Cleanup duplicate pileup jet ID configurations

[ahinzmann]

A cleanup of the PU jet ID configs which are duplicate as a followup of #10398.

The configs under 1. are kept while those under 2. and 3. are dropped and their content is integrated into 1.:

This one is from 2013 and kept in sync with MVA MET in 2013:
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/PileupJetID_cfi.py
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/PileupJetIDParams_cfi.py
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/PileupJetIDCutParams_cfi.py
It is used in MET.

This one was introduced in 2014 and kept in sync with MET in 2014:
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/pileupjetidproducer_cfi.py
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/puJetIDParams_cfi.py
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/RecoJets/JetProducers/python/puJetIDAlgo_cff.py
It is used in MET and DQM.

This one was introduced in 2014 for MiniAOD and is yet another copy:
https://github.com/jbrands/cmssw/blob/newOfflinePuID_new/PhysicsTools/PatAlgos/python/slimming/pileupJetId_cfi.py

[cmsbuild]

A new Pull Request was created by @ahinzmann for CMSSW_7_6_X.

Cleanup duplicate pileup jet ID configurations

It involves the following packages:

DQM/PhysicsHWW
DQMOffline/JetMET
PhysicsTools/PatAlgos
RecoJets/JetProducers
RecoMET/METPUSubtraction

@cvuosalo, @monttj, @cmsbuild, @deguio, @slava77, @vadler, @danduggan can you please review it and eventually sign? Thanks.
@rappoccio, @imarches, @yslai, @acaudron, @mmarionncern, @jdolen, @nhanvtran, @schoef, @ferencek, @gpetruc, @mariadalfonso, @pvmulder, @TaiSakuma, @rociovilar this is something you requested to watch as well.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.
If you are a L2 or a release manager you can ask for tests by saying 'please test' in the first line of a comment.
@Degano you are the release manager for this.
You can merge this pull request by typing 'merge' in the first line of your comment.

[slava77]

@cmsbuild please test

[cmsbuild]

The tests are being triggered in jenkins.

[cmsbuild]

+1
Tested at: a83e421
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-10625/7057/summary.html

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-10625/7057/summary.html

[cvuosalo]

+1

For #10625 a83e421

Clean-up of duplicate pile-up jet ID configurations. There should be no change in monitored quantities.

The code changes are satisfactory, and Jenkins tests against baseline CMSSW_7_6_X_2015-08-07-1100 show no significant differences, as expected. A check of the expanded RECO config for workflow 25202.0 confirms that the code changes do appear in the config.

[gpetruc]

When this PR is ready, can you prepare the backports for this + #10398?
Thanks,
Giovanni

[ahinzmann]

FYI: There is currently a bug in this new offline jet ID code. Only once this is fixed and validated, we can make the backport.

[ahinzmann]

Bug fixed in #10730. Validation ongoing.

[deguio]

+1

[slava77]

@monttj please check this one
Thank you

[monttj]

+1

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next CMSSW_7_6_X IBs (tests are also fine). This pull request requires discussion in the ORP meeting before it's merged. @davidlange6, @Degano, @smuzaffar

[gpetruc]

backports?

[ahinzmann]

While software-wise everything is validated, the training weights are still being validated and until this is done, backport to 74/75 is on hold.
Still need to check

1. this training is optimal for ak4 (for mvaMET) and ak4CHS (for jets) my comparing two separate trainings.
2. this training performs well for mvaMET.

[gpetruc]

Is what we currently have in 74X good for physics?
If that one isn't good anyway, then I'd rather have this one if it has a
chance of being good: at worst, if the validation fails, we will have just
replaced one random number with another random number, but if instead the
validation is successful people will be able to use it.
This is especially true for the more basic scenario of people using it on
the ak4CHS jets: mvaMET anyway is currently not in MiniAOD and so the few
people who use need to run more complex recipes and they can recompute a
new PuJetID if needed, while the same computation would add some burden to
the much larger number of users of the slimmedJets.

On Mon, Aug 24, 2015 at 2:58 PM, ahinzmann notifications@github.com wrote:

While software-wise everything is validated, the training weights are
still being validated and until this is done, backport to 74/75 is on hold.
Still need to check

1. this training is optimal for ak4 (for mvaMET) and ak4CHS (for jets) my
   comparing two separate trainings.
2. this training performs well for mvaMET.

—
Reply to this email directly or view it on GitHub
#10625 (comment).

[ahinzmann]

ok, if mvaMET is not in MiniAOD, including this one would make sense.
@jbrands : Since you started validating this one yesterday, could you post the new ROC curve comparison of 53 vs 74 here? (based on the current weights, not the new ones which are still in the works)

[ahinzmann]

Johannes confirmed that this discriminant performs better than the 53X training in 76X (though not reaching the Run I performance yet).

Starting backport to 74 and 75.