Refactor ``clip_box`` logic to clarify inner workings

After speaking with @Manangka about #1744

The logic required for the clip_box method is now spread over several utilities. There is even some included in the imod/mf6/model.py, where it shouldn't be. Furthermore the data handling is convoluted now and it should be clarified in some way. I remember while working on #1744 that it took me a bit of time to find where and when processing steps took place and what was happening, which isn't a good sign.

The logic required is complex regardless, as there are quite some edge cases to handle, so we shouldn't expect such clarity that a new dev would understand every step at first glance, but some improvement is required.

What would help:

1. Move functions in imod/mf6/model.py to clearer namespace
2. Add description of algorithm to clarify inner workings
3. Refactor clip_box logic, possibly add some objects for internal use

Read #1744 first, then trace the clip_box method and related utilities, including functions currently in imod/mf6/model.py. Map where processing occurs before deciding on clearer namespaces or internal objects; done means the logic is easier to follow, its algorithm is documented, and the listed responsibilities are reorganized.