some experimental scratch work on generic typing - don't look @ me lol

src/numpydantic/interface/interface.py

@@ -333,8 +333,11 @@ class Interface(ABC, Generic[T]):
             :class:`~numpydantic.exceptions.ShapeError`
         """
         if not valid:
+            from numpydantic.validation.shape import to_shape
+
+            _shape = to_shape(self.shape)
             raise ShapeError(
-                f"Invalid shape! expected shape {self.shape.prepared_args}, "
+                f"Invalid shape! expected shape {_shape.prepared_args}, "
                 f"got shape {shape}"
             )
 

src/numpydantic/ndarray_generic.py

@@ -0,0 +1,72 @@
+from typing import Protocol, TypeVar, runtime_checkable
+
+from typing_extensions import Unpack
+
+from numpydantic.types import DtypeType
+
+# Shape = TypeVarTuple("Shape")
+# Shape = tuple[int, ...]
+Shape = TypeVar("Shape", bound=tuple[int, ...])
+DType = TypeVar("DType", bound=DtypeType)
+
+
+@runtime_checkable
+class NDArray(Protocol[Shape, DType]):
+    """v2 generic protocol ndarray"""
+
+    @property
+    def dtype(self) -> DType:
+        """dtype"""
+
+    @property
+    def shape(self) -> Unpack[Shape]:
+        """shape"""
+
+
+#
+#
+# def __get_pydantic_core_schema__(
+#     typ: Type, handler: CallbackGetCoreSchemaHandler
+# ) -> core_schema.CoreSchema:
+#     args = get_args(typ)
+#     if len(args) == 0:
+#         shape, dtype = Any, Any
+#     elif len(args) == 1:
+#         shape, dtype = args[0], Any
+#     elif len(args) == 2:
+#         shape, dtype = args[0], args[1]
+#     else:
+#         shape, dtype = args[:-1], args[-1]
+#
+#     json_schema = make_json_schema(shape, dtype, handler)
+#     return core_schema.with_info_plain_validator_function(
+#         get_validate_interface(shape, dtype),
+#         serialization=core_schema.plain_serializer_function_ser_schema(
+#             jsonize_array, when_used="json", info_arg=True
+#         ),
+#         metadata=json_schema,
+#     )
+
+
+#
+# def __get_pydantic_json_schema__(
+#     schema: core_schema.CoreSchema, handler: GetJsonSchemaHandler
+# ) -> core_schema.JsonSchema:
+#     # shape, dtype = cls.__args__
+#     json_schema = handler(schema["metadata"])
+#     json_schema = handler.resolve_ref_schema(json_schema)
+#
+#     # if not isinstance(dtype, tuple) and dtype.__module__ not in (
+#     #     "builtins",
+#     #     "typing",
+#     # ):
+#     #     json_schema["dtype"] = ".".join([dtype.__module__, dtype.__name__])
+#
+#     return json_schema
+
+
+# NDArray = Annotated[
+#     _NDArray[Unpack[Shape], DType],
+#     GetPydanticSchema(__get_pydantic_core_schema__),
+#     # GetJsonSchemaFunction(__get_pydantic_json_schema__),
+# ]

src/numpydantic/schema.py

@@ -113,7 +113,9 @@ def list_of_lists_schema(shape: "Shape", array_type: CoreSchema) -> ListSchema:
         array_type ( :class:`pydantic_core.CoreSchema` ): The pre-rendered pydantic
             core schema to use in the innermost list entry
     """
-    from numpydantic.validation.shape import _is_range
+    from numpydantic.validation.shape import _is_range, to_shape
+
+    shape = to_shape(shape)
 
     shape_parts = [part.strip() for part in shape.__args__[0].split(",")]
     # labels, if present

src/numpydantic/validation/shape.py

@@ -91,6 +91,16 @@ class Shape(NPTypingType, ABC, metaclass=ShapeMeta):
     prepared_args = ("*", "...")
 
 
+def to_shape(shape) -> "Shape":
+    from numpydantic import Shape
+
+    if isinstance(shape, int):
+        shape = Shape[f"{shape}"]
+    elif isinstance(shape, tuple):
+        shape = Shape[f"{', '.join([s for s in shape])}"]
+    return shape
+
+
 def validate_shape_expression(shape_expression: Union[ShapeExpression, Any]) -> None:
     """
     CHANGES FROM NPTYPING: Allow ranges
@@ -112,6 +122,7 @@ def validate_shape(shape: ShapeTuple, target: "Shape") -> bool:
     :param target: the shape expression to which shape is tested.
     :return: True if the given shape corresponds to shape_expression.
     """
+    target = to_shape(target)
     target_shape = _handle_ellipsis(shape, target.prepared_args)
     return _check_dimensions_against_shape(shape, target_shape)
 

tests/test_generic.py

@@ -0,0 +1,19 @@
+from numpydantic.ndarray_generic import NDArray
+from pydantic import BaseModel
+from typing import Literal as L
+import numpy as np
+
+
+class MyClass(BaseModel):
+    array: NDArray[L[4, 5], int]
+
+
+model = MyClass(array=np.array([1, 2, 3, 4]))
+
+model2 = MyClass(array=np.array([1, 2, 3]))
+
+model3 = MyClass(array=(1, 2))
+
+
+array: NDArray[L[4], np.int64] = np.array([1, 2, 3])
+array2: NDArray[L[3], np.int64] = [1, 2, 3]

tests/test_ndarray.py

@@ -1,16 +1,15 @@
-import pytest
-
-from typing import Union, Optional, Any
 import json
+from typing import Any, Optional, Union
 
 import numpy as np
-from pydantic import BaseModel, ValidationError, Field
+import pytest
+from pydantic import BaseModel, Field, ValidationError
 
-
+# from numpydantic import NDArray, Shape
-from numpydantic import NDArray, Shape
+from numpydantic import Shape, dtype
-from numpydantic.exceptions import ShapeError, DtypeError
-from numpydantic import dtype
 from numpydantic.dtype import Number
+from numpydantic.exceptions import DtypeError
+from numpydantic.ndarray_generic import NDArray
 
 
 @pytest.mark.json_schema
@@ -177,9 +176,10 @@ def test_shape_ellipsis():
     """
 
     class MyModel(BaseModel):
-        array: NDArray[Shape["1, 2, ..."], Number]
+        array: NDArray[1, 2, ..., Number]
 
     _ = MyModel(array=np.zeros((1, 2, 3, 4, 5)))
+    _ = MyModel(array="hey")
 
 
 @pytest.mark.serialization