Akka(35): Http:Server side streaming
来源:互联网 发布:js点击按钮旋转图片 编辑:程序博客网 时间:2024/05/17 23:50
在前面几篇讨论里我们都提到过:Akka-http是一项系统集成工具库。它是以数据交换的形式进行系统集成的。所以,Akka-http的核心功能应该是数据交换的实现了:应该能通过某种公开的数据格式和传输标准比较方便的实现包括异类系统之间通过网上进行的数据交换。覆盖包括:数据编码、发送和数据接收、解析全过程。Akka-http提供了许多网上传输标准数据的概括模型以及数据类型转换方法,可以使编程人员很方便的构建网上往来的Request和Response。但是,现实中的数据交换远远不止针对request和response操作能够满足的。系统之间数据交换经常涉及文件或者数据库表类型的数据上传下载。虽然在Http标准中描述了如何通过MultiPart消息类型进行批量数据的传输,但是这个标准涉及的实现细节包括数据内容描述、数据分段方式、消息数据长度计算等等简直可以立即令人却步。Akka-http是基于Akka-stream开发的:不但它的工作流程可以用Akka-stream来表达,它还支持stream化的数据传输。我们知道:Akka-stream提供了功能强大的FileIO和Data-Streaming,可以用Stream-Source代表文件或数据库数据源。简单来说:Akka-http的消息数据内容HttpEntity可以支持理论上无限长度的data-stream。最可贵的是:这个Source是个Reactive-Stream-Source,具备了back-pressure机制,可以有效应付数据交换参与两方Reactive端点不同的数据传输速率。
Akka-http的stream类型数据内容是以Source[T,_]类型表示的。首先,Akka-stream通过FileIO对象提供了足够多的file-io操作函数,其中有个fromPath函数可以用某个文件内容数据构建一个Source类型:
/** * Creates a Source from a files contents. * Emitted elements are `chunkSize` sized [[akka.util.ByteString]] elements, * except the final element, which will be up to `chunkSize` in size. * * You can configure the default dispatcher for this Source by changing the `akka.stream.blocking-io-dispatcher` or * set it for a given Source by using [[akka.stream.ActorAttributes]]. * * It materializes a [[Future]] of [[IOResult]] containing the number of bytes read from the source file upon completion, * and a possible exception if IO operation was not completed successfully. * * @param f the file path to read from * @param chunkSize the size of each read operation, defaults to 8192 */ def fromPath(f: Path, chunkSize: Int = 8192): Source[ByteString, Future[IOResult]] = fromPath(f, chunkSize, startPosition = 0)
这个函数构建了Source[ByteString,Future[IOResult]],我们需要把ByteString转化成MessageEntity。首先需要在implicit-scope内提供Marshaller[ByteString,MessageEntity]类型的隐式实例:
trait JsonCodec extends Json4sSupport { import org.json4s.DefaultFormats import org.json4s.ext.JodaTimeSerializers implicit val serilizer = jackson.Serialization implicit val formats = DefaultFormats ++ JodaTimeSerializers.all}object JsConverters extends JsonCodecobject ServerStreaming extends App { import JsConverters._...
我们还需要Json-Streaming支持:
implicit val jsonStreamingSupport = EntityStreamingSupport.json() .withParallelMarshalling(parallelism = 8, unordered = false)
FileIO是blocking操作,我们还可以选用独立的线程供blocking操作使用:
FileIO.fromPath(file, 256) .withAttributes(ActorAttributes.dispatcher("akka.http.blocking-ops-dispatcher"))
现在我们可以从在server上用一个文件构建Source然后再转成Response:
val route = get { path("files"/Remaining) { name => complete(loadFile(name)) } } def loadFile(path: String) = { // implicit val ec = httpSys.dispatchers.lookup("akka.http.blocking-ops-dispatcher") val file = Paths.get("/Users/tiger/"+path) FileIO.fromPath(file, 256) .withAttributes(ActorAttributes.dispatcher("akka.http.blocking-ops-dispatcher")) .map(_.utf8String) }
同样,我们也可以把数据库表内数据转成Akka-Stream-Source,然后再实现到MessageEntity的转换。转换过程包括用Query读取数据库表内数据后转成Reactive-Publisher,然后把publisher转成Akka-Stream-Source,如下:
object SlickDAO { import slick.jdbc.H2Profile.api._ val dbConfig: slick.basic.DatabaseConfig[slick.jdbc.H2Profile] = slick.basic.DatabaseConfig.forConfig("slick.h2") val db = dbConfig.db case class CountyModel(id: Int, name: String) case class CountyTable(tag: Tag) extends Table[CountyModel](tag,"COUNTY") { def id = column[Int]("ID",O.AutoInc,O.PrimaryKey) def name = column[String]("NAME",O.Length(64)) def * = (id,name)<>(CountyModel.tupled,CountyModel.unapply) } val CountyQuery = TableQuery[CountyTable] def loadTable(filter: String) = { // implicit val ec = httpSys.dispatchers.lookup("akka.http.blocking-ops-dispatcher") val qry = CountyQuery.filter {_.name.toUpperCase like s"%${filter.toUpperCase}%"} val publisher = db.stream(qry.result) Source.fromPublisher(publisher = publisher) .withAttributes(ActorAttributes.dispatcher("akka.http.blocking-ops-dispatcher")) }}
然后进行到MessageEntity的转换:
val route = get { path("files"/Remaining) { name => complete(loadFile(name)) } ~ path("tables"/Segment) { t => complete(SlickDAO.loadTable(t)) } }
下面是本次示范的完整源代码:
import java.nio.file._import akka.actor._import akka.stream._import akka.stream.scaladsl._import akka.http.scaladsl.Httpimport akka.http.scaladsl.server.Directives._import akka.http.scaladsl.common._import de.heikoseeberger.akkahttpjson4s.Json4sSupportimport org.json4s.jacksonobject SlickDAO { import slick.jdbc.H2Profile.api._ val dbConfig: slick.basic.DatabaseConfig[slick.jdbc.H2Profile] = slick.basic.DatabaseConfig.forConfig("slick.h2") val db = dbConfig.db case class CountyModel(id: Int, name: String) case class CountyTable(tag: Tag) extends Table[CountyModel](tag,"COUNTY") { def id = column[Int]("ID",O.AutoInc,O.PrimaryKey) def name = column[String]("NAME",O.Length(64)) def * = (id,name)<>(CountyModel.tupled,CountyModel.unapply) } val CountyQuery = TableQuery[CountyTable] def loadTable(filter: String) = { // implicit val ec = httpSys.dispatchers.lookup("akka.http.blocking-ops-dispatcher") val qry = CountyQuery.filter {_.name.toUpperCase like s"%${filter.toUpperCase}%"} val publisher = db.stream(qry.result) Source.fromPublisher(publisher = publisher) .withAttributes(ActorAttributes.dispatcher("akka.http.blocking-ops-dispatcher")) }}trait JsonCodec extends Json4sSupport { import org.json4s.DefaultFormats import org.json4s.ext.JodaTimeSerializers implicit val serilizer = jackson.Serialization implicit val formats = DefaultFormats ++ JodaTimeSerializers.all}object JsConverters extends JsonCodecobject ServerStreaming extends App { import JsConverters._ implicit val httpSys = ActorSystem("httpSystem") implicit val httpMat = ActorMaterializer() implicit val httpEC = httpSys.dispatcher implicit val jsonStreamingSupport = EntityStreamingSupport.json() .withParallelMarshalling(parallelism = 8, unordered = false) val (port, host) = (8011,"localhost") val route = get { path("files"/Remaining) { name => complete(loadFile(name)) } ~ path("tables"/Segment) { t => complete(SlickDAO.loadTable(t)) } } def loadFile(path: String) = { // implicit val ec = httpSys.dispatchers.lookup("akka.http.blocking-ops-dispatcher") val file = Paths.get("/Users/tiger/"+path) FileIO.fromPath(file, 256) .withAttributes(ActorAttributes.dispatcher("akka.http.blocking-ops-dispatcher")) .map(_.utf8String) } val bindingFuture = Http().bindAndHandle(route,host,port) println(s"Server running at $host $port. Press any key to exit ...") scala.io.StdIn.readLine() bindingFuture.flatMap(_.unbind()) .onComplete(_ => httpSys.terminate())}
resource/application.conf
akka { http { blocking-ops-dispatcher { type = Dispatcher executor = "thread-pool-executor" thread-pool-executor { // or in Akka 2.4.2+ fixed-pool-size = 16 } throughput = 100 } }}slick { h2 { driver = "slick.driver.H2Driver$" db { url = "jdbc:h2:~/slickdemo;mv_store=false" driver = "org.h2.Driver" connectionPool = HikariCP numThreads = 48 maxConnections = 48 minConnections = 12 keepAliveConnection = true } }}
- Akka(35): Http:Server side streaming
- Akka(29): Http:Server-Side-Api,Low-Level-Api
- Akka(29): Http:Server-Side-Api,Low-Level-Api
- Akka(41): Http:DBTable-rows streaming
- Akka(36): Http:Client-side-Api,Client-Connections
- Akka(39): Http:File streaming-文件交换
- Akka(28): Http:About Akka-Http
- Akka(43): Http:SSE-Server Sent Event
- 七牛-Pili Streaming Cloud Server-Side Library For PHP
- Akka(42): Http:身份验证
- SSI(Server Side Include)
- SSI(Server Side Include)
- Akka(33): Http:Marshalling,to Json
- Akka(34): Http:Unmarshalling,from Json
- Akka(37): Http:客户端操作模式
- Akka(40): Http:Marshalling reviewed
- Akka边学边写(2)-- Echo Server
- client-side and server-side
- 25岁Java工程师如何转型学习人工智能?
- jtag相关
- Oracle中的记录(Record)
- 老外的个人网站原来是这样设计的:17个最佳范例给你灵感
- 【转载】Jdk jre jvm的区别和联系
- Akka(35): Http:Server side streaming
- Java标识符
- js 数组与对象
- 微信java开发3——获取access_token
- C++ 接口(抽象类)
- 面试心得与总结:BAT、网易、蘑菇街
- dpdk-ixgbe 详解
- redis内部数据结构
- 【Java作业】2017.10.28 / 10.29