Persist muon hits in SelReports

The goal of this MR is to add muon tracks to the event model and persist muon hits in the SelReports. This has two major uses:

1. Information about the muon hits can be used in HLT1 selections, such as the number of muon hits shared between two tracks.
2. The stored hits can be used to study the IsMuon decision in the downstream software.

The muon tile IDs are now recorded as part of the IsMuon algorithm. If multiple hits are found in a station, only the hit closest to the track extrapolation (normalized by the tile dimensions) is kept.

To-do:

• Create a consolidated muon event model
• Save muon tile IDs in IsMuon
• Create consolidated muon track views
• Add muon segments to BasicParticle
• Fix validation sequence, which now gives a segfault on the CPU

Should be tested with LHCb!4413 (merged)

FYI @acasaisv

assigned to @thboettc

added RTA label

added 2 commits

• c248bacc - Add muon segments to long tracks.
• 9030507f - Add algorithm to consolidate hits from IsMuon

Compare with previous version

Great to see this developing @thboettc!

Thanks a lot for starting this @thboettc !!

added 1 commit

• c3a77944 - Add ConsolidateIsMuon to sequence

Compare with previous version

marked the checklist item Create consolidated muon track views as completed

marked the checklist item Add muon segments to BasicParticle as completed

added 2 commits

• 0e41b1ff - Fixed formatting
• 3c14b95d - Add copyright

Compare with previous version

added 1 commit

• de417a1a - Keep only the best matching muon hit.

Compare with previous version

added 1 commit

• aeefb12f - Fixed formatting

Compare with previous version

Tracks were often hitting the maximum number of hits, so I reworked the logic here so only the "best" hit is saved for each station. I define "best" as the hit with the smallest

(x_{\rm true}-x_{\rm extrap})^2/dx^2 + (y_{\rm true} - y_{\rm extrap})^2/dy^2

, where

dx

and

dy

are the tile dimensions. I'm not sure if this is the best choice. Ideally we'd be able to match the strategy for building muon tracks in HLT2 as closely as possible.

@acasaisv any ideas? We should probably also test how much this affects the shared muon cut (i.e. is it still useful if we only save 1 hit per station?).

added 1 commit

• b901925e - Expand CheckerTrack id array length to include muon IDs.

Compare with previous version

The changes to efficiencies look reasonable to me: https://gitlab.cern.ch/lhcb/Allen/-/jobs/33916327. There are some slight changes to the checker output that are probably due to the additional muon hits affecting the track matching. What do you think @dovombru? We could always just omit the muon hits from the checker tracks if this is an issue, but to me it doesn't seem like a big deal.

marked the checklist item Fix validation sequence, which now gives a segfault on the CPU as completed

resolved all threads

changed the description

added 1 commit

• 517f8ca4 - Add type conversion for MuonTileID to LHCbID

Compare with previous version