Akka并发编程——第六节：Actor模型（五）

本章主要内容：
1. !消息发送，Fire-and-Forget消息模型
2. ？消息发送，Send-And-Receive-Future消息模型

Akka提供了两种消息模型：fire-and-forget和Send-And-Receive-Future。fire-and-forget是一种单向消息发送模型，指的是异步发送消息，通过异步发送消息且消息发送后可以立即返回，Akka中使用?方法进行fire-and-forget消息发送，如stringActor!”Creating Actors with implicit val context”，它的意思是当前发送方Aactor向stringActor发送字符串消息”Creating Actors with implicit val context”，发送完该消息后立即返回，而无需等待stringActor的返回，!还有个重载的方法tell；Send-And-Receive-Future指的是异步发送消息则是一种双向消息发送模型，向目标Actor发送完消息后，然后返回一个Future作为后期可能的返回，当前发送方Actor将等待目标Actor的返回，Akka中使用?方法进行Send-And-Receive-Future消息的发送，它也同样有一个重载的方法ask

1. !消息发送，Fire-and-Forget消息模型

/**  * 消息处理：!(Fire-Forget)  */object Example12 extends  App{  import akka.actor.Actor  import akka.actor.Props  import akka.event.Logging  import akka.actor.ActorSystem //定义几种不同的消息  case class Start(var msg:String)  case class Run(var msg:String)  case class Stop(var msg:String)  class ExampleActor extends Actor {    val other = context.actorOf(Props[OtherActor], "OtherActor")    val log = Logging(context.system, this)    def receive={      //使用fire-and-forget消息模型向OtherActor发送消息，隐式地传递sender      case Start(msg) => other ! msg      //使用fire-and-forget消息模型向OtherActor发送消息，直接调用tell方法，显式指定sender      case Run(msg) => other.tell(msg, sender)    }  }  class OtherActor extends  Actor{    val log = Logging(context.system, this)    def receive ={      case s:String=>log.info("received message:\n"+s)      case _      ⇒ log.info("received unknown message")    }  }  //创建ActorSystem,ActorSystem为创建和查找Actor的入口  //ActorSystem管理的Actor共享配置信息如分发器(dispatchers)、部署（deployments）等  val system = ActorSystem("MessageProcessingSystem")  //创建ContextActor  val exampleActor = system.actorOf(Props[ExampleActor],name="ExampleActor")  //使用fire-and-forget消息模型向exampleActor发送消息  exampleActor!Run("Running")  exampleActor!Start("Starting")  //关闭ActorSystem  system.shutdown()}

代码运行结果如下：

[INFO] [03/20/2016 20:57:43.665] [MessageProcessingSystem-akka.actor.default-dispatcher-5] [akka://MessageProcessingSystem/user/ExampleActor/OtherActor] received message:Running[INFO] [03/20/2016 20:57:43.672] [MessageProcessingSystem-akka.actor.default-dispatcher-5] [akka://MessageProcessingSystem/user/ExampleActor/OtherActor] received message:Starting

在ExampleActor中，通过隐式变量context创建了OtherActor实例：val other = context.actorOf(Props[OtherActor], “OtherActor”)，在ExampleActor的receive方法中，处理两种不同类型的消息例如：

  //使用fire-and-forget消息模型向OtherActor发送消息，隐式地传递sender  case Start(msg) => other ! msg  //使用fire-and-forget消息模型向OtherActor发送消息，直接调用tell方法，显式指定sender  case Run(msg) => other.tell(msg, sender)
1

处理Start类型的消息时，直接使用!进行消息发送，而处理Run类型的消息时，使用的是tell方法，可以看到使用tell方法需要显式地指定其sender，而使用!进行消息发送则不需要，事实上!方法通过隐式值传入需要的Sender，对比!与tell方法的定义便很容易理解

//!方法的定义def !(message: Any)(implicit sender: ActorRef = Actor.noSender): Unit//tell方法的定义final def tell(msg: Any, sender: ActorRef): Unit = this.!(msg)(sender)

可以看到，tell方法的实现依赖于!方法。如果在一个Actor当中使用!方法时，例如ExampleActor中使用的other ! msg向OtherActor发送消息，则sender隐式为ExampleActor，如果不是在Actor中使用则默认为Actor.noSender，即sender为null。

2. ？消息发送，Send-And-Receive-Future消息模型

理解了Fire-And-Forget消息模型后，接着对Send-And-Receive-Future消息模型进行介绍，下面的代码给出了其使用示例。

/**  * 消息处理：?(Send-And-Receive-Future)  */object Example13 extends  App{  import akka.actor.Actor  import akka.actor.Props  import akka.event.Logging  import akka.actor.ActorSystem  import scala.concurrent.Future  import akka.pattern.ask  import akka.util.Timeout  import scala.concurrent.duration._  import akka.pattern.pipe  import scala.concurrent.ExecutionContext.Implicits.global  //消息：个人基础信息  case class BasicInfo(id:Int,val name:String, age:Int)  //消息：个人兴趣信息  case class InterestInfo(id:Int,val interest:String)  //消息: 完整个人信息  case class Person(basicInfo: BasicInfo,interestInfo: InterestInfo)  //基础信息对应Actor  class BasicInfoActor extends Actor{    val log = Logging(context.system, this)    def receive = {      //处理送而来的用户ID，然后将结果发送给sender（本例中对应CombineActor）      case id:Int ⇒log.info("id="+id);sender!new BasicInfo(id,"John",19)      case _      ⇒ log.info("received unknown message")    }  }  //兴趣爱好对应Actor  class InterestInfoActor extends Actor{    val log = Logging(context.system, this)    def receive = {      //处理发送而来的用户ID，然后将结果发送给sender（本例中对应CombineActor）      case id:Int ⇒log.info("id="+id);sender!new InterestInfo(id,"足球")      case _      ⇒ log.info("received unknown message")    }  }  //Person完整信息对应Actor  class PersonActor extends Actor{    val log = Logging(context.system, this)    def receive = {      case person: Person =>log.info("Person="+person)      case _      ⇒ log.info("received unknown message")    }  }  class CombineActor extends Actor{    implicit val timeout = Timeout(5 seconds)    val basicInfoActor = context.actorOf(Props[BasicInfoActor],name="BasicInfoActor")    val interestInfoActor = context.actorOf(Props[InterestInfoActor],name="InterestInfoActor")    val personActor = context.actorOf(Props[PersonActor],name="PersonActor")    def receive = {      case id: Int =>        val combineResult: Future[Person] =          for {            //向basicInfoActor发送Send-And-Receive-Future消息，mapTo方法将返回结果映射为BasicInfo类型            basicInfo <- ask(basicInfoActor, id).mapTo[BasicInfo]            //向interestInfoActor发送Send-And-Receive-Future消息，mapTo方法将返回结果映射为InterestInfo类型            interestInfo <- ask(interestInfoActor, id).mapTo[InterestInfo]          } yield Person(basicInfo, interestInfo)        //将Future结果发送给PersonActor       pipe(combineResult).to(personActor)    }  }  val _system = ActorSystem("Send-And-Receive-Future")  val combineActor = _system.actorOf(Props[CombineActor],name="CombineActor")  combineActor ! 12345  Thread.sleep(5000)  _system.shutdown}

代码运行结果如下：

[INFO] [03/20/2016 22:55:11.208] [Send-And-Receive-Future-akka.actor.default-dispatcher-3] [akka://Send-And-Receive-Future/user/CombineActor/BasicInfoActor] id=12345[INFO] [03/20/2016 22:55:11.220] [Send-And-Receive-Future-akka.actor.default-dispatcher-2] [akka://Send-And-Receive-Future/user/CombineActor/InterestInfoActor] id=12345[INFO] [03/20/2016 22:55:11.223] [Send-And-Receive-Future-akka.actor.default-dispatcher-4] [akka://Send-And-Receive-Future/user/CombineActor/PersonActor] Person=Person(BasicInfo(12345,John,19),InterestInfo(12345,足球))
1

代码中定义了3种类型的消息，分别是个人基础信息case class BasicInfo(id:Int,val name:String, age:Int)、个人兴趣信息case class InterestInfo(id:Int,val interest:String)以及完整个人信息case class Person(basicInfo: BasicInfo,interestInfo: InterestInfo)，然后为这3种类型的消息定义了相应的Actor即BasicInfoActor、InterestInfoActor和PersonActor，在CombineActor分别创建相应Actor的实例，receive方法中使用ask向BasicInfoActor、InterestInfoActor发送Send-And-Receive-Future模型消息，BasicInfoActor、InterestInfoActor中的receive方法接收到发送来的Int类型消息并分别使用!向CombineActor发送BasicInfo、InterestInfo消息，将结果保存在Future[Person]中，然后通过代码pipe(combineResult).to(personActor)将结果发送给PersonActor。