Complete removal of ECAL PF cleaning from reco and HLT, and verify its effect on data

[perrotta]

With PR #18438 (now merged in 91X) PF cluster cleaning in ECAL clusters got removed from the Reco sequences.

This followed the discussion started in #17859 (and continued in the reco meeting of Apr 21) where it was realized that modifications to the PF spike cleaning originally meant for HF could have affected ECAL PF cleaning as well. In the discussion, it was stated that PF cleaning is not actually needed for ECAL, because spike cleaning is already performed at RecHit level: the decision to remove it from the reco and hlt sequences followed.

ECAL cleaning is also applied at HLT, and therefore another pull request with the removal at HLT is also needed: @Sam-Harper @Martin-Grunewald, @silviodonato : any plan for it? The HF spike cleaning pull request will be integrated as soon as that HLT pull request will also be available.

Furthermore, when looking at the effect of #18438 it was realized (see #18438 (comment)) that it is indeed tiny, but not completely transparent, even for MC samples. I was not able to test on a data sample with actual spikes in ECAL, in order to see which is the effect on them. At some point one should verify that really such a PF cleaning was useless for ECAL spike cleaning (I would say, especially for HLT): @Sam-Harper @rafaellopesdesa could you please verify by running with and without it on some appropriate data sample, and report here the outcome?

Attention: @slava77
Attention: @knash @emanueledimarco @jainshilpi @fwyzard

[cmsbuild]

A new Issue was created by @perrotta .

@davidlange6, @Dr15Jones, @smuzaffar can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[Sam-Harper]

I'll file a JIRA today for the HLT part. As for testing for data, I have no time to do this. Again, I consider the existence of this bug and I'm frankly annoyed that PF guys felt that they had to reinvent the wheel. I agree there is a possibility that it is hiding and cancelling out another bug.

I hate to buck pass, but this is a really a PF problem (its in the generic PFClusters) and PF should test and deal with it. They should justify why they were re-implimenting spike cleaning instead of using the ECAL flags.

[Sam-Harper]

Btw when I remove this, I'm removing this for all paths. It'll also effect JetMET as well. Just so people are aware this is a global problem with PFClustering that effects everything.

[Dr15Jones]

assign reconstruction

[cmsbuild]

New categories assigned: reconstruction

@slava77,@perrotta you have been requested to review this Pull request/Issue and eventually sign? Thanks

[rafaellopesdesa]

@perrotta Thanks. I will try to check the effect of this change in EGM in the coming days (probably not in the next 2 or 3, but after that I should have time... I apologize for not being able to do immediately).

[perrotta]

Thank you @rafaellopesdesa : the time schedule you propose is fine. I just would like to have it documented, and we can wait a few days for it.

[Sam-Harper]

JIRA filed: https://its.cern.ch/jira/browse/CMSHLT-1242

[slava77]

@perrotta
I suppose we can "+1" and close this issue now

[perrotta]

+1
JIRA ticket CMSHLT-1242 integrated in the HLT menu since /dev/CMSSW_9_1_0/HLT/V12

[cmsbuild]

This issue is fully signed and ready to be closed.