Missing a boundary check for sorter argument in dpctl.tensor.searchsorted

[antonwolfy]

In below example sorter has an index which is out of the range of input array a, but no validation exception raised by dpctl:

a = dpt.asarray([5, 2, 1, 3, 4])
v = dpt.asarray(4)
sorter = dpt.asarray([0, 1, 2, 3, 5])

dpt.searchsorted(a, v, sorter=sorter)
# Out: usm_ndarray(4)

a = numpy.asarray([5, 2, 1, 3, 4])
v = numpy.asarray(4)
sorter = numpy.asarray([0, 1, 2, 3, 5])
numpy.searchsorted(a, v, sorter=sorter)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[10], line 1
----> 1 numpy.searchsorted(a, v, sorter=sorter)

File ~/miniconda3/envs/dpnp_dev/lib/python3.9/site-packages/numpy/core/fromnumeric.py:1400, in searchsorted(a, v, side, sorter)
   1332 @array_function_dispatch(_searchsorted_dispatcher)
   1333 def searchsorted(a, v, side='left', sorter=None):
   1334     """
   1335     Find indices where elements should be inserted to maintain order.
   1336
   (...)
   1398
   1399     """
-> 1400     return _wrapfunc(a, 'searchsorted', v, side=side, sorter=sorter)

File ~/miniconda3/envs/dpnp_dev/lib/python3.9/site-packages/numpy/core/fromnumeric.py:59, in _wrapfunc(obj, method, *args, **kwds)
     56     return _wrapit(obj, method, *args, **kwds)
     58 try:
---> 59     return bound(*args, **kwds)
     60 except TypeError:
     61     # A TypeError occurs if the object does have such a method in its
     62     # class, but its signature is not identical to that of NumPy's. This
   (...)
     66     # Call _wrapit from within the except clause to ensure a potential
     67     # exception has a traceback chain.
     68     return _wrapit(obj, method, *args, **kwds)

ValueError: Sorter index out of range.

[oleksandr-pavlyk]

Yes, we use default index wrap mode. Perhaps this needs to be explicitly documented, but it is done for performance reasons. It is consistent with behavior of __getitem__ and of dpctl.tensor.take.

dpnp can use dpt.min, dpt.max to implement its own validation, if deemed necessary.

I will expand the documentation string.