Custom Node

Custom node's primary job is data transformation. This type of node is meant for processes which do not send requests. Let's create a simple custom node for data transformation.

Prerequisites

• Installed and running Orchesty.
• Connected SDK.

Creating Custom Node

Typescript

First, we create a new class in the worker in the src folder that will inherit from ACommonNode. The class doesn't do anything yet, we only set the name of the action that it will perform.

import ACommonNode from '@orchesty/nodejs-sdk/dist/lib/Commons/ACommonNode';
import ProcessDto from '@orchesty/nodejs-sdk/dist/lib/Utils/ProcessDto';

export const NAME = 'hello-world';

export default class HelloWorld extends ACommonNode {

    public getName(): string {
        return NAME;
    }

    public processAction(dto: ProcessDto): ProcessDto {
        return dto;
    }

}

PHP

First, we create a new class in the worker in the src folder that will inherit from CommonNodeAbstract. The class doesn't do anything yet, we only set the name of the action that it will perform.

namespace Pipes\PhpSdk\Mapper;

use Hanaboso\CommonsBundle\Process\ProcessDto;
use Hanaboso\PipesPhpSdk\CustomNode\CommonNodeAbstract;

final class HelloWorld extends CommonNodeAbstract
{

    public const NAME = 'hello-world';

    function getName(): string
    {
        return self::NAME;
    }

    function processAction(ProcessDto $dto): ProcessDto
    {
        return $dto;
    }

}

Note

Each process node must have a name identificator by which they are both registered and used within orchestration layer. By names, they are also listed in Admin, so it's best to keep them human-readable.

Data transformation

The ProcessDto object represents the data structure of the message that flows through the topology. Each node of the topology receives and resends data through this object. Thus, in the processAction method, we insert data into ProcessDto:

Typescript

    // ...
    public processAction(dto: ProcessDto): ProcessDto {
        return dto.setJsonData({ message: 'Hello world!' });
    }
    // ...

PHP

    // ...
    function processAction(ProcessDto $dto): ProcessDto
    {
        return $dto->setJsonData(['message' => 'Hello world']);
    }
    // ...

Registering into SDK container

Typescript

Now you need to register a new class in SDK container. This will make it available to the orchestration layer and we will be able to use it in topologies. We open the index.ts file in the src folder and register our HelloWorld class:

// ...
import { container } from '@orchesty/nodejs-sdk';
import HelloWorld from './HelloWorld';
// ...

export default function prepare(): void {
    // ...
    container.setCustomNode(new HelloWorld());
    // ...
}

PHP

Now you need to register a new class in SDK container. This will make it available to the orchestration layer and we will be able to use it in topologies. In the config folder, we create a custom_node.yaml file and register our HelloWorld class:

# ./config/custom_node.yaml
services:
    _defaults:
        public: '%public.services%'
        
    hbpf.custom_node.hello-world:
        class: Pipes\PhpSdk\Mapper\HelloWorld
        arguments:
            - '@hbpf.application_install.repository'

Using in orchestration layer

Let's go back in Admin to our topology that we have already prepared. Select custom node in the toolbar and add it to the topology. In the right sidebar, we will now release the settings for the action that the node should perform. Here we have the worker we registered in the previous tutorial and also the hello-world action we created in it.

At the end of the topology, we add a user task to check the transformation.

Now we save the topology. Since we have edited the published topology, we can see that a new version of it was automatically created, which will need to be published and enabled again in order to run it.

info

Orchesty always creates a new version when editing a published topology. When it is run, we then inactivate the previous active version. This stops receiving new messages, but all running process instances continue processing.

So we start the topology, but this time we don't put any data into it. Switching to the User Tasks tab, we can see that there is a message with an empty body in the user node.

We use the Approve button to send the message to the next node. The message should now appear in the user2 node. When we look at the body of the message, we can see the data that our first custom node inserted.

That's all. The principle of custom nodes is simple and its use is up to us. It can be used to prepare data mappings for integration processes. But using this principle, we can also orchestrate the microservice architecture and use the actions in custom nodes instead of the REST APIs of the individual microservices.

In the following tutorial, we will show how to create a connector to call the REST API of an external service.