42.python 进程间通信Queue/Pipe

猿说编程[Python和C]

修改于 2021-04-19 14:50:03

3K0

修改于 2021-04-19 14:50:03

一.前言

1.在前一篇文章 python进程Process与线程threading区别 中讲到线程threading共享内存地址，进程与进程Peocess之间相互独立，互不影响（相当于深拷贝）；

2.在线程间通信的时候可以使用Queue模块完成，进程间通信也可以通过Queue完成，但是此Queue并非线程的Queue，进程间通信Queue是将数据 pickle 后传给另一个进程的 Queue，用于父进程与子进程之间的通信或同一父进程的子进程之间通信；

使用Queue线程间通信：

#导入线程相关模块
import threading
import queue  
 
q = queue.Queue()

使用Queue进程间通信，适用于多个进程之间通信：

# 导入进程相关模块
from multiprocessing import Process
from multiprocessing import Queue 
 
q = Queue()

使用Pipe进程间通信，适用于两个进程之间通信（一对一）：

# 导入进程相关模块 
from multiprocessing import Process 
from multiprocessing import Pipe 
 
pipe = Pipe()

二.python进程间通信Queue/Pipe使用

python提供了多种进程通信的方式，主要Queue和Pipe这两种方式，Queue用于多个进程间实现通信，Pipe用于两个进程的通信；

1.使用Queue进程间通信，Queue包含两个方法：

put()：以插入数据到队列中，他还有两个可选参数：blocked和timeout。详情自行百度
get()：从队列读取并且删除一个元素。同样，他还有两个可选参数：blocked和timeout。详情自行百度

# !usr/bin/env python
# -*- coding:utf-8 _*-
"""
@Author:何以解忧
@Blog(个人博客地址): https://www.codersrc.com/
 
@File:python_process_queue.py
@Time:2019/12/21 21:25
 
@Motto:不积跬步无以至千里，不积小流无以成江海，程序人生的精彩需要坚持不懈地积累！
"""
 
from multiprocessing import Process
from multiprocessing import Queue
import os,time,random
 
#写数据进程执行的代码
def proc_write(q,urls):
    print ('Process is write....')
    for url in urls:
        q.put(url)
        print ('put %s to queue... ' %url)
        time.sleep(random.random())
 
#读数据进程的代码
def proc_read(q):
    print('Process is reading...')
    while True:
        url = q.get(True)
        print('Get %s from queue' %url)
 
if __name__ == '__main__':
    #父进程创建Queue，并传给各个子进程
    q = Queue()
    proc_write1 = Process(target=proc_write,args=(q,['url_1','url_2','url_3']))
    proc_write2 = Process(target=proc_write,args=(q,['url_4','url_5','url_6']))
    proc_reader = Process(target=proc_read,args=(q,))
    #启动子进程，写入
    proc_write1.start()
    proc_write2.start()
 
    proc_reader.start()
    #等待proc_write1结束
    proc_write1.join()
    proc_write2.join()
    #proc_raader进程是死循环，强制结束
    proc_reader.terminate()
    print("mian")

输出结果：

Process is write....
put url_1 to queue... 
Process is write....
put url_4 to queue... 
Process is reading...
Get url_1 from queue
Get url_4 from queue
put url_5 to queue... 
Get url_5 from queue
put url_2 to queue... 
Get url_2 from queue
put url_3 to queue... 
Get url_3 from queue
put url_6 to queue... 
Get url_6 from queue
mian

2.使用Pipe进程间通信

Pipe常用于两个进程，两个进程分别位于管道的两端 * Pipe方法返回（conn1,conn2）代表一个管道的两个端，Pipe方法有duplex参数，默认为True，即全双工模式，若为FALSE，conn1只负责接收信息，conn2负责发送,Pipe同样也包含两个方法：

send() : 发送信息;

recv() : 接收信息;

from multiprocessing import Process
from multiprocessing import Pipe
import os,time,random
#写数据进程执行的代码
def proc_send(pipe,urls):
    #print 'Process is write....'
    for url in urls:
 
        print ('Process is send :%s' %url)
        pipe.send(url)
        time.sleep(random.random())
 
#读数据进程的代码
def proc_recv(pipe):
    while True:
        print('Process rev:%s' %pipe.recv())
        time.sleep(random.random())
 
if __name__ == '__main__':
    #父进程创建pipe，并传给各个子进程
    pipe = Pipe()
    p1 = Process(target=proc_send,args=(pipe[0],['url_'+str(i) for i in range(10) ]))
    p2 = Process(target=proc_recv,args=(pipe[1],))
    #启动子进程，写入
    p1.start()
    p2.start()
 
    p1.join()
    p2.terminate()
    print("mian")

输出结果：

Process is send :url_0
Process rev:url_0
Process is send :url_1
Process rev:url_1
Process is send :url_2
Process rev:url_2
Process is send :url_3
Process rev:url_3
Process is send :url_4
Process rev:url_4
Process is send :url_5
Process is send :url_6
Process is send :url_7
Process rev:url_5
Process is send :url_8
Process is send :url_9
Process rev:url_6
mian

三.测试queue.Queue来完成进程间通信能否成功？

当然我们也可以尝试使用线程threading的Queue是否能完成线程间通信，示例代码如下：

from multiprocessing import Process
# from multiprocessing import Queue     # 进程间通信Queue，两者不要混淆
import queue                            # 线程间通信queue.Queue，两者不要混淆
import time
 
def p_put(q,*args):
    q.put(args)
    print('Has put %s' % args)
 
 
def p_get(q,*args):
    print('%s wait to get...' % args)
 
    print(q.get())
    print('%s got it' % args)
 
 
 
 
if __name__ == "__main__":
    q = queue.Queue()
    p1 = Process(target=p_put, args=(q,'p1', ))
    p2 = Process(target=p_get, args=(q,'p2', ))
    p1.start()
    p2.start()

直接异常报错：

Traceback (most recent call last):
  File "E:/Project/python_project/untitled10/123.py", line 38, in <module>
    p1.start()
  File "G:\ProgramData\Anaconda3\lib\multiprocessing\process.py", line 105, in start
    self._popen = self._Popen(self)
  File "G:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 223, in _Popen
    return _default_context.get_context().Process._Popen(process_obj)
  File "G:\ProgramData\Anaconda3\lib\multiprocessing\context.py", line 322, in _Popen
    return Popen(process_obj)
  File "G:\ProgramData\Anaconda3\lib\multiprocessing\popen_spawn_win32.py", line 65, in __init__
    reduction.dump(process_obj, to_child)
  File "G:\ProgramData\Anaconda3\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: can't pickle _thread.lock objects