Skip to content

Latest commit

 

History

History
331 lines (239 loc) · 9.86 KB

README.md

File metadata and controls

331 lines (239 loc) · 9.86 KB
Raw

Fn Java Functions Developer Kit (FDK)

CircleCI

This project adds support for writing functions in Java on the Fn platform, with full support for Java 9 as the default out of the box.

FAQ

Some common questions are answered in our FAQ.

Quick Start Tutorial

By following this step-by-step guide you will learn to create, run and deploy a simple app written in Java on Fn.

Pre-requisites

Before you get started you will need the following things:

Install the Fn CLI tool

To install the Fn CLI tool, just run the following:

curl -LSs https://raw.githubusercontent.com/fnproject/cli/master/install | sh

This will download a shell script and execute it. If the script asks for a password, that is because it invokes sudo.

Your first Function

1. Create your first Java Function:

$ mkdir hello-java-function && cd hello-java-function
$ fn init --runtime=java --name your_dockerhub_account/hello
Runtime: java
function boilerplate generated.
func.yaml created

This creates the boilerplate for a new Java Function based on Maven and Oracle Java 9. The pom.xml includes a dependency on the latest version of the Fn Java FDK that is useful for developing your Java functions.

You can now import this project into your favourite IDE as normal.

2. Deep dive into your first Java Function:

We'll now take a look at what makes up our new Java Function. First, lets take a look at the func.yaml:

$ cat func.yaml
name: your_dockerhub_account/hello
version: 0.0.1
runtime: java
cmd: com.example.fn.HelloFunction::handleRequest

The cmd field determines which method is called when your funciton is invoked. In the generated Function, the func.yaml references com.example.fn.HelloFunction::handleRequest. Your functions will likely live in different classes, and this field should always point to the method to execute, with the following syntax:

cmd: <fully qualified class name>::<method name>

For more information about the fields in func.yaml, refer to the Fn platform documentation about it.

Let's also have a brief look at the source: src/main/java/com/example/fn/HelloFunction.java:

package com.example.fn;

public class HelloFunction {

    public String handleRequest(String input) {
        String name = (input == null || input.isEmpty()) ? "world"  : input;

        return "Hello, " + name + "!";
    }

}

The function takes some optional input and returns a greeting dependent on it.

3. Run your first Java Function:

You are now ready to run your Function locally using the Fn CLI tool.

$ fn build
Building image your_dockerhub_account/hello:0.0.1
Sending build context to Docker daemon  14.34kB
Step 1/11 : FROM fnproject/fn-java-fdk-build:jdk9-latest as build-stage
 ---> 5435658a63ac
Step 2/11 : WORKDIR /function
 ---> 37340c5aa451

...

Step 5/11 : RUN mvn package dependency:copy-dependencies -DincludeScope=runtime -DskipTests=true -Dmdep.prependGroupId=true -DoutputDirectory=target --fail-never
---> Running in 58b3b1397ba2
[INFO] Scanning for projects...
Downloading: https://repo.maven.apache.org/maven2/org/apache/maven/plugins/maven-compiler-plugin/3.3/maven-compiler-plugin-3.3.pom
Downloaded: https://repo.maven.apache.org/maven2/org/apache/maven/plugins/maven-compiler-plugin/3.3/maven-compiler-plugin-3.3.pom (11 kB at 21 kB/s)

...

[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 2.228 s
[INFO] Finished at: 2017-06-27T12:06:59Z
[INFO] Final Memory: 18M/143M
[INFO] ------------------------------------------------------------------------

...

Function your_dockerhub_account/hello:0.0.1 built successfully.

$ fn run
Hello, world!

The next time you run this, it will execute much quicker as your dependencies are cached. Try passing in some input this time:

$ echo -n "Universe" | fn run
...
Hello, Universe!

4. Testing your function

The Fn Java FDK includes a testing library providing useful JUnit 4 rules to test functions. Look at the test in src/test/java/com/example/fn/HelloFunctionTest.java:

package com.example.fn;

import com.fnproject.fn.testing.*;
import org.junit.*;

import static org.junit.Assert.*;

public class HelloFunctionTest {

    @Rule
    public final FnTestingRule testing = FnTestingRule.createDefault();

    @Test
    public void shouldReturnGreeting() {
        testing.givenEvent().enqueue();
        testing.thenRun(HelloFunction.class, "handleRequest");

        FnResult result = testing.getOnlyResult();
        assertEquals("Hello, world!", result.getBodyAsString());
    }

}

This test is very simple: it just enqueues an event with empty input and then runs the function, checking its output. Under the hood, the FnTestingRule is actually instantiating the same runtime wrapping function invocations, so that during the test your function will be invoked in exactly the same way that it would when deployed.

There is much more functionality to construct tests in the testing library. Testing functions is covered in more detail in Testing Functions.

5. Run using HTTP and the local Fn server

The previous example used fn run to run a function directly via docker, you can also use the Fn server locally to test the deployment of your function and the HTTP calls to your functions.

Open another terminal and start the Fn server:

$ fn start

Then in your original terminal create an app:

$ fn apps create java-app
Successfully created app: java-app

Now deploy your Function using the fn deploy command. This will bump the function's version up, rebuild it, and push the image to the Docker registry, ready to be used in the function deployment. Finally it will create a route on the local Fn server, corresponding to your function.

We are using the --local flag to tell fn to skip pushing the image anywhere as we are just going to run this on our local fn server that we started with fn start above.

$ fn deploy --app java-app --local
...
Bumped to version 0.0.2
Building image hello:0.0.2
Sending build context to Docker daemon  14.34kB

...

Successfully built bf2b7fa55520
Successfully tagged your_dockerhub_account/hello:0.0.2
Updating route /hello-java-function using image your_dockerhub_account/hello:0.0.2...

Call the Function via the Fn CLI:

$ fn call java-app /hello-java-function
Hello, world!

You can also call the Function via curl:

$ curl http://localhost:8080/r/java-app/hello-java-function
Hello, world!

6. Something more interesting

The Fn Java FDK supports flexible data binding to make it easier for you to map function input and output data to Java objects.

Below is an example to of a Function that returns a POJO which will be serialized to JSON using Jackson:

package com.example.fn;

public class PojoFunction {

    public static class Greeting {
        public final String name;
        public final String salutation;

        public Greeting(String salutation, String name) {
            this.salutation = salutation;
            this.name = name;
        }
    }

    public Greeting greet(String name) {
        if (name == null || name.isEmpty())
            name = "World";

        return new Greeting("Hello", name);
    }

}

Update your func.yaml to reference the new method:

cmd: com.example.fn.PojoFunction::greet

Now run your new function:

$ fn run
...
{"name":"World","salutation":"Hello"}

$ echo -n Michael | fn run
...
{"name":"Michael","salutation":"Hello"}

7. Where do I go from here?

Learn more about the Fn Java FDK by reading the next tutorials in the series. Also check out the examples in the examples directory for some functions demonstrating different features of the Fn Java FDK.

Configuring your function

If you want to set up the state of your function object before the function is invoked, and to use external configuration variables that you can set up with the Fn tool, have a look at the Function Configuration tutorial.

Input and output bindings

You have the option of taking more control of how serialization and deserialization is performed by defining your own bindings.

See the Data Binding tutorial for other out-of-the-box options and the Extending Data Binding tutorial for how to define and use your own bindings.

Asynchronous workflows

Suppose you want to call out to some other function from yours - perhaps a function written in a different language, or even one maintained by a different team. Maybe you then want to do some processing on the result. Or even have your function interact asynchronously with a completely different system. Perhaps you also need to maintain some state for the duration of your function, but you don't want to pay for execution time while you're waiting for someone else to do their work.

If this sounds like you, then have a look at the Fn Flow quickstart.

Get help

Contributing

Please see CONTRIBUTING.md.