Sum of genweights as a dataset metadata parameter

Created by: mmarchegiani

With the present PR we introduce a new scheme for rescaling MC samples by the sum of genweights. Instead of computing the sum of the genweights each time the ttHbbBaseProcessor is run and rescale MC in postprocess(), the sum of genweights are computed with a separate processor genWeightsProcessor once for all and included in the datasets definitions as an additional metadata parameter.

These changes allow to label different datasets to belong to the same sample by keeping track of their cross-section and sum of genweights through the metadata. Additionally, an extra key part is added to the metadata to specify which part of the sample the dataset belongs to (e.g. H_T bin, s-/t-channel, ...).