Define which additional objects need to be passed for each RECO probability.

[namapane]

In a follow up to #345, it was stated that "for each probability, one needs to pass to MELA only the additional obects that are needed for that probability". This is not the case in the current implementation which now always passes extra jets and leptons (up to 2 each), regardless of which probability is being computed [here].

In #345, @AlessandroTarabini indeed noticed that passing 1 or 2 jets gives different results for D_VBF_1jet.

Action items:

1. If the above is confirmed, we will need some extra configuration flags for probabilities to specify which objects should be considered, and up to how many; or otherwise a mechanism to specify which ones needs to be taken into account in MELA, as was done in the miniAOD fwk; and the corresponding logic in the code.
2. We now pass only up to 2 jets and 2 leptons (the leading one passing selection cuts, regardless of flavour/charge in case of leptons). We need to confirm whether for some probability that may not be enough - in that case, some extra updates would be needed.

Note: for the case of D_VBF_1jet, this we concluded that this should be looked at only for genuine 1-jet events, so it does not matter of what happens for 2-jet events. But then we should store -1 (or the like) for these events, not a possibly misleading value (and save the time of the computation)

@AlessandroTarabini does this make sense to you?
@pianonick411 it would possibly be useful to check the above with other MELA experts at JHU.

[pianonick411]

@namapane, I will confer with others here and report tomorrow.