Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

EGM - PFPhoton Selection / Particle Based Isolation / New Soft-Electron ID / New Converted Brem Finder / ECAL Regressions for PF-SCs+Electrons+Photons #1801

Merged
merged 14 commits into from
Dec 18, 2013

Conversation

lgray
Copy link
Contributor

@lgray lgray commented Dec 13, 2013

built on top of CMSSW_7_0_0_pre10

This is the last major update to the EGM software for 700 with the aim of bringing it to an 'analysis ready' state for the HPA exercise in 2014.

You will need to manually download the new MVA weight file, for the classifiers, in the following packages:
RecoParticleFlow/PFTracking
RecoEgamma/ElectronIdentification

with the usual cat download.url | xargs wget

Changes can occur in:

  • PFBlocks, due to the brem finder and changes to photon importing requirements
  • electron efficiency and fake rates, from the new electron ID and corresponding preselection of tracker-driven electrons
  • particle based isolation for gedGsfElectrons and gedPhotons (these are a value map in the event, we'll post the plots for you)
  • Improved energy resolution for Mustache-SCs and for gedGsfElectrons & gedPhotons p4, from the regressions.
    • There can also be changes to the energy resolution for electrons from the new converted brem finder.

The PR for setting the GED RECO to default will come on another PR, based on top of this one.

There will be a follow up after this PR is merged for the GT update in #1835.
This will just be removing the ESPrefer statements for the PF-SCs, electrons and photons.

@cmsbuild
Copy link
Contributor

A new Pull Request was created by @lgray (Lindsey Gray) for CMSSW_7_0_X.

EGM - PFPhoton Selection / Particle Based Isolation / New Soft-Electron ID / New Converted Brem Finder / ECAL Regressions for PF-SCs+Electrons+Photons

It involves the following packages:

Validation/RecoEgamma
RecoEgamma/EgammaElectronProducers
RecoEgamma/ElectronIdentification
RecoEgamma/EgammaPhotonAlgos
DataFormats/EgammaCandidates
RecoEgamma/Configuration
RecoParticleFlow/PFTracking
RecoEgamma/EgammaIsolationAlgos
RecoEcal/EgammaClusterProducers
RecoEgamma/EgammaPhotonProducers
RecoEgamma/EgammaElectronAlgos
RecoParticleFlow/PFProducer
RecoParticleFlow/PFClusterProducer
RecoEgamma/PhotonIdentification
RecoParticleFlow/Configuration
RecoEgamma/EgammaTools

@nclopezo, @danduggan, @rovere, @cmsbuild, @thspeer, @deguio, @slava77, @eliasron can you please review it and eventually sign? Thanks.
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.
@ktf you are the release manager for this.

@dbenedet
Copy link
Contributor

isolation_sumpt_zee
isolation_sumpt_hgg
isolation_dr_zee
isolation_dr_hgg

dbenedet and others added 9 commits December 13, 2013 15:26
(collapsed 5 commits to one for brevity)
… GT, so temporary)

Add hooks to get gedphoton regressions from database in python.
Additionally, implement some comments from previous code review.
Update particle based isolation: remove the pair from the map.
Stick in just a vector of PFCandidates- only the candidates which needs to be excluded from the isolation sums, make the cone size configurable and fix the value to infinite cone
port of PF-ECAL SCs to new regression calculator
add DB source and ES prefer for gedelectron ECAL regressions, fix a crash
working on regression common regression calculation framework
@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

@cmsbuild
Copy link
Contributor

-1
When I ran the RelVals I found an error in the following worklfows:
5.1 step1

runTheMatrix-results/5.1_TTbar+TTbarFS+HARVESTFS/step1_TTbar+TTbarFS+HARVESTFS.log
----- Begin Fatal Exception 14-Dec-2013 20:46:40 CET-----------------------
An exception of category 'BadAlloc' occurred while
   [0] Constructing the EventProcessor
Exception Message:
A std::bad_alloc exception was thrown.
The job has probably exhausted the virtual memory available to the process.
----- End Fatal Exception -------------------------------------------------

1001.0 step2

runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD/step2_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD.log
----- Begin Fatal Exception 14-Dec-2013 20:47:04 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD
----- End Fatal Exception -------------------------------------------------

1000.0 step2

runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT/step2_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT.log
----- Begin Fatal Exception 14-Dec-2013 20:47:10 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT
----- End Fatal Exception -------------------------------------------------

401.0 step1

runTheMatrix-results/401.0_TTbarNewMix+TTbarFSPU2+HARVESTFS/step1_TTbarNewMix+TTbarFSPU2+HARVESTFS.log

1306.0 step3

runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15/step3_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15.log
----- Begin Fatal Exception 14-Dec-2013 20:50:05 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15
----- End Fatal Exception -------------------------------------------------

4.53 step3

runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT/step3_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT.log
----- Begin Fatal Exception 14-Dec-2013 20:50:24 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT
----- End Fatal Exception -------------------------------------------------

25.0 step3

runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT/step3_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT.log
----- Begin Fatal Exception 14-Dec-2013 20:58:21 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT
----- End Fatal Exception -------------------------------------------------

you can see the results of the tests here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-1801/1764/summary.html

@dbenedet
Copy link
Contributor

Until the new MVA classifier files are not attached to release (see
comment from Lindsey and Slava) they need to be downloaded manually:
RecoParticleFlow/PFTracking/data
RecoEgamma/ElectronIdentification/data
with the usual cat download.url | xargs wget

On 14/12/13 20:58, cmsbuild wrote:

-1
When I ran the RelVals I found an error in the following worklfows:
5.1 step1

runTheMatrix-results/5.1_TTbar+TTbarFS+HARVESTFS/step1_TTbar+TTbarFS+HARVESTFS.log

----- Begin Fatal Exception 14-Dec-2013 20:46:40 CET-----------------------

An exception of category 'BadAlloc' occurred while

[0] Constructing the EventProcessor

Exception Message:

A std::bad_alloc exception was thrown.

The job has probably exhausted the virtual memory available to the process.

----- End Fatal Exception -------------------------------------------------

1001.0 step2

runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD/step2_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD.log

----- Begin Fatal Exception 14-Dec-2013 20:47:04 CET-----------------------

An exception of category 'FileInPathError' occurred while

[0] Constructing the EventProcessor

[1] Constructing module: class=PFElecTkProducer label='pfTrackElec'

Exception Message:

edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.

The search path is defined by: CMSSW_SEARCH_PATH

${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data

Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD

----- End Fatal Exception -------------------------------------------------

1000.0 step2

runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT/step2_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT.log

----- Begin Fatal Exception 14-Dec-2013 20:47:10 CET-----------------------

An exception of category 'FileInPathError' occurred while

[0] Constructing the EventProcessor

[1] Constructing module: class=PFElecTkProducer label='pfTrackElec'

Exception Message:

edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.

The search path is defined by: CMSSW_SEARCH_PATH

${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data

Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT

----- End Fatal Exception -------------------------------------------------

401.0 step1

runTheMatrix-results/401.0_TTbarNewMix+TTbarFSPU2+HARVESTFS/step1_TTbarNewMix+TTbarFSPU2+HARVESTFS.log

1306.0 step3

runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15/step3_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15.log

----- Begin Fatal Exception 14-Dec-2013 20:50:05 CET-----------------------

An exception of category 'FileInPathError' occurred while

[0] Constructing the EventProcessor

[1] Constructing module: class=PFElecTkProducer label='pfTrackElec'

Exception Message:

edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.

The search path is defined by: CMSSW_SEARCH_PATH

${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data

Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15

----- End Fatal Exception -------------------------------------------------

4.53 step3

runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT/step3_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT.log

----- Begin Fatal Exception 14-Dec-2013 20:50:24 CET-----------------------

An exception of category 'FileInPathError' occurred while

[0] Constructing the EventProcessor

[1] Constructing module: class=PFElecTkProducer label='pfTrackElec'

Exception Message:

edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.

The search path is defined by: CMSSW_SEARCH_PATH

${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data

Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT

----- End Fatal Exception -------------------------------------------------

25.0 step3

runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT/step3_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT.log

----- Begin Fatal Exception 14-Dec-2013 20:58:21 CET-----------------------

An exception of category 'FileInPathError' occurred while

[0] Constructing the EventProcessor

[1] Constructing module: class=PFElecTkProducer label='pfTrackElec'

Exception Message:

edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.

The search path is defined by: CMSSW_SEARCH_PATH

${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol0/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-14-0200/external/slc5_amd64_gcc481/data

Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT

----- End Fatal Exception -------------------------------------------------

you can see the results of the tests here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-1801/1764/summary.html


Reply to this email directly or view it on GitHub
#1801 (comment).

@lgray
Copy link
Contributor Author

lgray commented Dec 16, 2013

@sdevissc @beaudett I've integrated your changes. Florian, I modified yours slightly to use the weights from the database directly. Please check that I haven't clobbered anything subtle.

@slava77 At this point this PR is closed for new additions. Please start testing.

@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @anton-a, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

ecalWeightsFromDB = cms.bool(True),
# if not from DB. Otherwise, keep empty
ecalRefinedRegressionWeightFiles = cms.vstring(),
combinationWeightsFromDB = cms.bool(True),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi Lindsey,

is that the change you mentioned (together with the label for the combination DB label) ? In that case, you were perfectly right to do it. Thanks

Florian

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi Florian,

Yes it is, and the changes at the top of the file to get the weights from the DB (until they are in the GT).

-Lindsey

@cmsbuild
Copy link
Contributor

-1
When I ran the RelVals I found an error in the following worklfows:
1000.0 step2

runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT/step2_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT.log
----- Begin Fatal Exception 16-Dec-2013 14:07:34 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1000.0_RunMinBias2011A+RunMinBias2011A+TIER0+SKIMD+HARVESTDfst2+ALCASPLIT
----- End Fatal Exception -------------------------------------------------

5.1 step1

runTheMatrix-results/5.1_TTbar+TTbarFS+HARVESTFS/step1_TTbar+TTbarFS+HARVESTFS.log
----- Begin Fatal Exception 16-Dec-2013 14:07:34 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/5.1_TTbar+TTbarFS+HARVESTFS
----- End Fatal Exception -------------------------------------------------

401.0 step1

runTheMatrix-results/401.0_TTbarNewMix+TTbarFSPU2+HARVESTFS/step1_TTbarNewMix+TTbarFSPU2+HARVESTFS.log
----- Begin Fatal Exception 16-Dec-2013 14:07:37 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/401.0_TTbarNewMix+TTbarFSPU2+HARVESTFS
----- End Fatal Exception -------------------------------------------------

1001.0 step2

runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD/step2_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD.log
----- Begin Fatal Exception 16-Dec-2013 14:07:39 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1001.0_RunMinBias2011A+RunMinBias2011A+TIER0EXP+ALCAEXP+ALCAHARVD
----- End Fatal Exception -------------------------------------------------

1003.0 step2

runTheMatrix-results/1003.0_RunMinBias2012A+RunMinBias2012A+RECODDQM+HARVESTDDQM/step2_RunMinBias2012A+RunMinBias2012A+RECODDQM+HARVESTDDQM.log
----- Begin Fatal Exception 16-Dec-2013 14:07:48 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1003.0_RunMinBias2012A+RunMinBias2012A+RECODDQM+HARVESTDDQM
----- End Fatal Exception -------------------------------------------------

1306.0 step3

runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15/step3_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15.log
----- Begin Fatal Exception 16-Dec-2013 14:10:38 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/1306.0_SingleMuPt1_UP15+SingleMuPt1_UP15+DIGIUP15+RECOUP15+HARVESTUP15
----- End Fatal Exception -------------------------------------------------

4.53 step3

runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT/step3_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT.log
----- Begin Fatal Exception 16-Dec-2013 14:10:53 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/4.53_RunPhoton2012B+RunPhoton2012B+HLTD+RECODreHLT+HARVESTDreHLT
----- End Fatal Exception -------------------------------------------------

25.0 step3

runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT/step3_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT.log
----- Begin Fatal Exception 16-Dec-2013 14:18:40 CET-----------------------
An exception of category 'FileInPathError' occurred while
   [0] Constructing the EventProcessor
   [1] Constructing module: class=PFElecTkProducer label='pfTrackElec'
Exception Message:
edm::FileInPath unable to find file RecoParticleFlow/PFTracking/data/TMVAClassification_ConvBremFinder_Testetlt20absetalt1_479_BDT.weights.xml anywhere in the search path.
The search path is defined by: CMSSW_SEARCH_PATH
${CMSSW_SEARCH_PATH} is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/src:/build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/src:/afs/cern.ch/cms/sw/ReleaseCandidates/vol1/slc5_amd64_gcc481/cms/cmssw/CMSSW_7_0_X_2013-12-16-0200/external/slc5_amd64_gcc481/data
Current directory is: /build/cmsbuild/jenkins-workarea/workspace/Pull-Request-Integration/ARCHITECTURE/slc5_amd64_gcc481/runTheMatrix-results/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT
----- End Fatal Exception -------------------------------------------------

you can see the results of the tests here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-1801/1786/summary.html

@lgray
Copy link
Contributor Author

lgray commented Dec 16, 2013

@davidlt Can you please test this one manually, it passes all matrix tests for me. There are missing files from CMSDIST that have to be downloaded by hand.

@nclopezo
Copy link
Contributor

@lgray

Hi Linsey,

I am afraid you were mentioning the wrong David, I will run them by hand.

@lgray
Copy link
Contributor Author

lgray commented Dec 16, 2013

@davidlt sorry!
@nclopezo Thanks!

@deguio
Copy link
Contributor

deguio commented Dec 16, 2013

+1
after downloading the weights by hand.

@@ -59,7 +61,7 @@
'keep recoConversions_conversions_*_*',
'keep recoConversions_mustacheConversions_*_*',
'drop *_conversions_uncleanedConversions_*',
'drop *_gedPhotons_valMapPFEgammaCandToPhoton_*',
'drop *_gedPhotonsTmp_valMapPFEgammaCandToPhoton_*',
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

if the gedPhotonsTmp are not kept (commented out a few lines above), what's the reason to keep gedPhotonsTmp_valMapPFEgammaCandToPhoton ?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@nancymarinelli Please comment?

@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @anton-a, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

ktf added a commit to ktf/cmsdist that referenced this pull request Dec 17, 2013
ktf added a commit to ktf/cmsdist that referenced this pull request Dec 17, 2013
@nclopezo
Copy link
Contributor

+1
I ran by hand the short matrix including the workflows 401.0,1306.0,101.0,1003.0 and all passed.

@slava77
Copy link
Contributor

slava77 commented Dec 17, 2013

This is before the last fix:
testing #1801 91ec9d6
in CMSSW_7_0_X_2013-12-16-0200

regular mode (non-GED)
The largest visible changes begin in the particleFlowSuperClusterECAL, then in the ged electrons and photons.
Of all the changes, it seems to me that gedGsfElectron trackMomentumError is wrong. The rest of the diffs are probably reasonable (see ~random picks below)

In 1000.0 (minbias 2011)
all_sign281vsorig_runminbias2011awf1000p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy

gamma 10
all_sign281vsorig_singlegammapt10wf18p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy

The electron energy error change looks quite significant: is it expected? (the old one has ~the same shape as the regular gsfElectrons)
all_sign281vsorig_singlegammapt10wf18p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror

The GED photon et is now above the generated 10GeV (is this the "around threshold" effect?)
all_sign281vsorig_singlegammapt10wf18p0c_recophotons_gedphotons__reco_obj_et

Gamma 35:

  • electron error degrades as well, except for the very large values
    all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror
  • is track momentum error filled in correctly? (all entries at low/non-default values are gone)
    all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror
  • ged photon et looks more centered on 35 . Good.
    all_sign281vsorig_singlegammapt35wf19p0c_recophotons_gedphotons__reco_obj_et

Electron 35 sample

  • ged electron energy error changes are more reasonable now
    all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_ecalenergyerror
  • track momentum error looks bad as well
    all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror

QCD 15-3000 sample

  • (more visible here) low end of the supercluster ecal changes to higher values. This is just the regression, right?
    all_sign281vsorig_qcdflatpt15s3000wf38p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy
  • the number of GED electrons goes down by quite a bit. This is preselection, OK.
    all_sign281vsorig_qcdflatpt15s3000wf38p0c_recogsfelectrons_gedgsfelectrons__reco_obj_classification

@lgray
Copy link
Contributor Author

lgray commented Dec 17, 2013

@beaudett Looks like we may have some issue with the calculation of the various uncertainty estimator MVAs at low pT. Could you please have a look at this?

@beaudett
Copy link
Contributor

Hello,

what is "reference", "new version" and "new reference" in these plots ?

Thanks
Florian

On 17/12/13 1:56 PM, slava77 wrote:

This is before the last fix:
testing #1801 #1801 91ec9d6
91ec9d6
in CMSSW_7_0_X_2013-12-16-0200

regular mode (non-GED)
The largest visible changes begin in the particleFlowSuperClusterECAL,
then in the ged electrons and photons.
Of all the changes, it seems to me that gedGsfElectron
trackMomentumError is wrong. The rest of the diffs are probably
reasonable (see ~random picks below)

In 1000.0 (minbias 2011)
all_sign281vsorig_runminbias2011awf1000p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy
https://f.cloud.github.com/assets/4676718/1763707/6753af3e-6715-11e3-9376-cec81b478212.png

gamma 10
all_sign281vsorig_singlegammapt10wf18p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy
https://f.cloud.github.com/assets/4676718/1763716/a49c88d4-6715-11e3-8c9d-bcb8734da911.png

The electron energy error change looks quite significant: is it
expected? (the old one has ~the same shape as the regular gsfElectrons)
all_sign281vsorig_singlegammapt10wf18p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror
https://f.cloud.github.com/assets/4676718/1763724/d3a71aae-6715-11e3-8d88-c96e22d43842.png

The GED photon et is now above the generated 10GeV (is this the "around
threshold" effect?)
all_sign281vsorig_singlegammapt10wf18p0c_recophotons_gedphotons__reco_obj_et
https://f.cloud.github.com/assets/4676718/1763759/6849759e-6716-11e3-9e75-b5ac1105584b.png

Gamma 35:

electron error degrades as well, except for the very large values
all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror
<https://f.cloud.github.com/assets/4676718/1763777/d6b4a9ea-6716-11e3-83c9-7b5eb7d37e43.png>
is track momentum error filled in correctly? (all entries at
low/non-default values are gone)
all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror
<https://f.cloud.github.com/assets/4676718/1763796/48d0b92e-6717-11e3-91df-2d40e945604c.png>
ged photon et looks more centered on 35 . Good.
all_sign281vsorig_singlegammapt35wf19p0c_recophotons_gedphotons__reco_obj_et
<https://f.cloud.github.com/assets/4676718/1763809/ab2f3e56-6717-11e3-9bb8-becd020147d7.png>

Electron 35 sample

ged electron energy error changes are more reasonable now
all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_ecalenergyerror
<https://f.cloud.github.com/assets/4676718/1763817/ee532ed6-6717-11e3-99c8-d60a4b663a53.png>
track momentum error looks bad as well
all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror
<https://f.cloud.github.com/assets/4676718/1763822/1809fd36-6718-11e3-8f30-ca208ece421f.png>

QCD 15-3000 sample

(more visible here) low end of the supercluster ecal changes to
higher values. This is just the regression, right?
all_sign281vsorig_qcdflatpt15s3000wf38p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy
<https://f.cloud.github.com/assets/4676718/1763868/abc0650a-6719-11e3-8cff-7661a9413b39.png>
the number of GED electrons goes down by quite a bit. This is
preselection, OK.
all_sign281vsorig_qcdflatpt15s3000wf38p0c_recogsfelectrons_gedgsfelectrons__reco_obj_classification
<https://f.cloud.github.com/assets/4676718/1763874/c237b6c6-6719-11e3-8154-ef08bc069a7a.png>


Reply to this email directly or view it on GitHub
#1801 (comment).

@beaudett
Copy link
Contributor

Ah, it is new-minus-reference Ok. Got it.
F.

On 17/12/13 2:09 PM, Florian Beaudette wrote:

Hello,

what is "reference", "new version" and "new reference" in these plots ?

Thanks
Florian

On 17/12/13 1:56 PM, slava77 wrote:

This is before the last fix:
testing #1801 #1801 91ec9d6
91ec9d6
in CMSSW_7_0_X_2013-12-16-0200

regular mode (non-GED)
The largest visible changes begin in the particleFlowSuperClusterECAL,
then in the ged electrons and photons.
Of all the changes, it seems to me that gedGsfElectron
trackMomentumError is wrong. The rest of the diffs are probably
reasonable (see ~random picks below)

In 1000.0 (minbias 2011)
all_sign281vsorig_runminbias2011awf1000p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy

https://f.cloud.github.com/assets/4676718/1763707/6753af3e-6715-11e3-9376-cec81b478212.png

gamma 10
all_sign281vsorig_singlegammapt10wf18p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy

https://f.cloud.github.com/assets/4676718/1763716/a49c88d4-6715-11e3-8c9d-bcb8734da911.png

The electron energy error change looks quite significant: is it
expected? (the old one has ~the same shape as the regular gsfElectrons)
all_sign281vsorig_singlegammapt10wf18p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror

https://f.cloud.github.com/assets/4676718/1763724/d3a71aae-6715-11e3-8d88-c96e22d43842.png

The GED photon et is now above the generated 10GeV (is this the "around
threshold" effect?)
all_sign281vsorig_singlegammapt10wf18p0c_recophotons_gedphotons__reco_obj_et

https://f.cloud.github.com/assets/4676718/1763759/6849759e-6716-11e3-9e75-b5ac1105584b.png

Gamma 35:

electron error degrades as well, except for the very large values

all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_correctedecalenergyerror

https://f.cloud.github.com/assets/4676718/1763777/d6b4a9ea-6716-11e3-83c9-7b5eb7d37e43.png

is track momentum error filled in correctly? (all entries at
low/non-default values are gone)

all_sign281vsorig_singlegammapt35wf19p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror

https://f.cloud.github.com/assets/4676718/1763796/48d0b92e-6717-11e3-91df-2d40e945604c.png

ged photon et looks more centered on 35 . Good.

all_sign281vsorig_singlegammapt35wf19p0c_recophotons_gedphotons__reco_obj_et

https://f.cloud.github.com/assets/4676718/1763809/ab2f3e56-6717-11e3-9bb8-becd020147d7.png

Electron 35 sample

ged electron energy error changes are more reasonable now

all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_ecalenergyerror

https://f.cloud.github.com/assets/4676718/1763817/ee532ed6-6717-11e3-99c8-d60a4b663a53.png

track momentum error looks bad as well

all_sign281vsorig_singleelectron35wf17p0c_recogsfelectrons_gedgsfelectrons__reco_obj_trackmomentumerror

https://f.cloud.github.com/assets/4676718/1763822/1809fd36-6718-11e3-8f30-ca208ece421f.png

QCD 15-3000 sample

(more visible here) low end of the supercluster ecal changes to
higher values. This is just the regression, right?

all_sign281vsorig_qcdflatpt15s3000wf38p0c_log10recosuperclusters_particleflowsuperclusterecal_particleflowsuperclusterecalbarrel_reco_obj_energy

https://f.cloud.github.com/assets/4676718/1763868/abc0650a-6719-11e3-8cff-7661a9413b39.png

the number of GED electrons goes down by quite a bit. This is
preselection, OK.

all_sign281vsorig_qcdflatpt15s3000wf38p0c_recogsfelectrons_gedgsfelectrons__reco_obj_classification

https://f.cloud.github.com/assets/4676718/1763874/c237b6c6-6719-11e3-8154-ef08bc069a7a.png


Reply to this email directly or view it on GitHub
#1801 (comment).

@slava77
Copy link
Contributor

slava77 commented Dec 17, 2013

reference is CMSSW_7_0_X_2013-12-16-0200
New version is #1801 91ec9d6
merged on top of CMSSW_7_0_X_2013-12-16-0200

Here are the ged photon energy errors
gamma 10
all_sign281vsorig_singlegammapt10wf18p0c_recophotons_gedphotons__reco_obj_energycorrections_regression1energyerror

gamma35
all_sign281vsorig_singlegammapt35wf19p0c_recophotons_gedphotons__reco_obj_energycorrections_regression1energyerror

@slava77
Copy link
Contributor

slava77 commented Dec 17, 2013

GED mode outputs

The rate of the following errors in TTBarPU (202.0) went up by ~x2 (to 45 in 200 events)
%MSG-e PFEGammaAlgo: PFEGammaProducer:particleFlowEGamma 16-Dec-2013 21:07:15 CET Run: 1 Event: 521
Encountered the known GSF-SC splitting bug in PFBlockAlgo! We should really fix this!

electron 35: response is much better
screen shot 2013-12-17 at 3 18 26 pm

QCD pt 15 to 3000

  • MET seems OK

Other changes in GED mode are in line with changes in GED electrons and photons mentioned above.

I note that the pf candidates don't changes significantly in yields, which means the preselections made are not cutting into the pf candidate selections (as expected).

@slava77
Copy link
Contributor

slava77 commented Dec 17, 2013

So, is the trackMomentumError was forgotten just in the output electrons, or did it affect other intermediate/derived values (momentum combination etc)?

@beaudett
Copy link
Contributor

Hi,

the computation of this quantity is done in a place that I didn't expect and which was no longer "called" in GED mode. More precisely, it is done in a helper class while I would have expected it to be done in the main algo.
It seems that the regression-based combination was smart enough for not using the track information in that case. So, I expect the regression combination to be improved after the fix; we'll know for sure in the coming hours.

F.

@deguio
Copy link
Contributor

deguio commented Dec 17, 2013

my understanding is that this pull request is superseded by #1834. Am I correct?

@lgray
Copy link
Contributor Author

lgray commented Dec 17, 2013

Not really, it is more that 1834 is based on this and needs pieces from this one to function properly.
Slava wanted to factor the default reconstruction flip into a different PR.

@beaudett
Copy link
Contributor

Hi,

so I have put back the computation of the momentum error. As expected, it has an impact on the regression.
Below reco-PT/35 for single electrons of 35 GeV before (top) and after(bottom). The number of events processed is different between the two cases.
before
after
Lindsey is already integrating the fix in his branch.
Thanks Slava for catching this.

F.

@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @anton-a, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

1 similar comment
@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @anton-a, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

@lgray
Copy link
Contributor Author

lgray commented Dec 17, 2013

@slava77 concerning the shape differences. The new energy error estimates are based on a BDT where the old are based off of some parameterizations. I am not entirely sure they're directly comparable.

@cmsbuild
Copy link
Contributor

Pull request #1801 was updated. @nclopezo, @danduggan, @rovere, @cmsbuild, @anton-a, @thspeer, @deguio, @slava77, @eliasron can you please check and sign again.

@slava77
Copy link
Contributor

slava77 commented Dec 17, 2013

+1

tested 5b92d89 in CMSSW_7_0_X_2013-12-16-0200

Compared to the previous iteration: the momentumError is fixed, even though it doesn't introduce a very significant effect on the combined momentum itself.

This should be merged assuming cms-sw/cmsdist#299 is included in the IB
(the weight/classification files)

@deguio
Copy link
Contributor

deguio commented Dec 17, 2013

+1

@cmsbuild
Copy link
Contributor

This pull request is fully signed and it will be integrated in one of the next IBs unless changes or unless it breaks tests. @ktf can you please take care of it?

ktf added a commit that referenced this pull request Dec 18, 2013
RecoUpdates -- EGM - PFPhoton Selection / Particle Based Isolation / New Soft-Electron ID / New Converted Brem Finder / ECAL Regressions for PF-SCs+Electrons+Photons
@ktf ktf merged commit f50bb9e into cms-sw:CMSSW_7_0_X Dec 18, 2013
@ktf
Copy link
Contributor

ktf commented Dec 18, 2013

Skip tests.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.