Merge a53b4004cff69bca0f206bbd0b4a3003038568c7 into da7223d4aa42ff9fc680b0951d043dd03cec2d30

* improve extraction of global "useful data" Array from player JS
* also handle tv-player and add tests: thx seproDev (yt-dlp/yt-dlp#12684)

Co-Authored-By: sepro <sepro@sepr0.com>

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 +223,42 @@ _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/643afba4/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
+    ),
+    (
+        'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js',
+        'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ',
+    ),
+    (
+        'https://www.youtube.com/s/player/d50f54ef/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'XFTb2HoeOE5MHg',
+    ),
+    (
+        'https://www.youtube.com/s/player/074a8365/player_ias_tce.vflset/en_US/base.js',
+        'Ha7507LzRmH3Utygtj', 'ufTsrE0IVYrkl8v',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'N5uAlLqm0eg1GyHO', 'dCBQOejdq5s-ww',
+    ),
+    (
+        'https://www.youtube.com/s/player/69f581a5/tv-player-ias.vflset/tv-player-ias.js',
+        '-qIP447rVlTTwaZjY', 'KNcGOksBAvwqQg',
+    ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
+    ),
 ]
 
 
@@ -284,7 +320,7 @@ 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 +328,10 @@ 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:
@@ -3205,7 +3268,7 @@ except AttributeError:
 
 # 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/extractors.py

@@ -1238,10 +1238,7 @@ from .svt import (
 from .swrmediathek import SWRMediathekIE
 from .syfy import SyfyIE
 from .sztvhu import SztvHuIE
-from .tagesschau import (
+from .tagesschau import TagesschauIE
-    TagesschauPlayerIE,
-    TagesschauIE,
-)
 from .tass import TassIE
 from .tbs import TBSIE
 from .tdslifeway import TDSLifewayIE

youtube_dl/extractor/tagesschau.py

@@ -5,127 +5,54 @@ import re
 
 from .common import InfoExtractor
 from ..utils import (
+    bool_or_none,
+    compat_str,
     determine_ext,
-    js_to_json,
+    ExtractorError,
-    parse_iso8601,
+    parse_duration,
     parse_filesize,
+    remove_quotes,
+    strip_or_none,
+    try_get,
+    unescapeHTML,
+    unified_timestamp,
+    url_or_none,
 )
 
+# Note that there are tagesschau.de/api and tagesschau.de/api2 endpoints, which
+# may be useful, but not all pages and not all formats can be easily accessed
+# by API.
 
-class TagesschauPlayerIE(InfoExtractor):
-    IE_NAME = 'tagesschau:player'
-    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de/multimedia/(?P<kind>audio|video)/(?P=kind)-(?P<id>\d+)~player(?:_[^/?#&]+)?\.html'
 
-    _TESTS = [{
+_FORMATS = {
-        'url': 'http://www.tagesschau.de/multimedia/video/video-179517~player.html',
-        'md5': '8d09548d5c15debad38bee3a4d15ca21',
-        'info_dict': {
-            'id': '179517',
-            'ext': 'mp4',
-            'title': 'Marie Kristin Boese, ARD Berlin, über den zukünftigen Kurs der AfD',
-            'thumbnail': r're:^https?:.*\.jpg$',
-            'formats': 'mincount:6',
-        },
-    }, {
-        'url': 'https://www.tagesschau.de/multimedia/audio/audio-29417~player.html',
-        'md5': '76e6eec6ebd40740671cf0a2c88617e5',
-        'info_dict': {
-            'id': '29417',
-            'ext': 'mp3',
-            'title': 'Trabi - Bye, bye Rennpappe',
-            'thumbnail': r're:^https?:.*\.jpg$',
-            'formats': 'mincount:2',
-        },
-    }, {
-        'url': 'http://www.tagesschau.de/multimedia/audio/audio-29417~player_autoplay-true.html',
-        'only_matching': True,
-    }]
-
-    _FORMATS = {
     'xs': {'quality': 0},
     's': {'width': 320, 'height': 180, 'quality': 1},
+    'sm': {'width': 480, 'height': 270, 'quality': 1},
     'm': {'width': 512, 'height': 288, 'quality': 2},
+    'ml': {'width': 640, 'height': 360, 'quality': 2},
     'l': {'width': 960, 'height': 540, 'quality': 3},
     'xl': {'width': 1280, 'height': 720, 'quality': 4},
     'xxl': {'quality': 5},
-    }
+    'mp3': {'abr': 64, 'vcodec': 'none', 'quality': 0},
+    'hi.mp3': {'abr': 192, 'vcodec': 'none', 'quality': 1},
+}
 
-    def _extract_via_api(self, kind, video_id):
+_FIELD_PREFERENCE = ('height', 'width', 'vbr', 'abr')
-        info = self._download_json(
-            'https://www.tagesschau.de/api/multimedia/{0}/{0}-{1}.json'.format(kind, video_id),
-            video_id)
-        title = info['headline']
-        formats = []
-        for media in info['mediadata']:
-            for format_id, format_url in media.items():
-                if determine_ext(format_url) == 'm3u8':
-                    formats.extend(self._extract_m3u8_formats(
-                        format_url, video_id, 'mp4',
-                        entry_protocol='m3u8_native', m3u8_id='hls'))
-                else:
-                    formats.append({
-                        'url': format_url,
-                        'format_id': format_id,
-                        'vcodec': 'none' if kind == 'audio' else None,
-                    })
-        self._sort_formats(formats)
-        timestamp = parse_iso8601(info.get('date'))
-        return {
-            'id': video_id,
-            'title': title,
-            'timestamp': timestamp,
-            'formats': formats,
-        }
 
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id')
 
-        # kind = mobj.group('kind').lower()
+def _normalize_format_id(format_id, ext):
-        # if kind == 'video':
+    if format_id:
-        #     return self._extract_via_api(kind, video_id)
+        m = re.match(r"web([^.]+)\.[^.]+$", format_id)
-
+        if m:
-        # JSON api does not provide some audio formats (e.g. ogg) thus
+            format_id = m.group(1)
-        # extracting audio via webpage
+        if format_id == 'hi' and ext:
-
+            # high-quality audio files
-        webpage = self._download_webpage(url, video_id)
+            format_id = '%s.%s' % (format_id, ext)
-
+    return format_id
-        title = self._og_search_title(webpage).strip()
-        formats = []
-
-        for media_json in re.findall(r'({src\s*:\s*["\']http[^}]+type\s*:[^}]+})', webpage):
-            media = self._parse_json(js_to_json(media_json), video_id, fatal=False)
-            if not media:
-                continue
-            src = media.get('src')
-            if not src:
-                return
-            quality = media.get('quality')
-            kind = media.get('type', '').split('/')[0]
-            ext = determine_ext(src)
-            f = {
-                'url': src,
-                'format_id': '%s_%s' % (quality, ext) if quality else ext,
-                'ext': ext,
-                'vcodec': 'none' if kind == 'audio' else None,
-            }
-            f.update(self._FORMATS.get(quality, {}))
-            formats.append(f)
-
-        self._sort_formats(formats)
-
-        thumbnail = self._og_search_thumbnail(webpage)
-
-        return {
-            'id': video_id,
-            'title': title,
-            'thumbnail': thumbnail,
-            'formats': formats,
-        }
 
 
 class TagesschauIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de/(?P<path>[^/]+/(?:[^/]+/)*?(?P<id>[^/#?]+?(?:-?[0-9]+)?))(?:~_?[^/#?]+?)?\.html'
+    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de(?:/?|/(?P<path>[^?#]+?(?:/(?P<id>[^/#?]+?(?:-?[0-9]+))(?:~_?[^/#?]+?)?(?:\.html)?)?))(?:[#?].*)?$'
 
     _TESTS = [{
         'url': 'http://www.tagesschau.de/multimedia/video/video-102143.html',
@@ -134,48 +61,111 @@ class TagesschauIE(InfoExtractor):
             'id': 'video-102143',
             'ext': 'mp4',
             'title': 'Regierungsumbildung in Athen: Neue Minister in Griechenland vereidigt',
-            'description': '18.07.2015 20:10 Uhr',
+            'description': '18.07.2015 20:10',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'upload_date': '20150718',
+            'duration': 138,
+            'timestamp': 1437250200,
+            'uploader': 'ARD',
+        },
+    }, {
+        # with player
+        'url': 'http://www.tagesschau.de/multimedia/video/video-102143~player.html',
+        'md5': 'f7c27a0eff3bfe8c7727e65f8fe1b1e6',
+        'info_dict': {
+            'id': 'video-102143',
+            'ext': 'mp4',
+            'title': 'Regierungsumbildung in Athen: Neue Minister in Griechenland vereidigt',
+            'description': '18.07.2015 20:10',
+            'thumbnail': r're:^https?:.*\.jpg$',
+            'upload_date': '20150718',
+            'timestamp': 1437250200,
+            'uploader': 'ARD',
         },
     }, {
         'url': 'http://www.tagesschau.de/multimedia/sendung/ts-5727.html',
         'md5': '3c54c1f6243d279b706bde660ceec633',
         'info_dict': {
-            'id': 'ts-5727',
+            'id': 'video-45741',
             'ext': 'mp4',
-            'title': 'Sendung: tagesschau \t04.12.2014 20:00 Uhr',
+            'title': 'tagesschau 20 Uhr - 04.12.14 20:00',
-            'description': 'md5:695c01bfd98b7e313c501386327aea59',
+            'description': '04.12.2014 20:00',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'uploader': 'tagesschau',
+            'timestamp': 1417723200,
+            'upload_date': '20141204',
+            'subtitles': dict,
         },
     }, {
         # exclusive audio
-        'url': 'http://www.tagesschau.de/multimedia/audio/audio-29417.html',
+        'url': 'https://www.tagesschau.de/multimedia/audio/audio-103205.html',
-        'md5': '76e6eec6ebd40740671cf0a2c88617e5',
+        'md5': 'c8e7b72aeca664031db0ba198519b09a',
         'info_dict': {
-            'id': 'audio-29417',
+            'id': 'audio-103205',
             'ext': 'mp3',
-            'title': 'Trabi - Bye, bye Rennpappe',
+            'title': 'Die USA: ein Impfwunder?',
-            'description': 'md5:8687dda862cbbe2cfb2df09b56341317',
+            'description': '06.03.2021 06:07',
+            'timestamp': 1615010820,
+            'upload_date': '20210306',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'uploader': 'Jule Käppel, ARD Washington',
+            'creator': 'ARD',
+            'channel': 'tagesschau.de',
+            'is_live': False,
         },
     }, {
         # audio in article
-        'url': 'http://www.tagesschau.de/inland/bnd-303.html',
+        'url': 'https://www.tagesschau.de/ausland/amerika/biden-versoehnung-101.html',
-        'md5': 'e0916c623e85fc1d2b26b78f299d3958',
+        'md5': '4c46b0283719d97aa976037e1ecb7b73',
         'info_dict': {
-            'id': 'bnd-303',
+            'id': 'audio-103429',
+            'title': 'Bidens Versöhnungswerk kommt nicht voran',
             'ext': 'mp3',
-            'title': 'Viele Baustellen für neuen BND-Chef',
+            'timestamp': 1615444860,
-            'description': 'md5:1e69a54be3e1255b2b07cdbce5bcd8b4',
+            'uploader': 'Sebastian Hesse, ARD Washington',
-            'thumbnail': r're:^https?:.*\.jpg$',
+            'description': '11.03.2021 06:41',
+            'upload_date': '20210311',
+            'creator': 'ARD',
+            'channel': 'tagesschau.de',
         },
     }, {
-        'url': 'http://www.tagesschau.de/inland/afd-parteitag-135.html',
+        # playlist in article
+        'url': 'https://www.tagesschau.de/ausland/impfungen-coronavirus-usa-101.html',
         'info_dict': {
-            'id': 'afd-parteitag-135',
+            'id': 'impfungen-coronavirus-usa-101',
-            'title': 'Möchtegern-Underdog mit Machtanspruch',
+            'title': 'Kampf gegen das Coronavirus: Impfwunder USA?',
         },
         'playlist_count': 2,
+    }, {
+        # article without videos
+        'url': 'https://www.tagesschau.de/wirtschaft/ukraine-russland-kredit-101.html',
+        'info_dict': {
+            'id': 'ukraine-russland-kredit-101',
+            'title': 'Ukraine stoppt Rückzahlung russischer Kredite',
+        },
+        'playlist_count': 0,
+    }, {
+        # legacy website
+        'url': 'https://www.tagesschau.de/multimedia/video/video-102303~_bab-sendung-211.html',
+        'md5': 'ab6d190c8147560d6429a467566affe6',
+        'info_dict': {
+            'id': 'video-102303',
+            'ext': 'mp4',
+            'title': 'Bericht aus Berlin: Sommerinterview mit Angela Merkel',
+            'description': '19.07.2015 19:05 Uhr',
+        }
+    }, {
+        # handling of generic title
+        'url': 'https://www.tagesschau.de/multimedia/video/video-835681.html',
+        'info_dict': {
+            'id': 'video-835681',
+            'ext': 'mp4',
+            'title': 'Tagesschau in 100 Sekunden - 13.03.21 17:35',
+            'upload_date': '20210313',
+            'uploader': 'Tagesschau24',
+            'description': '13.03.2021 17:35',
+            'timestamp': 1615656900,
+        }
     }, {
         'url': 'http://www.tagesschau.de/multimedia/sendung/tsg-3771.html',
         'only_matching': True,
@@ -204,13 +194,176 @@ class TagesschauIE(InfoExtractor):
         # playlist article with collapsing sections
         'url': 'http://www.tagesschau.de/wirtschaft/faq-freihandelszone-eu-usa-101.html',
         'only_matching': True,
+    }, {
+        'url': 'https://www.tagesschau.de/',
+        'only_matching': True,
     }]
 
-    @classmethod
+    def _video_id_from_url(self, url):
-    def suitable(cls, url):
+        if url:
-        return False if TagesschauPlayerIE.suitable(url) else super(TagesschauIE, cls).suitable(url)
+            mobj = re.match(self._VALID_URL, url)
+            if mobj:
+                return mobj.group('id')
 
-    def _extract_formats(self, download_text, media_kind):
+    def _handle_generic_titles(self, title, pixelConf):
+        if strip_or_none(title, '').lower() not in ('ganze sendung', '100 sekunden',
+                                                    'tagesschau in 100 sekunden'):
+            return title
+        # otherwise find more meaningful title than the generic Ganze Sendung/100 Sekunden
+        for item in pixelConf:
+            if item.get('tracker') == 'AGFdebug':
+                s = try_get(item, lambda x: x['clipData']['program'], compat_str)
+                if s:
+                    # extract date and time
+                    parts = (try_get(item, lambda x: x['clipData']['title'], compat_str)
+                             or '').split('_')[-2:]
+                    if len(parts) == 2:
+                        title = "%s - %s" % (s, ' '.join(parts))
+                    else:
+                        title = s
+                break
+        return title
+
+    def _extract_from_player(self, player_div, video_id_fallback, title_fallback):
+        player_data = unescapeHTML(self._search_regex(
+            r'data-config=(?P<quote>["\'])(?P<data>[^"\']*)(?P=quote)',
+            player_div, 'data-config', group='data'))
+
+        meta = self._parse_json(player_data, video_id_fallback, fatal=False)
+        mc = try_get(meta, lambda x: x['mc'], dict)
+        if not mc:
+            # fallback if parsing json fails, as tagesschau API sometimes sends
+            # invalid json
+            stream_hls = remove_quotes(self._search_regex(
+                r'"http[^"]+?\.m3u8"', player_data, '.m3u8-url', group=0))
+            formats = self._extract_m3u8_formats(stream_hls, video_id_fallback,
+                                                 ext='mp4', m3u8_id='hls',
+                                                 entry_protocol='m3u8_native')
+            self._sort_formats(formats, field_preference=_FIELD_PREFERENCE)
+            return {
+                'id': video_id_fallback,
+                'title': title_fallback,
+                'formats': formats,
+            }
+
+        # this url is more permanent than the original link
+        webpage_url = url_or_none(try_get(mc, lambda x: x['_sharing']['link']))
+
+        video_id = self._video_id_from_url(webpage_url)
+        duration = None
+        pixelConf = try_get(meta, lambda x: x['pc']['_pixelConfig'], list) or []
+        for item in pixelConf:
+            video_id = (video_id or try_get(item,
+                        [lambda x: x['playerID'],
+                         lambda x: x['clipData']['playerId']], compat_str))
+            duration = (duration or parse_duration(try_get(item,
+                        [lambda x: x['clipData']['length'],
+                         lambda x: x['clipData']['duration']])))
+        if not video_id:
+            video_id = video_id_fallback
+
+        formats = []
+        for elem in mc.get('_mediaArray', []):
+            for d in elem.get('_mediaStreamArray', []):
+                link_url = url_or_none(d.get('_stream'))
+                if not link_url:
+                    continue
+                ext = determine_ext(link_url)
+                if ext == "m3u8":
+                    formats.extend(self._extract_m3u8_formats(
+                        link_url, video_id_fallback, ext='mp4',
+                        entry_protocol='m3u8_native',
+                        m3u8_id='hls', fatal=False))
+                elif ext == "f4m":
+                    formats.extend(self._extract_f4m_formats(
+                        link_url, video_id_fallback, f4m_id='hds', fatal=False))
+                else:
+                    format_id = _normalize_format_id(self._search_regex(
+                        r'.*/[^/.]+\.([^/]+)\.[^/.]+$', link_url, 'format ID',
+                        default=ext, fatal=False),
+                        ext)
+                    fmt = {
+                        'format_id': format_id,
+                        'url': link_url,
+                        'format_name': ext,
+                    }
+                    fmt.update(_FORMATS.get(format_id, {}))
+                    formats.append(fmt)
+        self._sort_formats(formats, field_preference=_FIELD_PREFERENCE)
+        if not formats:
+            raise ExtractorError("could not extract formats from json")
+
+        # note that mc['_title'] can be very different from actual title,
+        # such as an image description in case of audio files
+        title = (try_get(mc, [lambda x: x['_info']['clipTitle'],
+                              lambda x: x['_download']['title']], compat_str)
+                 or title_fallback)
+        title = self._handle_generic_titles(title, pixelConf)
+
+        sub_url = url_or_none(mc.get('_subtitleUrl'))
+        subs = {'de': [{'ext': 'ttml', 'url': sub_url}]} if sub_url else None
+
+        images = try_get(mc, lambda x: x['_previewImage'], dict) or {}
+        thumbnails = [{
+            'url': url_or_none('https://www.tagesschau.de/%s'
+                               % (images[format_id],)),
+            'preference': _FORMATS.get(format_id, {}).get('quality'),
+        } for format_id in images] or None
+
+        return {
+            'id': video_id,
+            'title': title,
+            'formats': formats,
+            'webpage_url': webpage_url,
+            'subtitles': subs,
+            'thumbnails': thumbnails,
+            'duration': duration,
+            'timestamp': unified_timestamp(try_get(mc, [lambda x: x['_download']['date'],
+                                                        lambda x: x['_info']['clipDate']])),
+            'is_live': bool_or_none(mc.get('_isLive')),
+            'channel': try_get(mc, lambda x: x['_download']['channel'], compat_str),
+            'uploader': try_get(mc, lambda x: x['_info']['channelTitle'], compat_str),
+            'creator': try_get(mc, lambda x: x['_info']['clipContentSrc'], compat_str),
+            'description': try_get(mc, lambda x: x['_info']['clipDate'], compat_str),
+        }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id') or mobj.group('path')
+        display_id = video_id.lstrip('-') if video_id else 'tagesschau.de'
+
+        webpage = self._download_webpage(url, display_id)
+
+        title = (self._og_search_title(webpage, default=None)
+                 or self._html_search_regex(
+                     [r'<span[^>]*class="headline"[^>]*>(.+?)</span>',
+                      r'<title[^>]*>(.+?)</title>'],
+                     webpage, 'title'))
+
+        webpage_type = self._og_search_property('type', webpage, default=None)
+
+        player_pattern = r'<div[^>]+data-ts_component=(?P<quote>["\'])ts-mediaplayer(?P=quote)[^>]*>'
+        players = [m.group(0) for m in re.finditer(player_pattern, webpage)]
+        if not players:
+            # assume old website format
+            return self._legacy_extract(webpage, display_id, title, webpage_type)
+        elif (len(players) > 1
+              and not self._downloader.params.get('noplaylist')
+              and (webpage_type == 'website' or not mobj.group('id'))):
+            # article or playlist
+            entries = []
+            seen = set()
+            for player in players:
+                entry = self._extract_from_player(player, video_id, title)
+                if entry['id'] not in seen:
+                    entries.append(entry)
+                    seen.add(entry['id'])
+            return self.playlist_result(entries, display_id, title)
+        else:
+            # single video/audio
+            return self._extract_from_player(players[0], video_id, title)
+
+    def _legacy_extract_formats(self, download_text, media_kind):
         links = re.finditer(
             r'<div class="button" title="(?P<title>[^"]*)"><a href="(?P<url>[^"]+)">(?P<name>.+?)</a></div>',
             download_text)
@@ -219,9 +372,10 @@ class TagesschauIE(InfoExtractor):
             link_url = l.group('url')
             if not link_url:
                 continue
-            format_id = self._search_regex(
+            ext = determine_ext(link_url)
+            format_id = _normalize_format_id(self._search_regex(
                 r'.*/[^/.]+\.([^/]+)\.[^/.]+$', link_url, 'format ID',
-                default=determine_ext(link_url))
+                default=ext), ext)
             format = {
                 'format_id': format_id,
                 'url': l.group('url'),
@@ -262,39 +416,30 @@ class TagesschauIE(InfoExtractor):
         self._sort_formats(formats)
         return formats
 
-    def _real_extract(self, url):
+    # Some old pages still use the old format, so we keep the previous
-        mobj = re.match(self._VALID_URL, url)
+    # extractor for now.
-        video_id = mobj.group('id') or mobj.group('path')
+    def _legacy_extract(self, webpage, display_id, title, webpage_type):
-        display_id = video_id.lstrip('-')
+        DOWNLOAD_REGEX = r'<p>Wir bieten dieses (?P<kind>Video|Audio) in folgenden Formaten zum Download an:</p>\s*<div class="controls">(?P<links>.*?)</div>\s*<p>'
 
-        webpage = self._download_webpage(url, display_id)
-
-        title = self._html_search_regex(
-            r'<span[^>]*class="headline"[^>]*>(.+?)</span>',
-            webpage, 'title', default=None) or self._og_search_title(webpage)
-
-        DOWNLOAD_REGEX = r'(?s)<p>Wir bieten dieses (?P<kind>Video|Audio) in folgenden Formaten zum Download an:</p>\s*<div class="controls">(?P<links>.*?)</div>\s*<p>'
-
-        webpage_type = self._og_search_property('type', webpage, default=None)
         if webpage_type == 'website':  # Article
             entries = []
             for num, (entry_title, media_kind, download_text) in enumerate(re.findall(
-                    r'(?s)<p[^>]+class="infotext"[^>]*>\s*(?:<a[^>]+>)?\s*<strong>(.+?)</strong>.*?</p>.*?%s' % DOWNLOAD_REGEX,
+                    r'<p[^>]+class="infotext"[^>]*>\s*(?:<a[^>]+>)?\s*<strong>(.+?)</strong>.*?</p>.*?%s' % DOWNLOAD_REGEX,
-                    webpage), 1):
+                    webpage, flags=re.S), 1):
                 entries.append({
                     'id': '%s-%d' % (display_id, num),
                     'title': '%s' % entry_title,
-                    'formats': self._extract_formats(download_text, media_kind),
+                    'formats': self._legacy_extract_formats(download_text, media_kind),
                 })
-            if len(entries) > 1:
+            if len(entries) != 1:
                 return self.playlist_result(entries, display_id, title)
             formats = entries[0]['formats']
         else:  # Assume single video
             download_text = self._search_regex(
-                DOWNLOAD_REGEX, webpage, 'download links', group='links')
+                DOWNLOAD_REGEX, webpage, 'download links', flags=re.S, group='links')
             media_kind = self._search_regex(
-                DOWNLOAD_REGEX, webpage, 'media kind', default='Video', group='kind')
+                DOWNLOAD_REGEX, webpage, 'media kind', default='Video', flags=re.S, group='kind')
-            formats = self._extract_formats(download_text, media_kind)
+            formats = self._legacy_extract_formats(download_text, media_kind)
         thumbnail = self._og_search_thumbnail(webpage)
         description = self._html_search_regex(
             r'(?s)<p class="teasertext">(.*?)</p>',

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,
 )
 
 
@@ -89,12 +91,12 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'IOS',
-                    'clientVersion': '19.45.4',
+                    'clientVersion': '20.10.4',
                     'deviceMake': 'Apple',
                     'deviceModel': 'iPhone16,2',
-                    'userAgent': 'com.google.ios.youtube/19.45.4 (iPhone16,2; U; CPU iOS 18_1_0 like Mac OS X;)',
+                    'userAgent': 'com.google.ios.youtube/20.10.4 (iPhone16,2; U; CPU iOS 18_3_2 like Mac OS X;)',
                     'osName': 'iPhone',
-                    'osVersion': '18.1.0.22B83',
+                    'osVersion': '18.3.2.22D82',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 5,
@@ -107,7 +109,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'MWEB',
-                    'clientVersion': '2.20241202.07.00',
+                    'clientVersion': '2.20250311.03.00',
                     # mweb previously did not require PO Token with this UA
                     'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)',
                 },
@@ -120,7 +122,8 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'TVHTML5',
-                    'clientVersion': '7.20241201.18.00',
+                    'clientVersion': '7.20250312.16.00',
+                    'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/Version',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 7,
@@ -130,7 +133,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
             'INNERTUBE_CONTEXT': {
                 'client': {
                     'clientName': 'WEB',
-                    'clientVersion': '2.20241126.01.00',
+                    'clientVersion': '2.20250312.04.00',
                 },
             },
             'INNERTUBE_CONTEXT_CLIENT_NAME': 1,
@@ -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': {
@@ -669,7 +692,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/player',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
         r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
         r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
     )
@@ -1829,11 +1852,21 @@ 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\("\W+"\))(?=\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])))
 
+        if player_id:
             self.cache.store('youtube-nsig', player_id, func_code)
         return jsi, player_id, 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')