Consider using np.random.default_rng for self-contained randomness

[micimize]

Setting np.random.seed is alright, but IMO it would be better to add a seed + RNG to each generator using np.random.default_rng(seed) so that their RNG is self contained.

This is because a lot of other code mazelib gets integrated with may make use of the "global" rng, making reproducibility more challenging for depending projects.

[john-science]

That seem like a cool idea, sure.

I should probably read the docs there and play with the two, just to make sure. But from your description, I agree that would be a useful improvement.

I'm for it.

[cclauss]

I am not a user of Numpy... Would these be worth fixing?
% ruff check --select=NPY --output-format=concise

mazelib/generate/Kruskal.py:44:9: NPY002 Replace legacy `np.random.shuffle` call with `np.random.Generator`
mazelib/generate/MazeGenAlgo.py:57:9: NPY002 Replace legacy `np.random.shuffle` call with `np.random.Generator`
mazelib/mazelib.py:35:13: NPY002 Replace legacy `np.random.seed` call with `np.random.Generator`
mazelib/solve/MazeSolveAlgo.py:83:9: NPY002 Replace legacy `np.random.shuffle` call with `np.random.Generator`
mazelib/transmute/MazeTransmuteAlgo.py:64:9: NPY002 Replace legacy `np.random.shuffle` call with `np.random.Generator`
mazelib/transmute/MazeTransmuteAlgo.py:89:9: NPY002 Replace legacy `np.random.shuffle` call with `np.random.Generator`
Found 6 errors.