[Feature] Added Pure Random Algo to OrderDisorderedStructureTransformation

[jmmshn]

Added Pure Random Algo to OrderDisorderedStructureTransformation

The current implementation assume some oxidation and tries to evaluate and Ewald energy to get the ordered representation with the lowest energy.

pymatgen/src/pymatgen/transformations/standard_transformations.py

Lines 558 to 559 in da607e8

	matrix = EwaldSummation(struct).total_energy_matrix
	ewald_m = EwaldMinimizer(matrix, manipulations, n_to_return, self.algo)

For large cells this has some disadvantages:

matrix = EwaldSummation(struct).total_energy_matrix  # <<- This can run out of memory for cells with ~2000 atoms
ewald_m = EwaldMinimizer(matrix, manipulations, n_to_return, self.algo) # <<- This takes hours for cells with ~200 atoms

We need some way to indiscriminately generate random structures without Ewald evaluation. Since for larger cells the value proposition of explicitly enumerating all the ordered representations and ranking them is dubious.

This PR adds a purely random sampling method that will sample all allowed structures with equal probability.

Major changes:

• Added ALGO_RANDOM=-1 for purely random sampling
• Allows user to define the occupation tolerance occ_tol for if a ordered structure is valid representation of a disordered structure

Checklist

• Google format doc strings added. Check with ruff.
• Type annotations included. Check with mypy.
• Tests added for new features/fixes.
• If applicable, new classes/functions/modules have duecredit @due.dcite decorators to reference relevant papers by DOI (example)

[shyuep]

Thanks for the contribution.