Made sure to include absolute value in uncertainty for quantities that may be...

Made sure to include absolute value in uncertainty for quantities that may be negative (I believe my code was running into an issue and generating negative tp's). Added logic for taking the relevant higher order system directly from the database if the compositions are within a (user editable) tolerance, taken by default to be 0.025.This fixed the density prediction for FLiNaK and FLiBe, and now all plots are consistent with references.