Make usecases runnable under pytest

[JCGoran]

This is based on #1294 so I don't need to solve future merge conflicts.

To outline how it works:

• all Python files need to have the prefix test (hence simulate.py becomes test_sim.py or whatever) so pytest can pick them up
• all Python objects we want to test using pytest need to also have the prefix test
• the usecases dir needs an __init__.py (because pytest makes it basically impossible to use something like fixtures) since that's where we define a common function for compiling mod files, compile_modfile
• due to Python's import mechanics, each usecase itself needs an __init__.py (can be empty, but the file must exist) so we can actually import the compile_modfile function
• we parametrize two cases, one with NOCMODL (default), and one with NMODL (for which we need to specify a path to the executable, hence the arg is either None or some Path instance), and run those in turn (as a discussion point, should the argument be just nmodl/nocmodl, and then it looks up the location dynamically? Then we can parametrize the function with just ['nocmodl', 'nmodl'] rather than the more verbose [None, shutil.which("nmodl")] which requires remembering to import the shutil module)
• if the file is compiled successfully, we use load_mechanisms to actually make it visible for NEURON (otherwise insert fails)

[bbpbuildbot]

Logfiles from GitLab pipeline #215181 (:white_check_mark:) have been uploaded here!

Status and direct links:

• :white_check_mark: spack_setup
• :white_check_mark: build:intel
• :white_check_mark: build:nvhpc
• :white_check_mark: test:intel
• :white_check_mark: test:nvhpc

[1uc]

As discussed offline, the simplicity of simply having a mod file and an associated script is valuable because it mixes well with many important debugging tools. It also allows the usecases to double as examples. Finally, we often test that things are in the correct global state, before running "init". Therefore, we need to run tests in separate processes.