Merge branch 'yt-dlp:master' into feat/ie-next-flight

test/test_InfoExtractor.py

@@ -1975,6 +1975,137 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
         self.assertEqual(self.ie._search_nextjs_v13_data('', None, fatal=False), [])
         self.assertEqual(self.ie._search_nextjs_v13_data(None, None, fatal=False), [])
 
+    def test_search_nuxt_json(self):
+        HTML_TMPL = '<script data-ssr="true" id="__NUXT_DATA__" type="application/json">[{}]</script>'
+        VALID_DATA = '''
+            ["ShallowReactive",1],
+            {"data":2,"state":21,"once":25,"_errors":28,"_server_errors":30},
+            ["ShallowReactive",3],
+            {"$abcdef123456":4},
+            {"podcast":5,"activeEpisodeData":7},
+            {"podcast":6,"seasons":14},
+            {"title":10,"id":11},
+            ["Reactive",8],
+            {"episode":9,"creators":18,"empty_list":20},
+            {"title":12,"id":13,"refs":34,"empty_refs":35},
+            "Series Title",
+            "podcast-id-01",
+            "Episode Title",
+            "episode-id-99",
+            [15,16,17],
+            1,
+            2,
+            3,
+            [19],
+            "Podcast Creator",
+            [],
+            {"$ssite-config":22},
+            {"env":23,"name":24,"map":26,"numbers":14},
+            "production",
+            "podcast-website",
+            ["Set"],
+            ["Reactive",27],
+            ["Map"],
+            ["ShallowReactive",29],
+            {},
+            ["NuxtError",31],
+            {"status":32,"message":33},
+            503,
+            "Service Unavailable",
+            [36,37],
+            [38,39],
+            ["Ref",40],
+            ["ShallowRef",41],
+            ["EmptyRef",42],
+            ["EmptyShallowRef",43],
+            "ref",
+            "shallow_ref",
+            "{\\"ref\\":1}",
+            "{\\"shallow_ref\\":2}"
+        '''
+        PAYLOAD = {
+            'data': {
+                '$abcdef123456': {
+                    'podcast': {
+                        'podcast': {
+                            'title': 'Series Title',
+                            'id': 'podcast-id-01',
+                        },
+                        'seasons': [1, 2, 3],
+                    },
+                    'activeEpisodeData': {
+                        'episode': {
+                            'title': 'Episode Title',
+                            'id': 'episode-id-99',
+                            'refs': ['ref', 'shallow_ref'],
+                            'empty_refs': [{'ref': 1}, {'shallow_ref': 2}],
+                        },
+                        'creators': ['Podcast Creator'],
+                        'empty_list': [],
+                    },
+                },
+            },
+            'state': {
+                '$ssite-config': {
+                    'env': 'production',
+                    'name': 'podcast-website',
+                    'map': [],
+                    'numbers': [1, 2, 3],
+                },
+            },
+            'once': [],
+            '_errors': {},
+            '_server_errors': {
+                'status': 503,
+                'message': 'Service Unavailable',
+            },
+        }
+        PARTIALLY_INVALID = [(
+            '''
+            {"data":1},
+            {"invalid_raw_list":2},
+            [15,16,17]
+            ''',
+            {'data': {'invalid_raw_list': [None, None, None]}},
+        ), (
+            '''
+            {"data":1},
+            ["EmptyRef",2],
+            "not valid JSON"
+            ''',
+            {'data': None},
+        ), (
+            '''
+            {"data":1},
+            ["EmptyShallowRef",2],
+            "not valid JSON"
+            ''',
+            {'data': None},
+        )]
+        INVALID = [
+            '''
+                []
+            ''',
+            '''
+                ["unsupported",1],
+                {"data":2},
+                {}
+            ''',
+        ]
+        DEFAULT = object()
+
+        self.assertEqual(self.ie._search_nuxt_json(HTML_TMPL.format(VALID_DATA), None), PAYLOAD)
+        self.assertEqual(self.ie._search_nuxt_json('', None, fatal=False), {})
+        self.assertIs(self.ie._search_nuxt_json('', None, default=DEFAULT), DEFAULT)
+
+        for data, expected in PARTIALLY_INVALID:
+            self.assertEqual(
+                self.ie._search_nuxt_json(HTML_TMPL.format(data), None, fatal=False), expected)
+
+        for data in INVALID:
+            self.assertIs(
+                self.ie._search_nuxt_json(HTML_TMPL.format(data), None, default=DEFAULT), DEFAULT)
+
 
 if __name__ == '__main__':
     unittest.main()

test/test_devalue.py

@@ -0,0 +1,235 @@
+#!/usr/bin/env python3
+
+# Allow direct execution
+import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+
+import datetime as dt
+import json
+import math
+import re
+import unittest
+
+from yt_dlp.utils.jslib import devalue
+
+
+TEST_CASES_EQUALS = [{
+    'name': 'int',
+    'unparsed': [-42],
+    'parsed': -42,
+}, {
+    'name': 'str',
+    'unparsed': ['woo!!!'],
+    'parsed': 'woo!!!',
+}, {
+    'name': 'Number',
+    'unparsed': [['Object', 42]],
+    'parsed': 42,
+}, {
+    'name': 'String',
+    'unparsed': [['Object', 'yar']],
+    'parsed': 'yar',
+}, {
+    'name': 'Infinity',
+    'unparsed': -4,
+    'parsed': math.inf,
+}, {
+    'name': 'negative Infinity',
+    'unparsed': -5,
+    'parsed': -math.inf,
+}, {
+    'name': 'negative zero',
+    'unparsed': -6,
+    'parsed': -0.0,
+}, {
+    'name': 'RegExp',
+    'unparsed': [['RegExp', 'regexp', 'gim']],  # XXX: flags are ignored
+    'parsed': re.compile('regexp'),
+}, {
+    'name': 'Date',
+    'unparsed': [['Date', '2001-09-09T01:46:40.000Z']],
+    'parsed': dt.datetime.fromtimestamp(1e9, tz=dt.timezone.utc),
+}, {
+    'name': 'Array',
+    'unparsed': [[1, 2, 3], 'a', 'b', 'c'],
+    'parsed': ['a', 'b', 'c'],
+}, {
+    'name': 'Array (empty)',
+    'unparsed': [[]],
+    'parsed': [],
+}, {
+    'name': 'Array (sparse)',
+    'unparsed': [[-2, 1, -2], 'b'],
+    'parsed': [None, 'b', None],
+}, {
+    'name': 'Object',
+    'unparsed': [{'foo': 1, 'x-y': 2}, 'bar', 'z'],
+    'parsed': {'foo': 'bar', 'x-y': 'z'},
+}, {
+    'name': 'Set',
+    'unparsed': [['Set', 1, 2, 3], 1, 2, 3],
+    'parsed': [1, 2, 3],
+}, {
+    'name': 'Map',
+    'unparsed': [['Map', 1, 2], 'a', 'b'],
+    'parsed': [['a', 'b']],
+}, {
+    'name': 'BigInt',
+    'unparsed': [['BigInt', '1']],
+    'parsed': 1,
+}, {
+    'name': 'Uint8Array',
+    'unparsed': [['Uint8Array', 'AQID']],
+    'parsed': [1, 2, 3],
+}, {
+    'name': 'ArrayBuffer',
+    'unparsed': [['ArrayBuffer', 'AQID']],
+    'parsed': [1, 2, 3],
+}, {
+    'name': 'str (repetition)',
+    'unparsed': [[1, 1], 'a string'],
+    'parsed': ['a string', 'a string'],
+}, {
+    'name': 'None (repetition)',
+    'unparsed': [[1, 1], None],
+    'parsed': [None, None],
+}, {
+    'name': 'dict (repetition)',
+    'unparsed': [[1, 1], {}],
+    'parsed': [{}, {}],
+}, {
+    'name': 'Object without prototype',
+    'unparsed': [['null']],
+    'parsed': {},
+}, {
+    'name': 'cross-realm POJO',
+    'unparsed': [{}],
+    'parsed': {},
+}]
+
+TEST_CASES_IS = [{
+    'name': 'bool',
+    'unparsed': [True],
+    'parsed': True,
+}, {
+    'name': 'Boolean',
+    'unparsed': [['Object', False]],
+    'parsed': False,
+}, {
+    'name': 'undefined',
+    'unparsed': -1,
+    'parsed': None,
+}, {
+    'name': 'null',
+    'unparsed': [None],
+    'parsed': None,
+}, {
+    'name': 'NaN',
+    'unparsed': -3,
+    'parsed': math.nan,
+}]
+
+TEST_CASES_INVALID = [{
+    'name': 'empty string',
+    'unparsed': '',
+    'error': ValueError,
+    'pattern': r'expected int or list as input',
+}, {
+    'name': 'hole',
+    'unparsed': -2,
+    'error': ValueError,
+    'pattern': r'invalid integer input',
+}, {
+    'name': 'string',
+    'unparsed': 'hello',
+    'error': ValueError,
+    'pattern': r'expected int or list as input',
+}, {
+    'name': 'number',
+    'unparsed': 42,
+    'error': ValueError,
+    'pattern': r'invalid integer input',
+}, {
+    'name': 'boolean',
+    'unparsed': True,
+    'error': ValueError,
+    'pattern': r'expected int or list as input',
+}, {
+    'name': 'null',
+    'unparsed': None,
+    'error': ValueError,
+    'pattern': r'expected int or list as input',
+}, {
+    'name': 'object',
+    'unparsed': {},
+    'error': ValueError,
+    'pattern': r'expected int or list as input',
+}, {
+    'name': 'empty array',
+    'unparsed': [],
+    'error': ValueError,
+    'pattern': r'expected a non-empty list as input',
+}, {
+    'name': 'Python negative indexing',
+    'unparsed': [[1, 2, 3, 4, 5, 6, 7, -7], 1, 2, 3, 4, 5, 6, 7],
+    'error': IndexError,
+    'pattern': r'invalid index: -7',
+}]
+
+
+class TestDevalue(unittest.TestCase):
+    def test_devalue_parse_equals(self):
+        for tc in TEST_CASES_EQUALS:
+            self.assertEqual(devalue.parse(tc['unparsed']), tc['parsed'], tc['name'])
+
+    def test_devalue_parse_is(self):
+        for tc in TEST_CASES_IS:
+            self.assertIs(devalue.parse(tc['unparsed']), tc['parsed'], tc['name'])
+
+    def test_devalue_parse_invalid(self):
+        for tc in TEST_CASES_INVALID:
+            with self.assertRaisesRegex(tc['error'], tc['pattern'], msg=tc['name']):
+                devalue.parse(tc['unparsed'])
+
+    def test_devalue_parse_cyclical(self):
+        name = 'Map (cyclical)'
+        result = devalue.parse([['Map', 1, 0], 'self'])
+        self.assertEqual(result[0][0], 'self', name)
+        self.assertIs(result, result[0][1], name)
+
+        name = 'Set (cyclical)'
+        result = devalue.parse([['Set', 0, 1], 42])
+        self.assertEqual(result[1], 42, name)
+        self.assertIs(result, result[0], name)
+
+        result = devalue.parse([[0]])
+        self.assertIs(result, result[0], 'Array (cyclical)')
+
+        name = 'Object (cyclical)'
+        result = devalue.parse([{'self': 0}])
+        self.assertIs(result, result['self'], name)
+
+        name = 'Object with null prototype (cyclical)'
+        result = devalue.parse([['null', 'self', 0]])
+        self.assertIs(result, result['self'], name)
+
+        name = 'Objects (cyclical)'
+        result = devalue.parse([[1, 2], {'second': 2}, {'first': 1}])
+        self.assertIs(result[0], result[1]['first'], name)
+        self.assertIs(result[1], result[0]['second'], name)
+
+    def test_devalue_parse_revivers(self):
+        self.assertEqual(
+            devalue.parse([['indirect', 1], {'a': 2}, 'b'], revivers={'indirect': lambda x: x}),
+            {'a': 'b'}, 'revivers (indirect)')
+
+        self.assertEqual(
+            devalue.parse([['parse', 1], '{"a":0}'], revivers={'parse': lambda x: json.loads(x)}),
+            {'a': 0}, 'revivers (parse)')
+
+
+if __name__ == '__main__':
+    unittest.main()

test/test_pot/test_pot_builtin_utils.py

@@ -11,7 +11,7 @@ class TestGetWebPoContentBinding:
 
     @pytest.mark.parametrize('client_name, context, is_authenticated, expected', [
         *[(client, context, is_authenticated, expected) for client in [
-            'WEB', 'MWEB', 'TVHTML5', 'WEB_EMBEDDED_PLAYER', 'WEB_CREATOR', 'TVHTML5_SIMPLY_EMBEDDED_PLAYER']
+            'WEB', 'MWEB', 'TVHTML5', 'WEB_EMBEDDED_PLAYER', 'WEB_CREATOR', 'TVHTML5_SIMPLY_EMBEDDED_PLAYER', 'TVHTML5_SIMPLY']
           for context, is_authenticated, expected in [
             (PoTokenContext.GVS, False, ('example-visitor-data', ContentBindingType.VISITOR_DATA)),
             (PoTokenContext.PLAYER, False, ('example-video-id', ContentBindingType.VIDEO_ID)),

test/test_pot/test_pot_builtin_webpospec.py

@@ -49,7 +49,7 @@ class TestWebPoPCSP:
 
     @pytest.mark.parametrize('client_name, context, is_authenticated, remote_host, source_address, request_proxy, expected', [
         *[(client, context, is_authenticated, remote_host, source_address, request_proxy, expected) for client in [
-            'WEB', 'MWEB', 'TVHTML5', 'WEB_EMBEDDED_PLAYER', 'WEB_CREATOR', 'TVHTML5_SIMPLY_EMBEDDED_PLAYER']
+            'WEB', 'MWEB', 'TVHTML5', 'WEB_EMBEDDED_PLAYER', 'WEB_CREATOR', 'TVHTML5_SIMPLY_EMBEDDED_PLAYER', 'TVHTML5_SIMPLY']
           for context, is_authenticated, remote_host, source_address, request_proxy, expected in [
             (PoTokenContext.GVS, False, 'example-remote-host', 'example-source-address', 'example-request-proxy', {'t': 'webpo', 'ip': 'example-remote-host', 'sa': 'example-source-address', 'px': 'example-request-proxy', 'cb': '123abcXYZ_-', 'cbt': 'visitor_id'}),
             (PoTokenContext.PLAYER, False, 'example-remote-host', 'example-source-address', 'example-request-proxy', {'t': 'webpo', 'ip': 'example-remote-host', 'sa': 'example-source-address', 'px': 'example-request-proxy', 'cb': '123abcXYZ_-', 'cbt': 'video_id'}),

test/test_traversal.py

@@ -416,18 +416,8 @@ class TestTraversal:
             '`any` should allow further branching'
 
     def test_traversal_morsel(self):
-        values = {
-            'expires': 'a',
-            'path': 'b',
-            'comment': 'c',
-            'domain': 'd',
-            'max-age': 'e',
-            'secure': 'f',
-            'httponly': 'g',
-            'version': 'h',
-            'samesite': 'i',
-        }
         morsel = http.cookies.Morsel()
+        values = dict(zip(morsel, 'abcdefghijklmnop'))
         morsel.set('item_key', 'item_value', 'coded_value')
         morsel.update(values)
         values['key'] = 'item_key'

test/test_youtube_signature.py

@@ -320,6 +320,14 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/59b252b9/player_ias.vflset/en_US/base.js',
         'D3XWVpYgwhLLKNK4AGX', 'aZrQ1qWJ5yv5h',
     ),
+    (
+        'https://www.youtube.com/s/player/fc2a56a5/player_ias.vflset/en_US/base.js',
+        'qTKWg_Il804jd2kAC', 'OtUAm2W6gyzJjB9u',
+    ),
+    (
+        'https://www.youtube.com/s/player/fc2a56a5/tv-player-ias.vflset/tv-player-ias.js',
+        'qTKWg_Il804jd2kAC', 'OtUAm2W6gyzJjB9u',
+    ),
 ]