0d usm_ndarray must allow conversion to Python scalars

[oleksandr-pavlyk]

In [2]: import dpctl

In [3]: import dpctl.tensor as dpt

In [4]: dpt.usm_ndarray(tuple(), dtype='d')
Out[4]: <dpctl.tensor._usmarray.usm_ndarray at 0x7f562c2859f0>

In [5]: float(_)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-5-a06b00cb7231> in <module>
----> 1 float(_)

TypeError: float() argument must be a string or a number, not 'dpctl.tensor._usmarray.usm_ndarray'

In[5] should work.

Same for Python integers, and Python bools.