sigmaFromCanonical 0: divide by zero

Since ppDivPowerSum doesn't handle k=0 correctly, the right solution seems to be mimicking the behavior of sigma as below. There doesn't seem to be a bug with ppDivPowerSum k=1, but it might produce a small speedup?

sigmaFromCanonical 0 = tauFromCanonical sigmaFromCanonical 1 = divisorSumFromCanonical