EgammaPerformance

Introduction

This page describes the egammaPerformance package used to monitor the reconstruction and identification of electrons and photons at Tier0.

People involved:

• Kamal Benslama (Univ. Regina),
• Yao Ming (Univ. Regina)

Contact person: Kamal Benslama

Package Description

The egammaPerformance code is to be executed during the T0 processing. It makes a set of monitoring histograms and saves them into a ROOT file for later examination. The package contains three main classes:

        - electronMonTool - a tool for monitoring electron reconstruction and identification

        - photonMonTool - a tool for monitoring photon reconstruction and identification

        - PhysicsMonTool - a tool for monitoring physics process such as,  Z->ee, J/Psi->ee, Upsilon->ee

electronMonTool

The electronMonTool produces histograms in three classes (ROOT directories):

      - Generic - general electron candidate properties, such as eta,phi,et,shower shape variables, E/P,... etc.
      - BarrelSamplings - electron eta,phi,et in the barrel samplings: presampler, first,second and third LAr layers
      - EndcapSamplings - electron eta,phi,et in the endcap samplings: presampler, first,second and third LAr layers

The histograms in the two directories, BarrelSamplings and EndcapSamplings are not shown on the monitoring display, however, E-gamma
experts can access these two directories for more detailed information. 

Configuration: the electronMonTool can be configured with following python code:

-------------------------------------------------------------------------------------------------------
from egammaPerformance.egammaPerformanceConf import electronMonTool
        elMonTool = electronMonTool(name= "elMonTool",
                                    ElectronContainer       ="ElectronCollection",
                                    EMTrackMatchContainer   ="",
                                    EMShowerContainer      ="",
                                    EMConvertContainer      ="",
                                    Electron_Trigger_Items = ["EF_e10_medium"],
                                    Electron_Selection_Items = ["all","loose","tight"])
    
        ToolSvc+=elMonTool
        monManEgamma.AthenaMonTools += [ "electronMonTool/elMonTool" ]

-------------------------------------------------------------------------------------------------------

where:

ElectronContainer        - is the name of the electron EDM container
EMTrackMatchContainer    - is the name of the EMTrackMatch EDM container, containing details of the reconstructed egamma objects.
                           If the name is set to empty string: "", the relevant container is found automatically by the egamma object
                           (This is important, as the detail containers are different for the egamma objects created during the offline
 reconstruction and Event Filter)
EMShowerContainer        - is the name of the EMShower EDM container. Comments similar as for the EMTrackMatchContainer  
EMConvertContainer       - is the name of the EMConvert EDM container. Comments similar as for the EMTrackMatchContainer  
Electron_Trigger_Items   - is the list of electron trigger items. The histograms will be made separately for each trigger item, using the egamma 
                           objects attached to the EF TriggerElement. Only passed TEs are considered
Electron_Selection_Items - list of custom selections. The histograms will be made separately for each selection item. Only "all" option 
                           implemented so far, where "all" means all the electron candidates are used to fill the histograms 

photonMonTool

The photonMonTool produces histograms in three classes (ROOT directories):

      - Generic - general photon candidate properties, like eta,phi,et,shower shape variables,.. etc.
      - BarrelSamplings - photon eta,phi,et in the barrel samplings: presampler, first, second and third LAr layers
      - EndcapSamplings - photon eta,phi,et in the endcap samplings: presampler, first, second and third LAr layers

Here again, only the histograms in the Generic directory are shown on the monitoring display.

Configuration: the photonMonTool can be configured with following python code:

-------------------------------------------------------------------------------------------------------
from egammaPerformance.egammaPerformanceConf import photonMonTool
        phMonTool = photonMonTool(name= "phMonTool",
                                  PhotonContainer      ="PhotonCollection",
                                  EMShowerContainer      ="",
                                  EMConvertContainer      ="",
                                  Photon_Trigger_Items = "EF_g20",
                                  Photon_Selection_Items = ["all","tight"])
        
        ToolSvc+=phMonTool
        monManEgamma.AthenaMonTools += [ "photonMonTool/phMonTool" ]
-------------------------------------------------------------------------------------------------------

where:

PhotonContainer         - is the name of the photon EDM container
EMShowerContainer       - is the name of the EMShower EDM container; comments similar as for the EMTrackMatchContainer  
EMConvertContainer      - is the name of the EMConvert EDM container; comments similar as for the EMTrackMatchContainer  
Photon_Trigger_Items    - is the list of electron trigger items. The histograms will be made separately for each trigger item,
                          using the egamma objects attached to the EF TriggerElement. Only passed TEs are considered

Photon_Selection_Items  - list of custom selections. The histograms will be made separately for each selection item.
                          Only "all" option implemented so far, where "all" means all the electron candidates are used to fill the histograms 

physicsMonTool

The physicsMonTool produces histograms in two classes (ROOT directories):

      - Kinematics: Et, eta and phi of candidates before and after applying the identification criteria. 
                 Di-electron mass distributions in many eta regions (example: both clusters in full eta range, one cluster in barrel and the other one in endcap, ...etc.). The mass 
                     histograms are filled with mass values shifted by a constant value.
      - Efficiency - electron identification efficiency (for loose, medium and tight) using a simple tag&probe method implementation
                     The efficiency is plotted as a function of candidates et,eta or phi

This tool selects the two leading electron candidates in the event, and fills the monitoring histograms with their properties. For pairs passing a mass selection criteria, a tag&probe method is used to estimate the efficiency. It is possible to have more than one physicsMonTool, configured for different physics processes with di-electrons in the final state. By default, the tool is configured for Z->ee, J/Psi->ee, and Upsilon->ee

Configuration: the physicsMonTool can be configured with following python code:

-------------------------------------------------------------------------------------------------------
from egammaPerformance.egammaPerformanceConf import physicsMonTool
        ZeeMonTool = physicsMonTool(name= "ZeeMonTool",
                                    ElectronContainer      ="ElectronCollection",
                                    Trigger_Items = [],
                                    ProcessName = "Zee",
                                    Selection_Items = ["all"],
                                    massShift = 91188,
                                    massElectronClusterEtCut = 15000,
                                    massLowerCut = 70000,
                                    massUpperCut = 110000)
        
        ToolSvc+=ZeeMonTool
        monManEgamma.AthenaMonTools += [ "physicsMonTool/ZeeMonTool" ]

-------------------------------------------------------------------------------------------------------

where:

ElectronContainer         - is the name of the electron EDM container
Trigger_Items             - is the list of trigger items. A full set of histograms will be made for each item from the list, using only events passing relevant trigger item
Selection_Items           - is the list of custom selections. The histograms will be made separately for each selection item. Only "all" option is implemented so far, where "all" means
 all events
massShift                 - shift of the reconstructed mass used to fill the histograms  
massElectronClusterEtCut  - symmetric Et cut on the two electron candidates
massLowerCut              - lower cut on the reconstructed di-electron mass, used for the tag&probe selection 
massUpperCut              - upper cut on the reconstructed di-electron mass, used for the tag&probe selection 

The computed efficiency using the tag&probe method is defined as:

where:

denominator: number of "tagged" events passing the following selections:

                                            
                        - two leading electrons candidates in the event, with opposite charge                                            
                        - both electrons pass the same cluster Et cut
                     - invariant mass of the electron pair is within a specific mass range                                           
                        - one of the two electrons passes the tight identification criteria. 
                          The other electron (noted here as "e_probe" )is tested against the loose, medium and tight cuts 

numerator: number of events where the second electron "e_probe" further passes a given set of identification cuts (loose, medium, or tight)

The efficiencies are computed at the end of the run.

How to use and test the package

The package can be tested using the RTT procedure on 10 tt RDO events:

source RecExCommon_links.sh
setupLocalDBReplica_CERN.sh
athena.py  -c 'DetDescrVersion="ATLAS-CSC-01-02-00"' -b egammaPerformance/egammaPerformance_RTT_topOptions.py | tee a.out

list of the Monitoring histograms

The following histograms are made from a mixture of 5000 Z->ee events, 5000 J/spi-> event, 5000 W->enu events 5000 Upsilon 1S->ee , 5000Upsilon 2S-> events and 20,000 Jet events.

Electrons passing EF_e10_medium

Histogram name	Description	Possible problems	Action	Reference
hClusterEt	Electron cluster ET
hClusterEta	Electron cluster \x{03b7}
hClusterPhi	Electron cluster \x{03c6}
hCoreEM	Electron core energy in EM calorimeter
hDeltaEta1	Electron track match \x{0394} \x{03b7} (1st sampling)
hDeltaPhi2	Electron track match \x{0394} \x{03c6} (2st sampling)
hEhad1	Electron energy leakage in 1st sampling of hadronic cal.
hEoverP	Electron match track E over P
hEoverPvsEta	Electron E/p vs \x{03b7}
hEoverPvsPhi	Electron E/p vs \x{03c6}
hEoverPvsPhiEle2	Electron E/p for E/p>2
hEoverPvsPhiPos2	Positron E/p for E/p>2
hEtVsEta	Electron Et vs. \x{03b7}
hEtVsPhi	Electron Et vs. \x{03c6}
hEtaCorrMag	Electron match track \x{03b7} corr. magnitude
hF1	Electron fractional energy in 1st sampling
hF2	Electron fractional energy in 2nd sampling
hF3	Electron fractional energy in 3rd sampling
hHighEtRanges	Electrons with ET > 100, 500 and 1000 GeV
hIsEM	Electron IsEM
hNOfBLayerHits	Electron number of track B-Layer Hits
hNOfHighTRTHits	Electron number of high threshold TRT Hits
hNOfPixelHits	Electron number of track pixel Hits
hNOfTRTHits	Electron number of TRT Hits
hNOfTrackSiHits	Electron number of track precision Hits
hRe233e237	Electron uncor. energy fraction in 3x3/3x7 cells in em sampling 2
hRe237e277	Electron uncor. energy fraction in 3x7/7x7 cells in em sampling 2

Photons passing EF_g20

Histogram name	Description	Possible problems	Action	Reference
hClusterEt	Photon cluster ET
hClusterEta	Photon cluster \x{03b7}
hClusterPhi	Photon cluster \x{03c6}
hConvAngleMatch	Photon convAnglMatch flag
hConvTrkMatch	Photon convTrackMatch flag
hCoreEM	Photon core energy in EM calorimeter
hEhad1	Photon energy leakage in 1st sampling of hadronic cal.
hEtVsEta	Photon Et vs. \x{03b7}
hEtVsPhi	Photon Et vs. \x{03c6}
hF1	Photon fractional energy in 1st sampling
hF2	Photon fractional energy in 2nd sampling
hF3	Photon fractional energy in 3rd sampling
hRe233e237	Photon uncor. energy fraction in 3x3/3x7 cells in em sampling 2
hRe237e277	Photon uncor. energy fraction in 3x7/7x7 cells in em sampling 2

Efficiency of Jpsi passing EF_2e5_medium.

Histogram name	Description	Possible problems	Action	Reference
hLooseIdEffVsEt	Loose identification efficiency vs. ET for candidates passing ET, and mass cuts
hLooseIdEffVsEta	Loose identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hLooseIdEffVsPhi	Loose identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hMediumIdEffVsEt	Medium identification efficiency vs. ET for candidates passing ET, and mass cuts
hMediumIdEffVsEta	Medium identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hMediumIdEffVsPhi	Medium identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hTightIdEffVsEt	Tight identification efficiency vs. ET for candidates passing ET, and mass cuts
hTightIdEffVsEta	Tight identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hTightIdEffVsPhi	Tight identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts

Efficiency of Upsilon1S passing EF_2e5_medium.

Histogram name	Description	Possible problems	Action	Reference
hLooseIdEffVsEt	Loose identification efficiency vs. ET for candidates passing ET, and mass cuts
hLooseIdEffVsEta	Loose identification efficiency vs. #eta for candidates passing ET, and mass cuts
hLooseIdEffVsPhi	Loose identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hMediumIdEffVsEt	Medium identification efficiency vs. ET for candidates passing ET, and mass cuts
hMediumIdEffVsEta	Medium identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hMediumIdEffVsPhi	Medium identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hTightIdEffVsEt	Tight identification efficiency vs. ET for candidates passing ET, and mass cuts
hTightIdEffVsEta	Tight identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hTightIdEffVsPhi	Tight identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts

Efficiency of Upsilon2S passing EF_2e5_medium.

Histogram name	Description	Possible problems	Action	Reference
hLooseIdEffVsEt	Loose identification efficiency vs. ET for candidates passing ET, and mass cuts
hLooseIdEffVsEta	Loose identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hLooseIdEffVsPhi	Loose identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hMediumIdEffVsEt	Medium identification efficiency vs. ET for candidates passing ET, and mass cuts
hMediumIdEffVsEta	Medium identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hMediumIdEffVsPhi	Medium identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hTightIdEffVsEt	Tight identification efficiency vs. ET for candidates passing ET, and mass cuts
hTightIdEffVsEta	Tight identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hTightIdEffVsPhi	Tight identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts

Efficiency of Zee passing EF_e20_loose.

Histogram name	Description	Possible problems	Action	Reference
hLooseIdEffVsEt	Loose identification efficiency vs. ET for candidates passing ET, and mass cuts
hLooseIdEffVsEta	Loose identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hLooseIdEffVsPhi	Loose identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hMediumIdEffVsEt	Medium identification efficiency vs. ET for candidates passing ET, and mass cuts
hMediumIdEffVsEta	Medium identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hMediumIdEffVsPhi	Medium identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts
hTightIdEffVsEt	Tight identification efficiency vs. ET for candidates passing ET, and mass cuts
hTightIdEffVsEta	Tight identification efficiency vs. \x{03b7} for candidates passing ET, and mass cuts
hTightIdEffVsPhi	Tight identification efficiency vs. \x{03c6} for candidates passing ET, and mass cuts

W->enu passing EF_e25i_loose and medium cut.

Histogram name	Description	Possible problems	Action	Reference
MtvsdeltaPhi	Transverse mass of W vs \x{0394}\x{03c6} of leading electron
MtvsPhi	Transverse mass of W vs \x{03c6} of leading electron
MtvsEta	Transverse mass of W vs \x{03b7} of leading electron
hClusterEt	ET of leading cluster
hClusterEta	\x{03b7} of leading cluster
hClusterPhi	\x{03c6} of leading cluster
hIsEM	Wenu IsEM
hMassWenu	Transverse mass of (W)

What you will see if you are on shift

-

-

Monitoring Display

-

-

- WhiteBoard