Marcel Rieger

We have various (mostly cms-specific) task array functions that use correctionlib: - calibrators: jec, jer, met_phi - selectors: jet_veto_map - producers: sfs for btag, electron, muon, weights for pu, scale,...

Refactoring for 0.3 release

1

## Preface This PR refactors a portion of our code base. Initially, two developments happened in parallel: multi config tasks (for plotting, datacards, ml trainings) and a cleanup of the...

Documentation outdated

1

In the scope of the refactoring happening in #628 I overhauled some parts of the documentation, but mostly what was directly concerned with the updates therein. I noticed that a...

In a producer/selector/..., when declaring dependent columns via ```python @producer( uses={"some_column"}, ) def ... ``` the input file is checked for the presence of `some_column`. However, there are cases in...

Distinguish ML model parameters for training and evaluation

1

As discussed in our previous meeting, in our ML model interface we should distinguish between parameters used for training and evaluation. - Allows for different configurations being used in the...

We discussed several necessary updates to our plots that I want to document here: - **Rebinning option not considered in y-axis label bin width** - We should not show the...