helpers

class InterfaceCase

An interface test helper that allows a given interface to generate and validate arrays in one of its formats.

Each instance of “interface test case” should be considered one of the potentially multiple realizations of a given interface. If an interface has multiple formats (eg. zarr’s different store s), then it should have several test helpers.

abstract property interface: Interface

The interface that this helper is for

classmethod array_from_case(case: ValidationCase, path: Path | None = None) → NDArrayType | None

Generate an array from the given validation case.

Returns None if an array can’t be generated for a specific case.

abstract classmethod make_array(shape: tuple[int, ...] = (10, 10), dtype: str | type | ~typing.Any | ~numpy.generic = <class 'float'>, path: ~pathlib.Path | None = None, array: ~numpydantic.types.NDArrayType | None = None) → NDArrayType | None

Make an array from a shape and dtype, and a path if needed

Parameters:

• shape – shape of the array

• dtype – dtype of the array

• path – Path, if needed to generate on disk

• array – Rather than passing shape and dtype, pass a literal arraylike thing

classmethod validate_case(case: ValidationCase, path: Path) → bool

Validate a generated array against the annotation in the validation case.

Kept in the InterfaceCase in case an interface has specific needs aside from just validating against a model, but typically left as is.

If an array can’t be generated for a given case, returns None so that the calling function can know to skip rather than fail the case.

Raises exceptions if validation fails (or succeeds when it shouldn’t)

Parameters:

• case (ValidationCase) – The validation case to validate.

• path (Path) – Path to generate arrays into, if any.

Returns:

True if array is valid and was supposed to be,

or invalid and wasn’t supposed to be

classmethod skip(shape: tuple[int, ...], dtype: str | type | Any | generic) → bool

Whether a given interface should be skipped for the case

pydantic model ValidationCase

Test case for validating an array.

Contains both the validating model and the parameterization for an array to test in a given interface

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Config:

• arbitrary_types_allowed: bool = True

Fields:

• annotation_dtype (str | type | Any | numpy.generic | collections.abc.Sequence[str | type | Any | numpy.generic])

• annotation_shape (tuple[int | str, ...] | tuple[tuple[int | str, ...], ...])

• dtype (type | numpy.dtype)

• id (str | None)

• interface (type[numpydantic.testing.helpers.InterfaceCase] | None)

• marks (set[str])

• passes (bool)

• path (pathlib.Path | None)

• shape (tuple[int, ...])

field annotation_dtype: str | type | Any | generic | Sequence[str | type | Any | generic] = (<class 'numpy.float16'>, <class 'numpy.float32'>, <class 'numpy.float64'>, <class 'numpy.float32'>, <class 'numpy.float64'>)

Dtype to use in computed annotation used to validate against

field annotation_shape: tuple[int | str, ...] | tuple[tuple[int | str, ...], ...] = (10, 10, '*', '*')

Shape to use in computed annotation used to validate against

field dtype: type | dtype = <class 'float'>

Dtype of the array to validate

field id: str | None = None

String identifying the validation case

field interface: type[InterfaceCase] | None = None

The interface test case to generate and validate the array with

field marks: set[str] [Optional]

pytest marks to set for this test case

field passes: bool = False

Whether the validation should pass or not

field path: Path | None = None

The path to generate arrays into, if any.

field shape: tuple[int, ...] = (10, 10, 2, 2)

Shape of the array to validate

array(path: Path) → NDArrayType

Generate an array for the validation case if we have an interface to do so

merge(other: ValidationCase | Sequence[ValidationCase]) → ValidationCase

Merge two validation cases

Dump both, excluding any unset fields, and merge, preferring other.

valid is True if and only if it is True in both.

skip() → bool

Whether this case should be skipped (eg. due to the interface case being incompatible with the requested dtype or shape)

validate_case(path: Path | None = None) → bool

Whether the generated array correctly validated against the annotation, given the interface

Parameters:

path (pathlib.Path) – Directory to generate array into, if on disk.

Raises:

ValueError – if an interface is missing

property annotation: NDArray[Any, Any]

Annotation used in the model we validate against

property model: type[BaseModel]

A model with a field array with the given annotation

property pytest_marks: list[MarkDecorator]

Instantiated pytest marks from ValidationCase.marks plus the interface name.

merge_cases(*args: ValidationCase) → ValidationCase

Merge multiple validation cases

merged_product(*args: Sequence[ValidationCase], conditions: dict = None) → list[ValidationCase]

Generator for the product of the iterators of validation cases, merging each tuple, and respecting if they should be ValidationCase.skip() or not.

Examples

shape_cases = [
    ValidationCase(shape=(10, 10, 10), passes=True, id="valid shape"),
    ValidationCase(shape=(10, 10), passes=False, id="missing dimension"),
]
dtype_cases = [
    ValidationCase(dtype=float, passes=True, id="float"),
    ValidationCase(dtype=int, passes=False, id="int"),
]

iterator = merged_product(shape_cases, dtype_cases))
next(iterator)
# ValidationCase(
#     shape=(10, 10, 10),
#     dtype=float,
#     passes=True,
#     id="valid shape-float"
# )
next(iterator)
# ValidationCase(
#     shape=(10, 10, 10),
#     dtype=int,
#     passes=False,
#     id="valid shape-int"
# )