Mirrors/yt-dlp
2
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-12-04 18:39:29 +00:00
Code Issues Projects Releases Packages Wiki Activity

[test] traversal: Separate traversal tests (#9574)

Browse source 
Authored by: Grub4K
This commit is contained in:
Simon Sawicki 2024-03-30 19:32:07 +01:00 committed by GitHub
parent 58dd0f8d1e
commit 979ce2e786
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 379 additions and 379 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

379
test/test_traversal.py Normal file
View file

@ -0,0 +1,379 @@
import re
import xml.etree.ElementTree
import pytest
from yt_dlp.utils import dict_get, int_or_none, str_or_none
from yt_dlp.utils.traversal import traverse_obj
_TEST_DATA = {
100: 100,
1.2: 1.2,
'str': 'str',
'None': None,
'...': ...,
'urls': [
{'index': 0, 'url': 'https://www.example.com/0'},
{'index': 1, 'url': 'https://www.example.com/1'},
],
'data': (
{'index': 2},
{'index': 3},
),
'dict': {},
}
class TestTraversal:
def test_dict_get(self):
FALSE_VALUES = {
'none': None,
'false': False,
'zero': 0,
'empty_string': '',
'empty_list': [],
}
d = {**FALSE_VALUES, 'a': 42}
assert dict_get(d, 'a') == 42
assert dict_get(d, 'b') is None
assert dict_get(d, 'b', 42) == 42
assert dict_get(d, ('a',)) == 42
assert dict_get(d, ('b', 'a')) == 42
assert dict_get(d, ('b', 'c', 'a', 'd')) == 42
assert dict_get(d, ('b', 'c')) is None
assert dict_get(d, ('b', 'c'), 42) == 42
for key, false_value in FALSE_VALUES.items():
assert dict_get(d, ('b', 'c', key)) is None
assert dict_get(d, ('b', 'c', key), skip_false_values=False) == false_value
def test_traversal_base(self):
assert traverse_obj(_TEST_DATA, ('str',)) == 'str', \
'allow tuple path'
assert traverse_obj(_TEST_DATA, ['str']) == 'str', \
'allow list path'
assert traverse_obj(_TEST_DATA, (value for value in ("str",))) == 'str', \
'allow iterable path'
assert traverse_obj(_TEST_DATA, 'str') == 'str', \
'single items should be treated as a path'
assert traverse_obj(_TEST_DATA, 100) == 100, \
'allow int path'
assert traverse_obj(_TEST_DATA, 1.2) == 1.2, \
'allow float path'
assert traverse_obj(_TEST_DATA, None) == _TEST_DATA, \
'`None` should not perform any modification'
def test_traversal_ellipsis(self):
assert traverse_obj(_TEST_DATA, ...) == [x for x in _TEST_DATA.values() if x not in (None, {})], \
'`...` should give all non discarded values'
assert traverse_obj(_TEST_DATA, ('urls', 0, ...)) == list(_TEST_DATA['urls'][0].values()), \
'`...` selection for dicts should select all values'
assert traverse_obj(_TEST_DATA, (..., ..., 'url')) == ['https://www.example.com/0', 'https://www.example.com/1'], \
'nested `...` queries should work'
assert traverse_obj(_TEST_DATA, (..., ..., 'index')) == list(range(4)), \
'`...` query result should be flattened'
assert traverse_obj(iter(range(4)), ...) == list(range(4)), \
'`...` should accept iterables'
def test_traversal_function(self):
filter_func = lambda x, y: x == 'urls' and isinstance(y, list)
assert traverse_obj(_TEST_DATA, filter_func) == [_TEST_DATA['urls']], \
'function as query key should perform a filter based on (key, value)'
assert traverse_obj(_TEST_DATA, lambda _, x: isinstance(x[0], str)) == ['str'], \
'exceptions in the query function should be catched'
assert traverse_obj(iter(range(4)), lambda _, x: x % 2 == 0) == [0, 2], \
'function key should accept iterables'
# Wrong function signature should raise (debug mode)
with pytest.raises(Exception):
traverse_obj(_TEST_DATA, lambda a: ...)
with pytest.raises(Exception):
traverse_obj(_TEST_DATA, lambda a, b, c: ...)
def test_traversal_set(self):
# transformation/type, like `expected_type`
assert traverse_obj(_TEST_DATA, (..., {str.upper}, )) == ['STR'], \
'Function in set should be a transformation'
assert traverse_obj(_TEST_DATA, (..., {str})) == ['str'], \
'Type in set should be a type filter'
assert traverse_obj(_TEST_DATA, {dict}) == _TEST_DATA, \
'A single set should be wrapped into a path'
assert traverse_obj(_TEST_DATA, (..., {str.upper})) == ['STR'], \
'Transformation function should not raise'
expected = [x for x in map(str_or_none, _TEST_DATA.values()) if x is not None]
assert traverse_obj(_TEST_DATA, (..., {str_or_none})) == expected, \
'Function in set should be a transformation'
assert traverse_obj(_TEST_DATA, ('fail', {lambda _: 'const'})) == 'const', \
'Function in set should always be called'
# Sets with length != 1 should raise in debug
with pytest.raises(Exception):
traverse_obj(_TEST_DATA, set())
with pytest.raises(Exception):
traverse_obj(_TEST_DATA, {str.upper, str})
def test_traversal_slice(self):
_SLICE_DATA = [0, 1, 2, 3, 4]
assert traverse_obj(_TEST_DATA, ('dict', slice(1))) is None, \
'slice on a dictionary should not throw'
assert traverse_obj(_SLICE_DATA, slice(1)) == _SLICE_DATA[:1], \
'slice key should apply slice to sequence'
assert traverse_obj(_SLICE_DATA, slice(1, 2)) == _SLICE_DATA[1:2], \
'slice key should apply slice to sequence'
assert traverse_obj(_SLICE_DATA, slice(1, 4, 2)) == _SLICE_DATA[1:4:2], \
'slice key should apply slice to sequence'
def test_traversal_alternatives(self):
assert traverse_obj(_TEST_DATA, 'fail', 'str') == 'str', \
'multiple `paths` should be treated as alternative paths'
assert traverse_obj(_TEST_DATA, 'str', 100) == 'str', \
'alternatives should exit early'
assert traverse_obj(_TEST_DATA, 'fail', 'fail') is None, \
'alternatives should return `default` if exhausted'
assert traverse_obj(_TEST_DATA, (..., 'fail'), 100) == 100, \
'alternatives should track their own branching return'
assert traverse_obj(_TEST_DATA, ('dict', ...), ('data', ...)) == list(_TEST_DATA['data']), \
'alternatives on empty objects should search further'
def test_traversal_branching_nesting(self):
assert traverse_obj(_TEST_DATA, ('urls', (3, 0), 'url')) == ['https://www.example.com/0'], \
'tuple as key should be treated as branches'
assert traverse_obj(_TEST_DATA, ('urls', [3, 0], 'url')) == ['https://www.example.com/0'], \
'list as key should be treated as branches'
assert traverse_obj(_TEST_DATA, ('urls', ((1, 'fail'), (0, 'url')))) == ['https://www.example.com/0'], \
'double nesting in path should be treated as paths'
assert traverse_obj(['0', [1, 2]], [(0, 1), 0]) == [1], \
'do not fail early on branching'
expected = ['https://www.example.com/0', 'https://www.example.com/1']
assert traverse_obj(_TEST_DATA, ('urls', ((0, ('fail', 'url')), (1, 'url')))) == expected, \
'tripple nesting in path should be treated as branches'
assert traverse_obj(_TEST_DATA, ('urls', ('fail', (..., 'url')))) == expected, \
'ellipsis as branch path start gets flattened'
def test_traversal_dict(self):
assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}) == {0: 100, 1: 1.2}, \
'dict key should result in a dict with the same keys'
expected = {0: 'https://www.example.com/0'}
assert traverse_obj(_TEST_DATA, {0: ('urls', 0, 'url')}) == expected, \
'dict key should allow paths'
expected = {0: ['https://www.example.com/0']}
assert traverse_obj(_TEST_DATA, {0: ('urls', (3, 0), 'url')}) == expected, \
'tuple in dict path should be treated as branches'
assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, 'fail'), (0, 'url')))}) == expected, \
'double nesting in dict path should be treated as paths'
expected = {0: ['https://www.example.com/1', 'https://www.example.com/0']}
assert traverse_obj(_TEST_DATA, {0: ('urls', ((1, ('fail', 'url')), (0, 'url')))}) == expected, \
'tripple nesting in dict path should be treated as branches'
assert traverse_obj(_TEST_DATA, {0: 'fail'}) == {}, \
'remove `None` values when top level dict key fails'
assert traverse_obj(_TEST_DATA, {0: 'fail'}, default=...) == {0: ...}, \
'use `default` if key fails and `default`'
assert traverse_obj(_TEST_DATA, {0: 'dict'}) == {}, \
'remove empty values when dict key'
assert traverse_obj(_TEST_DATA, {0: 'dict'}, default=...) == {0: ...}, \
'use `default` when dict key and `default`'
assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}) == {}, \
'remove empty values when nested dict key fails'
assert traverse_obj(None, {0: 'fail'}) == {}, \
'default to dict if pruned'
assert traverse_obj(None, {0: 'fail'}, default=...) == {0: ...}, \
'default to dict if pruned and default is given'
assert traverse_obj(_TEST_DATA, {0: {0: 'fail'}}, default=...) == {0: {0: ...}}, \
'use nested `default` when nested dict key fails and `default`'
assert traverse_obj(_TEST_DATA, {0: ('dict', ...)}) == {}, \
'remove key if branch in dict key not successful'
def test_traversal_default(self):
_DEFAULT_DATA = {'None': None, 'int': 0, 'list': []}
assert traverse_obj(_DEFAULT_DATA, 'fail') is None, \
'default value should be `None`'
assert traverse_obj(_DEFAULT_DATA, 'fail', 'fail', default=...) == ..., \
'chained fails should result in default'
assert traverse_obj(_DEFAULT_DATA, 'None', 'int') == 0, \
'should not short cirquit on `None`'
assert traverse_obj(_DEFAULT_DATA, 'fail', default=1) == 1, \
'invalid dict key should result in `default`'
assert traverse_obj(_DEFAULT_DATA, 'None', default=1) == 1, \
'`None` is a deliberate sentinel and should become `default`'
assert traverse_obj(_DEFAULT_DATA, ('list', 10)) is None, \
'`IndexError` should result in `default`'
assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=1) == 1, \
'if branched but not successful return `default` if defined, not `[]`'
assert traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=None) is None, \
'if branched but not successful return `default` even if `default` is `None`'
assert traverse_obj(_DEFAULT_DATA, (..., 'fail')) == [], \
'if branched but not successful return `[]`, not `default`'
assert traverse_obj(_DEFAULT_DATA, ('list', ...)) == [], \
'if branched but object is empty return `[]`, not `default`'
assert traverse_obj(None, ...) == [], \
'if branched but object is `None` return `[]`, not `default`'
assert traverse_obj({0: None}, (0, ...)) == [], \
'if branched but state is `None` return `[]`, not `default`'
@pytest.mark.parametrize('path', [
('fail', ...),
(..., 'fail'),
100 * ('fail',) + (...,),
(...,) + 100 * ('fail',),
])
def test_traversal_branching(self, path):
assert traverse_obj({}, path) == [], \
'if branched but state is `None`, return `[]` (not `default`)'
assert traverse_obj({}, 'fail', path) == [], \
'if branching in last alternative and previous did not match, return `[]` (not `default`)'
assert traverse_obj({0: 'x'}, 0, path) == 'x', \
'if branching in last alternative and previous did match, return single value'
assert traverse_obj({0: 'x'}, path, 0) == 'x', \
'if branching in first alternative and non-branching path does match, return single value'
assert traverse_obj({}, path, 'fail') is None, \
'if branching in first alternative and non-branching path does not match, return `default`'
def test_traversal_expected_type(self):
_EXPECTED_TYPE_DATA = {'str': 'str', 'int': 0}
assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=str) == 'str', \
'accept matching `expected_type` type'
assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=int) is None, \
'reject non matching `expected_type` type'
assert traverse_obj(_EXPECTED_TYPE_DATA, 'int', expected_type=lambda x: str(x)) == '0', \
'transform type using type function'
assert traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=lambda _: 1 / 0) is None, \
'wrap expected_type fuction in try_call'
assert traverse_obj(_EXPECTED_TYPE_DATA, ..., expected_type=str) == ['str'], \
'eliminate items that expected_type fails on'
assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}, expected_type=int) == {0: 100}, \
'type as expected_type should filter dict values'
assert traverse_obj(_TEST_DATA, {0: 100, 1: 1.2, 2: 'None'}, expected_type=str_or_none) == {0: '100', 1: '1.2'}, \
'function as expected_type should transform dict values'
assert traverse_obj(_TEST_DATA, ({0: 1.2}, 0, {int_or_none}), expected_type=int) == 1, \
'expected_type should not filter non final dict values'
assert traverse_obj(_TEST_DATA, {0: {0: 100, 1: 'str'}}, expected_type=int) == {0: {0: 100}}, \
'expected_type should transform deep dict values'
assert traverse_obj(_TEST_DATA, [({0: '...'}, {0: '...'})], expected_type=type(...)) == [{0: ...}, {0: ...}], \
'expected_type should transform branched dict values'
assert traverse_obj({1: {3: 4}}, [(1, 2), 3], expected_type=int) == [4], \
'expected_type regression for type matching in tuple branching'
assert traverse_obj(_TEST_DATA, ['data', ...], expected_type=int) == [], \
'expected_type regression for type matching in dict result'
def test_traversal_get_all(self):
_GET_ALL_DATA = {'key': [0, 1, 2]}
assert traverse_obj(_GET_ALL_DATA, ('key', ...), get_all=False) == 0, \
'if not `get_all`, return only first matching value'
assert traverse_obj(_GET_ALL_DATA, ..., get_all=False) == [0, 1, 2], \
'do not overflatten if not `get_all`'
def test_traversal_casesense(self):
_CASESENSE_DATA = {
'KeY': 'value0',
0: {
'KeY': 'value1',
0: {'KeY': 'value2'},
},
}
assert traverse_obj(_CASESENSE_DATA, 'key') is None, \
'dict keys should be case sensitive unless `casesense`'
assert traverse_obj(_CASESENSE_DATA, 'keY', casesense=False) == 'value0', \
'allow non matching key case if `casesense`'
assert traverse_obj(_CASESENSE_DATA, [0, ('keY',)], casesense=False) == ['value1'], \
'allow non matching key case in branch if `casesense`'
assert traverse_obj(_CASESENSE_DATA, [0, ([0, 'keY'],)], casesense=False) == ['value2'], \
'allow non matching key case in branch path if `casesense`'
def test_traversal_traverse_string(self):
_TRAVERSE_STRING_DATA = {'str': 'str', 1.2: 1.2}
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0)) is None, \
'do not traverse into string if not `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0), traverse_string=True) == 's', \
'traverse into string if `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, (1.2, 1), traverse_string=True) == '.', \
'traverse into converted data if `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', ...), traverse_string=True) == 'str', \
'`...` should result in string (same value) if `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', slice(0, None, 2)), traverse_string=True) == 'sr', \
'`slice` should result in string if `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', lambda i, v: i or v == "s"), traverse_string=True) == 'str', \
'function should result in string if `traverse_string`'
assert traverse_obj(_TRAVERSE_STRING_DATA, ('str', (0, 2)), traverse_string=True) == ['s', 'r'], \
'branching should result in list if `traverse_string`'
assert traverse_obj({}, (0, ...), traverse_string=True) == [], \
'branching should result in list if `traverse_string`'
assert traverse_obj({}, (0, lambda x, y: True), traverse_string=True) == [], \
'branching should result in list if `traverse_string`'
assert traverse_obj({}, (0, slice(1)), traverse_string=True) == [], \
'branching should result in list if `traverse_string`'
def test_traversal_re(self):
mobj = re.fullmatch(r'0(12)(?P<group>3)(4)?', '0123')
assert traverse_obj(mobj, ...) == [x for x in mobj.groups() if x is not None], \
'`...` on a `re.Match` should give its `groups()`'
assert traverse_obj(mobj, lambda k, _: k in (0, 2)) == ['0123', '3'], \
'function on a `re.Match` should give groupno, value starting at 0'
assert traverse_obj(mobj, 'group') == '3', \
'str key on a `re.Match` should give group with that name'
assert traverse_obj(mobj, 2) == '3', \
'int key on a `re.Match` should give group with that name'
assert traverse_obj(mobj, 'gRoUp', casesense=False) == '3', \
'str key on a `re.Match` should respect casesense'
assert traverse_obj(mobj, 'fail') is None, \
'failing str key on a `re.Match` should return `default`'
assert traverse_obj(mobj, 'gRoUpS', casesense=False) is None, \
'failing str key on a `re.Match` should return `default`'
assert traverse_obj(mobj, 8) is None, \
'failing int key on a `re.Match` should return `default`'
assert traverse_obj(mobj, lambda k, _: k in (0, 'group')) == ['0123', '3'], \
'function on a `re.Match` should give group name as well'
def test_traversal_xml_etree(self):
etree = xml.etree.ElementTree.fromstring('''<?xml version="1.0"?>
<data>
<country name="Liechtenstein">
<rank>1</rank>
<year>2008</year>
<gdppc>141100</gdppc>
<neighbor name="Austria" direction="E"/>
<neighbor name="Switzerland" direction="W"/>
</country>
<country name="Singapore">
<rank>4</rank>
<year>2011</year>
<gdppc>59900</gdppc>
<neighbor name="Malaysia" direction="N"/>
</country>
<country name="Panama">
<rank>68</rank>
<year>2011</year>
<gdppc>13600</gdppc>
<neighbor name="Costa Rica" direction="W"/>
<neighbor name="Colombia" direction="E"/>
</country>
</data>''')
assert traverse_obj(etree, '') == etree, \
'empty str key should return the element itself'
assert traverse_obj(etree, 'country') == list(etree), \
'str key should lead all children with that tag name'
assert traverse_obj(etree, ...) == list(etree), \
'`...` as key should return all children'
assert traverse_obj(etree, lambda _, x: x[0].text == '4') == [etree[1]], \
'function as key should get element as value'
assert traverse_obj(etree, lambda i, _: i == 1) == [etree[1]], \
'function as key should get index as key'
assert traverse_obj(etree, 0) == etree[0], \
'int key should return the nth child'
expected = ['Austria', 'Switzerland', 'Malaysia', 'Costa Rica', 'Colombia']
assert traverse_obj(etree, './/neighbor/@name') == expected, \
'`@<attribute>` at end of path should give that attribute'
assert traverse_obj(etree, '//neighbor/@fail') == [None, None, None, None, None], \
'`@<nonexistant>` at end of path should give `None`'
assert traverse_obj(etree, ('//neighbor/@', 2)) == {'name': 'Malaysia', 'direction': 'N'}, \
'`@` should give the full attribute dict'
assert traverse_obj(etree, '//year/text()') == ['2008', '2011', '2011'], \
'`text()` at end of path should give the inner text'
assert traverse_obj(etree, '//*[@direction]/@direction') == ['E', 'W', 'N', 'W', 'E'], \
'full Python xpath features should be supported'
assert traverse_obj(etree, (0, '@name')) == 'Liechtenstein', \
'special transformations should act on current element'
assert traverse_obj(etree, ('country', 0, ..., 'text()', {int_or_none})) == [1, 2008, 141100], \
'special transformations should act on current element'

379
test/test_utils.py
View file

@ -2,7 +2,6 @@
# Allow direct execution
import os
import re
import sys
import unittest
import warnings
@ -45,7 +44,6 @@ from yt_dlp.utils import (
determine_ext,
determine_file_encoding,
dfxp2srt,
dict_get,
encode_base_n,
encode_compat_str,
encodeFilename,
@ -106,13 +104,11 @@ from yt_dlp.utils import (
sanitize_url,
shell_quote,
smuggle_url,
str_or_none,
str_to_int,
strip_jsonp,
strip_or_none,
subtitles_filename,
timeconvert,
traverse_obj,
try_call,
unescapeHTML,
unified_strdate,
@ -755,28 +751,6 @@ class TestUtil(unittest.TestCase):
self.assertRaises(
ValueError, multipart_encode, {b'field': b'value'}, boundary='value')
def test_dict_get(self):
FALSE_VALUES = {
'none': None,
'false': False,
'zero': 0,
'empty_string': '',
'empty_list': [],
}
d = FALSE_VALUES.copy()
d['a'] = 42
self.assertEqual(dict_get(d, 'a'), 42)
self.assertEqual(dict_get(d, 'b'), None)
self.assertEqual(dict_get(d, 'b', 42), 42)
self.assertEqual(dict_get(d, ('a', )), 42)
self.assertEqual(dict_get(d, ('b', 'a', )), 42)
self.assertEqual(dict_get(d, ('b', 'c', 'a', 'd', )), 42)
self.assertEqual(dict_get(d, ('b', 'c', )), None)
self.assertEqual(dict_get(d, ('b', 'c', ), 42), 42)
for key, false_value in FALSE_VALUES.items():
self.assertEqual(dict_get(d, ('b', 'c', key, )), None)
self.assertEqual(dict_get(d, ('b', 'c', key, ), skip_false_values=False), false_value)
def test_merge_dicts(self):
self.assertEqual(merge_dicts({'a': 1}, {'b': 2}), {'a': 1, 'b': 2})
self.assertEqual(merge_dicts({'a': 1}, {'a': 2}), {'a': 1})
@ -2039,359 +2013,6 @@ Line 1
warnings.simplefilter('ignore')
self.assertEqual(variadic('spam', allowed_types=[dict]), 'spam')
def test_traverse_obj(self):
_TEST_DATA = {
100: 100,
1.2: 1.2,
'str': 'str',
'None': None,
'...': ...,
'urls': [
{'index': 0, 'url': 'https://www.example.com/0'},
{'index': 1, 'url': 'https://www.example.com/1'},
],
'data': (
{'index': 2},
{'index': 3},
),
'dict': {},
}
# Test base functionality
self.assertEqual(traverse_obj(_TEST_DATA, ('str',)), 'str',
msg='allow tuple path')
self.assertEqual(traverse_obj(_TEST_DATA, ['str']), 'str',
msg='allow list path')
self.assertEqual(traverse_obj(_TEST_DATA, (value for value in ("str",))), 'str',
msg='allow iterable path')
self.assertEqual(traverse_obj(_TEST_DATA, 'str'), 'str',
msg='single items should be treated as a path')
self.assertEqual(traverse_obj(_TEST_DATA, None), _TEST_DATA)
self.assertEqual(traverse_obj(_TEST_DATA, 100), 100)
self.assertEqual(traverse_obj(_TEST_DATA, 1.2), 1.2)
# Test Ellipsis behavior
self.assertCountEqual(traverse_obj(_TEST_DATA, ...),
(item for item in _TEST_DATA.values() if item not in (None, {})),
msg='`...` should give all non discarded values')
self.assertCountEqual(traverse_obj(_TEST_DATA, ('urls', 0, ...)), _TEST_DATA['urls'][0].values(),
msg='`...` selection for dicts should select all values')
self.assertEqual(traverse_obj(_TEST_DATA, (..., ..., 'url')),
['https://www.example.com/0', 'https://www.example.com/1'],
msg='nested `...` queries should work')
self.assertCountEqual(traverse_obj(_TEST_DATA, (..., ..., 'index')), range(4),
msg='`...` query result should be flattened')
self.assertEqual(traverse_obj(iter(range(4)), ...), list(range(4)),
msg='`...` should accept iterables')
# Test function as key
self.assertEqual(traverse_obj(_TEST_DATA, lambda x, y: x == 'urls' and isinstance(y, list)),
[_TEST_DATA['urls']],
msg='function as query key should perform a filter based on (key, value)')
self.assertCountEqual(traverse_obj(_TEST_DATA, lambda _, x: isinstance(x[0], str)), {'str'},
msg='exceptions in the query function should be catched')
self.assertEqual(traverse_obj(iter(range(4)), lambda _, x: x % 2 == 0), [0, 2],
msg='function key should accept iterables')
if __debug__:
with self.assertRaises(Exception, msg='Wrong function signature should raise in debug'):
traverse_obj(_TEST_DATA, lambda a: ...)
with self.assertRaises(Exception, msg='Wrong function signature should raise in debug'):
traverse_obj(_TEST_DATA, lambda a, b, c: ...)
# Test set as key (transformation/type, like `expected_type`)
self.assertEqual(traverse_obj(_TEST_DATA, (..., {str.upper}, )), ['STR'],
msg='Function in set should be a transformation')
self.assertEqual(traverse_obj(_TEST_DATA, (..., {str})), ['str'],
msg='Type in set should be a type filter')
self.assertEqual(traverse_obj(_TEST_DATA, {dict}), _TEST_DATA,
msg='A single set should be wrapped into a path')
self.assertEqual(traverse_obj(_TEST_DATA, (..., {str.upper})), ['STR'],
msg='Transformation function should not raise')
self.assertEqual(traverse_obj(_TEST_DATA, (..., {str_or_none})),
[item for item in map(str_or_none, _TEST_DATA.values()) if item is not None],
msg='Function in set should be a transformation')
self.assertEqual(traverse_obj(_TEST_DATA, ('fail', {lambda _: 'const'})), 'const',
msg='Function in set should always be called')
if __debug__:
with self.assertRaises(Exception, msg='Sets with length != 1 should raise in debug'):
traverse_obj(_TEST_DATA, set())
with self.assertRaises(Exception, msg='Sets with length != 1 should raise in debug'):
traverse_obj(_TEST_DATA, {str.upper, str})
# Test `slice` as a key
_SLICE_DATA = [0, 1, 2, 3, 4]
self.assertEqual(traverse_obj(_TEST_DATA, ('dict', slice(1))), None,
msg='slice on a dictionary should not throw')
self.assertEqual(traverse_obj(_SLICE_DATA, slice(1)), _SLICE_DATA[:1],
msg='slice key should apply slice to sequence')
self.assertEqual(traverse_obj(_SLICE_DATA, slice(1, 2)), _SLICE_DATA[1:2],
msg='slice key should apply slice to sequence')
self.assertEqual(traverse_obj(_SLICE_DATA, slice(1, 4, 2)), _SLICE_DATA[1:4:2],
msg='slice key should apply slice to sequence')
# Test alternative paths
self.assertEqual(traverse_obj(_TEST_DATA, 'fail', 'str'), 'str',
msg='multiple `paths` should be treated as alternative paths')
self.assertEqual(traverse_obj(_TEST_DATA, 'str', 100), 'str',
msg='alternatives should exit early')
self.assertEqual(traverse_obj(_TEST_DATA, 'fail', 'fail'), None,
msg='alternatives should return `default` if exhausted')
self.assertEqual(traverse_obj(_TEST_DATA, (..., 'fail'), 100), 100,
msg='alternatives should track their own branching return')
self.assertEqual(traverse_obj(_TEST_DATA, ('dict', ...), ('data', ...)), list(_TEST_DATA['data']),
msg='alternatives on empty objects should search further')
# Test branch and path nesting
self.assertEqual(traverse_obj(_TEST_DATA, ('urls', (3, 0), 'url')), ['https://www.example.com/0'],
msg='tuple as key should be treated as branches')
self.assertEqual(traverse_obj(_TEST_DATA, ('urls', [3, 0], 'url')), ['https://www.example.com/0'],
msg='list as key should be treated as branches')
self.assertEqual(traverse_obj(_TEST_DATA, ('urls', ((1, 'fail'), (0, 'url')))), ['https://www.example.com/0'],
msg='double nesting in path should be treated as paths')
self.assertEqual(traverse_obj(['0', [1, 2]], [(0, 1), 0]), [1],
msg='do not fail early on branching')
self.assertCountEqual(traverse_obj(_TEST_DATA, ('urls', ((1, ('fail', 'url')), (0, 'url')))),
['https://www.example.com/0', 'https://www.example.com/1'],
msg='tripple nesting in path should be treated as branches')
self.assertEqual(traverse_obj(_TEST_DATA, ('urls', ('fail', (..., 'url')))),
['https://www.example.com/0', 'https://www.example.com/1'],
msg='ellipsis as branch path start gets flattened')
# Test dictionary as key
self.assertEqual(traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}), {0: 100, 1: 1.2},
msg='dict key should result in a dict with the same keys')
self.assertEqual(traverse_obj(_TEST_DATA, {0: ('urls', 0, 'url')}),
{0: 'https://www.example.com/0'},
msg='dict key should allow paths')
self.assertEqual(traverse_obj(_TEST_DATA, {0: ('urls', (3, 0), 'url')}),
{0: ['https://www.example.com/0']},
msg='tuple in dict path should be treated as branches')
self.assertEqual(traverse_obj(_TEST_DATA, {0: ('urls', ((1, 'fail'), (0, 'url')))}),
{0: ['https://www.example.com/0']},
msg='double nesting in dict path should be treated as paths')
self.assertEqual(traverse_obj(_TEST_DATA, {0: ('urls', ((1, ('fail', 'url')), (0, 'url')))}),
{0: ['https://www.example.com/1', 'https://www.example.com/0']},
msg='tripple nesting in dict path should be treated as branches')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 'fail'}), {},
msg='remove `None` values when top level dict key fails')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 'fail'}, default=...), {0: ...},
msg='use `default` if key fails and `default`')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 'dict'}), {},
msg='remove empty values when dict key')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 'dict'}, default=...), {0: ...},
msg='use `default` when dict key and `default`')
self.assertEqual(traverse_obj(_TEST_DATA, {0: {0: 'fail'}}), {},
msg='remove empty values when nested dict key fails')
self.assertEqual(traverse_obj(None, {0: 'fail'}), {},
msg='default to dict if pruned')
self.assertEqual(traverse_obj(None, {0: 'fail'}, default=...), {0: ...},
msg='default to dict if pruned and default is given')
self.assertEqual(traverse_obj(_TEST_DATA, {0: {0: 'fail'}}, default=...), {0: {0: ...}},
msg='use nested `default` when nested dict key fails and `default`')
self.assertEqual(traverse_obj(_TEST_DATA, {0: ('dict', ...)}), {},
msg='remove key if branch in dict key not successful')
# Testing default parameter behavior
_DEFAULT_DATA = {'None': None, 'int': 0, 'list': []}
self.assertEqual(traverse_obj(_DEFAULT_DATA, 'fail'), None,
msg='default value should be `None`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, 'fail', 'fail', default=...), ...,
msg='chained fails should result in default')
self.assertEqual(traverse_obj(_DEFAULT_DATA, 'None', 'int'), 0,
msg='should not short cirquit on `None`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, 'fail', default=1), 1,
msg='invalid dict key should result in `default`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, 'None', default=1), 1,
msg='`None` is a deliberate sentinel and should become `default`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, ('list', 10)), None,
msg='`IndexError` should result in `default`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=1), 1,
msg='if branched but not successful return `default` if defined, not `[]`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, (..., 'fail'), default=None), None,
msg='if branched but not successful return `default` even if `default` is `None`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, (..., 'fail')), [],
msg='if branched but not successful return `[]`, not `default`')
self.assertEqual(traverse_obj(_DEFAULT_DATA, ('list', ...)), [],
msg='if branched but object is empty return `[]`, not `default`')
self.assertEqual(traverse_obj(None, ...), [],
msg='if branched but object is `None` return `[]`, not `default`')
self.assertEqual(traverse_obj({0: None}, (0, ...)), [],
msg='if branched but state is `None` return `[]`, not `default`')
branching_paths = [
('fail', ...),
(..., 'fail'),
100 * ('fail',) + (...,),
(...,) + 100 * ('fail',),
]
for branching_path in branching_paths:
self.assertEqual(traverse_obj({}, branching_path), [],
msg='if branched but state is `None`, return `[]` (not `default`)')
self.assertEqual(traverse_obj({}, 'fail', branching_path), [],
msg='if branching in last alternative and previous did not match, return `[]` (not `default`)')
self.assertEqual(traverse_obj({0: 'x'}, 0, branching_path), 'x',
msg='if branching in last alternative and previous did match, return single value')
self.assertEqual(traverse_obj({0: 'x'}, branching_path, 0), 'x',
msg='if branching in first alternative and non-branching path does match, return single value')
self.assertEqual(traverse_obj({}, branching_path, 'fail'), None,
msg='if branching in first alternative and non-branching path does not match, return `default`')
# Testing expected_type behavior
_EXPECTED_TYPE_DATA = {'str': 'str', 'int': 0}
self.assertEqual(traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=str),
'str', msg='accept matching `expected_type` type')
self.assertEqual(traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=int),
None, msg='reject non matching `expected_type` type')
self.assertEqual(traverse_obj(_EXPECTED_TYPE_DATA, 'int', expected_type=lambda x: str(x)),
'0', msg='transform type using type function')
self.assertEqual(traverse_obj(_EXPECTED_TYPE_DATA, 'str', expected_type=lambda _: 1 / 0),
None, msg='wrap expected_type fuction in try_call')
self.assertEqual(traverse_obj(_EXPECTED_TYPE_DATA, ..., expected_type=str),
['str'], msg='eliminate items that expected_type fails on')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 100, 1: 1.2}, expected_type=int),
{0: 100}, msg='type as expected_type should filter dict values')
self.assertEqual(traverse_obj(_TEST_DATA, {0: 100, 1: 1.2, 2: 'None'}, expected_type=str_or_none),
{0: '100', 1: '1.2'}, msg='function as expected_type should transform dict values')
self.assertEqual(traverse_obj(_TEST_DATA, ({0: 1.2}, 0, {int_or_none}), expected_type=int),
1, msg='expected_type should not filter non final dict values')
self.assertEqual(traverse_obj(_TEST_DATA, {0: {0: 100, 1: 'str'}}, expected_type=int),
{0: {0: 100}}, msg='expected_type should transform deep dict values')
self.assertEqual(traverse_obj(_TEST_DATA, [({0: '...'}, {0: '...'})], expected_type=type(...)),
[{0: ...}, {0: ...}], msg='expected_type should transform branched dict values')
self.assertEqual(traverse_obj({1: {3: 4}}, [(1, 2), 3], expected_type=int),
[4], msg='expected_type regression for type matching in tuple branching')
self.assertEqual(traverse_obj(_TEST_DATA, ['data', ...], expected_type=int),
[], msg='expected_type regression for type matching in dict result')
# Test get_all behavior
_GET_ALL_DATA = {'key': [0, 1, 2]}
self.assertEqual(traverse_obj(_GET_ALL_DATA, ('key', ...), get_all=False), 0,
msg='if not `get_all`, return only first matching value')
self.assertEqual(traverse_obj(_GET_ALL_DATA, ..., get_all=False), [0, 1, 2],
msg='do not overflatten if not `get_all`')
# Test casesense behavior
_CASESENSE_DATA = {
'KeY': 'value0',
0: {
'KeY': 'value1',
0: {'KeY': 'value2'},
},
}
self.assertEqual(traverse_obj(_CASESENSE_DATA, 'key'), None,
msg='dict keys should be case sensitive unless `casesense`')
self.assertEqual(traverse_obj(_CASESENSE_DATA, 'keY',
casesense=False), 'value0',
msg='allow non matching key case if `casesense`')
self.assertEqual(traverse_obj(_CASESENSE_DATA, (0, ('keY',)),
casesense=False), ['value1'],
msg='allow non matching key case in branch if `casesense`')
self.assertEqual(traverse_obj(_CASESENSE_DATA, (0, ((0, 'keY'),)),
casesense=False), ['value2'],
msg='allow non matching key case in branch path if `casesense`')
# Test traverse_string behavior
_TRAVERSE_STRING_DATA = {'str': 'str', 1.2: 1.2}
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0)), None,
msg='do not traverse into string if not `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', 0),
traverse_string=True), 's',
msg='traverse into string if `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, (1.2, 1),
traverse_string=True), '.',
msg='traverse into converted data if `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', ...),
traverse_string=True), 'str',
msg='`...` should result in string (same value) if `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', slice(0, None, 2)),
traverse_string=True), 'sr',
msg='`slice` should result in string if `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', lambda i, v: i or v == "s"),
traverse_string=True), 'str',
msg='function should result in string if `traverse_string`')
self.assertEqual(traverse_obj(_TRAVERSE_STRING_DATA, ('str', (0, 2)),
traverse_string=True), ['s', 'r'],
msg='branching should result in list if `traverse_string`')
self.assertEqual(traverse_obj({}, (0, ...), traverse_string=True), [],
msg='branching should result in list if `traverse_string`')
self.assertEqual(traverse_obj({}, (0, lambda x, y: True), traverse_string=True), [],
msg='branching should result in list if `traverse_string`')
self.assertEqual(traverse_obj({}, (0, slice(1)), traverse_string=True), [],
msg='branching should result in list if `traverse_string`')
# Test re.Match as input obj
mobj = re.fullmatch(r'0(12)(?P<group>3)(4)?', '0123')
self.assertEqual(traverse_obj(mobj, ...), [x for x in mobj.groups() if x is not None],
msg='`...` on a `re.Match` should give its `groups()`')
self.assertEqual(traverse_obj(mobj, lambda k, _: k in (0, 2)), ['0123', '3'],
msg='function on a `re.Match` should give groupno, value starting at 0')
self.assertEqual(traverse_obj(mobj, 'group'), '3',
msg='str key on a `re.Match` should give group with that name')
self.assertEqual(traverse_obj(mobj, 2), '3',
msg='int key on a `re.Match` should give group with that name')
self.assertEqual(traverse_obj(mobj, 'gRoUp', casesense=False), '3',
msg='str key on a `re.Match` should respect casesense')
self.assertEqual(traverse_obj(mobj, 'fail'), None,
msg='failing str key on a `re.Match` should return `default`')
self.assertEqual(traverse_obj(mobj, 'gRoUpS', casesense=False), None,
msg='failing str key on a `re.Match` should return `default`')
self.assertEqual(traverse_obj(mobj, 8), None,
msg='failing int key on a `re.Match` should return `default`')
self.assertEqual(traverse_obj(mobj, lambda k, _: k in (0, 'group')), ['0123', '3'],
msg='function on a `re.Match` should give group name as well')
# Test xml.etree.ElementTree.Element as input obj
etree = xml.etree.ElementTree.fromstring('''<?xml version="1.0"?>
<data>
<country name="Liechtenstein">
<rank>1</rank>
<year>2008</year>
<gdppc>141100</gdppc>
<neighbor name="Austria" direction="E"/>
<neighbor name="Switzerland" direction="W"/>
</country>
<country name="Singapore">
<rank>4</rank>
<year>2011</year>
<gdppc>59900</gdppc>
<neighbor name="Malaysia" direction="N"/>
</country>
<country name="Panama">
<rank>68</rank>
<year>2011</year>
<gdppc>13600</gdppc>
<neighbor name="Costa Rica" direction="W"/>
<neighbor name="Colombia" direction="E"/>
</country>
</data>''')
self.assertEqual(traverse_obj(etree, ''), etree,
msg='empty str key should return the element itself')
self.assertEqual(traverse_obj(etree, 'country'), list(etree),
msg='str key should lead all children with that tag name')
self.assertEqual(traverse_obj(etree, ...), list(etree),
msg='`...` as key should return all children')
self.assertEqual(traverse_obj(etree, lambda _, x: x[0].text == '4'), [etree[1]],
msg='function as key should get element as value')
self.assertEqual(traverse_obj(etree, lambda i, _: i == 1), [etree[1]],
msg='function as key should get index as key')
self.assertEqual(traverse_obj(etree, 0), etree[0],
msg='int key should return the nth child')
self.assertEqual(traverse_obj(etree, './/neighbor/@name'),
['Austria', 'Switzerland', 'Malaysia', 'Costa Rica', 'Colombia'],
msg='`@<attribute>` at end of path should give that attribute')
self.assertEqual(traverse_obj(etree, '//neighbor/@fail'), [None, None, None, None, None],
msg='`@<nonexistant>` at end of path should give `None`')
self.assertEqual(traverse_obj(etree, ('//neighbor/@', 2)), {'name': 'Malaysia', 'direction': 'N'},
msg='`@` should give the full attribute dict')
self.assertEqual(traverse_obj(etree, '//year/text()'), ['2008', '2011', '2011'],
msg='`text()` at end of path should give the inner text')
self.assertEqual(traverse_obj(etree, '//*[@direction]/@direction'), ['E', 'W', 'N', 'W', 'E'],
msg='full Python xpath features should be supported')
self.assertEqual(traverse_obj(etree, (0, '@name')), 'Liechtenstein',
msg='special transformations should act on current element')
self.assertEqual(traverse_obj(etree, ('country', 0, ..., 'text()', {int_or_none})), [1, 2008, 141100],
msg='special transformations should act on current element')
def test_http_header_dict(self):
headers = HTTPHeaderDict()
headers['ytdl-test'] = b'0'