Numpy Type Annotation

May 20, 2026

Numpy type annotation numpy.typing.ArrayLike is often used on the input variables of a function. This doesn’t cover all cases a function might accept. For example, numerous Numpy functions can accept Python datetime that becomes an opaque object when passed to a function. It can be better to pass datetime into Numpy functions as numpy.datetime64, which is a subclass of numpy.ndarray and thus can be annotated as numpy.typing.NDArray[numpy.datetime64], especially as the return type of a function. Currently, Numpy type annotation doesn’t handle the shape of arrays, just the data type.