2 # -*- coding: utf-8 -*-
4 from __future__ import unicode_literals
33 import xml.etree.ElementTree
40 compat_etree_fromstring,
42 compat_html_entities_html5,
47 compat_socket_create_connection,
53 compat_urllib_parse_urlencode,
54 compat_urllib_parse_urlparse,
55 compat_urllib_parse_unquote_plus,
56 compat_urllib_request,
67 def register_socks_protocols():
68 # "Register" SOCKS protocols
69 # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
70 # URLs with protocols not in urlparse.uses_netloc are not handled correctly
71 for scheme in ('socks', 'socks4', 'socks4a', 'socks5'):
72 if scheme not in compat_urlparse.uses_netloc:
73 compat_urlparse.uses_netloc.append(scheme)
76 # This is not clearly defined otherwise
77 compiled_regex_type = type(re.compile(''))
80 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/47.0 (Chrome)',
81 'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
82 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
83 'Accept-Encoding': 'gzip, deflate',
84 'Accept-Language': 'en-us,en;q=0.5',
90 ENGLISH_MONTH_NAMES = [
91 'January', 'February', 'March', 'April', 'May', 'June',
92 'July', 'August', 'September', 'October', 'November', 'December']
95 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
96 'flv', 'f4v', 'f4a', 'f4b',
97 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
107 'f4f', 'f4m', 'm3u8', 'smil')
109 # needed for sanitizing filenames in restricted mode
110 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
111 itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUYP', ['ss'],
112 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuypy')))
128 '%Y-%m-%d %H:%M:%S.%f',
131 '%Y-%m-%dT%H:%M:%SZ',
132 '%Y-%m-%dT%H:%M:%S.%fZ',
133 '%Y-%m-%dT%H:%M:%S.%f0Z',
135 '%Y-%m-%dT%H:%M:%S.%f',
139 DATE_FORMATS_DAY_FIRST = list(DATE_FORMATS)
140 DATE_FORMATS_DAY_FIRST.extend([
149 DATE_FORMATS_MONTH_FIRST = list(DATE_FORMATS)
150 DATE_FORMATS_MONTH_FIRST.extend([
159 def preferredencoding():
160 """Get preferred encoding.
162 Returns the best encoding scheme for the system, based on
163 locale.getpreferredencoding() and some further tweaks.
166 pref = locale.getpreferredencoding()
174 def write_json_file(obj, fn):
175 """ Encode obj as JSON and write it to fn, atomically if possible """
177 fn = encodeFilename(fn)
178 if sys.version_info < (3, 0) and sys.platform != 'win32':
179 encoding = get_filesystem_encoding()
180 # os.path.basename returns a bytes object, but NamedTemporaryFile
181 # will fail if the filename contains non ascii characters unless we
182 # use a unicode object
183 path_basename = lambda f: os.path.basename(fn).decode(encoding)
184 # the same for os.path.dirname
185 path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
187 path_basename = os.path.basename
188 path_dirname = os.path.dirname
192 'prefix': path_basename(fn) + '.',
193 'dir': path_dirname(fn),
197 # In Python 2.x, json.dump expects a bytestream.
198 # In Python 3.x, it writes to a character stream
199 if sys.version_info < (3, 0):
207 tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
212 if sys.platform == 'win32':
213 # Need to remove existing file on Windows, else os.rename raises
214 # WindowsError or FileExistsError.
219 os.rename(tf.name, fn)
228 if sys.version_info >= (2, 7):
229 def find_xpath_attr(node, xpath, key, val=None):
230 """ Find the xpath xpath[@key=val] """
231 assert re.match(r'^[a-zA-Z_-]+$', key)
232 expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
233 return node.find(expr)
235 def find_xpath_attr(node, xpath, key, val=None):
236 for f in node.findall(compat_xpath(xpath)):
237 if key not in f.attrib:
239 if val is None or f.attrib.get(key) == val:
243 # On python2.6 the xml.etree.ElementTree.Element methods don't support
244 # the namespace parameter
247 def xpath_with_ns(path, ns_map):
248 components = [c.split(':') for c in path.split('/')]
252 replaced.append(c[0])
255 replaced.append('{%s}%s' % (ns_map[ns], tag))
256 return '/'.join(replaced)
259 def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
260 def _find_xpath(xpath):
261 return node.find(compat_xpath(xpath))
263 if isinstance(xpath, (str, compat_str)):
264 n = _find_xpath(xpath)
272 if default is not NO_DEFAULT:
275 name = xpath if name is None else name
276 raise ExtractorError('Could not find XML element %s' % name)
282 def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
283 n = xpath_element(node, xpath, name, fatal=fatal, default=default)
284 if n is None or n == default:
287 if default is not NO_DEFAULT:
290 name = xpath if name is None else name
291 raise ExtractorError('Could not find XML element\'s text %s' % name)
297 def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
298 n = find_xpath_attr(node, xpath, key)
300 if default is not NO_DEFAULT:
303 name = '%s[@%s]' % (xpath, key) if name is None else name
304 raise ExtractorError('Could not find XML attribute %s' % name)
310 def get_element_by_id(id, html):
311 """Return the content of the tag with the specified ID in the passed HTML document"""
312 return get_element_by_attribute('id', id, html)
315 def get_element_by_class(class_name, html):
316 return get_element_by_attribute(
317 'class', r'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
318 html, escape_value=False)
321 def get_element_by_attribute(attribute, value, html, escape_value=True):
322 """Return the content of the tag with the specified attribute in the passed HTML document"""
324 value = re.escape(value) if escape_value else value
326 m = re.search(r'''(?xs)
328 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
330 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
334 ''' % (re.escape(attribute), value), html)
338 res = m.group('content')
340 if res.startswith('"') or res.startswith("'"):
343 return unescapeHTML(res)
346 class HTMLAttributeParser(compat_HTMLParser):
347 """Trivial HTML parser to gather the attributes for a single element"""
350 compat_HTMLParser.__init__(self)
352 def handle_starttag(self, tag, attrs):
353 self.attrs = dict(attrs)
356 def extract_attributes(html_element):
357 """Given a string for an HTML element such as
359 a="foo" B="bar" c="&98;az" d=boz
360 empty= noval entity="&"
363 Decode and return a dictionary of attributes.
365 'a': 'foo', 'b': 'bar', c: 'baz', d: 'boz',
366 'empty': '', 'noval': None, 'entity': '&',
367 'sq': '"', 'dq': '\''
369 NB HTMLParser is stricter in Python 2.6 & 3.2 than in later versions,
370 but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
372 parser = HTMLAttributeParser()
373 parser.feed(html_element)
378 def clean_html(html):
379 """Clean an HTML snippet into a readable string"""
381 if html is None: # Convenience for sanitizing descriptions etc.
385 html = html.replace('\n', ' ')
386 html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
387 html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
389 html = re.sub('<.*?>', '', html)
390 # Replace html entities
391 html = unescapeHTML(html)
395 def sanitize_open(filename, open_mode):
396 """Try to open the given filename, and slightly tweak it if this fails.
398 Attempts to open the given filename. If this fails, it tries to change
399 the filename slightly, step by step, until it's either able to open it
400 or it fails and raises a final exception, like the standard open()
403 It returns the tuple (stream, definitive_file_name).
407 if sys.platform == 'win32':
409 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
410 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
411 stream = open(encodeFilename(filename), open_mode)
412 return (stream, filename)
413 except (IOError, OSError) as err:
414 if err.errno in (errno.EACCES,):
417 # In case of error, try to remove win32 forbidden chars
418 alt_filename = sanitize_path(filename)
419 if alt_filename == filename:
422 # An exception here should be caught in the caller
423 stream = open(encodeFilename(alt_filename), open_mode)
424 return (stream, alt_filename)
427 def timeconvert(timestr):
428 """Convert RFC 2822 defined time string into system timestamp"""
430 timetuple = email.utils.parsedate_tz(timestr)
431 if timetuple is not None:
432 timestamp = email.utils.mktime_tz(timetuple)
436 def sanitize_filename(s, restricted=False, is_id=False):
437 """Sanitizes a string so it could be used as part of a filename.
438 If restricted is set, use a stricter subset of allowed characters.
439 Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
441 def replace_insane(char):
442 if restricted and char in ACCENT_CHARS:
443 return ACCENT_CHARS[char]
444 if char == '?' or ord(char) < 32 or ord(char) == 127:
447 return '' if restricted else '\''
449 return '_-' if restricted else ' -'
450 elif char in '\\/|*<>':
452 if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
454 if restricted and ord(char) > 127:
459 s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
460 result = ''.join(map(replace_insane, s))
462 while '__' in result:
463 result = result.replace('__', '_')
464 result = result.strip('_')
465 # Common case of "Foreign band name - English song title"
466 if restricted and result.startswith('-_'):
468 if result.startswith('-'):
469 result = '_' + result[len('-'):]
470 result = result.lstrip('.')
476 def sanitize_path(s):
477 """Sanitizes and normalizes path on Windows"""
478 if sys.platform != 'win32':
480 drive_or_unc, _ = os.path.splitdrive(s)
481 if sys.version_info < (2, 7) and not drive_or_unc:
482 drive_or_unc, _ = os.path.splitunc(s)
483 norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
487 path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|[\s.]$)', '#', path_part)
488 for path_part in norm_path]
490 sanitized_path.insert(0, drive_or_unc + os.path.sep)
491 return os.path.join(*sanitized_path)
494 # Prepend protocol-less URLs with `http:` scheme in order to mitigate the number of
495 # unwanted failures due to missing protocol
496 def sanitize_url(url):
497 return 'http:%s' % url if url.startswith('//') else url
500 def sanitized_Request(url, *args, **kwargs):
501 return compat_urllib_request.Request(sanitize_url(url), *args, **kwargs)
504 def orderedSet(iterable):
505 """ Remove all duplicates from the input iterable """
513 def _htmlentity_transform(entity_with_semicolon):
514 """Transforms an HTML entity to a character."""
515 entity = entity_with_semicolon[:-1]
517 # Known non-numeric HTML entity
518 if entity in compat_html_entities.name2codepoint:
519 return compat_chr(compat_html_entities.name2codepoint[entity])
521 # TODO: HTML5 allows entities without a semicolon. For example,
522 # 'Éric' should be decoded as 'Éric'.
523 if entity_with_semicolon in compat_html_entities_html5:
524 return compat_html_entities_html5[entity_with_semicolon]
526 mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
528 numstr = mobj.group(1)
529 if numstr.startswith('x'):
531 numstr = '0%s' % numstr
534 # See https://github.com/rg3/youtube-dl/issues/7518
536 return compat_chr(int(numstr, base))
540 # Unknown entity in name, return its literal representation
541 return '&%s;' % entity
547 assert type(s) == compat_str
550 r'&([^;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
553 def get_subprocess_encoding():
554 if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
555 # For subprocess calls, encode with locale encoding
556 # Refer to http://stackoverflow.com/a/9951851/35070
557 encoding = preferredencoding()
559 encoding = sys.getfilesystemencoding()
565 def encodeFilename(s, for_subprocess=False):
567 @param s The name of the file
570 assert type(s) == compat_str
572 # Python 3 has a Unicode API
573 if sys.version_info >= (3, 0):
576 # Pass '' directly to use Unicode APIs on Windows 2000 and up
577 # (Detecting Windows NT 4 is tricky because 'major >= 4' would
578 # match Windows 9x series as well. Besides, NT 4 is obsolete.)
579 if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
582 # Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
583 if sys.platform.startswith('java'):
586 return s.encode(get_subprocess_encoding(), 'ignore')
589 def decodeFilename(b, for_subprocess=False):
591 if sys.version_info >= (3, 0):
594 if not isinstance(b, bytes):
597 return b.decode(get_subprocess_encoding(), 'ignore')
600 def encodeArgument(s):
601 if not isinstance(s, compat_str):
602 # Legacy code that uses byte strings
603 # Uncomment the following line after fixing all post processors
604 # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
605 s = s.decode('ascii')
606 return encodeFilename(s, True)
609 def decodeArgument(b):
610 return decodeFilename(b, True)
613 def decodeOption(optval):
616 if isinstance(optval, bytes):
617 optval = optval.decode(preferredencoding())
619 assert isinstance(optval, compat_str)
623 def formatSeconds(secs):
625 return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
627 return '%d:%02d' % (secs // 60, secs % 60)
632 def make_HTTPS_handler(params, **kwargs):
633 opts_no_check_certificate = params.get('nocheckcertificate', False)
634 if hasattr(ssl, 'create_default_context'): # Python >= 3.4 or 2.7.9
635 context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
636 if opts_no_check_certificate:
637 context.check_hostname = False
638 context.verify_mode = ssl.CERT_NONE
640 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
643 # (create_default_context present but HTTPSHandler has no context=)
646 if sys.version_info < (3, 2):
647 return YoutubeDLHTTPSHandler(params, **kwargs)
649 context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
650 context.verify_mode = (ssl.CERT_NONE
651 if opts_no_check_certificate
652 else ssl.CERT_REQUIRED)
653 context.set_default_verify_paths()
654 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
657 def bug_reports_message():
658 if ytdl_is_updateable():
659 update_cmd = 'type youtube-dl -U to update'
661 update_cmd = 'see https://yt-dl.org/update on how to update'
662 msg = '; please report this issue on https://yt-dl.org/bug .'
663 msg += ' Make sure you are using the latest version; %s.' % update_cmd
664 msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
668 class ExtractorError(Exception):
669 """Error during info extraction."""
671 def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
672 """ tb, if given, is the original traceback (so that it can be printed out).
673 If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
676 if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
678 if video_id is not None:
679 msg = video_id + ': ' + msg
681 msg += ' (caused by %r)' % cause
683 msg += bug_reports_message()
684 super(ExtractorError, self).__init__(msg)
687 self.exc_info = sys.exc_info() # preserve original exception
689 self.video_id = video_id
691 def format_traceback(self):
692 if self.traceback is None:
694 return ''.join(traceback.format_tb(self.traceback))
697 class UnsupportedError(ExtractorError):
698 def __init__(self, url):
699 super(UnsupportedError, self).__init__(
700 'Unsupported URL: %s' % url, expected=True)
704 class RegexNotFoundError(ExtractorError):
705 """Error when a regex didn't match"""
709 class DownloadError(Exception):
710 """Download Error exception.
712 This exception may be thrown by FileDownloader objects if they are not
713 configured to continue on errors. They will contain the appropriate
717 def __init__(self, msg, exc_info=None):
718 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
719 super(DownloadError, self).__init__(msg)
720 self.exc_info = exc_info
723 class SameFileError(Exception):
724 """Same File exception.
726 This exception will be thrown by FileDownloader objects if they detect
727 multiple files would have to be downloaded to the same file on disk.
732 class PostProcessingError(Exception):
733 """Post Processing exception.
735 This exception may be raised by PostProcessor's .run() method to
736 indicate an error in the postprocessing task.
739 def __init__(self, msg):
743 class MaxDownloadsReached(Exception):
744 """ --max-downloads limit has been reached. """
748 class UnavailableVideoError(Exception):
749 """Unavailable Format exception.
751 This exception will be thrown when a video is requested
752 in a format that is not available for that video.
757 class ContentTooShortError(Exception):
758 """Content Too Short exception.
760 This exception may be raised by FileDownloader objects when a file they
761 download is too small for what the server announced first, indicating
762 the connection was probably interrupted.
765 def __init__(self, downloaded, expected):
767 self.downloaded = downloaded
768 self.expected = expected
771 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
772 # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
773 # expected HTTP responses to meet HTTP/1.0 or later (see also
774 # https://github.com/rg3/youtube-dl/issues/6727)
775 if sys.version_info < (3, 0):
776 kwargs[b'strict'] = True
777 hc = http_class(*args, **kwargs)
778 source_address = ydl_handler._params.get('source_address')
779 if source_address is not None:
780 sa = (source_address, 0)
781 if hasattr(hc, 'source_address'): # Python 2.7+
782 hc.source_address = sa
784 def _hc_connect(self, *args, **kwargs):
785 sock = compat_socket_create_connection(
786 (self.host, self.port), self.timeout, sa)
788 self.sock = ssl.wrap_socket(
789 sock, self.key_file, self.cert_file,
790 ssl_version=ssl.PROTOCOL_TLSv1)
793 hc.connect = functools.partial(_hc_connect, hc)
798 def handle_youtubedl_headers(headers):
799 filtered_headers = headers
801 if 'Youtubedl-no-compression' in filtered_headers:
802 filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding')
803 del filtered_headers['Youtubedl-no-compression']
805 return filtered_headers
808 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
809 """Handler for HTTP requests and responses.
811 This class, when installed with an OpenerDirector, automatically adds
812 the standard headers to every HTTP request and handles gzipped and
813 deflated responses from web servers. If compression is to be avoided in
814 a particular request, the original request in the program code only has
815 to include the HTTP header "Youtubedl-no-compression", which will be
816 removed before making the real request.
818 Part of this code was copied from:
820 http://techknack.net/python-urllib2-handlers/
822 Andrew Rowls, the author of that code, agreed to release it to the
826 def __init__(self, params, *args, **kwargs):
827 compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
828 self._params = params
830 def http_open(self, req):
831 conn_class = compat_http_client.HTTPConnection
833 socks_proxy = req.headers.get('Ytdl-socks-proxy')
835 conn_class = make_socks_conn_class(conn_class, socks_proxy)
836 del req.headers['Ytdl-socks-proxy']
838 return self.do_open(functools.partial(
839 _create_http_connection, self, conn_class, False),
845 return zlib.decompress(data, -zlib.MAX_WBITS)
847 return zlib.decompress(data)
850 def addinfourl_wrapper(stream, headers, url, code):
851 if hasattr(compat_urllib_request.addinfourl, 'getcode'):
852 return compat_urllib_request.addinfourl(stream, headers, url, code)
853 ret = compat_urllib_request.addinfourl(stream, headers, url)
857 def http_request(self, req):
858 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
859 # always respected by websites, some tend to give out URLs with non percent-encoded
860 # non-ASCII characters (see telemb.py, ard.py [#3412])
861 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
862 # To work around aforementioned issue we will replace request's original URL with
863 # percent-encoded one
864 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
865 # the code of this workaround has been moved here from YoutubeDL.urlopen()
866 url = req.get_full_url()
867 url_escaped = escape_url(url)
869 # Substitute URL if any change after escaping
870 if url != url_escaped:
871 req = update_Request(req, url=url_escaped)
873 for h, v in std_headers.items():
874 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
875 # The dict keys are capitalized because of this bug by urllib
876 if h.capitalize() not in req.headers:
879 req.headers = handle_youtubedl_headers(req.headers)
881 if sys.version_info < (2, 7) and '#' in req.get_full_url():
882 # Python 2.6 is brain-dead when it comes to fragments
883 req._Request__original = req._Request__original.partition('#')[0]
884 req._Request__r_type = req._Request__r_type.partition('#')[0]
888 def http_response(self, req, resp):
891 if resp.headers.get('Content-encoding', '') == 'gzip':
892 content = resp.read()
893 gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
895 uncompressed = io.BytesIO(gz.read())
896 except IOError as original_ioerror:
897 # There may be junk add the end of the file
898 # See http://stackoverflow.com/q/4928560/35070 for details
899 for i in range(1, 1024):
901 gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
902 uncompressed = io.BytesIO(gz.read())
907 raise original_ioerror
908 resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
909 resp.msg = old_resp.msg
910 del resp.headers['Content-encoding']
912 if resp.headers.get('Content-encoding', '') == 'deflate':
913 gz = io.BytesIO(self.deflate(resp.read()))
914 resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
915 resp.msg = old_resp.msg
916 del resp.headers['Content-encoding']
917 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
918 # https://github.com/rg3/youtube-dl/issues/6457).
919 if 300 <= resp.code < 400:
920 location = resp.headers.get('Location')
922 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
923 if sys.version_info >= (3, 0):
924 location = location.encode('iso-8859-1').decode('utf-8')
926 location = location.decode('utf-8')
927 location_escaped = escape_url(location)
928 if location != location_escaped:
929 del resp.headers['Location']
930 if sys.version_info < (3, 0):
931 location_escaped = location_escaped.encode('utf-8')
932 resp.headers['Location'] = location_escaped
935 https_request = http_request
936 https_response = http_response
939 def make_socks_conn_class(base_class, socks_proxy):
940 assert issubclass(base_class, (
941 compat_http_client.HTTPConnection, compat_http_client.HTTPSConnection))
943 url_components = compat_urlparse.urlparse(socks_proxy)
944 if url_components.scheme.lower() == 'socks5':
945 socks_type = ProxyType.SOCKS5
946 elif url_components.scheme.lower() in ('socks', 'socks4'):
947 socks_type = ProxyType.SOCKS4
948 elif url_components.scheme.lower() == 'socks4a':
949 socks_type = ProxyType.SOCKS4A
951 def unquote_if_non_empty(s):
954 return compat_urllib_parse_unquote_plus(s)
958 url_components.hostname, url_components.port or 1080,
960 unquote_if_non_empty(url_components.username),
961 unquote_if_non_empty(url_components.password),
964 class SocksConnection(base_class):
966 self.sock = sockssocket()
967 self.sock.setproxy(*proxy_args)
968 if type(self.timeout) in (int, float):
969 self.sock.settimeout(self.timeout)
970 self.sock.connect((self.host, self.port))
972 if isinstance(self, compat_http_client.HTTPSConnection):
973 if hasattr(self, '_context'): # Python > 2.6
974 self.sock = self._context.wrap_socket(
975 self.sock, server_hostname=self.host)
977 self.sock = ssl.wrap_socket(self.sock)
979 return SocksConnection
982 class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
983 def __init__(self, params, https_conn_class=None, *args, **kwargs):
984 compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
985 self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
986 self._params = params
988 def https_open(self, req):
990 conn_class = self._https_conn_class
992 if hasattr(self, '_context'): # python > 2.6
993 kwargs['context'] = self._context
994 if hasattr(self, '_check_hostname'): # python 3.x
995 kwargs['check_hostname'] = self._check_hostname
997 socks_proxy = req.headers.get('Ytdl-socks-proxy')
999 conn_class = make_socks_conn_class(conn_class, socks_proxy)
1000 del req.headers['Ytdl-socks-proxy']
1002 return self.do_open(functools.partial(
1003 _create_http_connection, self, conn_class, True),
1007 class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
1008 def __init__(self, cookiejar=None):
1009 compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
1011 def http_response(self, request, response):
1012 # Python 2 will choke on next HTTP request in row if there are non-ASCII
1013 # characters in Set-Cookie HTTP header of last response (see
1014 # https://github.com/rg3/youtube-dl/issues/6769).
1015 # In order to at least prevent crashing we will percent encode Set-Cookie
1016 # header before HTTPCookieProcessor starts processing it.
1017 # if sys.version_info < (3, 0) and response.headers:
1018 # for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
1019 # set_cookie = response.headers.get(set_cookie_header)
1021 # set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
1022 # if set_cookie != set_cookie_escaped:
1023 # del response.headers[set_cookie_header]
1024 # response.headers[set_cookie_header] = set_cookie_escaped
1025 return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
1027 https_request = compat_urllib_request.HTTPCookieProcessor.http_request
1028 https_response = http_response
1031 def extract_timezone(date_str):
1033 r'^.{8,}?(?P<tz>Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
1036 timezone = datetime.timedelta()
1038 date_str = date_str[:-len(m.group('tz'))]
1039 if not m.group('sign'):
1040 timezone = datetime.timedelta()
1042 sign = 1 if m.group('sign') == '+' else -1
1043 timezone = datetime.timedelta(
1044 hours=sign * int(m.group('hours')),
1045 minutes=sign * int(m.group('minutes')))
1046 return timezone, date_str
1049 def parse_iso8601(date_str, delimiter='T', timezone=None):
1050 """ Return a UNIX timestamp from the given date """
1052 if date_str is None:
1055 date_str = re.sub(r'\.[0-9]+', '', date_str)
1057 if timezone is None:
1058 timezone, date_str = extract_timezone(date_str)
1061 date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
1062 dt = datetime.datetime.strptime(date_str, date_format) - timezone
1063 return calendar.timegm(dt.timetuple())
1068 def date_formats(day_first=True):
1069 return DATE_FORMATS_DAY_FIRST if day_first else DATE_FORMATS_MONTH_FIRST
1072 def unified_strdate(date_str, day_first=True):
1073 """Return a string with the date in the format YYYYMMDD"""
1075 if date_str is None:
1079 date_str = date_str.replace(',', ' ')
1080 # Remove AM/PM + timezone
1081 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
1082 _, date_str = extract_timezone(date_str)
1084 for expression in date_formats(day_first):
1086 upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
1089 if upload_date is None:
1090 timetuple = email.utils.parsedate_tz(date_str)
1093 upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
1096 if upload_date is not None:
1097 return compat_str(upload_date)
1100 def unified_timestamp(date_str, day_first=True):
1101 if date_str is None:
1104 date_str = date_str.replace(',', ' ')
1106 pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0
1107 timezone, date_str = extract_timezone(date_str)
1109 # Remove AM/PM + timezone
1110 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
1112 for expression in date_formats(day_first):
1114 dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
1115 return calendar.timegm(dt.timetuple())
1118 timetuple = email.utils.parsedate_tz(date_str)
1120 return calendar.timegm(timetuple) + pm_delta * 3600
1123 def determine_ext(url, default_ext='unknown_video'):
1126 guess = url.partition('?')[0].rpartition('.')[2]
1127 if re.match(r'^[A-Za-z0-9]+$', guess):
1129 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
1130 elif guess.rstrip('/') in KNOWN_EXTENSIONS:
1131 return guess.rstrip('/')
1136 def subtitles_filename(filename, sub_lang, sub_format):
1137 return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
1140 def date_from_str(date_str):
1142 Return a datetime object from a string in the format YYYYMMDD or
1143 (now|today)[+-][0-9](day|week|month|year)(s)?"""
1144 today = datetime.date.today()
1145 if date_str in ('now', 'today'):
1147 if date_str == 'yesterday':
1148 return today - datetime.timedelta(days=1)
1149 match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
1150 if match is not None:
1151 sign = match.group('sign')
1152 time = int(match.group('time'))
1155 unit = match.group('unit')
1156 # A bad approximation?
1160 elif unit == 'year':
1164 delta = datetime.timedelta(**{unit: time})
1165 return today + delta
1166 return datetime.datetime.strptime(date_str, '%Y%m%d').date()
1169 def hyphenate_date(date_str):
1171 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1172 match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
1173 if match is not None:
1174 return '-'.join(match.groups())
1179 class DateRange(object):
1180 """Represents a time interval between two dates"""
1182 def __init__(self, start=None, end=None):
1183 """start and end must be strings in the format accepted by date"""
1184 if start is not None:
1185 self.start = date_from_str(start)
1187 self.start = datetime.datetime.min.date()
1189 self.end = date_from_str(end)
1191 self.end = datetime.datetime.max.date()
1192 if self.start > self.end:
1193 raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
1197 """Returns a range that only contains the given day"""
1198 return cls(day, day)
1200 def __contains__(self, date):
1201 """Check if the date is in the range"""
1202 if not isinstance(date, datetime.date):
1203 date = date_from_str(date)
1204 return self.start <= date <= self.end
1207 return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
1210 def platform_name():
1211 """ Returns the platform name as a compat_str """
1212 res = platform.platform()
1213 if isinstance(res, bytes):
1214 res = res.decode(preferredencoding())
1216 assert isinstance(res, compat_str)
1220 def _windows_write_string(s, out):
1221 """ Returns True if the string was written using special methods,
1222 False if it has yet to be written out."""
1223 # Adapted from http://stackoverflow.com/a/3259271/35070
1226 import ctypes.wintypes
1234 fileno = out.fileno()
1235 except AttributeError:
1236 # If the output stream doesn't have a fileno, it's virtual
1238 except io.UnsupportedOperation:
1239 # Some strange Windows pseudo files?
1241 if fileno not in WIN_OUTPUT_IDS:
1244 GetStdHandle = ctypes.WINFUNCTYPE(
1245 ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
1246 (b'GetStdHandle', ctypes.windll.kernel32))
1247 h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
1249 WriteConsoleW = ctypes.WINFUNCTYPE(
1250 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
1251 ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
1252 ctypes.wintypes.LPVOID)((b'WriteConsoleW', ctypes.windll.kernel32))
1253 written = ctypes.wintypes.DWORD(0)
1255 GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b'GetFileType', ctypes.windll.kernel32))
1256 FILE_TYPE_CHAR = 0x0002
1257 FILE_TYPE_REMOTE = 0x8000
1258 GetConsoleMode = ctypes.WINFUNCTYPE(
1259 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
1260 ctypes.POINTER(ctypes.wintypes.DWORD))(
1261 (b'GetConsoleMode', ctypes.windll.kernel32))
1262 INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
1264 def not_a_console(handle):
1265 if handle == INVALID_HANDLE_VALUE or handle is None:
1267 return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
1268 GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
1270 if not_a_console(h):
1273 def next_nonbmp_pos(s):
1275 return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
1276 except StopIteration:
1280 count = min(next_nonbmp_pos(s), 1024)
1282 ret = WriteConsoleW(
1283 h, s, count if count else 2, ctypes.byref(written), None)
1285 raise OSError('Failed to write string')
1286 if not count: # We just wrote a non-BMP character
1287 assert written.value == 2
1290 assert written.value > 0
1291 s = s[written.value:]
1295 def write_string(s, out=None, encoding=None):
1298 assert type(s) == compat_str
1300 if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
1301 if _windows_write_string(s, out):
1304 if ('b' in getattr(out, 'mode', '') or
1305 sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr
1306 byt = s.encode(encoding or preferredencoding(), 'ignore')
1308 elif hasattr(out, 'buffer'):
1309 enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
1310 byt = s.encode(enc, 'ignore')
1311 out.buffer.write(byt)
1317 def bytes_to_intlist(bs):
1320 if isinstance(bs[0], int): # Python 3
1323 return [ord(c) for c in bs]
1326 def intlist_to_bytes(xs):
1329 return compat_struct_pack('%dB' % len(xs), *xs)
1332 # Cross-platform file locking
1333 if sys.platform == 'win32':
1334 import ctypes.wintypes
1337 class OVERLAPPED(ctypes.Structure):
1339 ('Internal', ctypes.wintypes.LPVOID),
1340 ('InternalHigh', ctypes.wintypes.LPVOID),
1341 ('Offset', ctypes.wintypes.DWORD),
1342 ('OffsetHigh', ctypes.wintypes.DWORD),
1343 ('hEvent', ctypes.wintypes.HANDLE),
1346 kernel32 = ctypes.windll.kernel32
1347 LockFileEx = kernel32.LockFileEx
1348 LockFileEx.argtypes = [
1349 ctypes.wintypes.HANDLE, # hFile
1350 ctypes.wintypes.DWORD, # dwFlags
1351 ctypes.wintypes.DWORD, # dwReserved
1352 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1353 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1354 ctypes.POINTER(OVERLAPPED) # Overlapped
1356 LockFileEx.restype = ctypes.wintypes.BOOL
1357 UnlockFileEx = kernel32.UnlockFileEx
1358 UnlockFileEx.argtypes = [
1359 ctypes.wintypes.HANDLE, # hFile
1360 ctypes.wintypes.DWORD, # dwReserved
1361 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1362 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1363 ctypes.POINTER(OVERLAPPED) # Overlapped
1365 UnlockFileEx.restype = ctypes.wintypes.BOOL
1366 whole_low = 0xffffffff
1367 whole_high = 0x7fffffff
1369 def _lock_file(f, exclusive):
1370 overlapped = OVERLAPPED()
1371 overlapped.Offset = 0
1372 overlapped.OffsetHigh = 0
1373 overlapped.hEvent = 0
1374 f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1375 handle = msvcrt.get_osfhandle(f.fileno())
1376 if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1377 whole_low, whole_high, f._lock_file_overlapped_p):
1378 raise OSError('Locking file failed: %r' % ctypes.FormatError())
1380 def _unlock_file(f):
1381 assert f._lock_file_overlapped_p
1382 handle = msvcrt.get_osfhandle(f.fileno())
1383 if not UnlockFileEx(handle, 0,
1384 whole_low, whole_high, f._lock_file_overlapped_p):
1385 raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1388 # Some platforms, such as Jython, is missing fcntl
1392 def _lock_file(f, exclusive):
1393 fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1395 def _unlock_file(f):
1396 fcntl.flock(f, fcntl.LOCK_UN)
1398 UNSUPPORTED_MSG = 'file locking is not supported on this platform'
1400 def _lock_file(f, exclusive):
1401 raise IOError(UNSUPPORTED_MSG)
1403 def _unlock_file(f):
1404 raise IOError(UNSUPPORTED_MSG)
1407 class locked_file(object):
1408 def __init__(self, filename, mode, encoding=None):
1409 assert mode in ['r', 'a', 'w']
1410 self.f = io.open(filename, mode, encoding=encoding)
1413 def __enter__(self):
1414 exclusive = self.mode != 'r'
1416 _lock_file(self.f, exclusive)
1422 def __exit__(self, etype, value, traceback):
1424 _unlock_file(self.f)
1431 def write(self, *args):
1432 return self.f.write(*args)
1434 def read(self, *args):
1435 return self.f.read(*args)
1438 def get_filesystem_encoding():
1439 encoding = sys.getfilesystemencoding()
1440 return encoding if encoding is not None else 'utf-8'
1443 def shell_quote(args):
1445 encoding = get_filesystem_encoding()
1447 if isinstance(a, bytes):
1448 # We may get a filename encoded with 'encodeFilename'
1449 a = a.decode(encoding)
1450 quoted_args.append(pipes.quote(a))
1451 return ' '.join(quoted_args)
1454 def smuggle_url(url, data):
1455 """ Pass additional data in a URL for internal use. """
1457 url, idata = unsmuggle_url(url, {})
1459 sdata = compat_urllib_parse_urlencode(
1460 {'__youtubedl_smuggle': json.dumps(data)})
1461 return url + '#' + sdata
1464 def unsmuggle_url(smug_url, default=None):
1465 if '#__youtubedl_smuggle' not in smug_url:
1466 return smug_url, default
1467 url, _, sdata = smug_url.rpartition('#')
1468 jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1469 data = json.loads(jsond)
1473 def format_bytes(bytes):
1476 if type(bytes) is str:
1477 bytes = float(bytes)
1481 exponent = int(math.log(bytes, 1024.0))
1482 suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1483 converted = float(bytes) / float(1024 ** exponent)
1484 return '%.2f%s' % (converted, suffix)
1487 def lookup_unit_table(unit_table, s):
1488 units_re = '|'.join(re.escape(u) for u in unit_table)
1490 r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re, s)
1493 num_str = m.group('num').replace(',', '.')
1494 mult = unit_table[m.group('unit')]
1495 return int(float(num_str) * mult)
1498 def parse_filesize(s):
1502 # The lower-case forms are of course incorrect and unofficial,
1503 # but we support those too
1541 return lookup_unit_table(_UNIT_TABLE, s)
1550 if re.match(r'^[\d,.]+$', s):
1551 return str_to_int(s)
1562 return lookup_unit_table(_UNIT_TABLE, s)
1565 def month_by_name(name):
1566 """ Return the number of a month by (locale-independently) English name """
1569 return ENGLISH_MONTH_NAMES.index(name) + 1
1574 def month_by_abbreviation(abbrev):
1575 """ Return the number of a month by (locale-independently) English
1579 return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
1584 def fix_xml_ampersands(xml_str):
1585 """Replace all the '&' by '&' in XML"""
1587 r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1592 def setproctitle(title):
1593 assert isinstance(title, compat_str)
1595 # ctypes in Jython is not complete
1596 # http://bugs.jython.org/issue2148
1597 if sys.platform.startswith('java'):
1601 libc = ctypes.cdll.LoadLibrary('libc.so.6')
1604 title_bytes = title.encode('utf-8')
1605 buf = ctypes.create_string_buffer(len(title_bytes))
1606 buf.value = title_bytes
1608 libc.prctl(15, buf, 0, 0, 0)
1609 except AttributeError:
1610 return # Strange libc, just skip this
1613 def remove_start(s, start):
1614 return s[len(start):] if s is not None and s.startswith(start) else s
1617 def remove_end(s, end):
1618 return s[:-len(end)] if s is not None and s.endswith(end) else s
1621 def remove_quotes(s):
1622 if s is None or len(s) < 2:
1624 for quote in ('"', "'", ):
1625 if s[0] == quote and s[-1] == quote:
1630 def url_basename(url):
1631 path = compat_urlparse.urlparse(url).path
1632 return path.strip('/').split('/')[-1]
1635 class HEADRequest(compat_urllib_request.Request):
1636 def get_method(self):
1640 class PUTRequest(compat_urllib_request.Request):
1641 def get_method(self):
1645 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1648 v = getattr(v, get_attr, None)
1654 return int(v) * invscale // scale
1659 def str_or_none(v, default=None):
1660 return default if v is None else compat_str(v)
1663 def str_to_int(int_str):
1664 """ A more relaxed version of int_or_none """
1667 int_str = re.sub(r'[,\.\+]', '', int_str)
1671 def float_or_none(v, scale=1, invscale=1, default=None):
1675 return float(v) * invscale / scale
1680 def strip_or_none(v):
1681 return None if v is None else v.strip()
1684 def parse_duration(s):
1685 if not isinstance(s, compat_basestring):
1690 days, hours, mins, secs, ms = [None] * 5
1691 m = re.match(r'(?:(?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?(?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?$', s)
1693 days, hours, mins, secs, ms = m.groups()
1698 (?P<days>[0-9]+)\s*d(?:ays?)?\s*
1701 (?P<hours>[0-9]+)\s*h(?:ours?)?\s*
1704 (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?\s*
1707 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
1710 days, hours, mins, secs, ms = m.groups()
1712 m = re.match(r'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)$', s)
1714 hours, mins = m.groups()
1720 duration += float(secs)
1722 duration += float(mins) * 60
1724 duration += float(hours) * 60 * 60
1726 duration += float(days) * 24 * 60 * 60
1728 duration += float(ms)
1732 def prepend_extension(filename, ext, expected_real_ext=None):
1733 name, real_ext = os.path.splitext(filename)
1735 '{0}.{1}{2}'.format(name, ext, real_ext)
1736 if not expected_real_ext or real_ext[1:] == expected_real_ext
1737 else '{0}.{1}'.format(filename, ext))
1740 def replace_extension(filename, ext, expected_real_ext=None):
1741 name, real_ext = os.path.splitext(filename)
1742 return '{0}.{1}'.format(
1743 name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
1747 def check_executable(exe, args=[]):
1748 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1749 args can be a list of arguments for a short output (like -version) """
1751 subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1757 def get_exe_version(exe, args=['--version'],
1758 version_re=None, unrecognized='present'):
1759 """ Returns the version of the specified executable,
1760 or False if the executable is not present """
1762 out, _ = subprocess.Popen(
1763 [encodeArgument(exe)] + args,
1764 stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1767 if isinstance(out, bytes): # Python 2.x
1768 out = out.decode('ascii', 'ignore')
1769 return detect_exe_version(out, version_re, unrecognized)
1772 def detect_exe_version(output, version_re=None, unrecognized='present'):
1773 assert isinstance(output, compat_str)
1774 if version_re is None:
1775 version_re = r'version\s+([-0-9._a-zA-Z]+)'
1776 m = re.search(version_re, output)
1783 class PagedList(object):
1785 # This is only useful for tests
1786 return len(self.getslice())
1789 class OnDemandPagedList(PagedList):
1790 def __init__(self, pagefunc, pagesize, use_cache=False):
1791 self._pagefunc = pagefunc
1792 self._pagesize = pagesize
1793 self._use_cache = use_cache
1797 def getslice(self, start=0, end=None):
1799 for pagenum in itertools.count(start // self._pagesize):
1800 firstid = pagenum * self._pagesize
1801 nextfirstid = pagenum * self._pagesize + self._pagesize
1802 if start >= nextfirstid:
1807 page_results = self._cache.get(pagenum)
1808 if page_results is None:
1809 page_results = list(self._pagefunc(pagenum))
1811 self._cache[pagenum] = page_results
1814 start % self._pagesize
1815 if firstid <= start < nextfirstid
1819 ((end - 1) % self._pagesize) + 1
1820 if (end is not None and firstid <= end <= nextfirstid)
1823 if startv != 0 or endv is not None:
1824 page_results = page_results[startv:endv]
1825 res.extend(page_results)
1827 # A little optimization - if current page is not "full", ie. does
1828 # not contain page_size videos then we can assume that this page
1829 # is the last one - there are no more ids on further pages -
1830 # i.e. no need to query again.
1831 if len(page_results) + startv < self._pagesize:
1834 # If we got the whole page, but the next page is not interesting,
1835 # break out early as well
1836 if end == nextfirstid:
1841 class InAdvancePagedList(PagedList):
1842 def __init__(self, pagefunc, pagecount, pagesize):
1843 self._pagefunc = pagefunc
1844 self._pagecount = pagecount
1845 self._pagesize = pagesize
1847 def getslice(self, start=0, end=None):
1849 start_page = start // self._pagesize
1851 self._pagecount if end is None else (end // self._pagesize + 1))
1852 skip_elems = start - start_page * self._pagesize
1853 only_more = None if end is None else end - start
1854 for pagenum in range(start_page, end_page):
1855 page = list(self._pagefunc(pagenum))
1857 page = page[skip_elems:]
1859 if only_more is not None:
1860 if len(page) < only_more:
1861 only_more -= len(page)
1863 page = page[:only_more]
1870 def uppercase_escape(s):
1871 unicode_escape = codecs.getdecoder('unicode_escape')
1873 r'\\U[0-9a-fA-F]{8}',
1874 lambda m: unicode_escape(m.group(0))[0],
1878 def lowercase_escape(s):
1879 unicode_escape = codecs.getdecoder('unicode_escape')
1881 r'\\u[0-9a-fA-F]{4}',
1882 lambda m: unicode_escape(m.group(0))[0],
1886 def escape_rfc3986(s):
1887 """Escape non-ASCII characters as suggested by RFC 3986"""
1888 if sys.version_info < (3, 0) and isinstance(s, compat_str):
1889 s = s.encode('utf-8')
1890 return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
1893 def escape_url(url):
1894 """Escape URL as suggested by RFC 3986"""
1895 url_parsed = compat_urllib_parse_urlparse(url)
1896 return url_parsed._replace(
1897 netloc=url_parsed.netloc.encode('idna').decode('ascii'),
1898 path=escape_rfc3986(url_parsed.path),
1899 params=escape_rfc3986(url_parsed.params),
1900 query=escape_rfc3986(url_parsed.query),
1901 fragment=escape_rfc3986(url_parsed.fragment)
1905 def read_batch_urls(batch_fd):
1907 if not isinstance(url, compat_str):
1908 url = url.decode('utf-8', 'replace')
1909 BOM_UTF8 = '\xef\xbb\xbf'
1910 if url.startswith(BOM_UTF8):
1911 url = url[len(BOM_UTF8):]
1913 if url.startswith(('#', ';', ']')):
1917 with contextlib.closing(batch_fd) as fd:
1918 return [url for url in map(fixup, fd) if url]
1921 def urlencode_postdata(*args, **kargs):
1922 return compat_urllib_parse_urlencode(*args, **kargs).encode('ascii')
1925 def update_url_query(url, query):
1928 parsed_url = compat_urlparse.urlparse(url)
1929 qs = compat_parse_qs(parsed_url.query)
1931 return compat_urlparse.urlunparse(parsed_url._replace(
1932 query=compat_urllib_parse_urlencode(qs, True)))
1935 def update_Request(req, url=None, data=None, headers={}, query={}):
1936 req_headers = req.headers.copy()
1937 req_headers.update(headers)
1938 req_data = data or req.data
1939 req_url = update_url_query(url or req.get_full_url(), query)
1940 req_get_method = req.get_method()
1941 if req_get_method == 'HEAD':
1942 req_type = HEADRequest
1943 elif req_get_method == 'PUT':
1944 req_type = PUTRequest
1946 req_type = compat_urllib_request.Request
1948 req_url, data=req_data, headers=req_headers,
1949 origin_req_host=req.origin_req_host, unverifiable=req.unverifiable)
1950 if hasattr(req, 'timeout'):
1951 new_req.timeout = req.timeout
1955 def dict_get(d, key_or_keys, default=None, skip_false_values=True):
1956 if isinstance(key_or_keys, (list, tuple)):
1957 for key in key_or_keys:
1958 if key not in d or d[key] is None or skip_false_values and not d[key]:
1962 return d.get(key_or_keys, default)
1965 def try_get(src, getter, expected_type=None):
1968 except (AttributeError, KeyError, TypeError, IndexError):
1971 if expected_type is None or isinstance(v, expected_type):
1975 def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
1976 return string if isinstance(string, compat_str) else compat_str(string, encoding, errors)
1988 TV_PARENTAL_GUIDELINES = {
1998 def parse_age_limit(s):
2000 return s if 0 <= s <= 21 else None
2001 if not isinstance(s, compat_basestring):
2003 m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
2005 return int(m.group('age'))
2007 return US_RATINGS[s]
2008 return TV_PARENTAL_GUIDELINES.get(s)
2011 def strip_jsonp(code):
2013 r'(?s)^[a-zA-Z0-9_.$]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
2016 def js_to_json(code):
2019 if v in ('true', 'false', 'null'):
2021 elif v.startswith('/*') or v == ',':
2024 if v[0] in ("'", '"'):
2025 v = re.sub(r'(?s)\\.|"', lambda m: {
2030 }.get(m.group(0), m.group(0)), v[1:-1])
2033 (r'^0[xX][0-9a-fA-F]+', 16),
2037 for regex, base in INTEGER_TABLE:
2038 im = re.match(regex, v)
2040 i = int(im.group(0), base)
2041 return '"%d":' % i if v.endswith(':') else '%d' % i
2045 return re.sub(r'''(?sx)
2046 "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
2047 '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
2048 /\*.*?\*/|,(?=\s*[\]}])|
2049 [a-zA-Z_][.a-zA-Z_0-9]*|
2050 \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:\s*:)?|
2055 def qualities(quality_ids):
2056 """ Get a numeric quality value out of a list of possible values """
2059 return quality_ids.index(qid)
2065 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
2068 def limit_length(s, length):
2069 """ Add ellipses to overly long strings """
2074 return s[:length - len(ELLIPSES)] + ELLIPSES
2078 def version_tuple(v):
2079 return tuple(int(e) for e in re.split(r'[-.]', v))
2082 def is_outdated_version(version, limit, assume_new=True):
2084 return not assume_new
2086 return version_tuple(version) < version_tuple(limit)
2088 return not assume_new
2091 def ytdl_is_updateable():
2092 """ Returns if youtube-dl can be updated with -U """
2093 from zipimport import zipimporter
2095 return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
2098 def args_to_str(args):
2099 # Get a short string representation for a subprocess command
2100 return ' '.join(compat_shlex_quote(a) for a in args)
2103 def error_to_compat_str(err):
2105 # On python 2 error byte string must be decoded with proper
2106 # encoding rather than ascii
2107 if sys.version_info[0] < 3:
2108 err_str = err_str.decode(preferredencoding())
2112 def mimetype2ext(mt):
2118 # Per RFC 3003, audio/mpeg can be .mp1, .mp2 or .mp3. Here use .mp3 as
2119 # it's the most popular one
2120 'audio/mpeg': 'mp3',
2125 _, _, res = mt.rpartition('/')
2130 'smptett+xml': 'tt',
2136 'x-mp4-fragmented': 'mp4',
2139 'x-mpegurl': 'm3u8',
2140 'vnd.apple.mpegurl': 'm3u8',
2145 'vnd.ms-sstr+xml': 'ism',
2149 def parse_codecs(codecs_str):
2150 # http://tools.ietf.org/html/rfc6381
2153 splited_codecs = list(filter(None, map(
2154 lambda str: str.strip(), codecs_str.strip().strip(',').split(','))))
2155 vcodec, acodec = None, None
2156 for full_codec in splited_codecs:
2157 codec = full_codec.split('.')[0]
2158 if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v'):
2161 elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac'):
2165 write_string('WARNING: Unknown codec %s' % full_codec, sys.stderr)
2166 if not vcodec and not acodec:
2167 if len(splited_codecs) == 2:
2172 elif len(splited_codecs) == 1:
2179 'vcodec': vcodec or 'none',
2180 'acodec': acodec or 'none',
2185 def urlhandle_detect_ext(url_handle):
2186 getheader = url_handle.headers.get
2188 cd = getheader('Content-Disposition')
2190 m = re.match(r'attachment;\s*filename="(?P<filename>[^"]+)"', cd)
2192 e = determine_ext(m.group('filename'), default_ext=None)
2196 return mimetype2ext(getheader('Content-Type'))
2199 def encode_data_uri(data, mime_type):
2200 return 'data:%s;base64,%s' % (mime_type, base64.b64encode(data).decode('ascii'))
2203 def age_restricted(content_limit, age_limit):
2204 """ Returns True iff the content should be blocked """
2206 if age_limit is None: # No limit set
2208 if content_limit is None:
2209 return False # Content available for everyone
2210 return age_limit < content_limit
2213 def is_html(first_bytes):
2214 """ Detect whether a file contains HTML by examining its first bytes. """
2217 (b'\xef\xbb\xbf', 'utf-8'),
2218 (b'\x00\x00\xfe\xff', 'utf-32-be'),
2219 (b'\xff\xfe\x00\x00', 'utf-32-le'),
2220 (b'\xff\xfe', 'utf-16-le'),
2221 (b'\xfe\xff', 'utf-16-be'),
2223 for bom, enc in BOMS:
2224 if first_bytes.startswith(bom):
2225 s = first_bytes[len(bom):].decode(enc, 'replace')
2228 s = first_bytes.decode('utf-8', 'replace')
2230 return re.match(r'^\s*<', s)
2233 def determine_protocol(info_dict):
2234 protocol = info_dict.get('protocol')
2235 if protocol is not None:
2238 url = info_dict['url']
2239 if url.startswith('rtmp'):
2241 elif url.startswith('mms'):
2243 elif url.startswith('rtsp'):
2246 ext = determine_ext(url)
2252 return compat_urllib_parse_urlparse(url).scheme
2255 def render_table(header_row, data):
2256 """ Render a list of rows, each as a list of values """
2257 table = [header_row] + data
2258 max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
2259 format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
2260 return '\n'.join(format_str % tuple(row) for row in table)
2263 def _match_one(filter_part, dct):
2264 COMPARISON_OPERATORS = {
2272 operator_rex = re.compile(r'''(?x)\s*
2274 \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
2276 (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
2277 (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
2280 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
2281 m = operator_rex.search(filter_part)
2283 op = COMPARISON_OPERATORS[m.group('op')]
2284 if m.group('strval') is not None:
2285 if m.group('op') not in ('=', '!='):
2287 'Operator %s does not support string values!' % m.group('op'))
2288 comparison_value = m.group('strval')
2291 comparison_value = int(m.group('intval'))
2293 comparison_value = parse_filesize(m.group('intval'))
2294 if comparison_value is None:
2295 comparison_value = parse_filesize(m.group('intval') + 'B')
2296 if comparison_value is None:
2298 'Invalid integer value %r in filter part %r' % (
2299 m.group('intval'), filter_part))
2300 actual_value = dct.get(m.group('key'))
2301 if actual_value is None:
2302 return m.group('none_inclusive')
2303 return op(actual_value, comparison_value)
2306 '': lambda v: v is not None,
2307 '!': lambda v: v is None,
2309 operator_rex = re.compile(r'''(?x)\s*
2310 (?P<op>%s)\s*(?P<key>[a-z_]+)
2312 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
2313 m = operator_rex.search(filter_part)
2315 op = UNARY_OPERATORS[m.group('op')]
2316 actual_value = dct.get(m.group('key'))
2317 return op(actual_value)
2319 raise ValueError('Invalid filter part %r' % filter_part)
2322 def match_str(filter_str, dct):
2323 """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
2326 _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
2329 def match_filter_func(filter_str):
2330 def _match_func(info_dict):
2331 if match_str(filter_str, info_dict):
2334 video_title = info_dict.get('title', info_dict.get('id', 'video'))
2335 return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
2339 def parse_dfxp_time_expr(time_expr):
2343 mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
2345 return float(mobj.group('time_offset'))
2347 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
2349 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
2352 def srt_subtitles_timecode(seconds):
2353 return '%02d:%02d:%02d,%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
2356 def dfxp2srt(dfxp_data):
2357 _x = functools.partial(xpath_with_ns, ns_map={
2358 'ttml': 'http://www.w3.org/ns/ttml',
2359 'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
2360 'ttaf1_0604': 'http://www.w3.org/2006/04/ttaf1',
2363 class TTMLPElementParser(object):
2366 def start(self, tag, attrib):
2367 if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
2373 def data(self, data):
2377 return self.out.strip()
2379 def parse_node(node):
2380 target = TTMLPElementParser()
2381 parser = xml.etree.ElementTree.XMLParser(target=target)
2382 parser.feed(xml.etree.ElementTree.tostring(node))
2383 return parser.close()
2385 dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
2387 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall(_x('.//ttaf1_0604:p')) or dfxp.findall('.//p')
2390 raise ValueError('Invalid dfxp/TTML subtitle')
2392 for para, index in zip(paras, itertools.count(1)):
2393 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
2394 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
2395 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
2396 if begin_time is None:
2401 end_time = begin_time + dur
2402 out.append('%d\n%s --> %s\n%s\n\n' % (
2404 srt_subtitles_timecode(begin_time),
2405 srt_subtitles_timecode(end_time),
2411 def cli_option(params, command_option, param):
2412 param = params.get(param)
2413 return [command_option, param] if param is not None else []
2416 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
2417 param = params.get(param)
2418 assert isinstance(param, bool)
2420 return [command_option + separator + (true_value if param else false_value)]
2421 return [command_option, true_value if param else false_value]
2424 def cli_valueless_option(params, command_option, param, expected_value=True):
2425 param = params.get(param)
2426 return [command_option] if param == expected_value else []
2429 def cli_configuration_args(params, param, default=[]):
2430 ex_args = params.get(param)
2433 assert isinstance(ex_args, list)
2437 class ISO639Utils(object):
2438 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
2627 def short2long(cls, code):
2628 """Convert language code from ISO 639-1 to ISO 639-2/T"""
2629 return cls._lang_map.get(code[:2])
2632 def long2short(cls, code):
2633 """Convert language code from ISO 639-2/T to ISO 639-1"""
2634 for short_name, long_name in cls._lang_map.items():
2635 if long_name == code:
2639 class ISO3166Utils(object):
2640 # From http://data.okfn.org/data/core/country-list
2642 'AF': 'Afghanistan',
2643 'AX': 'Åland Islands',
2646 'AS': 'American Samoa',
2651 'AG': 'Antigua and Barbuda',
2668 'BO': 'Bolivia, Plurinational State of',
2669 'BQ': 'Bonaire, Sint Eustatius and Saba',
2670 'BA': 'Bosnia and Herzegovina',
2672 'BV': 'Bouvet Island',
2674 'IO': 'British Indian Ocean Territory',
2675 'BN': 'Brunei Darussalam',
2677 'BF': 'Burkina Faso',
2683 'KY': 'Cayman Islands',
2684 'CF': 'Central African Republic',
2688 'CX': 'Christmas Island',
2689 'CC': 'Cocos (Keeling) Islands',
2693 'CD': 'Congo, the Democratic Republic of the',
2694 'CK': 'Cook Islands',
2696 'CI': 'Côte d\'Ivoire',
2701 'CZ': 'Czech Republic',
2705 'DO': 'Dominican Republic',
2708 'SV': 'El Salvador',
2709 'GQ': 'Equatorial Guinea',
2713 'FK': 'Falkland Islands (Malvinas)',
2714 'FO': 'Faroe Islands',
2718 'GF': 'French Guiana',
2719 'PF': 'French Polynesia',
2720 'TF': 'French Southern Territories',
2735 'GW': 'Guinea-Bissau',
2738 'HM': 'Heard Island and McDonald Islands',
2739 'VA': 'Holy See (Vatican City State)',
2746 'IR': 'Iran, Islamic Republic of',
2749 'IM': 'Isle of Man',
2759 'KP': 'Korea, Democratic People\'s Republic of',
2760 'KR': 'Korea, Republic of',
2763 'LA': 'Lao People\'s Democratic Republic',
2769 'LI': 'Liechtenstein',
2773 'MK': 'Macedonia, the Former Yugoslav Republic of',
2780 'MH': 'Marshall Islands',
2786 'FM': 'Micronesia, Federated States of',
2787 'MD': 'Moldova, Republic of',
2798 'NL': 'Netherlands',
2799 'NC': 'New Caledonia',
2800 'NZ': 'New Zealand',
2805 'NF': 'Norfolk Island',
2806 'MP': 'Northern Mariana Islands',
2811 'PS': 'Palestine, State of',
2813 'PG': 'Papua New Guinea',
2816 'PH': 'Philippines',
2820 'PR': 'Puerto Rico',
2824 'RU': 'Russian Federation',
2826 'BL': 'Saint Barthélemy',
2827 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
2828 'KN': 'Saint Kitts and Nevis',
2829 'LC': 'Saint Lucia',
2830 'MF': 'Saint Martin (French part)',
2831 'PM': 'Saint Pierre and Miquelon',
2832 'VC': 'Saint Vincent and the Grenadines',
2835 'ST': 'Sao Tome and Principe',
2836 'SA': 'Saudi Arabia',
2840 'SL': 'Sierra Leone',
2842 'SX': 'Sint Maarten (Dutch part)',
2845 'SB': 'Solomon Islands',
2847 'ZA': 'South Africa',
2848 'GS': 'South Georgia and the South Sandwich Islands',
2849 'SS': 'South Sudan',
2854 'SJ': 'Svalbard and Jan Mayen',
2857 'CH': 'Switzerland',
2858 'SY': 'Syrian Arab Republic',
2859 'TW': 'Taiwan, Province of China',
2861 'TZ': 'Tanzania, United Republic of',
2863 'TL': 'Timor-Leste',
2867 'TT': 'Trinidad and Tobago',
2870 'TM': 'Turkmenistan',
2871 'TC': 'Turks and Caicos Islands',
2875 'AE': 'United Arab Emirates',
2876 'GB': 'United Kingdom',
2877 'US': 'United States',
2878 'UM': 'United States Minor Outlying Islands',
2882 'VE': 'Venezuela, Bolivarian Republic of',
2884 'VG': 'Virgin Islands, British',
2885 'VI': 'Virgin Islands, U.S.',
2886 'WF': 'Wallis and Futuna',
2887 'EH': 'Western Sahara',
2894 def short2full(cls, code):
2895 """Convert an ISO 3166-2 country code to the corresponding full name"""
2896 return cls._country_map.get(code.upper())
2899 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
2900 def __init__(self, proxies=None):
2901 # Set default handlers
2902 for type in ('http', 'https'):
2903 setattr(self, '%s_open' % type,
2904 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
2905 meth(r, proxy, type))
2906 return compat_urllib_request.ProxyHandler.__init__(self, proxies)
2908 def proxy_open(self, req, proxy, type):
2909 req_proxy = req.headers.get('Ytdl-request-proxy')
2910 if req_proxy is not None:
2912 del req.headers['Ytdl-request-proxy']
2914 if proxy == '__noproxy__':
2915 return None # No Proxy
2916 if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
2917 req.add_header('Ytdl-socks-proxy', proxy)
2918 # youtube-dl's http/https handlers do wrapping the socket with socks
2920 return compat_urllib_request.ProxyHandler.proxy_open(
2921 self, req, proxy, type)
2924 def ohdave_rsa_encrypt(data, exponent, modulus):
2926 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
2929 data: data to encrypt, bytes-like object
2930 exponent, modulus: parameter e and N of RSA algorithm, both integer
2931 Output: hex string of encrypted data
2933 Limitation: supports one block encryption only
2936 payload = int(binascii.hexlify(data[::-1]), 16)
2937 encrypted = pow(payload, exponent, modulus)
2938 return '%x' % encrypted
2941 def encode_base_n(num, n, table=None):
2942 FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
2944 table = FULL_TABLE[:n]
2947 raise ValueError('base %d exceeds table length %d' % (n, len(table)))
2954 ret = table[num % n] + ret
2959 def decode_packed_codes(code):
2961 r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)",
2963 obfucasted_code, base, count, symbols = mobj.groups()
2966 symbols = symbols.split('|')
2971 base_n_count = encode_base_n(count, base)
2972 symbol_table[base_n_count] = symbols[count] or base_n_count
2975 r'\b(\w+)\b', lambda mobj: symbol_table[mobj.group(0)],
2979 def parse_m3u8_attributes(attrib):
2981 for (key, val) in re.findall(r'(?P<key>[A-Z0-9-]+)=(?P<val>"[^"]+"|[^",]+)(?:,|$)', attrib):
2982 if val.startswith('"'):
2988 def urshift(val, n):
2989 return val >> n if val >= 0 else (val + 0x100000000) >> n
2992 # Based on png2str() written by @gdkchan and improved by @yokrysty
2993 # Originally posted at https://github.com/rg3/youtube-dl/issues/9706
2994 def decode_png(png_data):
2995 # Reference: https://www.w3.org/TR/PNG/
2996 header = png_data[8:]
2998 if png_data[:8] != b'\x89PNG\x0d\x0a\x1a\x0a' or header[4:8] != b'IHDR':
2999 raise IOError('Not a valid PNG file.')
3001 int_map = {1: '>B', 2: '>H', 4: '>I'}
3002 unpack_integer = lambda x: compat_struct_unpack(int_map[len(x)], x)[0]
3007 length = unpack_integer(header[:4])
3010 chunk_type = header[:4]
3013 chunk_data = header[:length]
3014 header = header[length:]
3016 header = header[4:] # Skip CRC
3024 ihdr = chunks[0]['data']
3026 width = unpack_integer(ihdr[:4])
3027 height = unpack_integer(ihdr[4:8])
3031 for chunk in chunks:
3032 if chunk['type'] == b'IDAT':
3033 idat += chunk['data']
3036 raise IOError('Unable to read PNG data.')
3038 decompressed_data = bytearray(zlib.decompress(idat))
3043 def _get_pixel(idx):
3048 for y in range(height):
3049 basePos = y * (1 + stride)
3050 filter_type = decompressed_data[basePos]
3054 pixels.append(current_row)
3056 for x in range(stride):
3057 color = decompressed_data[1 + basePos + x]
3058 basex = y * stride + x
3063 left = _get_pixel(basex - 3)
3065 up = _get_pixel(basex - stride)
3067 if filter_type == 1: # Sub
3068 color = (color + left) & 0xff
3069 elif filter_type == 2: # Up
3070 color = (color + up) & 0xff
3071 elif filter_type == 3: # Average
3072 color = (color + ((left + up) >> 1)) & 0xff
3073 elif filter_type == 4: # Paeth
3079 c = _get_pixel(basex - stride - 3)
3087 if pa <= pb and pa <= pc:
3088 color = (color + a) & 0xff
3090 color = (color + b) & 0xff
3092 color = (color + c) & 0xff
3094 current_row.append(color)
3096 return width, height, pixels