Added struphy-plots integration

[max-models]

Merge after #406

Integrate the optional plasma-plots package with Output, so creating an output handle automatically enables whole-run plots and diagnostics through out.plot and out.analysis, plus the .plasma accessors on xarray products. Also enable automatic post-processing under MPI and direct selection of particle products by their full catalog keys.

Install plasma-plots with: pip install plasma-plots or pip install "struphy[pproc]".

Solves the following issue(s):

Makes plasma-plots available directly from the Struphy output API and removes the need for a separate pproc() call before first accessing unprocessed products under MPI.

Core changes:

• Add the optional struphy[plots] extra, which installs plasma-plots. Output loads the package when installed; output reading remains available without it. Accessing plotting or analysis without the package gives installation guidance.
• Let evaluate() and pproc() choose parallel processing automatically when the current MPI job has the same number of ranks as the saved simulation, and more than one rank. Otherwise, rank 0 processes the run while the remaining ranks wait. An explicit parallel=True or False overrides the selection.
• Synchronize ranks before they read newly written products and serialize serial processing with a filesystem lock to avoid competing writers. Use POSIX record locks where available, with an exclusive-file fallback.
• Accept full particle-product keys such as out.evaluate("kinetic_ions/e1_v1_density/delta_f"), including coordinate selections. Reject combining a full key with dataset=.

Model-specific changes:

None.

Documentation changes:

Update the README, quickstart, user guide, and output API documentation with plasma-plots installation, accessor examples, and MPI processing behavior.