Rework of Particle class, and Particle data creation

[VeckoTheGecko]

Changes:

• Cleares up Variable and Particle classes, making the dictinction between a ParticleClass (a type of Particle), and a KernelParticle (an accessor class used to read and set particle data in the arrays). This clarification allows for clearer abstractions and consolidation of responsibility (helping with testing).
• Removed lonlatdepth_dtype from particleset init
  • Created a public helper function get_default_particle(dtype: np.float32 | np.float64) which allows users to create particles with the right precision from the beginning (rather than redefining after the fact).
• Consolidate particle data creation to a function create_particle_data that works from the particle class, ngrids, time_interval, initial
• Remove MPI related code from ParticleFile, and some cleanup of the ParticleFile class (particlefile remains untested at the moment - more changes will come in a future PR)

Would be good to also discuss variable naming - let me know if you have any suggestions.

• Chose the correct base branch (v4-dev for v4 changes)
• Fixes #
• Added tests

[erikvansebille]

So the code (see below) to temporarily set time to np.nat until we know the sign of dt in execute() stays as it is?
https://github.com/OceanParcels/Parcels/blob/30aec22d2939e285bfbd44f6617e36019002f6d3/parcels/particleset.py#L100-L103
and
https://github.com/OceanParcels/Parcels/blob/30aec22d2939e285bfbd44f6617e36019002f6d3/parcels/particleset.py#L536-L540

That would be fine with me; but it could perhaps be a bit more transparently coded?

[VeckoTheGecko]

I see, something to keep in mind when we refactor the particleset execute loop

[VeckoTheGecko]

@erikvansebille after syncing this branch with v4-dev, I am getting some strange infinite loop bugs in the following tests to do with the RK45 kernel ..... I'm not really sure how to go about solving them at the moment, perhaps if you have a second tomorrow we can look through together since you're familiar with the rk45 kernel? (or can pytest.skip them for now until we have time to work through this. I think the pfile stuff needs to remain my priority for now

pytest 'tests/v4/test_advection.py::test_moving_eddy[RK45-1e-05]' 'tests/v4/test_advection.py::test_decaying_moving_eddy[RK45-1e-05]' 'tests/v4/test_advection.py::test_gyre_flowfields[stommel_gyre-A-RK45-1]' 'tests/v4/test_advection.py::test_gyre_flowfields[peninsula-A-RK45-1]'