test/test_traversal.py

   1 import http.cookies
   2 import re
   3 import xml.etree.ElementTree
   4
   5 import pytest
   6
   7 from yt_dlp.utils import (
   8     ExtractorError,
   9     determine_ext,
  10     dict_get,
  11     int_or_none,
  12     str_or_none,
  13 )
  14 from yt_dlp.utils.traversal import (
  15     traverse_obj,
  16     require,
  17     subs_list_to_dict,
  18 )
  19
  20 _TEST_DATA = {
  21     100: 100,
  22     1.2: 1.2,
  23     'str': 'str',
  24     'None': None,
  25     '...': ...,
  26     'urls': [
  27         {'index': 0, 'url': 'https://www.example.com/0'},
  28         {'index': 1, 'url': 'https://www.example.com/1'},
  29     ],
  30     'data': (
  31         {'index': 2},
  32         {'index': 3},
  33     ),
  34     'dict': {},
  35 }
  36
  37
  38 class TestTraversal:
  39     def test_traversal_base(self):
  40         assert traverse_obj(_TEST_DATA, ('str',)) == 'str', \
  41             'allow tuple path'
  42         assert traverse_obj(_TEST_DATA, ['str']) == 'str', \
  43             'allow list path'
  44         assert traverse_obj(_TEST_DATA, (value for value in ('str',))) == 'str', \
  45             'allow iterable path'
  46         assert traverse_obj(_TEST_DATA, 'str') == 'str', \
  47             'single items should be treated as a path'
  48         assert traverse_obj(_TEST_DATA, 100) == 100, \
  49             'allow int path'
  50         assert traverse_obj(_TEST_DATA, 1.2) == 1.2, \
  51             'allow float path'
  52         assert traverse_obj(_TEST_DATA, None) == _TEST_DATA, \
  53             '`None` should not perform any modification'
  54
  55     def test_traversal_ellipsis(self):
  56         assert traverse_obj(_TEST_DATA, ...) == [x for x in _TEST_DATA.values() if x not in (None, {})], \
  57             '`...` should give all non discarded values'
  58         assert traverse_obj(_TEST_DATA, ('urls', 0, ...)) == list(_TEST_DATA['urls'][0].values()), \
  59             '`...` selection for dicts should select all values'
  60         assert traverse_obj(_TEST_DATA, (..., ..., 'url')) == ['https://www.example.com/0', 'https://www.example.com/1'], \
  61             'nested `...` queries should work'
  62         assert traverse_obj(_TEST_DATA, (..., ..., 'index')) == list(range(4)), \
  63             '`...` query result should be flattened'
  64         assert traverse_obj(iter(range(4)), ...) == list(range(4)), \
  65             '`...` should accept iterables'
  66
  67     def test_traversal_function(self):
  68         filter_func = lambda x, y: x == 'urls' and isinstance(y, list)
  69         assert traverse_obj(_TEST_DATA, filter_func) == [_TEST_DATA['urls']], \
  70             'function as query key should perform a filter based on (key, value)'
  71         assert traverse_obj(_TEST_DATA, lambda _, x: isinstance(x[0], str)) == ['str'], \
  72             'exceptions in the query function should be catched'
  73         assert traverse_obj(iter(range(4)), lambda _, x: x % 2 == 0) == [0, 2], \
  74             'function key should accept iterables'
  75         # Wrong function signature should raise (debug mode)
  76         with pytest.raises(Exception):
  77             traverse_obj(_TEST_DATA, lambda a: ...)
  78         with pytest.raises(Exception):
  79             traverse_obj(_TEST_DATA, lambda a, b, c: ...)
  80
  81     def test_traversal_set(self):
  82         # transformation/type, like `expected_type`
  83         assert traverse_obj(_TEST_DATA, (..., {str.upper})) == ['STR'], \
  84             'Function in set should be a transformation'
  85         assert traverse_obj(_TEST_DATA, (..., {str})) == ['str'], \
  86             'Type in set should be a type filter'
  87         assert traverse_obj(_TEST_DATA, (..., {str, int})) == [100, 'str'], \
  88             'Multiple types in set should be a type filter'
  89         assert traverse_obj(_TEST_DATA, {dict}) == _TEST_DATA, \
  90             'A single set should be wrapped into a path'
  91         assert traverse_obj(_TEST_DATA, (..., {str.upper})) == ['STR'], \
  92             'Transformation function should not raise'
  93         expected = [x for x in map(str_or_none, _TEST_DATA.values()) if x is not None]
  94         assert traverse_obj(_TEST_DATA, (..., {str_or_none})) == expected, \
  95             'Function in set should be a transformation'
  96         assert traverse_obj(_TEST_DATA, ('fail', {lambda _: 'const'})) == 'const', \
  97             'Function in set should always be called'
  98         # Sets with length < 1 or > 1 not including only types should raise
  99         with pytest.raises(Exception):
 100             traverse_obj(_TEST_DATA, set())
 101         with pytest.raises(Exception):
 102             traverse_obj(_TEST_DATA, {str.upper, str})
 103
 104     def test_traversal_slice(self):
 105         _SLICE_DATA = [0, 1, 2, 3, 4]
 106
 107         assert traverse_obj(_TEST_DATA, ('dict', slice(1))) is None, \
 108             'slice on a dictionary should not throw'
 109         assert traverse_obj(_SLICE_DATA, slice(1)) == _SLICE_DATA[:1], \
 110             'slice key should apply slice to sequence'
 111         assert traverse_obj(_SLICE_DATA, slice(1, 2)) == _SLICE_DATA[1:2], \
 112             'slice key should apply slice to sequence'
 113         assert traverse_obj(_SLICE_DATA, slice(1, 4, 2)) == _SLICE_DATA[1:4:2], \
 114             'slice key should apply slice to sequence'
 115
 116     def test_traversal_alternatives(self):
 117         assert traverse_obj(_TEST_DATA, 'fail', 'str') == 'str', \
 118             'multiple `paths` should be treated as alternative paths'
 119         assert traverse_obj(_TEST_DATA, 'str', 100) == 'str', \
 120             'alternatives should exit early'
 121         assert traverse_obj(_TEST_DATA, 'fail', 'fail') is None, \
 122             'alternatives should return `default` if exhausted'
 123         assert traverse_obj(_TEST_DATA, (..., 'fail'), 100) == 100, \
 124             'alternatives should track their own branching return'
 125         assert traverse_obj(_TEST_DATA, ('dict', ...), ('data', ...)) == list(_TEST_DATA['data']), \
 126             'alternatives on empty objects should search further'
 127
 128     def test_traversal_branching_nesting(self):
 129         assert traverse_obj(_TEST_DATA, ('urls', (3, 0), 'url')) == ['https://www.example.com/0'], \
 130             'tuple as key should be treated as branches'
 131         assert traverse_obj(_TEST_DATA, ('urls', [3, 0], 'url')) == ['https://www.example.com/0'], \
 132             'list as key should be treated as branches'
 133         assert traverse_obj(_TEST_DATA, ('urls', ((1, 'fail'), (0, 'url')))) == ['https://www.example.com/0'], \
 134             'double nesting in path should be treated as paths'
 135         assert traverse_obj(['0', [1, 2]], [(0, 1), 0]) == [1], \
 136             'do not fail early on branching'
 137         expected = ['https://www.example.com/0', 'https://www.example.com/1']
 138         assert traverse_obj(_TEST_DATA, ('urls', ((0, ('fail', 'url')), (1, 'url')))) == expected, \
 139             'tripple nesting in path should be treated as branches'
 140         assert traverse_obj(_TEST_DATA, ('urls', ('fail', (..., 'url')))) == expected, \
 141             'ellipsis as branch path start gets flattened'
 142
 143     def test_traversal_dict(self):
 144         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}) == {0: 100, 1: 1.2}, \
 145             'dict key should result in a dict with the same keys'
 146         expected = {0: 'https://www.example.com/0'}
 147         assert traverse_obj(_TEST_DATA, {0: ('urls', 0, 'url')}) == expected, \
 148             'dict key should allow paths'
 149         expected = {0: ['https://www.example.com/0']}
 150         assert traverse_obj(_TEST_DATA, {0: ('urls', (3, 0), 'url')}) == expected, \
 151             'tuple in dict path should be treated as branches'
 152         assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, 'fail'), (0, 'url')))}) == expected, \
 153             'double nesting in dict path should be treated as paths'
 154         expected = {0: ['https://www.example.com/1', 'https://www.example.com/0']}
 155         assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, ('fail', 'url')), (0, 'url')))}) == expected, \
 156             'tripple nesting in dict path should be treated as branches'
 157         assert traverse_obj(_TEST_DATA, {0: 'fail'}) == {}, \
 158             'remove `None` values when top level dict key fails'
 159         assert traverse_obj(_TEST_DATA, {0: 'fail'}, default=...) == {0: ...}, \
 160             'use `default` if key fails and `default`'
 161         assert traverse_obj(_TEST_DATA, {0: 'dict'}) == {}, \
 162             'remove empty values when dict key'
 163         assert traverse_obj(_TEST_DATA, {0: 'dict'}, default=...) == {0: ...}, \
 164             'use `default` when dict key and `default`'
 165         assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}) == {}, \
 166             'remove empty values when nested dict key fails'
 167         assert traverse_obj(None, {0: 'fail'}) == {}, \
 168             'default to dict if pruned'
 169         assert traverse_obj(None, {0: 'fail'}, default=...) == {0: ...}, \
 170             'default to dict if pruned and default is given'
 171         assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}, default=...) == {0: {0: ...}}, \
 172             'use nested `default` when nested dict key fails and `default`'
 173         assert traverse_obj(_TEST_DATA, {0: ('dict', ...)}) == {}, \
 174             'remove key if branch in dict key not successful'
 175
 176     def test_traversal_default(self):
 177         _DEFAULT_DATA = {'None': None, 'int': 0, 'list': []}
 178
 179         assert traverse_obj(_DEFAULT_DATA, 'fail') is None, \
 180             'default value should be `None`'
 181         assert traverse_obj(_DEFAULT_DATA, 'fail', 'fail', default=...) == ..., \
 182             'chained fails should result in default'
 183         assert traverse_obj(_DEFAULT_DATA, 'None', 'int') == 0, \
 184             'should not short cirquit on `None`'
 185         assert traverse_obj(_DEFAULT_DATA, 'fail', default=1) == 1, \
 186             'invalid dict key should result in `default`'
 187         assert traverse_obj(_DEFAULT_DATA, 'None', default=1) == 1, \
 188             '`None` is a deliberate sentinel and should become `default`'
 189         assert traverse_obj(_DEFAULT_DATA, ('list', 10)) is None, \
 190             '`IndexError` should result in `default`'
 191         assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=1) == 1, \
 192             'if branched but not successful return `default` if defined, not `[]`'
 193         assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=None) is None, \
 194             'if branched but not successful return `default` even if `default` is `None`'
 195         assert traverse_obj(_DEFAULT_DATA, (..., 'fail')) == [], \
 196             'if branched but not successful return `[]`, not `default`'
 197         assert traverse_obj(_DEFAULT_DATA, ('list', ...)) == [], \
 198             'if branched but object is empty return `[]`, not `default`'
 199         assert traverse_obj(None, ...) == [], \
 200             'if branched but object is `None` return `[]`, not `default`'
 201         assert traverse_obj({0: None}, (0, ...)) == [], \
 202             'if branched but state is `None` return `[]`, not `default`'
 203
 204     @pytest.mark.parametrize('path', [
 205         ('fail', ...),
 206         (..., 'fail'),
 207         100 * ('fail',) + (...,),
 208         (...,) + 100 * ('fail',),
 209     ])
 210     def test_traversal_branching(self, path):
 211         assert traverse_obj({}, path) == [], \
 212             'if branched but state is `None`, return `[]` (not `default`)'
 213         assert traverse_obj({}, 'fail', path) == [], \
 214             'if branching in last alternative and previous did not match, return `[]` (not `default`)'
 215         assert traverse_obj({0: 'x'}, 0, path) == 'x', \
 216             'if branching in last alternative and previous did match, return single value'
 217         assert traverse_obj({0: 'x'}, path, 0) == 'x', \
 218             'if branching in first alternative and non-branching path does match, return single value'
 219         assert traverse_obj({}, path, 'fail') is None, \
 220             'if branching in first alternative and non-branching path does not match, return `default`'
 221
 222     def test_traversal_expected_type(self):
 223         _EXPECTED_TYPE_DATA = {'str': 'str', 'int': 0}
 224
 225         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=str) == 'str', \
 226             'accept matching `expected_type` type'
 227         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=int) is None, \
 228             'reject non matching `expected_type` type'
 229         assert traverse_obj(_EXPECTED_TYPE_DATA, 'int', expected_type=lambda x: str(x)) == '0', \
 230             'transform type using type function'
 231         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=lambda _: 1 / 0) is None, \
 232             'wrap expected_type fuction in try_call'
 233         assert traverse_obj(_EXPECTED_TYPE_DATA, ..., expected_type=str) == ['str'], \
 234             'eliminate items that expected_type fails on'
 235         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}, expected_type=int) == {0: 100}, \
 236             'type as expected_type should filter dict values'
 237         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2, 2: 'None'}, expected_type=str_or_none) == {0: '100', 1: '1.2'}, \
 238             'function as expected_type should transform dict values'
 239         assert traverse_obj(_TEST_DATA, ({0: 1.2}, 0, {int_or_none}), expected_type=int) == 1, \
 240             'expected_type should not filter non final dict values'
 241         assert traverse_obj(_TEST_DATA, {0: {0: 100, 1: 'str'}}, expected_type=int) == {0: {0: 100}}, \
 242             'expected_type should transform deep dict values'
 243         assert traverse_obj(_TEST_DATA, [({0: '...'}, {0: '...'})], expected_type=type(...)) == [{0: ...}, {0: ...}], \
 244             'expected_type should transform branched dict values'
 245         assert traverse_obj({1: {3: 4}}, [(1, 2), 3], expected_type=int) == [4], \
 246             'expected_type regression for type matching in tuple branching'
 247         assert traverse_obj(_TEST_DATA, ['data', ...], expected_type=int) == [], \
 248             'expected_type regression for type matching in dict result'
 249
 250     def test_traversal_get_all(self):
 251         _GET_ALL_DATA = {'key': [0, 1, 2]}
 252
 253         assert traverse_obj(_GET_ALL_DATA, ('key', ...), get_all=False) == 0, \
 254             'if not `get_all`, return only first matching value'
 255         assert traverse_obj(_GET_ALL_DATA, ..., get_all=False) == [0, 1, 2], \
 256             'do not overflatten if not `get_all`'
 257
 258     def test_traversal_casesense(self):
 259         _CASESENSE_DATA = {
 260             'KeY': 'value0',
 261             0: {
 262                 'KeY': 'value1',
 263                 0: {'KeY': 'value2'},
 264             },
 265         }
 266
 267         assert traverse_obj(_CASESENSE_DATA, 'key') is None, \
 268             'dict keys should be case sensitive unless `casesense`'
 269         assert traverse_obj(_CASESENSE_DATA, 'keY', casesense=False) == 'value0', \
 270             'allow non matching key case if `casesense`'
 271         assert traverse_obj(_CASESENSE_DATA, [0, ('keY',)], casesense=False) == ['value1'], \
 272             'allow non matching key case in branch if `casesense`'
 273         assert traverse_obj(_CASESENSE_DATA, [0, ([0, 'keY'],)], casesense=False) == ['value2'], \
 274             'allow non matching key case in branch path if `casesense`'
 275
 276     def test_traversal_traverse_string(self):
 277         _TRAVERSE_STRING_DATA = {'str': 'str', 1.2: 1.2}
 278
 279         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0)) is None, \
 280             'do not traverse into string if not `traverse_string`'
 281         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0), traverse_string=True) == 's', \
 282             'traverse into string if `traverse_string`'
 283         assert traverse_obj(_TRAVERSE_STRING_DATA, (1.2, 1), traverse_string=True) == '.', \
 284             'traverse into converted data if `traverse_string`'
 285         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', ...), traverse_string=True) == 'str', \
 286             '`...` should result in string (same value) if `traverse_string`'
 287         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', slice(0, None, 2)), traverse_string=True) == 'sr', \
 288             '`slice` should result in string if `traverse_string`'
 289         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', lambda i, v: i or v == 's'), traverse_string=True) == 'str', \
 290             'function should result in string if `traverse_string`'
 291         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', (0, 2)), traverse_string=True) == ['s', 'r'], \
 292             'branching should result in list if `traverse_string`'
 293         assert traverse_obj({}, (0, ...), traverse_string=True) == [], \
 294             'branching should result in list if `traverse_string`'
 295         assert traverse_obj({}, (0, lambda x, y: True), traverse_string=True) == [], \
 296             'branching should result in list if `traverse_string`'
 297         assert traverse_obj({}, (0, slice(1)), traverse_string=True) == [], \
 298             'branching should result in list if `traverse_string`'
 299
 300     def test_traversal_re(self):
 301         mobj = re.fullmatch(r'0(12)(?P<group>3)(4)?', '0123')
 302         assert traverse_obj(mobj, ...) == [x for x in mobj.groups() if x is not None], \
 303             '`...` on a `re.Match` should give its `groups()`'
 304         assert traverse_obj(mobj, lambda k, _: k in (0, 2)) == ['0123', '3'], \
 305             'function on a `re.Match` should give groupno, value starting at 0'
 306         assert traverse_obj(mobj, 'group') == '3', \
 307             'str key on a `re.Match` should give group with that name'
 308         assert traverse_obj(mobj, 2) == '3', \
 309             'int key on a `re.Match` should give group with that name'
 310         assert traverse_obj(mobj, 'gRoUp', casesense=False) == '3', \
 311             'str key on a `re.Match` should respect casesense'
 312         assert traverse_obj(mobj, 'fail') is None, \
 313             'failing str key on a `re.Match` should return `default`'
 314         assert traverse_obj(mobj, 'gRoUpS', casesense=False) is None, \
 315             'failing str key on a `re.Match` should return `default`'
 316         assert traverse_obj(mobj, 8) is None, \
 317             'failing int key on a `re.Match` should return `default`'
 318         assert traverse_obj(mobj, lambda k, _: k in (0, 'group')) == ['0123', '3'], \
 319             'function on a `re.Match` should give group name as well'
 320
 321     def test_traversal_xml_etree(self):
 322         etree = xml.etree.ElementTree.fromstring('''<?xml version="1.0"?>
 323         <data>
 324             <country name="Liechtenstein">
 325                 <rank>1</rank>
 326                 <year>2008</year>
 327                 <gdppc>141100</gdppc>
 328                 <neighbor name="Austria" direction="E"/>
 329                 <neighbor name="Switzerland" direction="W"/>
 330             </country>
 331             <country name="Singapore">
 332                 <rank>4</rank>
 333                 <year>2011</year>
 334                 <gdppc>59900</gdppc>
 335                 <neighbor name="Malaysia" direction="N"/>
 336             </country>
 337             <country name="Panama">
 338                 <rank>68</rank>
 339                 <year>2011</year>
 340                 <gdppc>13600</gdppc>
 341                 <neighbor name="Costa Rica" direction="W"/>
 342                 <neighbor name="Colombia" direction="E"/>
 343             </country>
 344         </data>''')
 345         assert traverse_obj(etree, '') == etree, \
 346             'empty str key should return the element itself'
 347         assert traverse_obj(etree, 'country') == list(etree), \
 348             'str key should lead all children with that tag name'
 349         assert traverse_obj(etree, ...) == list(etree), \
 350             '`...` as key should return all children'
 351         assert traverse_obj(etree, lambda _, x: x[0].text == '4') == [etree[1]], \
 352             'function as key should get element as value'
 353         assert traverse_obj(etree, lambda i, _: i == 1) == [etree[1]], \
 354             'function as key should get index as key'
 355         assert traverse_obj(etree, 0) == etree[0], \
 356             'int key should return the nth child'
 357         expected = ['Austria', 'Switzerland', 'Malaysia', 'Costa Rica', 'Colombia']
 358         assert traverse_obj(etree, './/neighbor/@name') == expected, \
 359             '`@<attribute>` at end of path should give that attribute'
 360         assert traverse_obj(etree, '//neighbor/@fail') == [None, None, None, None, None], \
 361             '`@<nonexistant>` at end of path should give `None`'
 362         assert traverse_obj(etree, ('//neighbor/@', 2)) == {'name': 'Malaysia', 'direction': 'N'}, \
 363             '`@` should give the full attribute dict'
 364         assert traverse_obj(etree, '//year/text()') == ['2008', '2011', '2011'], \
 365             '`text()` at end of path should give the inner text'
 366         assert traverse_obj(etree, '//*[@direction]/@direction') == ['E', 'W', 'N', 'W', 'E'], \
 367             'full Python xpath features should be supported'
 368         assert traverse_obj(etree, (0, '@name')) == 'Liechtenstein', \
 369             'special transformations should act on current element'
 370         assert traverse_obj(etree, ('country', 0, ..., 'text()', {int_or_none})) == [1, 2008, 141100], \
 371             'special transformations should act on current element'
 372
 373     def test_traversal_unbranching(self):
 374         assert traverse_obj(_TEST_DATA, [(100, 1.2), all]) == [100, 1.2], \
 375             '`all` should give all results as list'
 376         assert traverse_obj(_TEST_DATA, [(100, 1.2), any]) == 100, \
 377             '`any` should give the first result'
 378         assert traverse_obj(_TEST_DATA, [100, all]) == [100], \
 379             '`all` should give list if non branching'
 380         assert traverse_obj(_TEST_DATA, [100, any]) == 100, \
 381             '`any` should give single item if non branching'
 382         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100), all]) == [100], \
 383             '`all` should filter `None` and empty dict'
 384         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100), any]) == 100, \
 385             '`any` should filter `None` and empty dict'
 386         assert traverse_obj(_TEST_DATA, [{
 387             'all': [('dict', 'None', 100, 1.2), all],
 388             'any': [('dict', 'None', 100, 1.2), any],
 389         }]) == {'all': [100, 1.2], 'any': 100}, \
 390             '`all`/`any` should apply to each dict path separately'
 391         assert traverse_obj(_TEST_DATA, [{
 392             'all': [('dict', 'None', 100, 1.2), all],
 393             'any': [('dict', 'None', 100, 1.2), any],
 394         }], get_all=False) == {'all': [100, 1.2], 'any': 100}, \
 395             '`all`/`any` should apply to dict regardless of `get_all`'
 396         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), all, {float}]) is None, \
 397             '`all` should reset branching status'
 398         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), any, {float}]) is None, \
 399             '`any` should reset branching status'
 400         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), all, ..., {float}]) == [1.2], \
 401             '`all` should allow further branching'
 402         assert traverse_obj(_TEST_DATA, [('dict', 'None', 'urls', 'data'), any, ..., 'index']) == [0, 1], \
 403             '`any` should allow further branching'
 404
 405     def test_traversal_morsel(self):
 406         values = {
 407             'expires': 'a',
 408             'path': 'b',
 409             'comment': 'c',
 410             'domain': 'd',
 411             'max-age': 'e',
 412             'secure': 'f',
 413             'httponly': 'g',
 414             'version': 'h',
 415             'samesite': 'i',
 416         }
 417         morsel = http.cookies.Morsel()
 418         morsel.set('item_key', 'item_value', 'coded_value')
 419         morsel.update(values)
 420         values['key'] = 'item_key'
 421         values['value'] = 'item_value'
 422
 423         for key, value in values.items():
 424             assert traverse_obj(morsel, key) == value, \
 425                 'Morsel should provide access to all values'
 426         assert traverse_obj(morsel, ...) == list(values.values()), \
 427             '`...` should yield all values'
 428         assert traverse_obj(morsel, lambda k, v: True) == list(values.values()), \
 429             'function key should yield all values'
 430         assert traverse_obj(morsel, [(None,), any]) == morsel, \
 431             'Morsel should not be implicitly changed to dict on usage'
 432
 433     def test_traversal_filter(self):
 434         data = [None, False, True, 0, 1, 0.0, 1.1, '', 'str', {}, {0: 0}, [], [1]]
 435
 436         assert traverse_obj(data, [..., filter]) == [True, 1, 1.1, 'str', {0: 0}, [1]], \
 437             '`filter` should filter falsy values'
 438
 439
 440 class TestTraversalHelpers:
 441     def test_traversal_require(self):
 442         with pytest.raises(ExtractorError):
 443             traverse_obj(_TEST_DATA, ['None', {require('value')}])
 444         assert traverse_obj(_TEST_DATA, ['str', {require('value')}]) == 'str', \
 445             '`require` should pass through non `None` values'
 446
 447     def test_subs_list_to_dict(self):
 448         assert traverse_obj([
 449             {'name': 'de', 'url': 'https://example.com/subs/de.vtt'},
 450             {'name': 'en', 'url': 'https://example.com/subs/en1.ass'},
 451             {'name': 'en', 'url': 'https://example.com/subs/en2.ass'},
 452         ], [..., {
 453             'id': 'name',
 454             'url': 'url',
 455         }, all, {subs_list_to_dict}]) == {
 456             'de': [{'url': 'https://example.com/subs/de.vtt'}],
 457             'en': [
 458                 {'url': 'https://example.com/subs/en1.ass'},
 459                 {'url': 'https://example.com/subs/en2.ass'},
 460             ],
 461         }, 'function should build subtitle dict from list of subtitles'
 462         assert traverse_obj([
 463             {'name': 'de', 'url': 'https://example.com/subs/de.ass'},
 464             {'name': 'de'},
 465             {'name': 'en', 'content': 'content'},
 466             {'url': 'https://example.com/subs/en'},
 467         ], [..., {
 468             'id': 'name',
 469             'data': 'content',
 470             'url': 'url',
 471         }, all, {subs_list_to_dict}]) == {
 472             'de': [{'url': 'https://example.com/subs/de.ass'}],
 473             'en': [{'data': 'content'}],
 474         }, 'subs with mandatory items missing should be filtered'
 475         assert traverse_obj([
 476             {'url': 'https://example.com/subs/de.ass', 'name': 'de'},
 477             {'url': 'https://example.com/subs/en', 'name': 'en'},
 478         ], [..., {
 479             'id': 'name',
 480             'ext': ['url', {lambda x: determine_ext(x, default_ext=None)}],
 481             'url': 'url',
 482         }, all, {subs_list_to_dict(ext='ext')}]) == {
 483             'de': [{'url': 'https://example.com/subs/de.ass', 'ext': 'ass'}],
 484             'en': [{'url': 'https://example.com/subs/en', 'ext': 'ext'}],
 485         }, '`ext` should set default ext but leave existing value untouched'
 486         assert traverse_obj([
 487             {'name': 'en', 'url': 'https://example.com/subs/en2', 'prio': True},
 488             {'name': 'en', 'url': 'https://example.com/subs/en1', 'prio': False},
 489         ], [..., {
 490             'id': 'name',
 491             'quality': ['prio', {int}],
 492             'url': 'url',
 493         }, all, {subs_list_to_dict(ext='ext')}]) == {'en': [
 494             {'url': 'https://example.com/subs/en1', 'ext': 'ext'},
 495             {'url': 'https://example.com/subs/en2', 'ext': 'ext'},
 496         ]}, '`quality` key should sort subtitle list accordingly'
 497
 498
 499 class TestDictGet:
 500     def test_dict_get(self):
 501         FALSE_VALUES = {
 502             'none': None,
 503             'false': False,
 504             'zero': 0,
 505             'empty_string': '',
 506             'empty_list': [],
 507         }
 508         d = {**FALSE_VALUES, 'a': 42}
 509         assert dict_get(d, 'a') == 42
 510         assert dict_get(d, 'b') is None
 511         assert dict_get(d, 'b', 42) == 42
 512         assert dict_get(d, ('a',)) == 42
 513         assert dict_get(d, ('b', 'a')) == 42
 514         assert dict_get(d, ('b', 'c', 'a', 'd')) == 42
 515         assert dict_get(d, ('b', 'c')) is None
 516         assert dict_get(d, ('b', 'c'), 42) == 42
 517         for key, false_value in FALSE_VALUES.items():
 518             assert dict_get(d, ('b', 'c', key)) is None
 519             assert dict_get(d, ('b', 'c', key), skip_false_values=False) == false_value