search

eic / EICrecon

EIC Reconstruction - JANA based
https://eic.github.io/EICrecon
GNU Lesser General Public License v3.0
5 stars 26 forks source link

Error: [DRICHIrtCherenkovParticleID] radiators have differing numbers of TrackSegments #990

Closed c-dilks closed 10 months ago

c-dilks commented 10 months ago

Environment: (where does this bug occur, have you tried other environments)

Steps to reproduce: (give a step by step account of how to trigger the bug)

  1. run standard reconstruction for DIS events (see CI job eicrecon-dis)

Expected Result: (what do you expect when you execute the steps above)

No errors from dRICH reconstruction

Actual Result: (what do you get when you execute the steps above)

https://github.com/eic/EICrecon/actions/runs/6215248891/job/16868173477#step:5:700

Error: DRICHMergedTracks] [error] input track collections do not have the same size; cannot merge
Error: DRICHIrtCherenkovParticleID] [error] radiators have differing numbers of TrackSegments

Context:

From #984 comments:

I suspect these are charged particles which are outside the aerogel acceptance, but still inside the gas acceptance. After https://github.com/eic/EICrecon/pull/735 we simply don't have aerogel TrackkPoints for such tracks, hence the error (prior behavior was to propagate and fail, but still produce a track with zero TrackPoints (wasting time))

This is a bad issue though: if there are any tracks which only pass through gas and not aerogel, no dRICH PID will be done for the event!

A quick fix to restore behavior before https://github.com/eic/EICrecon/pull/735 would be to write out a TrackSegment with zero TrackPoints for such tracks, if and only if the track would pass through the backplane of the gas; see the filter surface implementation in https://github.com/eic/EICrecon/pull/735, which depends on the radiator (aerogel has smaller radius than gas, try making both filters be the gas back plane disc).

c-dilks commented 10 months ago

This is a bad issue though: if there are any tracks which only pass through gas and not aerogel, no dRICH PID will be done for the event!

If this is really true, then dRICH PID in production 23.09 is limited only to events which have charged particles in the aerogel acceptance, and NO charged particles in the gas-only acceptance region (outside aerogel, but still inside gas radiator volume).