Flexible sample database location, column definitions, and Run 3-specific weights and selections

[moritzmolch]

This pull request adds some new convenience features regarding the sample database location and the definition of new columns, as well as some hard-coded selections and weights for Run 3 analyses.

Sample database configuration

The path to the sample database can be provided in the common settings configuration file with the key sample_database. Providing this parameter is optional. By default, the path points to the datasets sublodule in the TauFakeFactors package.

Column definitions

New columns can be added from the preselection configuration, using the following syntax:

column_definitions:
  nbtag:
    expression: id_wgt_bjet_pnet_shape
    exclude_processes:  # optional, 
        - data

This definition will translate into the python code:

rdf.Define("nbtag", "id_wgt_bjet_pnet_shape")

If the process that a processed sample is assigned to appears in the optional list exclude_processes, the Define declaration is skipped.

Hard-coded selections and weights for Run 3 analyses

Some hard-coded selections and weights are added:

• Flavor filters for DYto2L* and DYto2Tau* samples: Only the Z → ee, μμ components are selected from DYto2L* samples, while the Z → ττ component is taken from the DYto2Tau* samples.
• Luminosity weights for the 2022 and 2023 eras

[moritzmolch]

Tested, preselection step works with both the usual nmssm Run 3 and the nmssm_boosted config.

[nshadskiy]

Looks fine to me.