[PWGHF,PWGJE,Trigger] Divide candidate data model into skimming, reconstruction, aliases

[vkucera]

Move parts of PWGHF/DataModel/CandidateReconstructionTables.h into:

• PWGHF/DataModel/CandidateSkimmingTables.h
  • Tables and flags used in the track index skimming
• PWGHF/DataModel/AliasTables.h
  • Definitions of joined table types
• PWGHF/Core/DecayChannelsLegacy.h
  • Legacy MC flags that should be replaced.
• PWGHF/ALICE3/Core/DecayChannelsLegacy.h
  • Legacy MC flags used only for ALICE3.

The goal is to factorise header dependencies and to reduce recompilations when modifying parts of the data model.

[github-actions]

O2 linter results: ❌ 112 errors, ⚠️ 24 warnings, 🔕 28 disabled

[fgrosa]

I would refrain from using the term "skimming" to avoid confusion, since in ALICE skimming is referring to the software triggers (I know that we called "skimCreator" our HF vertex finder, but this would also not be a proper name, since we are not skimming anything, but actually adding information to the AO2Ds). I propose to use the TrackIndexSkimCreatorTables.h or something similar, to make it a bit more clear that it refers to the trackIndexSkimCreator task.

[vkucera]

• The name is neither AliceSkimmingTables.h nor SkimmingTables.h, it's CandidateSkimmingTables.h. I think that makes the context clear enough.
• Why do you say that we are not skimming anything? We are skimming the combinations of track indices which are our decay candidates.
• Is TrackIndexSkimmingTables.h fine?

[fgrosa]

I think it's not "skimming" because if it was it, it would imply a reduction of the information while in our case we add the information of candidates starting from tracks (so it's vertex finding, similar to the SVertexer in the reconstruction). In any case I think that TrackIndexSkimmingTables makes it less ambiguous, thanks!

[vkucera]

• Renamed
• I always understood the "track index skimming" in the meaning of to read or consider something quickly in order to understand the main points, without studying it in detail, i.e. to scan the track table and save the decay candidate combinations without doing the full candidate reconstruction. I don't know what other meaning, that involves reduction of information, people have in mind.

[vkucera]

@fgrosa Good to merge?

[fgrosa]

@vkucera good for me, thanks!