viniciusccarvalho / kodein-cloud-functions

Kodein adapters for several FaaS providers

Introduction

This project provides adapters to different cloud providers allowing developers to write Kotlin functions that are not tied to any specific cloud vendor.

Although it is highly inspired on the amazing Spring Cloud Function, this project is far from providing anything even close to the functionality provided by the Spring project. This is simply an idea that may fly with others one day

How to use

I took a lot of liberty on using Kodein as the DI framework for the project, many of the decisions I made may not be the best, and hopefuly someone will correct it via a PR :)

In order to use the project you must first import it on your pom:

<dependency>
    <groupId>io.igx.kotlin.cloud.functions</groupId>
    <artifactId>kodein-cloud-functions-aws-adapter</artifactId>
</dependency>

The project uses java ServiceLoader mechanism to find Kodein modules on the classpath. You need to create a class that implements io.igx.kotlin.cloud.functions.ModuleLoader such as

class SampleModuleLoader : ModuleLoader {
    override fun getModules(): List<Kodein.Module> {
        val module = Kodein.Module("samples") {
            bind<Function1<Foo, Bar>>(tag = "fooBar") with singleton { FooBar() } (1)
            bind(tag ="reqRes") from singleton { {r:Request -> Response(200)} } (2)
        }
        return listOf(module)
    }
}

1. Binding a top level class such as class FooBar : (Foo) → (Bar) {

2. Binding of a high level function, no type declared

And as with any ServiceLoader implementation you need to create a file named META-INF/services/io.igx.kotlin.cloud.functions.ModuleLoader with the name of all the ServiceLoader implementations you have.

That’s pretty much all it takes, no dependency on the cloud provider contract.

On AWS when deploying make sure you set the Handler to io.igx.kotlin.cloud.functions.aws.KodeinStreamHandler and that you set a FUNCTION_NAME environment variable to match the tag attribute of your bind

Next steps

• Improve docs a lot

• Better reflection handling

• Azure, OpenWhisk support