[Bug] Upwards flowing systems still pour from the top

[ZoeRichter]

Describe the bug.

Problem (bug?) - Ghastly has a flag for upward flow, and that flag correctly switches the "gravity" vector to the upwards direction, but in the functions that determine the bounding box and pour regions in LAMMPS, it still does this from the top, which will not work. Upwards flow will also require adding a term for having the top surface of a core element be closed (so the pebbles don't float away).

What is the expected behavior?

It should have an insertion region at the bottom, and bounds that are expanded in the -z direction accordingly. The core_intake region being at a more positive z-direction than core_main and core_outtake is not a requirement, so the current method of "removing" the intake region to make space while pouring should still work. This also means that even when using upwards flow, the intake zone can still actually be the intake (instead of some weird interaction where the core_outtake is actually the intake because it happens to be the lowest in the z-direction).

Additional context.

Core filling currently works for downward flows.

How can this issue be closed?

Adding the above fixes so that core filling can happen for upward-flowing systems.

[ZoeRichter]

This is also where you can address the PR comment from #17 on if the scaling factors for the box bounds should be hard-coded.