PMT timing corrections should not be applied on simulation

[PetrilloAtWork]

The code in PR #440 applies a time correction to the reconstructed PMT hits read from a database according to the run their event belongs to.
In case of simulation, this correction should be disabled completely, the service interfacing to that database should not be used nor required.
Ideally, that service itself should be able also to return a neutral correction when dealing with Monte Carlo events.

[miquelnebot]

@PetrilloAtWork what's the status of this one?

[PetrilloAtWork]

Still to-do.

[PetrilloAtWork]

It should be pointed out that the MC and data workflows are different, and currently Stage0 is written so that Monte Carlo jobs generate optical hits in ophit directly with the hit finding algorithm, while data jobs generate optical hits in ophituncorrected and then apply the correction creating ophit.
In my opinion this is also the best way to avoid unnecessary dependencies (the MC job does not need the database service to be configured).
Nevertheless it would be good to neuter the timing correction module to just copy MC hits.

[mvicenzi]

#892 will resolve this issue. Timing corrections will be fully disabled in the legacy MC flow; while in the new MC-Overlays flow they will be used in both detsim and stage0 to properly simulate the timing delays and then correct for them after overlaying the data waveforms.