Change fitOrder depending on the FLUXSTD distribution

Some visits have highly skewed distribution of FLUXSTDs over the focal plane. In these visits, we want to reduce the polynomial fitting order. I do not think we need to be very flexible here (because this not a fluxCal problem but a fiber configuration problem), and my current plan is to implement a simple algorithm to characterize the skewness, and if the distribution is highly inhomogeneous, we set fitOrder=1. Otherwise, the default fitOrder is used.