2 # -*- coding: utf-8 -*-
4 from __future__ import unicode_literals
34 import xml.etree.ElementTree
41 compat_etree_fromstring,
46 compat_socket_create_connection,
50 compat_urllib_parse_urlparse,
51 compat_urllib_request,
57 # This is not clearly defined otherwise
58 compiled_regex_type = type(re.compile(''))
61 'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/44.0 (Chrome)',
62 'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
63 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
64 'Accept-Encoding': 'gzip, deflate',
65 'Accept-Language': 'en-us,en;q=0.5',
71 ENGLISH_MONTH_NAMES = [
72 'January', 'February', 'March', 'April', 'May', 'June',
73 'July', 'August', 'September', 'October', 'November', 'December']
76 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
77 'flv', 'f4v', 'f4a', 'f4b',
78 'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
88 'f4f', 'f4m', 'm3u8', 'smil')
91 def preferredencoding():
92 """Get preferred encoding.
94 Returns the best encoding scheme for the system, based on
95 locale.getpreferredencoding() and some further tweaks.
98 pref = locale.getpreferredencoding()
106 def write_json_file(obj, fn):
107 """ Encode obj as JSON and write it to fn, atomically if possible """
109 fn = encodeFilename(fn)
110 if sys.version_info < (3, 0) and sys.platform != 'win32':
111 encoding = get_filesystem_encoding()
112 # os.path.basename returns a bytes object, but NamedTemporaryFile
113 # will fail if the filename contains non ascii characters unless we
114 # use a unicode object
115 path_basename = lambda f: os.path.basename(fn).decode(encoding)
116 # the same for os.path.dirname
117 path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
119 path_basename = os.path.basename
120 path_dirname = os.path.dirname
124 'prefix': path_basename(fn) + '.',
125 'dir': path_dirname(fn),
129 # In Python 2.x, json.dump expects a bytestream.
130 # In Python 3.x, it writes to a character stream
131 if sys.version_info < (3, 0):
139 tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
144 if sys.platform == 'win32':
145 # Need to remove existing file on Windows, else os.rename raises
146 # WindowsError or FileExistsError.
151 os.rename(tf.name, fn)
160 if sys.version_info >= (2, 7):
161 def find_xpath_attr(node, xpath, key, val=None):
162 """ Find the xpath xpath[@key=val] """
163 assert re.match(r'^[a-zA-Z_-]+$', key)
164 expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
165 return node.find(expr)
167 def find_xpath_attr(node, xpath, key, val=None):
168 # Here comes the crazy part: In 2.6, if the xpath is a unicode,
169 # .//node does not match if a node is a direct child of . !
170 if isinstance(xpath, compat_str):
171 xpath = xpath.encode('ascii')
173 for f in node.findall(xpath):
174 if key not in f.attrib:
176 if val is None or f.attrib.get(key) == val:
180 # On python2.6 the xml.etree.ElementTree.Element methods don't support
181 # the namespace parameter
184 def xpath_with_ns(path, ns_map):
185 components = [c.split(':') for c in path.split('/')]
189 replaced.append(c[0])
192 replaced.append('{%s}%s' % (ns_map[ns], tag))
193 return '/'.join(replaced)
196 def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
197 def _find_xpath(xpath):
198 if sys.version_info < (2, 7): # Crazy 2.6
199 xpath = xpath.encode('ascii')
200 return node.find(xpath)
202 if isinstance(xpath, (str, compat_str)):
203 n = _find_xpath(xpath)
211 if default is not NO_DEFAULT:
214 name = xpath if name is None else name
215 raise ExtractorError('Could not find XML element %s' % name)
221 def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
222 n = xpath_element(node, xpath, name, fatal=fatal, default=default)
223 if n is None or n == default:
226 if default is not NO_DEFAULT:
229 name = xpath if name is None else name
230 raise ExtractorError('Could not find XML element\'s text %s' % name)
236 def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
237 n = find_xpath_attr(node, xpath, key)
239 if default is not NO_DEFAULT:
242 name = '%s[@%s]' % (xpath, key) if name is None else name
243 raise ExtractorError('Could not find XML attribute %s' % name)
249 def get_element_by_id(id, html):
250 """Return the content of the tag with the specified ID in the passed HTML document"""
251 return get_element_by_attribute('id', id, html)
254 def get_element_by_attribute(attribute, value, html):
255 """Return the content of the tag with the specified attribute in the passed HTML document"""
257 m = re.search(r'''(?xs)
259 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
261 (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]+|="[^"]+"|='[^']+'))*?
265 ''' % (re.escape(attribute), re.escape(value)), html)
269 res = m.group('content')
271 if res.startswith('"') or res.startswith("'"):
274 return unescapeHTML(res)
276 class HTMLAttributeParser(compat_HTMLParser):
277 """Trivial HTML parser to gather the attributes for a single element"""
280 compat_HTMLParser.__init__(self)
282 def handle_starttag(self, tag, attrs):
283 self.attrs = dict(attrs)
285 def extract_attributes(html_element):
286 """Given a string for an HTML element such as
288 a="foo" B="bar" c="&98;az" d=boz
289 empty= noval entity="&"
292 Decode and return a dictionary of attributes.
294 'a': 'foo', 'b': 'bar', c: 'baz', d: 'boz',
295 'empty': '', 'noval': None, 'entity': '&',
296 'sq': '"', 'dq': '\''
298 NB HTMLParser is stricter in Python 2.6 & 3.2 than in later versions,
299 but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
301 parser = HTMLAttributeParser()
302 parser.feed(html_element)
306 def clean_html(html):
307 """Clean an HTML snippet into a readable string"""
309 if html is None: # Convenience for sanitizing descriptions etc.
313 html = html.replace('\n', ' ')
314 html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
315 html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
317 html = re.sub('<.*?>', '', html)
318 # Replace html entities
319 html = unescapeHTML(html)
323 def sanitize_open(filename, open_mode):
324 """Try to open the given filename, and slightly tweak it if this fails.
326 Attempts to open the given filename. If this fails, it tries to change
327 the filename slightly, step by step, until it's either able to open it
328 or it fails and raises a final exception, like the standard open()
331 It returns the tuple (stream, definitive_file_name).
335 if sys.platform == 'win32':
337 msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
338 return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
339 stream = open(encodeFilename(filename), open_mode)
340 return (stream, filename)
341 except (IOError, OSError) as err:
342 if err.errno in (errno.EACCES,):
345 # In case of error, try to remove win32 forbidden chars
346 alt_filename = sanitize_path(filename)
347 if alt_filename == filename:
350 # An exception here should be caught in the caller
351 stream = open(encodeFilename(alt_filename), open_mode)
352 return (stream, alt_filename)
355 def timeconvert(timestr):
356 """Convert RFC 2822 defined time string into system timestamp"""
358 timetuple = email.utils.parsedate_tz(timestr)
359 if timetuple is not None:
360 timestamp = email.utils.mktime_tz(timetuple)
364 def sanitize_filename(s, restricted=False, is_id=False):
365 """Sanitizes a string so it could be used as part of a filename.
366 If restricted is set, use a stricter subset of allowed characters.
367 Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
369 def replace_insane(char):
370 if char == '?' or ord(char) < 32 or ord(char) == 127:
373 return '' if restricted else '\''
375 return '_-' if restricted else ' -'
376 elif char in '\\/|*<>':
378 if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
380 if restricted and ord(char) > 127:
385 s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
386 result = ''.join(map(replace_insane, s))
388 while '__' in result:
389 result = result.replace('__', '_')
390 result = result.strip('_')
391 # Common case of "Foreign band name - English song title"
392 if restricted and result.startswith('-_'):
394 if result.startswith('-'):
395 result = '_' + result[len('-'):]
396 result = result.lstrip('.')
402 def sanitize_path(s):
403 """Sanitizes and normalizes path on Windows"""
404 if sys.platform != 'win32':
406 drive_or_unc, _ = os.path.splitdrive(s)
407 if sys.version_info < (2, 7) and not drive_or_unc:
408 drive_or_unc, _ = os.path.splitunc(s)
409 norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
413 path_part if path_part in ['.', '..'] else re.sub('(?:[/<>:"\\|\\\\?\\*]|[\s.]$)', '#', path_part)
414 for path_part in norm_path]
416 sanitized_path.insert(0, drive_or_unc + os.path.sep)
417 return os.path.join(*sanitized_path)
420 # Prepend protocol-less URLs with `http:` scheme in order to mitigate the number of
421 # unwanted failures due to missing protocol
422 def sanitized_Request(url, *args, **kwargs):
423 return compat_urllib_request.Request(
424 'http:%s' % url if url.startswith('//') else url, *args, **kwargs)
427 def orderedSet(iterable):
428 """ Remove all duplicates from the input iterable """
436 def _htmlentity_transform(entity):
437 """Transforms an HTML entity to a character."""
438 # Known non-numeric HTML entity
439 if entity in compat_html_entities.name2codepoint:
440 return compat_chr(compat_html_entities.name2codepoint[entity])
442 mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
444 numstr = mobj.group(1)
445 if numstr.startswith('x'):
447 numstr = '0%s' % numstr
450 # See https://github.com/rg3/youtube-dl/issues/7518
452 return compat_chr(int(numstr, base))
456 # Unknown entity in name, return its literal representation
457 return '&%s;' % entity
463 assert type(s) == compat_str
466 r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s)
469 def get_subprocess_encoding():
470 if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
471 # For subprocess calls, encode with locale encoding
472 # Refer to http://stackoverflow.com/a/9951851/35070
473 encoding = preferredencoding()
475 encoding = sys.getfilesystemencoding()
481 def encodeFilename(s, for_subprocess=False):
483 @param s The name of the file
486 assert type(s) == compat_str
488 # Python 3 has a Unicode API
489 if sys.version_info >= (3, 0):
492 # Pass '' directly to use Unicode APIs on Windows 2000 and up
493 # (Detecting Windows NT 4 is tricky because 'major >= 4' would
494 # match Windows 9x series as well. Besides, NT 4 is obsolete.)
495 if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
498 # Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
499 if sys.platform.startswith('java'):
502 return s.encode(get_subprocess_encoding(), 'ignore')
505 def decodeFilename(b, for_subprocess=False):
507 if sys.version_info >= (3, 0):
510 if not isinstance(b, bytes):
513 return b.decode(get_subprocess_encoding(), 'ignore')
516 def encodeArgument(s):
517 if not isinstance(s, compat_str):
518 # Legacy code that uses byte strings
519 # Uncomment the following line after fixing all post processors
520 # assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
521 s = s.decode('ascii')
522 return encodeFilename(s, True)
525 def decodeArgument(b):
526 return decodeFilename(b, True)
529 def decodeOption(optval):
532 if isinstance(optval, bytes):
533 optval = optval.decode(preferredencoding())
535 assert isinstance(optval, compat_str)
539 def formatSeconds(secs):
541 return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
543 return '%d:%02d' % (secs // 60, secs % 60)
548 def make_HTTPS_handler(params, **kwargs):
549 opts_no_check_certificate = params.get('nocheckcertificate', False)
550 if hasattr(ssl, 'create_default_context'): # Python >= 3.4 or 2.7.9
551 context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
552 if opts_no_check_certificate:
553 context.check_hostname = False
554 context.verify_mode = ssl.CERT_NONE
556 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
559 # (create_default_context present but HTTPSHandler has no context=)
562 if sys.version_info < (3, 2):
563 return YoutubeDLHTTPSHandler(params, **kwargs)
565 context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
566 context.verify_mode = (ssl.CERT_NONE
567 if opts_no_check_certificate
568 else ssl.CERT_REQUIRED)
569 context.set_default_verify_paths()
570 return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
573 def bug_reports_message():
574 if ytdl_is_updateable():
575 update_cmd = 'type youtube-dl -U to update'
577 update_cmd = 'see https://yt-dl.org/update on how to update'
578 msg = '; please report this issue on https://yt-dl.org/bug .'
579 msg += ' Make sure you are using the latest version; %s.' % update_cmd
580 msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
584 class ExtractorError(Exception):
585 """Error during info extraction."""
587 def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
588 """ tb, if given, is the original traceback (so that it can be printed out).
589 If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
592 if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
594 if video_id is not None:
595 msg = video_id + ': ' + msg
597 msg += ' (caused by %r)' % cause
599 msg += bug_reports_message()
600 super(ExtractorError, self).__init__(msg)
603 self.exc_info = sys.exc_info() # preserve original exception
605 self.video_id = video_id
607 def format_traceback(self):
608 if self.traceback is None:
610 return ''.join(traceback.format_tb(self.traceback))
613 class UnsupportedError(ExtractorError):
614 def __init__(self, url):
615 super(UnsupportedError, self).__init__(
616 'Unsupported URL: %s' % url, expected=True)
620 class RegexNotFoundError(ExtractorError):
621 """Error when a regex didn't match"""
625 class DownloadError(Exception):
626 """Download Error exception.
628 This exception may be thrown by FileDownloader objects if they are not
629 configured to continue on errors. They will contain the appropriate
633 def __init__(self, msg, exc_info=None):
634 """ exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
635 super(DownloadError, self).__init__(msg)
636 self.exc_info = exc_info
639 class SameFileError(Exception):
640 """Same File exception.
642 This exception will be thrown by FileDownloader objects if they detect
643 multiple files would have to be downloaded to the same file on disk.
648 class PostProcessingError(Exception):
649 """Post Processing exception.
651 This exception may be raised by PostProcessor's .run() method to
652 indicate an error in the postprocessing task.
655 def __init__(self, msg):
659 class MaxDownloadsReached(Exception):
660 """ --max-downloads limit has been reached. """
664 class UnavailableVideoError(Exception):
665 """Unavailable Format exception.
667 This exception will be thrown when a video is requested
668 in a format that is not available for that video.
673 class ContentTooShortError(Exception):
674 """Content Too Short exception.
676 This exception may be raised by FileDownloader objects when a file they
677 download is too small for what the server announced first, indicating
678 the connection was probably interrupted.
681 def __init__(self, downloaded, expected):
683 self.downloaded = downloaded
684 self.expected = expected
687 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
688 # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
689 # expected HTTP responses to meet HTTP/1.0 or later (see also
690 # https://github.com/rg3/youtube-dl/issues/6727)
691 if sys.version_info < (3, 0):
692 kwargs[b'strict'] = True
693 hc = http_class(*args, **kwargs)
694 source_address = ydl_handler._params.get('source_address')
695 if source_address is not None:
696 sa = (source_address, 0)
697 if hasattr(hc, 'source_address'): # Python 2.7+
698 hc.source_address = sa
700 def _hc_connect(self, *args, **kwargs):
701 sock = compat_socket_create_connection(
702 (self.host, self.port), self.timeout, sa)
704 self.sock = ssl.wrap_socket(
705 sock, self.key_file, self.cert_file,
706 ssl_version=ssl.PROTOCOL_TLSv1)
709 hc.connect = functools.partial(_hc_connect, hc)
714 def handle_youtubedl_headers(headers):
715 filtered_headers = headers
717 if 'Youtubedl-no-compression' in filtered_headers:
718 filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding')
719 del filtered_headers['Youtubedl-no-compression']
721 return filtered_headers
724 class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
725 """Handler for HTTP requests and responses.
727 This class, when installed with an OpenerDirector, automatically adds
728 the standard headers to every HTTP request and handles gzipped and
729 deflated responses from web servers. If compression is to be avoided in
730 a particular request, the original request in the program code only has
731 to include the HTTP header "Youtubedl-no-compression", which will be
732 removed before making the real request.
734 Part of this code was copied from:
736 http://techknack.net/python-urllib2-handlers/
738 Andrew Rowls, the author of that code, agreed to release it to the
742 def __init__(self, params, *args, **kwargs):
743 compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
744 self._params = params
746 def http_open(self, req):
747 return self.do_open(functools.partial(
748 _create_http_connection, self, compat_http_client.HTTPConnection, False),
754 return zlib.decompress(data, -zlib.MAX_WBITS)
756 return zlib.decompress(data)
759 def addinfourl_wrapper(stream, headers, url, code):
760 if hasattr(compat_urllib_request.addinfourl, 'getcode'):
761 return compat_urllib_request.addinfourl(stream, headers, url, code)
762 ret = compat_urllib_request.addinfourl(stream, headers, url)
766 def http_request(self, req):
767 # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
768 # always respected by websites, some tend to give out URLs with non percent-encoded
769 # non-ASCII characters (see telemb.py, ard.py [#3412])
770 # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
771 # To work around aforementioned issue we will replace request's original URL with
772 # percent-encoded one
773 # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
774 # the code of this workaround has been moved here from YoutubeDL.urlopen()
775 url = req.get_full_url()
776 url_escaped = escape_url(url)
778 # Substitute URL if any change after escaping
779 if url != url_escaped:
780 req_type = HEADRequest if req.get_method() == 'HEAD' else compat_urllib_request.Request
782 url_escaped, data=req.data, headers=req.headers,
783 origin_req_host=req.origin_req_host, unverifiable=req.unverifiable)
784 new_req.timeout = req.timeout
787 for h, v in std_headers.items():
788 # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
789 # The dict keys are capitalized because of this bug by urllib
790 if h.capitalize() not in req.headers:
793 req.headers = handle_youtubedl_headers(req.headers)
795 if sys.version_info < (2, 7) and '#' in req.get_full_url():
796 # Python 2.6 is brain-dead when it comes to fragments
797 req._Request__original = req._Request__original.partition('#')[0]
798 req._Request__r_type = req._Request__r_type.partition('#')[0]
802 def http_response(self, req, resp):
805 if resp.headers.get('Content-encoding', '') == 'gzip':
806 content = resp.read()
807 gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
809 uncompressed = io.BytesIO(gz.read())
810 except IOError as original_ioerror:
811 # There may be junk add the end of the file
812 # See http://stackoverflow.com/q/4928560/35070 for details
813 for i in range(1, 1024):
815 gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
816 uncompressed = io.BytesIO(gz.read())
821 raise original_ioerror
822 resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
823 resp.msg = old_resp.msg
824 del resp.headers['Content-encoding']
826 if resp.headers.get('Content-encoding', '') == 'deflate':
827 gz = io.BytesIO(self.deflate(resp.read()))
828 resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
829 resp.msg = old_resp.msg
830 del resp.headers['Content-encoding']
831 # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
832 # https://github.com/rg3/youtube-dl/issues/6457).
833 if 300 <= resp.code < 400:
834 location = resp.headers.get('Location')
836 # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
837 if sys.version_info >= (3, 0):
838 location = location.encode('iso-8859-1').decode('utf-8')
839 location_escaped = escape_url(location)
840 if location != location_escaped:
841 del resp.headers['Location']
842 resp.headers['Location'] = location_escaped
845 https_request = http_request
846 https_response = http_response
849 class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
850 def __init__(self, params, https_conn_class=None, *args, **kwargs):
851 compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
852 self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
853 self._params = params
855 def https_open(self, req):
857 if hasattr(self, '_context'): # python > 2.6
858 kwargs['context'] = self._context
859 if hasattr(self, '_check_hostname'): # python 3.x
860 kwargs['check_hostname'] = self._check_hostname
861 return self.do_open(functools.partial(
862 _create_http_connection, self, self._https_conn_class, True),
866 class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
867 def __init__(self, cookiejar=None):
868 compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
870 def http_response(self, request, response):
871 # Python 2 will choke on next HTTP request in row if there are non-ASCII
872 # characters in Set-Cookie HTTP header of last response (see
873 # https://github.com/rg3/youtube-dl/issues/6769).
874 # In order to at least prevent crashing we will percent encode Set-Cookie
875 # header before HTTPCookieProcessor starts processing it.
876 # if sys.version_info < (3, 0) and response.headers:
877 # for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
878 # set_cookie = response.headers.get(set_cookie_header)
880 # set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
881 # if set_cookie != set_cookie_escaped:
882 # del response.headers[set_cookie_header]
883 # response.headers[set_cookie_header] = set_cookie_escaped
884 return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
886 https_request = compat_urllib_request.HTTPCookieProcessor.http_request
887 https_response = http_response
890 def parse_iso8601(date_str, delimiter='T', timezone=None):
891 """ Return a UNIX timestamp from the given date """
896 date_str = re.sub(r'\.[0-9]+', '', date_str)
900 r'(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
903 timezone = datetime.timedelta()
905 date_str = date_str[:-len(m.group(0))]
906 if not m.group('sign'):
907 timezone = datetime.timedelta()
909 sign = 1 if m.group('sign') == '+' else -1
910 timezone = datetime.timedelta(
911 hours=sign * int(m.group('hours')),
912 minutes=sign * int(m.group('minutes')))
914 date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
915 dt = datetime.datetime.strptime(date_str, date_format) - timezone
916 return calendar.timegm(dt.timetuple())
921 def unified_strdate(date_str, day_first=True):
922 """Return a string with the date in the format YYYYMMDD"""
928 date_str = date_str.replace(',', ' ')
929 # %z (UTC offset) is only supported in python>=3.2
930 if not re.match(r'^[0-9]{1,2}-[0-9]{1,2}-[0-9]{4}$', date_str):
931 date_str = re.sub(r' ?(\+|-)[0-9]{2}:?[0-9]{2}$', '', date_str)
932 # Remove AM/PM + timezone
933 date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
935 format_expressions = [
948 '%Y-%m-%d %H:%M:%S.%f',
951 '%Y-%m-%dT%H:%M:%SZ',
952 '%Y-%m-%dT%H:%M:%S.%fZ',
953 '%Y-%m-%dT%H:%M:%S.%f0Z',
955 '%Y-%m-%dT%H:%M:%S.%f',
959 format_expressions.extend([
967 format_expressions.extend([
974 for expression in format_expressions:
976 upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
979 if upload_date is None:
980 timetuple = email.utils.parsedate_tz(date_str)
982 upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
983 if upload_date is not None:
984 return compat_str(upload_date)
987 def determine_ext(url, default_ext='unknown_video'):
990 guess = url.partition('?')[0].rpartition('.')[2]
991 if re.match(r'^[A-Za-z0-9]+$', guess):
993 # Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
994 elif guess.rstrip('/') in KNOWN_EXTENSIONS:
995 return guess.rstrip('/')
1000 def subtitles_filename(filename, sub_lang, sub_format):
1001 return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
1004 def date_from_str(date_str):
1006 Return a datetime object from a string in the format YYYYMMDD or
1007 (now|today)[+-][0-9](day|week|month|year)(s)?"""
1008 today = datetime.date.today()
1009 if date_str in ('now', 'today'):
1011 if date_str == 'yesterday':
1012 return today - datetime.timedelta(days=1)
1013 match = re.match('(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
1014 if match is not None:
1015 sign = match.group('sign')
1016 time = int(match.group('time'))
1019 unit = match.group('unit')
1020 # A bad approximation?
1024 elif unit == 'year':
1028 delta = datetime.timedelta(**{unit: time})
1029 return today + delta
1030 return datetime.datetime.strptime(date_str, '%Y%m%d').date()
1033 def hyphenate_date(date_str):
1035 Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
1036 match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
1037 if match is not None:
1038 return '-'.join(match.groups())
1043 class DateRange(object):
1044 """Represents a time interval between two dates"""
1046 def __init__(self, start=None, end=None):
1047 """start and end must be strings in the format accepted by date"""
1048 if start is not None:
1049 self.start = date_from_str(start)
1051 self.start = datetime.datetime.min.date()
1053 self.end = date_from_str(end)
1055 self.end = datetime.datetime.max.date()
1056 if self.start > self.end:
1057 raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
1061 """Returns a range that only contains the given day"""
1062 return cls(day, day)
1064 def __contains__(self, date):
1065 """Check if the date is in the range"""
1066 if not isinstance(date, datetime.date):
1067 date = date_from_str(date)
1068 return self.start <= date <= self.end
1071 return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
1074 def platform_name():
1075 """ Returns the platform name as a compat_str """
1076 res = platform.platform()
1077 if isinstance(res, bytes):
1078 res = res.decode(preferredencoding())
1080 assert isinstance(res, compat_str)
1084 def _windows_write_string(s, out):
1085 """ Returns True if the string was written using special methods,
1086 False if it has yet to be written out."""
1087 # Adapted from http://stackoverflow.com/a/3259271/35070
1090 import ctypes.wintypes
1098 fileno = out.fileno()
1099 except AttributeError:
1100 # If the output stream doesn't have a fileno, it's virtual
1102 except io.UnsupportedOperation:
1103 # Some strange Windows pseudo files?
1105 if fileno not in WIN_OUTPUT_IDS:
1108 GetStdHandle = ctypes.WINFUNCTYPE(
1109 ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
1110 (b'GetStdHandle', ctypes.windll.kernel32))
1111 h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
1113 WriteConsoleW = ctypes.WINFUNCTYPE(
1114 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
1115 ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
1116 ctypes.wintypes.LPVOID)((b'WriteConsoleW', ctypes.windll.kernel32))
1117 written = ctypes.wintypes.DWORD(0)
1119 GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b'GetFileType', ctypes.windll.kernel32))
1120 FILE_TYPE_CHAR = 0x0002
1121 FILE_TYPE_REMOTE = 0x8000
1122 GetConsoleMode = ctypes.WINFUNCTYPE(
1123 ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
1124 ctypes.POINTER(ctypes.wintypes.DWORD))(
1125 (b'GetConsoleMode', ctypes.windll.kernel32))
1126 INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
1128 def not_a_console(handle):
1129 if handle == INVALID_HANDLE_VALUE or handle is None:
1131 return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
1132 GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
1134 if not_a_console(h):
1137 def next_nonbmp_pos(s):
1139 return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
1140 except StopIteration:
1144 count = min(next_nonbmp_pos(s), 1024)
1146 ret = WriteConsoleW(
1147 h, s, count if count else 2, ctypes.byref(written), None)
1149 raise OSError('Failed to write string')
1150 if not count: # We just wrote a non-BMP character
1151 assert written.value == 2
1154 assert written.value > 0
1155 s = s[written.value:]
1159 def write_string(s, out=None, encoding=None):
1162 assert type(s) == compat_str
1164 if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
1165 if _windows_write_string(s, out):
1168 if ('b' in getattr(out, 'mode', '') or
1169 sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr
1170 byt = s.encode(encoding or preferredencoding(), 'ignore')
1172 elif hasattr(out, 'buffer'):
1173 enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
1174 byt = s.encode(enc, 'ignore')
1175 out.buffer.write(byt)
1181 def bytes_to_intlist(bs):
1184 if isinstance(bs[0], int): # Python 3
1187 return [ord(c) for c in bs]
1190 def intlist_to_bytes(xs):
1193 return struct_pack('%dB' % len(xs), *xs)
1196 # Cross-platform file locking
1197 if sys.platform == 'win32':
1198 import ctypes.wintypes
1201 class OVERLAPPED(ctypes.Structure):
1203 ('Internal', ctypes.wintypes.LPVOID),
1204 ('InternalHigh', ctypes.wintypes.LPVOID),
1205 ('Offset', ctypes.wintypes.DWORD),
1206 ('OffsetHigh', ctypes.wintypes.DWORD),
1207 ('hEvent', ctypes.wintypes.HANDLE),
1210 kernel32 = ctypes.windll.kernel32
1211 LockFileEx = kernel32.LockFileEx
1212 LockFileEx.argtypes = [
1213 ctypes.wintypes.HANDLE, # hFile
1214 ctypes.wintypes.DWORD, # dwFlags
1215 ctypes.wintypes.DWORD, # dwReserved
1216 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1217 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1218 ctypes.POINTER(OVERLAPPED) # Overlapped
1220 LockFileEx.restype = ctypes.wintypes.BOOL
1221 UnlockFileEx = kernel32.UnlockFileEx
1222 UnlockFileEx.argtypes = [
1223 ctypes.wintypes.HANDLE, # hFile
1224 ctypes.wintypes.DWORD, # dwReserved
1225 ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
1226 ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
1227 ctypes.POINTER(OVERLAPPED) # Overlapped
1229 UnlockFileEx.restype = ctypes.wintypes.BOOL
1230 whole_low = 0xffffffff
1231 whole_high = 0x7fffffff
1233 def _lock_file(f, exclusive):
1234 overlapped = OVERLAPPED()
1235 overlapped.Offset = 0
1236 overlapped.OffsetHigh = 0
1237 overlapped.hEvent = 0
1238 f._lock_file_overlapped_p = ctypes.pointer(overlapped)
1239 handle = msvcrt.get_osfhandle(f.fileno())
1240 if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
1241 whole_low, whole_high, f._lock_file_overlapped_p):
1242 raise OSError('Locking file failed: %r' % ctypes.FormatError())
1244 def _unlock_file(f):
1245 assert f._lock_file_overlapped_p
1246 handle = msvcrt.get_osfhandle(f.fileno())
1247 if not UnlockFileEx(handle, 0,
1248 whole_low, whole_high, f._lock_file_overlapped_p):
1249 raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
1252 # Some platforms, such as Jython, is missing fcntl
1256 def _lock_file(f, exclusive):
1257 fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
1259 def _unlock_file(f):
1260 fcntl.flock(f, fcntl.LOCK_UN)
1262 UNSUPPORTED_MSG = 'file locking is not supported on this platform'
1264 def _lock_file(f, exclusive):
1265 raise IOError(UNSUPPORTED_MSG)
1267 def _unlock_file(f):
1268 raise IOError(UNSUPPORTED_MSG)
1271 class locked_file(object):
1272 def __init__(self, filename, mode, encoding=None):
1273 assert mode in ['r', 'a', 'w']
1274 self.f = io.open(filename, mode, encoding=encoding)
1277 def __enter__(self):
1278 exclusive = self.mode != 'r'
1280 _lock_file(self.f, exclusive)
1286 def __exit__(self, etype, value, traceback):
1288 _unlock_file(self.f)
1295 def write(self, *args):
1296 return self.f.write(*args)
1298 def read(self, *args):
1299 return self.f.read(*args)
1302 def get_filesystem_encoding():
1303 encoding = sys.getfilesystemencoding()
1304 return encoding if encoding is not None else 'utf-8'
1307 def shell_quote(args):
1309 encoding = get_filesystem_encoding()
1311 if isinstance(a, bytes):
1312 # We may get a filename encoded with 'encodeFilename'
1313 a = a.decode(encoding)
1314 quoted_args.append(pipes.quote(a))
1315 return ' '.join(quoted_args)
1318 def smuggle_url(url, data):
1319 """ Pass additional data in a URL for internal use. """
1321 sdata = compat_urllib_parse.urlencode(
1322 {'__youtubedl_smuggle': json.dumps(data)})
1323 return url + '#' + sdata
1326 def unsmuggle_url(smug_url, default=None):
1327 if '#__youtubedl_smuggle' not in smug_url:
1328 return smug_url, default
1329 url, _, sdata = smug_url.rpartition('#')
1330 jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
1331 data = json.loads(jsond)
1335 def format_bytes(bytes):
1338 if type(bytes) is str:
1339 bytes = float(bytes)
1343 exponent = int(math.log(bytes, 1024.0))
1344 suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
1345 converted = float(bytes) / float(1024 ** exponent)
1346 return '%.2f%s' % (converted, suffix)
1349 def lookup_unit_table(unit_table, s):
1350 units_re = '|'.join(re.escape(u) for u in unit_table)
1352 r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)' % units_re, s)
1355 num_str = m.group('num').replace(',', '.')
1356 mult = unit_table[m.group('unit')]
1357 return int(float(num_str) * mult)
1360 def parse_filesize(s):
1364 # The lower-case forms are of course incorrect and unofficial,
1365 # but we support those too
1403 return lookup_unit_table(_UNIT_TABLE, s)
1412 if re.match(r'^[\d,.]+$', s):
1413 return str_to_int(s)
1424 return lookup_unit_table(_UNIT_TABLE, s)
1427 def month_by_name(name):
1428 """ Return the number of a month by (locale-independently) English name """
1431 return ENGLISH_MONTH_NAMES.index(name) + 1
1436 def month_by_abbreviation(abbrev):
1437 """ Return the number of a month by (locale-independently) English
1441 return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
1446 def fix_xml_ampersands(xml_str):
1447 """Replace all the '&' by '&' in XML"""
1449 r'&(?!amp;|lt;|gt;|apos;|quot;|#x[0-9a-fA-F]{,4};|#[0-9]{,4};)',
1454 def setproctitle(title):
1455 assert isinstance(title, compat_str)
1457 # ctypes in Jython is not complete
1458 # http://bugs.jython.org/issue2148
1459 if sys.platform.startswith('java'):
1463 libc = ctypes.cdll.LoadLibrary('libc.so.6')
1466 title_bytes = title.encode('utf-8')
1467 buf = ctypes.create_string_buffer(len(title_bytes))
1468 buf.value = title_bytes
1470 libc.prctl(15, buf, 0, 0, 0)
1471 except AttributeError:
1472 return # Strange libc, just skip this
1475 def remove_start(s, start):
1476 if s.startswith(start):
1477 return s[len(start):]
1481 def remove_end(s, end):
1483 return s[:-len(end)]
1487 def remove_quotes(s):
1488 if s is None or len(s) < 2:
1490 for quote in ('"', "'", ):
1491 if s[0] == quote and s[-1] == quote:
1496 def url_basename(url):
1497 path = compat_urlparse.urlparse(url).path
1498 return path.strip('/').split('/')[-1]
1501 class HEADRequest(compat_urllib_request.Request):
1502 def get_method(self):
1506 def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
1509 v = getattr(v, get_attr, None)
1515 return int(v) * invscale // scale
1520 def str_or_none(v, default=None):
1521 return default if v is None else compat_str(v)
1524 def str_to_int(int_str):
1525 """ A more relaxed version of int_or_none """
1528 int_str = re.sub(r'[,\.\+]', '', int_str)
1532 def float_or_none(v, scale=1, invscale=1, default=None):
1536 return float(v) * invscale / scale
1541 def parse_duration(s):
1542 if not isinstance(s, compat_basestring):
1550 (?P<only_mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*|
1551 (?P<only_hours>[0-9.]+)\s*(?:hours?)|
1553 \s*(?P<hours_reversed>[0-9]+)\s*(?:[:h]|hours?)\s*(?P<mins_reversed>[0-9]+)\s*(?:[:m]|mins?\.?|minutes?)\s*|
1556 (?:(?P<days>[0-9]+)\s*(?:[:d]|days?)\s*)?
1557 (?P<hours>[0-9]+)\s*(?:[:h]|hours?)\s*
1559 (?P<mins>[0-9]+)\s*(?:[:m]|mins?|minutes?)\s*
1561 (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*(?:s|secs?|seconds?)?
1566 if m.group('only_mins'):
1567 return float_or_none(m.group('only_mins'), invscale=60)
1568 if m.group('only_hours'):
1569 return float_or_none(m.group('only_hours'), invscale=60 * 60)
1571 res += int(m.group('secs'))
1572 if m.group('mins_reversed'):
1573 res += int(m.group('mins_reversed')) * 60
1575 res += int(m.group('mins')) * 60
1576 if m.group('hours'):
1577 res += int(m.group('hours')) * 60 * 60
1578 if m.group('hours_reversed'):
1579 res += int(m.group('hours_reversed')) * 60 * 60
1581 res += int(m.group('days')) * 24 * 60 * 60
1583 res += float(m.group('ms'))
1587 def prepend_extension(filename, ext, expected_real_ext=None):
1588 name, real_ext = os.path.splitext(filename)
1590 '{0}.{1}{2}'.format(name, ext, real_ext)
1591 if not expected_real_ext or real_ext[1:] == expected_real_ext
1592 else '{0}.{1}'.format(filename, ext))
1595 def replace_extension(filename, ext, expected_real_ext=None):
1596 name, real_ext = os.path.splitext(filename)
1597 return '{0}.{1}'.format(
1598 name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
1602 def check_executable(exe, args=[]):
1603 """ Checks if the given binary is installed somewhere in PATH, and returns its name.
1604 args can be a list of arguments for a short output (like -version) """
1606 subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
1612 def get_exe_version(exe, args=['--version'],
1613 version_re=None, unrecognized='present'):
1614 """ Returns the version of the specified executable,
1615 or False if the executable is not present """
1617 out, _ = subprocess.Popen(
1618 [encodeArgument(exe)] + args,
1619 stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
1622 if isinstance(out, bytes): # Python 2.x
1623 out = out.decode('ascii', 'ignore')
1624 return detect_exe_version(out, version_re, unrecognized)
1627 def detect_exe_version(output, version_re=None, unrecognized='present'):
1628 assert isinstance(output, compat_str)
1629 if version_re is None:
1630 version_re = r'version\s+([-0-9._a-zA-Z]+)'
1631 m = re.search(version_re, output)
1638 class PagedList(object):
1640 # This is only useful for tests
1641 return len(self.getslice())
1644 class OnDemandPagedList(PagedList):
1645 def __init__(self, pagefunc, pagesize, use_cache=False):
1646 self._pagefunc = pagefunc
1647 self._pagesize = pagesize
1648 self._use_cache = use_cache
1652 def getslice(self, start=0, end=None):
1654 for pagenum in itertools.count(start // self._pagesize):
1655 firstid = pagenum * self._pagesize
1656 nextfirstid = pagenum * self._pagesize + self._pagesize
1657 if start >= nextfirstid:
1662 page_results = self._cache.get(pagenum)
1663 if page_results is None:
1664 page_results = list(self._pagefunc(pagenum))
1666 self._cache[pagenum] = page_results
1669 start % self._pagesize
1670 if firstid <= start < nextfirstid
1674 ((end - 1) % self._pagesize) + 1
1675 if (end is not None and firstid <= end <= nextfirstid)
1678 if startv != 0 or endv is not None:
1679 page_results = page_results[startv:endv]
1680 res.extend(page_results)
1682 # A little optimization - if current page is not "full", ie. does
1683 # not contain page_size videos then we can assume that this page
1684 # is the last one - there are no more ids on further pages -
1685 # i.e. no need to query again.
1686 if len(page_results) + startv < self._pagesize:
1689 # If we got the whole page, but the next page is not interesting,
1690 # break out early as well
1691 if end == nextfirstid:
1696 class InAdvancePagedList(PagedList):
1697 def __init__(self, pagefunc, pagecount, pagesize):
1698 self._pagefunc = pagefunc
1699 self._pagecount = pagecount
1700 self._pagesize = pagesize
1702 def getslice(self, start=0, end=None):
1704 start_page = start // self._pagesize
1706 self._pagecount if end is None else (end // self._pagesize + 1))
1707 skip_elems = start - start_page * self._pagesize
1708 only_more = None if end is None else end - start
1709 for pagenum in range(start_page, end_page):
1710 page = list(self._pagefunc(pagenum))
1712 page = page[skip_elems:]
1714 if only_more is not None:
1715 if len(page) < only_more:
1716 only_more -= len(page)
1718 page = page[:only_more]
1725 def uppercase_escape(s):
1726 unicode_escape = codecs.getdecoder('unicode_escape')
1728 r'\\U[0-9a-fA-F]{8}',
1729 lambda m: unicode_escape(m.group(0))[0],
1733 def lowercase_escape(s):
1734 unicode_escape = codecs.getdecoder('unicode_escape')
1736 r'\\u[0-9a-fA-F]{4}',
1737 lambda m: unicode_escape(m.group(0))[0],
1741 def escape_rfc3986(s):
1742 """Escape non-ASCII characters as suggested by RFC 3986"""
1743 if sys.version_info < (3, 0) and isinstance(s, compat_str):
1744 s = s.encode('utf-8')
1745 return compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")
1748 def escape_url(url):
1749 """Escape URL as suggested by RFC 3986"""
1750 url_parsed = compat_urllib_parse_urlparse(url)
1751 return url_parsed._replace(
1752 path=escape_rfc3986(url_parsed.path),
1753 params=escape_rfc3986(url_parsed.params),
1754 query=escape_rfc3986(url_parsed.query),
1755 fragment=escape_rfc3986(url_parsed.fragment)
1759 struct.pack('!I', 0)
1761 # In Python 2.6 (and some 2.7 versions), struct requires a bytes argument
1762 def struct_pack(spec, *args):
1763 if isinstance(spec, compat_str):
1764 spec = spec.encode('ascii')
1765 return struct.pack(spec, *args)
1767 def struct_unpack(spec, *args):
1768 if isinstance(spec, compat_str):
1769 spec = spec.encode('ascii')
1770 return struct.unpack(spec, *args)
1772 struct_pack = struct.pack
1773 struct_unpack = struct.unpack
1776 def read_batch_urls(batch_fd):
1778 if not isinstance(url, compat_str):
1779 url = url.decode('utf-8', 'replace')
1780 BOM_UTF8 = '\xef\xbb\xbf'
1781 if url.startswith(BOM_UTF8):
1782 url = url[len(BOM_UTF8):]
1784 if url.startswith(('#', ';', ']')):
1788 with contextlib.closing(batch_fd) as fd:
1789 return [url for url in map(fixup, fd) if url]
1792 def urlencode_postdata(*args, **kargs):
1793 return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
1796 def update_url_query(url, query):
1797 parsed_url = compat_urlparse.urlparse(url)
1798 qs = compat_parse_qs(parsed_url.query)
1800 qs = encode_dict(qs)
1801 return compat_urlparse.urlunparse(parsed_url._replace(
1802 query=compat_urllib_parse.urlencode(qs, True)))
1805 def encode_dict(d, encoding='utf-8'):
1807 return v.encode(encoding) if isinstance(v, compat_basestring) else v
1808 return dict((encode(k), encode(v)) for k, v in d.items())
1811 def dict_get(d, key_or_keys, default=None, skip_false_values=True):
1812 if isinstance(key_or_keys, (list, tuple)):
1813 for key in key_or_keys:
1814 if key not in d or d[key] is None or skip_false_values and not d[key]:
1818 return d.get(key_or_keys, default)
1821 def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
1822 return string if isinstance(string, compat_str) else compat_str(string, encoding, errors)
1834 def parse_age_limit(s):
1837 m = re.match(r'^(?P<age>\d{1,2})\+?$', s)
1838 return int(m.group('age')) if m else US_RATINGS.get(s)
1841 def strip_jsonp(code):
1843 r'(?s)^[a-zA-Z0-9_.]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
1846 def js_to_json(code):
1849 if v in ('true', 'false', 'null'):
1851 if v.startswith('"'):
1852 v = re.sub(r"\\'", "'", v[1:-1])
1853 elif v.startswith("'"):
1855 v = re.sub(r"\\\\|\\'|\"", lambda m: {
1862 res = re.sub(r'''(?x)
1863 "(?:[^"\\]*(?:\\\\|\\['"nu]))*[^"\\]*"|
1864 '(?:[^'\\]*(?:\\\\|\\['"nu]))*[^'\\]*'|
1865 [a-zA-Z_][.a-zA-Z_0-9]*
1867 res = re.sub(r',(\s*[\]}])', lambda m: m.group(1), res)
1871 def qualities(quality_ids):
1872 """ Get a numeric quality value out of a list of possible values """
1875 return quality_ids.index(qid)
1881 DEFAULT_OUTTMPL = '%(title)s-%(id)s.%(ext)s'
1884 def limit_length(s, length):
1885 """ Add ellipses to overly long strings """
1890 return s[:length - len(ELLIPSES)] + ELLIPSES
1894 def version_tuple(v):
1895 return tuple(int(e) for e in re.split(r'[-.]', v))
1898 def is_outdated_version(version, limit, assume_new=True):
1900 return not assume_new
1902 return version_tuple(version) < version_tuple(limit)
1904 return not assume_new
1907 def ytdl_is_updateable():
1908 """ Returns if youtube-dl can be updated with -U """
1909 from zipimport import zipimporter
1911 return isinstance(globals().get('__loader__'), zipimporter) or hasattr(sys, 'frozen')
1914 def args_to_str(args):
1915 # Get a short string representation for a subprocess command
1916 return ' '.join(shlex_quote(a) for a in args)
1919 def error_to_compat_str(err):
1921 # On python 2 error byte string must be decoded with proper
1922 # encoding rather than ascii
1923 if sys.version_info[0] < 3:
1924 err_str = err_str.decode(preferredencoding())
1928 def mimetype2ext(mt):
1935 _, _, res = mt.rpartition('/')
1939 'smptett+xml': 'tt',
1945 'x-mp4-fragmented': 'mp4',
1950 def urlhandle_detect_ext(url_handle):
1953 getheader = lambda h: url_handle.headers[h]
1954 except AttributeError: # Python < 3
1955 getheader = url_handle.info().getheader
1957 cd = getheader('Content-Disposition')
1959 m = re.match(r'attachment;\s*filename="(?P<filename>[^"]+)"', cd)
1961 e = determine_ext(m.group('filename'), default_ext=None)
1965 return mimetype2ext(getheader('Content-Type'))
1968 def encode_data_uri(data, mime_type):
1969 return 'data:%s;base64,%s' % (mime_type, base64.b64encode(data).decode('ascii'))
1972 def age_restricted(content_limit, age_limit):
1973 """ Returns True iff the content should be blocked """
1975 if age_limit is None: # No limit set
1977 if content_limit is None:
1978 return False # Content available for everyone
1979 return age_limit < content_limit
1982 def is_html(first_bytes):
1983 """ Detect whether a file contains HTML by examining its first bytes. """
1986 (b'\xef\xbb\xbf', 'utf-8'),
1987 (b'\x00\x00\xfe\xff', 'utf-32-be'),
1988 (b'\xff\xfe\x00\x00', 'utf-32-le'),
1989 (b'\xff\xfe', 'utf-16-le'),
1990 (b'\xfe\xff', 'utf-16-be'),
1992 for bom, enc in BOMS:
1993 if first_bytes.startswith(bom):
1994 s = first_bytes[len(bom):].decode(enc, 'replace')
1997 s = first_bytes.decode('utf-8', 'replace')
1999 return re.match(r'^\s*<', s)
2002 def determine_protocol(info_dict):
2003 protocol = info_dict.get('protocol')
2004 if protocol is not None:
2007 url = info_dict['url']
2008 if url.startswith('rtmp'):
2010 elif url.startswith('mms'):
2012 elif url.startswith('rtsp'):
2015 ext = determine_ext(url)
2021 return compat_urllib_parse_urlparse(url).scheme
2024 def render_table(header_row, data):
2025 """ Render a list of rows, each as a list of values """
2026 table = [header_row] + data
2027 max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
2028 format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
2029 return '\n'.join(format_str % tuple(row) for row in table)
2032 def _match_one(filter_part, dct):
2033 COMPARISON_OPERATORS = {
2041 operator_rex = re.compile(r'''(?x)\s*
2043 \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
2045 (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
2046 (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
2049 ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
2050 m = operator_rex.search(filter_part)
2052 op = COMPARISON_OPERATORS[m.group('op')]
2053 if m.group('strval') is not None:
2054 if m.group('op') not in ('=', '!='):
2056 'Operator %s does not support string values!' % m.group('op'))
2057 comparison_value = m.group('strval')
2060 comparison_value = int(m.group('intval'))
2062 comparison_value = parse_filesize(m.group('intval'))
2063 if comparison_value is None:
2064 comparison_value = parse_filesize(m.group('intval') + 'B')
2065 if comparison_value is None:
2067 'Invalid integer value %r in filter part %r' % (
2068 m.group('intval'), filter_part))
2069 actual_value = dct.get(m.group('key'))
2070 if actual_value is None:
2071 return m.group('none_inclusive')
2072 return op(actual_value, comparison_value)
2075 '': lambda v: v is not None,
2076 '!': lambda v: v is None,
2078 operator_rex = re.compile(r'''(?x)\s*
2079 (?P<op>%s)\s*(?P<key>[a-z_]+)
2081 ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
2082 m = operator_rex.search(filter_part)
2084 op = UNARY_OPERATORS[m.group('op')]
2085 actual_value = dct.get(m.group('key'))
2086 return op(actual_value)
2088 raise ValueError('Invalid filter part %r' % filter_part)
2091 def match_str(filter_str, dct):
2092 """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
2095 _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
2098 def match_filter_func(filter_str):
2099 def _match_func(info_dict):
2100 if match_str(filter_str, info_dict):
2103 video_title = info_dict.get('title', info_dict.get('id', 'video'))
2104 return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
2108 def parse_dfxp_time_expr(time_expr):
2112 mobj = re.match(r'^(?P<time_offset>\d+(?:\.\d+)?)s?$', time_expr)
2114 return float(mobj.group('time_offset'))
2116 mobj = re.match(r'^(\d+):(\d\d):(\d\d(?:(?:\.|:)\d+)?)$', time_expr)
2118 return 3600 * int(mobj.group(1)) + 60 * int(mobj.group(2)) + float(mobj.group(3).replace(':', '.'))
2121 def srt_subtitles_timecode(seconds):
2122 return '%02d:%02d:%02d,%03d' % (seconds / 3600, (seconds % 3600) / 60, seconds % 60, (seconds % 1) * 1000)
2125 def dfxp2srt(dfxp_data):
2126 _x = functools.partial(xpath_with_ns, ns_map={
2127 'ttml': 'http://www.w3.org/ns/ttml',
2128 'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
2131 class TTMLPElementParser(object):
2134 def start(self, tag, attrib):
2135 if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
2141 def data(self, data):
2145 return self.out.strip()
2147 def parse_node(node):
2148 target = TTMLPElementParser()
2149 parser = xml.etree.ElementTree.XMLParser(target=target)
2150 parser.feed(xml.etree.ElementTree.tostring(node))
2151 return parser.close()
2153 dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
2155 paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall('.//p')
2158 raise ValueError('Invalid dfxp/TTML subtitle')
2160 for para, index in zip(paras, itertools.count(1)):
2161 begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
2162 end_time = parse_dfxp_time_expr(para.attrib.get('end'))
2163 dur = parse_dfxp_time_expr(para.attrib.get('dur'))
2164 if begin_time is None:
2169 end_time = begin_time + dur
2170 out.append('%d\n%s --> %s\n%s\n\n' % (
2172 srt_subtitles_timecode(begin_time),
2173 srt_subtitles_timecode(end_time),
2179 def cli_option(params, command_option, param):
2180 param = params.get(param)
2181 return [command_option, param] if param is not None else []
2184 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
2185 param = params.get(param)
2186 assert isinstance(param, bool)
2188 return [command_option + separator + (true_value if param else false_value)]
2189 return [command_option, true_value if param else false_value]
2192 def cli_valueless_option(params, command_option, param, expected_value=True):
2193 param = params.get(param)
2194 return [command_option] if param == expected_value else []
2197 def cli_configuration_args(params, param, default=[]):
2198 ex_args = params.get(param)
2201 assert isinstance(ex_args, list)
2205 class ISO639Utils(object):
2206 # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
2395 def short2long(cls, code):
2396 """Convert language code from ISO 639-1 to ISO 639-2/T"""
2397 return cls._lang_map.get(code[:2])
2400 def long2short(cls, code):
2401 """Convert language code from ISO 639-2/T to ISO 639-1"""
2402 for short_name, long_name in cls._lang_map.items():
2403 if long_name == code:
2407 class ISO3166Utils(object):
2408 # From http://data.okfn.org/data/core/country-list
2410 'AF': 'Afghanistan',
2411 'AX': 'Ã…land Islands',
2414 'AS': 'American Samoa',
2419 'AG': 'Antigua and Barbuda',
2436 'BO': 'Bolivia, Plurinational State of',
2437 'BQ': 'Bonaire, Sint Eustatius and Saba',
2438 'BA': 'Bosnia and Herzegovina',
2440 'BV': 'Bouvet Island',
2442 'IO': 'British Indian Ocean Territory',
2443 'BN': 'Brunei Darussalam',
2445 'BF': 'Burkina Faso',
2451 'KY': 'Cayman Islands',
2452 'CF': 'Central African Republic',
2456 'CX': 'Christmas Island',
2457 'CC': 'Cocos (Keeling) Islands',
2461 'CD': 'Congo, the Democratic Republic of the',
2462 'CK': 'Cook Islands',
2464 'CI': 'Côte d\'Ivoire',
2469 'CZ': 'Czech Republic',
2473 'DO': 'Dominican Republic',
2476 'SV': 'El Salvador',
2477 'GQ': 'Equatorial Guinea',
2481 'FK': 'Falkland Islands (Malvinas)',
2482 'FO': 'Faroe Islands',
2486 'GF': 'French Guiana',
2487 'PF': 'French Polynesia',
2488 'TF': 'French Southern Territories',
2503 'GW': 'Guinea-Bissau',
2506 'HM': 'Heard Island and McDonald Islands',
2507 'VA': 'Holy See (Vatican City State)',
2514 'IR': 'Iran, Islamic Republic of',
2517 'IM': 'Isle of Man',
2527 'KP': 'Korea, Democratic People\'s Republic of',
2528 'KR': 'Korea, Republic of',
2531 'LA': 'Lao People\'s Democratic Republic',
2537 'LI': 'Liechtenstein',
2541 'MK': 'Macedonia, the Former Yugoslav Republic of',
2548 'MH': 'Marshall Islands',
2554 'FM': 'Micronesia, Federated States of',
2555 'MD': 'Moldova, Republic of',
2566 'NL': 'Netherlands',
2567 'NC': 'New Caledonia',
2568 'NZ': 'New Zealand',
2573 'NF': 'Norfolk Island',
2574 'MP': 'Northern Mariana Islands',
2579 'PS': 'Palestine, State of',
2581 'PG': 'Papua New Guinea',
2584 'PH': 'Philippines',
2588 'PR': 'Puerto Rico',
2592 'RU': 'Russian Federation',
2594 'BL': 'Saint Barthélemy',
2595 'SH': 'Saint Helena, Ascension and Tristan da Cunha',
2596 'KN': 'Saint Kitts and Nevis',
2597 'LC': 'Saint Lucia',
2598 'MF': 'Saint Martin (French part)',
2599 'PM': 'Saint Pierre and Miquelon',
2600 'VC': 'Saint Vincent and the Grenadines',
2603 'ST': 'Sao Tome and Principe',
2604 'SA': 'Saudi Arabia',
2608 'SL': 'Sierra Leone',
2610 'SX': 'Sint Maarten (Dutch part)',
2613 'SB': 'Solomon Islands',
2615 'ZA': 'South Africa',
2616 'GS': 'South Georgia and the South Sandwich Islands',
2617 'SS': 'South Sudan',
2622 'SJ': 'Svalbard and Jan Mayen',
2625 'CH': 'Switzerland',
2626 'SY': 'Syrian Arab Republic',
2627 'TW': 'Taiwan, Province of China',
2629 'TZ': 'Tanzania, United Republic of',
2631 'TL': 'Timor-Leste',
2635 'TT': 'Trinidad and Tobago',
2638 'TM': 'Turkmenistan',
2639 'TC': 'Turks and Caicos Islands',
2643 'AE': 'United Arab Emirates',
2644 'GB': 'United Kingdom',
2645 'US': 'United States',
2646 'UM': 'United States Minor Outlying Islands',
2650 'VE': 'Venezuela, Bolivarian Republic of',
2652 'VG': 'Virgin Islands, British',
2653 'VI': 'Virgin Islands, U.S.',
2654 'WF': 'Wallis and Futuna',
2655 'EH': 'Western Sahara',
2662 def short2full(cls, code):
2663 """Convert an ISO 3166-2 country code to the corresponding full name"""
2664 return cls._country_map.get(code.upper())
2667 class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
2668 def __init__(self, proxies=None):
2669 # Set default handlers
2670 for type in ('http', 'https'):
2671 setattr(self, '%s_open' % type,
2672 lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
2673 meth(r, proxy, type))
2674 return compat_urllib_request.ProxyHandler.__init__(self, proxies)
2676 def proxy_open(self, req, proxy, type):
2677 req_proxy = req.headers.get('Ytdl-request-proxy')
2678 if req_proxy is not None:
2680 del req.headers['Ytdl-request-proxy']
2682 if proxy == '__noproxy__':
2683 return None # No Proxy
2684 return compat_urllib_request.ProxyHandler.proxy_open(
2685 self, req, proxy, type)
2688 def ohdave_rsa_encrypt(data, exponent, modulus):
2690 Implement OHDave's RSA algorithm. See http://www.ohdave.com/rsa/
2693 data: data to encrypt, bytes-like object
2694 exponent, modulus: parameter e and N of RSA algorithm, both integer
2695 Output: hex string of encrypted data
2697 Limitation: supports one block encryption only
2700 payload = int(binascii.hexlify(data[::-1]), 16)
2701 encrypted = pow(payload, exponent, modulus)
2702 return '%x' % encrypted
2705 def encode_base_n(num, n, table=None):
2706 FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
2708 table = FULL_TABLE[:n]
2711 raise ValueError('base %d exceeds table length %d' % (n, len(table)))
2718 ret = table[num % n] + ret
2723 def decode_packed_codes(code):
2725 r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)",
2727 obfucasted_code, base, count, symbols = mobj.groups()
2730 symbols = symbols.split('|')
2735 base_n_count = encode_base_n(count, base)
2736 symbol_table[base_n_count] = symbols[count] or base_n_count
2739 r'\b(\w+)\b', lambda mobj: symbol_table[mobj.group(0)],