Reimplement ATLAS Z0 7TEV 46FB Dataset by ecole41 · Pull Request #2237 · NNPDF/nnpdf

ecole41 · 2024-12-04T13:29:47Z

This pull request is merging the two CC and CF datasets into one filter.py script.
Functions haves been added to produce the data central, kinematic and uncertainties yaml files for both of these datasets.

Old vs New Data

CC Dataset: https://vp.nnpdf.science/SpiOHitcS-2WggfpKcLW3Q==/

CF Dataset: https://vp.nnpdf.science/D3121rJ6RWG3IgyqViXJMw==/

Compatibility Check:

from validphys.api import API
import numpy as np
 
inp1 = {"dataset_input": {"dataset": "ATLASZRAP11CC"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
inp2 = {"dataset_input": {"dataset": "ATLASZRAP11CC", "variant": "legacy"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
 
covmat1 = API.covmat_from_systematics(**inp1)
covmat2 = API.covmat_from_systematics(**inp2)
 
t0_covmat1 = API.t0_covmat_from_systematics(**inp1)
t0_covmat2 = API.t0_covmat_from_systematics(**inp2)
 
result = np.all(np.isclose(covmat1, covmat2))
result_2 = np.all(np.isclose(t0_covmat1, t0_covmat2))

print('covmat', result)
print('t0_covmat', result_2)

inp3 = {"dataset_input": {"dataset": "ATLASWZRAP11CF"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
inp4 = {"dataset_input": {"dataset": "ATLASWZRAP11CF", "variant": "legacy"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
 
covmat3 = API.covmat_from_systematics(**inp3)
covmat4 = API.covmat_from_systematics(**inp4)
 
t0_covmat3 = API.t0_covmat_from_systematics(**inp3)
t0_covmat4 = API.t0_covmat_from_systematics(**inp4)
 
result3 = np.all(np.isclose(covmat3, covmat4))
result_4 = np.all(np.isclose(t0_covmat3, t0_covmat4))

print('covmat', result3)
print('t0_covmat', result_4)

[Out]:

covmat True
t0_covmat True
covmat True
t0_covmat True

scarlehoff · 2024-12-04T15:26:34Z

Hi @ecole41 this is not ready for rewview yet right? I see that the variables are still called k1/k2/k_i etc for instance.

ecole41 · 2024-12-04T15:31:58Z

Hi @ecole41 this is not ready for rewview yet right? I see that the variables are still called k1/k2/k_i etc for instance.

No, this is not ready yet. I will keep working on it

scarlehoff · 2024-12-04T15:33:39Z

Ok! No problem! Is any of the PR finished? There are now many and I got a bit lost. So that I can review (and hopefully merge) the ones that are

ecole41 · 2024-12-05T09:57:30Z

Ok! No problem! Is any of the PR finished? There are now many and I got a bit lost. So that I can review (and hopefully merge) the ones that are

Yes, #2178 and #2202 should be ready unless there is something that I have missed.

…ot sure on process type

…0_7TEV_46FB_CC_AND_CF

ecole41 · 2024-12-11T11:29:05Z

This branch should be complete, but is failing the checks. I have merged from the master so am not sure why this is happening

jacoterh · 2024-12-18T14:08:09Z

This branch should be complete, but is failing the checks. I have merged from the master so am not sure why this is happening

Indeed, it's really odd - I don't see what's wrong with your metadata. The yaml file cannot be parsed for some reason. I'm looking into it.

jacoterh · 2024-12-18T14:29:54Z

      data_uncertainties:
      - uncertainties_legacy_CC-Y.yaml
-  data_central: data_legacy_CC-Y.yaml
+      data_central:


I noticed that if you remove data_central from variants: legacy the tests pass again

Ok, I will just check if changing the formatting of this helps

The tests pass now, excellent!

…0_7TEV_46FB_CC_AND_CF

jacoterh

Looks like this is good to go after the minor comments I left are implemented. Thanks!

scarlehoff · 2025-01-15T13:27:21Z

This can be merged, right? (I see it is approved, but just checking!)

scarlehoff

Just the comment on the labels, for both datasets (like in #2223 I see the check was done with the old names though, here's a link with the check with the new names: https://vp.nnpdf.science/eS7CgNDEQMmflyqxTbBmdQ==

scarlehoff · 2025-01-21T10:35:32Z

-        label: k1
+      abs_eta:
+        description: Absolute dilepton rapidity
+        label: abs_eta


Eventually this will be the label to be used in plots, so it would be better to use |\eta| or m_{ll}^2 (below) etc.

Ok, I have changed the format in the metadata. Is this format correct?

Yes, that works. Since it is rapidity, like in the other PR it might make sense to change the labels to "y" already.

Also, for sqrts, I think you are missing the \ in sqrt but these are minor points.

Please have a look at the conflict and then this can be merged I believe.

ecole41 · 2025-01-21T14:52:21Z

-        label: k1
+      abs_eta:
+        description: Absolute dilepton rapidity
+        label: "|y|"


@scarlehoff Just wanted to check if this is now the correct labelling.

I think in the one below (m_ll) you need to also put $m_{ll}^2$ ($) for matplotlib to render it correctly. Not sure about the |. But other than that, yes, looks good.

In the xlabel though you might want to remove (in both here and the other observable) the \eta part so that it is only the y.

Ok, I have just changed this

ecole41 · 2025-01-21T15:09:26Z

After updating the compatibility test to use the new dataset name, I get that the matrices are compatible within 0.1% except the covariance matrix for the CF set which is compatible within 1%, so there are some changes in the new implementation :

from validphys.api import API
import numpy as np
 
inp1 = {"dataset_input": {"dataset": "ATLAS_Z0_7TEV_46FB_CC-Y"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
inp2 = {"dataset_input": {"dataset": "ATLASZRAP11CC", "variant": "legacy"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
 
covmat1 = API.covmat_from_systematics(**inp1)
covmat2 = API.covmat_from_systematics(**inp2)
 
t0_covmat1 = API.t0_covmat_from_systematics(**inp1)
t0_covmat2 = API.t0_covmat_from_systematics(**inp2)

result = np.all(np.isclose(covmat1, covmat2,rtol=1e-3))
result_2 = np.all(np.isclose(t0_covmat1, t0_covmat2,rtol=1e-3))

print('covmat', result)
print('t0_covmat', result_2)

inp3 = {"dataset_input": {"dataset": "ATLAS_Z0_7TEV_46FB_CF-Y"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
inp4 = {"dataset_input": {"dataset": "ATLASWZRAP11CF", "variant": "legacy"}, "theoryid": 40_000_000, "use_cuts": "internal", "t0pdfset": "NNPDF40_nnlo_as_01180", "use_t0": True}
 
covmat3 = API.covmat_from_systematics(**inp3)
covmat4 = API.covmat_from_systematics(**inp4)
 
t0_covmat3 = API.t0_covmat_from_systematics(**inp3)
t0_covmat4 = API.t0_covmat_from_systematics(**inp4)
 
result3 = np.all(np.isclose(covmat3, covmat4,rtol=1e-3))
result3_2 = np.all(np.isclose(t0_covmat3, t0_covmat4,rtol=1e-2))
result_4 = np.all(np.isclose(t0_covmat3, t0_covmat4,rtol=1e-3))

print('covmat 0.001', result3)
print('covmat 0.01', result3_2)
print('t0_covmat', result_4)

[OUT]:
covmat True
t0_covmat True
covmat 0.001 False
covmat 0.01 True
t0_covmat True

Here are the updated old vs new reports:

scarlehoff · 2025-01-21T17:26:36Z

Thanks. If you solve the conflict I think this can be merged.

ecole41 added 2 commits December 3, 2024 12:28

Added raw data and edited metadata

d27acde

Added function in filter.py to poduce data.yaml files

6e92dec

ecole41 added data toolchain ATLAS_DY_DATA labels Dec 4, 2024

ecole41 requested review from comane and scarlehoff December 4, 2024 13:29

ecole41 added 3 commits December 6, 2024 10:41

Added function to produce kinematics.yaml file and edited metadata, n…

b9a07fa

…ot sure on process type

Added uncertainties matrices in csv format

d4cda69

Added function in filter to produce uncertainties yaml files

a346ed6

scarlehoff mentioned this pull request Dec 6, 2024

Final revision of the 4.0 dataset #2242

Closed

5 tasks

ecole41 added 2 commits December 9, 2024 09:49

Added m_ll2 as a variable in process, removed legacy from metadata

7273e27

Merge branch 'master' into reimplement_ATLAS_Z0_7TEV_46FB_CC_AND_CF

36e4dbc

ecole41 added the regenerate-data label Dec 9, 2024

ecole41 added 2 commits December 11, 2024 11:20

Added legacy back

3d176e1

Merge remote-tracking branch 'origin/master' into reimplement_ATLAS_Z…

95cc376

…0_7TEV_46FB_CC_AND_CF

RoyStegeman requested review from jacoterh and removed request for jacoterh December 11, 2024 15:31

Removed comment from metadata

851eb44

jacoterh reviewed Dec 18, 2024

View reviewed changes

ecole41 added 2 commits December 18, 2024 14:33

changed data legacy in metadata

2602844

Merge remote-tracking branch 'origin/master' into reimplement_ATLAS_Z…

f2079cd

…0_7TEV_46FB_CC_AND_CF

ecole41 changed the title ~~[WIP] Reimplement ATLAS Z0 7TEV 46FB Dataset~~ Reimplement ATLAS Z0 7TEV 46FB Dataset Dec 18, 2024

jacoterh reviewed Dec 18, 2024

View reviewed changes

Comment thread nnpdf_data/nnpdf_data/commondata/ATLAS_Z0_7TEV_46FB/metadata.yaml Outdated

jacoterh reviewed Dec 18, 2024

View reviewed changes

Comment thread nnpdf_data/nnpdf_data/commondata/ATLAS_Z0_7TEV_46FB/metadata.yaml Outdated

jacoterh reviewed Dec 18, 2024

View reviewed changes

Comment thread nnpdf_data/nnpdf_data/commondata/ATLAS_Z0_7TEV_46FB/metadata.yaml Outdated

jacoterh approved these changes Dec 18, 2024

View reviewed changes

ecole41 added 2 commits December 18, 2024 15:50

Removed comment from metadat

4e6ace0

Added inspire link

0a9c7f9

scarlehoff approved these changes Jan 21, 2025

View reviewed changes

ecole41 added 4 commits January 21, 2025 10:44

changed version to 1 and edited labels

a380771

Merge branch 'master' into reimplement_ATLAS_Z0_7TEV_46FB_CC_AND_CF

7b29e0e

changed format of labels

60a0b0b

changed kin labels in metadata

43c4bb0

ecole41 commented Jan 21, 2025

View reviewed changes

label edit

bd98ca3

merged

1a73d86

scarlehoff merged commit bf1a67a into master Jan 22, 2025

scarlehoff deleted the reimplement_ATLAS_Z0_7TEV_46FB_CC_AND_CF branch January 22, 2025 04:15

Conversation

ecole41 commented Dec 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Old vs New Data

Compatibility Check:

Uh oh!

scarlehoff commented Dec 4, 2024

Uh oh!

ecole41 commented Dec 4, 2024

Uh oh!

scarlehoff commented Dec 4, 2024

Uh oh!

ecole41 commented Dec 5, 2024

Uh oh!

ecole41 commented Dec 11, 2024

Uh oh!

jacoterh commented Dec 18, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jacoterh left a comment

Choose a reason for hiding this comment

Uh oh!

scarlehoff commented Jan 15, 2025

Uh oh!

scarlehoff left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ecole41 commented Jan 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

scarlehoff commented Jan 21, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ecole41 commented Dec 4, 2024 •

edited

Loading

jacoterh commented Dec 18, 2024 •

edited

Loading

scarlehoff left a comment •

edited

Loading

ecole41 commented Jan 21, 2025 •

edited

Loading