Skip to content

Add delta mass functor

Mengzhen Wang requested to merge mengzhen_Thor_DeltaMass into master

To be tested with LHCb!4012 (merged) and DaVinci!867 (merged)

This is for solving Moore#289

Decomposition of IS_ID and IS_ABS_ID using the more basic one PARTICLE_PROPERTY. This latter is used in composition to make PDG_MASS., SIGNED_DELTA_MASS and ABS_DELTA_MASS functors.

Performance test (2022/10/12)

Tested by inserting some additional selections at detached Jpsi line:

option 1:

F.MASS < 3106.9 * MeV, F.MASS > 3086.9 * MeV

option 2:

F.DELTA_MASS("J/psi(1S)") < 10 * MeV

Then 1000 Lb2JpsiLambda events are used to test the efficiency. Both options result in 220 selected events.

Impact on existing functions (2022/10/12)

This MR may also influence the existing functions IS_ID, IS_ABS_ID. To ensure they still work properly, a test is performed using the detached Jpsi line:

If requiring F.IS_ID("J/psi(1S)"), F.IS_ABS_ID("J/psi(1S)"), the efficiency keeps unchanged; if requiring F.IS_ID("pi+"), F.IS_ABS_ID("pi+"), the efficiency goes to zero.

Exclude the new function from TestFunctors.cpp

It seems that TestFunctors.cpp does not support functions using

        ServiceHandle<LHCb::IParticlePropertySvc> pp_svc{top_level.algorithm(), top_level.generate_property_name(),

They would trigger an error message of

unknown location(0): fatal error: in "test_particle_functors": GaudiException: generate_property_name called with a nullptr algorithm
../Phys/FunctorCore/tests/src/TestFunctors.cpp(1534): last checkpoint: "test_particle_functors" test entry

This seems to be a known issue, and can also be triggered by IS_ID("pi+), as documented here:

So the new function are excluded from the checklist of TestFunctors.cpp

A local run of ./run build.x86_64_v2-centos7-gcc11-opt/Phys/FunctorCore/TestFunctors end up with no error message about existing functions.

Edited by Tommaso Fulghesu

Merge request reports