KLUB cleaning about klubanalysis HOT 9 CLOSED

This is where actually I wanted to comment. Can we add to the removal list the old DY scale factors?

from klubanalysis.

Hi Bruno!
These variables were used to train the mutlticlassificator developed by Marcel to improve the sensitivity in the VBF category.
I do not think that we will us it in the resonant since we do not need extra-sensitivity to this production mode.
If so then we can remove the variables and branches.

from klubanalysis.

The first and third bullet points have been implemented in #315.
@portalesHEP Let me know when the second is finished, so we can add this PR to the list of PRs to merge :)

from klubanalysis.

Thank you Bruno! I really agree on the second bullet!

from klubanalysis.

I'm not sure the second point is so relevant in fact; looking a bit closer, it seems that only few of the lines can be removed from the SF file (size goes from ~3.8MB to ~3.5MB removing everything I found that is not used in KLUB). This could be done anyway, but the speedup should be marginal

from klubanalysis.

Ma noooooo, that's too bad! Well I think we can leave them as they are then :)

from klubanalysis.

Can we add to the removal list the old DY scale factors?

We absolutely should!

from klubanalysis.

Added to the list, and will be applied to #315 asap.

from klubanalysis.

I've now removed all DY-related scale factors and associated variables in #315.

from klubanalysis.