Using the Key4hep-Stack for CLIC Simulation and Reconstruction

This assumes that you have access to an installation of the Key4hep-stack, either via CVMFS or spack install. To setup the installation on cvmfs, do:

source /cvmfs/sw.hsf.org/key4hep/setup.sh

These commands will explain how one can run the CLIC detector simulation and reconstruction using the Key4hep-Stack. First we will obtain all the necessary steering and input files for CLIC, simulate a few events and run the reconstruction both with Marlin and k4run. These steps can be adapted to simulate or run other Marlin processors as well.

The CLICPerformance repository contains the steering and input files.

git clone https://github.com/iLCSoft/CLICPerformance

Simulation

Simulating a few events with ddsim can produce output in EDM4hep or LCIO format.

To produce events in EDM4hep format one can run indicating --outputFile <name>_edm4hep.root to produce the output in such format:

cd CLICPerformance/clicConfig

ddsim --compactFile $K4GEO/CLIC/compact/CLIC_o3_v14/CLIC_o3_v14.xml \
      --outputFile ttbar_edm4hep.root \
      --steeringFile clic_steer.py \
      --inputFiles ../Tests/yyxyev_000.stdhep \
      --numberOfEvents 3

To produce events in LCIO format one can run indicating --outputFile <name>.slcio to produce the output file in such format:

cd CLICPerformance/clicConfig

ddsim --compactFile $K4GEO/CLIC/compact/CLIC_o3_v14/CLIC_o3_v14.xml \
      --outputFile ttbar.slcio \
      --steeringFile clic_steer.py \
      --inputFiles ../Tests/yyxyev_000.stdhep \
      --numberOfEvents 3

Reconstruction

Reconstruction with Marlin

To run the reconstruction with Marlin:

cd CLICPerformance/clicConfig

Marlin clicReconstruction.xml \
       --InitDD4hep.DD4hepXMLFile=$K4GEO/CLIC/compact/CLIC_o3_v14/CLIC_o3_v14.xml \
       --global.LCIOInputFiles=ttbar.slcio \
       --global.MaxRecordNumber=3

Reconstruction with Gaudi through k4MarlinWrapper

We can convert the xml steering file to a Gaudi steering file (python):

cd CLICPerformance/clicConfig

convertMarlinSteeringToGaudi.py clicReconstruction.xml clicReconstruction.py

Reconstruction can be performed with LCIO or EDM4hep input, depending on the output format of the events produced during Simulation.

Reconstruction with LCIO input

When using LCIO format for the input events to be used in reconstruction:
- Modify the clicReconstruction.py file to point to the ttbar.slcio input file, and change the DD4hepXMLFile parameter for the InitDD4hep algorithm. In addition the two processors with the comment # Config.OverlayFalse and # Config.TrackingConformal should be enabled by uncommenting their line in the algList at the end of the file.

cd CLICPerformance/clicConfig

sed -i '1s/^/import os\n/' clicReconstruction.py
sed -i 's;read.Files = \[".*"\];read.Files = \["ttbar.slcio"\];' clicReconstruction.py
sed -i 's;EvtMax   = 10,;EvtMax   = 3,;' clicReconstruction.py
sed -i 's;"MaxRecordNumber": ["10"],;"MaxRecordNumber": ["3"],;' clicReconstruction.py
sed -i 's;# algList.append(OverlayFalse);algList.append(OverlayFalse);' clicReconstruction.py
sed -i 's;# algList.append(MyConformalTracking);algList.append(MyConformalTracking);' clicReconstruction.py
sed -i 's;# algList.append(ClonesAndSplitTracksFinder);algList.append(ClonesAndSplitTracksFinder);' clicReconstruction.py
sed -i 's;# algList.append(RenameCollection);algList.append(RenameCollection);' clicReconstruction.py
sed -i 's;"DD4hepXMLFile": \[".*"\],; "DD4hepXMLFile": \[os.environ["LCGEO"]+"/CLIC/compact/CLIC_o3_v14/CLIC_o3_v14.xml"\],;' clicReconstruction.py

Then the reconstruction using the k4MarlinWrapper can be run with

cd CLICPerformance/clicConfig

k4run clicReconstruction.py

Reconstruction with EDM4hep input

When using EDM4hep format for the input events to be used in reconstruction, refer to the EDM converters included with k4MarlinWrapper. Note that:
- MarlinProcessorWrappers need input in LCIO format: EDM4hep collections need to be converted to LCIO
- The output collections of MarlinProcessorWrappers may be used later by other algorithms:
  - Output collections of MarlinProcessorWrappers will be in LCIO format unless these are explicitly converted
  - Some MarlinProcessorWrappers may modify collections instead of producing new ones: the original EDM4hep collection wont be updated in this case and would need conversion from LCIO to EDM4hep.
To run clicReconstruction with EDM4hep format, use the steering file found in the examples folder of k4MarlinWrapper: k4MarlinWrapper/examples/clicRec_e4h_input.py (this also gets installed to $K4MARLINWRAPPER/examples in Key4hep releases)
- Change the line where evtsvc.input is defined to point to the location of your input file.
- At the bottom of the file, in the ApplicationMgr parameters, change EvtMax = 3, to the number of events to run.

This can be run in the following way.

cd CLICPerformance/clicConfig

cp $K4MARLINWRAPPER/examples/clicRec_e4h_input.py .

k4run clicRec_e4h_input.py --EventDataSvc.input ttbar_edm4hep.root

DD4hep Geometry Information

The MarlinDD4hep::InitializeDD4hep processor can be replaced by the k4SimGeant4::GeoSvc and the TrackingCellIDEncodingSvc the latter of which is part of the k4MarlinWrapper repository. This requires removing the wrapped InitDD4hep processor from the algList and the two new processed be appended to the ExtSvc argument in the ApplicationMgr.

We will create another list, svcList for this. In the space following

import os
from Configurables import k4DataSvc, PodioInput
evtsvc = k4DataSvc('EventDataSvc')
evtsvc.input = os.path.join('$TEST_DIR/inputFiles/', os.environ.get("INPUTFILE", "ttbar_edm4hep.root"))

we can start the svcList list and add the GeoSvc and TrackingCellIDEncodingSvc with;

svcList = []
svcList.append(evtsvc)


import os
from Gaudi.Configuration import INFO
from Configurables import GeoSvc, TrackingCellIDEncodingSvc

geoservice = GeoSvc("GeoSvc")
geoservice.detectors = [os.environ["K4GEO"]+"/CLIC/compact/CLIC_o3_v15/CLIC_o3_v15.xml"]
geoservice.OutputLevel = INFO
geoservice.EnableGeant4Geo = False
svcList.append(geoservice)


cellIDSvc = TrackingCellIDEncodingSvc("CellIDSvc")
cellIDSvc.EncodingStringParameterName = "GlobalTrackerReadoutID"
cellIDSvc.GeoSvcName = geoservice.name()
cellIDSvc.OutputLevel = INFO
svcList.append(cellIDSvc)

Then all that is left is to pass that to the ExtSvc argument. At the bottom of the file:

ApplicationMgr( TopAlg = algList,
                EvtSel = 'NONE',
                EvtMax   = 3,
-               ExtSvc = [evtsvc],
+               ExtSvc = svcList,
                OutputLevel=WARNING
              )