SimCalorimeterHit and CaloHitContribution: missing fields used in EIC production

[wdconinc]

Two points for discussion in calorimeter hit datatypes:

• SimCalorimeterHit does not have a time field (that info is only available in the CaloHitContributions, which are hidden behind a relation that is hard to navigate outside ROOT), however this information is needed for digitization and a fairly pertinent quantity for a SimCalorimeterHit. This adds the need for an expensive CaloHitContributions loop only to find the min(time). Other quantities technically already in CaloHitContributions (i.e. energy) are provided in SimCalorimeterHit.
• For Birks' constant saturation treatment outside of repeated geant4 running (and during digitization), CaloHitContribution would benefit from a length field to allow applying this correction. (https://en.wikipedia.org/wiki/Birks'_law). Our transition to edm4hep required us to ditch the digitization treatment of Birks' saturation.

[tmadlener]

I think we are in principle fairly open to these suggestions, and more detailed discussions can probably be done in a corresponding pull request (or several to keep things separated if necessary). The main thing that currently makes us a bit reluctant in merging these already is the missing schema evolution that we are still working on in podio.

[andresailer]

• From datamodel POV simple to add it
• Some physics questions behind it, e.g. electronics, timing. I.e. what do you actually fill into that field
  • Could also store the MC truth time
• Could also depend on simulation, e.g. do you need each Geant4 hit (digitial calorimeters)
• Not necessarily the lowest time of all CaloHitContributions
• LCIO also covers use case from detailed simulations for Calorimeter testbeam studies. In the linear collider world SimCalorimeterHit is not yet digitized, and just holds all data from the Geant4 simulation that migth be necessary for digitization (which is done at reconstruction time).