Extend th covmat plots

[giacomomagni]

This PR is to extend the N3LO th covmat plots @andreab1997

[andreab1997]

@giacomomagni I would say that in this PR we could also aadd the pull plots as done in #1832 and then close #1832 . Do you agree?

[andreab1997]

Given that in 04665ab I moved the pull plots functions that were in #1832, I would say we can now close it.

[giacomomagni]

@andreab1997 I added a similar function, to allow for multiple prescriptions in the same diag plot and keep the grouping by process.

See for instance:
plot_diag_cov_comparison_by_process.pdf

if this is okay we can drop the old function, otherwise we can keep both.

[andreab1997]

@andreab1997 I added a similar function, to allow for multiple prescriptions in the same diag plot and keep the grouping by process.

See for instance: plot_diag_cov_comparison_by_process.pdf

if this is okay we can drop the old function, otherwise we can keep both.

For me it is even fine but I am worried that in this way it would not be possible anymore to reproduce the plots that are currently in the MHOU paper

[andreab1997]

btw @giacomomagni I guess you need to merge master to fix the tests

[scarlehoff]

If you can rebase instead of merge it would be better, that way it goes for sure on top of the last commit.

[scarlehoff]

@andreab1997 (since I guess @giacomomagni is not around this week?) do you know whether this final and should be merged? Could you have a look?

[andreab1997]

@andreab1997 (since I guess @giacomomagni is not around this week?) do you know whether this final and should be merged? Could you have a look?

This is final (it is just missing a review). I also wanted to add regression tests but if it is urgent, we can merge this and I will add regression tests later on.

[scarlehoff]

Nono, it is not. Please review it (and add regression tests if you can of course!)

[andreab1997]

@andreab1997 please rebase this

Ok did it, sorry for the delay but I was travelling. Now I can also add some more meaningfull regression tests so wait until tomorrow to start your review please (I should be able to do everything I want today).

[RoyStegeman]

Thanks! Somehow in the web interface I can still see the old commits, but when I check the commits locally it seems fine so I guess it's alright.

P.S. @achiefa is planning to have a first go at reviewing this

[andreab1997]

Ok I added what I wanted to add, so please go on with the review

[RoyStegeman]

Is there a report that produces plots using this PR?

[giacomomagni]

for the diagonal plots you can try:

Diagonal elements of covariance matrices
========================================

{@theoryid@}
------------------
   {@plot_diag_cov_comparison_by_experiment@}
   {@plot_diag_cov_comparison_by_process@}

and this runcard:

meta:
   author: Giacomo Magni
   keywords: [theory uncertainties, n3lo ad var, newth, covmat_plots]
   title: Comparison of (th)covariance matrices with and without n3lo anomalous dimension uncertanties 
   
normalize_to: 1

q2min: 3.49
w2min: 12.5

theoryid: 722
fit: 210715-n3fit-1000-001
pdf:
  from_: fit
use_cuts: internal

dataspecs:

  - theoryid: 722
    speclabel: "Anom. Dim. only IHOU" # "$\\gamma_{IHOU}$"
    point_prescription: 'n3lo ad covmat'
    plot_total: False
    theoryids:
      from_: scale_variation_theories


  - theoryid: 722
    speclabel: "Total IHOU"
    point_prescription: 'n3lo ihou'
    plot_total: False
    theoryids:
      from_: scale_variation_theories



dataset_inputs:
      - {dataset: NMCPD_dw_ite}
      # - {dataset: NMC}
      # - {dataset: SLACP_dwsh}
      # - {dataset: SLACD_dw_ite}
      # - {dataset: BCDMSP_dwsh}
      # - {dataset: BCDMSD_dw_ite}
      # - {dataset: CHORUSNUPb_dw_ite, }
      # - {dataset: CHORUSNBPb_dw_ite, }
      # - {dataset: NTVNUDMNFe_dw_ite,  cfac: [MAS]}
      # - {dataset: NTVNBDMNFe_dw_ite,  cfac: [MAS]}
      # - {dataset: HERACOMBNCEM}
      # - {dataset: HERACOMBNCEP460}
      # - {dataset: HERACOMBNCEP575}
      # - {dataset: HERACOMBNCEP820}
      # - {dataset: HERACOMBNCEP920}
      - {dataset: HERACOMBCCEM, }
      # - {dataset: HERACOMBCCEP, }
      # - {dataset: HERACOMB_SIGMARED_B}
      # - {dataset: HERACOMB_SIGMARED_C}

template: template_diag_plots_plots.md

dataset_report:
   meta: Null

actions_:
  - report(main=true)

[RoyStegeman]

Thanks

[RoyStegeman]

Thanks, this looks good. Just some minor comments

[andreab1997]

If noone complains in the next hour I will merge this