当前位置:首页 >> 脚本专栏

Python RabbitMQ消息队列实现rpc

上个项目中用到了ActiveMQ,只是简单应用,安装完成后直接是用就可以了。由于新项目中一些硬件的限制,需要把消息队列换成RabbitMQ。

RabbitMQ中的几种模式和机制比ActiveMQ多多了,根据业务需要,使用RPC实现功能,其中踩过的一些坑,有必要记录一下了。

Python RabbitMQ消息队列实现rpc

上代码,目录结构分为 c_server、c_client、c_hanlder:

c_server:

#!/usr/bin/env python
# -*- coding:utf-8 -*-
import pika
import time
import json
import io
import yaml

s_exchange = input("请输入交换机名称-").decode('utf-8').strip()
s_queue = input("输入消息队列名称-").decode('utf-8').strip()
credentials = pika.PlainCredentials('system', 'manager')
connection = pika.BlockingConnection(pika.ConnectionParameters(host='XXX.XXX.XXX.XXX',credentials=credentials))
# 定义
channel = connection.channel()
channel.exchange_declare(exchange=s_exchange, exchange_type='direct')
channel.queue_declare(queue=s_queue, exclusive=True)
channel.queue_bind(queue=s_queue, exchange=s_exchange)

def s_manage(content):
 # 解决unicode转码问题 json.JSONDecoder().decode(content)
 str_content = yaml.safe_load(json.loads(content,encoding='utf-8'))
 str_res = {
  "errorid": 0,
  "resp": str_content['cmd'],
  "errorcont": "成功"
 }
 return json.dumps(str_res)

def on_request(ch, method, props, body):
 response = s_manage(body)
 ch.basic_publish(exchange='',
      routing_key=props.reply_to,
      properties=pika.BasicProperties(correlation_id =                props.correlation_id),
      body=response)
 ch.basic_ack(delivery_tag = method.delivery_tag)

channel.basic_qos(prefetch_count=1)
channel.basic_consume(on_request, queue=s_queue)

print(" [x] Awaiting RPC requests")
channel.start_consuming()

c_client:

#!/usr/bin/env python
# -*- coding:utf-8 -*-

import pika
import uuid
import json
import io

class RpcClient(object):
  def __init__(self):
    self.credentials = pika.PlainCredentials('guest', 'guest')
    self.connection = pika.BlockingConnection(pika.ConnectionParameters(host='XXX.XXX.XXX.XXX',
                                credentials=self.credentials))
    self.channel = self.connection.channel()

  def on_response(self, ch, method, props, body):
    if self.callback_id == props.correlation_id:
      self.response = body
    ch.basic_ack(delivery_tag=method.delivery_tag)

  def get_response(self, callback_queue, callback_id):
    '''取队列里的值,获取callback_queued的执行结果'''
    self.callback_id = callback_id
    self.response = None
    self.channel.queue_declare('q_manager', durable=True)
    self.channel.basic_consume(self.on_response, # 只要收到消息就执行on_response
                  queue=callback_queue)
    while self.response is None:
      self.connection.process_data_events() # 非阻塞版的start_consuming
    return self.response

  def call(self, queue_name, command, exchange,rout_key): # 命令下发
    '''队列里发送数据'''
    # result = self.channel.queue_declare(exclusive=False) #exclusive=False 必须这样写
    self.callback_queue = 'q_manager' # result.method.queue
    self.corr_id = str(uuid.uuid4())
    self.channel.basic_publish(exchange=exchange,
                  routing_key=queue_name,
                  properties=pika.BasicProperties(
                    reply_to=self.callback_queue, # 发送返回信息的队列name
                    correlation_id=self.corr_id, # 发送uuid 相当于验证码
                  ),
                  body=command)
    return self.callback_queue,self.corr_id

client

c_handler:

#!/usr/bin/env python
# -*- coding:utf-8 -*-

from c_client import *
import random, time
import threading
import json
import sys

class Handler(object):
  def __init__(self):
    self.information = {}  # 后台进程信息

  def check_all(self, *args):
    '''查看所有信息'''
    time.sleep(2)
    print('获取消息')
    for key in self.information:
      print("cid【%s】\t 队列【%s】\t 命令【%s】"%(key, self.information[key][0],
                               self.information[key][1]))

  def check_task(self, cmd):
    '''查看task_id执行结果'''
    time.sleep(2)
    try:
      task_id = int(cmd)
      print(task_id)
      callback_queue= self.information[task_id][2]
      callback_id= self.information[task_id][3]
      client = RpcClient()
      response = client.get_response(callback_queue, callback_id)
      print(response)
      # print(response.decode())
      del self.information[task_id]

    except KeyError as e :
      print("error: [%s]" % e)
    except IndexError as e:
      print("error: [%s]" % e)

  def run(self, user_cmd, host, exchange='', rout_key='',que=''):
    try:
      time.sleep(2)
      command = user_cmd
      task_id = random.randint(10000, 99999)
      client = RpcClient()
      response = client.call(queue_name=host, command=command,exchange=exchange,rout_key=que)
      self.information[task_id] = [host, command, response[0], response[1]]
    except IndexError as e:
      print("[error]:%s"%e)

  def reflect(self, str,cmd,host,exchange,que):
    '''反射'''
    if hasattr(self, str):
      getattr(self, str)(cmd,host,exchange,que)

  def start(self, m,cmd, host, exchange,que):
    while True:
      user_resp = input("输入处理消息内容ID-").decode('utf-8').strip()
      self.check_task(user_resp)
      str = m
      print(self.information)
      t1 = threading.Thread(target=self.reflect, args=(str,cmd,host,exchange,que)) #多线程
      t1.start()

s_exchange = input("请输入交换机名称-").decode('utf-8').strip()
s_queue = input("输入消息队列名称-").decode('utf-8').strip()
d_cmd_state =input("输入json命令-").decode('utf-8').strip()
s_cmd = json.dumps(d_cmd_state)
handler = Handler()
handler.start('run',s_cmd, s_queue, s_exchange, s_queue)

handler

注意要点:1、c_client 发布消息到rabbitmq 需要携带 服务器返回的队列名称,及corr_id

2、c_handler 做了处理,每次发送的内容都会放到task列表中,直到显示ID号,就可以查询返回的内容,调用如下:

Python RabbitMQ消息队列实现rpc

Python RabbitMQ消息队列实现rpc

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持。