Fix union type bug in pipeline type-checker and enhance Entity ID type

lenskit/lenskit/data/query.py

@@ -71,7 +71,7 @@ def create(cls, data: QueryInput) -> RecQuery:
             assert_never(f"invalid type {type(data)}")
 
 
-QueryInput: TypeAlias = RecQuery | EntityId | ItemList | np.integer | None
+QueryInput: TypeAlias = RecQuery | EntityId | ItemList | None
 """
 Types that can be converted to a query by :meth:`RecQuery.create`.
 """

lenskit/lenskit/data/types.py

@@ -20,7 +20,7 @@
 FeedbackType: TypeAlias = Literal["explicit", "implicit"]
 "Types of feedback supported."
 
-EntityId: TypeAlias = int | str | bytes
+EntityId: TypeAlias = int | str | bytes | np.integer[Any] | np.string_
 "Allowable entity identifier types."
 NPEntityId: TypeAlias = np.integer[Any] | np.str_ | np.bytes_ | np.object_
 "Allowable entity identifier types (NumPy version)"

lenskit/lenskit/pipeline/types.py

@@ -10,7 +10,7 @@
 import re
 import warnings
 from importlib import import_module
-from types import GenericAlias, NoneType
+from types import GenericAlias, NoneType, UnionType
 from typing import (  # type: ignore
     Generic,
     Protocol,
@@ -128,7 +128,8 @@ def is_compatible_data(obj: object, *targets: type | TypeVar) -> bool:
         except TypeError:
             pass
 
-        if get_origin(target) == Union:
+        origin = get_origin(target)
+        if origin == UnionType or origin == Union:
             types = get_args(target)
             if is_compatible_data(obj, *types):
                 return True

lenskit/tests/pipeline/test_types.py

@@ -12,7 +12,7 @@
 from collections.abc import Iterable, Sequence
 from pathlib import Path
 from types import NoneType
-from typing import TypeVar
+from typing import Any, TypeVar
 
 import numpy as np
 import pandas as pd
@@ -102,6 +102,14 @@ def test_numpy_typecheck():
     assert not is_compatible_data(np.arange(10), NDArray[np.float64])
 
 
+def test_numpy_scalar_typecheck():
+    assert is_compatible_data(np.int32(4270), np.integer[Any])
+
+
+def test_numpy_scalar_typecheck2():
+    assert is_compatible_data(np.int32(4270), np.integer[Any] | int)
+
+
 def test_pandas_typecheck():
     assert is_compatible_data(pd.Series(["a", "b"]), ArrayLike)