Add unit check for sum vs integral in line flux + moments

[rosteen]

Currently, line flux does an integral (well, a sum multiplying by dx, see #712 for more discussion), whereas the moment calculations are doing simple sums. I'm fixing the moment calculations to match line flux in #1141, but really for both of these we should have a check to do the simple sum in the appropriate case of flux units that are actually flux (e.g. W/m2) rather than the usual case of flux density (e.g. Jy) that requires the integral.