Add Plaid in spack

[williampiat3]

Spack may be preferable for HPC clusters rather than uv/conda/pypi: it increases performance of librairies at the cost of reducing portability

ONGOING:

• Add all dependencies in spack-packages: only pycgns is missing, PR ongoing py-pycgns: new package spack/spack-packages#4888: Just waiting for PyCGNS developers to state on their involvment in the recipe: https://github.com/orgs/pyCGNS/discussions/159. Otherwise it already works
• Making a Plaid recipe: https://github.com/williampiat3/spack-packages/blob/Xkblas/repos/spack_repo/builtin/packages/py_plaid/package.py. Currently operational, waiting for py_cgns to be merged in spack-packages

TODO:

• For plaid 1.0.0 update setuptools and python default version to provide more flexibility to spack: see 📦 allow 3.14 builds + update setuptools #394. Still need to perform additionnal tests on release pipelines

[fbordeu]

for the moment Muscat is not fully available on python3.14 in conda-forge.
working on the solution.
conda-forge/muscat-split-feedstock#61