Handle weights for nonresonant samples

[OlivierBondu]

There is a root file available with weights at this link. We would need to store / apply these weights both because there is some slight out-of-sync in between the central CMS samples and their latest paper draft.

For each of the 12 anomalous nonresonant samples we would have to store (12 weights for the new benchmarks points) x (1 nominal + 6 'outliers' to study acceptance variations within a cluster). Then for plots / histograms limits we would:

• merge all the anomalous nonresonant samples together
• apply the weight corresponding to the scenario we want to study

[OlivierBondu]

done in https://github.com/cp3-llbb/HHTools/pull/30 as a more appropriate place