Disclaimer: this code is hosted for legal uses only, such as downloading content you own the rights to but no longer have a local copy of, downloading content under a permissive license, educational use, etc. Also RIAA can **** right off.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

4014 lines
123 KiB

#!/usr/bin/env python
# coding: utf-8
from __future__ import unicode_literals
import base64
import binascii
import calendar
import codecs
import contextlib
import ctypes
import datetime
import email.utils
import email.header
import errno
import functools
import gzip
import io
import itertools
import json
import locale
import math
import operator
import os
import platform
import random
import re
import socket
import ssl
import subprocess
import sys
import tempfile
import traceback
import xml.etree.ElementTree
import zlib
from .compat import (
from .socks import (
def register_socks_protocols():
# "Register" SOCKS protocols
# In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
# URLs with protocols not in urlparse.uses_netloc are not handled correctly
for scheme in ('socks', 'socks4', 'socks4a', 'socks5'):
if scheme not in compat_urlparse.uses_netloc:
# This is not clearly defined otherwise
compiled_regex_type = type(re.compile(''))
std_headers = {
'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:59.0) Gecko/20100101 Firefox/59.0',
'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
'Accept-Encoding': 'gzip, deflate',
'Accept-Language': 'en-us,en;q=0.5',
'Safari': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 Safari/533.20.27',
NO_DEFAULT = object()
'January', 'February', 'March', 'April', 'May', 'June',
'July', 'August', 'September', 'October', 'November', 'December']
'fr': [
'janvier', 'février', 'mars', 'avril', 'mai', 'juin',
'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
'flv', 'f4v', 'f4a', 'f4b',
'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
'mkv', 'mka', 'mk3d',
'avi', 'divx',
'asf', 'wmv', 'wma',
'3gp', '3g2',
'f4f', 'f4m', 'm3u8', 'smil')
# needed for sanitizing filenames in restricted mode
ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
'%d %B %Y',
'%d %b %Y',
'%B %d %Y',
'%B %dst %Y',
'%B %dnd %Y',
'%B %dth %Y',
'%b %d %Y',
'%b %dst %Y',
'%b %dnd %Y',
'%b %dth %Y',
'%b %dst %Y %I:%M',
'%b %dnd %Y %I:%M',
'%b %dth %Y %I:%M',
'%Y %m %d',
'%Y/%m/%d %H:%M',
'%Y/%m/%d %H:%M:%S',
'%Y-%m-%d %H:%M',
'%Y-%m-%d %H:%M:%S',
'%Y-%m-%d %H:%M:%S.%f',
'%d.%m.%Y %H:%M',
'%d.%m.%Y %H.%M',
'%b %d %Y at %H:%M',
'%b %d %Y at %H:%M:%S',
'%B %d %Y at %H:%M',
'%B %d %Y at %H:%M:%S',
'%d/%m/%Y %H:%M:%S',
'%m/%d/%Y %H:%M:%S',
PACKED_CODES_RE = r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
JSON_LD_RE = r'(?is)<script[^>]+type=(["\']?)application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>'
def preferredencoding():
"""Get preferred encoding.
Returns the best encoding scheme for the system, based on
locale.getpreferredencoding() and some further tweaks.
pref = locale.getpreferredencoding()
except Exception:
pref = 'UTF-8'
return pref
def write_json_file(obj, fn):
""" Encode obj as JSON and write it to fn, atomically if possible """
fn = encodeFilename(fn)
if sys.version_info < (3, 0) and sys.platform != 'win32':
encoding = get_filesystem_encoding()
# os.path.basename returns a bytes object, but NamedTemporaryFile
# will fail if the filename contains non ascii characters unless we
# use a unicode object
path_basename = lambda f: os.path.basename(fn).decode(encoding)
# the same for os.path.dirname
path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
path_basename = os.path.basename
path_dirname = os.path.dirname
args = {
'suffix': '.tmp',
'prefix': path_basename(fn) + '.',
'dir': path_dirname(fn),
'delete': False,
# In Python 2.x, json.dump expects a bytestream.
# In Python 3.x, it writes to a character stream
if sys.version_info < (3, 0):
args['mode'] = 'wb'
'mode': 'w',
'encoding': 'utf-8',
tf = tempfile.NamedTemporaryFile(**compat_kwargs(args))
with tf:
json.dump(obj, tf)
if sys.platform == 'win32':
# Need to remove existing file on Windows, else os.rename raises
# WindowsError or FileExistsError.
except OSError:
os.rename(tf.name, fn)
except Exception:
except OSError:
if sys.version_info >= (2, 7):
def find_xpath_attr(node, xpath, key, val=None):
""" Find the xpath xpath[@key=val] """
assert re.match(r'^[a-zA-Z_-]+$', key)
expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))
return node.find(expr)
def find_xpath_attr(node, xpath, key, val=None):
for f in node.findall(compat_xpath(xpath)):
if key not in f.attrib:
if val is None or f.attrib.get(key) == val:
return f
return None
# On python2.6 the xml.etree.ElementTree.Element methods don't support
# the namespace parameter
def xpath_with_ns(path, ns_map):
components = [c.split(':') for c in path.split('/')]
replaced = []
for c in components:
if len(c) == 1:
ns, tag = c
replaced.append('{%s}%s' % (ns_map[ns], tag))
return '/'.join(replaced)
def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
def _find_xpath(xpath):
return node.find(compat_xpath(xpath))
if isinstance(xpath, (str, compat_str)):
n = _find_xpath(xpath)
for xp in xpath:
n = _find_xpath(xp)
if n is not None:
if n is None:
if default is not NO_DEFAULT:
return default
elif fatal:
name = xpath if name is None else name
raise ExtractorError('Could not find XML element %s' % name)
return None
return n
def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
n = xpath_element(node, xpath, name, fatal=fatal, default=default)
if n is None or n == default:
return n
if n.text is None:
if default is not NO_DEFAULT:
return default
elif fatal:
name = xpath if name is None else name
raise ExtractorError('Could not find XML element\'s text %s' % name)
return None
return n.text
def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
n = find_xpath_attr(node, xpath, key)
if n is None:
if default is not NO_DEFAULT:
return default
elif fatal:
name = '%s[@%s]' % (xpath, key) if name is None else name
raise ExtractorError('Could not find XML attribute %s' % name)
return None
return n.attrib[key]
def get_element_by_id(id, html):
"""Return the content of the tag with the specified ID in the passed HTML document"""
return get_element_by_attribute('id', id, html)
def get_element_by_class(class_name, html):
"""Return the content of the first tag with the specified class in the passed HTML document"""
retval = get_elements_by_class(class_name, html)
return retval[0] if retval else None
def get_element_by_attribute(attribute, value, html, escape_value=True):
retval = get_elements_by_attribute(attribute, value, html, escape_value)
return retval[0] if retval else None
def get_elements_by_class(class_name, html):
"""Return the content of all tags with the specified class in the passed HTML document as a list"""
return get_elements_by_attribute(
'class', r'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
html, escape_value=False)
def get_elements_by_attribute(attribute, value, html, escape_value=True):
"""Return the content of the tag with the specified attribute in the passed HTML document"""
value = re.escape(value) if escape_value else value
retlist = []
for m in re.finditer(r'''(?xs)
''' % (re.escape(attribute), value), html):
res = m.group('content')
if res.startswith('"') or res.startswith("'"):
res = res[1:-1]
return retlist
class HTMLAttributeParser(compat_HTMLParser):
"""Trivial HTML parser to gather the attributes for a single element"""
def __init__(self):
self.attrs = {}
def handle_starttag(self, tag, attrs):
self.attrs = dict(attrs)
def extract_attributes(html_element):
"""Given a string for an HTML element such as
a="foo" B="bar" c="&98;az" d=boz
empty= noval entity="&amp;"
sq='"' dq="'"
Decode and return a dictionary of attributes.
'a': 'foo', 'b': 'bar', c: 'baz', d: 'boz',
'empty': '', 'noval': None, 'entity': '&',
'sq': '"', 'dq': '\''
NB HTMLParser is stricter in Python 2.6 & 3.2 than in later versions,
but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
parser = HTMLAttributeParser()
# Older Python may throw HTMLParseError in case of malformed HTML
except compat_HTMLParseError:
return parser.attrs
def clean_html(html):
"""Clean an HTML snippet into a readable string"""
if html is None: # Convenience for sanitizing descriptions etc.
return html
# Newline vs <br />
html = html.replace('\n', ' ')
html = re.sub(r'(?u)\s*<\s*br\s*/?\s*>\s*', '\n', html)
html = re.sub(r'(?u)<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
# Strip html tags
html = re.sub('<.*?>', '', html)
# Replace html entities
html = unescapeHTML(html)
return html.strip()
def sanitize_open(filename, open_mode):
"""Try to open the given filename, and slightly tweak it if this fails.
Attempts to open the given filename. If this fails, it tries to change
the filename slightly, step by step, until it's either able to open it
or it fails and raises a final exception, like the standard open()
It returns the tuple (stream, definitive_file_name).
if filename == '-':
if sys.platform == 'win32':
import msvcrt
msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
stream = open(encodeFilename(filename), open_mode)
return (stream, filename)
except (IOError, OSError) as err:
if err.errno in (errno.EACCES,):
# In case of error, try to remove win32 forbidden chars
alt_filename = sanitize_path(filename)
if alt_filename == filename:
# An exception here should be caught in the caller
stream = open(encodeFilename(alt_filename), open_mode)
return (stream, alt_filename)
def timeconvert(timestr):
"""Convert RFC 2822 defined time string into system timestamp"""
timestamp = None
timetuple = email.utils.parsedate_tz(timestr)
if timetuple is not None:
timestamp = email.utils.mktime_tz(timetuple)
return timestamp
def sanitize_filename(s, restricted=False, is_id=False):
"""Sanitizes a string so it could be used as part of a filename.
If restricted is set, use a stricter subset of allowed characters.
Set is_id if this is not an arbitrary string, but an ID that should be kept
if possible.
def replace_insane(char):
if restricted and char in ACCENT_CHARS:
return ACCENT_CHARS[char]
if char == '?' or ord(char) < 32 or ord(char) == 127:
return ''
elif char == '"':
return '' if restricted else '\''
elif char == ':':
return '_-' if restricted else ' -'
elif char in '\\/|*<>':
return '_'
if restricted and (char in '!&\'()[]{}$;`^,#' or char.isspace()):
return '_'
if restricted and ord(char) > 127:
return '_'
return char
# Handle timestamps
s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
result = ''.join(map(replace_insane, s))
if not is_id:
while '__' in result:
result = result.replace('__', '_')
result = result.strip('_')
# Common case of "Foreign band name - English song title"
if restricted and result.startswith('-_'):
result = result[2:]
if result.startswith('-'):
result = '_' + result[len('-'):]
result = result.lstrip('.')
if not result:
result = '_'
return result
def sanitize_path(s):
"""Sanitizes and normalizes path on Windows"""
if sys.platform != 'win32':
return s
drive_or_unc, _ = os.path.splitdrive(s)
if sys.version_info < (2, 7) and not drive_or_unc:
drive_or_unc, _ = os.path.splitunc(s)
norm_path = os.path.normpath(remove_start(s, drive_or_unc)).split(os.path.sep)
if drive_or_unc:
sanitized_path = [
path_part if path_part in ['.', '..'] else re.sub(r'(?:[/<>:"\|\\?\*]|[\s.]$)', '#', path_part)
for path_part in norm_path]
if drive_or_unc:
sanitized_path.insert(0, drive_or_unc + os.path.sep)
return os.path.join(*sanitized_path)
def sanitize_url(url):
# Prepend protocol-less URLs with `http:` scheme in order to mitigate
# the number of unwanted failures due to missing protocol
if url.startswith('//'):
return 'http:%s' % url
# Fix some common typos seen so far
# https://github.com/ytdl-org/youtube-dl/issues/15649
(r'^httpss://', r'https://'),
# https://bx1.be/lives/direct-tv/
(r'^rmtp([es]?)://', r'rtmp\1://'),
for mistake, fixup in COMMON_TYPOS:
if re.match(mistake, url):
return re.sub(mistake, fixup, url)
return url
def sanitized_Request(url, *args, **kwargs):
return compat_urllib_request.Request(sanitize_url(url), *args, **kwargs)
def expand_path(s):
"""Expand shell variables and ~"""
return os.path.expandvars(compat_expanduser(s))
def orderedSet(iterable):
""" Remove all duplicates from the input iterable """
res = []
for el in iterable:
if el not in res:
return res
def _htmlentity_transform(entity_with_semicolon):
"""Transforms an HTML entity to a character."""
entity = entity_with_semicolon[:-1]
# Known non-numeric HTML entity
if entity in compat_html_entities.name2codepoint:
return compat_chr(compat_html_entities.name2codepoint[entity])
# TODO: HTML5 allows entities without a semicolon. For example,
# '&Eacuteric' should be decoded as 'Éric'.
if entity_with_semicolon in compat_html_entities_html5:
return compat_html_entities_html5[entity_with_semicolon]
mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
if mobj is not None:
numstr = mobj.group(1)
if numstr.startswith('x'):
base = 16
numstr = '0%s' % numstr
base = 10
# See https://github.com/ytdl-org/youtube-dl/issues/7518
return compat_chr(int(numstr, base))
except ValueError:
# Unknown entity in name, return its literal representation
return '&%s;' % entity
def unescapeHTML(s):
if s is None:
return None
assert type(s) == compat_str
return re.sub(
r'&([^&;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
def get_subprocess_encoding():
if sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
# For subprocess calls, encode with locale encoding
# Refer to http://stackoverflow.com/a/9951851/35070
encoding = preferredencoding()
encoding = sys.getfilesystemencoding()
if encoding is None:
encoding = 'utf-8'
return encoding
def encodeFilename(s, for_subprocess=False):
@param s The name of the file
assert type(s) == compat_str
# Python 3 has a Unicode API
if sys.version_info >= (3, 0):
return s
# Pass '' directly to use Unicode APIs on Windows 2000 and up
# (Detecting Windows NT 4 is tricky because 'major >= 4' would
# match Windows 9x series as well. Besides, NT 4 is obsolete.)
if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
return s
# Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
if sys.platform.startswith('java'):
return s
return s.encode(get_subprocess_encoding(), 'ignore')
def decodeFilename(b, for_subprocess=False):
if sys.version_info >= (3, 0):
return b
if not isinstance(b, bytes):
return b
return b.decode(get_subprocess_encoding(), 'ignore')
def encodeArgument(s):
if not isinstance(s, compat_str):
# Legacy code that uses byte strings
# Uncomment the following line after fixing all post processors
# assert False, 'Internal error: %r should be of type %r, is %r' % (s, compat_str, type(s))
s = s.decode('ascii')
return encodeFilename(s, True)
def decodeArgument(b):
return decodeFilename(b, True)
def decodeOption(optval):
if optval is None:
return optval
if isinstance(optval, bytes):
optval = optval.decode(preferredencoding())
assert isinstance(optval, compat_str)
return optval
def formatSeconds(secs):
if secs > 3600:
return '%d:%02d:%02d' % (secs // 3600, (secs % 3600) // 60, secs % 60)
elif secs > 60:
return '%d:%02d' % (secs // 60, secs % 60)
return '%d' % secs
def make_HTTPS_handler(params, **kwargs):
opts_no_check_certificate = params.get('nocheckcertificate', False)
if hasattr(ssl, 'create_default_context'): # Python >= 3.4 or 2.7.9
context = ssl.create_default_context(ssl.Purpose.SERVER_AUTH)
if opts_no_check_certificate:
context.check_hostname = False
context.verify_mode = ssl.CERT_NONE
return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
except TypeError:
# Python 2.7.8
# (create_default_context present but HTTPSHandler has no context=)
if sys.version_info < (3, 2):
return YoutubeDLHTTPSHandler(params, **kwargs)
else: # Python < 3.4
context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
context.verify_mode = (ssl.CERT_NONE
if opts_no_check_certificate
return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
def bug_reports_message():
if ytdl_is_updateable():
update_cmd = 'type youtube-dl -U to update'
update_cmd = 'see https://yt-dl.org/update on how to update'
msg = '; please report this issue on https://yt-dl.org/bug .'
msg += ' Make sure you are using the latest version; %s.' % update_cmd
msg += ' Be sure to call youtube-dl with the --verbose flag and include its complete output.'
return msg
class YoutubeDLError(Exception):
"""Base exception for YoutubeDL errors."""
class ExtractorError(YoutubeDLError):
"""Error during info extraction."""
def __init__(self, msg, tb=None, expected=False, cause=None, video_id=None):
""" tb, if given, is the original traceback (so that it can be printed out).
If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
expected = True
if video_id is not None:
msg = video_id + ': ' + msg
if cause:
msg += ' (caused by %r)' % cause
if not expected:
msg += bug_reports_message()
super(ExtractorError, self).__init__(msg)
self.traceback = tb
self.exc_info = sys.exc_info() # preserve original exception
self.cause = cause
self.video_id = video_id
def format_traceback(self):
if self.traceback is None:
return None
return ''.join(traceback.format_tb(self.traceback))
class UnsupportedError(ExtractorError):
def __init__(self, url):
super(UnsupportedError, self).__init__(
'Unsupported URL: %s' % url, expected=True)
self.url = url
class RegexNotFoundError(ExtractorError):
"""Error when a regex didn't match"""
class GeoRestrictedError(ExtractorError):
"""Geographic restriction Error exception.
This exception may be thrown when a video is not available from your
geographic location due to geographic restrictions imposed by a website.
def __init__(self, msg, countries=None):
super(GeoRestrictedError, self).__init__(msg, expected=True)
self.msg = msg
self.countries = countries
class DownloadError(YoutubeDLError):
"""Download Error exception.
This exception may be thrown by FileDownloader objects if they are not
configured to continue on errors. They will contain the appropriate
error message.
def __init__(self, msg, exc_info=None):
""" exc_info, if given, is the original exception that caused the trouble (as returned by sys.exc_info()). """
super(DownloadError, self).__init__(msg)
self.exc_info = exc_info
class SameFileError(YoutubeDLError):
"""Same File exception.
This exception will be thrown by FileDownloader objects if they detect
multiple files would have to be downloaded to the same file on disk.
class PostProcessingError(YoutubeDLError):
"""Post Processing exception.
This exception may be raised by PostProcessor's .run() method to
indicate an error in the postprocessing task.
def __init__(self, msg):
super(PostProcessingError, self).__init__(msg)
self.msg = msg
class MaxDownloadsReached(YoutubeDLError):
""" --max-downloads limit has been reached. """
class UnavailableVideoError(YoutubeDLError):
"""Unavailable Format exception.
This exception will be thrown when a video is requested
in a format that is not available for that video.
class ContentTooShortError(YoutubeDLError):
"""Content Too Short exception.
This exception may be raised by FileDownloader objects when a file they
download is too small for what the server announced first, indicating
the connection was probably interrupted.
def __init__(self, downloaded, expected):
super(ContentTooShortError, self).__init__(
'Downloaded {0} bytes, expected {1} bytes'.format(downloaded, expected)
# Both in bytes
self.downloaded = downloaded
self.expected = expected
class XAttrMetadataError(YoutubeDLError):
def __init__(self, code=None, msg='Unknown error'):
super(XAttrMetadataError, self).__init__(msg)
self.code = code
self.msg = msg
# Parsing code and msg
if (self.code in (errno.ENOSPC, errno.EDQUOT)
or 'No space left' in self.msg or 'Disk quota excedded' in self.msg):
self.reason = 'NO_SPACE'
elif self.code == errno.E2BIG or 'Argument list too long' in self.msg:
self.reason = 'VALUE_TOO_LONG'
self.reason = 'NOT_SUPPORTED'
class XAttrUnavailableError(YoutubeDLError):
def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
# Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
# expected HTTP responses to meet HTTP/1.0 or later (see also
# https://github.com/ytdl-org/youtube-dl/issues/6727)
if sys.version_info < (3, 0):
kwargs['strict'] = True
hc = http_class(*args, **compat_kwargs(kwargs))
source_address = ydl_handler._params.get('source_address')
if source_address is not None:
# This is to workaround _create_connection() from socket where it will try all
# address data from getaddrinfo() including IPv6. This filters the result from
# getaddrinfo() based on the source_address value.
# This is based on the cpython socket.create_connection() function.
# https://github.com/python/cpython/blob/master/Lib/socket.py#L691
def _create_connection(address, timeout=socket._GLOBAL_DEFAULT_TIMEOUT, source_address=None):
host, port = address
err = None
addrs = socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM)
af = socket.AF_INET if '.' in source_address[0] else socket.AF_INET6
ip_addrs = [addr for addr in addrs if addr[0] == af]
if addrs and not ip_addrs:
ip_version = 'v4' if af == socket.AF_INET else 'v6'
raise socket.error(
"No remote IP%s addresses available for connect, can't use '%s' as source address"
% (ip_version, source_address[0]))
for res in ip_addrs:
af, socktype, proto, canonname, sa = res
sock = None
sock = socket.socket(af, socktype, proto)
if timeout is not socket._GLOBAL_DEFAULT_TIMEOUT:
err = None # Explicitly break reference cycle
return sock
except socket.error as _:
err = _
if sock is not None:
if err is not None:
raise err
raise socket.error('getaddrinfo returns an empty list')
if hasattr(hc, '_create_connection'):
hc._create_connection = _create_connection
sa = (source_address, 0)
if hasattr(hc, 'source_address'): # Python 2.7+
hc.source_address = sa
else: # Python 2.6
def _hc_connect(self, *args, **kwargs):
sock = _create_connection(
(self.host, self.port), self.timeout, sa)
if is_https:
self.sock = ssl.wrap_socket(
sock, self.key_file, self.cert_file,
self.sock = sock
hc.connect = functools.partial(_hc_connect, hc)
return hc
def handle_youtubedl_headers(headers):
filtered_headers = headers
if 'Youtubedl-no-compression' in filtered_headers:
filtered_headers = dict((k, v) for k, v in filtered_headers.items() if k.lower() != 'accept-encoding')
del filtered_headers['Youtubedl-no-compression']
return filtered_headers
class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
"""Handler for HTTP requests and responses.
This class, when installed with an OpenerDirector, automatically adds
the standard headers to every HTTP request and handles gzipped and
deflated responses from web servers. If compression is to be avoided in
a particular request, the original request in the program code only has
to include the HTTP header "Youtubedl-no-compression", which will be
removed before making the real request.
Part of this code was copied from:
Andrew Rowls, the author of that code, agreed to release it to the
public domain.
def __init__(self, params, *args, **kwargs):
compat_urllib_request.HTTPHandler.__init__(self, *args, **kwargs)
self._params = params
def http_open(self, req):
conn_class = compat_http_client.HTTPConnection
socks_proxy = req.headers.get('Ytdl-socks-proxy')
if socks_proxy:
conn_class = make_socks_conn_class(conn_class, socks_proxy)
del req.headers['Ytdl-socks-proxy']
return self.do_open(functools.partial(
_create_http_connection, self, conn_class, False),
def deflate(data):
return zlib.decompress(data, -zlib.MAX_WBITS)
except zlib.error:
return zlib.decompress(data)
def http_request(self, req):
# According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
# always respected by websites, some tend to give out URLs with non percent-encoded
# non-ASCII characters (see telemb.py, ard.py [#3412])
# urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
# To work around aforementioned issue we will replace request's original URL with
# percent-encoded one
# Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
# the code of this workaround has been moved here from YoutubeDL.urlopen()
url = req.get_full_url()
url_escaped = escape_url(url)
# Substitute URL if any change after escaping
if url != url_escaped:
req = update_Request(req, url=url_escaped)
for h, v in std_headers.items():
# Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
# The dict keys are capitalized because of this bug by urllib
if h.capitalize() not in req.headers:
req.add_header(h, v)
req.headers = handle_youtubedl_headers(req.headers)
if sys.version_info < (2, 7) and '#' in req.get_full_url():
# Python 2.6 is brain-dead when it comes to fragments
req._Request__original = req._Request__original.partition('#')[0]
req._Request__r_type = req._Request__r_type.partition('#')[0]
return req
def http_response(self, req, resp):
old_resp = resp
# gzip
if resp.headers.get('Content-encoding', '') == 'gzip':
content = resp.read()
gz = gzip.GzipFile(fileobj=io.BytesIO(content), mode='rb')
uncompressed = io.BytesIO(gz.read())
except IOError as original_ioerror:
# There may be junk add the end of the file
# See http://stackoverflow.com/q/4928560/35070 for details
for i in range(1, 1024):
gz = gzip.GzipFile(fileobj=io.BytesIO(content[:-i]), mode='rb')
uncompressed = io.BytesIO(gz.read())
except IOError:
raise original_ioerror
resp = compat_urllib_request.addinfourl(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
resp.msg = old_resp.msg
del resp.headers['Content-encoding']
# deflate
if resp.headers.get('Content-encoding', '') == 'deflate':
gz = io.BytesIO(self.deflate(resp.read()))
resp = compat_urllib_request.addinfourl(gz, old_resp.headers, old_resp.url, old_resp.code)
resp.msg = old_resp.msg
del resp.headers['Content-encoding']
# Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
# https://github.com/ytdl-org/youtube-dl/issues/6457).
if 300 <= resp.code < 400:
location = resp.headers.get('Location')
if location:
# As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
if sys.version_info >= (3, 0):
location = location.encode('iso-8859-1').decode('utf-8')
location = location.decode('utf-8')
location_escaped = escape_url(location)
if location != location_escaped:
del resp.headers['Location']
if sys.version_info < (3, 0):
location_escaped = location_escaped.encode('utf-8')
resp.headers['Location'] = location_escaped
return resp
https_request = http_request
https_response = http_response
def make_socks_conn_class(base_class, socks_proxy):
assert issubclass(base_class, (
compat_http_client.HTTPConnection, compat_http_client.HTTPSConnection))
url_components = compat_urlparse.urlparse(socks_proxy)
if url_components.scheme.lower() == 'socks5':
socks_type = ProxyType.SOCKS5
elif url_components.scheme.lower() in ('socks', 'socks4'):
socks_type = ProxyType.SOCKS4
elif url_components.scheme.lower() == 'socks4a':
socks_type = ProxyType.SOCKS4A
def unquote_if_non_empty(s):
if not s:
return s
return compat_urllib_parse_unquote_plus(s)
proxy_args = (
url_components.hostname, url_components.port or 1080,
True, # Remote DNS
class SocksConnection(base_class):
def connect(self):
self.sock = sockssocket()
if type(self.timeout) in (int, float):
self.sock.connect((self.host, self.port))
if isinstance(self, compat_http_client.HTTPSConnection):
if hasattr(self, '_context'): # Python > 2.6
self.sock = self._context.wrap_socket(
self.sock, server_hostname=self.host)
self.sock = ssl.wrap_socket(self.sock)
return SocksConnection
class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
def __init__(self, params, https_conn_class=None, *args, **kwargs):
compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
self._https_conn_class = https_conn_class or compat_http_client.HTTPSConnection
self._params = params
def https_open(self, req):
kwargs = {}
conn_class = self._https_conn_class
if hasattr(self, '_context'): # python > 2.6
kwargs['context'] = self._context
if hasattr(self, '_check_hostname'): # python 3.x
kwargs['check_hostname'] = self._check_hostname
socks_proxy = req.headers.get('Ytdl-socks-proxy')
if socks_proxy:
conn_class = make_socks_conn_class(conn_class, socks_proxy)
del req.headers['Ytdl-socks-proxy']
return self.do_open(functools.partial(
_create_http_connection, self, conn_class, True),
req, **kwargs)
class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
def save(self, filename=None, ignore_discard=False, ignore_expires=False):
# Store session cookies with `expires` set to 0 instead of an empty
# string
for cookie in self:
if cookie.expires is None:
cookie.expires = 0
compat_cookiejar.MozillaCookieJar.save(self, filename, ignore_discard, ignore_expires)
def load(self, filename=None, ignore_discard=False, ignore_expires=False):
"""Load cookies from a file."""
if filename is None:
if self.filename is not None:
filename = self.filename
raise ValueError(compat_cookiejar.MISSING_FILENAME_TEXT)
cf = io.StringIO()
with open(filename) as f:
for line in f:
if line.startswith(self._HTTPONLY_PREFIX):
line = line[len(self._HTTPONLY_PREFIX):]
self._really_load(cf, filename, ignore_discard, ignore_expires)
# Session cookies are denoted by either `expires` field set to
# an empty string or 0. MozillaCookieJar only recognizes the former
# (see [1]). So we need force the latter to be recognized as session
# cookies on our own.
# Session cookies may be important for cookies-based authentication,
# e.g. usually, when user does not check 'Remember me' check box while
# logging in on a site, some important cookies are stored as session
# cookies so that not recognizing them will result in failed login.
# 1. https://bugs.python.org/issue17164
for cookie in self:
# Treat `expires=0` cookies as session cookies
if cookie.expires == 0:
cookie.expires = None
cookie.discard = True
class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
def __init__(self, cookiejar=None):
compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
def http_response(self, request, response):
# Python 2 will choke on next HTTP request in row if there are non-ASCII
# characters in Set-Cookie HTTP header of last response (see
# https://github.com/ytdl-org/youtube-dl/issues/6769).
# In order to at least prevent crashing we will percent encode Set-Cookie
# header before HTTPCookieProcessor starts processing it.
# if sys.version_info < (3, 0) and response.headers:
# for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
# set_cookie = response.headers.get(set_cookie_header)
# if set_cookie:
# set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
# if set_cookie != set_cookie_escaped:
# del response.headers[set_cookie_header]
# response.headers[set_cookie_header] = set_cookie_escaped
return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
https_request = compat_urllib_request.HTTPCookieProcessor.http_request
https_response = http_response
def extract_timezone(date_str):
m = re.search(
r'^.{8,}?(?P<tz>Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
if not m:
timezone = datetime.timedelta()
date_str = date_str[:-len(m.group('tz'))]
if not m.group('sign'):
timezone = datetime.timedelta()
sign = 1 if m.group('sign') == '+' else -1
timezone = datetime.timedelta(
hours=sign * int(m.group('hours')),
minutes=sign * int(m.group('minutes')))
return timezone, date_str
def parse_iso8601(date_str, delimiter='T', timezone=None):
""" Return a UNIX timestamp from the given date """
if date_str is None:
return None
date_str = re.sub(r'\.[0-9]+', '', date_str)
if timezone is None:
timezone, date_str = extract_timezone(date_str)
date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
dt = datetime.datetime.strptime(date_str, date_format) - timezone
return calendar.timegm(dt.timetuple())
except ValueError:
def date_formats(day_first=True):
def unified_strdate(date_str, day_first=True):
"""Return a string with the date in the format YYYYMMDD"""
if date_str is None:
return None
upload_date = None
# Replace commas
date_str = date_str.replace(',', ' ')
# Remove AM/PM + timezone
date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
_, date_str = extract_timezone(date_str)
for expression in date_formats(day_first):
upload_date = datetime.datetime.strptime(date_str, expression).strftime('%Y%m%d')
except ValueError:
if upload_date is None:
timetuple = email.utils.parsedate_tz(date_str)
if timetuple:
upload_date = datetime.datetime(*timetuple[:6]).strftime('%Y%m%d')
except ValueError:
if upload_date is not None:
return compat_str(upload_date)
def unified_timestamp(date_str, day_first=True):
if date_str is None:
return None
date_str = re.sub(r'[,|]', '', date_str)
pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0
timezone, date_str = extract_timezone(date_str)
# Remove AM/PM + timezone
date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
# Remove unrecognized timezones from ISO 8601 alike timestamps
m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str)
if m:
date_str = date_str[:-len(m.group('tz'))]
# Python only supports microseconds, so remove nanoseconds
m = re.search(r'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str)
if m:
date_str = m.group(1)
for expression in date_formats(day_first):
dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
return calendar.timegm(dt.timetuple())
except ValueError:
timetuple = email.utils.parsedate_tz(date_str)
if timetuple:
return calendar.timegm(timetuple) + pm_delta * 3600
def determine_ext(url, default_ext='unknown_video'):
if url is None or '.' not in url:
return default_ext
guess = url.partition('?')[0].rpartition('.')[2]
if re.match(r'^[A-Za-z0-9]+$', guess):
return guess
# Try extract ext from URLs like http://example.com/foo/bar.mp4/?download
elif guess.rstrip('/') in KNOWN_EXTENSIONS:
return guess.rstrip('/')
return default_ext
def subtitles_filename(filename, sub_lang, sub_format):
return filename.rsplit('.', 1)[0] + '.' + sub_lang + '.' + sub_format
def date_from_str(date_str):
Return a datetime object from a string in the format YYYYMMDD or
today = datetime.date.today()
if date_str in ('now', 'today'):
return today
if date_str == 'yesterday':
return today - datetime.timedelta(days=1)
match = re.match(r'(now|today)(?P<sign>[+-])(?P<time>\d+)(?P<unit>day|week|month|year)(s)?', date_str)
if match is not None:
sign = match.group('sign')
time = int(match.group('time'))
if sign == '-':
time = -time
unit = match.group('unit')
# A bad approximation?
if unit == 'month':
unit = 'day'
time *= 30
elif unit == 'year':
unit = 'day'
time *= 365
unit += 's'
delta = datetime.timedelta(**{unit: time})
return today + delta
return datetime.datetime.strptime(date_str, '%Y%m%d').date()
def hyphenate_date(date_str):
Convert a date in 'YYYYMMDD' format to 'YYYY-MM-DD' format"""
match = re.match(r'^(\d\d\d\d)(\d\d)(\d\d)$', date_str)
if match is not None:
return '-'.join(match.groups())
return date_str
class DateRange(object):
"""Represents a time interval between two dates"""
def __init__(self, start=None, end=None):
"""start and end must be strings in the format accepted by date"""
if start is not None:
self.start = date_from_str(start)
self.start = datetime.datetime.min.date()
if end is not None:
self.end = date_from_str(end)
self.end = datetime.datetime.max.date()
if self.start > self.end:
raise ValueError('Date range: "%s" , the start date must be before the end date' % self)
def day(cls, day):
"""Returns a range that only contains the given day"""
return cls(day, day)
def __contains__(self, date):
"""Check if the date is in the range"""
if not isinstance(date, datetime.date):
date = date_from_str(date)
return self.start <= date <= self.end
def __str__(self):
return '%s - %s' % (self.start.isoformat(), self.end.isoformat())
def platform_name():
""" Returns the platform name as a compat_str """
res = platform.platform()
if isinstance(res, bytes):
res = res.decode(preferredencoding())
assert isinstance(res, compat_str)
return res
def _windows_write_string(s, out):
""" Returns True if the string was written using special methods,
False if it has yet to be written out."""
# Adapted from http://stackoverflow.com/a/3259271/35070
import ctypes
import ctypes.wintypes
1: -11,
2: -12,
fileno = out.fileno()
except AttributeError:
# If the output stream doesn't have a fileno, it's virtual
return False
except io.UnsupportedOperation:
# Some strange Windows pseudo files?
return False
if fileno not in WIN_OUTPUT_IDS:
return False
GetStdHandle = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
('GetStdHandle', ctypes.windll.kernel32))
h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
WriteConsoleW = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
ctypes.wintypes.LPVOID)(('WriteConsoleW', ctypes.windll.kernel32))
written = ctypes.wintypes.DWORD(0)
GetFileType = compat_ctypes_WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(('GetFileType', ctypes.windll.kernel32))
GetConsoleMode = compat_ctypes_WINFUNCTYPE(
ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
('GetConsoleMode', ctypes.windll.kernel32))
INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
def not_a_console(handle):
if handle == INVALID_HANDLE_VALUE or handle is None:
return True
return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR
or GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
if not_a_console(h):
return False
def next_nonbmp_pos(s):
return next(i for i, c in enumerate(s) if ord(c) > 0xffff)
except StopIteration:
return len(s)
while s:
count = min(next_nonbmp_pos(s), 1024)
ret = WriteConsoleW(
h, s, count if count else 2, ctypes.byref(written), None)
if ret == 0:
raise OSError('Failed to write string')
if not count: # We just wrote a non-BMP character
assert written.value == 2
s = s[1:]
assert written.value > 0
s = s[written.value:]
return True
def write_string(s, out=None, encoding=None):
if out is None:
out = sys.stderr
assert type(s) == compat_str
if sys.platform == 'win32' and encoding is None and hasattr(out, 'fileno'):
if _windows_write_string(s, out):
if ('b' in getattr(out, 'mode', '')
or sys.version_info[0] < 3): # Python 2 lies about mode of sys.stderr
byt = s.encode(encoding or preferredencoding(), 'ignore')
elif hasattr(out, 'buffer'):
enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
byt = s.encode(enc, 'ignore')
def bytes_to_intlist(bs):
if not bs:
return []
if isinstance(bs[0], int): # Python 3
return list(bs)
return [ord(c) for c in bs]
def intlist_to_bytes(xs):
if not xs:
return b''
return compat_struct_pack('%dB' % len(xs), *xs)
# Cross-platform file locking
if sys.platform == 'win32':
import ctypes.wintypes
import msvcrt
class OVERLAPPED(ctypes.Structure):
_fields_ = [
('Internal', ctypes.wintypes.LPVOID),
('InternalHigh', ctypes.wintypes.LPVOID),
('Offset', ctypes.wintypes.DWORD),
('OffsetHigh', ctypes.wintypes.DWORD),
('hEvent', ctypes.wintypes.HANDLE),
kernel32 = ctypes.windll.kernel32
LockFileEx = kernel32.LockFileEx
LockFileEx.argtypes = [
ctypes.wintypes.HANDLE, # hFile
ctypes.wintypes.DWORD, # dwFlags
ctypes.wintypes.DWORD, # dwReserved
ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
ctypes.POINTER(OVERLAPPED) # Overlapped
LockFileEx.restype = ctypes.wintypes.BOOL
UnlockFileEx = kernel32.UnlockFileEx
UnlockFileEx.argtypes = [
ctypes.wintypes.HANDLE, # hFile
ctypes.wintypes.DWORD, # dwReserved
ctypes.wintypes.DWORD, # nNumberOfBytesToLockLow
ctypes.wintypes.DWORD, # nNumberOfBytesToLockHigh
ctypes.POINTER(OVERLAPPED) # Overlapped
UnlockFileEx.restype = ctypes.wintypes.BOOL
whole_low = 0xffffffff
whole_high = 0x7fffffff
def _lock_file(f, exclusive):
overlapped = OVERLAPPED()
overlapped.Offset = 0
overlapped.OffsetHigh = 0
overlapped.hEvent = 0
f._lock_file_overlapped_p = ctypes.pointer(overlapped)
handle = msvcrt.get_osfhandle(f.fileno())
if not LockFileEx(handle, 0x2 if exclusive else 0x0, 0,
whole_low, whole_high, f._lock_file_overlapped_p):
raise OSError('Locking file failed: %r' % ctypes.FormatError())
def _unlock_file(f):
assert f._lock_file_overlapped_p
handle = msvcrt.get_osfhandle(f.fileno())
if not UnlockFileEx(handle, 0,
whole_low, whole_high, f._lock_file_overlapped_p):
raise OSError('Unlocking file failed: %r' % ctypes.FormatError())
# Some platforms, such as Jython, is missing fcntl
import fcntl
def _lock_file(f, exclusive):
fcntl.flock(f, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
def _unlock_file(f):
fcntl.flock(f, fcntl.LOCK_UN)
except ImportError:
UNSUPPORTED_MSG = 'file locking is not supported on this platform'
def _lock_file(f, exclusive):
def _unlock_file(f):
class locked_file(object):
def __init__(self, filename, mode, encoding=None):
assert mode in ['r', 'a', 'w']
self.f = io.open(filename, mode, encoding=encoding)
self.mode = mode
def __enter__(self):
exclusive = self.mode != 'r'
_lock_file(self.f, exclusive)
except IOError:
return self
def __exit__(self, etype, value, traceback):
def __iter__(self):
return iter(self.f)
def write(self, *args):
return self.f.write(*args)
def read(self, *args):
return self.f.read(*args)
def get_filesystem_encoding():
encoding = sys.getfilesystemencoding()
return encoding if encoding is not None else 'utf-8'
def shell_quote(args):
quoted_args = []
encoding = get_filesystem_encoding()
for a in args:
if isinstance(a, bytes):
# We may get a filename encoded with 'encodeFilename'
a = a.decode(encoding)
return ' '.join(quoted_args)
def smuggle_url(url, data):
""" Pass additional data in a URL for internal use. """
url, idata = unsmuggle_url(url, {})
sdata = compat_urllib_parse_urlencode(
{'__youtubedl_smuggle': json.dumps(data)})
return url + '#' + sdata
def unsmuggle_url(smug_url, default=None):
if '#__youtubedl_smuggle' not in smug_url:
return smug_url, default
url, _, sdata = smug_url.rpartition('#')
jsond = compat_parse_qs(sdata)['__youtubedl_smuggle'][0]
data = json.loads(jsond)
return url, data
def format_bytes(bytes):
if bytes is None:
return 'N/A'
if type(bytes) is str:
bytes = float(bytes)
if bytes == 0.0:
exponent = 0
exponent = int(math.log(bytes, 1024.0))
suffix = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB'][exponent]
converted = float(bytes) / float(1024 ** exponent)
return '%.2f%s' % (converted, suffix)
def lookup_unit_table(unit_table, s):
units_re = '|'.join(re.escape(u) for u in unit_table)
m = re.match(
r'(?P<num>[0-9]+(?:[,.][0-9]*)?)\s*(?P<unit>%s)\b' % units_re, s)
if not m:
return None
num_str = m.group('num').replace(',', '.')
mult = unit_table[m.group('unit')]
return int(float(num_str) * mult)
def parse_filesize(s):
if s is None:
return None
# The lower-case forms are of course incorrect and unofficial,
# but we support those too
'B': 1,
'b': 1,
'bytes': 1,
'KiB': 1024,
'KB': 1000,
'kB': 1024,
'Kb': 1000,
'kb': 1000,
'kilobytes': 1000,
'kibibytes': 1024,
'MiB': 1024 ** 2,
'MB': 1000 ** 2,
'mB': 1024 ** 2,
'Mb': 1000 ** 2,
'mb': 1000 ** 2,
'megabytes': 1000 ** 2,
'mebibytes': 1024 ** 2,
'GiB': 1024 ** 3,
'GB': 1000 ** 3,
'gB': 1024 ** 3,
'Gb': 1000 ** 3,
'gb': 1000 ** 3,
'gigabytes': 1000 ** 3,
'gibibytes': 1024 ** 3,
'TiB': 1024 ** 4,
'TB': 1000 ** 4,
'tB': 1024 ** 4,
'Tb': 1000 ** 4,
'tb': 1000 ** 4,
'terabytes': 1000 ** 4,
'tebibytes': 1024 ** 4,
'PiB': 1024 ** 5,
'PB': 1000 ** 5,
'pB': 1024 ** 5,
'Pb': 1000 ** 5,
'pb': 1000 ** 5,
'petabytes': 1000 ** 5,
'pebibytes': 1024 ** 5,
'EiB': 1024 ** 6,
'EB': 1000 ** 6,
'eB': 1024 ** 6,
'Eb': 1000 ** 6,
'eb': 1000 ** 6,
'exabytes': 1000 ** 6,
'exbibytes': 1024 ** 6,
'ZiB': 1024 ** 7,
'ZB': 1000 ** 7,
'zB': 1024 ** 7,
'Zb': 1000 ** 7,
'zb': 1000 ** 7,
'zettabytes': 1000 ** 7,
'zebibytes': 1024 ** 7,
'YiB': 1024 ** 8,
'YB': 1000 ** 8,
'yB': 1024 ** 8,
'Yb': 1000 ** 8,
'yb': 1000 ** 8,
'yottabytes': 1000 ** 8,
'yobibytes': 1024 ** 8,
return lookup_unit_table(_UNIT_TABLE, s)
def parse_count(s):
if s is None:
return None
s = s.strip()
if re.match(r'^[\d,.]+$', s):
return str_to_int(s)
'k': 1000,
'K': 1000,
'm': 1000 ** 2,
'M': 1000 ** 2,
'kk': 1000 ** 2,
'KK': 1000 ** 2,
return lookup_unit_table(_UNIT_TABLE, s)
def parse_resolution(s):
if s is None:
return {}
mobj = re.search(r'\b(?P<w>\d+)\s*[xX×]\s*(?P<h>\d+)\b', s)
if mobj:
return {
'width': int(mobj.group('w')),
'height': int(mobj.group('h')),
mobj = re.search(r'\b(\d+)[pPiI]\b', s)
if mobj:
return {'height': int(mobj.group(1))}
mobj = re.search(r'\b([48])[kK]\b', s)
if mobj:
return {'height': int(mobj.group(1)) * 540}
return {}
def parse_bitrate(s):
if not isinstance(s, compat_str):
mobj = re.search(r'\b(\d+)\s*kbps', s)
if mobj:
return int(mobj.group(1))
def month_by_name(name, lang='en'):
""" Return the number of a month by (locale-independently) English name """
month_names = MONTH_NAMES.get(lang, MONTH_NAMES['en'])
return month_names.index(name) + 1
except ValueError:
return None
def month_by_abbreviation(abbrev):
""" Return the number of a month by (locale-independently) English
abbreviations """
return [s[:3] for s in ENGLISH_MONTH_NAMES].index(abbrev) + 1
except ValueError:
return None
def fix_xml_ampersands(xml_str):
"""Replace all the '&' by '&amp;' in XML"""
return re.sub(
def setproctitle(title):
assert isinstance(title, compat_str)
# ctypes in Jython is not complete
# http://bugs.jython.org/issue2148
if sys.platform.startswith('java'):
libc = ctypes.cdll.LoadLibrary('libc.so.6')
except OSError:
except TypeError:
# LoadLibrary in Windows Python 2.7.13 only expects
# a bytestring, but since unicode_literals turns
# every string into a unicode string, it fails.
title_bytes = title.encode('utf-8')
buf = ctypes.create_string_buffer(len(title_bytes))
buf.value = title_bytes
libc.prctl(15, buf, 0, 0, 0)
except AttributeError:
return # Strange libc, just skip this
def remove_start(s, start):
return s[len(start):] if s is not None and s.startswith(start) else s
def remove_end(s, end):
return s[:-len(end)] if s is not None and s.endswith(end) else s
def remove_quotes(s):
if s is None or len(s) < 2:
return s
for quote in ('"', "'", ):
if s[0] == quote and s[-1] == quote:
return s[1:-1]
return s
def url_basename(url):
path = compat_urlparse.urlparse(url).path
return path.strip('/').split('/')[-1]
def base_url(url):
return re.match(r'https?://[^?#&]+/', url).group()
def urljoin(base, path):
if isinstance(path, bytes):
path = path.decode('utf-8')
if not isinstance(path, compat_str) or not path:
return None
if re.match(r'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path):
return path
if isinstance(base, bytes):
base = base.decode('utf-8')
if not isinstance(base, compat_str) or not re.match(
r'^(?:https?:)?//', base):
return None
return compat_urlparse.urljoin(base, path)
class HEADRequest(compat_urllib_request.Request):
def get_method(self):
return 'HEAD'
class PUTRequest(compat_urllib_request.Request):
def get_method(self):
return 'PUT'
def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
if get_attr:
if v is not None:
v = getattr(v, get_attr, None)
if v == '':
v = None
if v is None:
return default
return int(v) * invscale // scale
except (ValueError, TypeError):
return default
def str_or_none(v, default=None):
return default if v is None else compat_str(v)
def str_to_int(int_str):
""" A more relaxed version of int_or_none """
if int_str is None:
return None
int_str = re.sub(r'[,\.\+]', '', int_str)
return int(int_str)
def float_or_none(v, scale=1, invscale=1, default=None):
if v is None:
return default
return float(v) * invscale / scale
except (ValueError, TypeError):
return default
def bool_or_none(v, default=None):
return v if isinstance(v, bool) else default
def strip_or_none(v):
return None if v is None else v.strip()
def url_or_none(url):
if not url or not isinstance(url, compat_str):
return None
url = url.strip()
return url if re.match(r'^(?:[a-zA-Z][\da-zA-Z.+-]*:)?//', url) else None
def parse_duration(s):
if not isinstance(s, compat_basestring):
return None
s = s.strip()
days, hours, mins, secs, ms = [None] * 5
m = re.match(r'(?:(?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?(?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?Z?$', s)
if m:
days, hours, mins, secs, ms = m.groups()
m = re.match(
)?Z?$''', s)
if m:
days, hours, mins, secs, ms = m.groups()
m = re.match(r'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s)
if m:
hours, mins = m.groups()
return None
duration = 0
if secs:
duration += float(secs)
if mins:
duration += float(mins) * 60
if hours:
duration += float(hours) * 60 * 60
if days:
duration += float(days) * 24 * 60 * 60
if ms:
duration += float(ms)
return duration
def prepend_extension(filename, ext, expected_real_ext=None):
name, real_ext = os.path.splitext(filename)
return (
'{0}.{1}{2}'.format(name, ext, real_ext)
if not expected_real_ext or real_ext[1:] == expected_real_ext
else '{0}.{1}'.format(filename, ext))
def replace_extension(filename, ext, expected_real_ext=None):
name, real_ext = os.path.splitext(filename)
return '{0}.{1}'.format(
name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
def check_executable(exe, args=[]):
""" Checks if the given binary is installed somewhere in PATH, and returns its name.
args can be a list of arguments for a short output (like -version) """
subprocess.Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
except OSError:
return False
return exe
def get_exe_version(exe, args=['--version'],
version_re=None, unrecognized='present'):
""" Returns the version of the specified executable,
or False if the executable is not present """
# STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
# SIGTTOU if youtube-dl is run in the background.
# See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
out, _ = subprocess.Popen(
[encodeArgument(exe)] + args,
stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
except OSError:
return False
if isinstance(out, bytes): # Python 2.x
out = out.decode('ascii', 'ignore')
return detect_exe_version(out, version_re, unrecognized)
def detect_exe_version(output, version_re=None, unrecognized='present'):
assert isinstance(output, compat_str)
if version_re is None:
version_re = r'version\s+([-0-9._a-zA-Z]+)'
m = re.search(version_re, output)
if m:
return m.group(1)
return unrecognized
class PagedList(object):
def __len__(self):
# This is only useful for tests
return len(self.getslice())
class OnDemandPagedList(PagedList):
def __init__(self, pagefunc, pagesize, use_cache=True):
self._pagefunc = pagefunc
self._pagesize = pagesize
self._use_cache = use_cache
if use_cache:
self._cache = {}
def getslice(self, start=0, end=None):
res = []
for pagenum in itertools.count(start // self._pagesize):
firstid = pagenum * self._pagesize
nextfirstid = pagenum * self._pagesize + self._pagesize
if start >= nextfirstid:
page_results = None
if self._use_cache:
page_results = self._cache.get(pagenum)
if page_results is None:
page_results = list(self._pagefunc(pagenum))
if self._use_cache:
self._cache[pagenum] = page_results
startv = (
start % self._pagesize
if firstid <= start < nextfirstid
else 0)
endv = (
((end - 1) % self._pagesize) + 1
if (end is not None and firstid <= end <= nextfirstid)
else None)
if startv != 0 or endv is not None:
page_results = page_results[startv:endv]
# A little optimization - if current page is not "full", ie. does
# not contain page_size videos then we can assume that this page
# is the last one - there are no more ids on further pages -
# i.e. no need to query again.
if len(page_results) + startv < self._pagesize:
# If we got the whole page, but the next page is not interesting,
# break out early as well
if end == nextfirstid:
return res
class InAdvancePagedList(PagedList):
def __init__(self, pagefunc, pagecount, pagesize):