• RabbitMQ(pika模块)


    RabbitMQ

    基础

    关于MQ:

    MQ全称为Message Queue, 消息队列(MQ)是一种应用程序对应用程序的通信方法。应用程序通过读写出入队列的消息(针对应用程序的数据)来通信,而无需专用连接来链接它们。消息传递指的是程序之间通过在消息中发送数据进行通信,而不是通过直接调用彼此来通信,直接调用通常是用于诸如远程过程调用的技术。排队指的是应用程序通过队列来通信。队列的使用除去了接收和发送应用程序同时执行的要求。

    RabbitMQ安装

    1
    2
    3
    4
    5
    6
    7
    8
    安装配置epel源
       $ rpm -ivh http://dl.fedoraproject.org/pub/epel/6/i386/epel-release-6-8.noarch.rpm
      
    安装erlang
       $ yum -y install erlang
      
    安装RabbitMQ
       $ yum -y install rabbitmq-server

    启动/停止:

    1
    systemctl start/stop rabbitmq

    安装python-API:

    1
    2
    3
    4
    5
    6
    7
    pip install pika
    or
    easy_install pika
    or
    源码
      
    https://pypi.python.org/pypi/pika


    API基础操作


    先来看看使用RabbitMQ之前,怎么实现消息队列:利用Queue和Thread,每线程往内存里的队列里put一个数,另一个程序再去内存队列里取数。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    import Queue
    import threading
     
     
    message = Queue.Queue(10)
     
     
    def producer(i):
        while True:
            message.put(i)
     
     
    def consumer(i):
        while True:
            msg = message.get()
     
     
    for i in range(12):
        t = threading.Thread(target=producer, args=(i,))
        t.start()
     
    for i in range(10):
        t = threading.Thread(target=consumer, args=(i,))
        t.start()

    对于RabbitMQ来说,生产和消费不再针对内存里的一个Queue对象,而是某台服务器上的RabbitMQ Server实现的消息队列。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    import pika
     
    # ######################### 生产者 #########################
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()    #创建通道
     
    channel.queue_declare(queue='hello')    #队列名称
     
    channel.basic_publish(exchange='',
                          routing_key='hello',  #路由名称
                          body='Hello World!'#发送内容
    print(" [x] Sent 'Hello World!'")
    connection.close()

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    import pika
     
    # ########################## 消费者 ##########################
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    channel.queue_declare(queue='hello')    #声明,队列名称,和producer创建的重复没有关系
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
     
    channel.basic_consume(callback,         #获取body后执行回调函数
                          queue='hello',
                          no_ack=True)                #自动应答开启,会给MQ服务器发送一个ack:‘已经收到了’。
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

    消费者运行起来后会和RabbitMQ建立长连接,一旦生产者放数据到队列里,消费者就能获取到该值,并进行处理。

    1
    2
    [root@localhost ~]# netstat -ntp |grep beam
    tcp6       0      0 192.168.136.8:5672      192.168.136.1:52587     ESTABLISHED 1146/beam


    消息安全

    1、no-ack = False(自动应答关闭)

    如果生产者遇到情况(its channel is closed, connection is closed, or TCP connection is lost)挂掉了,那么,RabbitMQ会重新将该任务添加到队列中。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    import pika
    #no-ack
    ########################### 消费者 ##########################
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    channel.queue_declare(queue='hello')
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print 'ok'
        ch.basic_ack(delivery_tag = method.delivery_tag)    #主动发送ack
        #打印‘ok’后才告诉MQ,这个消息已经处理完了。
     
    channel.basic_consume(callback,
                          queue='hello',
                          no_ack=False)     #自动应答关闭,与channel.basic_ack共同使用
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

    2、durable  

    make message persistent 使消息持久化

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    import pika
     
    #durable
    ########################## 生产者 #########################
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    channel.queue_declare(queue='hello', durable=True#开启持久化
     
    channel.basic_publish(exchange='',
                          routing_key='hello',
                          body='Hello World!',
                          properties=pika.BasicProperties(
                              delivery_mode=2, # make message persistent
                          ))
    print(" [x] Sent 'Hello World!'")
    connection.close()
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    import pika
    #durable
    ########################## 消费者 #########################
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    # make message persistent
    channel.queue_declare(queue='hello', durable=True)
     
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print 'ok'
        ch.basic_ack(delivery_tag = method.delivery_tag)
     
    channel.basic_consume(callback,
                          queue='hello',
                          no_ack=False)
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()


    消息获取顺序

    默认消息队列里的数据是按照奇偶顺序被消费者拿走,例如:消费者1 去队列中获取 奇数 序列的任务,消费者2去队列中获取 偶数 序列的任务。

    channel.basic_qos(prefetch_count=1) 表示谁来谁取,不再按照奇偶数排列.

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='10.211.55.4'))
    channel = connection.channel()
     
    # make message persistent
    channel.queue_declare(queue='hello')
     
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print 'ok'
        ch.basic_ack(delivery_tag = method.delivery_tag)
     
    channel.basic_qos(prefetch_count=1)    #增加这行
     
    channel.basic_consume(callback,
                          queue='hello',
                          no_ack=False)
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()


    发布&订阅

    与消息队列区别:

    消息队列中的数据只要被消费一次便消失。

    创建队列的数量:

    同一份消息,有多少订阅者,就要创建多少个队列。(RabbitMQ实现发布和订阅时,会为每一个订阅者创建一个队列,而发布者发布消息时,会将消息放置在所有相关队列中。)

    语法:

    exchange type = fanout        #fanout==>输出到很多

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    # ######################### 发布者 #########################
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='fanout_name',type='fanout')
     
    message = ' '.join(sys.argv[1:]) or "info: Hello World!"
    channel.basic_publish(exchange='fanout_name'#自命名exchange
                          routing_key='',
                          body=message)
    print(" [x] Sent %r" % message)
    connection.close()
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    # ########################## 订阅者1 ##########################
     
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='fanout_name',type='fanout')    #创建exchange(if not exist)
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue        #获取队列名称
     
    channel.queue_bind(exchange='fanout_name',queue=queue_name)    #通过上面两个值绑定队列
     
    print(' [*] Waiting for fanout_name. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r" % body)
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()

    创建多个订阅者,能更好的体现它的效果。

    运行结果总结:

    每个订阅者创建一个exchange队列,名称自定,发布者会把数据发送给所有叫这个名字的队列。因为数据只能被消费一次,所以有多少个订阅者,就有多少个队列。


    发送到指定(not 固定)队列

    之前事例,发送消息时明确指定某个队列并向其中发送消息,RabbitMQ还支持根据关键字发送

    1、按关键字寻找队列发送

    exchange type = direct

    队列绑定关键字,发送者将数据根据关键字发送到消息exchange,exchange根据 关键字 判定应该将数据发送至指定队列。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    # ######################### 生产者 #########################
    #关键字发送
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='direct_logs',
                             type='direct')
     
     
    message = 'Hello World!'
    channel.basic_publish(exchange='direct_logs',
                          routing_key="yes", #"yes","no","db"
                          body=message)
    print(" [x] Sent %r" % (message))
    connection.close()

    模拟两个消费者,一个消费者的队列是("yes","db"),另一个消费者队列("no","db")。如果生产者发送的队列关键字是"yes"or"no",其一匹配;如果生产者发送的队列关键字是"db",则两个消费者都能接收到。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    ########################### 消费者1 ##########################
    import pika
    import sys
      
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
      
    channel.exchange_declare(exchange='direct_logs',
                             type='direct')
      
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    channel.queue_bind(exchange='direct_logs',
                           queue=queue_name,
                           routing_key='yes')
    channel.queue_bind(exchange='direct_logs',
                           queue=queue_name,
                           routing_key='db')
      
    print(' [*] Waiting for logs. To exit press CTRL+C')
      
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
      
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
      
    channel.start_consuming()

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    ########################### 消费者2 ##########################
    import pika
    import sys
      
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
      
    channel.exchange_declare(exchange='direct_logs',
                             type='direct')
      
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    channel.queue_bind(exchange='direct_logs',
                           queue=queue_name,
                           routing_key='no')
    channel.queue_bind(exchange='direct_logs',
                           queue=queue_name,
                           routing_key='db')
      
    print(' [*] Waiting for logs. To exit press CTRL+C')
      
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
      
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
      
    channel.start_consuming()

    2、模糊匹配

     exchange type = topic

    在topic类型下,可以让队列绑定几个模糊的关键字,之后发送者将数据发送到exchange,exchange将传入”路由值“和 ”关键字“进行匹配,匹配成功,则将数据发送到指定队列。

    • # 表示可以匹配 0 个 或 多个 单词

    • *  表示只能匹配 一个 单词

    1
    2
    3
    发送者路由值              队列中
    python.topic.cn          python.*  -- 不匹配
    python.topic.cn          python.#  -- 匹配
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    # ######################### 生产者 #########################
    #模糊匹配
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='topic_logs',
                             type='topic')
     
    message = 'Hello World!'
    channel.basic_publish(exchange='topic_logs',
                          routing_key="python.topic",
                          body=message)
    print(" [x] Sent %r" % (message))
    connection.close()

    消费者1是‘*’匹配,消费者2是‘#’匹配:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    ########################### 消费者1 ##########################
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='topic_logs',
                             type='topic')
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    channel.queue_bind(exchange='topic_logs',
                           queue=queue_name,
                           routing_key='python.*')     #只匹配python.后有一个单词的
     
    print(' [*] Waiting for topic_logs. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    ########################### 消费者2 ##########################
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(
            host='192.168.136.8'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='topic_logs',
                             type='topic')
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    channel.queue_bind(exchange='topic_logs',
                           queue=queue_name,
                           routing_key='python.#')    #匹配python.后所有单词
     
    print(' [*] Waiting for topic_logs. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()

    从结果得出结论,如果生产者发送的routing_key是:

    • python.topic.cn    -->    只有消费者2能接收到

    • python.cn            -->    消费者1和消费者2都能接收到

    • python.                -->    消费者1和消费者2都能接收到

    • python                 -->    只有消费者2能接收到


    网络搜索的概念:

    Topic Exchange – 主题式交换器,通过消息的路由关键字和绑定关键字的模式匹配,将消息路由到被绑定的队列中。

    这种路由器类型可以被用来支持经典的发布/订阅消息传输模型——使用主题名字空间作为消息寻址模式,将消息传递给那些部分或者全部匹配主题模式的多个消费者。

    主题交换器类型的工作方式如下: 绑定关键字用零个或多个标记构成,每一个标记之间用“.”字符分隔

    绑定关键字必须用这种形式明确说明,并支持通配符:“*”匹配一个词组,“#”零个或多个词组。

    因此绑定关键字“*.stock.#”匹配路由关键字“usd.stock”和“eur.stock.db”,但是不匹配“stock.nasdaq”


    参考来源:http://www.cnblogs.com/wupeiqi/














  • 相关阅读:
    JS实现添加至购物车功能
    python定制数据类型(继承与授权)两种方式
    模拟数据,控制账号
    绑定知识拓展
    面向对象之多态练习
    面向对象之继承与派生(学生管理系统)
    面向对象之组合(学生管理系统)
    一次小的上机试题
    面向对象之self classmethod staticmethod
    haproxy.conf文件操作(基于函数方式)
  • 原文地址:https://www.cnblogs.com/daliangtou/p/5147730.html
Copyright © 2020-2023  润新知