CPAlgorithms: allow running Full JER on data

This MR introduces the possibility to run systematics on data, which is only suggested for the "Full"/"All" models of JER.

Note that with the current setup, running the same config with e.g. "Full" JER enabled, would produce systematics on data AND the "pseudo-data" variations on MC. Getting both in the same ntuple production is probably not ideal, storage-wise. Maybe someone wants to suggest a preferred approach? e.g. making JER-on-data an expert option that is off by default?

--> introduced expert option runJERsystematicsOnData

@tdado @fballi @jdandoy @tstreble @skorn @thsteven @ztao @dbaronmo