Using Caffeine’s `AsyncLoadingCache` in a Vert.x application

This document will show you how to use Caffeine’s AsyncLoadingCache in a Vert.x application.

What you will build

You will build an application that rotates images in a web page.

The images will be downloaded by the server from a public API exposed at https://http.cat.

Each image represents an HTTP status code as a 🐱.

Ready?

Note	Images created by Tomomi Imura (@girlie_mac)

What you need

A text editor or IDE,
Java 11 or higher,
Maven or Gradle.

Create a project

The code of this project contains Maven and Gradle build files that are functionally equivalent.

Using Maven

Here is the content of the pom.xml file you should be using:

Maven pom.xml

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">

  <modelVersion>4.0.0</modelVersion>

  <groupId>io.vertx.howtos</groupId>
  <artifactId>async-loading-cache-caffeine-howto</artifactId>
  <version>1.0-SNAPSHOT</version>

  <properties>
    <vertx.version>4.4.0</vertx.version>
    <maven.compiler.source>11</maven.compiler.source>
    <maven.compiler.target>11</maven.compiler.target>
  </properties>

  <dependencyManagement>
    <dependencies>
      <dependency>
        <groupId>io.vertx</groupId>
        <artifactId>vertx-stack-depchain</artifactId>
        <version>${vertx.version}</version>
        <type>pom</type>
        <scope>import</scope>
      </dependency>
    </dependencies>
  </dependencyManagement>

  <dependencies>
    <dependency>
      <groupId>io.vertx</groupId>
      <artifactId>vertx-core</artifactId>
    </dependency>
    <dependency>
      <groupId>io.vertx</groupId>
      <artifactId>vertx-web</artifactId>
    </dependency>
    <dependency>
      <groupId>io.vertx</groupId>
      <artifactId>vertx-web-client</artifactId>
    </dependency>
    <dependency>
      <groupId>com.github.ben-manes.caffeine</groupId>
      <artifactId>caffeine</artifactId>
      <version>3.1.5</version>
    </dependency>
  </dependencies>

  <build>
    <plugins>
      <plugin>
        <groupId>org.apache.maven.plugins</groupId>
        <artifactId>maven-compiler-plugin</artifactId>
        <version>3.10.1</version>
        <configuration>
          <release>11</release>
        </configuration>
      </plugin>
      <plugin>
        <groupId>org.codehaus.mojo</groupId>
        <artifactId>exec-maven-plugin</artifactId>
        <version>1.5.0</version>
        <configuration>
          <mainClass>io.vertx.howtos.caffeine.CatsVerticle</mainClass>
        </configuration>
      </plugin>
    </plugins>
  </build>

</project>

Using Gradle

Assuming you use Gradle with the Kotlin DSL, here is what your build.gradle.kts file should look like:

Gradle build.gradle.kts

plugins {
  java
  application
}

repositories {
  mavenCentral()
}

java {
  toolchain {
    languageVersion.set(JavaLanguageVersion.of(11))
  }
}

dependencies {
  implementation(platform("io.vertx:vertx-stack-depchain:4.4.0"))
  implementation("io.vertx:vertx-core")
  implementation("io.vertx:vertx-web")
  implementation("io.vertx:vertx-web-client")
  implementation("com.github.ben-manes.caffeine:caffeine:3.1.5")
}

application {
  mainClassName = "io.vertx.howtos.caffeine.CatsVerticle"
}

tasks.wrapper {
  gradleVersion = "7.4.1"
}

Web Page Implementation

The index.html web page consists mainly of:

an <img> tag in the body, and
a script that, after the page is loaded, changes the src attribute of the image.

index.html

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <title>Rotating Cats Page</title>
  <script type="application/javascript">
    window.onload = function () { // (2)
      const codes = [ // (3)
        100,
        200, 201, 204, 206,
        301, 302, 303, 304, 307, 308,
        401, 403, 404, 406, 407, 409, 410, 412, 416, 418, 425, 451,
        500, 501, 502, 503, 504,
      ]
      setInterval(function () { // (4)
        let img = document.getElementById("cat-img"); // (5)
        img.src = "/api/cats/" + codes[Math.floor(Math.random() * codes.length)]; // (6)
      }, 250)
    }
  </script>
</head>
<body>
<div>
  <img src="/api/cats/200" id="cat-img" alt="Cat image"/> <!--(1)-->
</div>
</body>
</html>

img tag in the body with id set to cat-img
run the script function when the page is loaded
define some HTTP status codes
schedule a function to execute periodically at a fixed-delay of 250 milliseconds
retrieve the img element from the DOM using its id
update the src attribute using an HTTP status code chosen randomly

Server Implementation

We will need a Vert.x Web Client instance to fetch images:

Web Client setup

    request = WebClient.create(vertx)
      .request(GET, new RequestOptions().setHost("http.cat").setPort(443).setSsl(true))
      .as(BodyCodec.buffer())
      .expect(ResponsePredicate.SC_OK);

We will also need a cache because we do not want to overload the backend API:

Caffeine setup

    cache = Caffeine.newBuilder() // (1)
      .expireAfterWrite(Duration.ofMinutes(1)) // (2)
      .recordStats() // (3)
      .executor(cmd -> context.runOnContext(v -> cmd.run())) // (4)
      .buildAsync((key, exec) -> CompletableFuture.supplyAsync(() -> { // (5)
        Future<Buffer> future = fetchCatImage(key); // (6)
        return future.toCompletionStage(); // (7)
      }, exec).thenComposeAsync(Function.identity(), exec));

    vertx.setPeriodic(20000, l -> { // (8)
      CacheStats stats = cache.synchronous().stats();
      log.info("Stats: " + stats);
    });

create a cache builder
configure the cache to expire items after 1 minute
enable statistics recording
define an executor which invokes tasks on the verticle context
create an asynchronous loader, which much return a CompletableFuture, using the cache executor
fetch the cat image
convert the Vert.x Future to a CompletionStage
log cache statistics periodically

Note

The executor definition and the complex loader implementation are not strictly needed here.

Indeed, we will deploy a single verticle instance, bind the cache to a field and always invoke its methods from the event-loop. If that is your use-case, you may simplify the setup to:

    cache = Caffeine.newBuilder()
      .expireAfterWrite(Duration.ofMinutes(1))
      .recordStats()
      .buildAsync((key, exec) -> {
        Future<Buffer> future = fetchCatImage(key);
        return future.toCompletionStage().toCompletableFuture();
      });

If, however, you plan to deploy several instances of the verticle and to share the cache between them, stick to the previous implementation. It guarantees that the asynchronous loader is always invoked on the right context.

Fetching the cat image consists in sending the request to the backend using the corresponding HTTP status code as URI:

Fetching the image

  private Future<Buffer> fetchCatImage(int code) {
    return request.uri("/" + code)
      .send()
      .map(HttpResponse::body);
  }

Using Vert.x Web, creating the HTTP server for our API and static file is pretty straightforward:

Server setup

    Router router = Router.router(vertx);
    router.get("/api/cats/:id").produces("image/*").handler(this::handleImageRequest);
    router.get().handler(StaticHandler.create());

    vertx.createHttpServer()
      .requestHandler(router)
      .listen(8080);

    log.info("Server started on port 8080");

Here is how we will implement image request handling:

Handling image requests

  private void handleImageRequest(RoutingContext rc) {
    Integer code = Integer.valueOf(rc.pathParam("id")); // (1)
    CompletableFuture<Buffer> completableFuture = cache.get(code); // (2)
    Future<Buffer> future = Future.fromCompletionStage(completableFuture, context); // (3)
    future.onComplete(ar -> { // (4)
      if (ar.succeeded()) {
        rc.response()
          .putHeader("Cache-Control", "no-store") // (5)
          .end(ar.result());
      } else {
        rc.fail(ar.cause());
      }
    });
  }

retrieve the specified code from the request path
invoke Caffeine (the image will be loaded from the backend transparently, if needed)
convert the CompletableFuture returned by Caffeine to a Vert.x Future
on completion, send the image bytes (or the failure) to the client
instruct the browser to disable caching of the image (otherwise, it would query our server only once for a given code!)

Running the application

The CatsVerticle needs a main method:

Main method

  public static void main(String[] args) {
    Vertx vertx = Vertx.vertx(); // (1)
    vertx.deployVerticle(new CatsVerticle()); // (2)
  }

Create a Vert.x instance
Deploy CatsVerticle

You can run the application from:

your IDE, by running the main method from the CatsVerticle class, or
with Maven: mvn compile exec:java, or
with Gradle: ./gradlew run (Linux, macOS) or gradle run (Windows).

Browse to http://localhost:8080.

You should see the cat images rotating in the web page:

After some time, inspect the program output. You should read something like:

Mar 22, 2022 3:45:17 PM io.vertx.howtos.caffeine.CatsVerticle lambda$start$4
INFO: Stats: CacheStats{hitCount=52, missCount=28, loadSuccessCount=28, loadFailureCount=0, totalLoadTime=2514949257, evictionCount=0, evictionWeight=0}
Mar 22, 2022 3:45:37 PM io.vertx.howtos.caffeine.CatsVerticle lambda$start$4
INFO: Stats: CacheStats{hitCount=132, missCount=28, loadSuccessCount=28, loadFailureCount=0, totalLoadTime=2514949257, evictionCount=0, evictionWeight=0}
Mar 22, 2022 3:45:57 PM io.vertx.howtos.caffeine.CatsVerticle lambda$start$4
INFO: Stats: CacheStats{hitCount=212, missCount=28, loadSuccessCount=28, loadFailureCount=0, totalLoadTime=2514949257, evictionCount=0, evictionWeight=0}
Mar 22, 2022 3:46:17 PM io.vertx.howtos.caffeine.CatsVerticle lambda$start$4
INFO: Stats: CacheStats{hitCount=267, missCount=53, loadSuccessCount=52, loadFailureCount=0, totalLoadTime=3337599348, evictionCount=28, evictionWeight=28}
Mar 22, 2022 3:46:37 PM io.vertx.howtos.caffeine.CatsVerticle lambda$start$4
INFO: Stats: CacheStats{hitCount=344, missCount=56, loadSuccessCount=56, loadFailureCount=0, totalLoadTime=3480880213, evictionCount=28, evictionWeight=28}

Notice the changes in hitCount, missCount, or evictionCount.

Summary

This document covered:

the Vert.x web client for making HTTP requests,
the Vert.x web server and router,
the integration of Caffeine’s asynchronous loading cache in a Vert.x application,
Vert.x periodic tasks.

Using Caffeine’s AsyncLoadingCache in a Vert.x application