[mattiasotgia]

Overview and references

This PR re-work the stage1 paths for Pandora + NuGraph2 + Pandora and SPINE reconstruction.
Details are in SBN DocDB 50214, and where discussed at the ICARUS sim/reco and meeting.

The idea is essentially to split the reconstruction chain into three paths:

• "vanilla" Pandora, which takes as input the hits produced using the updated hit finder tuning (Tuned hit finder parameters #912);
• Pandora + NG2 + Pandora, which take as input the hits produced using the updated hit finder tuning as well (but using cluster3D which is required to have SpacePoints);
• SPINE, which is using the "old" not-tuned hit finder parameters.

Key changes

Stage1 definitions (stage1_icarus_defs.fcl)

• Splits Cluster3D into two chains:
  • cluster3DCryoW/E is the SPINE version and reads the gaushitPT2dTPC* hits.
  • New pandoraCluster3DCryoW/E is for Pandora + NG2 + Pandora and reads the plain gaushit2dTPC* hits.
• Moves the Cluster3D settings into inline tables, replacing the old per-field overrides.
• The new modules are added to the filter sequences, to new icarus_reco_pandoraCluster3DCryo{W,E} sequences and to the Gauss2D sequences.
• The gaushitPT2dTPC* hit finders now take a shared spine_gaushit_parameters block instead of just LongMaxHits. It adds back the default values for the hit finder algorithm (MaxMultiHit: 5, hi2NDF: 500., LongPulseWidth: [10, 10, 10]) as well as the default for the candidate-hit tool (RoiThreshold: 5. and NextADCThreshold: 0).
• The default pandoraGausCryo* input is now pandoraCluster3DCryo*.

Stage1 data and MC job configs

• The 1D configs (*_1d_*) also set the gaushit1d* tags for pandoraCluster3D*, so Pandora with 1D signal processing uses them too.
• The Pandora-only configs (stage1_run2_icarus.fcl and its _MC version) point pandoraGausCryo* back to combineHits*.
• The NuGraph configs use pandoraCluster3D* for pandoraGausCryo*.

NuGraph (nugraph_icarus.fcl)

SpsLabel, ClusterLabel and the hitInput labels for the slice-hit, NCC-slice and multi-slice modules now use pandoraCluster3DCryo*, aligned with the Pandora-specific Cluster3D instance.

Notes

This PR overrides (and expands) the previous PR

• Reintroduced old hit finder parameters for SPINE reconstruction #934

[cerati]

looks good. There is still the opportunity to clean up a bit spine_gaushit_parameters but that can be done once those parameters are finalized