Rabbit tensorwriter for direct theory fit of sigmaUL

[lucalavezzo]

Relies on #671

The previous script, feedRabbitTheory.py, was creating tensors for rabbit for sigmaUL, as well as for W and the Ai's. The latter two -- W and Ai -- were duplicate efforts to what setupRabbit.py already does, since it's in fiducial phase space, they have to be read from MiNNLO. The sigmaUL, on the other hand, is in full lepton phase space, so it can be compared directly to the generators' predictions and variations.

The new structure is to have a single feedRabbitSigmaUL.py, and all it does is create the rabbit tensor for sigmaUL directly from the TheoryCorrection files.

The sigmaUL tensor from feedRabbitSigmaUL.py and the W,Ai tensors from setupRabbit.py can be combined, as different cards, if one wants to do a simultaneous fit.

[kdlong]

I still don't understand what the input of the script is: is it the rabbit tensor (e.g., ZMassDilepton.hdf5) or is it the corrections file from SCETlib+DYTurbo (wremnants-data/data/TheoryCorrections/*.pkl.lz4)

[lucalavezzo]

I still don't understand what the input of the script is: is it the rabbit tensor (e.g., ZMassDilepton.hdf5) or is it the corrections file from SCETlib+DYTurbo (wremnants-data/data/TheoryCorrections/*.pkl.lz4)

Hi @kdlong the inputs are:

1. unfodled sigmaUL
2. TheoryCorrections

the output is a tensor for rabbit to the generator-level extraction.

[davidwalter2]

The feedRabbitTheory script looks very good, I have a few comments to make it more integrated with the rest of the framework.