RecoParticleVertexAssociation: rename relation members for consistency between association

[fdplacido]

BEGINRELEASENOTES

• Rename last association to be consistent with all other associations.

ENDRELEASENOTES

This allows to use any association in a generic (templates) way, so all the methods will be called the same.

[vvolkl]

Makes sense to me.

[tmadlener]

Makes sense to me as well. Should we consider having more generic names than rec and sim then? I think all but this relation describe an association to an MC particle, so there it makes sense, but with the vertex it could be a bit confusing.

Depending on how many places this occurs in, one could also consider to have some if constexpr in place to handle the RecoParticleVertexAssociation just slightly different than the rest.

In any case this is likely a temporary solution until there is a more generic notion of relations/associations in podio, right?

[fdplacido]

Maybe we should rename all association elements to "from" and "to", which seems the most generic thing to put. Since those are part of a collection named "*Association" should be pretty clear what that refers to, no?

[tmadlener]

I think renaming to "to" and "from" could be done as a preparatory step before there is proper support for this in podio. The only concern I would have here is that it then becomes slightly harder to use these associations in "both directions"; e.g. in ILD workflows there is usually an association from MC to Reco and another one from Reco to MC with different weights. This is then used for example to calculate reconstruction efficiencies with one of those and fake rates with the other. But I don't think this would be a major showstopper at the moment.

[tmadlener]

This would not break I/O, but probably some code that still has rec and sim as relation names. I have no strong preference on whether this should go in before a new tag. We will have to fix the naming of the code that uses this in any case.

[tmadlener]

Can you quickly rebase, to get rid of the "out-of-date with the base branch"?

[tmadlener]

Suggested change

	- edm4hep::CalorimeterHit from // reference to the reconstructed hit
	- edm4hep::SimCalorimeterHit to // reference to the simulated hit
	- edm4hep::CalorimeterHit to // reference to the reconstructed hit
	- edm4hep::SimCalorimeterHit from // reference to the simulated hit

The name of the datatype suggests that the MC part is first, so I think it is more intuitive for it to be the "from" part of the association, right? A similar argument can be made for some of the other associations as well.

[hegner]

Outdated and decided not to pick up