How to run FATRAS, Digitization and Truth Tracking with a custom geometry

This twiki has been written to run FATRAS, FastDigitization and truth-seeded reconstruction using a general custom geometry.

Getting started

Setup 20.20.X-VAL:

Use latest stable release (replace rel_X with the correct release) from: http://atlas-nightlies-browser.cern.ch/~platinum/nightlies/info?tp=g&nightly=20.20.X-VAL

asetup 20.20.X-VAL,rel_X,slc6,gcc48,opt,here,runtime

Building geometry

How to implement fast definitions of realistic detector layouts in the ATLAS TrackingGeometry is described here: https://twiki.cern.ch/twiki/bin/view/Sandbox/CustomGeometryImplementationGuide

Running FATRAS

Once you have defined your geometry in the XML package you have to edit the jobOption (in the ISF_FatrasDetDescrAlpine package) that has to be pre-icluded when you run FATRAS.

Just check the ISF_FatrasDetDescrAlpine package out:

pkgco.py ISF_FatrasDetDescrAlpine

and edit the file share/CustomGeometrySettings_preInclude.py

Here you have to define the layout you want to use and all its specification in order to build it correctly. The reference layout is ATLAS Run2 geometry.

If you want to run FATRAS, you must have an EVNT input file you want to process. In case you don't have it, I have one file available for you (single muons, pT = 100 GeV) ;-).

Add another pre-include jobOption in your area with the remaining specifications. It should contain at least these lines (You can name it Sim_pre_jo.py):

from AthenaCommon.DetFlags import DetFlags 
DetFlags.ID_setOn() 
DetFlags.TRT_setOff() 
DetFlags.Calo_setOff() 
DetFlags.Muon_setOff() 
from TrkDetDescrSvc.TrkDetDescrJobProperties import TrkDetFlags 
TrkDetFlags.MaterialSource = 'Input' 
from AthenaCommon.AppMgr import ServiceMgr 
ServiceMgr.MessageSvc.Format = "% F%50W%S%7W%R%T %0W%M" 
# the magnetic field 
from MagFieldServices import SetupField 
from IOVDbSvc.CondDB import conddb 
conddb.addOverride('/GLOBAL/BField/Map','BFieldMap-FullAsym-09-solTil')

You can also add a post-include jobOption to add VP1 (You can name it Sim_post_jo.py):

from VP1Algs.VP1AlgsConf import VP1Alg 
topSequence += VP1Alg()

NB: To load correctly VP1 libraries you need to export the GL libraries first:

export LD_PRELOAD=/usr/lib64/libGL.so

Then run FATRAS:

Sim_tf.py --simulator 'Fatras_newExtrapolation_IDOnly' --randomSeed 0 --enableLooperKiller True --maxEvents 10 --skipEvents 0 --geometryVersion ATLAS-PX-ITK-00-00-00_VALIDATION --conditionsTag OFLCOND-SIM-00-00-00 --inputEVNTFile /afs/cern.ch/user/n/ncalace/public/EVNT.06603958._000029.pool.root.1 --outputHITSFile "Hits.pool.root" --preInclude "Simulation/ISF/ISF_Fatras/ISF_FatrasDetDescrAlpine/share/CustomGeometrySettings_preInclude.py,Sim_pre_jo.py" --postInclude "Sim_post_jo.py"

or

Sim_tf.py --simulator 'Fatras_newExtrapolation_IDOnly' --randomSeed 0 --enableLooperKiller True \
--maxEvents 10 --skipEvents 0 --geometryVersion ATLAS-PX-ITK-00-00-00_VALIDATION --conditionsTag OFLCOND-SIM-00-00-00 \
--inputEVNTFile /afs/cern.ch/user/n/ncalace/public/EVNT.06603958._000029.pool.root.1 --outputHITSFile "Hits.pool.root"\
--preInclude "Simulation/ISF/ISF_Fatras/ISF_FatrasDetDescrAlpine/share/CustomGeometrySettings_preInclude.py,Sim_pre_jo.py" \
--postInclude "Sim_post_jo.py"
 

Very Fast Digitization module and truth-seeded reconstruction

In order to run the very fast digitization module, you have to check out the FastSiDigitization package and compile it.

The jobOptions used to run are in the share folder of this package. In particular, you can use two jobOptions:

• share/SiSmearing_noPU.py --> To run digitization and produce clusters (for both Pixel and SCT).
• share/SiSmearing_noPU_CompletedjO.py --> To run digitization and truth-tracking in one step.

athena Simulation/ISF/ISF_Fatras/ISF_FatrasDetDescrAlpine/share/CustomGeometrySettings_preInclude.py InnerDetector/InDetDigitization/FastSiDigitization/share/SiSmearing_noPU.py

or

athena Simulation/ISF/ISF_Fatras/ISF_FatrasDetDescrAlpine/share/CustomGeometrySettings_preInclude.py InnerDetector/InDetDigitization/FastSiDigitization/share/SiSmearing_noPU.py

The truth-seeded reconstruction is configured to used the Trk::KalmanFitter with the Trk::ExtrapolationEngine.

-- Main.NoemiCalace - 2015-07-01