Dynamic strip reco 74 x #10943

andrewj314 · 2015-08-25T20:45:13Z

Rebase of PR #10805 to 7_4_X

- added Loose, Medium and Tight tau ID discriminators with pileup weighted isolation - removal of deprecated tau ID discriminators from pat::Taus

- bug-fix: disable deltaBeta corrections for pileup weighted isolation discriminators

…ended by Yuta

pf charged hadrons and tracks -> solves rare problem with 3-prongs having charge +/-3.

…riminators with deltaBeta correction

- updated thresholds for isolation WPs: Loose = 2.5 GeV Medium = 1.5 GeV Tight = 0.8 GeV (no change)

…e was causing unit tests to fail.

…ry because of a rebase while debugging Previous PR at: cms-sw#10215

…on since they may not be modified. Now failing unit test 1330.0 as well as 135.4. Attempting to diagnose but updating the PR anyway to iterate with the experts.

…sts except 135.4 (missing module label: combinatoricRecoTaus)

…hem to import section instead. Now runTheMatrix passes all tests!!!

…605. Code compiles but segfaults on several matrix tests - currently attempting to diagnose with GDB but sharing with the experts anyway.

… to auto iterators

cmsbuild · 2015-08-27T10:01:41Z

Pull request #10943 was updated. @cmsbuild, @cvuosalo, @vadler, @monttj, @slava77 can you please check and sign again.

slava77 · 2015-08-27T12:20:03Z

Hi Christian,

On 8/27/15 3:26 AM, Christian Veelken wrote:

Hi Slava,

I am not sure I understand the comment that you make after "However" above:
the use case that we have in mind is to rerun the tau reconstruction
(PFTau sequence defined in
RecoTauTag/Configuration/python/RecoPFTauTag_cff.py) right at the
beginning of the miniAOD production sequence, i.e. using AOD inputs.

miniAOD configuration should be aware of it.
Maybe it's a one-liner, but it still should be added someplace.
Giovanni should make a suggestion where (since I don't think that all
tau reco should be rerun
in front of miniAOD by default).

Can you be a bit more explicite concerning where you see a problem/what
still needs to be done ?

I agree with you that PR #8095
#8095 should be included in this
branch.
AJ, can you take care of that, please ?

Thank you,

Christian

—
Reply to this email directly or view it on GitHub
#10943 (comment).

slava77 · 2015-08-27T12:28:16Z

@cmsbuild please test

cmsbuild · 2015-08-27T12:31:51Z

The tests are being triggered in jenkins.

gpetruc · 2015-08-27T12:59:57Z

I'd make it a customize for running on pre-existing AODs.
It could either be called remakeTausFromAOD or be a generic runOn74Xv1AOD
in case we need it.
I'll try code it later today both for 74X and for 75X (in case we do the
re-MiniAOD in 75X)

Giovanni

On Thu, Aug 27, 2015 at 2:31 PM, cmsbuild notifications@github.com wrote:

The tests are being triggered in jenkins.

—
Reply to this email directly or view it on GitHub
#10943 (comment).

veelken · 2015-08-27T13:16:27Z

Hi Giovanni,

please rerun the complete PFTau sequence defined in RecoTauTag/Configuration/python/RecoPFTauTag_cff.py at the beginning of miniAOD production. The dynamic strip reconstruction changes the way how the reco::PFTau objects are built, which is one of the first steps in the tau reconstruction. It is not enough to rerun part of the tau reconstruction sequence only.

Thank you,

Christian

cmsbuild · 2015-08-27T15:04:51Z

+1
Tested at: a95733f
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-10943/7692/summary.html

cmsbuild · 2015-08-27T16:03:08Z

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-10943/7692/summary.html

gpetruc · 2015-08-27T19:59:08Z

This would be my proposal: cms-tau-pog#10
Testing it now.

On Thu, Aug 27, 2015 at 6:03 PM, cmsbuild notifications@github.com wrote:

Comparison is ready

https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-10943/7692/summary.html

—
Reply to this email directly or view it on GitHub
#10943 (comment).

slava77 · 2015-08-27T21:57:08Z

+1

for #10943 a95733f

this correctly picks up the feature from added cut that each tau has at least one pixel hits (for at least one of... #8095
jenkins tests pass and show differences only in tau distributions
tested locally in CMSSW_7_4_10_patch1 /test area sign573/ comparing the previous commit results with the latest one (09467ef to a95733f): changes are small, as expected from tests of added cut that each tau has at least one pixel hits (for at least one of... #8095

gpetruc · 2015-08-28T12:34:48Z

I tested on some MC and Data, including this PR. Timing increases by a bit,
as expected, but within reason: I see about +10% in data when also
rerunning the taus ( before the cleanup of
gpetruc#5 )

On Thu, Aug 27, 2015 at 11:57 PM, Slava Krutelyov notifications@github.com
wrote:

+1

for #10943 #10943 a95733f
a95733f

this correctly picks up the feature from added cut that each tau has at least one pixel hits (for at least one of... #8095
added cut that each tau has at least one pixel hits (for at least one of... #8095

jenkins tests pass and show differences only in tau distributions

tested locally in CMSSW_7_4_10_patch1 /test area sign573/ comparing
the previous commit results with the latest one (09467ef
09467ef
to a95733f
a95733f):
changes are small, as expected from tests of added cut that each tau has at least one pixel hits (for at least one of... #8095
added cut that each tau has at least one pixel hits (for at least one of... #8095

—
Reply to this email directly or view it on GitHub
#10943 (comment).

monttj · 2015-08-30T23:13:50Z

+1
This was already approved in 76X.

cmsbuild · 2015-08-30T23:16:12Z

This pull request is fully signed and it will be integrated in one of the next CMSSW_7_4_X IBs once checked with relvals in the development release cycle of CMSSW (tests are also fine). This pull request requires discussion in the ORP meeting before it's merged. @davidlange6, @Degano, @smuzaffar

slava77 · 2015-08-31T21:09:33Z

@davidlange6 this one is in #10649 now
and can be closed

Christian and others added 24 commits August 25, 2015 22:25

- dynamic strip reconstruction for Yuta

87ebec5

- added Loose, Medium and Tight tau ID discriminators with pileup weighted isolation - removal of deprecated tau ID discriminators from pat::Taus

fixed typo in discriminator name

4742d79

bug-fix in storing of (raw) footprint correction

7c658b7

- store neutral pileup weighted isolation pT sum

6f5d602

- bug-fix: disable deltaBeta corrections for pileup weighted isolation discriminators

added pileup weighted neutral isolation

e54a3a6

fixed typo in discriminator name

3204063

updated parameters for dynamic strip reconstruction

de9331a

sutract 2 GeV from k offset values (except for 3prong0pi0), as recomm…

8e114c2

…ended by Yuta

New cleaner module that works also if tau is made up of combination of

b92676f

pf charged hadrons and tracks -> solves rare problem with 3-prongs having charge +/-3.

added functionality to make tau footprint correction dependent on tauPt

6a0927a

tightened cut on pTouter from 0.20*tauPt to 0.10*tauPt

6c0970e

apply cut on ptOuter also in case of HPS combined islation 3-hit disc…

7ce036f

…riminators with deltaBeta correction

- updated deltaBeta correction factor to 0.2

7ed86f7

- updated thresholds for isolation WPs: Loose = 2.5 GeV Medium = 1.5 GeV Tight = 0.8 GeV (no change)

updates to tau vertex producers from Arun

51106a1

updated patTau selection to use new discriminators

0f9d847

Added missing discriminators to updateHPSPFTaus_cff.py - their absenc…

2b9df30

…e was causing unit tests to fail.

Added import statements whose absence was causing the unit tests to fail

068059b

New attempt to add Dynamic Strip Reco to 7_6_X - a new PR was necessa…

332ef14

…ry because of a rebase while debugging Previous PR at: cms-sw#10215

Added fillDescriptions() method, reverted HLT dumps to previous versi…

d57288b

…on since they may not be modified. Now failing unit test 1330.0 as well as 135.4. Attempting to diagnose but updating the PR anyway to iterate with the experts.

Added Christian's alternate fix to the HLT issues. Now passing all te…

7c03241

…sts except 135.4 (missing module label: combinatoricRecoTaus)

Per Slava's suggestion, removed modules from cms.Sequence and added t…

1f90487

…hem to import section instead. Now runTheMatrix passes all tests!!!

Attempting to implement Slava's suggestions after his review of PR 10…

f298382

…605. Code compiles but segfaults on several matrix tests - currently attempting to diagnose with GDB but sharing with the experts anyway.

Applied Slava's suggestions in PR 10605+10802, now passing all tests!

79a5b3a

Bugfix adding absolute value to deltaPhi, changing BOOSTFOREACH loops…

09467ef

… to auto iterators

cmsbuild added this to the Next CMSSW_7_4_X milestone Aug 25, 2015

cmsbuild added reconstruction-pending analysis-pending pending-signatures tests-pending orp-pending labels Aug 25, 2015

cmsbuild added the comparison-pending label Aug 27, 2015

cmsbuild added tests-started and removed tests-pending labels Aug 27, 2015

cmsbuild added tests-approved and removed tests-started labels Aug 27, 2015

cmsbuild added comparison-available and removed comparison-pending labels Aug 27, 2015

cmsbuild added reconstruction-approved and removed reconstruction-pending labels Aug 27, 2015

gpetruc mentioned this pull request Aug 28, 2015

MiniAOD Spring15pass2 7.4.X (collection of all PRs) #10649

Merged

cmsbuild added analysis-approved fully-signed and removed analysis-pending pending-signatures labels Aug 30, 2015

davidlange6 closed this Sep 1, 2015

roger-wolf deleted the DynamicStripReco_74X branch March 24, 2016 22:15

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dynamic strip reco 74 x #10943

Dynamic strip reco 74 x #10943

andrewj314 commented Aug 25, 2015

cmsbuild commented Aug 27, 2015

slava77 commented Aug 27, 2015

slava77 commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

gpetruc commented Aug 27, 2015

veelken commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

gpetruc commented Aug 27, 2015

slava77 commented Aug 27, 2015

gpetruc commented Aug 28, 2015

monttj commented Aug 30, 2015

cmsbuild commented Aug 30, 2015

slava77 commented Aug 31, 2015

Dynamic strip reco 74 x #10943

Dynamic strip reco 74 x #10943

Conversation

andrewj314 commented Aug 25, 2015

cmsbuild commented Aug 27, 2015

slava77 commented Aug 27, 2015

slava77 commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

gpetruc commented Aug 27, 2015

veelken commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

cmsbuild commented Aug 27, 2015

gpetruc commented Aug 27, 2015

slava77 commented Aug 27, 2015

gpetruc commented Aug 28, 2015

monttj commented Aug 30, 2015

cmsbuild commented Aug 30, 2015

slava77 commented Aug 31, 2015