Automate and check "integral xs" option

[sethrj]

The integral cross section treatment (variously named use_integral_xs or integral_approach; aka G4EmParameters::Integral) is only valid for monotonic nonincreasing cross sections. We pull this parameter from Geant4 usually, but we should validate it ourselves. I think Geant4 recently added an option to handle integral rejection for single- and two-peak cross sections too.

We should either calculate this ourselves from the cross section tables, or pull it directly from Geant4.

[amandalund]

Could you clarify a bit what you mean by "validate [the parameter] ourselves" and "calculate this ourselves"?

[sethrj]

I mean validating whether the cross sections are monotonic... we could do it ourselves for tabulated cross sections but would have to rely on self-reporting for the on-the-fly methods (or validate at runtime that after slowing down, the cross section is no less than the starting XS).

[amandalund]

You mean monotonic over the entire energy range? I don't think that's necessary for the approach to be valid; as implemented I think it just needs to be unimodal. If there's more than one peak, it should be accurate as long as the cross section is monotonic (increasing or decreasing) within the pre- and post-step energy or that range contains the global maximum.