edm4eic.yaml: fix MCRecoTrackerHitAssociation to relate to a single sim hit at a time

[veprbl]

I believe, having sim hits as a OneToMany relation was a mistake. The idea behind having a weight is to specify level of contribution of a specific sim hit. If multiple sim hits are associated, user should expect multiple association records in the collection. We are only now introducing the actual use of this, so now is the good time to do a fix eic/EICrecon#1392 (comment)

cc @wdconinc @c-dilks

[c-dilks]

For the dRICH, simHits.size > 1 when two optical photons hit the same SiPM pixel, and weight is not used.

This PR requires:

• check for anything that assumes that one MCRecoTrackerHitAssociation corresponds to one unique raw hit, namely the EICrecon algorithm IrtCherenkovParticleID; update if needed
• update PhotoMultiplierHitDigi, which fills the collection
• updates may also be needed downstream of EICrecon, namely reconstruction_benchmarks/benchmarks/rich

All of these are feasible; please followup with @chchatte92 and @bleaktwig

[veprbl]

Thanks for overview, @c-dilks, I've submitted a PR to update EICrecon and https://eicweb.phy.anl.gov/EIC/benchmarks/reconstruction_benchmarks/-/merge_requests/327

[c-dilks]

Awesome, thanks!

[wdconinc]

Yes, this looks like the right direction to me. Also consistent with EDM4hep where all associations are one-to-one, https://github.com/key4hep/EDM4hep/blob/main/edm4hep.yaml#L490-L560.

[veprbl]

If there are no objections to this, let's merge.