Fixes #698

[oleksandr-pavlyk]

Fixes #698

When processing sequence of sequences, pass requested array data type
to numpy converter to ensure that the expected ValueError is raised:

In [2]: import dpctl.tensor as dpt

In [3]: dpt.asarray([float("nan"), -1, 1], dtype="i8")
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-3-5393cc22ddff> in <module>
----> 1 dpt.asarray([float("nan"), -1, 1], dtype="i8")

/.../dpctl/dpctl/tensor/_ctors.py in asarray(obj, dtype, device, copy, usm_type, sycl_queue, order)
    389         if devs == _host_set:
    390             return _asarray_from_numpy_ndarray(
--> 391                 np.asarray(obj, dtype=dtype, order=order),
    392                 dtype=dtype,
    393                 usm_type=usm_type,

~/..../site-packages/numpy/core/_asarray.py in asarray(a, dtype, order, like)
    100         return _asarray_with_like(a, dtype=dtype, order=order, like=like)
    101
--> 102     return array(a, dtype, copy=False, order=order)
    103
    104

ValueError: cannot convert float NaN to integer

In [4]: quit

[github-actions]

View rendered docs @ https://intelpython.github.io/dpctl/pulls/709/index.html

[coveralls]

Coverage remained the same at 81.194% when pulling 9c7426f on gh-698-asarray into 431bed8 on master.

[github-actions]

Deleted rendered PR docs from intelpython.github.com/dpctl, latest should be updated shortly. 🤞