Python/numpy: catch IEEE-754 “inexact” exception

Question

numpy allows one to handle IEEE-754 exceptions originating from floating-point by using np.seterr appropriately. However seterr only supports the f