play ws Download - play ws Source code download

play ws

Other source code

3.0.6

Download

Play WS Standalone

Play WS is a powerful HTTP Client library, originally developed by the Play team for use with Play Framework. It uses AsyncHttpClient for HTTP client functionality and has no Play dependencies.

We've provided some documentation here on how to use Play WS in your app (without Play). For more information on how to use Play WS in Play, please refer to the Play documentation.

Getting Started

To get started, you can add play-ahc-ws-standalone as a dependency in SBT:

libraryDependencies += "org.playframework" %% "play-ahc-ws-standalone" % "LATEST_VERSION"// Before version 3.0.0:libraryDependencies += "com.typesafe.play" %% "play-ahc-ws-standalone" % "LATEST_VERSION"

Where you replace LATEST_VERSION with the version shown in this image: .

This adds the standalone version of Play WS, backed by AsyncHttpClient. This library contains both the Scala and Java APIs, under play.api.libs.ws and play.libs.ws.

To add XML and JSON support using Play-JSON or Scala XML, add the following:

libraryDependencies += "org.playframework" %% "play-ws-standalone-xml" % playWsStandaloneVersion
libraryDependencies += "org.playframework" %% "play-ws-standalone-json" % playWsStandaloneVersion// Before version 3.0.0:libraryDependencies += "com.typesafe.play" %% "play-ws-standalone-xml" % playWsStandaloneVersion
libraryDependencies += "com.typesafe.play" %% "play-ws-standalone-json" % playWsStandaloneVersion

Shading

Play WS uses shaded versions of AsyncHttpClient and OAuth Signpost, repackaged under the play.shaded.ahc and play.shaded.oauth package names, respectively. Shading AsyncHttpClient means that the version of Netty used behind AsyncHttpClient is completely independent of the application and Play as a whole.

Specifically, shading AsyncHttpClient means that there are no version conflicts introduced between Netty 4.0 and Netty 4.1 using Play WS.

NOTE: If you are developing play-ws and publishing shaded-asynchttpclient and shaded-oauth using sbt publishLocal, you need to be aware that updating ~/.ivy2/local does not overwrite ~/.ivy2/cache and so you will not see your updated shaded code until you remove it from cache. See http://eed3si9n.com/field-test for more details. This bug has been filed as sbt/sbt#2687.

Shaded AHC Defaults

Because Play WS shades AsyncHttpClient, the default settings are also shaded and so do not adhere to the AHC documentation. This means that the settings in ahc-default.properties and the AsyncHttpClient system properties are prepended with play.shaded.ahc, for example the usePooledMemory setting in the shaded version of AsyncHttpClient is defined like this:

play.shaded.ahc.org.asynchttpclient.usePooledMemory=true

Typed Bodies

The type system in Play-WS has changed so that the request body and the response body can use richer types.

You can define your own BodyWritable or BodyReadable, but if you want to use the default out of the box settings, you can import the type mappings with the DefaultBodyReadables / DefaultBodyWritables.

Scala

import play.api.libs.ws.DefaultBodyReadables._import play.api.libs.ws.DefaultBodyWritables._

More likely you will want the XML and JSON support:

import play.api.libs.ws.XMLBodyReadables._import play.api.libs.ws.XMLBodyWritables._

or

import play.api.libs.ws.JsonBodyReadables._import play.api.libs.ws.JsonBodyWritables._

To use a BodyReadable in a response, you must type the response explicitly:

import scala.concurrent.{ ExecutionContext, Future }import play.api.libs.ws.StandaloneWSClientimport play.api.libs.ws.XMLBodyReadables._ // requireddef handleXml(ws: StandaloneWSClient)(  implicit ec: ExecutionContext): Future[scala.xml.Elem] =
  ws.url("...").get().map { response =>response.body[scala.xml.Elem]
  }

or using Play-JSON:

import scala.concurrent.{ ExecutionContext, Future }import play.api.libs.json.JsValueimport play.api.libs.ws.StandaloneWSClientimport play.api.libs.ws.JsonBodyReadables._ // requireddef handleJsonResp(ws: StandaloneWSClient)(  implicit ec: ExecutionContext): Future[JsValue] =
  ws.url("...").get().map { response =>response.body[JsValue]
  }

Note that there is a special case: when you are streaming the response, then you should get the body as a Source:

import scala.concurrent.ExecutionContextimport org.apache.pekko.util.ByteStringimport org.apache.pekko.stream.scaladsl.Sourceimport play.api.libs.ws.StandaloneWSClientdef useWSStream(ws: StandaloneWSClient)(implicit ec: ExecutionContext) =
  ws.url("...").stream().map { response => val source: Source[ByteString, _] = response.bodyAsSource     val _ = source // do something with source
  }

To POST, you should pass in a type which has an implicit class mapping of BodyWritable:

import scala.concurrent.ExecutionContextimport play.api.libs.ws.DefaultBodyWritables._ // requireddef postExampleString(ws: play.api.libs.ws.StandaloneWSClient)(  implicit ec: ExecutionContext) = {  val stringData = "Hello world"
  ws.url("...").post(stringData).map { response => /* do something */ }
}

You can also define your own custom BodyReadable:

import play.api.libs.ws.BodyReadableimport play.api.libs.ws.ahc.StandaloneAhcWSResponsecase class Foo(body: String)implicit val fooBodyReadable = BodyReadable[Foo] { response =>
  import play.shaded.ahc.org.asynchttpclient.{ Response => AHCResponse }  val ahcResponse = response.asInstanceOf[StandaloneAhcWSResponse].underlying[AHCResponse]  Foo(ahcResponse.getResponseBody)
}

or custom BodyWritable:

import org.apache.pekko.util.ByteStringimport play.api.libs.ws.{ BodyWritable, InMemoryBody }implicit val writeableOf_Foo: BodyWritable[Foo] = {  // https://tools.ietf.org/html/rfc6838#section-3.2
  BodyWritable(foo => InMemoryBody(ByteString.fromString(foo.toString)), "application/vnd.company.category+foo")
}

Java

To use the default type mappings in Java, you should use the following:

import play.libs.ws.DefaultBodyReadables;import play.libs.ws.DefaultBodyWritables;

followed by:

public class MyClient implements DefaultBodyWritables, DefaultBodyReadables {    public CompletionStage<String> doStuff() {      return client.url("http://example.com").post(body("hello world")).thenApply(response ->response.body(string())
      );
    }
}

Note that there is a special case: when you are using a stream, then you should get the body as a Source:

class MyClass {public CompletionStage<Source<ByteString, NotUsed>> readResponseAsStream() {return ws.url(url).stream().thenApply(response ->response.bodyAsSource()
        );
    }
}

You can also post a Source:

class MyClass {public CompletionStage<String> doStuff() {Source<ByteString, NotUsed> source = fromSource();return ws.url(url).post(body(source)).thenApply(response ->response.body()
        );
    }
}

You can define a custom BodyReadable:

import play.libs.ws.ahc.*;import play.shaded.ahc.org.asynchttpclient.Response;class FooReadable implements BodyReadable<StandaloneWSResponse, Foo> {public Foo apply(StandaloneWSResponse response) {Response ahcResponse = (Response) response.getUnderlying();return Foo.serialize(ahcResponse.getResponseBody(StandardCharsets.UTF_8));
    }
}

You can also define your own custom BodyWritable:

public class MyClient {private BodyWritable<String> someOtherMethod(String string) {org.apache.pekko.util.ByteString byteString = org.apache.pekko.util.ByteString.fromString(string);      return new DefaultBodyWritables.InMemoryBodyWritable(byteString, "text/plain");
    }
}

Instantiating a standalone client

The standalone client needs Pekko to handle streaming data internally:

Scala

In Scala, the way to call out to a web service and close down the client:

package playwsclientimport org.apache.pekko.actor.ActorSystemimport org.apache.pekko.stream._import play.api.libs.ws._import play.api.libs.ws.ahc._import scala.concurrent.Futureobject ScalaClient {  import DefaultBodyReadables._
  import scala.concurrent.ExecutionContext.Implicits._

  def main(args: Array[String]): Unit = {// Create Pekko system for thread and streaming managementimplicit val system = ActorSystem()
    system.registerOnTermination {      System.exit(0)
    }implicit val materializer = SystemMaterializer(system).materializer// Create the standalone WS client// no argument defaults to a AhcWSClientConfig created from// "AhcWSClientConfigFactory.forConfig(ConfigFactory.load, this.getClass.getClassLoader)"val wsClient = StandaloneAhcWSClient()

    call(wsClient)
      .andThen { case _ => wsClient.close() }
      .andThen { case _ => system.terminate() }
  }  def call(wsClient: StandaloneWSClient): Future[Unit] = {
    wsClient.url("http://www.google.com").get().map { response =>  val statusText: String = response.statusText      val body = response.body[String]
      println(s"Got a response $statusText: $body")
    }
  }
}

You can also create the standalone client directly from an AsyncHttpClient instance:

object ScalaClient {  def main(args: Array[String]): Unit = {// Use import play.shaded.ahc.org.asynchttpclient._val asyncHttpClientConfig = new DefaultAsyncHttpClientConfig.Builder()
      .setMaxRequestRetry(0)
      .setShutdownQuietPeriod(0)
      .setShutdownTimeout(0).buildval asyncHttpClient = new DefaultAsyncHttpClient(asyncHttpClientConfig)val wsClient = new StandaloneAhcWSClient(asyncHttpClient)/// ...
  }
}

This is useful when there is an AsyncHttpClient configuration option that is not available in the WS config layer.

Java

In Java the API is much the same:

package playwsclient;import org.apache.pekko.actor.ActorSystem;import org.apache.pekko.stream.*;import com.typesafe.config.ConfigFactory;import play.libs.ws.*;import play.libs.ws.ahc.*;public class JavaClient implements DefaultBodyReadables {private final StandaloneAhcWSClient client;private final ActorSystem system;public static void main(String[] args) {// Set up Pekko materializer to handle streamingfinal String name = "wsclient";ActorSystem system = ActorSystem.create(name);system.registerOnTermination(() -> System.exit(0));Materializer materializer = SystemMaterializer.get(system).materializer();// Create the WS client from the `application.conf` file, the current classloader and materializer.StandaloneAhcWSClient ws = StandaloneAhcWSClient.create(AhcWSClientConfigFactory.forConfig(ConfigFactory.load(), system.getClass().getClassLoader()),materializer);JavaClient javaClient = new JavaClient(system, ws);javaClient.run();
    }JavaClient(ActorSystem system, StandaloneAhcWSClient client) {this.system = system;this.client = client;
    }public void run() {client.url("http://www.google.com").get()
                .whenComplete((response, throwable) -> {String statusText = response.getStatusText();String body = response.getBody(string());System.out.println("Got a response " + statusText);
                })
                .thenRun(() -> {try {client.close();
                    } catch (Exception e) {e.printStackTrace();
                    }
                })
                .thenRun(system::terminate);
    }
}

Likewise, you can provide the AsyncHttpClient client explicitly from configuration:

public class JavaClient implements DefaultBodyReadables {     public static void main(String[] args) { // ...// Set up AsyncHttpClient directly from configAsyncHttpClientConfig asyncHttpClientConfig =new DefaultAsyncHttpClientConfig.Builder()
                .setMaxRequestRetry(0)
                .setShutdownQuietPeriod(0)
                .setShutdownTimeout(0)
                .build();AsyncHttpClient asyncHttpClient = new DefaultAsyncHttpClient(asyncHttpClientConfig);    // Set up WSClient instance directly from asynchttpclient.WSClient client = new AhcWSClient(asyncHttpClient, materializer);// ...}
}

Caching

Play WS implements HTTP Caching through CachingAsyncHttpClient, AhcHTTPCache and CacheControl, a minimal HTTP cache management library in Scala.

To create a standalone AHC client that uses caching, pass in an instance of AhcHttpCache with a cache adapter to the underlying implementation. For example, to use Caffeine as the underlying cache, you could use the following:

import scala.concurrent.Futureimport java.util.concurrent.TimeUnitimport com.github.benmanes.caffeine.cache.{ Caffeine, Ticker }import play.api.libs.ws.ahc.StandaloneAhcWSClientimport play.api.libs.ws.ahc.cache.{  AhcHttpCache, Cache, EffectiveURIKey, ResponseEntry}class CaffeineHttpCache extends Cache {  val underlying = Caffeine.newBuilder()
    .ticker(Ticker.systemTicker())
    .expireAfterWrite(365, TimeUnit.DAYS)
    .build[EffectiveURIKey, ResponseEntry]()  def remove(key: EffectiveURIKey) =Future.successful(Option(underlying.invalidate(key)))  def put(key: EffectiveURIKey, entry: ResponseEntry) =Future.successful(underlying.put(key, entry))  def get(key: EffectiveURIKey) =Future.successful(Option(underlying getIfPresent key ))  def close(): Unit = underlying.cleanUp()
}def withCache(implicit m: org.apache.pekko.stream.Materializer): StandaloneAhcWSClient = {  implicit def ec = m.executionContext  val cache = new CaffeineHttpCache()  StandaloneAhcWSClient(httpCache = Some(new AhcHttpCache(cache)))
}

There are a number of guides that help with putting together Cache-Control headers: