Fix usm_ndarray ctor when shape is integral numpy scalar

[oleksandr-pavlyk]

This enables dpt.usm_ndarray(np.prod((2,3,4))).

• Have you provided a meaningful PR description?
• Have you added a test, reproducer or referred to an issue with a reproducer?
• Have you tested your changes locally for CPU and GPU devices?
• Have you made sure that new changes do not introduce compiler warnings?
• Have you checked performance impact of proposed changes?
• If this PR is a work in progress, are you opening the PR as a draft?

[github-actions]

View rendered docs @ https://intelpython.github.io/dpctl/pulls/1467/index.html

[coveralls]

coverage: 85.783% (-0.002%) from 85.785%
when pulling da59476 on fix-usm-ndarray-ctor-when-shape-is-integral-numpy-scalar
into 9018745 on master.

[github-actions]

Array API standard conformance tests for dpctl=0.15.1dev1=py310ha25a700_4 ran successfully.
Passed: 935
Failed: 65
Skipped: 119

[ndgrigorian]

Perhaps the logic of this exception and message could be improved a bit. For instance:

In [5]: x = dpt.ones(np.prod((2, 3, 4), dtype="f4"), dtype="i8")
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
File ~/repos/dpctl/dpctl/tensor/_usmarray.pyx:192, in dpctl.tensor._usmarray.usm_ndarray.__cinit__()
    191 try:
--> 192     <Py_ssize_t> shape
    193     shape = [shape, ]

TypeError: 'float' object cannot be interpreted as an integer

During handling of the above exception, another exception occurred:

TypeError                                 Traceback (most recent call last)
Cell In[5], line 1
----> 1 x = dpt.ones(np.prod((2, 3, 4), dtype="f4"), dtype="i8")

File ~/repos/dpctl/dpctl/tensor/_ctors.py:968, in ones(shape, dtype, order, device, usm_type, sycl_queue)
    966 sycl_queue = normalize_queue_device(sycl_queue=sycl_queue, device=device)
    967 dtype = _get_dtype(dtype, sycl_queue)
--> 968 res = dpt.usm_ndarray(
    969     shape,
    970     dtype=dtype,
    971     buffer=usm_type,
    972     order=order,
    973     buffer_ctor_kwargs={"queue": sycl_queue},
    974 )
    975 hev, _ = ti._full_usm_ndarray(1, res, sycl_queue)
    976 hev.wait()

File ~/repos/dpctl/dpctl/tensor/_usmarray.pyx:195, in dpctl.tensor._usmarray.usm_ndarray.__cinit__()
    193     shape = [shape, ]
    194 except Exception:
--> 195     raise TypeError(
    196         "Argument shape must be a list or a tuple."
    197     )

TypeError: Argument shape must be a list or a tuple.

It seems a bit misleading at first, because it would work for np.prod((2, 3, 4)).

[github-actions]

Array API standard conformance tests for dpctl=0.15.1dev1=py310ha25a700_5 ran successfully.
Passed: 935
Failed: 65
Skipped: 119

[ndgrigorian]

LGTM!