Python Custom Process Pool Instance Analysis [Producer and Consumer Model Issues]

高洛峰
Release: 2017-02-22 16:14:22
Original
1237 people have browsed it

This article analyzes the Python custom process pool through examples. Share it with everyone for your reference, the details are as follows:

The code explains everything:

#encoding=utf-8
#author: walker
#date: 2014-05-21
#function: 自定义进程池遍历目录下文件
from multiprocessing import Process, Queue, Lock
import time, os
#消费者
class Consumer(Process):
  def __init__(self, queue, ioLock):
    super(Consumer, self).__init__()
    self.queue = queue
    self.ioLock = ioLock
  def run(self):
    while True:
      task = self.queue.get()  #队列中无任务时,会阻塞进程
      if isinstance(task, str) and task == 'quit':
        break;
      time.sleep(1)  #假定任务处理需要1秒钟
      self.ioLock.acquire()
      print( str(os.getpid()) + ' ' + task)
      self.ioLock.release()
    self.ioLock.acquire()
    print 'Bye-bye'
    self.ioLock.release()
#生产者
def Producer():
  queue = Queue()  #这个队列是进程/线程安全的
  ioLock = Lock()
  subNum = 4  #子进程数量
  workers = build_worker_pool(queue, ioLock, subNum)
  start_time = time.time()
  for parent, dirnames, filenames in os.walk(r'D:\test'):
    for filename in filenames:
      queue.put(filename)
      ioLock.acquire()
      print('qsize:' + str(queue.qsize()))
      ioLock.release()
      while queue.qsize() > subNum * 10: #控制队列中任务数量
        time.sleep(1)
  for worker in workers:
    queue.put('quit')
  for worker in workers:
    worker.join()
  ioLock.acquire()
  print('Done! Time taken: {}'.format(time.time() - start_time))
  ioLock.release()
#创建进程池
def build_worker_pool(queue, ioLock, size):
  workers = []
  for _ in range(size):
    worker = Consumer(queue, ioLock)
    worker.start()
    workers.append(worker)
  return workers
if __name__ == '__main__':
  Producer()
Copy after login

ps:

self.ioLock.acquire()
...
self.ioLock.release()
Copy after login

Available:

with self.ioLock:
  ...
Copy after login

Alternative.

Another fun example:

#encoding=utf-8
#author: walker
#date: 2016-01-06
#function: 一个多进程的好玩例子
import os, sys, time
from multiprocessing import Pool
cur_dir_fullpath = os.path.dirname(os.path.abspath(__file__))
g_List = ['a']
#修改全局变量g_List
def ModifyDict_1():
  global g_List
  g_List.append('b')
#修改全局变量g_List
def ModifyDict_2():
  global g_List
  g_List.append('c')
#处理一个
def ProcOne(num):
  print('ProcOne ' + str(num) + ', g_List:' + repr(g_List))
#处理所有
def ProcAll():
  pool = Pool(processes = 4)
  for i in range(1, 20):
    #ProcOne(i)
    #pool.apply(ProcOne, (i,))
    pool.apply_async(ProcOne, (i,))
  pool.close()
  pool.join()
ModifyDict_1() #修改全局变量g_List
if __name__ == '__main__':
  ModifyDict_2() #修改全局变量g_List
  print('In main g_List :' + repr(g_List))
  ProcAll()
Copy after login

The result of running under Windows 7:

λ python3 demo.py
In main g_List :['a', 'b', 'c']
ProcOne 1, g_List:['a', 'b']
ProcOne 2, g_List:['a', 'b']
ProcOne 3, g_List:['a', 'b']
ProcOne 4, g_List:['a', 'b']
ProcOne 5, g_List:['a', 'b']
ProcOne 6, g_List:['a', 'b']
ProcOne 7, g_List:['a', 'b']
ProcOne 8, g_List:['a', 'b']
ProcOne 9, g_List:['a', 'b']
ProcOne 10, g_List:['a', 'b']
ProcOne 11, g_List:['a', 'b']
ProcOne 12, g_List:['a', 'b']
ProcOne 13, g_List:['a', 'b']
ProcOne 14, g_List:['a', 'b']
ProcOne 15, g_List:['a', 'b']
ProcOne 16, g_List:['a', 'b']
ProcOne 17, g_List:['a', 'b']
ProcOne 18, g_List:['a', 'b']
ProcOne 19, g_List:['a', 'b']
Copy after login

The result of running under Ubuntu 14.04:

In main g_List :['a', 'b', 'c']
ProcOne 1, g_List:['a', 'b', 'c']
ProcOne 2, g_List:['a', 'b', 'c']
ProcOne 3, g_List:['a', 'b', 'c']
ProcOne 5, g_List:['a', 'b', 'c']
ProcOne 4, g_List:['a', 'b', 'c']
ProcOne 8, g_List:['a', 'b', 'c']
ProcOne 9, g_List:['a', 'b', 'c']
ProcOne 7, g_List:['a', 'b', 'c']
ProcOne 11, g_List:['a', 'b', 'c']
ProcOne 6, g_List:['a', 'b', 'c']
ProcOne 12, g_List:['a', 'b', 'c']
ProcOne 13, g_List:['a', 'b', 'c']
ProcOne 10, g_List:['a', 'b', 'c']
ProcOne 14, g_List:['a', 'b', 'c']
ProcOne 15, g_List:['a', 'b', 'c']
ProcOne 16, g_List:['a', 'b', 'c']
ProcOne 17, g_List:['a', 'b', 'c']
ProcOne 18, g_List:['a', 'b', 'c']
ProcOne 19, g_List:['a', 'b', 'c']
Copy after login

It can be seen that the second modification under Windows 7 was not successful, but the modification under Ubuntu was successful. According to limodou, the author of uliweb, the reason is that under Windows, the child process is implemented by restarting; under Linux, it is implemented by fork.

For more Python custom process pool instance analysis [producer and consumer model issues] related articles, please pay attention to the PHP Chinese website!


Related labels:
source:php.cn
Statement of this Website
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn
Popular Tutorials
More>
Latest Downloads
More>
Web Effects
Website Source Code
Website Materials
Front End Template
About us Disclaimer Sitemap
php.cn:Public welfare online PHP training,Help PHP learners grow quickly!