[ty] Fix subtyping of type[Any] / type[T] and protocols

[sharkdp]

Summary

This is a bugfix for subtyping of type[Any] / type[T] and protocols.

Test Plan

Regression test that will only be really meaningful once #21553 lands.

[astral-sh-bot]

Diagnostic diff on typing conformance tests

No changes detected when running ty on typing conformance tests ✅

[astral-sh-bot]

mypy_primer results

Changes were detected when running on open source projects

beartype (https://github.com/beartype/beartype)
+ beartype/claw/_package/clawpkgtrie.py:66:29: warning[unsupported-base] Unsupported class base with type `<class 'dict[str, PackagesTrieBlacklist]'> | <class 'dict[str, Divergent]'>`
+ beartype/claw/_package/clawpkgtrie.py:247:29: warning[unsupported-base] Unsupported class base with type `<class 'dict[str, PackagesTrieWhitelist]'> | <class 'dict[str, Divergent]'>`
- Found 498 diagnostics
+ Found 500 diagnostics

xarray (https://github.com/pydata/xarray)
- xarray/core/common.py:1702:22: error[invalid-assignment] Object of type `(_SupportsDType[dtype[Any]] & Top[Mapping[Unknown, object]]) | (str & Top[Mapping[Unknown, object]]) | (tuple[Any, SupportsIndex | Sequence[SupportsIndex]] & Top[Mapping[Unknown, object]]) | ... omitted 3 union elements` is not assignable to `Mapping[Any, dtype[Any] | None | _SupportsDType[dtype[Any]] | ... omitted 4 union elements]`
+ xarray/core/common.py:1702:22: error[invalid-assignment] Object of type `(type[Any] & Top[Mapping[Unknown, object]]) | (str & Top[Mapping[Unknown, object]]) | (tuple[Any, SupportsIndex | Sequence[SupportsIndex]] & Top[Mapping[Unknown, object]]) | ... omitted 4 union elements` is not assignable to `Mapping[Any, dtype[Any] | None | type[Any] | ... omitted 5 union elements]`
- xarray/tests/test_dataarray.py:4157:18: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_dataarray.py:4163:13: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_dataset.py:6905:18: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_dataset.py:6921:18: error[no-matching-overload] No overload of function `full_like` matches arguments
+ xarray/tests/test_plot.py:232:64: warning[unused-ignore-comment] Unused blanket `type: ignore` directive
+ xarray/tests/test_plot.py:1320:55: warning[unused-ignore-comment] Unused blanket `type: ignore` directive
- xarray/tests/test_variable.py:2837:34: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_variable.py:2844:13: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_variable.py:2863:13: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_variable.py:2883:26: error[no-matching-overload] No overload of function `zeros_like` matches arguments
- xarray/tests/test_variable.py:2883:55: error[no-matching-overload] No overload of function `full_like` matches arguments
- xarray/tests/test_variable.py:2890:26: error[no-matching-overload] No overload of function `ones_like` matches arguments
- xarray/tests/test_variable.py:2890:54: error[no-matching-overload] No overload of function `full_like` matches arguments
- Found 1786 diagnostics
+ Found 1777 diagnostics

scikit-build-core (https://github.com/scikit-build/scikit-build-core)
+ src/scikit_build_core/_logging.py:153:13: warning[unsupported-base] Unsupported class base with type `<class 'Mapping[str, Style]'> | <class 'Mapping[str, Divergent]'>`
- Found 41 diagnostics
+ Found 42 diagnostics

jax (https://github.com/google/jax)
- jax/_src/cudnn/scaled_matmul_stablehlo.py:689:54: error[invalid-argument-type] Argument to function `scaled_dot_general_transpose_rhs` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `Unknown | list[Unknown]`
+ jax/_src/cudnn/scaled_matmul_stablehlo.py:689:54: error[invalid-argument-type] Argument to function `scaled_dot_general_transpose_rhs` is incorrect: Expected `str | type[Any] | dtype[Any] | SupportsDType`, found `Unknown | list[Unknown]`
- jax/_src/dtypes.py:637:45: error[invalid-assignment] Object of type `(SupportsDType & tuple[object, ...]) | tuple[str | SupportsDType | dtype[Any], ...]` is not assignable to `tuple[str | SupportsDType | dtype[Any], ...]`
+ jax/_src/dtypes.py:637:45: error[invalid-assignment] Object of type `(SupportsDType & tuple[object, ...]) | tuple[str | type[Any] | dtype[Any] | SupportsDType, ...]` is not assignable to `tuple[str | type[Any] | dtype[Any] | SupportsDType, ...]`
- jax/_src/lax/lax.py:5901:54: error[invalid-argument-type] Argument to function `dot_algorithm_attr` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `str | SupportsDType | dtype[Any] | None`
+ jax/_src/lax/lax.py:5901:54: error[invalid-argument-type] Argument to function `dot_algorithm_attr` is incorrect: Expected `str | type[Any] | dtype[Any] | SupportsDType`, found `str | type[Any] | dtype[Any] | SupportsDType | None`
- jax/_src/lax/lax.py:5901:65: error[invalid-argument-type] Argument to function `dot_algorithm_attr` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `str | SupportsDType | dtype[Any] | None`
+ jax/_src/lax/lax.py:5901:65: error[invalid-argument-type] Argument to function `dot_algorithm_attr` is incorrect: Expected `str | type[Any] | dtype[Any] | SupportsDType`, found `str | type[Any] | dtype[Any] | SupportsDType | None`
- jax/_src/lax/linalg.py:2707:19: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/_src/lax/linalg.py:2712:19: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/_src/lax/other.py:224:7: error[invalid-argument-type] Argument to function `conv_general_dilated_patches` is incorrect: Expected `Precision | None`, found `None | Precision | str | SupportsDType | dtype[Any]`
+ jax/_src/lax/other.py:224:7: error[invalid-argument-type] Argument to function `conv_general_dilated_patches` is incorrect: Expected `Precision | None`, found `None | Precision | str | ... omitted 3 union elements`
- jax/_src/numpy/array_methods.py:566:41: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/_src/numpy/lax_numpy.py:5489:37: error[invalid-parameter-default] Default value of type `<class 'float'>` is not assignable to annotated parameter type `str | SupportsDType | dtype[Any]`
- jax/_src/numpy/lax_numpy.py:5637:21: error[invalid-parameter-default] Default value of type `<class 'float'>` is not assignable to annotated parameter type `str | SupportsDType | dtype[Any]`
- jax/_src/numpy/lax_numpy.py:5723:29: error[invalid-parameter-default] Default value of type `<class 'float'>` is not assignable to annotated parameter type `str | SupportsDType | dtype[Any]`
- jax/_src/numpy/lax_numpy.py:8269:60: error[invalid-argument-type] Argument to function `argmax` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'int'>`
- jax/_src/numpy/lax_numpy.py:8330:60: error[invalid-argument-type] Argument to function `argmin` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'int'>`
- jax/_src/numpy/lax_numpy.py:9021:40: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/_src/numpy/lax_numpy.py:9351:81: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `type`
- jax/_src/numpy/reductions.py:2542:35: error[invalid-argument-type] Argument to function `broadcasted_iota` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'int'>`
- jax/_src/numpy/ufuncs.py:1204:40: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/_src/numpy/ufuncs.py:1248:40: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/_src/pallas/mosaic_gpu/lowering.py:1407:13: error[invalid-assignment] Object of type `str | SupportsDType | dtype[Any] | None` is not assignable to `dtype[Any]`
+ jax/_src/pallas/mosaic_gpu/lowering.py:1407:13: error[invalid-assignment] Object of type `str | type[Any] | dtype[Any] | SupportsDType | None` is not assignable to `dtype[Any]`
- jax/_src/pallas/mosaic_gpu/lowering.py:2054:60: error[invalid-argument-type] Argument to function `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/_src/random.py:816:47: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/_src/scipy/linalg.py:2194:28: error[invalid-argument-type] Argument to function `broadcasted_iota` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'float'>`
- jax/_src/tpu/linalg/eigh.py:621:32: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/_src/tpu/linalg/eigh.py:623:48: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/_src/tpu/linalg/eigh.py:627:37: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/_src/tpu/linalg/eigh.py:629:53: error[invalid-argument-type] Argument to function `_tri` is incorrect: Expected `str | SupportsDType | dtype[Any]`, found `<class 'bool'>`
- jax/experimental/jax2tf/tests/flax_models/seq2seq_lstm.py:83:41: error[invalid-argument-type] Argument to function `zeros` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/experimental/pallas/ops/gpu/attention.py:662:68: error[invalid-argument-type] Argument to function `ones` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'bool'>`
- jax/experimental/pallas/ops/gpu/transposed_ragged_dot_mgpu.py:59:21: error[invalid-argument-type] Argument to function `zeros` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/pallas/ops/gpu/transposed_ragged_dot_mgpu.py:60:12: error[invalid-argument-type] Argument to bound method `astype` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:582:94: error[invalid-argument-type] Argument to function `zeros` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:1122:53: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:1123:53: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:1124:85: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:1125:85: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:1758:34: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- jax/experimental/sparse/bcoo.py:2238:46: error[invalid-argument-type] Argument to function `array` is incorrect: Expected `str | SupportsDType | dtype[Any] | None`, found `<class 'int'>`
- Found 2733 diagnostics
+ Found 2702 diagnostics

No memory usage changes detected ✅

[sharkdp]

These errors prevent us from running into the problem on main. But on #21553, these errors will go away, and then we need to properly understand all of this code here, or otherwise we see np.array calls failing all over the place.

[AlexWaygood]

Some suggestions, but they're untested, so accept them with care 😄

[astral-sh-bot]

ecosystem-analyzer results

Lint rule	Added	Removed	Changed
invalid-argument-type	0	28	4
no-matching-overload	0	11	0
invalid-assignment	0	0	3
invalid-parameter-default	0	3	0
unused-ignore-comment	2	0	0
Total	2	42	7

Full report with detailed diff (timing results)