Remove cross sections from the UI

cms-PdmV / cmsPdmV

CERN CMS McM repository

4 stars 10 forks source link

Remove cross sections from the UI #1126

Closed lmoureaux closed 5 months ago

lmoureaux commented 5 months ago

The cross sections shown in the McM UI were confusing users into thinking that they were correct, which is almost never the case. Remove them from the UI (but keep them in the database for now).

@ggonzr could you deploy this to the dev machine for testing?

ggonzr commented 5 months ago

Hi @lmoureaux, deployed in dev. By the way, is there any other source we can link so that users can see this attribute properly? Maybe the XSDB app from CMS GEN?

lmoureaux commented 5 months ago

Thanks for the deployment! It appears to work as intended: cross sections are no longer visible on the Request page (in the "Generator parameters" column) nor on the Request edit page.

is there any other source we can link so that users can see this attribute properly? Maybe the XSDB app from CMS GEN?

XSDB could be a link but there are TWiki pages that should have priority for analyzers. XSDB is less wrong than McM but still often far from perfect.

ggonzr commented 5 months ago

Okay, thanks for the info. Just one little question/request, is there any central page in the TWiki that groups/provides this information related to cross-section for the analyzers? If so, please provide the link as a comment.

From my side, I don’t have any other comments. I think @sunilUIET is the right person to approve this so I am changing the reviewer's list

Best regards Geovanny

lmoureaux commented 5 months ago

The cross sections are found here and in linked pages for most standard model processes (13TeV version, not sure about 13.6):

https://twiki.cern.ch/twiki/bin/viewauth/CMS/StandardModelCrossSectionsat13TeV

And here for Higgs:

https://twiki.cern.ch/twiki/bin/view/LHCPhysics/HiggsXSBR

sunilUIET commented 5 months ago

If I understand correctly, the cross-section value and efficiency parameters are set as per the fragment. If so what is the issue having these available?

lmoureaux commented 5 months ago

In most cases the fragment doesn't set any cross section. The values are filled in manually and there is no guarantee that they are correct. There is also no information about how they are derived (the validation script does for sure not run the GenXSecAnalyzer; even then it would be low stat).

In tutorials I've seen analysts use values from McM because they didn't know the recommended ways. I want to prevent this and I think the best solution is to just remove the least trustworthy source of information.

sunilUIET commented 5 months ago

Thanks Louis, it makes sense