iLCSoft / ILDConfig

Package for iLCSoft configuration files related to the ILD detector simulation
GNU General Public License v3.0
3 stars 23 forks source link

Muon reconstruction failing #88

Closed rete closed 4 years ago

rete commented 6 years ago

We see a non-negligible fraction of single muon events failing. This issue was first communicated by email:

Dear all,

now to the next problem, which concerns muons around costh~ -0.6 (and also +0.6, but I'm not yet convinced that these are identical...)

in single particle 20 gev mu- events, we sometimes get exactly 1 PandoraPFO, which is neutral and contains only MUON hits.
The MarlinTrack is there, as are ECAL/HCAL hits, but they don't make it into a PandoraPFO.
I think there may be a related effect at 90deg: at least the symptoms look similar.

I stripped off some events which created problems, and uploaded them here:   http://research.kek.jp/people/jeans/badmuons/
You can find the sim output, reco output, and reco logfile (with MarlinPandora in DEBUG mode, which shows some complaints/errors).

The behaviour in physics (ie multi-particle) events is different. If I understand correctly, Junping has noticed that his muons sometime/often result in charged PFOs without assigned muon hits, in the similar/same costh region.

The problematic error stack is the following:

[ VERBOSE "MyDDMarlinPandora"] ClusterFitHelper: linear fit to cluster failed. 
[ VERBOSE "MyDDMarlinPandora"] ClusterFitHelper::FitStart(pCluster, m_nClusterLayersToFit, clusterFitResult) return STATUS_CODE_INVALID_PARAMETER
[ VERBOSE "MyDDMarlinPandora"]     in function: AssociateMuonTracks
[ VERBOSE "MyDDMarlinPandora"]     in file:     /cvmfs/ilc.desy.de/sw/x86_64_gcc49_sl6/v02-00-01/PandoraPFANew/v03-09-00/LCContent-v03-01-03/src/LCParticleId/MuonReconstructionAlgorithm.cc line#: 96
[ VERBOSE "MyDDMarlinPandora"] this->AssociateMuonTracks(pMuonClusterList) return STATUS_CODE_INVALID_PARAMETER
[ VERBOSE "MyDDMarlinPandora"]     in function: Run
[ VERBOSE "MyDDMarlinPandora"]     in file:     /cvmfs/ilc.desy.de/sw/x86_64_gcc49_sl6/v02-00-01/PandoraPFANew/v03-09-00/LCContent-v03-01-03/src/LCParticleId/MuonReconstructionAlgorithm.cc line#: 55
[ VERBOSE "MyDDMarlinPandora"] iter->second->Run() throw STATUS_CODE_INVALID_PARAMETER
[ VERBOSE "MyDDMarlinPandora"]     in function: RunAlgorithm
[ VERBOSE "MyDDMarlinPandora"]     in file:     /cvmfs/ilc.desy.de/sw/x86_64_gcc49_sl6/v02-00-01/PandoraPFANew/v03-09-00/PandoraSDK-v03-02-01/src/Api/PandoraContentApiImpl.cc line#: 231
[ VERBOSE "MyDDMarlinPandora"] Failure in algorithm Alg0003, MuonReconstruction, STATUS_CODE_INVALID_PARAMETER
[ VERBOSE "MyDDMarlinPandora"] PhotonReconstructionAlgorithm::CreateRegionsOfInterests - no photon candidates available, no regions of interests created.
[ VERBOSE "MyDDMarlinPandora"] HighEnergyPhotonRecoveryAlgorithm: Failed to obtain cluster list PhotonClusters
[ VERBOSE "MyDDMarlinPandora"] SoftClusterMergingAlgorithm: Failed to obtain cluster list PhotonClusters
[ VERBOSE "MyDDMarlinPandora"] IsolatedHitMergingAlgorithm: Failed to obtain cluster list PhotonClusters
[ VERBOSE "MyDDMarlinPandora"] PandoraContentApi::ReplaceCurrentList<CaloHit>(*this, m_inputCaloHitListName) return STATUS_CODE_NOT_FOUND
[ VERBOSE "MyDDMarlinPandora"]     in function: Run
[ VERBOSE "MyDDMarlinPandora"]     in file:     /cvmfs/ilc.desy.de/sw/x86_64_gcc49_sl6/v02-00-01/PandoraPFANew/v03-09-00/LCContent-v03-01-03/src/LCClustering/ClusteringParentAlgorithm.cc line#: 34
[ VERBOSE "MyDDMarlinPandora"] iter->second->Run() throw STATUS_CODE_NOT_FOUND
[ VERBOSE "MyDDMarlinPandora"]     in function: RunAlgorithm
[ VERBOSE "MyDDMarlinPandora"]     in file:     /cvmfs/ilc.desy.de/sw/x86_64_gcc49_sl6/v02-00-01/PandoraPFANew/v03-09-00/PandoraSDK-v03-02-01/src/Api/PandoraContentApiImpl.cc line#: 231
[ VERBOSE "MyDDMarlinPandora"] Failure in algorithm Alg0075, ClusteringParent, STATUS_CODE_NOT_FOUND

The problem comes from the cluster fit failing thus making the muon clustering algorithm failing too.

The different resources for debugging, as pointed out from the mail are available here:

rete commented 6 years ago

Related issue: https://github.com/iLCSoft/lcgeo/issues/227

rete commented 6 years ago

Related fix: #89