Merge a4759ae8ea18bc13cb4342d8ec5d498a8a309e4a into 420d53387cff54ea1fccca061438d59bdb50a39c

* from yt-dlp/yt-dlp#12313
* also fix d7c2708

test/test_jsinterp.py

@@ -11,6 +11,7 @@ sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
 import math
 import re
+import time
 
 from youtube_dl.compat import compat_str as str
 from youtube_dl.jsinterp import JS_Undefined, JSInterpreter
@@ -208,6 +209,34 @@ class TestJSInterpreter(unittest.TestCase):
         self._test(jsi, 86000, args=['12/31/1969 18:01:26 MDT'])
         # epoch 0
         self._test(jsi, 0, args=['1 January 1970 00:00:00 UTC'])
+        # undefined
+        self._test(jsi, NaN, args=[JS_Undefined])
+        # y,m,d, ... - may fail with older dates lacking DST data
+        jsi = JSInterpreter(
+            'function f() { return new Date(%s); }'
+            % ('2024, 5, 29, 2, 52, 12, 42',))
+        self._test(jsi, (
+            1719625932042                           # UK value
+            + (
+                + 3600                              # back to GMT
+                + (time.altzone if time.daylight    # host's DST
+                   else time.timezone)
+            ) * 1000))
+        # no arg
+        self.assertAlmostEqual(JSInterpreter(
+            'function f() { return new Date() - 0; }').call_function('f'),
+            time.time() * 1000, delta=100)
+        # Date.now()
+        self.assertAlmostEqual(JSInterpreter(
+            'function f() { return Date.now(); }').call_function('f'),
+            time.time() * 1000, delta=100)
+        # Date.parse()
+        jsi = JSInterpreter('function f(dt) { return Date.parse(dt); }')
+        self._test(jsi, 0, args=['1 January 1970 00:00:00 UTC'])
+        # Date.UTC()
+        jsi = JSInterpreter('function f() { return Date.UTC(%s); }'
+                            % ('1970, 0, 1, 0, 0, 0, 0',))
+        self._test(jsi, 0)
 
     def test_call(self):
         jsi = JSInterpreter('''
@@ -463,6 +492,14 @@ class TestJSInterpreter(unittest.TestCase):
         self._test('function f(){return NaN << 42}', 0)
         self._test('function f(){return "21.9" << 1}', 42)
         self._test('function f(){return 21 << 4294967297}', 42)
+        self._test('function f(){return true << "5";}', 32)
+        self._test('function f(){return true << true;}', 2)
+        self._test('function f(){return "19" & "21.9";}', 17)
+        self._test('function f(){return "19" & false;}', 0)
+        self._test('function f(){return "11.0" >> "2.1";}', 2)
+        self._test('function f(){return 5 ^ 9;}', 12)
+        self._test('function f(){return 0.0 << NaN}', 0)
+        self._test('function f(){return null << undefined}', 0)
 
     def test_negative(self):
         self._test('function f(){return 2    *    -2.0    ;}', -4)

test/test_youtube_signature.py

@@ -223,6 +223,18 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/9c6dfc4a/player_ias.vflset/en_US/base.js',
         'jbu7ylIosQHyJyJV', 'uwI0ESiynAmhNg',
     ),
+    (
+        'https://www.youtube.com/s/player/f6e09c70/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
+    ),
+    (
+        'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
+    ),
+    (
+        'https://www.youtube.com/s/player/91201489/player_ias_tce.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'U48vOZHaeYS6vO',
+    ),
 ]
 
 
@@ -284,7 +296,7 @@ def t_factory(name, sig_func, url_pattern):
 
 
 def signature(jscode, sig_input):
-    func = YoutubeIE(FakeYDL())._parse_sig_js(jscode)
+    func = YoutubeIE(FakeYDL({'cachedir': False}))._parse_sig_js(jscode)
     src_sig = (
         compat_str(string.printable[:sig_input])
         if isinstance(sig_input, int) else sig_input)
@@ -292,9 +304,10 @@ def signature(jscode, sig_input):
 
 
 def n_sig(jscode, sig_input):
-    funcname = YoutubeIE(FakeYDL())._extract_n_function_name(jscode)
+    ie = YoutubeIE(FakeYDL({'cachedir': False}))
-    return JSInterpreter(jscode).call_function(
+    jsi = JSInterpreter(jscode)
-        funcname, sig_input, _ytdl_do_not_return=sig_input)
+    jsi, _, func_code = ie._extract_n_function_code_jsi(sig_input, jsi)
+    return ie._extract_n_function_from_code(jsi, func_code)(sig_input)
 
 
 make_sig_test = t_factory(

youtube_dl/__init__.py

@@ -18,7 +18,7 @@ from .compat import (
     compat_getpass,
     compat_register_utf8,
     compat_shlex_split,
-    workaround_optparse_bug9161,
+    _workaround_optparse_bug9161,
 )
 from .utils import (
     _UnsafeExtensionError,
@@ -50,7 +50,7 @@ def _real_main(argv=None):
     # Compatibility fix for Windows
     compat_register_utf8()
 
-    workaround_optparse_bug9161()
+    _workaround_optparse_bug9161()
 
     setproctitle('youtube-dl')
 

youtube_dl/compat.py

@@ -16,7 +16,6 @@ import os
 import platform
 import re
 import shlex
-import shutil
 import socket
 import struct
 import subprocess
@@ -24,11 +23,15 @@ import sys
 import types
 import xml.etree.ElementTree
 
+_IDENTITY = lambda x: x
+
 # naming convention
 # 'compat_' + Python3_name.replace('.', '_')
 # other aliases exist for convenience and/or legacy
+# wrap disposable test values in type() to reclaim storage
 
-# deal with critical unicode/str things first
+# deal with critical unicode/str things first:
+# compat_str, compat_basestring, compat_chr
 try:
     # Python 2
     compat_str, compat_basestring, compat_chr = (
@@ -39,18 +42,23 @@ except NameError:
         str, (str, bytes), chr
     )
 
-# casefold
+
+# compat_casefold
 try:
     compat_str.casefold
     compat_casefold = lambda s: s.casefold()
 except AttributeError:
     from .casefold import _casefold as compat_casefold
 
+
+# compat_collections_abc
 try:
     import collections.abc as compat_collections_abc
 except ImportError:
     import collections as compat_collections_abc
 
+
+# compat_urllib_request
 try:
     import urllib.request as compat_urllib_request
 except ImportError:  # Python 2
@@ -79,11 +87,15 @@ except TypeError:
     _add_init_method_arg(compat_urllib_request.Request)
     del _add_init_method_arg
 
+
+# compat_urllib_error
 try:
     import urllib.error as compat_urllib_error
 except ImportError:  # Python 2
     import urllib2 as compat_urllib_error
 
+
+# compat_urllib_parse
 try:
     import urllib.parse as compat_urllib_parse
 except ImportError:  # Python 2
@@ -98,17 +110,23 @@ except ImportError:  # Python 2
 compat_urlparse = compat_urllib_parse
 compat_urllib_parse_urlparse = compat_urllib_parse.urlparse
 
+
+# compat_urllib_response
 try:
     import urllib.response as compat_urllib_response
 except ImportError:  # Python 2
     import urllib as compat_urllib_response
 
+
+# compat_urllib_response.addinfourl
 try:
     compat_urllib_response.addinfourl.status
 except AttributeError:
     # .getcode() is deprecated in Py 3.
     compat_urllib_response.addinfourl.status = property(lambda self: self.getcode())
 
+
+# compat_http_cookiejar
 try:
     import http.cookiejar as compat_cookiejar
 except ImportError:  # Python 2
@@ -127,12 +145,16 @@ else:
     compat_cookiejar_Cookie = compat_cookiejar.Cookie
 compat_http_cookiejar_Cookie = compat_cookiejar_Cookie
 
+
+# compat_http_cookies
 try:
     import http.cookies as compat_cookies
 except ImportError:  # Python 2
     import Cookie as compat_cookies
 compat_http_cookies = compat_cookies
 
+
+# compat_http_cookies_SimpleCookie
 if sys.version_info[0] == 2 or sys.version_info < (3, 3):
     class compat_cookies_SimpleCookie(compat_cookies.SimpleCookie):
         def load(self, rawdata):
@@ -155,11 +177,15 @@ else:
     compat_cookies_SimpleCookie = compat_cookies.SimpleCookie
 compat_http_cookies_SimpleCookie = compat_cookies_SimpleCookie
 
+
+# compat_html_entities, probably useless now
 try:
     import html.entities as compat_html_entities
 except ImportError:  # Python 2
     import htmlentitydefs as compat_html_entities
 
+
+# compat_html_entities_html5
 try:  # Python >= 3.3
     compat_html_entities_html5 = compat_html_entities.html5
 except AttributeError:
@@ -2408,18 +2434,24 @@ except AttributeError:
     # Py < 3.1
     compat_http_client.HTTPResponse.getcode = lambda self: self.status
 
+
+# compat_urllib_HTTPError
 try:
     from urllib.error import HTTPError as compat_HTTPError
 except ImportError:  # Python 2
     from urllib2 import HTTPError as compat_HTTPError
 compat_urllib_HTTPError = compat_HTTPError
 
+
+# compat_urllib_request_urlretrieve
 try:
     from urllib.request import urlretrieve as compat_urlretrieve
 except ImportError:  # Python 2
     from urllib import urlretrieve as compat_urlretrieve
 compat_urllib_request_urlretrieve = compat_urlretrieve
 
+
+# compat_html_parser_HTMLParser, compat_html_parser_HTMLParseError
 try:
     from HTMLParser import (
         HTMLParser as compat_HTMLParser,
@@ -2432,22 +2464,33 @@ except ImportError:  # Python 3
         # HTMLParseError was deprecated in Python 3.3 and removed in
         # Python 3.5. Introducing dummy exception for Python >3.5 for compatible
         # and uniform cross-version exception handling
+
         class compat_HTMLParseError(Exception):
             pass
+
 compat_html_parser_HTMLParser = compat_HTMLParser
 compat_html_parser_HTMLParseError = compat_HTMLParseError
 
+
+# compat_subprocess_get_DEVNULL
 try:
     _DEVNULL = subprocess.DEVNULL
     compat_subprocess_get_DEVNULL = lambda: _DEVNULL
 except AttributeError:
     compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
 
+
+# compat_http_server
 try:
     import http.server as compat_http_server
 except ImportError:
     import BaseHTTPServer as compat_http_server
 
+
+# compat_urllib_parse_unquote_to_bytes,
+# compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus,
+# compat_urllib_parse_urlencode,
+# compat_urllib_parse_parse_qs
 try:
     from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes
     from urllib.parse import unquote as compat_urllib_parse_unquote
@@ -2598,6 +2641,8 @@ except ImportError:  # Python 2
 
 compat_urllib_parse_parse_qs = compat_parse_qs
 
+
+# compat_urllib_request_DataHandler
 try:
     from urllib.request import DataHandler as compat_urllib_request_DataHandler
 except ImportError:  # Python < 3.4
@@ -2632,16 +2677,20 @@ except ImportError:  # Python < 3.4
 
             return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url)
 
+
+# compat_xml_etree_ElementTree_ParseError
 try:
     from xml.etree.ElementTree import ParseError as compat_xml_parse_error
 except ImportError:  # Python 2.6
     from xml.parsers.expat import ExpatError as compat_xml_parse_error
 compat_xml_etree_ElementTree_ParseError = compat_xml_parse_error
 
-etree = xml.etree.ElementTree
+
+# compat_xml_etree_ElementTree_Element
+_etree = xml.etree.ElementTree
 
 
-class _TreeBuilder(etree.TreeBuilder):
+class _TreeBuilder(_etree.TreeBuilder):
     def doctype(self, name, pubid, system):
         pass
 
@@ -2650,7 +2699,7 @@ try:
     # xml.etree.ElementTree.Element is a method in Python <=2.6 and
     # the following will crash with:
     #  TypeError: isinstance() arg 2 must be a class, type, or tuple of classes and types
-    isinstance(None, etree.Element)
+    isinstance(None, _etree.Element)
     from xml.etree.ElementTree import Element as compat_etree_Element
 except TypeError:  # Python <=2.6
     from xml.etree.ElementTree import _ElementInterface as compat_etree_Element
@@ -2658,12 +2707,12 @@ compat_xml_etree_ElementTree_Element = compat_etree_Element
 
 if sys.version_info[0] >= 3:
     def compat_etree_fromstring(text):
-        return etree.XML(text, parser=etree.XMLParser(target=_TreeBuilder()))
+        return _etree.XML(text, parser=_etree.XMLParser(target=_TreeBuilder()))
 else:
     # python 2.x tries to encode unicode strings with ascii (see the
     # XMLParser._fixtext method)
     try:
-        _etree_iter = etree.Element.iter
+        _etree_iter = _etree.Element.iter
     except AttributeError:  # Python <=2.6
         def _etree_iter(root):
             for el in root.findall('*'):
@@ -2675,27 +2724,29 @@ else:
     # 2.7 source
     def _XML(text, parser=None):
         if not parser:
-            parser = etree.XMLParser(target=_TreeBuilder())
+            parser = _etree.XMLParser(target=_TreeBuilder())
         parser.feed(text)
         return parser.close()
 
     def _element_factory(*args, **kwargs):
-        el = etree.Element(*args, **kwargs)
+        el = _etree.Element(*args, **kwargs)
         for k, v in el.items():
             if isinstance(v, bytes):
                 el.set(k, v.decode('utf-8'))
         return el
 
     def compat_etree_fromstring(text):
-        doc = _XML(text, parser=etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory)))
+        doc = _XML(text, parser=_etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory)))
         for el in _etree_iter(doc):
             if el.text is not None and isinstance(el.text, bytes):
                 el.text = el.text.decode('utf-8')
         return doc
 
-if hasattr(etree, 'register_namespace'):
+
-    compat_etree_register_namespace = etree.register_namespace
+# compat_xml_etree_register_namespace
-else:
+try:
+    compat_etree_register_namespace = _etree.register_namespace
+except AttributeError:
     def compat_etree_register_namespace(prefix, uri):
         """Register a namespace prefix.
         The registry is global, and any existing mapping for either the
@@ -2704,14 +2755,16 @@ else:
         attributes in this namespace will be serialized with prefix if possible.
         ValueError is raised if prefix is reserved or is invalid.
         """
-        if re.match(r"ns\d+$", prefix):
+        if re.match(r'ns\d+$', prefix):
-            raise ValueError("Prefix format reserved for internal use")
+            raise ValueError('Prefix format reserved for internal use')
-        for k, v in list(etree._namespace_map.items()):
+        for k, v in list(_etree._namespace_map.items()):
             if k == uri or v == prefix:
-                del etree._namespace_map[k]
+                del _etree._namespace_map[k]
-        etree._namespace_map[uri] = prefix
+        _etree._namespace_map[uri] = prefix
 compat_xml_etree_register_namespace = compat_etree_register_namespace
 
+
+# compat_xpath, compat_etree_iterfind
 if sys.version_info < (2, 7):
     # Here comes the crazy part: In 2.6, if the xpath is a unicode,
     # .//node does not match if a node is a direct child of . !
@@ -2898,7 +2951,6 @@ if sys.version_info < (2, 7):
         def __init__(self, root):
             self.root = root
 
-    ##
     # Generate all matching objects.
 
     def compat_etree_iterfind(elem, path, namespaces=None):
@@ -2933,13 +2985,15 @@ if sys.version_info < (2, 7):
 
 
 else:
-    compat_xpath = lambda xpath: xpath
     compat_etree_iterfind = lambda element, match: element.iterfind(match)
+    compat_xpath = _IDENTITY
 
 
+# compat_os_name
 compat_os_name = os._name if os.name == 'java' else os.name
 
 
+# compat_shlex_quote
 if compat_os_name == 'nt':
     def compat_shlex_quote(s):
         return s if re.match(r'^[-_\w./]+$', s) else '"%s"' % s.replace('"', '\\"')
@@ -2954,6 +3008,7 @@ else:
                 return "'" + s.replace("'", "'\"'\"'") + "'"
 
 
+# compat_shlex.split
 try:
     args = shlex.split('中文')
     assert (isinstance(args, list)
@@ -2969,6 +3024,7 @@ except (AssertionError, UnicodeEncodeError):
         return list(map(lambda s: s.decode('utf-8'), shlex.split(s, comments, posix)))
 
 
+# compat_ord
 def compat_ord(c):
     if isinstance(c, int):
         return c
@@ -2976,6 +3032,7 @@ def compat_ord(c):
         return ord(c)
 
 
+# compat_getenv, compat_os_path_expanduser, compat_setenv
 if sys.version_info >= (3, 0):
     compat_getenv = os.getenv
     compat_expanduser = os.path.expanduser
@@ -3063,6 +3120,7 @@ else:
 compat_os_path_expanduser = compat_expanduser
 
 
+# compat_os_path_realpath
 if compat_os_name == 'nt' and sys.version_info < (3, 8):
     # os.path.realpath on Windows does not follow symbolic links
     # prior to Python 3.8 (see https://bugs.python.org/issue9949)
@@ -3076,6 +3134,7 @@ else:
 compat_os_path_realpath = compat_realpath
 
 
+# compat_print
 if sys.version_info < (3, 0):
     def compat_print(s):
         from .utils import preferredencoding
@@ -3086,6 +3145,7 @@ else:
         print(s)
 
 
+# compat_getpass_getpass
 if sys.version_info < (3, 0) and sys.platform == 'win32':
     def compat_getpass(prompt, *args, **kwargs):
         if isinstance(prompt, compat_str):
@@ -3098,22 +3158,22 @@ else:
 compat_getpass_getpass = compat_getpass
 
 
+# compat_input
 try:
     compat_input = raw_input
 except NameError:  # Python 3
     compat_input = input
 
 
+# compat_kwargs
 # Python < 2.6.5 require kwargs to be bytes
 try:
-    def _testfunc(x):
+    (lambda x: x)(**{'x': 0})
-        pass
-    _testfunc(**{'x': 0})
 except TypeError:
     def compat_kwargs(kwargs):
         return dict((bytes(k), v) for k, v in kwargs.items())
 else:
-    compat_kwargs = lambda kwargs: kwargs
+    compat_kwargs = _IDENTITY
 
 
 # compat_numeric_types
@@ -3132,6 +3192,8 @@ except NameError:  # Python 3
 # compat_int
 compat_int = compat_integer_types[-1]
 
+
+# compat_socket_create_connection
 if sys.version_info < (2, 7):
     def compat_socket_create_connection(address, timeout, source_address=None):
         host, port = address
@@ -3158,6 +3220,7 @@ else:
     compat_socket_create_connection = socket.create_connection
 
 
+# compat_contextlib_suppress
 try:
     from contextlib import suppress as compat_contextlib_suppress
 except ImportError:
@@ -3200,12 +3263,12 @@ except AttributeError:
                         # repeated .close() is OK, but just in case
                         with compat_contextlib_suppress(EnvironmentError):
                             f.close()
-                popen.wait()
+            popen.wait()
 
 
 # Fix https://github.com/ytdl-org/youtube-dl/issues/4223
 # See http://bugs.python.org/issue9161 for what is broken
-def workaround_optparse_bug9161():
+def _workaround_optparse_bug9161():
     op = optparse.OptionParser()
     og = optparse.OptionGroup(op, 'foo')
     try:
@@ -3224,9 +3287,10 @@ def workaround_optparse_bug9161():
         optparse.OptionGroup.add_option = _compat_add_option
 
 
-if hasattr(shutil, 'get_terminal_size'):  # Python >= 3.3
+# compat_shutil_get_terminal_size
-    compat_get_terminal_size = shutil.get_terminal_size
+try:
-else:
+    from shutil import get_terminal_size as compat_get_terminal_size  # Python >= 3.3
+except ImportError:
     _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
 
     def compat_get_terminal_size(fallback=(80, 24)):
@@ -3256,27 +3320,33 @@ else:
                 columns = _columns
             if lines is None or lines <= 0:
                 lines = _lines
+
         return _terminal_size(columns, lines)
 
+compat_shutil_get_terminal_size = compat_get_terminal_size
 
+
+# compat_itertools_count
 try:
-    itertools.count(start=0, step=1)
+    type(itertools.count(start=0, step=1))
     compat_itertools_count = itertools.count
-except TypeError:  # Python 2.6
+except TypeError:  # Python 2.6 lacks step
     def compat_itertools_count(start=0, step=1):
         while True:
             yield start
             start += step
 
 
+# compat_tokenize_tokenize
 if sys.version_info >= (3, 0):
     from tokenize import tokenize as compat_tokenize_tokenize
 else:
     from tokenize import generate_tokens as compat_tokenize_tokenize
 
 
+# compat_struct_pack, compat_struct_unpack, compat_Struct
 try:
-    struct.pack('!I', 0)
+    type(struct.pack('!I', 0))
 except TypeError:
     # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument
     # See https://bugs.python.org/issue19099
@@ -3308,8 +3378,10 @@ else:
         compat_Struct = struct.Struct
 
 
-# compat_map/filter() returning an iterator, supposedly the
+# builtins returning an iterator
-# same versioning as for zip below
+
+# compat_map, compat_filter
+# supposedly the same versioning as for zip below
 try:
     from future_builtins import map as compat_map
 except ImportError:
@@ -3326,6 +3398,7 @@ except ImportError:
     except ImportError:
         compat_filter = filter
 
+# compat_zip
 try:
     from future_builtins import zip as compat_zip
 except ImportError:  # not 2.6+ or is 3.x
@@ -3335,6 +3408,7 @@ except ImportError:  # not 2.6+ or is 3.x
         compat_zip = zip
 
 
+# compat_itertools_zip_longest
 # method renamed between Py2/3
 try:
     from itertools import zip_longest as compat_itertools_zip_longest
@@ -3342,7 +3416,8 @@ except ImportError:
     from itertools import izip_longest as compat_itertools_zip_longest
 
 
-# new class in collections
+# compat_collections_chain_map
+# collections.ChainMap: new class
 try:
     from collections import ChainMap as compat_collections_chain_map
     # Py3.3's ChainMap is deficient
@@ -3398,19 +3473,22 @@ except ImportError:
         def new_child(self, m=None, **kwargs):
             m = m or {}
             m.update(kwargs)
-            return compat_collections_chain_map(m, *self.maps)
+            # support inheritance !
+            return type(self)(m, *self.maps)
 
         @property
         def parents(self):
-            return compat_collections_chain_map(*(self.maps[1:]))
+            return type(self)(*(self.maps[1:]))
 
 
+# compat_re_Pattern, compat_re_Match
 # Pythons disagree on the type of a pattern (RegexObject, _sre.SRE_Pattern, Pattern, ...?)
 compat_re_Pattern = type(re.compile(''))
 # and on the type of a match
 compat_re_Match = type(re.match('a', 'a'))
 
 
+# compat_base64_b64decode
 if sys.version_info < (3, 3):
     def compat_b64decode(s, *args, **kwargs):
         if isinstance(s, compat_str):
@@ -3422,6 +3500,7 @@ else:
 compat_base64_b64decode = compat_b64decode
 
 
+# compat_ctypes_WINFUNCTYPE
 if platform.python_implementation() == 'PyPy' and sys.pypy_version_info < (5, 4, 0):
     # PyPy2 prior to version 5.4.0 expects byte strings as Windows function
     # names, see the original PyPy issue [1] and the youtube-dl one [2].
@@ -3440,6 +3519,7 @@ else:
         return ctypes.WINFUNCTYPE(*args, **kwargs)
 
 
+# compat_open
 if sys.version_info < (3, 0):
     # open(file, mode='r', buffering=- 1, encoding=None, errors=None, newline=None, closefd=True) not: opener=None
     def compat_open(file_, *args, **kwargs):
@@ -3467,18 +3547,28 @@ except AttributeError:
     def compat_datetime_timedelta_total_seconds(td):
         return (td.microseconds + (td.seconds + td.days * 24 * 3600) * 10**6) / 10**6
 
+
 # optional decompression packages
+# compat_brotli
 # PyPi brotli package implements 'br' Content-Encoding
 try:
     import brotli as compat_brotli
 except ImportError:
     compat_brotli = None
+# compat_ncompress
 # PyPi ncompress package implements 'compress' Content-Encoding
 try:
     import ncompress as compat_ncompress
 except ImportError:
     compat_ncompress = None
 
+# compat_zstandard
+# PyPi zstandard package implements 'zstd' Content-Encoding (RFC 8878 7.2)
+try:
+    import zstandard as compat_zstandard
+except ImportError:
+    compat_zstandard = None
+
 
 legacy = [
     'compat_HTMLParseError',
@@ -3495,6 +3585,7 @@ legacy = [
     'compat_getpass',
     'compat_parse_qs',
     'compat_realpath',
+    'compat_shlex_split',
     'compat_urllib_parse_parse_qs',
     'compat_urllib_parse_unquote',
     'compat_urllib_parse_unquote_plus',
@@ -3508,8 +3599,6 @@ legacy = [
 
 
 __all__ = [
-    'compat_html_parser_HTMLParseError',
-    'compat_html_parser_HTMLParser',
     'compat_Struct',
     'compat_base64_b64decode',
     'compat_basestring',
@@ -3518,13 +3607,9 @@ __all__ = [
     'compat_chr',
     'compat_collections_abc',
     'compat_collections_chain_map',
-    'compat_datetime_timedelta_total_seconds',
-    'compat_http_cookiejar',
-    'compat_http_cookiejar_Cookie',
-    'compat_http_cookies',
-    'compat_http_cookies_SimpleCookie',
     'compat_contextlib_suppress',
     'compat_ctypes_WINFUNCTYPE',
+    'compat_datetime_timedelta_total_seconds',
     'compat_etree_fromstring',
     'compat_etree_iterfind',
     'compat_filter',
@@ -3533,6 +3618,12 @@ __all__ = [
     'compat_getpass_getpass',
     'compat_html_entities',
     'compat_html_entities_html5',
+    'compat_html_parser_HTMLParseError',
+    'compat_html_parser_HTMLParser',
+    'compat_http_cookiejar',
+    'compat_http_cookiejar_Cookie',
+    'compat_http_cookies',
+    'compat_http_cookies_SimpleCookie',
     'compat_http_client',
     'compat_http_server',
     'compat_input',
@@ -3555,7 +3646,7 @@ __all__ = [
     'compat_register_utf8',
     'compat_setenv',
     'compat_shlex_quote',
-    'compat_shlex_split',
+    'compat_shutil_get_terminal_size',
     'compat_socket_create_connection',
     'compat_str',
     'compat_struct_pack',
@@ -3575,5 +3666,5 @@ __all__ = [
     'compat_xml_etree_register_namespace',
     'compat_xpath',
     'compat_zip',
-    'workaround_optparse_bug9161',
+    'compat_zstandard',
 ]

youtube_dl/extractor/arte.py

@@ -4,208 +4,272 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
-from ..compat import (
+from ..compat import compat_str
-    compat_str,
-    compat_urlparse,
-)
 from ..utils import (
     ExtractorError,
+    GeoRestrictedError,
     int_or_none,
-    qualities,
+    merge_dicts,
+    parse_iso8601,
+    parse_qs,
     strip_or_none,
-    try_get,
+    traverse_obj,
-    unified_strdate,
     url_or_none,
+    urljoin,
 )
 
 
 class ArteTVBaseIE(InfoExtractor):
     _ARTE_LANGUAGES = 'fr|de|en|es|it|pl'
-    _API_BASE = 'https://api.arte.tv/api/player/v1'
+    _API_BASE = 'https://api.arte.tv/api/player/v2'
+
+    # yt-dlp shims
+
+    @classmethod
+    def _match_valid_url(cls, url):
+        return re.match(cls._VALID_URL, url)
+
+    def _extract_m3u8_formats_and_subtitles(self, *args, **kwargs):
+        return self._extract_m3u8_formats(*args, **kwargs), {}
 
 
 class ArteTVIE(ArteTVBaseIE):
     _VALID_URL = r'''(?x)
-                    https?://
+                    (?:https?://
                         (?:
                             (?:www\.)?arte\.tv/(?P<lang>%(langs)s)/videos|
                             api\.arte\.tv/api/player/v\d+/config/(?P<lang_2>%(langs)s)
                         )
-                        /(?P<id>\d{6}-\d{3}-[AF])
+                    |arte://program)
+                        /(?P<id>\d{6}-\d{3}-[AF]|LIVE)
                     ''' % {'langs': ArteTVBaseIE._ARTE_LANGUAGES}
     _TESTS = [{
         'url': 'https://www.arte.tv/en/videos/088501-000-A/mexico-stealing-petrol-to-survive/',
-        'info_dict': {
+        'only_matching': True,
-            'id': '088501-000-A',
-            'ext': 'mp4',
-            'title': 'Mexico: Stealing Petrol to Survive',
-            'upload_date': '20190628',
-        },
     }, {
         'url': 'https://www.arte.tv/pl/videos/100103-000-A/usa-dyskryminacja-na-porodowce/',
-        'only_matching': True,
+        'info_dict': {
+            'id': '100103-000-A',
+            'title': 'USA: Dyskryminacja na porodówce',
+            'description': 'md5:242017b7cce59ffae340a54baefcafb1',
+            'alt_title': 'ARTE Reportage',
+            'timestamp': 1604417980,
+            'upload_date': '20201103',
+            'duration': 554,
+            # test format sort
+            'height': 720,
+            'thumbnail': r're:https://api-cdn\.arte\.tv/.+940x530',
+            'ext': 'mp4',
+        },
+        'params': {
+            'format': 'bestvideo',
+            'skip_download': 'm3u8',
+        },
+    }, {
+        'note': 'No alt_title',
+        'url': 'https://www.arte.tv/fr/videos/110371-000-A/la-chaleur-supplice-des-arbres-de-rue/',
+        'info_dict': {
+            'id': '110371-000-A',
+            'ext': 'mp4',
+            'upload_date': '20220718',
+            'duration': 154,
+            'timestamp': 1658162460,
+            'description': 'md5:5890f36fe7dccfadb8b7c0891de54786',
+            'title': 'La chaleur, supplice des arbres de rue',
+            'thumbnail': 'https://api-cdn.arte.tv/img/v2/image/CPE2sQDtD8GLQgt8DuYHLf/940x530',
+        },
+        'params': {
+            'format': 'bestvideo',
+            'skip_download': 'm3u8',
+        },
     }, {
         'url': 'https://api.arte.tv/api/player/v2/config/de/100605-013-A',
         'only_matching': True,
+    }, {
+        'url': 'https://api.arte.tv/api/player/v2/config/de/LIVE',
+        'only_matching': True,
     }]
 
+    _GEO_BYPASS = True
+
+    _LANG_MAP = {  # ISO639 -> French abbreviations
+        'r': 'F',
+        'de': 'A',
+        'en': 'E[ANG]',
+        'es': 'E[ESP]',
+        'it': 'E[ITA]',
+        'pl': 'E[POL]',
+        # XXX: probably means mixed; <https://www.arte.tv/en/videos/107710-029-A/dispatches-from-ukraine-local-journalists-report/>
+        # uses this code for audio that happens to be in Ukrainian, but the manifest uses the ISO code 'mul' (mixed)
+        'mul': 'EU',
+    }
+
+    _VERSION_CODE_RE = re.compile(r'''(?x)
+        V
+        (?P<original_voice>O?)
+        (?P<vlang>[FA]|E\[[A-Z]+\]|EU)?
+        (?P<audio_desc>AUD|)
+        (?:
+            (?P<has_sub>-ST)
+            (?P<sdh_sub>M?)
+            (?P<sub_lang>[FA]|E\[[A-Z]+\]|EU)
+        )?
+    ''')
+
+    # all obtained by exhaustive testing
+    _COUNTRIES_MAP = {
+        'DE_FR': (
+            'BL', 'DE', 'FR', 'GF', 'GP', 'MF', 'MQ', 'NC',
+            'PF', 'PM', 'RE', 'WF', 'YT',
+        ),
+        # with both of the below 'BE' sometimes works, sometimes doesn't
+        'EUR_DE_FR': (
+            'AT', 'BL', 'CH', 'DE', 'FR', 'GF', 'GP', 'LI',
+            'MC', 'MF', 'MQ', 'NC', 'PF', 'PM', 'RE', 'WF',
+            'YT',
+        ),
+        'SAT': (
+            'AD', 'AT', 'AX', 'BG', 'BL', 'CH', 'CY', 'CZ',
+            'DE', 'DK', 'EE', 'ES', 'FI', 'FR', 'GB', 'GF',
+            'GR', 'HR', 'HU', 'IE', 'IS', 'IT', 'KN', 'LI',
+            'LT', 'LU', 'LV', 'MC', 'MF', 'MQ', 'MT', 'NC',
+            'NL', 'NO', 'PF', 'PL', 'PM', 'PT', 'RE', 'RO',
+            'SE', 'SI', 'SK', 'SM', 'VA', 'WF', 'YT',
+        ),
+    }
+
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        mobj = self._match_valid_url(url)
         video_id = mobj.group('id')
         lang = mobj.group('lang') or mobj.group('lang_2')
+        language_code = self._LANG_MAP.get(lang)
 
-        info = self._download_json(
+        config = self._download_json('{0}/config/{1}/{2}'.format(self._API_BASE, lang, video_id), video_id)
-            '%s/config/%s/%s' % (self._API_BASE, lang, video_id), video_id)
-        player_info = info['videoJsonPlayer']
 
-        vsr = try_get(player_info, lambda x: x['VSR'], dict)
+        geoblocking = traverse_obj(config, ('data', 'attributes', 'restriction', 'geoblocking')) or {}
-        if not vsr:
+        if geoblocking.get('restrictedArea'):
-            error = None
+            raise GeoRestrictedError('Video restricted to {0!r}'.format(geoblocking['code']),
-            if try_get(player_info, lambda x: x['custom_msg']['type']) == 'error':
+                                     countries=self._COUNTRIES_MAP.get(geoblocking['code'], ('DE', 'FR')))
-                error = try_get(
-                    player_info, lambda x: x['custom_msg']['msg'], compat_str)
-            if not error:
-                error = 'Video %s is not available' % player_info.get('VID') or video_id
-            raise ExtractorError(error, expected=True)
 
-        upload_date_str = player_info.get('shootingDate')
+        if not traverse_obj(config, ('data', 'attributes', 'rights')):
-        if not upload_date_str:
+            # Eg: https://www.arte.tv/de/videos/097407-215-A/28-minuten
-            upload_date_str = (player_info.get('VRA') or player_info.get('VDA') or '').split(' ')[0]
+            # Eg: https://www.arte.tv/es/videos/104351-002-A/serviteur-du-peuple-1-23
+            raise ExtractorError(
+                'Video is not available in this language edition of Arte or broadcast rights expired', expected=True)
 
-        title = (player_info.get('VTI') or player_info['VID']).strip()
+        formats, subtitles = [], {}
-        subtitle = player_info.get('VSU', '').strip()
+        secondary_formats = []
-        if subtitle:
+        for stream in config['data']['attributes']['streams']:
-            title += ' - %s' % subtitle
+            # official player contains code like `e.get("versions")[0].eStat.ml5`
+            stream_version = stream['versions'][0]
+            stream_version_code = stream_version['eStat']['ml5']
 
-        qfunc = qualities(['MQ', 'HQ', 'EQ', 'SQ'])
+            lang_pref = -1
+            m = self._VERSION_CODE_RE.match(stream_version_code)
+            if m:
+                lang_pref = int(''.join('01'[x] for x in (
+                    m.group('vlang') == language_code,      # we prefer voice in the requested language
+                    not m.group('audio_desc'),              # and not the audio description version
+                    bool(m.group('original_voice')),        # but if voice is not in the requested language, at least choose the original voice
+                    m.group('sub_lang') == language_code,   # if subtitles are present, we prefer them in the requested language
+                    not m.group('has_sub'),                 # but we prefer no subtitles otherwise
+                    not m.group('sdh_sub'),                 # and we prefer not the hard-of-hearing subtitles if there are subtitles
+                )))
 
-        LANGS = {
+            short_label = traverse_obj(stream_version, 'shortLabel', expected_type=compat_str, default='?')
-            'fr': 'F',
+            if stream['protocol'].startswith('HLS'):
-            'de': 'A',
+                fmts, subs = self._extract_m3u8_formats_and_subtitles(
-            'en': 'E[ANG]',
+                    stream['url'], video_id=video_id, ext='mp4', m3u8_id=stream_version_code, fatal=False)
-            'es': 'E[ESP]',
+                for fmt in fmts:
-            'it': 'E[ITA]',
+                    fmt.update({
-            'pl': 'E[POL]',
+                        'format_note': '{0} [{1}]'.format(stream_version.get("label", "unknown"), short_label),
-        }
+                        'language_preference': lang_pref,
+                    })
+                if any(map(short_label.startswith, ('cc', 'OGsub'))):
+                    secondary_formats.extend(fmts)
+                else:
+                    formats.extend(fmts)
+                for sub in subs:
+                    subtitles = self._merge_subtitles(subtitles, sub)
 
-        langcode = LANGS.get(lang, lang)
+            elif stream['protocol'] in ('HTTPS', 'RTMP'):
+                formats.append({
+                    'format_id': '{0}-{1}'.format(stream["protocol"], stream_version_code),
+                    'url': stream['url'],
+                    'format_note': '{0} [{1}]'.format(stream_version.get("label", "unknown"), short_label),
+                    'language_preference': lang_pref,
+                    # 'ext': 'mp4',  # XXX: may or may not be necessary, at least for HTTPS
+                })
 
-        formats = []
-        for format_id, format_dict in vsr.items():
-            f = dict(format_dict)
-            format_url = url_or_none(f.get('url'))
-            streamer = f.get('streamer')
-            if not format_url and not streamer:
-                continue
-            versionCode = f.get('versionCode')
-            l = re.escape(langcode)
-
-            # Language preference from most to least priority
-            # Reference: section 6.8 of
-            # https://www.arte.tv/sites/en/corporate/files/complete-technical-guidelines-arte-geie-v1-07-1.pdf
-            PREFERENCES = (
-                # original version in requested language, without subtitles
-                r'VO{0}$'.format(l),
-                # original version in requested language, with partial subtitles in requested language
-                r'VO{0}-ST{0}$'.format(l),
-                # original version in requested language, with subtitles for the deaf and hard-of-hearing in requested language
-                r'VO{0}-STM{0}$'.format(l),
-                # non-original (dubbed) version in requested language, without subtitles
-                r'V{0}$'.format(l),
-                # non-original (dubbed) version in requested language, with subtitles partial subtitles in requested language
-                r'V{0}-ST{0}$'.format(l),
-                # non-original (dubbed) version in requested language, with subtitles for the deaf and hard-of-hearing in requested language
-                r'V{0}-STM{0}$'.format(l),
-                # original version in requested language, with partial subtitles in different language
-                r'VO{0}-ST(?!{0}).+?$'.format(l),
-                # original version in requested language, with subtitles for the deaf and hard-of-hearing in different language
-                r'VO{0}-STM(?!{0}).+?$'.format(l),
-                # original version in different language, with partial subtitles in requested language
-                r'VO(?:(?!{0}).+?)?-ST{0}$'.format(l),
-                # original version in different language, with subtitles for the deaf and hard-of-hearing in requested language
-                r'VO(?:(?!{0}).+?)?-STM{0}$'.format(l),
-                # original version in different language, without subtitles
-                r'VO(?:(?!{0}))?$'.format(l),
-                # original version in different language, with partial subtitles in different language
-                r'VO(?:(?!{0}).+?)?-ST(?!{0}).+?$'.format(l),
-                # original version in different language, with subtitles for the deaf and hard-of-hearing in different language
-                r'VO(?:(?!{0}).+?)?-STM(?!{0}).+?$'.format(l),
-            )
-
-            for pref, p in enumerate(PREFERENCES):
-                if re.match(p, versionCode):
-                    lang_pref = len(PREFERENCES) - pref
-                    break
             else:
-                lang_pref = -1
+                self.report_warning('Skipping stream with unknown protocol {0}'.format(stream["protocol"]))
 
-            media_type = f.get('mediaType')
+            # TODO: chapters from stream['segments']?
-            if media_type == 'hls':
+            # The JS also looks for chapters in config['data']['attributes']['chapters'],
-                m3u8_formats = self._extract_m3u8_formats(
+            # but I am yet to find a video having those
-                    format_url, video_id, 'mp4', entry_protocol='m3u8_native',
-                    m3u8_id=format_id, fatal=False)
-                for m3u8_format in m3u8_formats:
-                    m3u8_format['language_preference'] = lang_pref
-                formats.extend(m3u8_formats)
-                continue
-
-            format = {
-                'format_id': format_id,
-                'preference': -10 if f.get('videoFormat') == 'M3U8' else None,
-                'language_preference': lang_pref,
-                'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')),
-                'width': int_or_none(f.get('width')),
-                'height': int_or_none(f.get('height')),
-                'tbr': int_or_none(f.get('bitrate')),
-                'quality': qfunc(f.get('quality')),
-            }
-
-            if media_type == 'rtmp':
-                format['url'] = f['streamer']
-                format['play_path'] = 'mp4:' + f['url']
-                format['ext'] = 'flv'
-            else:
-                format['url'] = f['url']
-
-            formats.append(format)
 
+        formats.extend(secondary_formats)
+        self._remove_duplicate_formats(formats)
         self._sort_formats(formats)
 
+        metadata = config['data']['attributes']['metadata']
+
         return {
-            'id': player_info.get('VID') or video_id,
+            'id': metadata['providerId'],
-            'title': title,
+            'webpage_url': traverse_obj(metadata, ('link', 'url')),
-            'description': player_info.get('VDE'),
+            'title': traverse_obj(metadata, 'subtitle', 'title'),
-            'upload_date': unified_strdate(upload_date_str),
+            'alt_title': metadata.get('subtitle') and metadata.get('title'),
-            'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
+            'description': metadata.get('description'),
+            'duration': traverse_obj(metadata, ('duration', 'seconds')),
+            'language': metadata.get('language'),
+            'timestamp': traverse_obj(config, ('data', 'attributes', 'rights', 'begin'), expected_type=parse_iso8601),
+            'is_live': config['data']['attributes'].get('live', False),
             'formats': formats,
+            'subtitles': subtitles,
+            'thumbnails': [
+                {'url': image['url'], 'id': image.get('caption')}
+                for image in metadata.get('images') or [] if url_or_none(image.get('url'))
+            ],
         }
 
 
 class ArteTVEmbedIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?arte\.tv/player/v\d+/index\.php\?.*?\bjson_url=.+'
+    _EMBED_REGEX = [r'<(?:iframe|script)[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?arte\.tv/player/v\d+/index\.php\?.*?\bjson_url=.+?)\1']
     _TESTS = [{
         'url': 'https://www.arte.tv/player/v5/index.php?json_url=https%3A%2F%2Fapi.arte.tv%2Fapi%2Fplayer%2Fv2%2Fconfig%2Fde%2F100605-013-A&lang=de&autoplay=true&mute=0100605-013-A',
+        'only_matching': True,
+        'skip': 'Video is not available in this language edition of Arte or broadcast rights expired'
+    }, {
+        'url': 'https://www.arte.tv/player/v5/index.php?json_url=https%3A%2F%2Fapi.arte.tv%2Fapi%2Fplayer%2Fv2%2Fconfig%2Fpl%2F100103-000-A&lang=pl&autoplay=true&mute=100103-000-A',
         'info_dict': {
-            'id': '100605-013-A',
+            'id': '100103-000-A',
             'ext': 'mp4',
-            'title': 'United we Stream November Lockdown Edition #13',
+            'title': 'USA: Dyskryminacja na porodówce',
-            'description': 'md5:be40b667f45189632b78c1425c7c2ce1',
+            'timestamp': 1604417980,
-            'upload_date': '20201116',
+            'upload_date': '20201103',
+            'description': 'md5:242017b7cce59ffae340a54baefcafb1',
+            'duration': 554,
+        },
+        'params': {
+            'format': 'bestvideo',
+            'skip_download': 'm3u8',
         },
     }, {
         'url': 'https://www.arte.tv/player/v3/index.php?json_url=https://api.arte.tv/api/player/v2/config/de/100605-013-A',
         'only_matching': True,
     }]
 
-    @staticmethod
+    @classmethod
-    def _extract_urls(webpage):
+    def _extract_urls(cls, webpage):
-        return [url for _, url in re.findall(
+        import itertools    # just until this is lifted into IE
-            r'<(?:iframe|script)[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?arte\.tv/player/v\d+/index\.php\?.*?\bjson_url=.+?)\1',
+        return list(itertools.chain(*(
-            webpage)]
+            (url for _, url in re.findall(erx, webpage)) for erx in cls._EMBED_REGEX)
+        ))
 
     def _real_extract(self, url):
-        qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
+        qs = parse_qs(url)
-        json_url = qs['json_url'][0]
+        json_url = qs['json_url'][-1]
         video_id = ArteTVIE._match_id(json_url)
         return self.url_result(
             json_url, ie=ArteTVIE.ie_key(), video_id=video_id)
@@ -215,44 +279,45 @@ class ArteTVPlaylistIE(ArteTVBaseIE):
     _VALID_URL = r'https?://(?:www\.)?arte\.tv/(?P<lang>%s)/videos/(?P<id>RC-\d{6})' % ArteTVBaseIE._ARTE_LANGUAGES
     _TESTS = [{
         'url': 'https://www.arte.tv/en/videos/RC-016954/earn-a-living/',
-        'info_dict': {
+        'only_matching': True,
-            'id': 'RC-016954',
-            'title': 'Earn a Living',
-            'description': 'md5:d322c55011514b3a7241f7fb80d494c2',
-        },
-        'playlist_mincount': 6,
     }, {
         'url': 'https://www.arte.tv/pl/videos/RC-014123/arte-reportage/',
-        'only_matching': True,
+        'playlist_mincount': 100,
+        'info_dict': {
+            'description': 'md5:84e7bf1feda248bc325ebfac818c476e',
+            'id': 'RC-014123',
+            'title': 'ARTE Reportage - najlepsze reportaże',
+        },
+        'skip': '404 Not Found',
+    }, {
+        'url': 'https://www.arte.tv/en/videos/RC-016979/war-in-ukraine/',
+        'playlist_mincount': 79,
+        'info_dict': {
+            'id': 'RC-016979',
+            'title': 'War in Ukraine',
+            'description': 'On 24 February, Russian armed forces invaded Ukraine. We follow the war day by day and provide background information with special insights, reports and documentaries.',
+        },
     }]
 
     def _real_extract(self, url):
-        lang, playlist_id = re.match(self._VALID_URL, url).groups()
+        lang, playlist_id = self._match_valid_url(url).group('lang', 'id')
-        collection = self._download_json(
+        playlist = self._download_json(
-            '%s/collectionData/%s/%s?source=videos'
+            '{0}/playlist/{1}/{2}'.format(self._API_BASE, lang, playlist_id), playlist_id)['data']['attributes']
-            % (self._API_BASE, lang, playlist_id), playlist_id)
+
-        entries = []
+        entries = [{
-        for video in collection['videos']:
+            '_type': 'url_transparent',
-            if not isinstance(video, dict):
+            'url': video['config']['url'],
-                continue
+            'ie_key': ArteTVIE.ie_key(),
-            video_url = url_or_none(video.get('url')) or url_or_none(video.get('jsonUrl'))
+            'id': video.get('providerId'),
-            if not video_url:
+            'title': video.get('title'),
-                continue
+            'alt_title': video.get('subtitle'),
-            video_id = video.get('programId')
+            'thumbnail': url_or_none(traverse_obj(video, ('mainImage', 'url'))),
-            entries.append({
+            'duration': int_or_none(traverse_obj(video, ('duration', 'seconds'))),
-                '_type': 'url_transparent',
+        } for video in traverse_obj(playlist, ('items', lambda _, v: v['config']['url']))]
-                'url': video_url,
+
-                'id': video_id,
+        return self.playlist_result(entries, playlist_id,
-                'title': video.get('title'),
+                                    traverse_obj(playlist, ('metadata', 'title')),
-                'alt_title': video.get('subtitle'),
+                                    traverse_obj(playlist, ('metadata', 'description')))
-                'thumbnail': url_or_none(try_get(video, lambda x: x['mainImage']['url'], compat_str)),
-                'duration': int_or_none(video.get('durationSeconds')),
-                'view_count': int_or_none(video.get('views')),
-                'ie_key': ArteTVIE.ie_key(),
-            })
-        title = collection.get('title')
-        description = collection.get('shortDescription') or collection.get('teaserText')
-        return self.playlist_result(entries, playlist_id, title, description)
 
 
 class ArteTVCategoryIE(ArteTVBaseIE):
@@ -262,11 +327,10 @@ class ArteTVCategoryIE(ArteTVBaseIE):
         'info_dict': {
             'id': 'politics-and-society',
             'title': 'Politics and society',
-            'description': 'Investigative documentary series, geopolitical analysis, and international commentary',
+            'description': 'Watch documentaries and reportage about politics, society and current affairs.',
         },
         'playlist_mincount': 13,
-    },
+    }]
-    ]
 
     @classmethod
     def suitable(cls, url):
@@ -275,27 +339,23 @@ class ArteTVCategoryIE(ArteTVBaseIE):
             and super(ArteTVCategoryIE, cls).suitable(url))
 
     def _real_extract(self, url):
-        lang, playlist_id = re.match(self._VALID_URL, url).groups()
+        lang, playlist_id = self._match_valid_url(url).groups()
         webpage = self._download_webpage(url, playlist_id)
 
         items = []
         for video in re.finditer(
-                r'<a\b[^>]*?href\s*=\s*(?P<q>"|\'|\b)(?P<url>https?://www\.arte\.tv/%s/videos/[\w/-]+)(?P=q)' % lang,
+                r'<a\b[^>]+\bhref\s*=\s*(?P<q>"|\'|\b)(?P<url>(?:https?://www\.arte\.tv)?/%s/videos/[\w/-]+)(?P=q)' % lang,
                 webpage):
-            video = video.group('url')
+            video = urljoin(url, video.group('url'))
             if video == url:
                 continue
             if any(ie.suitable(video) for ie in (ArteTVIE, ArteTVPlaylistIE, )):
                 items.append(video)
 
-        if items:
+        title = (self._og_search_title(webpage, default=None)
-            title = (self._og_search_title(webpage, default=None)
+                 or self._html_search_regex(r'<title\b[^>]*>([^<]+)</title>', default=None))
-                     or self._html_search_regex(r'<title\b[^>]*>([^<]+)</title>', default=None))
+        title = strip_or_none(title.rsplit('|', 1)[0]) or self._generic_title(url)
-            title = strip_or_none(title.rsplit('|', 1)[0]) or self._generic_title(url)
 
-            result = self.playlist_from_matches(items, playlist_id=playlist_id, playlist_title=title)
+        return merge_dicts(
-            if result:
+            self.playlist_from_matches(items, playlist_id=playlist_id, playlist_title=title),
-                description = self._og_search_description(webpage, default=None)
+            {'description': self._og_search_description(webpage, default=None)})
-                if description:
-                    result['description'] = description
-                return result

youtube_dl/extractor/youtube.py

@@ -27,6 +27,7 @@ from ..compat import (
 )
 from ..jsinterp import JSInterpreter
 from ..utils import (
+    bug_reports_message,
     clean_html,
     dict_get,
     error_to_compat_str,
@@ -65,6 +66,7 @@ from ..utils import (
     url_or_none,
     urlencode_postdata,
     urljoin,
+    variadic,
 )
 
 
@@ -120,7 +122,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'TVHTML5',
-                    'clientVersion': '7.20241201.18.00',
+                    'clientVersion': '7.20250120.19.00',
+                    'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 7,
@@ -460,6 +463,26 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'uploader': uploader,
         }
 
+    @staticmethod
+    def _extract_thumbnails(data, *path_list, **kw_final_key):
+        """
+        Extract thumbnails from thumbnails dict
+        @param path_list: path list to level that contains 'thumbnails' key
+        """
+        final_key = kw_final_key.get('final_key', 'thumbnails')
+
+        return traverse_obj(data, ((
+            tuple(variadic(path) + (final_key, Ellipsis)
+                  for path in path_list or [()])), {
+            'url': ('url', T(url_or_none),
+                    # Sometimes youtube gives a wrong thumbnail URL. See:
+                    # https://github.com/yt-dlp/yt-dlp/issues/233
+                    # https://github.com/ytdl-org/youtube-dl/issues/28023
+                    T(lambda u: update_url(u, query=None) if u and 'maxresdefault' in u else u)),
+            'height': ('height', T(int_or_none)),
+            'width': ('width', T(int_or_none)),
+        }, T(lambda t: t if t.get('url') else None)))
+
     def _search_results(self, query, params):
         data = {
             'context': {
@@ -1829,12 +1852,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         if func_code:
             return jsi, player_id, func_code
+        return self._extract_n_function_code_jsi(video_id, jsi, player_id)
 
-        func_name = self._extract_n_function_name(jscode)
+    def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
+
+        var_ay = self._search_regex(
+            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"[^"]+"\s*\.\s*split\("\{"\))(?=\s*[,;])',
+            jsi.code, 'useful values', default='')
+
+        func_name = self._extract_n_function_name(jsi.code)
 
         func_code = jsi.extract_function_code(func_name)
+        if var_ay:
+            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
 
-        self.cache.store('youtube-nsig', player_id, func_code)
+        if player_id:
+            self.cache.store('youtube-nsig', player_id, func_code)
         return jsi, player_id, func_code
 
     def _extract_n_function_from_code(self, jsi, func_code):
@@ -3183,8 +3216,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
             expected_type=txt_or_none)
 
     def _grid_entries(self, grid_renderer):
-        for item in grid_renderer['items']:
+        for item in traverse_obj(grid_renderer, ('items', Ellipsis, T(dict))):
-            if not isinstance(item, dict):
+            lockup_view_model = traverse_obj(item, ('lockupViewModel', T(dict)))
+            if lockup_view_model:
+                entry = self._extract_lockup_view_model(lockup_view_model)
+                if entry:
+                    yield entry
                 continue
             renderer = self._extract_grid_item_renderer(item)
             if not isinstance(renderer, dict):
@@ -3268,6 +3305,25 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
                 continue
             yield self._extract_video(renderer)
 
+    def _extract_lockup_view_model(self, view_model):
+        content_id = view_model.get('contentId')
+        if not content_id:
+            return
+        content_type = view_model.get('contentType')
+        if content_type not in ('LOCKUP_CONTENT_TYPE_PLAYLIST', 'LOCKUP_CONTENT_TYPE_PODCAST'):
+            self.report_warning(
+                'Unsupported lockup view model content type "{0}"{1}'.format(content_type, bug_reports_message()), only_once=True)
+            return
+        return merge_dicts(self.url_result(
+            update_url_query('https://www.youtube.com/playlist', {'list': content_id}),
+            ie=YoutubeTabIE.ie_key(), video_id=content_id), {
+                'title': traverse_obj(view_model, (
+                    'metadata', 'lockupMetadataViewModel', 'title', 'content', T(compat_str))),
+                'thumbnails': self._extract_thumbnails(view_model, (
+                    'contentImage', 'collectionThumbnailViewModel', 'primaryThumbnail',
+                    'thumbnailViewModel', 'image'), final_key='sources'),
+        })
+
     def _video_entry(self, video_renderer):
         video_id = video_renderer.get('videoId')
         if video_id:

youtube_dl/jsinterp.py

@@ -1,10 +1,12 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import calendar
 import itertools
 import json
 import operator
 import re
+import time
 
 from functools import update_wrapper, wraps
 
@@ -12,8 +14,10 @@ from .utils import (
     error_to_compat_str,
     ExtractorError,
     float_or_none,
+    int_or_none,
     js_to_json,
     remove_quotes,
+    str_or_none,
     unified_timestamp,
     variadic,
     write_string,
@@ -150,6 +154,7 @@ def _js_to_primitive(v):
     )
 
 
+# more exact: yt-dlp/yt-dlp#12110
 def _js_toString(v):
     return (
         'undefined' if v is JS_Undefined
@@ -158,7 +163,7 @@ def _js_toString(v):
         else 'null' if v is None
         # bool <= int: do this first
         else ('false', 'true')[v] if isinstance(v, bool)
-        else '{0:.7f}'.format(v).rstrip('.0') if isinstance(v, compat_numeric_types)
+        else re.sub(r'(?<=\d)\.?0*$', '', '{0:.7f}'.format(v)) if isinstance(v, compat_numeric_types)
         else _js_to_primitive(v))
 
 
@@ -404,6 +409,7 @@ class JSInterpreter(object):
     class Exception(ExtractorError):
         def __init__(self, msg, *args, **kwargs):
             expr = kwargs.pop('expr', None)
+            msg = str_or_none(msg, default='"None"')
             if expr is not None:
                 msg = '{0} in: {1!r:.100}'.format(msg.rstrip(), expr)
             super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs)
@@ -431,6 +437,7 @@ class JSInterpreter(object):
                 flags, _ = self.regex_flags(flags)
             # First, avoid https://github.com/python/cpython/issues/74534
             self.__self = None
+            pattern_txt = str_or_none(pattern_txt) or '(?:)'
             self.__pattern_txt = pattern_txt.replace('[[', r'[\[')
             self.__flags = flags
 
@@ -475,6 +482,73 @@ class JSInterpreter(object):
                 flags |= cls.RE_FLAGS[ch]
             return flags, expr[idx + 1:]
 
+    class JS_Date(object):
+        _t = None
+
+        @staticmethod
+        def __ymd_etc(*args, **kw_is_utc):
+            # args: year, monthIndex, day, hours, minutes, seconds, milliseconds
+            is_utc = kw_is_utc.get('is_utc', False)
+
+            args = list(args[:7])
+            args += [0] * (9 - len(args))
+            args[1] += 1  # month 0..11 -> 1..12
+            ms = args[6]
+            for i in range(6, 9):
+                args[i] = -1  # don't know
+            if is_utc:
+                args[-1] = 1
+            # TODO: [MDN] When a segment overflows or underflows its expected
+            # range, it usually "carries over to" or "borrows from" the higher segment.
+            try:
+                mktime = calendar.timegm if is_utc else time.mktime
+                return mktime(time.struct_time(args)) * 1000 + ms
+            except (OverflowError, ValueError):
+                return None
+
+        @classmethod
+        def UTC(cls, *args):
+            t = cls.__ymd_etc(*args, is_utc=True)
+            return _NaN if t is None else t
+
+        @staticmethod
+        def parse(date_str, **kw_is_raw):
+            is_raw = kw_is_raw.get('is_raw', False)
+
+            t = unified_timestamp(str_or_none(date_str), False)
+            return int(t * 1000) if t is not None else t if is_raw else _NaN
+
+        @staticmethod
+        def now(**kw_is_raw):
+            is_raw = kw_is_raw.get('is_raw', False)
+
+            t = time.time()
+            return int(t * 1000) if t is not None else t if is_raw else _NaN
+
+        def __init__(self, *args):
+            if not args:
+                args = [self.now(is_raw=True)]
+            if len(args) == 1:
+                if isinstance(args[0], JSInterpreter.JS_Date):
+                    self._t = int_or_none(args[0].valueOf(), default=None)
+                else:
+                    arg_type = _js_typeof(args[0])
+                    if arg_type == 'string':
+                        self._t = self.parse(args[0], is_raw=True)
+                    elif arg_type == 'number':
+                        self._t = int(args[0])
+            else:
+                self._t = self.__ymd_etc(*args)
+
+        def toString(self):
+            try:
+                return time.strftime('%a %b %0d %Y %H:%M:%S %Z%z', self._t).rstrip()
+            except TypeError:
+                return "Invalid Date"
+
+        def valueOf(self):
+            return _NaN if self._t is None else self._t
+
     @classmethod
     def __op_chars(cls):
         op_chars = set(';,[')
@@ -599,14 +673,15 @@ class JSInterpreter(object):
         except Exception as e:
             raise self.Exception('Failed to evaluate {left_val!r:.50} {op} {right_val!r:.50}'.format(**locals()), expr, cause=e)
 
-    def _index(self, obj, idx, allow_undefined=True):
+    def _index(self, obj, idx, allow_undefined=None):
         if idx == 'length' and isinstance(obj, list):
             return len(obj)
         try:
             return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)]
         except (TypeError, KeyError, IndexError) as e:
-            if allow_undefined:
+            # allow_undefined is None gives correct behaviour
-                # when is not allowed?
+            if allow_undefined or (
+                    allow_undefined is None and not isinstance(e, TypeError)):
                 return JS_Undefined
             raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
@@ -715,7 +790,7 @@ class JSInterpreter(object):
 
         new_kw, _, obj = expr.partition('new ')
         if not new_kw:
-            for klass, konstr in (('Date', lambda x: int(unified_timestamp(x, False) * 1000)),
+            for klass, konstr in (('Date', lambda *x: self.JS_Date(*x).valueOf()),
                                   ('RegExp', self.JS_RegExp),
                                   ('Error', self.Exception)):
                 if not obj.startswith(klass + '('):
@@ -1034,6 +1109,7 @@ class JSInterpreter(object):
                     'String': compat_str,
                     'Math': float,
                     'Array': list,
+                    'Date': self.JS_Date,
                 }
                 obj = local_vars.get(variable)
                 if obj in (JS_Undefined, None):
@@ -1086,6 +1162,8 @@ class JSInterpreter(object):
                         assertion(len(argvals) == 2, 'takes two arguments')
                         return argvals[0] ** argvals[1]
                     raise self.Exception('Unsupported Math method ' + member, expr=expr)
+                elif obj is self.JS_Date:
+                    return getattr(obj, member)(*argvals)
 
                 if member == 'split':
                     assertion(len(argvals) <= 2, 'takes at most two arguments')