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     join_nonempty,
  13     str_or_none,
  14 )
  15 from yt_dlp.utils.traversal import (
  16     find_element,
  17     find_elements,
  18     require,
  19     subs_list_to_dict,
  20     traverse_obj,
  21     trim_str,
  22     unpack,
  23 )
  24
  25 _TEST_DATA = {
  26     100: 100,
  27     1.2: 1.2,
  28     'str': 'str',
  29     'None': None,
  30     '...': ...,
  31     'urls': [
  32         {'index': 0, 'url': 'https://www.example.com/0'},
  33         {'index': 1, 'url': 'https://www.example.com/1'},
  34     ],
  35     'data': (
  36         {'index': 2},
  37         {'index': 3},
  38     ),
  39     'dict': {},
  40 }
  41
  42 _TEST_HTML = '''<html><body>
  43     <div class="a">1</div>
  44     <div class="a" id="x" custom="z">2</div>
  45     <div class="b" data-id="y" custom="z">3</div>
  46     <p class="a">4</p>
  47     <p id="d" custom="e">5</p>
  48 </body></html>'''
  49
  50
  51 class TestTraversal:
  52     def test_traversal_base(self):
  53         assert traverse_obj(_TEST_DATA, ('str',)) == 'str', \
  54             'allow tuple path'
  55         assert traverse_obj(_TEST_DATA, ['str']) == 'str', \
  56             'allow list path'
  57         assert traverse_obj(_TEST_DATA, (value for value in ('str',))) == 'str', \
  58             'allow iterable path'
  59         assert traverse_obj(_TEST_DATA, 'str') == 'str', \
  60             'single items should be treated as a path'
  61         assert traverse_obj(_TEST_DATA, 100) == 100, \
  62             'allow int path'
  63         assert traverse_obj(_TEST_DATA, 1.2) == 1.2, \
  64             'allow float path'
  65         assert traverse_obj(_TEST_DATA, None) == _TEST_DATA, \
  66             '`None` should not perform any modification'
  67
  68     def test_traversal_ellipsis(self):
  69         assert traverse_obj(_TEST_DATA, ...) == [x for x in _TEST_DATA.values() if x not in (None, {})], \
  70             '`...` should give all non discarded values'
  71         assert traverse_obj(_TEST_DATA, ('urls', 0, ...)) == list(_TEST_DATA['urls'][0].values()), \
  72             '`...` selection for dicts should select all values'
  73         assert traverse_obj(_TEST_DATA, (..., ..., 'url')) == ['https://www.example.com/0', 'https://www.example.com/1'], \
  74             'nested `...` queries should work'
  75         assert traverse_obj(_TEST_DATA, (..., ..., 'index')) == list(range(4)), \
  76             '`...` query result should be flattened'
  77         assert traverse_obj(iter(range(4)), ...) == list(range(4)), \
  78             '`...` should accept iterables'
  79
  80     def test_traversal_function(self):
  81         filter_func = lambda x, y: x == 'urls' and isinstance(y, list)
  82         assert traverse_obj(_TEST_DATA, filter_func) == [_TEST_DATA['urls']], \
  83             'function as query key should perform a filter based on (key, value)'
  84         assert traverse_obj(_TEST_DATA, lambda _, x: isinstance(x[0], str)) == ['str'], \
  85             'exceptions in the query function should be catched'
  86         assert traverse_obj(iter(range(4)), lambda _, x: x % 2 == 0) == [0, 2], \
  87             'function key should accept iterables'
  88         # Wrong function signature should raise (debug mode)
  89         with pytest.raises(Exception):
  90             traverse_obj(_TEST_DATA, lambda a: ...)
  91         with pytest.raises(Exception):
  92             traverse_obj(_TEST_DATA, lambda a, b, c: ...)
  93
  94     def test_traversal_set(self):
  95         # transformation/type, like `expected_type`
  96         assert traverse_obj(_TEST_DATA, (..., {str.upper})) == ['STR'], \
  97             'Function in set should be a transformation'
  98         assert traverse_obj(_TEST_DATA, (..., {str})) == ['str'], \
  99             'Type in set should be a type filter'
 100         assert traverse_obj(_TEST_DATA, (..., {str, int})) == [100, 'str'], \
 101             'Multiple types in set should be a type filter'
 102         assert traverse_obj(_TEST_DATA, {dict}) == _TEST_DATA, \
 103             'A single set should be wrapped into a path'
 104         assert traverse_obj(_TEST_DATA, (..., {str.upper})) == ['STR'], \
 105             'Transformation function should not raise'
 106         expected = [x for x in map(str_or_none, _TEST_DATA.values()) if x is not None]
 107         assert traverse_obj(_TEST_DATA, (..., {str_or_none})) == expected, \
 108             'Function in set should be a transformation'
 109         assert traverse_obj(_TEST_DATA, ('fail', {lambda _: 'const'})) == 'const', \
 110             'Function in set should always be called'
 111         # Sets with length < 1 or > 1 not including only types should raise
 112         with pytest.raises(Exception):
 113             traverse_obj(_TEST_DATA, set())
 114         with pytest.raises(Exception):
 115             traverse_obj(_TEST_DATA, {str.upper, str})
 116
 117     def test_traversal_slice(self):
 118         _SLICE_DATA = [0, 1, 2, 3, 4]
 119
 120         assert traverse_obj(_TEST_DATA, ('dict', slice(1))) is None, \
 121             'slice on a dictionary should not throw'
 122         assert traverse_obj(_SLICE_DATA, slice(1)) == _SLICE_DATA[:1], \
 123             'slice key should apply slice to sequence'
 124         assert traverse_obj(_SLICE_DATA, slice(1, 2)) == _SLICE_DATA[1:2], \
 125             'slice key should apply slice to sequence'
 126         assert traverse_obj(_SLICE_DATA, slice(1, 4, 2)) == _SLICE_DATA[1:4:2], \
 127             'slice key should apply slice to sequence'
 128
 129     def test_traversal_alternatives(self):
 130         assert traverse_obj(_TEST_DATA, 'fail', 'str') == 'str', \
 131             'multiple `paths` should be treated as alternative paths'
 132         assert traverse_obj(_TEST_DATA, 'str', 100) == 'str', \
 133             'alternatives should exit early'
 134         assert traverse_obj(_TEST_DATA, 'fail', 'fail') is None, \
 135             'alternatives should return `default` if exhausted'
 136         assert traverse_obj(_TEST_DATA, (..., 'fail'), 100) == 100, \
 137             'alternatives should track their own branching return'
 138         assert traverse_obj(_TEST_DATA, ('dict', ...), ('data', ...)) == list(_TEST_DATA['data']), \
 139             'alternatives on empty objects should search further'
 140
 141     def test_traversal_branching_nesting(self):
 142         assert traverse_obj(_TEST_DATA, ('urls', (3, 0), 'url')) == ['https://www.example.com/0'], \
 143             'tuple as key should be treated as branches'
 144         assert traverse_obj(_TEST_DATA, ('urls', [3, 0], 'url')) == ['https://www.example.com/0'], \
 145             'list as key should be treated as branches'
 146         assert traverse_obj(_TEST_DATA, ('urls', ((1, 'fail'), (0, 'url')))) == ['https://www.example.com/0'], \
 147             'double nesting in path should be treated as paths'
 148         assert traverse_obj(['0', [1, 2]], [(0, 1), 0]) == [1], \
 149             'do not fail early on branching'
 150         expected = ['https://www.example.com/0', 'https://www.example.com/1']
 151         assert traverse_obj(_TEST_DATA, ('urls', ((0, ('fail', 'url')), (1, 'url')))) == expected, \
 152             'tripple nesting in path should be treated as branches'
 153         assert traverse_obj(_TEST_DATA, ('urls', ('fail', (..., 'url')))) == expected, \
 154             'ellipsis as branch path start gets flattened'
 155
 156     def test_traversal_dict(self):
 157         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}) == {0: 100, 1: 1.2}, \
 158             'dict key should result in a dict with the same keys'
 159         expected = {0: 'https://www.example.com/0'}
 160         assert traverse_obj(_TEST_DATA, {0: ('urls', 0, 'url')}) == expected, \
 161             'dict key should allow paths'
 162         expected = {0: ['https://www.example.com/0']}
 163         assert traverse_obj(_TEST_DATA, {0: ('urls', (3, 0), 'url')}) == expected, \
 164             'tuple in dict path should be treated as branches'
 165         assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, 'fail'), (0, 'url')))}) == expected, \
 166             'double nesting in dict path should be treated as paths'
 167         expected = {0: ['https://www.example.com/1', 'https://www.example.com/0']}
 168         assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, ('fail', 'url')), (0, 'url')))}) == expected, \
 169             'tripple nesting in dict path should be treated as branches'
 170         assert traverse_obj(_TEST_DATA, {0: 'fail'}) == {}, \
 171             'remove `None` values when top level dict key fails'
 172         assert traverse_obj(_TEST_DATA, {0: 'fail'}, default=...) == {0: ...}, \
 173             'use `default` if key fails and `default`'
 174         assert traverse_obj(_TEST_DATA, {0: 'dict'}) == {}, \
 175             'remove empty values when dict key'
 176         assert traverse_obj(_TEST_DATA, {0: 'dict'}, default=...) == {0: ...}, \
 177             'use `default` when dict key and `default`'
 178         assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}) == {}, \
 179             'remove empty values when nested dict key fails'
 180         assert traverse_obj(None, {0: 'fail'}) == {}, \
 181             'default to dict if pruned'
 182         assert traverse_obj(None, {0: 'fail'}, default=...) == {0: ...}, \
 183             'default to dict if pruned and default is given'
 184         assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}, default=...) == {0: {0: ...}}, \
 185             'use nested `default` when nested dict key fails and `default`'
 186         assert traverse_obj(_TEST_DATA, {0: ('dict', ...)}) == {}, \
 187             'remove key if branch in dict key not successful'
 188
 189     def test_traversal_default(self):
 190         _DEFAULT_DATA = {'None': None, 'int': 0, 'list': []}
 191
 192         assert traverse_obj(_DEFAULT_DATA, 'fail') is None, \
 193             'default value should be `None`'
 194         assert traverse_obj(_DEFAULT_DATA, 'fail', 'fail', default=...) == ..., \
 195             'chained fails should result in default'
 196         assert traverse_obj(_DEFAULT_DATA, 'None', 'int') == 0, \
 197             'should not short cirquit on `None`'
 198         assert traverse_obj(_DEFAULT_DATA, 'fail', default=1) == 1, \
 199             'invalid dict key should result in `default`'
 200         assert traverse_obj(_DEFAULT_DATA, 'None', default=1) == 1, \
 201             '`None` is a deliberate sentinel and should become `default`'
 202         assert traverse_obj(_DEFAULT_DATA, ('list', 10)) is None, \
 203             '`IndexError` should result in `default`'
 204         assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=1) == 1, \
 205             'if branched but not successful return `default` if defined, not `[]`'
 206         assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=None) is None, \
 207             'if branched but not successful return `default` even if `default` is `None`'
 208         assert traverse_obj(_DEFAULT_DATA, (..., 'fail')) == [], \
 209             'if branched but not successful return `[]`, not `default`'
 210         assert traverse_obj(_DEFAULT_DATA, ('list', ...)) == [], \
 211             'if branched but object is empty return `[]`, not `default`'
 212         assert traverse_obj(None, ...) == [], \
 213             'if branched but object is `None` return `[]`, not `default`'
 214         assert traverse_obj({0: None}, (0, ...)) == [], \
 215             'if branched but state is `None` return `[]`, not `default`'
 216
 217     @pytest.mark.parametrize('path', [
 218         ('fail', ...),
 219         (..., 'fail'),
 220         100 * ('fail',) + (...,),
 221         (...,) + 100 * ('fail',),
 222     ])
 223     def test_traversal_branching(self, path):
 224         assert traverse_obj({}, path) == [], \
 225             'if branched but state is `None`, return `[]` (not `default`)'
 226         assert traverse_obj({}, 'fail', path) == [], \
 227             'if branching in last alternative and previous did not match, return `[]` (not `default`)'
 228         assert traverse_obj({0: 'x'}, 0, path) == 'x', \
 229             'if branching in last alternative and previous did match, return single value'
 230         assert traverse_obj({0: 'x'}, path, 0) == 'x', \
 231             'if branching in first alternative and non-branching path does match, return single value'
 232         assert traverse_obj({}, path, 'fail') is None, \
 233             'if branching in first alternative and non-branching path does not match, return `default`'
 234
 235     def test_traversal_expected_type(self):
 236         _EXPECTED_TYPE_DATA = {'str': 'str', 'int': 0}
 237
 238         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=str) == 'str', \
 239             'accept matching `expected_type` type'
 240         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=int) is None, \
 241             'reject non matching `expected_type` type'
 242         assert traverse_obj(_EXPECTED_TYPE_DATA, 'int', expected_type=lambda x: str(x)) == '0', \
 243             'transform type using type function'
 244         assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=lambda _: 1 / 0) is None, \
 245             'wrap expected_type fuction in try_call'
 246         assert traverse_obj(_EXPECTED_TYPE_DATA, ..., expected_type=str) == ['str'], \
 247             'eliminate items that expected_type fails on'
 248         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}, expected_type=int) == {0: 100}, \
 249             'type as expected_type should filter dict values'
 250         assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2, 2: 'None'}, expected_type=str_or_none) == {0: '100', 1: '1.2'}, \
 251             'function as expected_type should transform dict values'
 252         assert traverse_obj(_TEST_DATA, ({0: 1.2}, 0, {int_or_none}), expected_type=int) == 1, \
 253             'expected_type should not filter non final dict values'
 254         assert traverse_obj(_TEST_DATA, {0: {0: 100, 1: 'str'}}, expected_type=int) == {0: {0: 100}}, \
 255             'expected_type should transform deep dict values'
 256         assert traverse_obj(_TEST_DATA, [({0: '...'}, {0: '...'})], expected_type=type(...)) == [{0: ...}, {0: ...}], \
 257             'expected_type should transform branched dict values'
 258         assert traverse_obj({1: {3: 4}}, [(1, 2), 3], expected_type=int) == [4], \
 259             'expected_type regression for type matching in tuple branching'
 260         assert traverse_obj(_TEST_DATA, ['data', ...], expected_type=int) == [], \
 261             'expected_type regression for type matching in dict result'
 262
 263     def test_traversal_get_all(self):
 264         _GET_ALL_DATA = {'key': [0, 1, 2]}
 265
 266         assert traverse_obj(_GET_ALL_DATA, ('key', ...), get_all=False) == 0, \
 267             'if not `get_all`, return only first matching value'
 268         assert traverse_obj(_GET_ALL_DATA, ..., get_all=False) == [0, 1, 2], \
 269             'do not overflatten if not `get_all`'
 270
 271     def test_traversal_casesense(self):
 272         _CASESENSE_DATA = {
 273             'KeY': 'value0',
 274             0: {
 275                 'KeY': 'value1',
 276                 0: {'KeY': 'value2'},
 277             },
 278         }
 279
 280         assert traverse_obj(_CASESENSE_DATA, 'key') is None, \
 281             'dict keys should be case sensitive unless `casesense`'
 282         assert traverse_obj(_CASESENSE_DATA, 'keY', casesense=False) == 'value0', \
 283             'allow non matching key case if `casesense`'
 284         assert traverse_obj(_CASESENSE_DATA, [0, ('keY',)], casesense=False) == ['value1'], \
 285             'allow non matching key case in branch if `casesense`'
 286         assert traverse_obj(_CASESENSE_DATA, [0, ([0, 'keY'],)], casesense=False) == ['value2'], \
 287             'allow non matching key case in branch path if `casesense`'
 288
 289     def test_traversal_traverse_string(self):
 290         _TRAVERSE_STRING_DATA = {'str': 'str', 1.2: 1.2}
 291
 292         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0)) is None, \
 293             'do not traverse into string if not `traverse_string`'
 294         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0), traverse_string=True) == 's', \
 295             'traverse into string if `traverse_string`'
 296         assert traverse_obj(_TRAVERSE_STRING_DATA, (1.2, 1), traverse_string=True) == '.', \
 297             'traverse into converted data if `traverse_string`'
 298         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', ...), traverse_string=True) == 'str', \
 299             '`...` should result in string (same value) if `traverse_string`'
 300         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', slice(0, None, 2)), traverse_string=True) == 'sr', \
 301             '`slice` should result in string if `traverse_string`'
 302         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', lambda i, v: i or v == 's'), traverse_string=True) == 'str', \
 303             'function should result in string if `traverse_string`'
 304         assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', (0, 2)), traverse_string=True) == ['s', 'r'], \
 305             'branching should result in list if `traverse_string`'
 306         assert traverse_obj({}, (0, ...), traverse_string=True) == [], \
 307             'branching should result in list if `traverse_string`'
 308         assert traverse_obj({}, (0, lambda x, y: True), traverse_string=True) == [], \
 309             'branching should result in list if `traverse_string`'
 310         assert traverse_obj({}, (0, slice(1)), traverse_string=True) == [], \
 311             'branching should result in list if `traverse_string`'
 312
 313     def test_traversal_re(self):
 314         mobj = re.fullmatch(r'0(12)(?P<group>3)(4)?', '0123')
 315         assert traverse_obj(mobj, ...) == [x for x in mobj.groups() if x is not None], \
 316             '`...` on a `re.Match` should give its `groups()`'
 317         assert traverse_obj(mobj, lambda k, _: k in (0, 2)) == ['0123', '3'], \
 318             'function on a `re.Match` should give groupno, value starting at 0'
 319         assert traverse_obj(mobj, 'group') == '3', \
 320             'str key on a `re.Match` should give group with that name'
 321         assert traverse_obj(mobj, 2) == '3', \
 322             'int key on a `re.Match` should give group with that name'
 323         assert traverse_obj(mobj, 'gRoUp', casesense=False) == '3', \
 324             'str key on a `re.Match` should respect casesense'
 325         assert traverse_obj(mobj, 'fail') is None, \
 326             'failing str key on a `re.Match` should return `default`'
 327         assert traverse_obj(mobj, 'gRoUpS', casesense=False) is None, \
 328             'failing str key on a `re.Match` should return `default`'
 329         assert traverse_obj(mobj, 8) is None, \
 330             'failing int key on a `re.Match` should return `default`'
 331         assert traverse_obj(mobj, lambda k, _: k in (0, 'group')) == ['0123', '3'], \
 332             'function on a `re.Match` should give group name as well'
 333
 334     def test_traversal_xml_etree(self):
 335         etree = xml.etree.ElementTree.fromstring('''<?xml version="1.0"?>
 336         <data>
 337             <country name="Liechtenstein">
 338                 <rank>1</rank>
 339                 <year>2008</year>
 340                 <gdppc>141100</gdppc>
 341                 <neighbor name="Austria" direction="E"/>
 342                 <neighbor name="Switzerland" direction="W"/>
 343             </country>
 344             <country name="Singapore">
 345                 <rank>4</rank>
 346                 <year>2011</year>
 347                 <gdppc>59900</gdppc>
 348                 <neighbor name="Malaysia" direction="N"/>
 349             </country>
 350             <country name="Panama">
 351                 <rank>68</rank>
 352                 <year>2011</year>
 353                 <gdppc>13600</gdppc>
 354                 <neighbor name="Costa Rica" direction="W"/>
 355                 <neighbor name="Colombia" direction="E"/>
 356             </country>
 357         </data>''')
 358         assert traverse_obj(etree, '') == etree, \
 359             'empty str key should return the element itself'
 360         assert traverse_obj(etree, 'country') == list(etree), \
 361             'str key should lead all children with that tag name'
 362         assert traverse_obj(etree, ...) == list(etree), \
 363             '`...` as key should return all children'
 364         assert traverse_obj(etree, lambda _, x: x[0].text == '4') == [etree[1]], \
 365             'function as key should get element as value'
 366         assert traverse_obj(etree, lambda i, _: i == 1) == [etree[1]], \
 367             'function as key should get index as key'
 368         assert traverse_obj(etree, 0) == etree[0], \
 369             'int key should return the nth child'
 370         expected = ['Austria', 'Switzerland', 'Malaysia', 'Costa Rica', 'Colombia']
 371         assert traverse_obj(etree, './/neighbor/@name') == expected, \
 372             '`@<attribute>` at end of path should give that attribute'
 373         assert traverse_obj(etree, '//neighbor/@fail') == [None, None, None, None, None], \
 374             '`@<nonexistant>` at end of path should give `None`'
 375         assert traverse_obj(etree, ('//neighbor/@', 2)) == {'name': 'Malaysia', 'direction': 'N'}, \
 376             '`@` should give the full attribute dict'
 377         assert traverse_obj(etree, '//year/text()') == ['2008', '2011', '2011'], \
 378             '`text()` at end of path should give the inner text'
 379         assert traverse_obj(etree, '//*[@direction]/@direction') == ['E', 'W', 'N', 'W', 'E'], \
 380             'full Python xpath features should be supported'
 381         assert traverse_obj(etree, (0, '@name')) == 'Liechtenstein', \
 382             'special transformations should act on current element'
 383         assert traverse_obj(etree, ('country', 0, ..., 'text()', {int_or_none})) == [1, 2008, 141100], \
 384             'special transformations should act on current element'
 385
 386     def test_traversal_unbranching(self):
 387         assert traverse_obj(_TEST_DATA, [(100, 1.2), all]) == [100, 1.2], \
 388             '`all` should give all results as list'
 389         assert traverse_obj(_TEST_DATA, [(100, 1.2), any]) == 100, \
 390             '`any` should give the first result'
 391         assert traverse_obj(_TEST_DATA, [100, all]) == [100], \
 392             '`all` should give list if non branching'
 393         assert traverse_obj(_TEST_DATA, [100, any]) == 100, \
 394             '`any` should give single item if non branching'
 395         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100), all]) == [100], \
 396             '`all` should filter `None` and empty dict'
 397         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100), any]) == 100, \
 398             '`any` should filter `None` and empty dict'
 399         assert traverse_obj(_TEST_DATA, [{
 400             'all': [('dict', 'None', 100, 1.2), all],
 401             'any': [('dict', 'None', 100, 1.2), any],
 402         }]) == {'all': [100, 1.2], 'any': 100}, \
 403             '`all`/`any` should apply to each dict path separately'
 404         assert traverse_obj(_TEST_DATA, [{
 405             'all': [('dict', 'None', 100, 1.2), all],
 406             'any': [('dict', 'None', 100, 1.2), any],
 407         }], get_all=False) == {'all': [100, 1.2], 'any': 100}, \
 408             '`all`/`any` should apply to dict regardless of `get_all`'
 409         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), all, {float}]) is None, \
 410             '`all` should reset branching status'
 411         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), any, {float}]) is None, \
 412             '`any` should reset branching status'
 413         assert traverse_obj(_TEST_DATA, [('dict', 'None', 100, 1.2), all, ..., {float}]) == [1.2], \
 414             '`all` should allow further branching'
 415         assert traverse_obj(_TEST_DATA, [('dict', 'None', 'urls', 'data'), any, ..., 'index']) == [0, 1], \
 416             '`any` should allow further branching'
 417
 418     def test_traversal_morsel(self):
 419         values = {
 420             'expires': 'a',
 421             'path': 'b',
 422             'comment': 'c',
 423             'domain': 'd',
 424             'max-age': 'e',
 425             'secure': 'f',
 426             'httponly': 'g',
 427             'version': 'h',
 428             'samesite': 'i',
 429         }
 430         morsel = http.cookies.Morsel()
 431         morsel.set('item_key', 'item_value', 'coded_value')
 432         morsel.update(values)
 433         values['key'] = 'item_key'
 434         values['value'] = 'item_value'
 435
 436         for key, value in values.items():
 437             assert traverse_obj(morsel, key) == value, \
 438                 'Morsel should provide access to all values'
 439         assert traverse_obj(morsel, ...) == list(values.values()), \
 440             '`...` should yield all values'
 441         assert traverse_obj(morsel, lambda k, v: True) == list(values.values()), \
 442             'function key should yield all values'
 443         assert traverse_obj(morsel, [(None,), any]) == morsel, \
 444             'Morsel should not be implicitly changed to dict on usage'
 445
 446     def test_traversal_filter(self):
 447         data = [None, False, True, 0, 1, 0.0, 1.1, '', 'str', {}, {0: 0}, [], [1]]
 448
 449         assert traverse_obj(data, [..., filter]) == [True, 1, 1.1, 'str', {0: 0}, [1]], \
 450             '`filter` should filter falsy values'
 451
 452
 453 class TestTraversalHelpers:
 454     def test_traversal_require(self):
 455         with pytest.raises(ExtractorError):
 456             traverse_obj(_TEST_DATA, ['None', {require('value')}])
 457         assert traverse_obj(_TEST_DATA, ['str', {require('value')}]) == 'str', \
 458             '`require` should pass through non `None` values'
 459
 460     def test_subs_list_to_dict(self):
 461         assert traverse_obj([
 462             {'name': 'de', 'url': 'https://example.com/subs/de.vtt'},
 463             {'name': 'en', 'url': 'https://example.com/subs/en1.ass'},
 464             {'name': 'en', 'url': 'https://example.com/subs/en2.ass'},
 465         ], [..., {
 466             'id': 'name',
 467             'url': 'url',
 468         }, all, {subs_list_to_dict}]) == {
 469             'de': [{'url': 'https://example.com/subs/de.vtt'}],
 470             'en': [
 471                 {'url': 'https://example.com/subs/en1.ass'},
 472                 {'url': 'https://example.com/subs/en2.ass'},
 473             ],
 474         }, 'function should build subtitle dict from list of subtitles'
 475         assert traverse_obj([
 476             {'name': 'de', 'url': 'https://example.com/subs/de.ass'},
 477             {'name': 'de'},
 478             {'name': 'en', 'content': 'content'},
 479             {'url': 'https://example.com/subs/en'},
 480         ], [..., {
 481             'id': 'name',
 482             'data': 'content',
 483             'url': 'url',
 484         }, all, {subs_list_to_dict(lang=None)}]) == {
 485             'de': [{'url': 'https://example.com/subs/de.ass'}],
 486             'en': [{'data': 'content'}],
 487         }, 'subs with mandatory items missing should be filtered'
 488         assert traverse_obj([
 489             {'url': 'https://example.com/subs/de.ass', 'name': 'de'},
 490             {'url': 'https://example.com/subs/en', 'name': 'en'},
 491         ], [..., {
 492             'id': 'name',
 493             'ext': ['url', {determine_ext(default_ext=None)}],
 494             'url': 'url',
 495         }, all, {subs_list_to_dict(ext='ext')}]) == {
 496             'de': [{'url': 'https://example.com/subs/de.ass', 'ext': 'ass'}],
 497             'en': [{'url': 'https://example.com/subs/en', 'ext': 'ext'}],
 498         }, '`ext` should set default ext but leave existing value untouched'
 499         assert traverse_obj([
 500             {'name': 'en', 'url': 'https://example.com/subs/en2', 'prio': True},
 501             {'name': 'en', 'url': 'https://example.com/subs/en1', 'prio': False},
 502         ], [..., {
 503             'id': 'name',
 504             'quality': ['prio', {int}],
 505             'url': 'url',
 506         }, all, {subs_list_to_dict(ext='ext')}]) == {'en': [
 507             {'url': 'https://example.com/subs/en1', 'ext': 'ext'},
 508             {'url': 'https://example.com/subs/en2', 'ext': 'ext'},
 509         ]}, '`quality` key should sort subtitle list accordingly'
 510         assert traverse_obj([
 511             {'name': 'de', 'url': 'https://example.com/subs/de.ass'},
 512             {'name': 'de'},
 513             {'name': 'en', 'content': 'content'},
 514             {'url': 'https://example.com/subs/en'},
 515         ], [..., {
 516             'id': 'name',
 517             'url': 'url',
 518             'data': 'content',
 519         }, all, {subs_list_to_dict(lang='en')}]) == {
 520             'de': [{'url': 'https://example.com/subs/de.ass'}],
 521             'en': [
 522                 {'data': 'content'},
 523                 {'url': 'https://example.com/subs/en'},
 524             ],
 525         }, 'optionally provided lang should be used if no id available'
 526         assert traverse_obj([
 527             {'name': 1, 'url': 'https://example.com/subs/de1'},
 528             {'name': {}, 'url': 'https://example.com/subs/de2'},
 529             {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
 530             {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
 531         ], [..., {
 532             'id': 'name',
 533             'url': 'url',
 534             'ext': 'ext',
 535         }, all, {subs_list_to_dict(lang=None)}]) == {
 536             'de': [
 537                 {'url': 'https://example.com/subs/de3'},
 538                 {'url': 'https://example.com/subs/de4'},
 539             ],
 540         }, 'non str types should be ignored for id and ext'
 541         assert traverse_obj([
 542             {'name': 1, 'url': 'https://example.com/subs/de1'},
 543             {'name': {}, 'url': 'https://example.com/subs/de2'},
 544             {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
 545             {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
 546         ], [..., {
 547             'id': 'name',
 548             'url': 'url',
 549             'ext': 'ext',
 550         }, all, {subs_list_to_dict(lang='de')}]) == {
 551             'de': [
 552                 {'url': 'https://example.com/subs/de1'},
 553                 {'url': 'https://example.com/subs/de2'},
 554                 {'url': 'https://example.com/subs/de3'},
 555                 {'url': 'https://example.com/subs/de4'},
 556             ],
 557         }, 'non str types should be replaced by default id'
 558
 559     def test_trim_str(self):
 560         with pytest.raises(TypeError):
 561             trim_str('positional')
 562
 563         assert callable(trim_str(start='a'))
 564         assert trim_str(start='ab')('abc') == 'c'
 565         assert trim_str(end='bc')('abc') == 'a'
 566         assert trim_str(start='a', end='c')('abc') == 'b'
 567         assert trim_str(start='ab', end='c')('abc') == ''
 568         assert trim_str(start='a', end='bc')('abc') == ''
 569         assert trim_str(start='ab', end='bc')('abc') == ''
 570         assert trim_str(start='abc', end='abc')('abc') == ''
 571         assert trim_str(start='', end='')('abc') == 'abc'
 572
 573     def test_unpack(self):
 574         assert unpack(lambda *x: ''.join(map(str, x)))([1, 2, 3]) == '123'
 575         assert unpack(join_nonempty)([1, 2, 3]) == '1-2-3'
 576         assert unpack(join_nonempty, delim=' ')([1, 2, 3]) == '1 2 3'
 577         with pytest.raises(TypeError):
 578             unpack(join_nonempty)()
 579         with pytest.raises(TypeError):
 580             unpack()
 581
 582     def test_find_element(self):
 583         for improper_kwargs in [
 584             dict(attr='data-id'),
 585             dict(value='y'),
 586             dict(attr='data-id', value='y', cls='a'),
 587             dict(attr='data-id', value='y', id='x'),
 588             dict(cls='a', id='x'),
 589             dict(cls='a', tag='p'),
 590             dict(cls='[ab]', regex=True),
 591         ]:
 592             with pytest.raises(AssertionError):
 593                 find_element(**improper_kwargs)(_TEST_HTML)
 594
 595         assert find_element(cls='a')(_TEST_HTML) == '1'
 596         assert find_element(cls='a', html=True)(_TEST_HTML) == '<div class="a">1</div>'
 597         assert find_element(id='x')(_TEST_HTML) == '2'
 598         assert find_element(id='[ex]')(_TEST_HTML) is None
 599         assert find_element(id='[ex]', regex=True)(_TEST_HTML) == '2'
 600         assert find_element(id='x', html=True)(_TEST_HTML) == '<div class="a" id="x" custom="z">2</div>'
 601         assert find_element(attr='data-id', value='y')(_TEST_HTML) == '3'
 602         assert find_element(attr='data-id', value='y(?:es)?')(_TEST_HTML) is None
 603         assert find_element(attr='data-id', value='y(?:es)?', regex=True)(_TEST_HTML) == '3'
 604         assert find_element(
 605             attr='data-id', value='y', html=True)(_TEST_HTML) == '<div class="b" data-id="y" custom="z">3</div>'
 606
 607     def test_find_elements(self):
 608         for improper_kwargs in [
 609             dict(tag='p'),
 610             dict(attr='data-id'),
 611             dict(value='y'),
 612             dict(attr='data-id', value='y', cls='a'),
 613             dict(cls='a', tag='div'),
 614             dict(cls='[ab]', regex=True),
 615         ]:
 616             with pytest.raises(AssertionError):
 617                 find_elements(**improper_kwargs)(_TEST_HTML)
 618
 619         assert find_elements(cls='a')(_TEST_HTML) == ['1', '2', '4']
 620         assert find_elements(cls='a', html=True)(_TEST_HTML) == [
 621             '<div class="a">1</div>', '<div class="a" id="x" custom="z">2</div>', '<p class="a">4</p>']
 622         assert find_elements(attr='custom', value='z')(_TEST_HTML) == ['2', '3']
 623         assert find_elements(attr='custom', value='[ez]')(_TEST_HTML) == []
 624         assert find_elements(attr='custom', value='[ez]', regex=True)(_TEST_HTML) == ['2', '3', '5']
 625
 626
 627 class TestDictGet:
 628     def test_dict_get(self):
 629         FALSE_VALUES = {
 630             'none': None,
 631             'false': False,
 632             'zero': 0,
 633             'empty_string': '',
 634             'empty_list': [],
 635         }
 636         d = {**FALSE_VALUES, 'a': 42}
 637         assert dict_get(d, 'a') == 42
 638         assert dict_get(d, 'b') is None
 639         assert dict_get(d, 'b', 42) == 42
 640         assert dict_get(d, ('a',)) == 42
 641         assert dict_get(d, ('b', 'a')) == 42
 642         assert dict_get(d, ('b', 'c', 'a', 'd')) == 42
 643         assert dict_get(d, ('b', 'c')) is None
 644         assert dict_get(d, ('b', 'c'), 42) == 42
 645         for key, false_value in FALSE_VALUES.items():
 646             assert dict_get(d, ('b', 'c', key)) is None
 647             assert dict_get(d, ('b', 'c', key), skip_false_values=False) == false_value