Implement corr smearings #54
Description
The EGM POG recommends to correlate s&s between the eras: https://egammapog.docs.cern.ch/Run3/SaS/#systematics-correlation
Currently, this is not done in our configs as the s&s are treated as scales and smears, which are by construction in the finalfits code considered uncorrelated. This cannot be changed at the datacard step since the nuisance correlation scheme for signal shape nuisances is determined at the signal step.
For scales, it is straightforward: Just change it to scalesCorr globally in the configs. For smears, smearsCorr needs to be implemented. Most important entry point here: https://github.com/JaLuka98/flashggFinalFit/blob/higgsdnafinalfit/Signal/tools/finalModel.py#L204-L224, probably there is also a change needed in the argparsing of the helper script that is executed.
FYI @AlessandroTarabini @TheRealLoliges486 @maxwrabetz and @Linuserdm (relevant for comb with H4l).
Furthermore, @AlessandroTarabini should confirm that it is fine to keep the separate scale nuisances for EB and EE. In general, careful evaluation of the nuisance correlation scheme should be performed given that a single scale nuisance across all eras and years might lead to an inflexible statistical model. For example, in Run 2 and especially the mass measurements, the s&s was split into more parts like lowR9/highR9 and a gain dependence was also present.