Transition benchmark analyses from Associations to Links

[Copilot]

The benchmark analyses were still consuming the legacy *Associations collections and the older getSimID()/getRecID() traversal pattern. EICrecon has moved to link-based truth/reco navigation, so the benchmark code needs to follow the current *Links model to remain compatible with generated output.

• Summary

  • This patch updates the benchmark code paths that read/reuse reconstructed-truth relations to the current EICrecon link API.
  • It covers the DRICH, tracking performance, and nHCal analyses that were still relying on deprecated association collections.

• What changed

  • Renamed collection usage from *Associations to *Links and updated the corresponding C++ collection types from MCReco*AssociationCollection to MCReco*LinkCollection.
  • Replaced legacy association traversal (getSimID(), getRecID()) with the link-based accessors (getTo(), getFrom()).
  • Updated the TTree branch names used by the tracking analyses to the current link-backed branches, including CentralCKFTrackLinks.weight and CentralCKFTruthSeededTrackLinks.weight.
  • Kept the fix scoped to the affected benchmark consumers rather than changing unrelated analysis logic.

• Example

// before
for (const auto& assoc : assocs) {
  if (assoc.getSimID() == mc_idx) {
    auto reco = recParts.at(assoc.getRecID());
    ...
  }
}

// after
for (const auto& link : links) {
  auto sim = link.getTo();
  auto reco = link.getFrom();
  if (sim.isAvailable() && reco.isAvailable()) {
    ...
  }
}

• Scope
  • RICH benchmark runner and analysis code
  • DIS tracking performance analysis
  • nHCal dimuon and pion rejection analyses

• Fixes Transition from Associations to Links #323

[Copilot]

Pull request overview

Updates detector benchmark analyses to consume EICrecon’s current link-based truth↔reco navigation model, replacing deprecated *Associations collections and legacy ID-based traversal so the benchmarks remain compatible with modern EICrecon PODIO output.

Changes:

• Switched benchmark inputs from *Associations to *Links collections (including collection/type updates in the RICH benchmark and collection lists in runner scripts).
• Updated DIS tracking benchmark configuration and analysis to read link-backed CentralCKF*TrackLinks.weight branches.
• Migrated nHCal analyses from association ID traversal (getSimID()/getRecID()) to link handle navigation (getTo()/getFrom()).

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
benchmarks/tracking_performances_dis/Snakefile	Requests CentralCKF*TrackLinks collections in produced PODIO output.
benchmarks/tracking_performances_dis/analysis/trk_dis_plots.cxx	Updates plot annotations/comments to refer to links rather than associations.
benchmarks/tracking_performances_dis/analysis/trk_dis_analysis.cxx	Reads CentralCKF*TrackLinks.weight branches instead of association weights.
benchmarks/rich/src/ReconstructedParticleAnalysis.cc	Switches reco↔MC traversal to MCRecoParticleLinkCollection using getFrom()/getTo().
benchmarks/rich/src/RawHitAnalysis.cc	Updates raw-hit truth navigation to MCRecoTrackerHitLinkCollection.
benchmarks/rich/src/benchmark.cc	Updates frame collection names/types to *Links for RICH processing.
benchmarks/rich/run_benchmark.rb	Updates requested output collections to DRICHRawHitsLinks and ReconstructedChargedParticleLinks.
benchmarks/rich/include/ReconstructedParticleAnalysis.h	Updates includes/signature from association to link collection.
benchmarks/rich/include/RawHitAnalysis.h	Updates includes/signature from association to link collection.
benchmarks/nhcal_pion_rejection/scripts/pion_rejection_analysis.cxx	Migrates MC↔reco matching to MCRecoParticleLinkCollection (getTo()/getFrom()).
benchmarks/nhcal_dimuon_photoproduction/scripts/dimuon_photoproduction_analysis.cxx	Migrates MC↔reco matching to MCRecoParticleLinkCollection (getTo()/getFrom()).

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[wdconinc]

@veprbl This passes a full CI run (for what that's worth). It would need to be uploaded into hydra with all channels activated for actual comparison since there are 1934 artifacts. But based on a download of the artifact archive here and for a reference run, and running ImageMagick's compare there are some strange differences. So don't merge yet :-)

[wdconinc]

E.g. https://eicweb.phy.anl.gov/EIC/benchmarks/detector_benchmarks/-/jobs/8389718/artifacts/file/results/emcal_barrel_electron_fsam.png vs https://eicweb.phy.anl.gov/EIC/benchmarks/detector_benchmarks/-/jobs/8390711/artifacts/file/results/emcal_barrel_electron_fsam.png

But that's generated by code not actually changed in this PR...

[wdconinc]

It's fine, this is apparently broken in main too now... 🆘 https://eicweb.phy.anl.gov/EIC/benchmarks/detector_benchmarks/-/jobs/8392454/artifacts/file/results/emcal_barrel_electron_fsam.png

[wdconinc]

https://eicweb.phy.anl.gov/EIC/benchmarks/detector_benchmarks/-/jobs/8400656/artifacts/file/results/emcal_barrel_electron_fsam.png looks better