异步io进程代理

如何解决异步io进程代理

我正在寻找一种方法或库来与不同进程的对象进行异步通信。这是一个示例，我想将 Solver 放在不同的进程上，但仍然能够经常与求解器通信。这里重要的是，Solver 的初始化非常昂贵，因此每个进程只能执行一次，并且应该可以用于多次 solve 调用。

class Solver:
    def __init__(self):
        # Initializing this object is very expensive so it should be done only once
        super_expensive_initialization(self)

    async def solve(self,task):
        # This function uses asyncio to do a lot of things in parallel that will overall 
        # use this process very well
        solution = asyncio.gather([self._highly_asyncronous_solve(task) for i in range(200)])

        return await solution


    async def _highly_asyncronous_solve(self,a):
        ...


# <<< This part is what I am looking a solution for
my_solver_proxy_a = run_in_process_and_return_proxy(lambda: Solver(config_a))
my_solver_proxyb = run_in_process_and_return_proxy(lambda: Solver(config_b))
# >>>

tasks = [task_1,task_2,...]

schedule_tasks_in_parallel(tasks,[my_solver_proxy_a,my_solver_proxyb])

我希望你能理解我想要实现的目标。 multiprocessing.Manager 似乎与我想要做的类似，但我需要通过代理进行非阻塞调用。