feat: parallel processing

[kstruempf]

Implements #7

[Copilot]

Pull Request Overview

This PR introduces parallel processing support for the Monte Carlo Tree Search (MCTS) algorithm by adding a new method, search_parallel, and updating tests and documentation accordingly.

• Added search_parallel functionality for parallel rollouts.
• Updated tests to verify parallel processing behavior and error handling for unsupported configurations.
• Updated README to include instructions on using parallel processing.

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
tests/test_mcts.py	Adds new tests for parallel search and NotImplementedError case
mcts/searcher/mcts.py	Implements search_parallel and related multiprocessing changes
README.md	Documents the parallel processing feature

Comments suppressed due to low confidence (1)

mcts/searcher/mcts.py:69

• [nitpick] Consider renaming the method 'all_child_have_at_least_one_visit' to something more grammatically clear such as 'all_children_visited' to improve readability.

    def all_child_have_at_least_one_visit(self,) -> bool:

[Copilot]

The 'copy' module is imported but not used. Removing unused imports can improve code clarity and maintainability.

Suggested change

import copy

[Copilot]

[nitpick] Consider removing commented-out alternative code to reduce clutter and make the current implementation clearer.

Suggested change

	#processes.append(Process(target=self.execute_rollout_parallel,args=(index,root_shared,return_dict)))
	for p in processes: p.start()
	for p in processes: p.join()
	for index, reward in return_dict.items():
	#for index, rez in return_dict.items():
	self.backpropagate(input_nodes[index], reward)
	#node, reward = rez
	#self.backpropagate(node, reward)
	best_child = self.get_best_child(self.root, 0)
	#best_child = self.get_best_child(root_shared, 0)
	action = (action for action, node in self.root.children.items() if node is best_child).__next__()
	#action = (action for action, node in root_shared.children.items() if node is best_child).__next__()
	for p in processes: p.start()
	for p in processes: p.join()
	for index, reward in return_dict.items():
	self.backpropagate(input_nodes[index], reward)
	best_child = self.get_best_child(self.root, 0)
	action = (action for action, node in self.root.children.items() if node is best_child).__next__()