Proposal for a Reorganization of Reco Category

[ruse-traveler]

One of the goals of the July 2024 UGM Reconstruction Workfest was to map out our reconstruction chain and develop a more coherent picture of our reconstruction's overall flow. For a summary of the discussion, see the corresponding summary talk here.

After "canonical reconstruction" -- reconstructing tracks, calorimeter clusters, and PID hypotheses -- is complete, there are still several stages of reconstruction where these types are combined into objects closer to concrete physical objects (e.g. combining tracks into vertices, particle flow, etc.). The proposed periodization of the above talk groups these latter stages into 2 periods, "early" and "late syntheses."

Right now, all algorithms that fall into the early and late syntheses are collected into the reco category of our reconstruction. This has lead to this category becoming fairly cluttered, collecting several different categories of reconstruction such as vertexing, jet finding, particle transformations, and more. And in turn, this has also led to the corresponding plugin becoming dense and tough to read since there are multiple strands of reconstruction happening simultaneously...

By divvying up the existing reco category into smaller, more focused categories might be beneficial: it might make the overall structure of our reconstruction more clear, help users more clearly identify where to make changes, and keep plugins concise.

Describe the solution you'd like

I presented a proposal for a reorganization during the Sep. 16th Reconstruction WG meeting which suggested to divide reco into 3 categories -- particle, jet, and kinematic -- and then move some algorithms into the existing meta category. I should have also included a category for vertexing (vertex) since vertexing is also currently grouped into reco.

So in total, the specific algorithms which would fall under each category are:

particle
  - ElectronReconstruction
  - ScatteredElectronEMinusPz
  - ScatteredElectronTruth
  - FarForwardNeutronReconstruction
  - ParticleFlow (when ready)
  - MatchClusters (should rename once rest of PF workflow is implemented)
  - TransformBreitFrame
  - UndoAfterBurner
jet
  - JetReconstruction
kinematic
  - HadronicFinalState
  - InclusiveKinematics{DA,ESigma,Electron,JB,Sigma,Truth}
vertex
  - PrimaryVertices
  - SecondaryVertices (when ready)
meta
  - Charged{MC,Reconstructed}ParticleSelector (could be subsumed by CollectionSubdivider)
  - MC2SmearedParticle (should be renamed since it doesn't smear particles)

Describe alternatives you've considered

This definitely isn't the only possible division! For example, we could also just split reco into an early and late category...