Improve random state management

[tautomer]

This PR addresses the random state management issues in #80 to improve reproducibility and eliminate repeated noise patterns or initial optimizer conditions during a campaign.

Implementation details:

• Introduced a centralized RNG initialization that occurs once per campaign, controlled by a single seed.
• Updated the code to pass RNG generator instances explicitly whenever possible to components requiring randomness, avoiding global RNG state resets.
• Use a context manager for temporary global seed override if explicitly passing a generator is not possible, for example, model.fit and optimizer.suggest.
• Added fallback behavior to maintain reproducibility even when no seed is provided by initializing RNG states with system entropy but tracking them internally.
• Added a package level toggle obsidian.USE_OLD_RNG_CONTROL for backward compatibility.

Other changes

Some other improvements were made while this new RNG control was implemented.

1. Introduced direct optimizer options passthrough to model fitting routine in campaign.fit. You can now do

   campaign.fit({"max_attempts": 10, "multi_starts": True})

   to fit with model with 10 different restarts.

2. Introduced direct optimizer options passthrough to acquisition function optimization in optimizer.suggest

   campaign.optimizer.suggest(
       ...,
       optim_options={"method": "Powell", "maxiter": 100}
   )

   This will tell scipy.optimize.minimize to use Powell and maximum iteration of 100.

Current status

The following tests have been performed, both on single objective (modified from Simple single objective.ipynb)

1. The new code, with obsidian.USE_OLD_RNG_CONTROL=True produces the same results as the current main branch.
2. With new RNG control, two runs with the same initial random states achieve identical results with categorical parameters.
3. With new RNG control, two runs with the same initial random states achieve identical results without categorical parameters. (fixed in 1064d6f).

TODO:

1. save_state and load_state have been implemented and functions as they should, but they are not part of larger scope (i.e., campaign level) yet.
2. More extensive tests needed. (Help appreciated.)