Integrating optical photons into restG4

[mariajmz]

This PR is related to the Issue #123.

I added G4OpticalPhysics list to PhysicsList and included the optical surfaces headers to the DetectorConstruction class as first stage to include optical photon simulations in restG4.
Also, I needed to update TRestPhysicsList::PhysicsListExists, which I have done in PR rest-for-physics/geant4lib#119.
So, to include optical physics in a restG4 simulation, just add <physicsList name="G4OpticalPhysics"></physicsList>

I have checked the simulation of Optical Boundary processes in restG4 (ie, reflection, refraction, absorption) with two basic geometries:

• The first, a gold layer as an xray mirror. Assigning refraction indexes and optical surfaces (dielectric_metal) via gdml.

An example of an absorbed photon:

And a reflected photon:

• The second, a veto geometry with dielectric_dielectric optical surfaces:
  This geometry is based on the one that can be found in https://github.com/iaxo/iaxo-geometry/blob/master/gdml/IAXO-D1/Veto.gdml
  Just an example of an event.
  

In this case, the red volume doesn't have refractive index so the optical photons are killed when they reach it.

The optical photon generator is configured via rml just using <source particle="opticalphoton">. Also, G4OpticalPhoton have polarization. This has not been implemented so, while running the restG4 simulation with debug verbose level, you will see this message:

As Optical Boundary processes are classified as 'Transportation' in GEANT4, all of the hits have processName 'Transportation'. Maybe it will be interesting to save more specific information of the hit (if it is refraction, reflection or absorption in dielectric_metal surfaces). As I undestood and did in GEANT4, for Optical Boundary processes this needs to be done differently, checking G4Step::GetPostStepPoint()::GetStepStatus(). Maybe @lobis,@jgalan have some ideas about this?

I haven't include the gdml and rml of these examples into this PR, just tell me in case you believe it's necessary.

[lobis]

Looks good to me!

This is a good first step towards full-stack optical simulations.

Regarding storing additional step information instead of just Transportation, I think the process could instead be replaced by the more informative Reflection, etc. Not sure if this has some limitations.

I would like a more general solution to this problem. For example for hadronic processes such as nCapture (neutron capture) it's interesting to know which isotope actually captured the particle. This information is also available in Geant4, however there is no place to store it in the current layout for the event data. We could just add a new field to the hits for each one of this needs but it would be too ineficient as it would be mostly empty for all hits (root compresses de data on disk, but this compression is not perfect). But this is a different issue, for now, storing it in place of the process should be enough, I guess.

By the way the link to the Geant4Lib PR is broken.

Also currently as it stands (I think) this PR is linked to the "big issue" #123 which has a larger scope, if this is merged, that issue will be closed and we don't want that, not until more features are added, examples and documentation included, etc.

[mariajmz]

Thank you @lobis! I fixed the link to the geant4lib PR.

So should I unlink this PR to issue #123?

[lobis]

So should I unlink this PR to issue #123?

Yes, it's enough to mention it. Linking the issue will close it when this PR is done and we don't want that.