2905 url_parsed
= compat_urllib_parse_urlparse(url
)
2906 return url_parsed
._replace
(
2907 netloc
=url_parsed
.netloc
.encode('idna').decode('ascii'),
2908 path
=escape_rfc3986(url_parsed
.path
),
2909 params
=escape_rfc3986(url_parsed
.params
),
2910 query
=escape_rfc3986(url_parsed
.query
),
2911 fragment
=escape_rfc3986(url_parsed
.fragment
)
2916 return compat_parse_qs(compat_urllib_parse_urlparse(url
).query
)
2919 def read_batch_urls(batch_fd
):
2921 if not isinstance(url
, compat_str
):
2922 url
= url
.decode('utf-8', 'replace')
2923 BOM_UTF8
= ('\xef\xbb\xbf', '\ufeff')
2924 for bom
in BOM_UTF8
:
2925 if url
.startswith(bom
):
2926 url
= url
[len(bom
):]
2928 if not url
or url
.startswith(('#', ';', ']')):
2930 # "#" cannot be stripped out since it is part of the URI
2931 # However, it can be safely stipped out if follwing a whitespace
2932 return re
.split(r
'\s#', url
, 1)[0].rstrip()
2934 with contextlib
.closing(batch_fd
) as fd
:
2935 return [url
for url
in map(fixup
, fd
) if url
]
2938 def urlencode_postdata(*args
, **kargs
):
2939 return compat_urllib_parse_urlencode(*args
, **kargs
).encode('ascii')
2942 def update_url_query(url
, query
):
2945 parsed_url
= compat_urlparse
.urlparse(url
)
2946 qs
= compat_parse_qs(parsed_url
.query
)
2948 return compat_urlparse
.urlunparse(parsed_url
._replace
(
2949 query
=compat_urllib_parse_urlencode(qs
, True)))
2952 def update_Request(req
, url
=None, data
=None, headers
={}, query={}
):
2953 req_headers
= req
.headers
.copy()
2954 req_headers
.update(headers
)
2955 req_data
= data
or req
.data
2956 req_url
= update_url_query(url
or req
.get_full_url(), query
)
2957 req_get_method
= req
.get_method()
2958 if req_get_method
== 'HEAD':
2959 req_type
= HEADRequest
2960 elif req_get_method
== 'PUT':
2961 req_type
= PUTRequest
2963 req_type
= compat_urllib_request
.Request
2965 req_url
, data
=req_data
, headers
=req_headers
,
2966 origin_req_host
=req
.origin_req_host
, unverifiable
=req
.unverifiable
)
2967 if hasattr(req
, 'timeout'):
2968 new_req
.timeout
= req
.timeout
2972 def _multipart_encode_impl(data
, boundary
):
2973 content_type
= 'multipart/form-data; boundary=%s' % boundary
2976 for k
, v
in data
.items():
2977 out
+= b
'--' + boundary
.encode('ascii') + b
'\r\n'
2978 if isinstance(k
, compat_str
):
2979 k
= k
.encode('utf-8')
2980 if isinstance(v
, compat_str
):
2981 v
= v
.encode('utf-8')
2982 # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
2983 # suggests sending UTF-8 directly. Firefox sends UTF-8, too
2984 content
= b
'Content-Disposition: form-data; name="' + k
+ b
'"\r\n\r\n' + v
+ b
'\r\n'
2985 if boundary
.encode('ascii') in content
:
2986 raise ValueError('Boundary overlaps with data')
2989 out
+= b
'--' + boundary
.encode('ascii') + b
'--\r\n'
2991 return out
, content_type
2994 def multipart_encode(data
, boundary
=None):
2996 Encode a dict to RFC 7578-compliant form-data
2999 A dict where keys and values can be either Unicode or bytes-like
3002 If specified a Unicode object, it's used as the boundary. Otherwise
3003 a random boundary is generated.
3005 Reference: https://tools.ietf.org/html/rfc7578
3007 has_specified_boundary
= boundary
is not None
3010 if boundary
is None:
3011 boundary
= '---------------' + str(random
.randrange(0x0fffffff, 0xffffffff))
3014 out
, content_type
= _multipart_encode_impl(data
, boundary
)
3017 if has_specified_boundary
:
3021 return out
, content_type
3024 def dict_get(d
, key_or_keys
, default
=None, skip_false_values
=True):
3025 if isinstance(key_or_keys
, (list, tuple)):
3026 for key
in key_or_keys
:
3027 if key
not in d
or d
[key
] is None or skip_false_values
and not d
[key
]:
3031 return d
.get(key_or_keys
, default
)
3034 def try_get(src
, getter
, expected_type
=None):
3035 for get
in variadic(getter
):
3038 except (AttributeError, KeyError, TypeError, IndexError):
3041 if expected_type
is None or isinstance(v
, expected_type
):
3045 def merge_dicts(*dicts
):
3047 for a_dict
in dicts
:
3048 for k
, v
in a_dict
.items():
3052 or (isinstance(v
, compat_str
) and v
3053 and isinstance(merged
[k
], compat_str
)
3054 and not merged
[k
])):
3059 def encode_compat_str(string
, encoding
=preferredencoding(), errors
='strict'):
3060 return string
if isinstance(string
, compat_str
) else compat_str(string
, encoding
, errors
)
3072 TV_PARENTAL_GUIDELINES
= {
3082 def parse_age_limit(s
):
3084 return s
if 0 <= s
<= 21 else None
3085 if not isinstance(s
, compat_basestring
):
3087 m
= re
.match(r
'^(?P<age>\d{1,2})\+?$', s
)
3089 return int(m
.group('age'))
3092 return US_RATINGS
[s
]
3093 m
= re
.match(r
'^TV[_-]?(%s)$' % '|'.join(k
[3:] for k
in TV_PARENTAL_GUIDELINES
), s
)
3095 return TV_PARENTAL_GUIDELINES
['TV-' + m
.group(1)]
3099 def strip_jsonp(code
):
3102 (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
3103 (?:\s*&&\s*(?P=func_name))?
3104 \s*\(\s*(?P<callback_data>.*)\);?
3105 \s*?(?://[^\n]*)*$''',
3106 r
'\g<callback_data>', code
)
3109 def js_to_json(code
, vars={}):
3110 # vars is a dict of var, val pairs to substitute
3111 COMMENT_RE
= r
'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
3112 SKIP_RE
= r
'\s*(?:{comment})?\s*'.format(comment
=COMMENT_RE
)
3114 (r
'(?s)^(0[xX][0-9a-fA-F]+){skip}:?$'.format(skip
=SKIP_RE
), 16),
3115 (r
'(?s)^(0+[0-7]+){skip}:?$'.format(skip
=SKIP_RE
), 8),
3120 if v
in ('true', 'false', 'null'):
3122 elif v
in ('undefined', 'void 0'):
3124 elif v
.startswith('/*') or v
.startswith('//') or v
.startswith('!') or v
== ',':
3127 if v
[0] in ("'", '"'):
3128 v
= re
.sub(r
'(?s)\\.|"', lambda m
: {
3133 }.get(m
.group(0), m
.group(0)), v
[1:-1])
3135 for regex
, base
in INTEGER_TABLE
:
3136 im
= re
.match(regex
, v
)
3138 i
= int(im
.group(1), base
)
3139 return '"%d":' % i
if v
.endswith(':') else '%d' % i
3146 code
= re
.sub(r
'new Date\((".+")\)', r
'\g<1>', code
)
3148 return re
.sub(r
'''(?sx)
3149 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
3150 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
3151 {comment}|,(?={skip}[\]}}])|
3152 void\s0|(?:(?<![0-9])[eE]|[a-df-zA-DF-Z_$])[.a-zA-Z_$0-9]*|
3153 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
3156 '''.format(comment
=COMMENT_RE
, skip
=SKIP_RE
), fix_kv
, code
)
3159 def qualities(quality_ids
):
3160 """ Get a numeric quality value out of a list of possible values """
3163 return quality_ids
.index(qid
)
3169 POSTPROCESS_WHEN
= {'pre_process', 'before_dl', 'after_move', 'post_process', 'after_video', 'playlist'}
3173 'default': '%(title)s [%(id)s].%(ext)s',
3174 'chapter': '%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s',
3180 'description': 'description',
3181 'annotation': 'annotations.xml',
3182 'infojson': 'info.json',
3185 'pl_thumbnail': None,
3186 'pl_description': 'description',
3187 'pl_infojson': 'info.json',
3190 # As of [1] format syntax is:
3191 # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
3192 # 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
3193 STR_FORMAT_RE_TMPL
= r
'''(?x)
3194 (?<!%)(?P<prefix>(?:%%)*)
3196 (?P<has_key>\((?P<key>{0})\))?
3198 (?P<conversion>[#0\-+ ]+)?
3200 (?P<precision>\.\d+)?
3201 (?P<len_mod>[hlL])? # unused in python
3202 {1} # conversion type
3207 STR_FORMAT_TYPES
= 'diouxXeEfFgGcrs'
3210 def limit_length(s
, length
):
3211 """ Add ellipses to overly long strings """
3216 return s
[:length
- len(ELLIPSES
)] + ELLIPSES
3220 def version_tuple(v
):
3221 return tuple(int(e
) for e
in re
.split(r
'[-.]', v
))
3224 def is_outdated_version(version
, limit
, assume_new
=True):
3226 return not assume_new
3228 return version_tuple(version
) < version_tuple(limit
)
3230 return not assume_new
3233 def ytdl_is_updateable():
3234 """ Returns if yt-dlp can be updated with -U """
3236 from .update
import is_non_updateable
3238 return not is_non_updateable()
3241 def args_to_str(args
):
3242 # Get a short string representation for a subprocess command
3243 return ' '.join(compat_shlex_quote(a
) for a
in args
)
3246 def error_to_compat_str(err
):
3248 # On python 2 error byte string must be decoded with proper
3249 # encoding rather than ascii
3250 if sys
.version_info
[0] < 3:
3251 err_str
= err_str
.decode(preferredencoding())
3255 def mimetype2ext(mt
):
3259 mt
, _
, params
= mt
.partition(';')
3264 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
3265 # it's the most popular one
3266 'audio/mpeg': 'mp3',
3267 'audio/x-wav': 'wav',
3269 'audio/wave': 'wav',
3272 ext
= FULL_MAP
.get(mt
)
3278 'smptett+xml': 'tt',
3282 'x-mp4-fragmented': 'mp4',
3283 'x-ms-sami': 'sami',
3286 'x-mpegurl': 'm3u8',
3287 'vnd.apple.mpegurl': 'm3u8',
3291 'vnd.ms-sstr+xml': 'ism',
3295 'filmstrip+json': 'fs',
3299 _
, _
, subtype
= mt
.rpartition('/')
3300 ext
= SUBTYPE_MAP
.get(subtype
.lower())
3311 _
, _
, suffix
= subtype
.partition('+')
3312 ext
= SUFFIX_MAP
.get(suffix
)
3316 return subtype
.replace('+', '.')
3319 def ext2mimetype(ext_or_url
):
3322 if '.' not in ext_or_url
:
3323 ext_or_url
= f
'file.{ext_or_url}'
3324 return mimetypes
.guess_type(ext_or_url
)[0]
3327 def parse_codecs(codecs_str
):
3328 # http://tools.ietf.org/html/rfc6381
3331 split_codecs
= list(filter(None, map(
3332 str.strip
, codecs_str
.strip().strip(',').split(','))))
3333 vcodec
, acodec
, tcodec
, hdr
= None, None, None, None
3334 for full_codec
in split_codecs
:
3335 parts
= full_codec
.split('.')
3336 codec
= parts
[0].replace('0', '')
3337 if codec
in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2',
3338 'h263', 'h264', 'mp4v', 'hvc1', 'av1', 'theora', 'dvh1', 'dvhe'):
3340 vcodec
= '.'.join(parts
[:4]) if codec
in ('vp9', 'av1', 'hvc1') else full_codec
3341 if codec
in ('dvh1', 'dvhe'):
3343 elif codec
== 'av1' and len(parts
) > 3 and parts
[3] == '10':
3345 elif full_codec
.replace('0', '').startswith('vp9.2'):
3347 elif codec
in ('flac', 'mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
3350 elif codec
in ('stpp', 'wvtt',):
3354 write_string('WARNING: Unknown codec %s\n' % full_codec
, sys
.stderr
)
3355 if vcodec
or acodec
or tcodec
:
3357 'vcodec': vcodec
or 'none',
3358 'acodec': acodec
or 'none',
3359 'dynamic_range': hdr
,
3360 **({'tcodec': tcodec}
if tcodec
is not None else {}),
3362 elif len(split_codecs
) == 2:
3364 'vcodec': split_codecs
[0],
3365 'acodec': split_codecs
[1],
3370 def urlhandle_detect_ext(url_handle
):
3371 getheader
= url_handle
.headers
.get
3373 cd
= getheader('Content-Disposition')
3375 m
= re
.match(r
'attachment;\s*filename="(?P<filename>[^"]+)"', cd
)
3377 e
= determine_ext(m
.group('filename'), default_ext
=None)
3381 return mimetype2ext(getheader('Content-Type'))
3384 def encode_data_uri(data
, mime_type
):
3385 return 'data:%s;base64,%s' % (mime_type
, base64
.b64encode(data
).decode('ascii'))
3388 def age_restricted(content_limit
, age_limit
):
3389 """ Returns True iff the content should be blocked """
3391 if age_limit
is None: # No limit set
3393 if content_limit
is None:
3394 return False # Content available for everyone
3395 return age_limit
< content_limit
3398 def is_html(first_bytes
):
3399 """ Detect whether a file contains HTML by examining its first bytes. """
3402 (b
'\xef\xbb\xbf', 'utf-8'),
3403 (b
'\x00\x00\xfe\xff', 'utf-32-be'),
3404 (b
'\xff\xfe\x00\x00', 'utf-32-le'),
3405 (b
'\xff\xfe', 'utf-16-le'),
3406 (b
'\xfe\xff', 'utf-16-be'),
3408 for bom
, enc
in BOMS
:
3409 if first_bytes
.startswith(bom
):
3410 s
= first_bytes
[len(bom
):].decode(enc
, 'replace')
3413 s
= first_bytes
.decode('utf-8', 'replace')
3415 return re
.match(r
'^\s*<', s
)
3418 def determine_protocol(info_dict
):
3419 protocol
= info_dict
.get('protocol')
3420 if protocol
is not None:
3423 url
= sanitize_url(info_dict
['url'])
3424 if url
.startswith('rtmp'):
3426 elif url
.startswith('mms'):
3428 elif url
.startswith('rtsp'):
3431 ext
= determine_ext(url
)
3437 return compat_urllib_parse_urlparse(url
).scheme
3440 def render_table(header_row
, data
, delim
=False, extra_gap
=0, hide_empty
=False):
3441 """ Render a list of rows, each as a list of values.
3442 Text after a \t will be right aligned """
3444 return len(remove_terminal_sequences(string
).replace('\t', ''))
3446 def get_max_lens(table
):
3447 return [max(width(str(v
)) for v
in col
) for col
in zip(*table
)]
3449 def filter_using_list(row
, filterArray
):
3450 return [col
for take
, col
in itertools
.zip_longest(filterArray
, row
, fillvalue
=True) if take
]
3452 max_lens
= get_max_lens(data
) if hide_empty
else []
3453 header_row
= filter_using_list(header_row
, max_lens
)
3454 data
= [filter_using_list(row
, max_lens
) for row
in data
]
3456 table
= [header_row
] + data
3457 max_lens
= get_max_lens(table
)
3460 table
= [header_row
, [delim
* (ml
+ extra_gap
) for ml
in max_lens
]] + data
3461 table
[1][-1] = table
[1][-1][:-extra_gap
] # Remove extra_gap from end of delimiter
3463 for pos
, text
in enumerate(map(str, row
)):
3465 row
[pos
] = text
.replace('\t', ' ' * (max_lens
[pos
] - width(text
))) + ' ' * extra_gap
3467 row
[pos
] = text
+ ' ' * (max_lens
[pos
] - width(text
) + extra_gap
)
3468 ret
= '\n'.join(''.join(row
).rstrip() for row
in table
)
3472 def _match_one(filter_part
, dct
, incomplete
):
3473 # TODO: Generalize code with YoutubeDL._build_format_filter
3474 STRING_OPERATORS
= {
3475 '*=': operator
.contains
,
3476 '^=': lambda attr
, value
: attr
.startswith(value
),
3477 '$=': lambda attr
, value
: attr
.endswith(value
),
3478 '~=': lambda attr
, value
: re
.search(value
, attr
),
3480 COMPARISON_OPERATORS
= {
3482 '<=': operator
.le
, # "<=" must be defined above "<"
3489 operator_rex
= re
.compile(r
'''(?x)\s*
3491 \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
3493 (?P<quote>["\'])(?P
<quotedstrval
>.+?
)(?P
=quote
)|
3497 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
3498 m = operator_rex.search(filter_part)
3501 unnegated_op = COMPARISON_OPERATORS[m['op']]
3503 op = lambda attr, value: not unnegated_op(attr, value)
3506 comparison_value = m['quotedstrval'] or m['strval'] or m['intval']
3508 comparison_value = comparison_value.replace(r'\%s' % m['quote'], m['quote'])
3509 actual_value = dct.get(m['key'])
3510 numeric_comparison = None
3511 if isinstance(actual_value, compat_numeric_types):
3512 # If the original field is a string and matching comparisonvalue is
3513 # a number we should respect the origin of the original field
3514 # and process comparison value as a string (see
3515 # https://github.com/ytdl-org/youtube-dl/issues/11082)
3517 numeric_comparison = int(comparison_value)
3519 numeric_comparison = parse_filesize(comparison_value)
3520 if numeric_comparison is None:
3521 numeric_comparison = parse_filesize(f'{comparison_value}B')
3522 if numeric_comparison is None:
3523 numeric_comparison = parse_duration(comparison_value)
3524 if numeric_comparison is not None and m['op'] in STRING_OPERATORS:
3525 raise ValueError('Operator %s only supports string values!' % m['op'])
3526 if actual_value is None:
3527 return incomplete or m['none_inclusive']
3528 return op(actual_value, comparison_value if numeric_comparison is None else numeric_comparison)
3531 '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
3532 '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
3534 operator_rex = re.compile(r'''(?x
)\s
*
3535 (?P
<op
>%s)\s
*(?P
<key
>[a
-z_
]+)
3537 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
3538 m = operator_rex.search(filter_part)
3540 op = UNARY_OPERATORS[m.group('op')]
3541 actual_value = dct.get(m.group('key'))
3542 if incomplete and actual_value is None:
3544 return op(actual_value)
3546 raise ValueError('Invalid filter part %r' % filter_part)
3549 def match_str(filter_str, dct, incomplete=False):
3550 """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false
3551 When incomplete, all conditions passes on missing fields
3554 _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
3555 for filter_part in re.split(r'(?<!\\)&', filter_str))
3558 def match_filter_func(filter_str):
3559 def _match_func(info_dict, *args, **kwargs):
3560 if match_str(filter_str, info_dict, *args, **kwargs):
3563 video_title = info_dict.get('title', info_dict.get('id', 'video'))
3564 return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
3568 def parse_dfxp_time_expr(time_expr):
3572 mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
3574 return float(mobj.group('time_offset'))
3576 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
3578 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
3581 def srt_subtitles_timecode(seconds):
3582 return '%02d:%02d:%02d,%03d' % timetuple_from_msec(seconds * 1000)
3585 def ass_subtitles_timecode(seconds):
3586 time = timetuple_from_msec(seconds * 1000)
3587 return '%01d:%02d:%02d.%02d' % (*time[:-1], time.milliseconds / 10)
3590 def dfxp2srt(dfxp_data):
3592 @param dfxp_data A
bytes-like
object containing DFXP data
3593 @returns A
unicode object containing converted SRT data
3595 LEGACY_NAMESPACES = (
3596 (b'http://www.w3.org/ns/ttml', [
3597 b'http://www.w3.org/2004/11/ttaf1',
3598 b'http://www.w3.org/2006/04/ttaf1',
3599 b'http://www.w3.org/2006/10/ttaf1',
3601 (b'http://www.w3.org/ns/ttml#styling', [
3602 b'http://www.w3.org/ns/ttml#style',
3606 SUPPORTED_STYLING = [
3615 _x = functools.partial(xpath_with_ns, ns_map={
3616 'xml': 'http://www.w3.org/XML/1998/namespace',
3617 'ttml': 'http://www.w3.org/ns/ttml',
3618 'tts': 'http://www.w3.org/ns/ttml#styling',
3624 class TTMLPElementParser(object):
3626 _unclosed_elements = []
3627 _applied_styles = []
3629 def start(self, tag, attrib):
3630 if tag in (_x('ttml:br'), 'br'):
3633 unclosed_elements = []
3635 element_style_id = attrib.get('style')
3637 style.update(default_style)
3638 if element_style_id:
3639 style.update(styles.get(element_style_id, {}))
3640 for prop in SUPPORTED_STYLING:
3641 prop_val = attrib.get(_x('tts:' + prop))
3643 style[prop] = prop_val
3646 for k, v in sorted(style.items()):
3647 if self._applied_styles and self._applied_styles[-1].get(k) == v:
3650 font += ' color="%s"' % v
3651 elif k == 'fontSize':
3652 font += ' size="%s"' % v
3653 elif k == 'fontFamily':
3654 font += ' face="%s"' % v
3655 elif k == 'fontWeight' and v == 'bold':
3657 unclosed_elements.append('b')
3658 elif k == 'fontStyle' and v == 'italic':
3660 unclosed_elements.append('i')
3661 elif k == 'textDecoration' and v == 'underline':
3663 unclosed_elements.append('u')
3665 self._out += '<font' + font + '>'
3666 unclosed_elements.append('font')
3668 if self._applied_styles:
3669 applied_style.update(self._applied_styles[-1])
3670 applied_style.update(style)
3671 self._applied_styles.append(applied_style)
3672 self._unclosed_elements.append(unclosed_elements)
3675 if tag not in (_x('ttml:br'), 'br'):
3676 unclosed_elements = self._unclosed_elements.pop()
3677 for element in reversed(unclosed_elements):
3678 self._out += '</%s>' % element
3679 if unclosed_elements and self._applied_styles:
3680 self._applied_styles.pop()
3682 def data(self, data):
3686 return self._out.strip()
3688 def parse_node(node):
3689 target = TTMLPElementParser()
3690 parser = xml.etree.ElementTree.XMLParser(target=target)
3691 parser.feed(xml.etree.ElementTree.tostring(node))
3692 return parser.close()
3694 for k, v in LEGACY_NAMESPACES:
3696 dfxp_data = dfxp_data.replace(ns, k)
3698 dfxp = compat_etree_fromstring(dfxp_data)
3700 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
3703 raise ValueError('Invalid dfxp/TTML subtitle')
3707 for style in dfxp.findall(_x('.//ttml:style')):
3708 style_id = style.get('id') or style.get(_x('xml:id'))
3711 parent_style_id = style.get('style')
3713 if parent_style_id not in styles:
3716 styles[style_id] = styles[parent_style_id].copy()
3717 for prop in SUPPORTED_STYLING:
3718 prop_val = style.get(_x('tts:' + prop))
3720 styles.setdefault(style_id, {})[prop] = prop_val
3726 for p in ('body', 'div'):
3727 ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
3730 style = styles.get(ele.get('style'))
3733 default_style.update(style)
3735 for para, index in zip(paras, itertools.count(1)):
3736 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
3737 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
3738 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
3739 if begin_time is None:
3744 end_time = begin_time + dur
3745 out.append('%d\n%s --> %s\n%s\n\n' % (
3747 srt_subtitles_timecode(begin_time),
3748 srt_subtitles_timecode(end_time),
3754 def cli_option(params, command_option, param):
3755 param = params.get(param)
3757 param = compat_str(param)
3758 return [command_option, param] if param is not None else []
3761 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
3762 param = params.get(param)
3765 assert isinstance(param, bool)
3767 return [command_option + separator + (true_value if param else false_value)]
3768 return [command_option, true_value if param else false_value]
3771 def cli_valueless_option(params, command_option, param, expected_value=True):
3772 param = params.get(param)
3773 return [command_option] if param == expected_value else []
3776 def cli_configuration_args(argdict, keys, default=[], use_compat=True):
3777 if isinstance(argdict, (list, tuple)): # for backward compatibility
3784 assert isinstance(argdict, dict)
3786 assert isinstance(keys, (list, tuple))
3787 for key_list in keys:
3788 arg_list = list(filter(
3789 lambda x: x is not None,
3790 [argdict.get(key.lower()) for key in variadic(key_list)]))
3792 return [arg for args in arg_list for arg in args]
3796 def _configuration_args(main_key, argdict, exe, keys=None, default=[], use_compat=True):
3797 main_key, exe = main_key.lower(), exe.lower()
3798 root_key = exe if main_key == exe else f'{main_key}+{exe}'
3799 keys = [f'{root_key}{k}' for k in (keys or [''])]
3800 if root_key in keys:
3802 keys.append((main_key, exe))
3803 keys.append('default')
3806 return cli_configuration_args(argdict, keys, default, use_compat)
3809 class ISO639Utils(object):
3810 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
3869 'iw': 'heb', # Replaced by he in 1989 revision
3879 'in': 'ind', # Replaced by id in 1989 revision
3994 'ji': 'yid', # Replaced by yi in 1989 revision
4002 def short2long(cls, code):
4003 """Convert language code from ISO 639-1 to ISO 639-2/T"""
4004 return cls._lang_map.get(code[:2])
4007 def long2short(cls, code):
4008 """Convert language code from ISO 639-2/T to ISO 639-1"""
4009 for short_name, long_name in cls._lang_map.items():
4010 if long_name == code:
4014 class ISO3166Utils(object):
4015 # From http://data.okfn.org/data/core/country-list
4017 'AF': 'Afghanistan',
4018 'AX': 'Åland Islands',
4021 'AS': 'American Samoa',
4026 'AG': 'Antigua and Barbuda',
4043 'BO': 'Bolivia, Plurinational State of',
4044 'BQ': 'Bonaire, Sint Eustatius and Saba',
4045 'BA': 'Bosnia and Herzegovina',
4047 'BV': 'Bouvet Island',
4049 'IO': 'British Indian Ocean Territory',
4050 'BN': 'Brunei Darussalam',
4052 'BF': 'Burkina Faso',
4058 'KY': 'Cayman Islands',
4059 'CF': 'Central African Republic',
4063 'CX': 'Christmas Island',
4064 'CC': 'Cocos (Keeling) Islands',
4068 'CD': 'Congo, the Democratic Republic of the',
4069 'CK': 'Cook Islands',
4071 'CI': 'Côte d\'Ivoire',
4076 'CZ': 'Czech Republic',
4080 'DO': 'Dominican Republic',
4083 'SV': 'El Salvador',
4084 'GQ': 'Equatorial Guinea',
4088 'FK': 'Falkland Islands (Malvinas)',
4089 'FO': 'Faroe Islands',
4093 'GF': 'French Guiana',
4094 'PF': 'French Polynesia',
4095 'TF': 'French Southern Territories',
4110 'GW': 'Guinea-Bissau',
4113 'HM': 'Heard Island and McDonald Islands',
4114 'VA': 'Holy See (Vatican City State)',
4121 'IR': 'Iran, Islamic Republic of',
4124 'IM': 'Isle of Man',
4134 'KP': 'Korea, Democratic People\'s Republic of',
4135 'KR': 'Korea, Republic of',
4138 'LA': 'Lao People\'s Democratic Republic',
4144 'LI': 'Liechtenstein',
4148 'MK': 'Macedonia, the Former Yugoslav Republic of',
4155 'MH': 'Marshall Islands',
4161 'FM': 'Micronesia, Federated States of',
4162 'MD': 'Moldova, Republic of',
4173 'NL': 'Netherlands',
4174 'NC': 'New Caledonia',
4175 'NZ': 'New Zealand',
4180 'NF': 'Norfolk Island',
4181 'MP': 'Northern Mariana Islands',
4186 'PS': 'Palestine, State of',
4188 'PG': 'Papua New Guinea',
4191 'PH': 'Philippines',
4195 'PR': 'Puerto Rico',
4199 'RU': 'Russian Federation',
4201 'BL': 'Saint Barthélemy',
4202 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
4203 'KN': 'Saint Kitts and Nevis',
4204 'LC': 'Saint Lucia',
4205 'MF': 'Saint Martin (French part)',
4206 'PM': 'Saint Pierre and Miquelon',
4207 'VC': 'Saint Vincent and the Grenadines',
4210 'ST': 'Sao Tome and Principe',
4211 'SA': 'Saudi Arabia',
4215 'SL': 'Sierra Leone',
4217 'SX': 'Sint Maarten (Dutch part)',
4220 'SB': 'Solomon Islands',
4222 'ZA': 'South Africa',
4223 'GS': 'South Georgia and the South Sandwich Islands',
4224 'SS': 'South Sudan',
4229 'SJ': 'Svalbard and Jan Mayen',
4232 'CH': 'Switzerland',
4233 'SY': 'Syrian Arab Republic',
4234 'TW': 'Taiwan, Province of China',
4236 'TZ': 'Tanzania, United Republic of',
4238 'TL': 'Timor-Leste',
4242 'TT': 'Trinidad and Tobago',
4245 'TM': 'Turkmenistan',
4246 'TC': 'Turks and Caicos Islands',
4250 'AE': 'United Arab Emirates',
4251 'GB': 'United Kingdom',
4252 'US': 'United States',
4253 'UM': 'United States Minor Outlying Islands',
4257 'VE': 'Venezuela, Bolivarian Republic of',
4259 'VG': 'Virgin Islands, British',
4260 'VI': 'Virgin Islands, U.S.',
4261 'WF': 'Wallis and Futuna',
4262 'EH': 'Western Sahara',
4269 def short2full(cls, code):
4270 """Convert an ISO 3166-2 country code to the corresponding full name"""
4271 return cls._country_map.get(code.upper())
4274 class GeoUtils(object):
4275 # Major IPv4 address blocks per country
4277 'AD': '46.172.224.0/19',
4278 'AE': '94.200.0.0/13',
4279 'AF': '149.54.0.0/17',
4280 'AG': '209.59.64.0/18',
4281 'AI': '204.14.248.0/21',
4282 'AL': '46.99.0.0/16',
4283 'AM': '46.70.0.0/15',
4284 'AO': '105.168.0.0/13',
4285 'AP': '182.50.184.0/21',
4286 'AQ': '23.154.160.0/24',
4287 'AR': '181.0.0.0/12',
4288 'AS': '202.70.112.0/20',
4289 'AT': '77.116.0.0/14',
4290 'AU': '1.128.0.0/11',
4291 'AW': '181.41.0.0/18',
4292 'AX': '185.217.4.0/22',
4293 'AZ': '5.197.0.0/16',
4294 'BA': '31.176.128.0/17',
4295 'BB': '65.48.128.0/17',
4296 'BD': '114.130.0.0/16',
4298 'BF': '102.178.0.0/15',
4299 'BG': '95.42.0.0/15',
4300 'BH': '37.131.0.0/17',
4301 'BI': '154.117.192.0/18',
4302 'BJ': '137.255.0.0/16',
4303 'BL': '185.212.72.0/23',
4304 'BM': '196.12.64.0/18',
4305 'BN': '156.31.0.0/16',
4306 'BO': '161.56.0.0/16',
4307 'BQ': '161.0.80.0/20',
4308 'BR': '191.128.0.0/12',
4309 'BS': '24.51.64.0/18',
4310 'BT': '119.2.96.0/19',
4311 'BW': '168.167.0.0/16',
4312 'BY': '178.120.0.0/13',
4313 'BZ': '179.42.192.0/18',
4314 'CA': '99.224.0.0/11',
4315 'CD': '41.243.0.0/16',
4316 'CF': '197.242.176.0/21',
4317 'CG': '160.113.0.0/16',
4318 'CH': '85.0.0.0/13',
4319 'CI': '102.136.0.0/14',
4320 'CK': '202.65.32.0/19',
4321 'CL': '152.172.0.0/14',
4322 'CM': '102.244.0.0/14',
4323 'CN': '36.128.0.0/10',
4324 'CO': '181.240.0.0/12',
4325 'CR': '201.192.0.0/12',
4326 'CU': '152.206.0.0/15',
4327 'CV': '165.90.96.0/19',
4328 'CW': '190.88.128.0/17',
4329 'CY': '31.153.0.0/16',
4330 'CZ': '88.100.0.0/14',
4332 'DJ': '197.241.0.0/17',
4333 'DK': '87.48.0.0/12',
4334 'DM': '192.243.48.0/20',
4335 'DO': '152.166.0.0/15',
4336 'DZ': '41.96.0.0/12',
4337 'EC': '186.68.0.0/15',
4338 'EE': '90.190.0.0/15',
4339 'EG': '156.160.0.0/11',
4340 'ER': '196.200.96.0/20',
4341 'ES': '88.0.0.0/11',
4342 'ET': '196.188.0.0/14',
4343 'EU': '2.16.0.0/13',
4344 'FI': '91.152.0.0/13',
4345 'FJ': '144.120.0.0/16',
4346 'FK': '80.73.208.0/21',
4347 'FM': '119.252.112.0/20',
4348 'FO': '88.85.32.0/19',
4350 'GA': '41.158.0.0/15',
4352 'GD': '74.122.88.0/21',
4353 'GE': '31.146.0.0/16',
4354 'GF': '161.22.64.0/18',
4355 'GG': '62.68.160.0/19',
4356 'GH': '154.160.0.0/12',
4357 'GI': '95.164.0.0/16',
4358 'GL': '88.83.0.0/19',
4359 'GM': '160.182.0.0/15',
4360 'GN': '197.149.192.0/18',
4361 'GP': '104.250.0.0/19',
4362 'GQ': '105.235.224.0/20',
4363 'GR': '94.64.0.0/13',
4364 'GT': '168.234.0.0/16',
4365 'GU': '168.123.0.0/16',
4366 'GW': '197.214.80.0/20',
4367 'GY': '181.41.64.0/18',
4368 'HK': '113.252.0.0/14',
4369 'HN': '181.210.0.0/16',
4370 'HR': '93.136.0.0/13',
4371 'HT': '148.102.128.0/17',
4372 'HU': '84.0.0.0/14',
4373 'ID': '39.192.0.0/10',
4374 'IE': '87.32.0.0/12',
4375 'IL': '79.176.0.0/13',
4376 'IM': '5.62.80.0/20',
4377 'IN': '117.192.0.0/10',
4378 'IO': '203.83.48.0/21',
4379 'IQ': '37.236.0.0/14',
4380 'IR': '2.176.0.0/12',
4381 'IS': '82.221.0.0/16',
4382 'IT': '79.0.0.0/10',
4383 'JE': '87.244.64.0/18',
4384 'JM': '72.27.0.0/17',
4385 'JO': '176.29.0.0/16',
4386 'JP': '133.0.0.0/8',
4387 'KE': '105.48.0.0/12',
4388 'KG': '158.181.128.0/17',
4389 'KH': '36.37.128.0/17',
4390 'KI': '103.25.140.0/22',
4391 'KM': '197.255.224.0/20',
4392 'KN': '198.167.192.0/19',
4393 'KP': '175.45.176.0/22',
4394 'KR': '175.192.0.0/10',
4395 'KW': '37.36.0.0/14',
4396 'KY': '64.96.0.0/15',
4397 'KZ': '2.72.0.0/13',
4398 'LA': '115.84.64.0/18',
4399 'LB': '178.135.0.0/16',
4400 'LC': '24.92.144.0/20',
4401 'LI': '82.117.0.0/19',
4402 'LK': '112.134.0.0/15',
4403 'LR': '102.183.0.0/16',
4404 'LS': '129.232.0.0/17',
4405 'LT': '78.56.0.0/13',
4406 'LU': '188.42.0.0/16',
4407 'LV': '46.109.0.0/16',
4408 'LY': '41.252.0.0/14',
4409 'MA': '105.128.0.0/11',
4410 'MC': '88.209.64.0/18',
4411 'MD': '37.246.0.0/16',
4412 'ME': '178.175.0.0/17',
4413 'MF': '74.112.232.0/21',
4414 'MG': '154.126.0.0/17',
4415 'MH': '117.103.88.0/21',
4416 'MK': '77.28.0.0/15',
4417 'ML': '154.118.128.0/18',
4418 'MM': '37.111.0.0/17',
4419 'MN': '49.0.128.0/17',
4420 'MO': '60.246.0.0/16',
4421 'MP': '202.88.64.0/20',
4422 'MQ': '109.203.224.0/19',
4423 'MR': '41.188.64.0/18',
4424 'MS': '208.90.112.0/22',
4425 'MT': '46.11.0.0/16',
4426 'MU': '105.16.0.0/12',
4427 'MV': '27.114.128.0/18',
4428 'MW': '102.70.0.0/15',
4429 'MX': '187.192.0.0/11',
4430 'MY': '175.136.0.0/13',
4431 'MZ': '197.218.0.0/15',
4432 'NA': '41.182.0.0/16',
4433 'NC': '101.101.0.0/18',
4434 'NE': '197.214.0.0/18',
4435 'NF': '203.17.240.0/22',
4436 'NG': '105.112.0.0/12',
4437 'NI': '186.76.0.0/15',
4438 'NL': '145.96.0.0/11',
4439 'NO': '84.208.0.0/13',
4440 'NP': '36.252.0.0/15',
4441 'NR': '203.98.224.0/19',
4442 'NU': '49.156.48.0/22',
4443 'NZ': '49.224.0.0/14',
4444 'OM': '5.36.0.0/15',
4445 'PA': '186.72.0.0/15',
4446 'PE': '186.160.0.0/14',
4447 'PF': '123.50.64.0/18',
4448 'PG': '124.240.192.0/19',
4449 'PH': '49.144.0.0/13',
4450 'PK': '39.32.0.0/11',
4451 'PL': '83.0.0.0/11',
4452 'PM': '70.36.0.0/20',
4453 'PR': '66.50.0.0/16',
4454 'PS': '188.161.0.0/16',
4455 'PT': '85.240.0.0/13',
4456 'PW': '202.124.224.0/20',
4457 'PY': '181.120.0.0/14',
4458 'QA': '37.210.0.0/15',
4459 'RE': '102.35.0.0/16',
4460 'RO': '79.112.0.0/13',
4461 'RS': '93.86.0.0/15',
4462 'RU': '5.136.0.0/13',
4463 'RW': '41.186.0.0/16',
4464 'SA': '188.48.0.0/13',
4465 'SB': '202.1.160.0/19',
4466 'SC': '154.192.0.0/11',
4467 'SD': '102.120.0.0/13',
4468 'SE': '78.64.0.0/12',
4469 'SG': '8.128.0.0/10',
4470 'SI': '188.196.0.0/14',
4471 'SK': '78.98.0.0/15',
4472 'SL': '102.143.0.0/17',
4473 'SM': '89.186.32.0/19',
4474 'SN': '41.82.0.0/15',
4475 'SO': '154.115.192.0/18',
4476 'SR': '186.179.128.0/17',
4477 'SS': '105.235.208.0/21',
4478 'ST': '197.159.160.0/19',
4479 'SV': '168.243.0.0/16',
4480 'SX': '190.102.0.0/20',
4482 'SZ': '41.84.224.0/19',
4483 'TC': '65.255.48.0/20',
4484 'TD': '154.68.128.0/19',
4485 'TG': '196.168.0.0/14',
4486 'TH': '171.96.0.0/13',
4487 'TJ': '85.9.128.0/18',
4488 'TK': '27.96.24.0/21',
4489 'TL': '180.189.160.0/20',
4490 'TM': '95.85.96.0/19',
4491 'TN': '197.0.0.0/11',
4492 'TO': '175.176.144.0/21',
4493 'TR': '78.160.0.0/11',
4494 'TT': '186.44.0.0/15',
4495 'TV': '202.2.96.0/19',
4496 'TW': '120.96.0.0/11',
4497 'TZ': '156.156.0.0/14',
4498 'UA': '37.52.0.0/14',
4499 'UG': '102.80.0.0/13',
4501 'UY': '167.56.0.0/13',
4502 'UZ': '84.54.64.0/18',
4503 'VA': '212.77.0.0/19',
4504 'VC': '207.191.240.0/21',
4505 'VE': '186.88.0.0/13',
4506 'VG': '66.81.192.0/20',
4507 'VI': '146.226.0.0/16',
4508 'VN': '14.160.0.0/11',
4509 'VU': '202.80.32.0/20',
4510 'WF': '117.20.32.0/21',
4511 'WS': '202.4.32.0/19',
4512 'YE': '134.35.0.0/16',
4513 'YT': '41.242.116.0/22',
4514 'ZA': '41.0.0.0/11',
4515 'ZM': '102.144.0.0/13',
4516 'ZW': '102.177.192.0/18',
4520 def random_ipv4(cls, code_or_block):
4521 if len(code_or_block) == 2:
4522 block = cls._country_ip_map.get(code_or_block.upper())
4526 block = code_or_block
4527 addr, preflen = block.split('/')
4528 addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
4529 addr_max = addr_min | (0xffffffff >> int(preflen))
4530 return compat_str(socket.inet_ntoa(
4531 compat_struct_pack('!L', random.randint(addr_min, addr_max))))
4534 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
4535 def __init__(self, proxies=None):
4536 # Set default handlers
4537 for type in ('http', 'https'):
4538 setattr(self, '%s_open' % type,
4539 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
4540 meth(r, proxy, type))
4541 compat_urllib_request.ProxyHandler.__init__(self, proxies)
4543 def proxy_open(self, req, proxy, type):
4544 req_proxy = req.headers.get('Ytdl-request-proxy')
4545 if req_proxy is not None:
4547 del req.headers['Ytdl-request-proxy']
4549 if proxy == '__noproxy__':
4550 return None # No Proxy
4551 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
4552 req.add_header('Ytdl-socks-proxy', proxy)
4553 # yt-dlp's http/https handlers do wrapping the socket with socks
4555 return compat_urllib_request.ProxyHandler.proxy_open(
4556 self, req, proxy, type)
4559 # Both long_to_bytes and bytes_to_long are adapted from PyCrypto, which is
4560 # released into Public Domain
4561 # https://github.com/dlitz/pycrypto/blob/master/lib/Crypto/Util/number.py#L387
4563 def long_to_bytes(n, blocksize=0):
4564 """long_to_bytes(n:long, blocksize:int) : string
4565 Convert a long integer to a byte string.
4567 If optional blocksize is given and greater than zero, pad the front of the
4568 byte string with binary zeros so that the length is a multiple of
4571 # after much testing, this algorithm was deemed to be the fastest
4575 s = compat_struct_pack('>I', n & 0xffffffff) + s
4577 # strip off leading zeros
4578 for i in range(len(s)):
4579 if s[i] != b'\000'[0]:
4582 # only happens when n == 0
4586 # add back some pad bytes. this could be done more efficiently w.r.t. the
4587 # de-padding being done above, but sigh...
4588 if blocksize > 0 and len(s) % blocksize:
4589 s = (blocksize - len(s) % blocksize) * b'\000' + s
4593 def bytes_to_long(s):
4594 """bytes_to_long(string) : long
4595 Convert a byte string to a long integer.
4597 This is (essentially) the inverse of long_to_bytes().
4602 extra = (4 - length % 4)
4603 s = b'\000' * extra + s
4604 length = length + extra
4605 for i in range(0, length, 4):
4606 acc = (acc << 32) + compat_struct_unpack('>I', s[i:i + 4])[0]
4610 def ohdave_rsa_encrypt(data, exponent, modulus):
4612 Implement OHDave
's RSA algorithm. See http://www.ohdave.com/rsa/
4615 data: data to encrypt, bytes-like object
4616 exponent, modulus: parameter e and N of RSA algorithm, both integer
4617 Output: hex string of encrypted data
4619 Limitation: supports one block encryption only
4622 payload = int(binascii.hexlify(data[::-1]), 16)
4623 encrypted = pow(payload, exponent, modulus)
4624 return '%x' % encrypted
4627 def pkcs1pad(data, length):
4629 Padding input data with PKCS#1 scheme
4631 @param {int[]} data input data
4632 @param {int} length target length
4633 @returns {int[]} padded data
4635 if len(data) > length - 11:
4636 raise ValueError('Input data too
long for PKCS
#1 padding')
4638 pseudo_random
= [random
.randint(0, 254) for _
in range(length
- len(data
) - 3)]
4639 return [0, 2] + pseudo_random
+ [0] + data
4642 def encode_base_n(num
, n
, table
=None):
4643 FULL_TABLE
= '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
4645 table
= FULL_TABLE
[:n
]
4648 raise ValueError('base %d exceeds table length %d' % (n
, len(table
)))
4655 ret
= table
[num
% n
] + ret
4660 def decode_packed_codes(code
):
4661 mobj
= re
.search(PACKED_CODES_RE
, code
)
4662 obfuscated_code
, base
, count
, symbols
= mobj
.groups()
4665 symbols
= symbols
.split('|')
4670 base_n_count
= encode_base_n(count
, base
)
4671 symbol_table
[base_n_count
] = symbols
[count
] or base_n_count
4674 r
'\b(\w+)\b', lambda mobj
: symbol_table
[mobj
.group(0)],
4678 def caesar(s
, alphabet
, shift
):
4683 alphabet
[(alphabet
.index(c
) + shift
) % l
] if c
in alphabet
else c
4688 return caesar(s
, r
'''!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~''', 47)
4691 def parse_m3u8_attributes(attrib
):
4693 for (key
, val
) in re
.findall(r
'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib
):
4694 if val
.startswith('"'):
4700 def urshift(val
, n
):
4701 return val
>> n
if val
>= 0 else (val
+ 0x100000000) >> n
4704 # Based on png2str() written by @gdkchan and improved by @yokrysty
4705 # Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
4706 def decode_png(png_data
):
4707 # Reference: https://www.w3.org/TR/PNG/
4708 header
= png_data
[8:]
4710 if png_data
[:8] != b
'\x89PNG\x0d\x0a\x1a\x0a' or header
[4:8] != b
'IHDR':
4711 raise IOError('Not a valid PNG file.')
4713 int_map
= {1: '>B', 2: '>H', 4: '>I'}
4714 unpack_integer
= lambda x
: compat_struct_unpack(int_map
[len(x
)], x
)[0]
4719 length
= unpack_integer(header
[:4])
4722 chunk_type
= header
[:4]
4725 chunk_data
= header
[:length
]
4726 header
= header
[length
:]
4728 header
= header
[4:] # Skip CRC
4736 ihdr
= chunks
[0]['data']
4738 width
= unpack_integer(ihdr
[:4])
4739 height
= unpack_integer(ihdr
[4:8])
4743 for chunk
in chunks
:
4744 if chunk
['type'] == b
'IDAT':
4745 idat
+= chunk
['data']
4748 raise IOError('Unable to read PNG data.')
4750 decompressed_data
= bytearray(zlib
.decompress(idat
))
4755 def _get_pixel(idx
):
4760 for y
in range(height
):
4761 basePos
= y
* (1 + stride
)
4762 filter_type
= decompressed_data
[basePos
]
4766 pixels
.append(current_row
)
4768 for x
in range(stride
):
4769 color
= decompressed_data
[1 + basePos
+ x
]
4770 basex
= y
* stride
+ x
4775 left
= _get_pixel(basex
- 3)
4777 up
= _get_pixel(basex
- stride
)
4779 if filter_type
== 1: # Sub
4780 color
= (color
+ left
) & 0xff
4781 elif filter_type
== 2: # Up
4782 color
= (color
+ up
) & 0xff
4783 elif filter_type
== 3: # Average
4784 color
= (color
+ ((left
+ up
) >> 1)) & 0xff
4785 elif filter_type
== 4: # Paeth
4791 c
= _get_pixel(basex
- stride
- 3)
4799 if pa
<= pb
and pa
<= pc
:
4800 color
= (color
+ a
) & 0xff
4802 color
= (color
+ b
) & 0xff
4804 color
= (color
+ c
) & 0xff
4806 current_row
.append(color
)
4808 return width
, height
, pixels
4811 def write_xattr(path
, key
, value
):
4812 # This mess below finds the best xattr tool for the job
4814 # try the pyxattr module...
4817 if hasattr(xattr
, 'set'): # pyxattr
4818 # Unicode arguments are not supported in python-pyxattr until
4820 # See https://github.com/ytdl-org/youtube-dl/issues/5498
4821 pyxattr_required_version
= '0.5.0'
4822 if version_tuple(xattr
.__version
__) < version_tuple(pyxattr_required_version
):
4823 # TODO: fallback to CLI tools
4824 raise XAttrUnavailableError(
4825 'python-pyxattr is detected but is too old. '
4826 'yt-dlp requires %s or above while your version is %s. '
4827 'Falling back to other xattr implementations' % (
4828 pyxattr_required_version
, xattr
.__version
__))
4830 setxattr
= xattr
.set
4832 setxattr
= xattr
.setxattr
4835 setxattr(path
, key
, value
)
4836 except EnvironmentError as e
:
4837 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4840 if compat_os_name
== 'nt':
4841 # Write xattrs to NTFS Alternate Data Streams:
4842 # http://en.wikipedia.org/wiki/NTFS#Alternate_data_streams_.28ADS.29
4843 assert ':' not in key
4844 assert os
.path
.exists(path
)
4846 ads_fn
= path
+ ':' + key
4848 with open(ads_fn
, 'wb') as f
:
4850 except EnvironmentError as e
:
4851 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4853 user_has_setfattr
= check_executable('setfattr', ['--version'])
4854 user_has_xattr
= check_executable('xattr', ['-h'])
4856 if user_has_setfattr
or user_has_xattr
:
4858 value
= value
.decode('utf-8')
4859 if user_has_setfattr
:
4860 executable
= 'setfattr'
4861 opts
= ['-n', key
, '-v', value
]
4862 elif user_has_xattr
:
4863 executable
= 'xattr'
4864 opts
= ['-w', key
, value
]
4866 cmd
= ([encodeFilename(executable
, True)]
4867 + [encodeArgument(o
) for o
in opts
]
4868 + [encodeFilename(path
, True)])
4872 cmd
, stdout
=subprocess
.PIPE
, stderr
=subprocess
.PIPE
, stdin
=subprocess
.PIPE
)
4873 except EnvironmentError as e
:
4874 raise XAttrMetadataError(e
.errno
, e
.strerror
)
4875 stdout
, stderr
= p
.communicate_or_kill()
4876 stderr
= stderr
.decode('utf-8', 'replace')
4877 if p
.returncode
!= 0:
4878 raise XAttrMetadataError(p
.returncode
, stderr
)
4881 # On Unix, and can't find pyxattr, setfattr, or xattr.
4882 if sys
.platform
.startswith('linux'):
4883 raise XAttrUnavailableError(
4884 "Couldn't find a tool to set the xattrs. "
4885 "Install either the python 'pyxattr' or 'xattr' "
4886 "modules, or the GNU 'attr' package "
4887 "(which contains the 'setfattr' tool).")
4889 raise XAttrUnavailableError(
4890 "Couldn't find a tool to set the xattrs. "
4891 "Install either the python 'xattr' module, "
4892 "or the 'xattr' binary.")
4895 def random_birthday(year_field
, month_field
, day_field
):
4896 start_date
= datetime
.date(1950, 1, 1)
4897 end_date
= datetime
.date(1995, 12, 31)
4898 offset
= random
.randint(0, (end_date
- start_date
).days
)
4899 random_date
= start_date
+ datetime
.timedelta(offset
)
4901 year_field
: str(random_date
.year
),
4902 month_field
: str(random_date
.month
),
4903 day_field
: str(random_date
.day
),
4907 # Templates for internet shortcut files, which are plain text files.
4908 DOT_URL_LINK_TEMPLATE
= '''
4913 DOT_WEBLOC_LINK_TEMPLATE
= '''
4914 <?xml version="1.0" encoding="UTF-8"?>
4915 <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
4916 <plist version="1.0">
4919 \t<string>%(url)s</string>
4924 DOT_DESKTOP_LINK_TEMPLATE
= '''
4934 'url': DOT_URL_LINK_TEMPLATE
,
4935 'desktop': DOT_DESKTOP_LINK_TEMPLATE
,
4936 'webloc': DOT_WEBLOC_LINK_TEMPLATE
,
4940 def iri_to_uri(iri
):
4942 Converts an IRI (Internationalized Resource Identifier, allowing Unicode characters) to a URI (Uniform Resource Identifier, ASCII-only).
4944 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.
4947 iri_parts
= compat_urllib_parse_urlparse(iri
)
4949 if '[' in iri_parts
.netloc
:
4950 raise ValueError('IPv6 URIs are not, yet, supported.')
4951 # Querying `.netloc`, when there's only one bracket, also raises a ValueError.
4953 # 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.
4956 if iri_parts
.username
:
4957 net_location
+= compat_urllib_parse_quote(iri_parts
.username
, safe
=r
"!$%&'()*+,~")
4958 if iri_parts
.password
is not None:
4959 net_location
+= ':' + compat_urllib_parse_quote(iri_parts
.password
, safe
=r
"!$%&'()*+,~")
4962 net_location
+= iri_parts
.hostname
.encode('idna').decode('utf-8') # Punycode for Unicode hostnames.
4963 # The 'idna' encoding produces ASCII text.
4964 if iri_parts
.port
is not None and iri_parts
.port
!= 80:
4965 net_location
+= ':' + str(iri_parts
.port
)
4967 return compat_urllib_parse_urlunparse(
4971 compat_urllib_parse_quote_plus(iri_parts
.path
, safe
=r
"!$%&'()*+,/:;=@|~"),
4973 # Unsure about the `safe` argument, since this is a legacy way of handling parameters.
4974 compat_urllib_parse_quote_plus(iri_parts
.params
, safe
=r
"!$%&'()*+,/:;=@|~"),
4976 # Not totally sure about the `safe` argument, since the source does not explicitly mention the query URI component.
4977 compat_urllib_parse_quote_plus(iri_parts
.query
, safe
=r
"!$%&'()*+,/:;=?@{|}~"),
4979 compat_urllib_parse_quote_plus(iri_parts
.fragment
, safe
=r
"!#$%&'()*+,/:;=?@{|}~")))
4981 # Source for `safe` arguments: https://url.spec.whatwg.org/#percent-encoded-bytes.
4984 def to_high_limit_path(path
):
4985 if sys
.platform
in ['win32', 'cygwin']:
4986 # Work around MAX_PATH limitation on Windows. The maximum allowed length for the individual path segments may still be quite limited.
4987 return r
'\\?\ '.rstrip() + os
.path
.abspath(path
)
4992 def format_field(obj
, field
=None, template
='%s', ignore
=(None, ''), default
='', func
=None):
4993 val
= traverse_obj(obj
, *variadic(field
))
4996 return template
% (func(val
) if func
else val
)
4999 def clean_podcast_url(url
):
5000 return re
.sub(r
'''(?x)
5004 media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
5007 (?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
5010 cn\.co| # https://podcorn.com/analytics-prefix/
5011 st\.fm # https://podsights.com/docs/
5016 _HEX_TABLE
= '0123456789abcdef'
5019 def random_uuidv4():
5020 return re
.sub(r
'[xy]', lambda x
: _HEX_TABLE
[random
.randint(0, 15)], 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx')
5023 def make_dir(path
, to_screen
=None):
5025 dn
= os
.path
.dirname(path
)
5026 if dn
and not os
.path
.exists(dn
):
5029 except (OSError, IOError) as err
:
5030 if callable(to_screen
) is not None:
5031 to_screen('unable to create directory ' + error_to_compat_str(err
))
5035 def get_executable_path():
5036 from zipimport
import zipimporter
5037 if hasattr(sys
, 'frozen'): # Running from PyInstaller
5038 path
= os
.path
.dirname(sys
.executable
)
5039 elif isinstance(globals().get('__loader__'), zipimporter
): # Running from ZIP
5040 path
= os
.path
.join(os
.path
.dirname(__file__
), '../..')
5042 path
= os
.path
.join(os
.path
.dirname(__file__
), '..')
5043 return os
.path
.abspath(path
)
5046 def load_plugins(name
, suffix
, namespace
):
5049 plugins_spec
= importlib
.util
.spec_from_file_location(
5050 name
, os
.path
.join(get_executable_path(), 'ytdlp_plugins', name
, '__init__.py'))
5051 plugins
= importlib
.util
.module_from_spec(plugins_spec
)
5052 sys
.modules
[plugins_spec
.name
] = plugins
5053 plugins_spec
.loader
.exec_module(plugins
)
5054 for name
in dir(plugins
):
5055 if name
in namespace
:
5057 if not name
.endswith(suffix
):
5059 klass
= getattr(plugins
, name
)
5060 classes
[name
] = namespace
[name
] = klass
5061 except FileNotFoundError
:
5067 obj
, *path_list
, default
=None, expected_type
=None, get_all
=True,
5068 casesense
=True, is_user_input
=False, traverse_string
=False):
5069 ''' Traverse nested list/dict/tuple
5070 @param path_list A list of paths which are checked one by one.
5071 Each path is a list of keys where each key is a string,
5072 a function, a tuple of strings/None or "...".
5073 When a fuction is given, it takes the key as argument and
5074 returns whether the key matches or not. When a tuple is given,
5075 all the keys given in the tuple are traversed, and
5076 "..." traverses all the keys in the object
5077 "None" returns the object without traversal
5078 @param default Default value to return
5079 @param expected_type Only accept final value of this type (Can also be any callable)
5080 @param get_all Return all the values obtained from a path or only the first one
5081 @param casesense Whether to consider dictionary keys as case sensitive
5082 @param is_user_input Whether the keys are generated from user input. If True,
5083 strings are converted to int/slice if necessary
5084 @param traverse_string Whether to traverse inside strings. If True, any
5085 non-compatible object will also be converted into a string
5089 _lower
= lambda k
: (k
.lower() if isinstance(k
, str) else k
)
5090 path_list
= (map(_lower
, variadic(path
)) for path
in path_list
)
5092 def _traverse_obj(obj
, path
, _current_depth
=0):
5094 path
= tuple(variadic(path
))
5095 for i
, key
in enumerate(path
):
5096 if None in (key
, obj
):
5098 if isinstance(key
, (list, tuple)):
5099 obj
= [_traverse_obj(obj
, sub_key
, _current_depth
) for sub_key
in key
]
5102 obj
= (obj
.values() if isinstance(obj
, dict)
5103 else obj
if isinstance(obj
, (list, tuple, LazyList
))
5104 else str(obj
) if traverse_string
else [])
5106 depth
= max(depth
, _current_depth
)
5107 return [_traverse_obj(inner_obj
, path
[i
+ 1:], _current_depth
) for inner_obj
in obj
]
5109 if isinstance(obj
, (list, tuple, LazyList
)):
5110 obj
= enumerate(obj
)
5111 elif isinstance(obj
, dict):
5114 if not traverse_string
:
5118 depth
= max(depth
, _current_depth
)
5119 return [_traverse_obj(v
, path
[i
+ 1:], _current_depth
) for k
, v
in obj
if key(k
)]
5120 elif isinstance(obj
, dict) and not (is_user_input
and key
== ':'):
5121 obj
= (obj
.get(key
) if casesense
or (key
in obj
)
5122 else next((v
for k
, v
in obj
.items() if _lower(k
) == key
), None))
5125 key
= (int_or_none(key
) if ':' not in key
5126 else slice(*map(int_or_none
, key
.split(':'))))
5127 if key
== slice(None):
5128 return _traverse_obj(obj
, (..., *path
[i
+ 1:]), _current_depth
)
5129 if not isinstance(key
, (int, slice)):
5131 if not isinstance(obj
, (list, tuple, LazyList
)):
5132 if not traverse_string
:
5141 if isinstance(expected_type
, type):
5142 type_test
= lambda val
: val
if isinstance(val
, expected_type
) else None
5143 elif expected_type
is not None:
5144 type_test
= expected_type
5146 type_test
= lambda val
: val
5148 for path
in path_list
:
5150 val
= _traverse_obj(obj
, path
)
5153 for _
in range(depth
- 1):
5154 val
= itertools
.chain
.from_iterable(v
for v
in val
if v
is not None)
5155 val
= [v
for v
in map(type_test
, val
) if v
is not None]
5157 return val
if get_all
else val
[0]
5159 val
= type_test(val
)
5165 def traverse_dict(dictn
, keys
, casesense
=True):
5166 write_string('DeprecationWarning: yt_dlp.utils.traverse_dict is deprecated '
5167 'and may be removed in a future version. Use yt_dlp.utils.traverse_obj instead')
5168 return traverse_obj(dictn
, keys
, casesense
=casesense
, is_user_input
=True, traverse_string
=True)
5171 def variadic(x
, allowed_types
=(str, bytes, dict)):
5172 return x
if isinstance(x
, collections
.abc
.Iterable
) and not isinstance(x
, allowed_types
) else (x
,)
5175 # create a JSON Web Signature (jws) with HS256 algorithm
5176 # the resulting format is in JWS Compact Serialization
5177 # implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html
5178 # implemented following JWS https://www.rfc-editor.org/rfc/rfc7515.html
5179 def jwt_encode_hs256(payload_data
, key
, headers
={}):
5185 header_data
.update(headers
)
5186 header_b64
= base64
.b64encode(json
.dumps(header_data
).encode('utf-8'))
5187 payload_b64
= base64
.b64encode(json
.dumps(payload_data
).encode('utf-8'))
5188 h
= hmac
.new(key
.encode('utf-8'), header_b64
+ b
'.' + payload_b64
, hashlib
.sha256
)
5189 signature_b64
= base64
.b64encode(h
.digest())
5190 token
= header_b64
+ b
'.' + payload_b64
+ b
'.' + signature_b64
5194 # can be extended in future to verify the signature and parse header and return the algorithm used if it's not HS256
5195 def jwt_decode_hs256(jwt
):
5196 header_b64
, payload_b64
, signature_b64
= jwt
.split('.')
5197 payload_data
= json
.loads(base64
.urlsafe_b64decode(payload_b64
))
5201 def supports_terminal_sequences(stream
):
5202 if compat_os_name
== 'nt':
5203 from .compat
import WINDOWS_VT_MODE
# Must be imported locally
5204 if not WINDOWS_VT_MODE
or get_windows_version() < (10, 0, 10586):
5206 elif not os
.getenv('TERM'):
5209 return stream
.isatty()
5210 except BaseException
:
5214 _terminal_sequences_re
= re
.compile('\033\\[[^m]+m')
5217 def remove_terminal_sequences(string
):
5218 return _terminal_sequences_re
.sub('', string
)
5221 def number_of_digits(number
):
5222 return len('%d' % number
)
5225 def join_nonempty(*values
, delim
='-', from_dict
=None):
5226 if from_dict
is not None:
5227 values
= map(from_dict
.get
, values
)
5228 return delim
.join(map(str, filter(None, values
)))
5234 __initialized
= False
5236 def __init__(self
, parser
, label
=None):
5237 self
._parser
, self
.label
= parser
, label
5238 self
._loaded
_paths
, self
.configs
= set(), []
5240 def init(self
, args
=None, filename
=None):
5241 assert not self
.__initialized
5244 location
= os
.path
.realpath(filename
)
5245 directory
= os
.path
.dirname(location
)
5246 if location
in self
._loaded
_paths
:
5248 self
._loaded
_paths
.add(location
)
5250 self
.__initialized
= True
5251 self
.own_args
, self
.filename
= args
, filename
5252 for location
in self
._parser
.parse_args(args
)[0].config_locations
or []:
5253 location
= os
.path
.join(directory
, expand_path(location
))
5254 if os
.path
.isdir(location
):
5255 location
= os
.path
.join(location
, 'yt-dlp.conf')
5256 if not os
.path
.exists(location
):
5257 self
._parser
.error(f
'config location {location} does not exist')
5258 self
.append_config(self
.read_file(location
), location
)
5262 label
= join_nonempty(
5263 self
.label
, 'config', f
'"{self.filename}"' if self
.filename
else '',
5265 return join_nonempty(
5266 self
.own_args
is not None and f
'{label[0].upper()}{label[1:]}: {self.hide_login_info(self.own_args)}',
5267 *(f
'\n{c}'.replace('\n', '\n| ')[1:] for c
in self
.configs
),
5271 def read_file(filename
, default
=[]):
5273 optionf
= open(filename
)
5275 return default
# silently skip if file is not present
5277 # FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
5278 contents
= optionf
.read()
5279 if sys
.version_info
< (3,):
5280 contents
= contents
.decode(preferredencoding())
5281 res
= compat_shlex_split(contents
, comments
=True)
5287 def hide_login_info(opts
):
5288 PRIVATE_OPTS
= set(['-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'])
5289 eqre
= re
.compile('^(?P<key>' + ('|'.join(re
.escape(po
) for po
in PRIVATE_OPTS
)) + ')=.+$')
5294 return m
.group('key') + '=PRIVATE'
5298 opts
= list(map(_scrub_eq
, opts
))
5299 for idx
, opt
in enumerate(opts
):
5300 if opt
in PRIVATE_OPTS
and idx
+ 1 < len(opts
):
5301 opts
[idx
+ 1] = 'PRIVATE'
5304 def append_config(self
, *args
, label
=None):
5305 config
= type(self
)(self
._parser
, label
)
5306 config
._loaded
_paths
= self
._loaded
_paths
5307 if config
.init(*args
):
5308 self
.configs
.append(config
)
5312 for config
in reversed(self
.configs
):
5313 yield from config
.all_args
5314 yield from self
.own_args
or []
5316 def parse_args(self
):
5317 return self
._parser
.parse_args(list(self
.all_args
))
5320 class WebSocketsWrapper():
5321 """Wraps websockets module to use in non-async scopes"""
5323 def __init__(self
, url
, headers
=None):
5324 self
.loop
= asyncio
.events
.new_event_loop()
5325 self
.conn
= compat_websockets
.connect(
5326 url
, extra_headers
=headers
, ping_interval
=None,
5327 close_timeout
=float('inf'), loop
=self
.loop
, ping_timeout
=float('inf'))
5329 def __enter__(self
):
5330 self
.pool
= self
.run_with_loop(self
.conn
.__aenter
__(), self
.loop
)
5333 def send(self
, *args
):
5334 self
.run_with_loop(self
.pool
.send(*args
), self
.loop
)
5336 def recv(self
, *args
):
5337 return self
.run_with_loop(self
.pool
.recv(*args
), self
.loop
)
5339 def __exit__(self
, type, value
, traceback
):
5341 return self
.run_with_loop(self
.conn
.__aexit
__(type, value
, traceback
), self
.loop
)
5344 self
.r_cancel_all_tasks(self
.loop
)
5346 # taken from https://github.com/python/cpython/blob/3.9/Lib/asyncio/runners.py with modifications
5347 # for contributors: If there's any new library using asyncio needs to be run in non-async, move these function out of this class
5349 def run_with_loop(main
, loop
):
5350 if not asyncio
.coroutines
.iscoroutine(main
):
5351 raise ValueError(f
'a coroutine was expected, got {main!r}')
5354 return loop
.run_until_complete(main
)
5356 loop
.run_until_complete(loop
.shutdown_asyncgens())
5357 if hasattr(loop
, 'shutdown_default_executor'):
5358 loop
.run_until_complete(loop
.shutdown_default_executor())
5361 def _cancel_all_tasks(loop
):
5362 to_cancel
= asyncio
.tasks
.all_tasks(loop
)
5367 for task
in to_cancel
:
5370 loop
.run_until_complete(
5371 asyncio
.tasks
.gather(*to_cancel
, loop
=loop
, return_exceptions
=True))
5373 for task
in to_cancel
:
5374 if task
.cancelled():
5376 if task
.exception() is not None:
5377 loop
.call_exception_handler({
5378 'message': 'unhandled exception during asyncio.run() shutdown',
5379 'exception': task
.exception(),
5384 has_websockets
= bool(compat_websockets
)