Validation against NumPy dtypes -- what's the least circuitous way to check values? Validation against NumPy dtypes -- what's the least circuitous way to check values? numpy numpy