feat(numpy backend): removed manual dtype casting.

[Madjid-CH]

PR Description

used script for dtypes checking:

import numpy as np

dtypes = [
    np.dtype("int8"),
    np.dtype("int16"),
    np.dtype("int32"),
    np.dtype("int64"),
    np.dtype("uint8"),
    np.dtype("uint16"),
    np.dtype("uint32"),
    np.dtype("uint64"),
    np.dtype("float16"),
    np.dtype("float32"),
    np.dtype("float64"),
    np.dtype("complex64"),
    np.dtype("complex128"),
    np.dtype("bool"),
]


def get_supported_and_unsupported_dtypes():
    supported_dtypes = set()
    unsupported_dtypes = set()

    for dtype in dtypes:
        try:
            x = np.array(
                [[1, 2], [2, 1]],
                dtype=dtype,
            )
            x1 = np.array([1, 1, 1], dtype=dtype)
            x2 = np.array([1, 2, 3], dtype=dtype)
            print(dtype,  cummax(x1))
            supported_dtypes.add(dtype)
        except Exception as e:
            print(e)
            unsupported_dtypes.add(dtype)

    return supported_dtypes, unsupported_dtypes


if __name__ == "__main__":
    supported_dtypes, unsupported_dtypes = get_supported_and_unsupported_dtypes()

    supported_dtypes = sorted(str(d) for d in supported_dtypes)
    unsupported_dtypes = sorted(str(d) for d in unsupported_dtypes)


    print("Supported Data Types:")
    m = f"""@with_supported_dtypes(
            {{"1.25.2 and below": {tuple(supported_dtypes)}}}, backend_version
        )"""
    print(m)

    print("\nUnsupported Data Types:")
    m = f"""@with_unsupported_dtypes(
            {{"1.25.2 and below": {tuple(unsupported_dtypes)}}}, backend_version
        )"""
    print(m)

[github-actions]

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

[Madjid-CH]

result of running the script on cummax:

int8 (array([1, 1, 1], dtype=int8), array([0, 1, 2], dtype=int64))
int16 (array([1, 1, 1], dtype=int16), array([0, 1, 2], dtype=int64))
int32 (array([1, 1, 1]), array([0, 1, 2], dtype=int64))
int64 (array([1, 1, 1], dtype=int64), array([0, 1, 2], dtype=int64))
uint8 (array([1, 1, 1], dtype=uint8), array([0, 1, 2], dtype=int64))
uint16 (array([1, 1, 1], dtype=uint16), array([0, 1, 2], dtype=int64))
uint32 (array([1, 1, 1], dtype=uint32), array([0, 1, 2], dtype=int64))
uint64 (array([1, 1, 1], dtype=uint64), array([0, 1, 2], dtype=int64))
float16 (array([1., 1., 1.], dtype=float16), array([0, 1, 2], dtype=int64))
float32 (array([1., 1., 1.], dtype=float32), array([0, 1, 2], dtype=int64))
float64 (array([1., 1., 1.]), array([0, 1, 2], dtype=int64))
complex64 (array([1.+0.j, 1.+0.j, 1.+0.j], dtype=complex64), array([0, 1, 2], dtype=int64))
complex128 (array([1.+0.j, 1.+0.j, 1.+0.j]), array([0, 1, 2], dtype=int64))
bool (array([ True,  True,  True]), array([0, 1, 2], dtype=int64))
Supported Data Types:
@with_supported_dtypes(
        {"1.25.2 and below": ('bool', 'complex128', 'complex64', 'float16', 'float32', 'float64', 'int16', 'int32', 'int64', 'int8', 'uint16', 'uint32', 'uint64', 'uint8')}, backend_version
    )
Unsupported Data Types:
@with_unsupported_dtypes(
        {"1.25.2 and below": ()}, backend_version
    )

[Madjid-CH]

same for cummin

int8 [1 1 1]
int16 [1 1 1]
int32 [1 1 1]
int64 [1 1 1]
uint8 [1 1 1]
uint16 [1 1 1]
uint32 [1 1 1]
uint64 [1 1 1]
float16 [1. 1. 1.]
float32 [1. 1. 1.]
float64 [1. 1. 1.]
complex64 [1.+0.j 1.+0.j 1.+0.j]
complex128 [1.+0.j 1.+0.j 1.+0.j]
bool [ True  True  True]
Supported Data Types:
@with_supported_dtypes(
        {"1.25.2 and below": ('bool', 'complex128', 'complex64', 'float16', 'float32', 'float64', 'int16', 'int32', 'int64', 'int8', 'uint16', 'uint32', 'uint64', 'uint8')}, backend_version
    )
Unsupported Data Types:
@with_unsupported_dtypes(
        {"1.25.2 and below": ()}, backend_version
    )

[Madjid-CH]

int8 [1 2 6]
int16 [1 2 6]
int32 [1 2 6]
int64 [1 2 6]
uint8 [1 2 6]
uint16 [1 2 6]
uint32 [1 2 6]
uint64 [1 2 6]
float16 [1. 2. 6.]
float32 [1. 2. 6.]
float64 [1. 2. 6.]
complex64 [1.+0.j 2.+0.j 6.+0.j]
complex128 [1.+0.j 2.+0.j 6.+0.j]
bool [ True  True  True]
Supported Data Types:
@with_supported_dtypes(
        {"1.25.2 and below": ('bool', 'complex128', 'complex64', 'float16', 'float32', 'float64', 'int16', 'int32', 'int64', 'int8', 'uint16', 'uint32', 'uint64', 'uint8')}, backend_version
    )
Unsupported Data Types:
@with_unsupported_dtypes(
        {"1.25.2 and below": ()}, backend_version
    )

[Madjid-CH]

int8 [1 3 6]
int16 [1 3 6]
int32 [1 3 6]
int64 [1 3 6]
uint8 [1 3 6]
uint16 [1 3 6]
uint32 [1 3 6]
uint64 [1 3 6]
float16 [1. 3. 6.]
float32 [1. 3. 6.]
float64 [1. 3. 6.]
complex64 [1.+0.j 3.+0.j 6.+0.j]
complex128 [1.+0.j 3.+0.j 6.+0.j]
bool [ True  True  True]
Supported Data Types:
@with_supported_dtypes(
        {"1.25.2 and below": ('bool', 'complex128', 'complex64', 'float16', 'float32', 'float64', 'int16', 'int32', 'int64', 'int8', 'uint16', 'uint32', 'uint64', 'uint8')}, backend_version
    )
Unsupported Data Types:
@with_unsupported_dtypes(
        {"1.25.2 and below": ()}, backend_version
    )

[vedpatwardhan]

lgtm! Feel free to merge, thanks @Madjid-CH 😄
(I'm assuming you've already run the tests for the functions changed and the tests pass or fail, whichever was the state of those tests before the changes. And even if it fails, the logs of the failure are the same as those before making the changes.)

[Madjid-CH]

lgtm! Feel free to merge, thanks @Madjid-CH 😄 (I'm assuming you've already run the tests for the functions changed and the tests pass or fail, whichever was the state of those tests before the changes. And even if it fails, the logs of the failure are the same as those before making the changes.)

yeah, the tests was failing because of a missing on_device attribute and now they are passing except for cummax, it test is still failing after modification.