Hlt2InclDetDiMuonLine not triggering on K*mumu when using real-time reconstruction
Discovered with the help of @mengzhen and @yafan.
When using signal samples of B \to K^{*}\mu\mu
, the Hlt2InclDetDiMuonLine
fails to fire (zero efficiency) when the on-the-fly reconstruction is used:
LAZY_AND: Hlt2InclDetDiMuonLine #=100 Sum=0 Eff=|( 0.000000 +- 0.00000 )%|
DeterministicPrescaler/DeterministicPrescaler #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
VoidFilter/require_pvs #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
FilterDesktop/FilterDesktop #=100 Sum=0 Eff=|( 0.000000 +- 0.00000 )%|
whereas if we use the reconstruction from file, we get a nice 33% efficiency on 100 events:
LAZY_AND: Hlt2InclDetDiMuonLine #=100 Sum=33 Eff=|( 33.00000 +- 4.70213 )%|
DeterministicPrescaler/DeterministicPrescaler #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackRecVertex/UnpackRecVertices #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackCaloHypo/UnpackCaloElectrons #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackCaloHypo/UnpackCaloPhotons #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackCaloHypo/UnpackCaloMergedPi0s #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackCaloHypo/UnpackCaloSplitPhotons #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
DataPacking__Unpack<LHCb__MuonPIDPacker>/UnpackMuonPIDs #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
DataPacking__Unpack<LHCb__RichPIDPacker>/UnpackRichPIDs #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackTrackFunctional/UnpackBestTracks #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackTrackFunctional/UnpackMuonTracks #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackProtoParticle/UnpackNeutralProtos #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
UnpackProtoParticle/UnpackChargedProtos #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
VoidFilter/require_pvs #=100 Sum=100 Eff=|( 100.0000 +- 0.00000 )%|
FilterDesktop/FilterDesktop #=100 Sum=33 Eff=|( 33.00000 +- 4.70213 )%|
leading me to believe something is wrong with this line. This can be reproduced with this options file; if you replace the lines
with reconstruction.bind(from_file=False), \
make_charged_protoparticles.bind(enable_muon_id=True):
run_moore(options, make_lines, public_tools)
with simply
run_moore(options, make_lines, public_tools)
you should see the same behaviour. Logs for completeness: kstmumu_reco_from_file.log kstmumu_reco_on_the_fly.log