dpctl.tensor.concat does not support axis=None mandated by array API

[oleksandr-pavlyk]

Array API provisions support for axis=None in concat function.

If axis is None, arrays must be flattened before concatenation.

This is presently not supported in dpctl 0.14.2:

Evidence of the lack thereof (excerpt of the IPython sessions)

In [1]: import dpctl.tensor as dpt

In [2]: x0 = dpt.ones((3,2,), dtype="i4")

In [3]: x1 = dpt.full((3,3,), 2, dtype="i4")

In [4]: dpt.concat((x0, x1), axis=1)
Out[4]:
usm_ndarray([[1, 1, 2, 2, 2],
             [1, 1, 2, 2, 2],
             [1, 1, 2, 2, 2]], dtype=int32)

In [5]: dpt.concat((x0, x1), axis=None)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Input In [5], in <cell line: 1>()
----> 1 dpt.concat((x0, x1), axis=None)

File ~/repos/dpctl/dpctl/tensor/_manipulation_functions.py:479, in concat(arrays, axis)
    476 n = len(arrays)
    477 X0 = arrays[0]
--> 479 axis = normalize_axis_index(axis, X0.ndim)
    480 X0_shape = X0.shape
    481 _check_same_shapes(X0_shape, axis, n, arrays)

TypeError: an integer is required (got type NoneType)