[DF] Implement weighted Filters

[eguiraud]

Serious analyses involve running on MC samples which are tipically assigned a weight.
If Filters could be informed of event weights, Reports could take that into account (as requested at https://root-forum.cern.ch/t/rdataframe-feature-request-weighted-cut-reports/53570).

Alternative we could add a mechanism to keep track of event weights separately, in a dedicated "virtual" column (that by default contains all ones, e.g., as discussed with G. Petrucciani from CMS a couple of times:

df.UpdateEventWeight("w1")
  .UpdateEventWeight("w2")
  .Filter("x > 0")
  .Histo1D("x", "rdfeventweight_")