NIKHEF PAT workshop Oct2006

From Atlas Wiki
Jump to navigation Jump to search

Set up CMT

  • Login to lxplus at CERN and create a working directory called "Tutorial". We will use the release 12.0.2 this tutorial:
ssh lxplus.cern.ch
cd scratch0
mkdir Tutorial
cd Tutorial
mkdir 12.0.2
  • Create a file called "requirements" and setup CMT. My requirements file looks like this where my working area for 12.0.2 is ${HOME}/scratch0/Workshop/12.0.2:
set   CMTSITE  CERN
set   SITEROOT /afs/cern.ch
macro ATLAS_DIST_AREA ${SITEROOT}/atlas/software/dist
                      
macro ATLAS_GROUP_AREA "/afs/cern.ch/atlas/groups/PAT/Tutorial"
 
macro ATLAS_TEST_AREA "" \
        12.0.2 "${HOME}/scratch0/Workshop/12.0.2"
 
 
apply_tag oneTest
 
use AtlasLogin AtlasLogin-* $(ATLAS_DIST_AREA)
  • Then do the following:
source /afs/cern.ch/sw/contrib/CMT/v1r18p20060301/mgr/setup.sh
cmt config

Setup for the release 12.0.2

  • Do the following:
source setup.sh -tag=12.0.2
  • Check that the CMT path is correct by doing this
echo $CMTPATH
  • Stuff like this should be printed to your screen:
/afs/cern.ch/user/o/ordonez/scratch0/Workshop/12.0.2:
/afs/cern.ch/atlas/software/builds/AtlasOffline/12.0.2
  • You see that the path to your working directory, the path to the release directory and the external are set correctly - do not proceed unless this is done correctly.

Run Time Setup

  • Go to your working area and check out the follow package:
cd 12.0.2
cmt co -r UserAnalysis-00-08-05 PhysicsAnalysis/AnalysisCommon/UserAnalysis
  • Now compile everything as follows:
cd PhysicsAnalysis/AnalysisCommon/UserAnalysis/cmt
cmt config
source setup.sh
cmt broadcast gmake
cd ../run
  • To test that everything has been done correctly, do:
get_files HelloWorldOptions.py
athena.py -b HelloWorldOptions.py
  • Stuff like this should be printed to the screen:
HelloWorld           INFO execute()
HelloWorld           INFO An INFO message
HelloWorld           WARNING A WARNING message
HelloWorld           ERROR An ERROR message
HelloWorld           FATAL A FATAL error message
AthenaEventLoopMgr   INFO   ===>>>  end of event 9    <<<===
HistorySvc           INFO Service finalised successfully
ChronoStatSvc.f...   INFO  Service finalized succesfully
ToolSvc              INFO Removing all tools created by ToolSvc
ApplicationMgr       INFO Application Manager Finalized successfully
ApplicationMgr       INFO Application Manager Terminated successfully

The Next Time You Login

  • The next time you login, just do the following:
cd ${HOME}/scratch0/Workshop
source setup.sh -tag=12.0.2
cd 12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/cmt
source setup.sh
cd ../run

Get The Necessary Files

  • We will use inclusive Z to mumu events (sample 4202), digitized in 11.0.41 but the ESD and AOD produced in the release 12.0.1. Copy the PoolFileCatalog.xml for the AOD, ESD and Raw Data (RDO) into your UserAnalysis "run" directory:
cp ~ordonez/scratch0/Workshop/12.0.2/data/PoolFileCatalog.xml .
cp ~ordonez/scratch0/Workshop/12.0.2/data/*.py .

Run the AnalysisSkeleton for 500 Events

  • get the necessary job options file from the release:
get_files AnalysisSkeleton_jobOptions.py
  • Edit AnalysisSkeleton_jobOptions.py and specify the input data to be the AOD files that you just copied in the previous step: to do this, replace these lines
EventSelector.InputCollections = [ 
  "AOD.pool.root"
                                 ]
  • with this line
include( "mc11.004201.ZeeJimmy.recon.AOD.v11000401.py" )
  • Check that in AnalysisSkeleton_jobOptions.py, the number of events to run is set to 500:
# Number of Events to process
theApp.EvtMax = 500
  • Now run the AnalysisSkeleton_jobOptions.py. It should produce a file called AnalysisSkeleton.aan.root which contains some ROOT histograms and an AthenaAwareNTuple:
athena.py AnalysisSkeleton_jobOptions.py | tee athenaout.log
  • You've just produced the AthenaAwareNTuple in a standalone way, directly on the AOD, or the ESD or even on the RDO. For Further details on how to produce the AthenaAwareNTuple in this way, follow this link: AthenaAwareNTuple.
  • Now, you may look at the the AnalysisSkeleton.h, the AnalysisSkeleton.cxx and AnalysisSkeleton_jobOptions.py to see how the ROOT histograms and the AthenaAwareNTuple of the output file AnalysisSkeleton.aan.root are defined, filled and saved: Follow this link to the AnalysisSkeleton. For further details on how to produces ROOT histograms and Tuples in ATHENA, look at the THistSvc.
  • You should open the output file AnalysisSkeleton.aan.root in ROOT and browse the histograms and the AthenaAwareNTuple:
root AnalysisSkeleton.aan.root
new TBrowser
  • You will notice that beside the user-defined NTuple, the AthenaAwareNTuple tools add references to the AOD, ESD and Raw Data, in addition the RunNumber and the EventNumber.

More Information

  • For further details on the analysis tools, look at the PhysicsAnalysisTools. Details on the content of the ESD, the AOD and the TAG are on StoregateKeysForESD, StoregateKeysForAOD and TagForEventSelection.

Z to mumu on AOD

  • In this exercise, we will do Z to mumu reconstruction on AOD: we will use the association objects Analysis::Muon and its associated TrackParticle. Proceed as follows: copy the code from my area, compiling and running it, then we go through it:
cd ../src
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/src/ZmumuOnAOD.cxx .
cd ../UserAnalysis
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/UserAnalysis/ZmumuOnAOD.h .
cd ../run
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/run/ZmumuOnAOD_jobOptions.py .
  • Now, you want to go the src/components directory and declare this new algorithm ZmumuOnAOD.
cd ../src/components
  • Edit UserAnalysis_entries.cxx and add these lines - follow the example that is already in there for the AnalysisSkeleton:
#include "UserAnalysis/ZmumuOnAOD.h"

DECLARE_ALGORITHM_FACTORY( ZmumuOnAOD )

  DECLARE_ALGORITHM( ZmumuOnAOD )
  • Save UserAnalysis_entries.cxx and go back the cmt directory and compile:
cd ../../cmt
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/cmt/requirements .
cmt config
cmt broadcast gmake
cd ../run
  • Now you may run the ZmumuOnAOD algorithm as follows:
athena.py ZmumuOnAOD_jobOptions.py
  • This should produce an AthenaAwareNTuple file, ZmumuOnAOD.aan.root. While Athena is running, let's go through the code by following this link: ZmumuOnAOD for the details of the algorithm and the selection cuts.
  • Now, open this file in ROOT and browse the AthenaAwareNTuple:
root -l
TFile *f = TFile::Open("ZmumuOnAOD.aan.root")
new TBrowser
TTree * zmumuOnAOD=(TTree*) gDirectory->Get("CollectionTree");
zmumuOnAOD->SetMarkerStyle(20)
zmumuOnAOD->SetMarkerSize(0.5)
zmumuOnAOD->SetMarkerColor(2)
zmumuOnAOD->SetLineColor(4)
zmumuOnAOD->SetLineWidth(2)
zmumuOnAOD->Draw("ZmmMass")
  • By default, we used the Muonboy/STACO/MuTag association objects in the AOD and the associated TrackParticles. But you may change it in the job options ZmumuOnAOD_jobOptions.py: the StoreGate keys for the various objects in the AOD are listed here: StoregateKeysForAOD.
  • Exercise: add a piece code to access the number of hits in the various stations of the Inner Detector and the Muon Spectrometer. Recompile, and run. Then in ROOT look at the NTuple in ROOT and check that the distributions look OK.


Z to ee on AOD

  • In this exercise, we will do Z to ee reconstruction on AOD: we will use the association objects Electron/Photon and its associated Electron TrackParticle. Proceed as follows: copy the code from my area, compiling and running it, then we go through it:
cd ../src
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/src/ZeeOnAOD.cxx .
cd ../UserAnalysis
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/UserAnalysis/ZeeOnAOD.h .
cd ../run
cp ~ordonez/scratch0/Workshop/12.0.2/PhysicsAnalysis/AnalysisCommon/UserAnalysis/run/ZeeOnAOD_jobOptions.py .
  • Now, you want to go the src/components directory and declare this new algorithm ZeeOnAOD.
cd ../src/components
  • Edit UserAnalysis_entries.cxx and add these lines - follow the example that is already in there for the AnalysisSkeleton:
#include "UserAnalysis/ZeeOnAOD.h"

DECLARE_ALGORITHM_FACTORY( ZeeOnAOD )

  DECLARE_ALGORITHM( ZeeOnAOD )
  • Save UserAnalysis_entries.cxx and go back the cmt directory and compile:
cd ../../cmt
cmt config
cmt broadcast gmake
cd ../run
  • Now you may run the ZeeOnAOD algorithm as follows:
athena.py -b ZeeOnAOD_jobOptions.py
  • This should produce an AthenaAwareNTuple file, ZeeOnAOD.aan.root. While Athena is running, let's go through the code by following this link: ZeeOnAOD for the details of the algorithm and the selection cuts.
  • Now, open this file in ROOT and browse the AthenaAwareNTuple:
root
TFile *_file0 = TFile::Open("ZeeOnAOD.aan.root")
new TBrowser
TTree * zmumuOnAOD=(TTree*) gDirectory->Get("CollectionTree");
zeeOnAOD->SetMarkerStyle(20)
zeeOnAOD->SetMarkerSize(0.5)
zeeOnAOD->SetMarkerColor(2)
zeeOnAOD->SetLineColor(4)
zeeOnAOD->SetLineWidth(2)
zeeOnAOD->Draw("ZeeMass")
ZeeOnAOD->Draw("ElectronIsEM")
  • The StoreGate keys for the various objects in the AOD are listed here: StoregateKeysForAOD.
  • Exercise: add 2 branches to you ZeeOnAOD NTuple to save the Electron cluster energy and TrackParticle momentum. Recompile and run. Then in ROOT check your NTuple that these variables are filled correctly.

The TrackParticle

  • In this exercise, we will access the a collection of TrackParticles directly from the AOD. Note that from the association objects, Analysis::Muon or Electron, you can access the access the various associated TrackParticles, e.g., look in AnalysisMuon.cxx. But in this exercise, we will use one collection of TrackParticles in the AOD. You may proceed as follows, then we go through the code:
cd ../src
cp ~/PhysicsAnalysis/AnalysisCommon/UserAnalysis/src/TrackExample.cxx .
cd ../UserAnalysis
cp ~/PhysicsAnalysis/AnalysisCommon/UserAnalysis/UserAnalysis/TrackExample.h .
cd ../run
cp ~/PhysicsAnalysis/AnalysisCommon/UserAnalysis/run/TrackExample_jobOptions.py .
  • Now, you want to go the src/components directory and declare this new algorithm TrackExample.
cd ../src/components
  • edit UserAnalysis_entries.cxx and add these lines

(follow the example that is already in there for the AnalysisSkeleton and AnalysisMuon): 

#include "UserAnalysis/TrackExample.h"

DECLARE_ALGORITHM_FACTORY( TrackExample )

  DECLARE_ALGORITHM( TrackExample )
  • Save UserAnalysis_entries.cxx and go back the cmt directory and compile:
cd ../../cmt
cmt config
cmt broadcast gmake
cd ../run
  • Now you may run the TrackExample algorithm as follows:
athena.py TrackExample_jobOptions.py
  • You should have an AthenaAwareNTuple file, TrackExample.aan.root. While Athena is running, let's go through the code by following this link: KeteviAnalysis. Note the log output to the screen: you will see that the perigee information and the track summary information are being printed to the screen. Look at TrackExample.cxx in details to see how to access this information and more from the TrackParticle.
  • Now, open this file in ROOT and browse the AthenaAwareNTuple:
root -l
TFile *_file0 = TFile::Open("TrackExample.aan.root")
new TBrowser
TTree * trackExample=(TTree*) gDirectory->Get("CollectionTree");
trackExample->SetMarkerStyle(20)
trackExample->SetMarkerSize(0.5)
trackExample->SetMarkerColor(2)
trackExample->SetLineColor(4)
trackExample->SetLineWidth(2)
trackExample->Draw("TrackPEta")
trackExample->Draw("TrackPD0")
trackExample->Draw("TrackPZ0:TrackPD0")
  • In this example, we used the Inner Detector collection of TrackParticle in AOD, with StoreGate key "TrackParticleCandidate": follow this link for the AOD content and the keys for various objects/containers int the AOD - StoregateKeysForAOD.

Event Selection with TAG or AAN

  • In this exercises, we will use the TAG and the AthenaAwareNTuple to select the final states of interest. For details on the TAG based event selection, follow this link: TagForEventSelection.

Event Selection with the TAG

  • In this exercise, we will use the TAG to select the final states of interest. The data that we are contains Z to mumu so there is not must selection final state we can do with the TAG on this data but this exercise will give a flavor of what you can to with the TAG and how to it. Consider the following job options.
  2. Author - Ketevi A. Assamagan

AllAlgs = False doHist = False doCBNT = False doWriteTAG = False doWriteESD = False readAOD = True doAOD = False doWriteAOD = True

  1. read the TAG as input to the job

readColl = True

  1. Number of Events

EvtMax = 20

  1. Use the TAG that you just produce to select the events
  2. define also the selection criteria

include ( "mc11.004202.ZmumuJimmy.recon.TAG.v11000401.py" ) CollInputQuery="NLooseMuon>0 && NLooseMuon<3 && abs(LooseMuonEta1)<2.5 && abs(LooseMuonEta2)<2.5 && LooseMuonPt2>10000 && LooseMuonPt2>10000"

  1. Create a new RDO, ESD and AOD that contain only selected events

PoolAODOutput = "mc11.004202.ZmumuJimmy.recon.AOD.TagSel.v11000401.root"

  1. main reconstruction job jobOptions

include ("RecExCommon/RecExCommon_topOptions.py")

Create a file called TagBasedEventSelection_topOptions.py and copy the above options into it and save it. We are using the TAG file as input and selecting the events at the AOD level. Our selection criteria are defined in the CollInputQuery. We are querying for the events having:

Two one or 2 loose muons Eta of each loose muons within 2.5 Pt of 2 loose muons above 10 GeV - unit in MeV

For the all the events that pass the selection, we create an AOD file that contains ONLY the selected events: for details on how to use the TAG and which TAG content or attributes you can make selection on, follow this link: TagForEventSelection. The new AOD file name should be mc11.004202.ZmumuJimmy.recon.AOD.TagSel.v11000401.root as specified in the above job options. To select events at the ESD level, one would simply add this lines:

readESD = True readAOD = False

Now run this job options and see that you have a new AOD file that contains only the selected events. An easy and a quick way to see the TAG attributes is to just open the TAG file in ROOT and browse. Copy a sample TAG file from my area and open it in ROOT:

cp ~ketevi/scratch0/Tutorial/12.0.1/PhysicsAnalysis/AnalysisCommon/UserAnalysis/run/TAG.fromAOD.pool.root . root TAG.fromAOD.pool.root b=TBrowser()

NOw you can do analysis on the new AOD that contains ONLY the selected events.

Event Selection with the AthenaAwareNTuple In this exercise, we will use the AthenaAwareNTuple that you produced in the introdcution: you should have an NTuple called AnalysisSkeleton.aan.root. Proceed as follows:

cp AnalysisSkeleton_jobOptions.py AanSelectionAnalysis_topOptions.py

Edit AanSelectionAnalysis_topOptions.py and replace this line

include( "mc11.004201.ZeeJimmy.recon.AOD.v11000401.py" )

with these lines

EventSelector.InputCollections = [ "AnalysisSkeleton.aan" ] EventSelector.Query="NElectrons>0 && NElectrons<3 && abs(ElectronEta[0])<2.5 && abs(ElectronEta[1])<2.5 && ElectronPt[0]>10000 && ElectronPt[1]>10000" EventSelector.CollectionType = "ExplicitROOT"

Here, we are using the AthenaAwareNTuple called AnalysisSkeleton.aan.root to selected the event at the AOD. Then, we are run the Analysis code, AnalysisSkeleton.cxx on the selected events. Note that when you use the TAG or the AthenaAwareNTuple to event selection, the file extension .root must not be specified: it is assumed and will be appended. In these exercises, we are using the ROOT version of the TAG --- also, our AthenaAwareNTuple is a ROOT file. This is why we specify the CollectionType to be ExplicitROOT. There is also a database version of the TAG: for details see TagForEventSelection. In AanSelectionAnalysis_topOptions.py, replace these lines

THistSvc.Output = ["AANT DATAFILE='AnalysisSkeleton.aan.root' OPT='RECREATE'"] AANTupleStream.OutputName = 'AnalysisSkeleton.aan.root'

with these lines:

THistSvc.Output = ["AANT DATAFILE='AnalysisSkeleton.sel.aan.root' OPT='RECREATE'"] AANTupleStream.OutputName = 'AnalysisSkeleton.sel.aan.root'

Now run:

athena.py -b AanSelectionAnalysis_topOptions.py

You should have a new AthenaAwareNTuple file called AnalysisSkeleton.sel.aan.root. That we've just done is the following: we use a previously available AthenaAwareNTuple or TAG to select events and we did analysis on the selected events creating a new AthenaAwareNTuple. We also saw that it is possible to create a new AOD that contains ONLY the selected events. Now check that the selection criteria that we specified above were indeed apply: open AnalysisSkeleton.sel.aan.root and plot the EventNumber and the NElectrons (this should be one or two) according to our selection criteria.

Retrieved from "https://wiki.nikhef.nl/atlas/index.php?title=NIKHEF_PAT_workshop_Oct2006&oldid=982"