Beta decays

[FrancescoMerlotti]

PR dedicated to update of the routines in SIMUnet to handle BETA_DECAYS dataset.

Note: never put ? in the experiment name in PLOTTING file.

Related to Issue #87

[ElieHammou]

Looks very reasonable @FrancescoMerlotti ! The fixes on covmats.py and n3fit_data.py come from another bug about the cuts right?

Did you have a chance to test it @ecole41 ? If it fixes the errors you had, I am happy to sign off on it.

[FrancescoMerlotti]

I run the fit with a simplified runcard, and all went just fine. @ecole41 and I found an error in the original SIMU file, and the numbers seem not to be sound to me.

[ElieHammou]

Ah ok, but that's likely a completely separated issue

Did you also add the stat uncertainties in the SYSTYPE file to have it working?

[FrancescoMerlotti]

The cuts fix is there because this branch is based on the fixed_prediction_pdf

Ah ok, but that's likely a completely separated issue

Did you also add the stat uncertainties in the SYSTYPE file to have it working?

The stat uncertainty were already there and sound, the issue was related to the absence of systematics which had to be cured introducing a np.zeros array when ndata is 0

[ElieHammou]

It looks all great to me then :)

[FrancescoMerlotti]

Do you think it is clever to leave the branch alive for further checks and fixes?

[ElieHammou]

Sure thing
I do not thing this will impact in any way the observables with systematics we already have, hence my approval, but
I leave the merging at your discretion in any case.
You are that it would be good to have reasonable physics results for the new beta decays datasets before releasing the feature.

[ElieHammou]

The current status is that we first need to fix the SIMU files before proceeding further, right?

[FrancescoMerlotti]

the simu file can be fixed independetly on the branch

[ElieHammou]

Yes but we need them fixed to run some reasonable fit and analysis if we want to make sure that this branch works as expected, right?

[FrancescoMerlotti]

right