[mhemmer-cern]

The EMCal track-matching veto misses many clusters made by conversion electrons (PCM photons counted twice in PCM-EMC pairs). The derived data cannot tell where the matching chain fails, so add an MC truth process that follows the leading MC particle of each cluster through it.

emcTmMonitor:

• New processMCTruth (off by default). For every cluster, take the leading MC contributor (largest energy fraction) and classify it as conversion e+-, primary e+-, other e+-, other charged, photon, neutral or other.
• For charged leading particles, follow the particle's best reconstructed track and record one status per cluster: no reco track, not extrapolated to the EMCal, no matching candidate (neither a global track nor a V0 leg), candidate not matched, matched as primary, matched as secondary.
• Histograms under MCTruth/:
  • status vs truth class, track type, V0-leg flag and MC pT
  • track-cluster residuals up to +-0.4 in the origin frame of the extrapolation, so tails beyond the matching windows are visible
  • own residual minus stored residual, to validate the frame conversion
  • status vs conversion radius
  • E/p of conversion-e+- clusters
• Needs emcal-correction-task with processMCWithSecondaries and the photon conversion builder (aod::V0Legs).
• The event selection moves into a helper shared by both process functions; the event histograms are filled only once.
• Charges come from the O2DatabasePDG service.
• Histograms with labelled axes are TH2 or THn, avoiding the TH3 label merge bug in ROOT 6.36.
• Add missing BC table to process functions for event selecton which needs it to get the runnumber via collision -> bc -> runnumber

twoPhotonMcTask:

• hOrigin: disable axis extension on the labelled z axis. TH1Merger::LabelMerge in ROOT 6.36 reads the z labels from the y axis, so merged content ended up in a new, unlabelled bin.
• Leg matching diagnostics: evaluate the primary-match status only for legs that pass the global track selection (rebuilt from the V0Legs columns); other legs get a new "not a global track" status instead of inflating "no match stored".

Co-Authored-By: Claude Opus 5.5

Fix missing BC table in process functions

[github-actions]

O2 linter results: ❌ 0 errors, ⚠️ 6 warnings, 🔕 25 disabled

[mhemmer-cern]

Did locally build and run, without any code check issues, so let's hope online its the same.

[alibuild]

Auto-approving on behalf of @mhemmer-cern.