Python多处理池队列通信

Question

我正在尝试实现一个并行运行并通过队列进行通信的两个进程池。

我们的目标是具有通过使用队列传递消息到读取器处理的写入处理。

每个过程都在终端上打印反馈以获得反馈。

这是代码：

#!/usr/bin/env python

import os
import time
import multiprocessing as mp
import Queue

def writer(queue):
    pid = os.getpid()
    for i in range(1,4):
        msg = i
        print "### writer ", pid, " -> ", msg
        queue.put(msg)
        time.sleep(1)
        msg = 'Done'
    print '### '+msg
    queue.put(msg)

def reader(queue):
    pid = os.getpid()
    time.sleep(0.5)
    while True:
        print "--- reader ", pid, " -> ",
        msg = queue.get()
        print msg
        if msg == 'Done':
            break

if __name__ == "__main__":
    print "Initialize the experiment PID: ", os.getpid()
    mp.freeze_support()

    queue = mp.Queue()

    pool = mp.Pool()
    pool.apply_async(writer, (queue)) 
    pool.apply_async(reader, (queue))

    pool.close()
    pool.join()

我期待的输出应该是这样的：

Initialize the experiment PID: 2341
writer 2342 -> 1
reader 2343 -> 1
writer 2342 -> 2
reader 2343 -> 2
writer 2342 -> 3
reader 2343 -> 3
Done

但是我只得到了这条线：

Initialize the experiment PID: 2341

然后脚本退出。

在通过队列进行通信的池中实现两个进程的进程间通信的正确方法是什么？

Answer 1

我使用mp.Manager().Queue()作为队列，因为我们无法直接传递Queue 。 尝试直接使用Queue导致异常但由于我们使用apply_async 。

我将您的代码更新为：

#!/usr/bin/env python

import os
import time
import multiprocessing as mp
import Queue

def writer(queue):
    pid = os.getpid()
    for i in range(1,4):
        msg = i
        print "### writer ", pid, " -> ", msg
        queue.put(msg)
        time.sleep(1)
        msg = 'Done'
    print '### '+msg
    queue.put(msg)

def reader(queue):
    pid = os.getpid()
    time.sleep(0.5)
    while True:
        print "--- reader ", pid, " -> ",
        msg = queue.get()
        print msg
        if msg == 'Done':
            break

if __name__ == "__main__":
    print "Initialize the experiment PID: ", os.getpid()
    manager = mp.Manager()

    queue = manager.Queue()

    pool = mp.Pool()
    pool.apply_async(writer, (queue,))
    pool.apply_async(reader, (queue,))

    pool.close()
    pool.join()

我得到了这个输出：

Initialize the experiment PID:  46182
### writer  46210  ->  1
--- reader  46211  ->  1
### writer  46210  ->  2
--- reader  46211  ->  2
### writer  46210  ->  3
--- reader  46211  ->  3
### Done
--- reader  46211  ->  Done

我相信这是你的期望。