numpy.sum() returns inconsistent types

With 0D FieldArray inputs, numpy.sum() returns a FieldArray, as expected:

import galois
import numpy

>>> f = galois.Field(2**64-59)
>>> a = f.Random()
>>> b = f.Random()
>>> numpy.sum([a, b], axis=0)
GF(5802771140473745109, order=18446744073709551557)

However, with 1D FieldArray inputs, numpy.sum() returns a numpy.ndarray:

>>> a = f.Random((2,))
>>> b = f.Random((2,))
>>> numpy.sum([a, b], axis=0)
array([27182153923255896342, 13808647080647953998], dtype=object)

I’ve verified that numpy.add.reduce() works the same, and sum() doesn’t have a subok argument. Is this expected behavior?

Cheers, Tim

[tshead2]

Yup, appreciate your help.

Cheers, Tim

[mhostetter]

No problem. Is this issue good to close?