import email.header
import email.utils
import errno
-import functools
import gzip
import hashlib
import hmac
import tempfile
import time
import traceback
+import types
import urllib.parse
import xml.etree.ElementTree
import zlib
+from .compat import asyncio, functools # isort: split
from .compat import (
- asyncio,
compat_chr,
compat_cookiejar,
compat_etree_fromstring,
NO_DEFAULT = object()
+IDENTITY = lambda x: x
ENGLISH_MONTH_NAMES = [
'January', 'February', 'March', 'April', 'May', 'June',
NUMBER_RE = r'\d+(?:\.\d+)?'
+@functools.cache
def preferredencoding():
"""Get preferred encoding.
return n.attrib[key]
-def get_element_by_id(id, html):
+def get_element_by_id(id, html, **kwargs):
"""Return the content of the tag with the specified ID in the passed HTML document"""
- return get_element_by_attribute('id', id, html)
+ return get_element_by_attribute('id', id, html, **kwargs)
-def get_element_html_by_id(id, html):
+def get_element_html_by_id(id, html, **kwargs):
"""Return the html of the tag with the specified ID in the passed HTML document"""
- return get_element_html_by_attribute('id', id, html)
+ return get_element_html_by_attribute('id', id, html, **kwargs)
def get_element_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)
+def get_element_by_attribute(attribute, value, html, **kwargs):
+ retval = get_elements_by_attribute(attribute, value, html, **kwargs)
return retval[0] if retval else None
-def get_element_html_by_attribute(attribute, value, html, escape_value=True):
- retval = get_elements_html_by_attribute(attribute, value, html, escape_value)
+def get_element_html_by_attribute(attribute, value, html, **kargs):
+ retval = get_elements_html_by_attribute(attribute, value, html, **kargs)
return retval[0] if retval else None
-def get_elements_by_class(class_name, html):
+def get_elements_by_class(class_name, html, **kargs):
"""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),
+ 'class', r'[^\'"]*(?<=[\'"\s])%s(?=[\'"\s])[^\'"]*' % re.escape(class_name),
html, escape_value=False)
def get_elements_html_by_class(class_name, html):
"""Return the html of all tags with the specified class in the passed HTML document as a list"""
return get_elements_html_by_attribute(
- 'class', r'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
+ 'class', r'[^\'"]*(?<=[\'"\s])%s(?=[\'"\s])[^\'"]*' % re.escape(class_name),
html, escape_value=False)
return html.strip()
+class LenientJSONDecoder(json.JSONDecoder):
+ def __init__(self, *args, transform_source=None, ignore_extra=False, **kwargs):
+ self.transform_source, self.ignore_extra = transform_source, ignore_extra
+ super().__init__(*args, **kwargs)
+
+ def decode(self, s):
+ if self.transform_source:
+ s = self.transform_source(s)
+ if self.ignore_extra:
+ return self.raw_decode(s.lstrip())[0]
+ return super().decode(s)
+
+
def sanitize_open(filename, open_mode):
"""Try to open the given filename, and slightly tweak it if this fails.
# Ref: https://github.com/yt-dlp/yt-dlp/issues/3124
raise LockingUnsupportedError()
stream = locked_file(filename, open_mode, block=False).__enter__()
- except LockingUnsupportedError:
+ except OSError:
stream = open(filename, open_mode)
- return (stream, filename)
+ return stream, filename
except OSError as err:
if attempt or err.errno in (errno.EACCES,):
raise
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('//'):
+ if url is None:
+ return
+ elif url.startswith('//'):
return 'http:%s' % url
# Fix some common typos seen so far
COMMON_TYPOS = (
parts.hostname if parts.port is None
else '%s:%d' % (parts.hostname, parts.port))))
auth_payload = base64.b64encode(
- ('%s:%s' % (parts.username, parts.password or '')).encode('utf-8'))
- return url, 'Basic ' + auth_payload.decode('utf-8')
+ ('%s:%s' % (parts.username, parts.password or '')).encode())
+ return url, f'Basic {auth_payload.decode()}'
def sanitized_Request(url, *args, **kwargs):
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:
- res.append(el)
- return res
+def orderedSet(iterable, *, lazy=False):
+ """Remove all duplicates from the input iterable"""
+ def _iter():
+ seen = [] # Do not use set since the items can be unhashable
+ for x in iterable:
+ if x not in seen:
+ seen.append(x)
+ yield x
+
+ return _iter() if lazy else list(_iter())
def _htmlentity_transform(entity_with_semicolon):
def process_communicate_or_kill(p, *args, **kwargs):
- try:
- return p.communicate(*args, **kwargs)
- except BaseException: # Including KeyboardInterrupt
- p.kill()
- p.wait()
- raise
+ write_string('DeprecationWarning: yt_dlp.utils.process_communicate_or_kill is deprecated '
+ 'and may be removed in a future version. Use yt_dlp.utils.Popen.communicate_or_kill instead')
+ return Popen.communicate_or_kill(p, *args, **kwargs)
class Popen(subprocess.Popen):
else:
_startupinfo = None
- def __init__(self, *args, **kwargs):
+ def __init__(self, *args, text=False, **kwargs):
+ if text is True:
+ kwargs['universal_newlines'] = True # For 3.6 compatibility
+ kwargs.setdefault('encoding', 'utf-8')
+ kwargs.setdefault('errors', 'replace')
super().__init__(*args, **kwargs, startupinfo=self._startupinfo)
def communicate_or_kill(self, *args, **kwargs):
- return process_communicate_or_kill(self, *args, **kwargs)
+ try:
+ return self.communicate(*args, **kwargs)
+ except BaseException: # Including KeyboardInterrupt
+ self.kill(timeout=None)
+ raise
+
+ def kill(self, *, timeout=0):
+ super().kill()
+ if timeout != 0:
+ self.wait(timeout=timeout)
+
+ @classmethod
+ def run(cls, *args, **kwargs):
+ with cls(*args, **kwargs) as proc:
+ stdout, stderr = proc.communicate_or_kill()
+ return stdout or '', stderr or '', proc.returncode
def get_subprocess_encoding():
context.check_hostname = opts_check_certificate
if params.get('legacyserverconnect'):
context.options |= 4 # SSL_OP_LEGACY_SERVER_CONNECT
+ # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
+ context.set_ciphers('DEFAULT')
+
context.verify_mode = ssl.CERT_REQUIRED if opts_check_certificate else ssl.CERT_NONE
if opts_check_certificate:
if has_certifi and 'no-certifi' not in params.get('compat_opts', []):
context.load_verify_locations(cafile=certifi.where())
- else:
- try:
- context.load_default_certs()
- # Work around the issue in load_default_certs when there are bad certificates. See:
- # https://github.com/yt-dlp/yt-dlp/issues/1060,
- # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
- except ssl.SSLError:
- # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
- if sys.platform == 'win32' and hasattr(ssl, 'enum_certificates'):
- # Create a new context to discard any certificates that were already loaded
- context = ssl.SSLContext(ssl.PROTOCOL_TLS_CLIENT)
- context.check_hostname, context.verify_mode = True, ssl.CERT_REQUIRED
- for storename in ('CA', 'ROOT'):
- _ssl_load_windows_store_certs(context, storename)
- context.set_default_verify_paths()
+ try:
+ context.load_default_certs()
+ # Work around the issue in load_default_certs when there are bad certificates. See:
+ # https://github.com/yt-dlp/yt-dlp/issues/1060,
+ # https://bugs.python.org/issue35665, https://bugs.python.org/issue45312
+ except ssl.SSLError:
+ # enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
+ if sys.platform == 'win32' and hasattr(ssl, 'enum_certificates'):
+ for storename in ('CA', 'ROOT'):
+ _ssl_load_windows_store_certs(context, storename)
+ context.set_default_verify_paths()
+
+ client_certfile = params.get('client_certificate')
+ if client_certfile:
+ try:
+ context.load_cert_chain(
+ client_certfile, keyfile=params.get('client_certificate_key'),
+ password=params.get('client_certificate_password'))
+ except ssl.SSLError:
+ raise YoutubeDLError('Unable to load client certificate')
+
+ # Some servers may reject requests if ALPN extension is not sent. See:
+ # https://github.com/python/cpython/issues/85140
+ # https://github.com/yt-dlp/yt-dlp/issues/3878
+ with contextlib.suppress(NotImplementedError):
+ context.set_alpn_protocols(['http/1.1'])
+
return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
def bug_reports_message(before=';'):
- msg = ('please report this issue on https://github.com/yt-dlp/yt-dlp/issues?q= , '
- 'filling out the appropriate issue template. '
- 'Confirm you are on the latest version using yt-dlp -U')
+ from .update import REPOSITORY
+
+ msg = (f'please report this issue on https://github.com/{REPOSITORY}/issues?q= , '
+ 'filling out the appropriate issue template. Confirm you are on the latest version using yt-dlp -U')
before = before.rstrip()
if not before or before.endswith(('.', '!', '?')):
self.video_id = video_id
self.ie = ie
self.exc_info = sys.exc_info() # preserve original exception
+ if isinstance(self.exc_info[1], ExtractorError):
+ self.exc_info = self.exc_info[1].exc_info
super().__init__(''.join((
- format_field(ie, template='[%s] '),
- format_field(video_id, template='%s: '),
+ format_field(ie, None, '[%s] '),
+ format_field(video_id, None, '%s: '),
msg,
- format_field(cause, template=' (caused by %r)'),
+ format_field(cause, None, ' (caused by %r)'),
'' if expected else bug_reports_message())))
def format_traceback(self):
location = resp.headers.get('Location')
if location:
# As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
- location = location.encode('iso-8859-1').decode('utf-8')
+ location = location.encode('iso-8859-1').decode()
location_escaped = escape_url(location)
if location != location_escaped:
del resp.headers['Location']
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)
+ try:
+ return self.do_open(
+ functools.partial(_create_http_connection, self, conn_class, True), req, **kwargs)
+ except urllib.error.URLError as e:
+ if (isinstance(e.reason, ssl.SSLError)
+ and getattr(e.reason, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
+ raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
+ raise
class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
'CookieFileEntry',
('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
- def save(self, filename=None, ignore_discard=False, ignore_expires=False):
+ def __init__(self, filename=None, *args, **kwargs):
+ super().__init__(None, *args, **kwargs)
+ if self.is_path(filename):
+ filename = os.fspath(filename)
+ self.filename = filename
+
+ @staticmethod
+ def _true_or_false(cndn):
+ return 'TRUE' if cndn else 'FALSE'
+
+ @staticmethod
+ def is_path(file):
+ return isinstance(file, (str, bytes, os.PathLike))
+
+ @contextlib.contextmanager
+ def open(self, file, *, write=False):
+ if self.is_path(file):
+ with open(file, 'w' if write else 'r', encoding='utf-8') as f:
+ yield f
+ else:
+ if write:
+ file.truncate(0)
+ yield file
+
+ def _really_save(self, f, ignore_discard=False, ignore_expires=False):
+ now = time.time()
+ for cookie in self:
+ if (not ignore_discard and cookie.discard
+ or not ignore_expires and cookie.is_expired(now)):
+ continue
+ name, value = cookie.name, cookie.value
+ if value is None:
+ # cookies.txt regards 'Set-Cookie: foo' as a cookie
+ # with no name, whereas http.cookiejar regards it as a
+ # cookie with no value.
+ name, value = '', name
+ f.write('%s\n' % '\t'.join((
+ cookie.domain,
+ self._true_or_false(cookie.domain.startswith('.')),
+ cookie.path,
+ self._true_or_false(cookie.secure),
+ str_or_none(cookie.expires, default=''),
+ name, value
+ )))
+
+ def save(self, filename=None, *args, **kwargs):
"""
Save cookies to a file.
+ Code is taken from CPython 3.6
+ https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
- Most of the code is taken from CPython 3.8 and slightly adapted
- to support cookie files with UTF-8 in both python 2 and 3.
- """
if filename is None:
if self.filename is not None:
filename = self.filename
else:
raise ValueError(compat_cookiejar.MISSING_FILENAME_TEXT)
- # Store session cookies with `expires` set to 0 instead of an empty
- # string
+ # 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
- with open(filename, 'w', encoding='utf-8') as f:
+ with self.open(filename, write=True) as f:
f.write(self._HEADER)
- now = time.time()
- for cookie in self:
- if not ignore_discard and cookie.discard:
- continue
- if not ignore_expires and cookie.is_expired(now):
- continue
- if cookie.secure:
- secure = 'TRUE'
- else:
- secure = 'FALSE'
- if cookie.domain.startswith('.'):
- initial_dot = 'TRUE'
- else:
- initial_dot = 'FALSE'
- if cookie.expires is not None:
- expires = compat_str(cookie.expires)
- else:
- expires = ''
- if cookie.value is None:
- # cookies.txt regards 'Set-Cookie: foo' as a cookie
- # with no name, whereas http.cookiejar regards it as a
- # cookie with no value.
- name = ''
- value = cookie.name
- else:
- name = cookie.name
- value = cookie.value
- f.write(
- '\t'.join([cookie.domain, initial_dot, cookie.path,
- secure, expires, name, value]) + '\n')
+ self._really_save(f, *args, **kwargs)
def load(self, filename=None, ignore_discard=False, ignore_expires=False):
"""Load cookies from a file."""
return line
cf = io.StringIO()
- with open(filename, encoding='utf-8') as f:
+ with self.open(filename) as f:
for line in f:
try:
cf.write(prepare_line(line))
CONTENT_HEADERS = ("content-length", "content-type")
# NB: don't use dict comprehension for python 2.6 compatibility
newheaders = {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
+
+ # A 303 must either use GET or HEAD for subsequent request
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
+ if code == 303 and m != 'HEAD':
+ m = 'GET'
+ # 301 and 302 redirects are commonly turned into a GET from a POST
+ # for subsequent requests by browsers, so we'll do the same.
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
+ if code in (301, 302) and m == 'POST':
+ m = 'GET'
+
return compat_urllib_request.Request(
newurl, headers=newheaders, origin_req_host=req.origin_req_host,
- unverifiable=True)
+ unverifiable=True, method=m)
def extract_timezone(date_str):
def datetime_from_str(date_str, precision='auto', format='%Y%m%d'):
- """
- Return a datetime object from a string in the format YYYYMMDD or
- (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
-
- format: string date format used to return datetime object from
- precision: round the time portion of a datetime object.
- auto|microsecond|second|minute|hour|day.
- auto: round to the unit provided in date_str (if applicable).
+ R"""
+ Return a datetime object from a string.
+ Supported format:
+ (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
+
+ @param format strftime format of DATE
+ @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
+ auto: round to the unit provided in date_str (if applicable).
"""
auto_precision = False
if precision == 'auto':
if date_str == 'yesterday':
return today - datetime.timedelta(days=1)
match = re.match(
- r'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)(s)?',
+ r'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
date_str)
if match is not None:
start_time = datetime_from_str(match.group('start'), precision, format)
def date_from_str(date_str, format='%Y%m%d', strict=False):
- """
- Return a datetime object from a string in the format YYYYMMDD or
- (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
-
- If "strict", only (now|today)[+-][0-9](day|week|month|year)(s)? is allowed
+ R"""
+ Return a date object from a string using datetime_from_str
- format: string date format used to return datetime object from
+ @param strict Restrict allowed patterns to "YYYYMMDD" and
+ (now|today|yesterday)(-\d+(day|week|month|year)s?)?
"""
- if strict and not re.fullmatch(r'\d{8}|(now|today)[+-]\d+(day|week|month|year)(s)?', date_str):
- raise ValueError(f'Invalid date format {date_str}')
+ if strict and not re.fullmatch(r'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str):
+ raise ValueError(f'Invalid date format "{date_str}"')
return datetime_from_str(date_str, precision='microsecond', format=format).date()
return res
+@functools.cache
def get_windows_version():
- ''' Get Windows version. None if it's not running on Windows '''
+ ''' Get Windows version. returns () if it's not running on Windows '''
if compat_os_name == 'nt':
return version_tuple(platform.win32_ver()[1])
else:
- return None
+ return ()
def write_string(s, out=None, encoding=None):
assert isinstance(s, str)
out = out or sys.stderr
- from .compat import WINDOWS_VT_MODE # Must be imported locally
- if WINDOWS_VT_MODE:
+ if compat_os_name == 'nt' and supports_terminal_sequences(out):
s = re.sub(r'([\r\n]+)', r' \1', s)
+ enc, buffer = None, out
if 'b' in getattr(out, 'mode', ''):
- byt = s.encode(encoding or preferredencoding(), 'ignore')
- out.write(byt)
+ enc = encoding or preferredencoding()
elif hasattr(out, 'buffer'):
+ buffer = out.buffer
enc = encoding or getattr(out, 'encoding', None) or preferredencoding()
- byt = s.encode(enc, 'ignore')
- out.buffer.write(byt)
- else:
- out.write(s)
+
+ buffer.write(s.encode(enc, 'ignore') if enc else s)
out.flush()
return compat_struct_pack('%dB' % len(xs), *xs)
-class LockingUnsupportedError(IOError):
- msg = 'File locking is not supported on this platform'
+class LockingUnsupportedError(OSError):
+ msg = 'File locking is not supported'
def __init__(self):
super().__init__(self.msg)
if not LockFileEx(msvcrt.get_osfhandle(f.fileno()),
(0x2 if exclusive else 0x0) | (0x0 if block else 0x1),
0, whole_low, whole_high, f._lock_file_overlapped_p):
- raise BlockingIOError('Locking file failed: %r' % ctypes.FormatError())
+ # NB: No argument form of "ctypes.FormatError" does not work on PyPy
+ raise BlockingIOError(f'Locking file failed: {ctypes.FormatError(ctypes.GetLastError())!r}')
def _unlock_file(f):
assert f._lock_file_overlapped_p
self.f.close()
raise
if 'w' in self.mode:
- self.f.truncate()
+ try:
+ self.f.truncate()
+ except OSError as e:
+ if e.errno not in (
+ errno.ESPIPE, # Illegal seek - expected for FIFO
+ errno.EINVAL, # Invalid argument - expected for /dev/null
+ ):
+ raise
return self
def unlock(self):
return iter(self.f)
+@functools.cache
def get_filesystem_encoding():
encoding = sys.getfilesystemencoding()
return encoding if encoding is not None else 'utf-8'
# a bytestring, but since unicode_literals turns
# every string into a unicode string, it fails.
return
- title_bytes = title.encode('utf-8')
+ title_bytes = title.encode()
buf = ctypes.create_string_buffer(len(title_bytes))
buf.value = title_bytes
try:
def urljoin(base, path):
if isinstance(path, bytes):
- path = path.decode('utf-8')
+ path = path.decode()
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')
+ base = base.decode()
if not isinstance(base, compat_str) or not re.match(
r'^(?:https?:)?//', base):
return None
""" 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) """
try:
- Popen([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate_or_kill()
+ Popen.run([exe] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
except OSError:
return False
return exe
# STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
# SIGTTOU if yt-dlp is run in the background.
# See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
- out, _ = Popen(
- [encodeArgument(exe)] + args, stdin=subprocess.PIPE,
- stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate_or_kill()
+ stdout, _, _ = Popen.run([encodeArgument(exe)] + args, text=True,
+ stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
except OSError:
return False
- if isinstance(out, bytes): # Python 2.x
- out = out.decode('ascii', 'ignore')
- return out
+ return stdout
def detect_exe_version(output, version_re=None, unrecognized='present'):
return detect_exe_version(out, version_re, unrecognized) if out else False
+def frange(start=0, stop=None, step=1):
+ """Float range"""
+ if stop is None:
+ start, stop = 0, start
+ sign = [-1, 1][step > 0] if step else 0
+ while sign * start < sign * stop:
+ yield start
+ start += step
+
+
class LazyList(collections.abc.Sequence):
- ''' Lazy immutable list from an iterable
- Note that slices of a LazyList are lists and not LazyList'''
+ """Lazy immutable list from an iterable
+ Note that slices of a LazyList are lists and not LazyList"""
class IndexError(IndexError):
pass
def __init__(self, iterable, *, reverse=False, _cache=None):
- self.__iterable = iter(iterable)
- self.__cache = [] if _cache is None else _cache
- self.__reversed = reverse
+ self._iterable = iter(iterable)
+ self._cache = [] if _cache is None else _cache
+ self._reversed = reverse
def __iter__(self):
- if self.__reversed:
+ if self._reversed:
# We need to consume the entire iterable to iterate in reverse
yield from self.exhaust()
return
- yield from self.__cache
- for item in self.__iterable:
- self.__cache.append(item)
+ yield from self._cache
+ for item in self._iterable:
+ self._cache.append(item)
yield item
- def __exhaust(self):
- self.__cache.extend(self.__iterable)
- # Discard the emptied iterable to make it pickle-able
- self.__iterable = []
- return self.__cache
+ def _exhaust(self):
+ self._cache.extend(self._iterable)
+ self._iterable = [] # Discard the emptied iterable to make it pickle-able
+ return self._cache
def exhaust(self):
- ''' Evaluate the entire iterable '''
- return self.__exhaust()[::-1 if self.__reversed else 1]
+ """Evaluate the entire iterable"""
+ return self._exhaust()[::-1 if self._reversed else 1]
@staticmethod
- def __reverse_index(x):
+ def _reverse_index(x):
return None if x is None else -(x + 1)
def __getitem__(self, idx):
if isinstance(idx, slice):
- if self.__reversed:
- idx = slice(self.__reverse_index(idx.start), self.__reverse_index(idx.stop), -(idx.step or 1))
+ if self._reversed:
+ idx = slice(self._reverse_index(idx.start), self._reverse_index(idx.stop), -(idx.step or 1))
start, stop, step = idx.start, idx.stop, idx.step or 1
elif isinstance(idx, int):
- if self.__reversed:
- idx = self.__reverse_index(idx)
+ if self._reversed:
+ idx = self._reverse_index(idx)
start, stop, step = idx, idx, 0
else:
raise TypeError('indices must be integers or slices')
or (stop is None and step > 0)):
# We need to consume the entire iterable to be able to slice from the end
# Obviously, never use this with infinite iterables
- self.__exhaust()
+ self._exhaust()
try:
- return self.__cache[idx]
+ return self._cache[idx]
except IndexError as e:
raise self.IndexError(e) from e
- n = max(start or 0, stop or 0) - len(self.__cache) + 1
+ n = max(start or 0, stop or 0) - len(self._cache) + 1
if n > 0:
- self.__cache.extend(itertools.islice(self.__iterable, n))
+ self._cache.extend(itertools.islice(self._iterable, n))
try:
- return self.__cache[idx]
+ return self._cache[idx]
except IndexError as e:
raise self.IndexError(e) from e
def __bool__(self):
try:
- self[-1] if self.__reversed else self[0]
+ self[-1] if self._reversed else self[0]
except self.IndexError:
return False
return True
def __len__(self):
- self.__exhaust()
- return len(self.__cache)
+ self._exhaust()
+ return len(self._cache)
def __reversed__(self):
- return type(self)(self.__iterable, reverse=not self.__reversed, _cache=self.__cache)
+ return type(self)(self._iterable, reverse=not self._reversed, _cache=self._cache)
def __copy__(self):
- return type(self)(self.__iterable, reverse=self.__reversed, _cache=self.__cache)
+ return type(self)(self._iterable, reverse=self._reversed, _cache=self._cache)
def __repr__(self):
# repr and str should mimic a list. So we exhaust the iterable
yield from page_results
+class PlaylistEntries:
+ MissingEntry = object()
+ is_exhausted = False
+
+ def __init__(self, ydl, info_dict):
+ self.ydl = ydl
+
+ # _entries must be assigned now since infodict can change during iteration
+ entries = info_dict.get('entries')
+ if entries is None:
+ raise EntryNotInPlaylist('There are no entries')
+ elif isinstance(entries, list):
+ self.is_exhausted = True
+
+ requested_entries = info_dict.get('requested_entries')
+ self.is_incomplete = bool(requested_entries)
+ if self.is_incomplete:
+ assert self.is_exhausted
+ self._entries = [self.MissingEntry] * max(requested_entries)
+ for i, entry in zip(requested_entries, entries):
+ self._entries[i - 1] = entry
+ elif isinstance(entries, (list, PagedList, LazyList)):
+ self._entries = entries
+ else:
+ self._entries = LazyList(entries)
+
+ PLAYLIST_ITEMS_RE = re.compile(r'''(?x)
+ (?P<start>[+-]?\d+)?
+ (?P<range>[:-]
+ (?P<end>[+-]?\d+|inf(?:inite)?)?
+ (?::(?P<step>[+-]?\d+))?
+ )?''')
+
+ @classmethod
+ def parse_playlist_items(cls, string):
+ for segment in string.split(','):
+ if not segment:
+ raise ValueError('There is two or more consecutive commas')
+ mobj = cls.PLAYLIST_ITEMS_RE.fullmatch(segment)
+ if not mobj:
+ raise ValueError(f'{segment!r} is not a valid specification')
+ start, end, step, has_range = mobj.group('start', 'end', 'step', 'range')
+ if int_or_none(step) == 0:
+ raise ValueError(f'Step in {segment!r} cannot be zero')
+ yield slice(int_or_none(start), float_or_none(end), int_or_none(step)) if has_range else int(start)
+
+ def get_requested_items(self):
+ playlist_items = self.ydl.params.get('playlist_items')
+ playlist_start = self.ydl.params.get('playliststart', 1)
+ playlist_end = self.ydl.params.get('playlistend')
+ # For backwards compatibility, interpret -1 as whole list
+ if playlist_end in (-1, None):
+ playlist_end = ''
+ if not playlist_items:
+ playlist_items = f'{playlist_start}:{playlist_end}'
+ elif playlist_start != 1 or playlist_end:
+ self.ydl.report_warning('Ignoring playliststart and playlistend because playlistitems was given', only_once=True)
+
+ for index in self.parse_playlist_items(playlist_items):
+ for i, entry in self[index]:
+ yield i, entry
+ try:
+ # TODO: Add auto-generated fields
+ self.ydl._match_entry(entry, incomplete=True, silent=True)
+ except (ExistingVideoReached, RejectedVideoReached):
+ return
+
+ def get_full_count(self):
+ if self.is_exhausted and not self.is_incomplete:
+ return len(self)
+ elif isinstance(self._entries, InAdvancePagedList):
+ if self._entries._pagesize == 1:
+ return self._entries._pagecount
+
+ @functools.cached_property
+ def _getter(self):
+ if isinstance(self._entries, list):
+ def get_entry(i):
+ try:
+ entry = self._entries[i]
+ except IndexError:
+ entry = self.MissingEntry
+ if not self.is_incomplete:
+ raise self.IndexError()
+ if entry is self.MissingEntry:
+ raise EntryNotInPlaylist(f'Entry {i} cannot be found')
+ return entry
+ else:
+ def get_entry(i):
+ try:
+ return type(self.ydl)._handle_extraction_exceptions(lambda _, i: self._entries[i])(self.ydl, i)
+ except (LazyList.IndexError, PagedList.IndexError):
+ raise self.IndexError()
+ return get_entry
+
+ def __getitem__(self, idx):
+ if isinstance(idx, int):
+ idx = slice(idx, idx)
+
+ # NB: PlaylistEntries[1:10] => (0, 1, ... 9)
+ step = 1 if idx.step is None else idx.step
+ if idx.start is None:
+ start = 0 if step > 0 else len(self) - 1
+ else:
+ start = idx.start - 1 if idx.start >= 0 else len(self) + idx.start
+
+ # NB: Do not call len(self) when idx == [:]
+ if idx.stop is None:
+ stop = 0 if step < 0 else float('inf')
+ else:
+ stop = idx.stop - 1 if idx.stop >= 0 else len(self) + idx.stop
+ stop += [-1, 1][step > 0]
+
+ for i in frange(start, stop, step):
+ if i < 0:
+ continue
+ try:
+ entry = self._getter(i)
+ except self.IndexError:
+ self.is_exhausted = True
+ if step > 0:
+ break
+ continue
+ yield i + 1, entry
+
+ def __len__(self):
+ return len(tuple(self[:]))
+
+ class IndexError(IndexError):
+ pass
+
+
def uppercase_escape(s):
unicode_escape = codecs.getdecoder('unicode_escape')
return re.sub(
for k, v in data.items():
out += b'--' + boundary.encode('ascii') + b'\r\n'
if isinstance(k, compat_str):
- k = k.encode('utf-8')
+ k = k.encode()
if isinstance(v, compat_str):
- v = v.encode('utf-8')
+ v = v.encode()
# RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
# suggests sending UTF-8 directly. Firefox sends UTF-8, too
content = b'Content-Disposition: form-data; name="' + k + b'"\r\n\r\n' + v + b'\r\n'
def parse_age_limit(s):
# isinstance(False, int) is True. So type() must be used instead
- if type(s) is int:
+ if type(s) is int: # noqa: E721
return s if 0 <= s <= 21 else None
elif not isinstance(s, str):
return None
return '"%s"' % v
+ def create_map(mobj):
+ return json.dumps(dict(json.loads(js_to_json(mobj.group(1) or '[]', vars=vars))))
+
code = re.sub(r'new Date\((".+")\)', r'\g<1>', code)
+ code = re.sub(r'new Map\((\[.*?\])?\)', create_map, code)
return re.sub(r'''(?sx)
"(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
return q
-POSTPROCESS_WHEN = ('pre_process', 'after_filter', 'before_dl', 'after_move', 'post_process', 'after_video', 'playlist')
+POSTPROCESS_WHEN = ('pre_process', 'after_filter', 'before_dl', 'post_process', 'after_move', 'after_video', 'playlist')
DEFAULT_OUTTMPL = {
(b'\xff\xfe', 'utf-16-le'),
(b'\xfe\xff', 'utf-16-be'),
]
+
+ encoding = 'utf-8'
for bom, enc in BOMS:
- if first_bytes.startswith(bom):
- s = first_bytes[len(bom):].decode(enc, 'replace')
- break
- else:
- s = first_bytes.decode('utf-8', 'replace')
+ while first_bytes.startswith(bom):
+ encoding, first_bytes = enc, first_bytes[len(bom):]
- return re.match(r'^\s*<', s)
+ return re.match(r'^\s*<', first_bytes.decode(encoding, 'replace'))
def determine_protocol(info_dict):
else:
is_incomplete = lambda k: k in incomplete
- operator_rex = re.compile(r'''(?x)\s*
+ operator_rex = re.compile(r'''(?x)
(?P<key>[a-z_]+)
\s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
(?:
(?P<quote>["\'])(?P<quotedstrval>.+?)(?P=quote)|
(?P<strval>.+?)
)
- \s*$
''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
- m = operator_rex.search(filter_part)
+ m = operator_rex.fullmatch(filter_part.strip())
if m:
m = m.groupdict()
unnegated_op = COMPARISON_OPERATORS[m['op']]
'': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
'!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
}
- operator_rex = re.compile(r'''(?x)\s*
+ operator_rex = re.compile(r'''(?x)
(?P<op>%s)\s*(?P<key>[a-z_]+)
- \s*$
''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
- m = operator_rex.search(filter_part)
+ m = operator_rex.fullmatch(filter_part.strip())
if m:
op = UNARY_OPERATORS[m.group('op')]
actual_value = dct.get(m.group('key'))
return _match_func
+def download_range_func(chapters, ranges):
+ def inner(info_dict, ydl):
+ warning = ('There are no chapters matching the regex' if info_dict.get('chapters')
+ else 'Cannot match chapters since chapter information is unavailable')
+ for regex in chapters or []:
+ for i, chapter in enumerate(info_dict.get('chapters') or []):
+ if re.search(regex, chapter['title']):
+ warning = None
+ yield {**chapter, 'index': i}
+ if chapters and warning:
+ ydl.to_screen(f'[info] {info_dict["id"]}: {warning}')
+
+ yield from ({'start_time': start, 'end_time': end} for start, end in ranges or [])
+
+ return inner
+
+
def parse_dfxp_time_expr(time_expr):
if not time_expr:
return
return ''.join(out)
-def cli_option(params, command_option, param):
+def cli_option(params, command_option, param, separator=None):
param = params.get(param)
- if param:
- param = compat_str(param)
- return [command_option, param] if param is not None else []
+ return ([] if param is None
+ else [command_option, str(param)] if separator is None
+ else [f'{command_option}{separator}{param}'])
def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
param = params.get(param)
- if param is None:
- return []
- assert isinstance(param, bool)
- if separator:
- return [command_option + separator + (true_value if param else false_value)]
- return [command_option, true_value if param else false_value]
+ assert param in (True, False, None)
+ return cli_option({True: true_value, False: false_value}, command_option, param, separator)
def cli_valueless_option(params, command_option, param, expected_value=True):
- param = params.get(param)
- return [command_option] if param == expected_value else []
+ return [command_option] if params.get(param) == expected_value else []
def cli_configuration_args(argdict, keys, default=[], use_compat=True):
'YE': 'Yemen',
'ZM': 'Zambia',
'ZW': 'Zimbabwe',
+ # Not ISO 3166 codes, but used for IP blocks
+ 'AP': 'Asia/Pacific Region',
+ 'EU': 'Europe',
}
@classmethod
return [0, 2] + pseudo_random + [0] + data
-def encode_base_n(num, n, table=None):
- FULL_TABLE = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
- if not table:
- table = FULL_TABLE[:n]
+def _base_n_table(n, table):
+ if not table and not n:
+ raise ValueError('Either table or n must be specified')
+ table = (table or '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ')[:n]
+
+ if n != len(table):
+ raise ValueError(f'base {n} exceeds table length {len(table)}')
+ return table
- if n > len(table):
- raise ValueError('base %d exceeds table length %d' % (n, len(table)))
- if num == 0:
+def encode_base_n(num, n=None, table=None):
+ """Convert given int to a base-n string"""
+ table = _base_n_table(n, table)
+ if not num:
return table[0]
- ret = ''
+ result, base = '', len(table)
while num:
- ret = table[num % n] + ret
- num = num // n
- return ret
+ result = table[num % base] + result
+ num = num // base
+ return result
+
+
+def decode_base_n(string, n=None, table=None):
+ """Convert given base-n string to int"""
+ table = {char: index for index, char in enumerate(_base_n_table(n, table))}
+ result, base = 0, len(table)
+ for char in string:
+ result = result * base + table[char]
+ return result
+
+
+def decode_base(value, digits):
+ write_string('DeprecationWarning: yt_dlp.utils.decode_base is deprecated '
+ 'and may be removed in a future version. Use yt_dlp.decode_base_n instead')
+ return decode_base_n(value, table=digits)
def decode_packed_codes(code):
'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
+ ('"xattr" binary' if sys.platform != 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
- value = value.decode('utf-8')
+ value = value.decode()
try:
- p = Popen(
+ _, stderr, returncode = Popen.run(
[exe, '-w', key, value, path] if exe == 'xattr' else [exe, '-n', key, '-v', value, path],
- stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)
+ text=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)
except OSError as e:
raise XAttrMetadataError(e.errno, e.strerror)
- stderr = p.communicate_or_kill()[1].decode('utf-8', 'replace')
- if p.returncode:
- raise XAttrMetadataError(p.returncode, stderr)
+ if returncode:
+ raise XAttrMetadataError(returncode, stderr)
def random_birthday(year_field, month_field, day_field):
net_location += ':' + urllib.parse.quote(iri_parts.password, safe=r"!$%&'()*+,~")
net_location += '@'
- net_location += iri_parts.hostname.encode('idna').decode('utf-8') # Punycode for Unicode hostnames.
+ net_location += iri_parts.hostname.encode('idna').decode() # Punycode for Unicode hostnames.
# The 'idna' encoding produces ASCII text.
if iri_parts.port is not None and iri_parts.port != 80:
net_location += ':' + str(iri_parts.port)
return path
-def format_field(obj, field=None, template='%s', ignore=(None, ''), default='', func=None):
+def format_field(obj, field=None, template='%s', ignore=NO_DEFAULT, default='', func=IDENTITY):
val = traverse_obj(obj, *variadic(field))
- if val in ignore:
+ if (not val and val != 0) if ignore is NO_DEFAULT else val in variadic(ignore):
return default
- return template % (func(val) if func else val)
+ return template % func(val)
def clean_podcast_url(url):
def get_executable_path():
- from zipimport import zipimporter
- if hasattr(sys, 'frozen'): # Running from PyInstaller
- path = os.path.dirname(sys.executable)
- elif isinstance(__loader__, zipimporter): # Running from ZIP
- path = os.path.join(os.path.dirname(__file__), '../..')
- else:
- path = os.path.join(os.path.dirname(__file__), '..')
- return os.path.abspath(path)
+ from .update import _get_variant_and_executable_path
+
+ return os.path.dirname(os.path.abspath(_get_variant_and_executable_path()[1]))
def load_plugins(name, suffix, namespace):
if isinstance(expected_type, type):
type_test = lambda val: val if isinstance(val, expected_type) else None
- elif expected_type is not None:
- type_test = expected_type
else:
- type_test = lambda val: val
+ type_test = expected_type or IDENTITY
for path in path_list:
depth = 0
return x if isinstance(x, collections.abc.Iterable) and not isinstance(x, allowed_types) else (x,)
-def decode_base(value, digits):
- # This will convert given base-x string to scalar (long or int)
- table = {char: index for index, char in enumerate(digits)}
- result = 0
- base = len(digits)
- for chr in value:
- result *= base
- result += table[chr]
- return result
-
-
def time_seconds(**kwargs):
t = datetime.datetime.now(datetime.timezone(datetime.timedelta(**kwargs)))
return t.timestamp()
}
if headers:
header_data.update(headers)
- header_b64 = base64.b64encode(json.dumps(header_data).encode('utf-8'))
- payload_b64 = base64.b64encode(json.dumps(payload_data).encode('utf-8'))
- h = hmac.new(key.encode('utf-8'), header_b64 + b'.' + payload_b64, hashlib.sha256)
+ header_b64 = base64.b64encode(json.dumps(header_data).encode())
+ payload_b64 = base64.b64encode(json.dumps(payload_data).encode())
+ h = hmac.new(key.encode(), header_b64 + b'.' + payload_b64, hashlib.sha256)
signature_b64 = base64.b64encode(h.digest())
token = header_b64 + b'.' + payload_b64 + b'.' + signature_b64
return token
return payload_data
+WINDOWS_VT_MODE = False if compat_os_name == 'nt' else None
+
+
+@functools.cache
def supports_terminal_sequences(stream):
if compat_os_name == 'nt':
- from .compat import WINDOWS_VT_MODE # Must be imported locally
- if not WINDOWS_VT_MODE or get_windows_version() < (10, 0, 10586):
+ if not WINDOWS_VT_MODE:
return False
elif not os.getenv('TERM'):
return False
return False
+def windows_enable_vt_mode(): # TODO: Do this the proper way https://bugs.python.org/issue30075
+ if get_windows_version() < (10, 0, 10586):
+ return
+ global WINDOWS_VT_MODE
+ try:
+ Popen.run('', shell=True)
+ except Exception:
+ return
+
+ WINDOWS_VT_MODE = True
+ supports_terminal_sequences.cache_clear()
+
+
_terminal_sequences_re = re.compile('\033\\[[^m]+m')
def join_nonempty(*values, delim='-', from_dict=None):
if from_dict is not None:
- values = map(from_dict.get, values)
+ values = (traverse_obj(from_dict, variadic(v)) for v in values)
return delim.join(map(str, filter(None, values)))
return int(crg.group(1)), int_or_none(crg.group(2)), int_or_none(crg.group(3))
+def read_stdin(what):
+ eof = 'Ctrl+Z' if compat_os_name == 'nt' else 'Ctrl+D'
+ write_string(f'Reading {what} from STDIN - EOF ({eof}) to end:\n')
+ return sys.stdin
+
+
class Config:
own_args = None
+ parsed_args = None
filename = None
__initialized = False
def __init__(self, parser, label=None):
- self._parser, self.label = parser, label
+ self.parser, self.label = parser, label
self._loaded_paths, self.configs = set(), []
def init(self, args=None, filename=None):
return False
self._loaded_paths.add(location)
- self.__initialized = True
- self.own_args, self.filename = args, filename
- for location in self._parser.parse_args(args)[0].config_locations or []:
+ self.own_args, self.__initialized = args, True
+ opts, _ = self.parser.parse_known_args(args)
+ self.parsed_args, self.filename = args, filename
+
+ for location in opts.config_locations or []:
+ if location == '-':
+ self.append_config(shlex.split(read_stdin('options'), comments=True), label='stdin')
+ continue
location = os.path.join(directory, expand_path(location))
if os.path.isdir(location):
location = os.path.join(location, 'yt-dlp.conf')
if not os.path.exists(location):
- self._parser.error(f'config location {location} does not exist')
+ self.parser.error(f'config location {location} does not exist')
self.append_config(self.read_file(location), location)
return True
# FIXME: https://github.com/ytdl-org/youtube-dl/commit/dfe5fa49aed02cf36ba9f743b11b0903554b5e56
contents = optionf.read()
res = shlex.split(contents, comments=True)
+ except Exception as err:
+ raise ValueError(f'Unable to parse "{filename}": {err}')
finally:
optionf.close()
return res
return opts
def append_config(self, *args, label=None):
- config = type(self)(self._parser, label)
+ config = type(self)(self.parser, label)
config._loaded_paths = self._loaded_paths
if config.init(*args):
self.configs.append(config)
def all_args(self):
for config in reversed(self.configs):
yield from config.all_args
- yield from self.own_args or []
+ yield from self.parsed_args or []
+
+ def parse_known_args(self, **kwargs):
+ return self.parser.parse_known_args(self.all_args, **kwargs)
def parse_args(self):
- return self._parser.parse_args(self.all_args)
+ return self.parser.parse_args(self.all_args)
class WebSocketsWrapper():
class classproperty:
- def __init__(self, f):
- self.f = f
+ """classmethod(property(func)) that works in py < 3.9"""
+
+ def __init__(self, func):
+ functools.update_wrapper(self, func)
+ self.func = func
def __get__(self, _, cls):
- return self.f(cls)
+ return self.func(cls)
+
+class Namespace(types.SimpleNamespace):
+ """Immutable namespace"""
-def Namespace(**kwargs):
- return collections.namedtuple('Namespace', kwargs)(**kwargs)
+ def __iter__(self):
+ return iter(self.__dict__.values())
+
+ @property
+ def items_(self):
+ return self.__dict__.items()
# Deprecated
projects / yt-dlp.git / blobdiff