Revision of CMS inclusive DY data

[achiefa]

This PR addresses the four datasets listed below:

• CMS_WPWM_7TEV_ELECTRON_ASY
• CMS_WPWM_7TEV_MUON_ASY
• CMS_Z0_7TEV_DIMUON_2D
• CMS_WPWM_8TEV_MUON_Y

These datasets were already re-implemented in the new format by TR. I think the things to do here are:

1. Check the filter file is working.
2. Update the process_type with the processes in process_options.
3. Update the name of the kinematics (and re-run the filter)

See the reports for these datasets before I apply any change: report

[achiefa]

And this is the report after these changes: report. Nothing has changed compared to the previous versions. I think this one is ready for merging. Let's wait for the tests.

[achiefa]

@scarlehoff I don't understand that error message that I get from the test

Run echo "Changed files: $CHANGED_FILES"
  echo "Changed files: $CHANGED_FILES"
  exit 1
  shell: /usr/bin/bash -e {0}
  env:
    CHANGED_FILES: nnpdf_data/nnpdf_data/commondata/CMS_WPWM_8TEV_MUON/uncertainties.yaml
Changed files: nnpdf_data/nnpdf_data/commondata/CMS_WPWM_8TEV_MUON/uncertainties.yaml
Error: Process completed with exit code 1.

Can you help me?

[scarlehoff]

It means that you need to use the label again. Basically it means that the bot would get results different than what you commited (@Radonirinaunimi and I were unable to make the -1 signs agree between computers so we decided to take the bot as the ultimate truth)

[achiefa]

(@Radonirinaunimi and I were unable to make the -1 signs agree between computers so we decided to take the bot as the ultimate truth)

How is that possible 🫤?

[scarlehoff]

I think it is the eigen value decomposition of the covmat in either scipy or numpy that is basically non reproducible (you would need to take the values and post process them), but having the bot generating all data was something we wanted to do anyway.

[achiefa]

Ok, I think now this is solved. Wait for the tests, and then I think it can be merged.

[scarlehoff]

Thanks!

Beside the comment about the kinematics_override this is good to go.

The only other thing that might be interesting to change is the Z0 variable from mass squared to just mass, so that the plots have the mass in the title

(right now the value is of the mass due to the kinematics override, but the units are of course mass squared)

Up to you if you want to change it, the information is of course the same, I just feel the non-squared mass is easier to read.

[scarlehoff]

The only thing missing is that the kinematics_override needs to be set to identity in all of the files, because otherwise validphys will sqrt the masses (see e.g, the kinematic coverage plot, the scales are too low)