Geometry improvements

Reporting the issue from [WenjieWu-Sci/FLArE#53](https://github.com/WenjieWu-Sci/FLArE/issues/53).

The current cavern geometry in the simulation needs to be revisited following the latest facility design as well as the almost-finalized integration of the detectors in the hall. Reference document is the [July 2024 PBC note](https://cds.cern.ch/record/2904086).

Cavern:

- [ ] Size and shape of the cavern. Detectors stay still at `(0,0)`, but walls need to be moved relative to them.
- [x] Length of the cavern. The detector side of the hall is roughly the same length as before, but shifted 10m away from the IP. This has consequences for the background muon flux: potentially add 10 m of rock in front and propagate the muon flux through that.
- [x] #9

Detectors:
- [x] Simplify nested volumes: avoid assemblies that complicate the GDML output file.
- [ ] Detector placements in the cavern: size and "final" position.
- [ ] Update default geometry (FORMOSA now at the end of the hall)
- [ ] Update alternative geometry options with latests developments (FASER2 magnet, etc.)
- [ ] #4
- [x] #6