Raw iso pt sum rework

[andrewj314]

Rebase of reworking of raw iso pt sum calculation to 8_0_X. No changes to reco quantities are anticipated - this is a simple fix that calculates the pt sum for isolation once and stores it so that it may be used for each iso WP, rather than have each iso WP calculate the pt sum separately.

[cmsbuild]

A new Pull Request was created by @andrewj314 (AJ Johnson) for CMSSW_8_0_X.

Raw iso pt sum rework

It involves the following packages:

CommonTools/ParticleFlow
RecoTauTag/Configuration
RecoTauTag/RecoTau

@cmsbuild, @cvuosalo, @vadler, @monttj, @slava77 can you please review it and eventually sign? Thanks.
@ahinzmann, @jdolen, @rappoccio 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' or '@cmsbuild, 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.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/9464/console

[cmsbuild]

+1
Tested at: 058df4f
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-12261/9464/summary.html

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-12261/9464/summary.html

[cvuosalo]

@andrewj314: Please add to the PR description the changes to reco quantities that are expected from this PR.

[andrewj314]

@cvuosalo Updated! See the first comment - we do not expect changes to reco quantities

[cvuosalo]

@andrewj314: Your description makes it sound like this PR will not change much. However, if you look at the Jenkins DQM plots, there are thousands of changes in tau isolation quantities, including efficiencies. This large number may be partly due to there being an unnecessarily large number of DQM plots for taus, but still it is hard to reconcile your description of no changes with the DQM plots showing that seemingly every quantity with tau isolation in the name is showing changes.

[andrewj314]

@cvuosalo This is indeed surprising - all that's happening is that the pt sum is calculated once, rather than for each WP. The pt sum should be the same for each WP, so I really don't know why this is changing physics results. I will look into it.
@veelken is there any reason we'd expect to see this simple rework change tau isolation quantities?

[andrewj314]

@cvuosalo Could you please point me to these DQM plots? I couldn't find them in the cmsbuild link. Thanks!

[cvuosalo]

@andrewj314: The DQM plots are here in alternative-comparisons:

https://cmssdt.cern.ch/SDT/jenkins-artifacts/baseLineComparisons/CMSSW_8_0_X_2015-11-04-1100+12261/11137/alternative-comparisons/

25202.0 and 50202.0 have especially large numbers of differences.

[cvuosalo]

Here are some examples of differences found in a test of workflow 25202.0_TTbar_13 with 70 events against baseline CMSSW_8_0_X_2015-11-03-2300.

[cvuosalo]

@andrewj314: When might this PR get updated?

[andrewj314]

@cvuosalo There are substantial relvals that need to be done in order to find out what's causing the large discrepancy in measured quantities. Given that the tau pog's priority is to get PR #12298 added to 76x in the next 7-10 days, we are going to focus on achieving that and then push this PR to 80x once we fix the errors.

[cvuosalo]

@andrewj314: When will work on this PR resume?

[andrewj314]

@cvuosalo Now that the new discriminators have been added to 76x (PR #12586) (or at least now that it's wrapping up) I will resume work on this. I will continue to try to get to the bottom of why it's changing measured quantities so severely.

[cvuosalo]

@andrewj314: Please give an update about the status of this PR.

[andrewj314]

On recommendation of the tau pog conveners, I'm closing this PR until we can get PR #12586 ported to 80x