Improve model evaluation to match XSPEC behavior

[renecotyfanboy]

I have identified several differences between XSPEC and jaxspec when it comes to evaluating models.

1. Determine how multiplicative components are evaluated. Some models (i.e. APEC) yield an integrated flux between two energy bins, so we can't evaluate the additive components multiplied by the multiplicative components at several energies, and integrate it. We must understand how XSPEC deals with this specific situation, and my guess is that it evaluates a mean absorption within the bin by averaging the tabulated values (or equivalent).