Adding ProbNNs for Downstream, Upstream and ProbNNghost

Goes with Rec!3926 (merged) and lhcb-datapkg/ParamFiles!99 (merged)

Downstream has essentially same PID inputs as Long, even for RICH1. Only for electrons the frequency is slightly different for brem emission (depends strongly on origin of the track). In the end it was chosen to keep the same models and training.

For upstream tracks PID is almost completely dominated by RICH, to have a smooth response it is therefore strongly advised to cut on having used RICH1 gas (RICH1_GAS_USED functor)

changed milestone to %RTA/2024.06.03-JuneTS

added PID RTA labels

changed title from Adding ProbNNs for Downstream and Upstream to Adding ProbNNs for Downstream/Upstream and ProbNNghost

added needs ref update label

added 4 commits

• 6b18eb9d - configuration for HLT2 PrVeloUT
• 820656cf - enable checking of Upstream track efficiencies
• f321f6e7 - change hlt2_reco_upstream_particles_low_momentum to use PrVeloUT HLT2 config
• 71508572 - move downstream to long trainings

Compare with previous version

added 1 commit

• e69a8121 - move downstream to long trainings

Compare with previous version

changed the description

mentioned in merge request lhcb-datapkg/ParamFiles!99 (merged)

changed title from Adding ProbNNs for Downstream/Upstream and ProbNNghost to Adding ProbNNs for Downstream, Upstream and ProbNNghost

mentioned in merge request Rec!3926 (merged)

• [2024-06-01 04:55] Validation started with lhcb-2024-patches-mr#623

mentioned in commit Alignment@8ed70f4b

mentioned in merge request Alignment!519 (merged)

approved this merge request

mentioned in commit 0446c711

merged

mentioned in commit Alignment@3fdf62c0

mentioned in commit Alignment@578be8f8

mentioned in merge request Alignment!510 (merged)

mentioned in merge request !3504 (merged)