Skip to main content
Version: Release 24.3

Build Custom Container Plugin using Python SDK

Digital.ai Release brings in the new integration SDK experience for you to run tasks as containers, using any language or third-party library.

This quick-start helps you learn how to build custom container plugins using the Python SDK.

Note: In this quick-start guide, you will learn how to use Docker Desktop with Kubernetes enabled to create a local Kubernetes cluster on your computer. This setup uses a local development environment for building, installing, and running your first container-based task. For production use, the steps to use the template repository, develop a plugin, build a plugin, and install the plugin will be the same. However, to run the container-based tasks, you will set up Digital.ai Release with the Digital.ai Release Runner in a Kubernetes cluster. For more information, see Run Release Digital.ai Release Runner on Kubernetes cluster.

Before you Begin​

  • A computer that runs Windows, macOS, or Linux. For macOS-based computers with the M1 or M2 chips, macOS Ventura is needed. Additionally, enable Rosetta in Docker Desktop from the Feature in development section.
  • Ensure that your internal memory meets the following requirements for running Docker effectively:
    • Minimum: 8GB of internal memory (RAM)
    • Recommended: 16GB of internal memory (RAM) for optimal performance
  • You must have Administrator privileges on your system, allowing you to edit the /etc/hosts file.
  • Docker Desktop
  • Digital.ai Release
  • GitHub as the repository that hosts the Digital.ai Release Integration Python SDK template to build the plugin
  • Python 3.11
  • The pip installer tool for Python
  • Use an IDE of your choice. However, PyCharm is recommended.

Here's a detailed step-by-step.

Step 1 — Set up GitHub Project for Building Container Plugins​

In this step, you will set up a new Git project that will contain the project files for building container-based integration plugins.

Step 1.1 — Create a New Repository​

Create a duplicate of the Digital.ai Release Integration Python SDK template project to start developing your own container-based integration plugins.

  1. In GitHub, go to the Digital.ai Release Integration Python SDK template.
  2. Click Use this template > Create a new repository to create a new repository with the contents of this sample template.
  3. Provide a repository name, based on the following naming convention.

The best practice is to use the following naming convention to develop container-based integration plugins:

[publisher]-release-[target]-integration

Sample:

acme-release-workshop-integration

For more details, see how to Create a Git repository.

Step 1.2 — Checkout the New Repository​

Open the terminal and navigate to a directory where you want to store your project, and run the following command:

git clone https://github.com/digital-ai/acme-release-workshop-integration.git

The repository is cloned in your local system.

Step 1.3 — Configure the project.properties File​

Navigate to the project.properties file in the repository, and update as follows:

PLUGIN=acme-release-workshop-integration

Note: This is the name of the repository you created in Step 1.1.

Step 2 — Run a Development Instance of Digital.ai Release​

In this step, you will learn how to set up a development environment with Release running on Docker using the Template repository.

  1. Navigate to the dev-environment folder in the cloned repository, and start the Release and Runner environments by running the following command:
docker compose up -d --build
  1. Once the docker command is executed, the containers will start. Go to Docker Desktop and check the logs of the dev-environment-digitalai-release-1 container for it to display the following lines:
2023-04-26 11:48:21 2023-04-26 09:48:21.377 [main] {activemq.broker=embedded-broker} INFO  c.x.x.s.jetty.JettyServerListener - Digital.ai Release has started.
2023-04-26 11:48:21 2023-04-26 09:48:21.378 [main] {activemq.broker=embedded-broker} INFO  c.x.x.s.jetty.JettyServerListener - You can now point your browser to http://host.docker.internal:5516/
  1. Open a browser, and go to http://localhost:5516. The Digital.ai Release login screen opens.
  2. Log in with username=admin and password=admin.

At the end of this step, you would have started Release. Additionally, a container registry is created, which will be used in the next steps.

Step 3 — Configure your hosts File​

Note: This set up will vary for a production environment based on your container registry.

For all the containers to find each other, the easiest way is to use the hosts file.

Note: You need sudo or Admin privileges to perform this step based on the OS.

To allow the Release server to locate the container images of the integration being developed, a registry is set up within Docker for the development environment. You must include the address of this registry in your local machine's host's file.

For Unix / macOS

Add the following lines in the hosts file in the following location: /etc/hosts

127.0.0.1 digitalai.release.local
127.0.0.1 container-registry
127.0.0.1 host.docker.internal

For Windows

Add the following lines in the hosts file in the following location: C:\Windows\System32\drivers\etc\hosts

127.0.0.1 digitalai.release.local
127.0.0.1 container-registry
127.0.0.1 host.docker.internal

Step 4 — Set up your Development Environment​

To keep all the dependencies of Python isolated and prevent interference with other projects on your computer, you can create a virtual environment in Python.

  1. Open your terminal.
  2. Navigate to the project directory where you want to create the virtual environment.
  3. Create a new Python virtual environment in the project directory by running the following command:
python -m venv venv

This will create a new directory called venv in your project directory, which will contain all the Python dependencies. 4. Activate the virtual environment by running the following command: For Linux / macOS

source venv/bin/activate

For Windows

venv\Scripts\activate

Step 4.1 — Set up requirements.txt file to Manage Dependencies​

  1. Install the required libraries by running the following command:
pip install -r requirements.txt

This command will install the required Python packages listed in the requirements.txt file within the virtual environment. 2. You now have a virtual environment with all the necessary dependencies installed for your project.

Step 4.2 — Set up IDE for Coding​

  1. Open PyCharm.
  2. Click Configure Python interpreter link > Python 3.11. The Add Python Interpreter dialog opens.
  3. Click Existing radio button.
  4. Click OK.

Step 5 — Develop a Custom Container Plugin​

Define and code a new task, based on the Hello example task.

Step 5.1 — Define a Type in type-definitions.yaml​

Firstly, you will define the name and properties of the task, so that Digital.ai Release can handle it.

Note: The type-definitions.yaml file is newly introduced to replace synthetic.xml for creating plugins. This quick-start is explained using type-definitions.yaml.

  1. Navigate to the type-definitions.yaml file.
  2. Rename the following:
  • containerExamples.BaseTask to workshop.BaseTask.
  • containerExamples.Hello to workshop.Greet.
  1. Remove the following:
  • containerExamples.SetSystemMessage
  • containerExamples.ServerQuery
  • containerExamples.Server
types:
  workshop.BaseTask:
    extends: xlrelease.ContainerTask
    virtual: true

    hidden-properties:
      image:
        default: "@registry.url@/@registry.org@/@project.name@:@project.version@"
        transient: true
      iconLocation: test.png
      taskColor: "#667385"


  workshop.Greet:
    extends: workshop.BaseTask  # Don't forget to extend the new base task
    description: "Simple greeter task"

    input-properties:
      yourName:
        description: The name to greet
        kind: string
        default: World

    output-properties:
      greeting:
        kind: string

With this metadata and other artifacts like icons, Release will be able to display the task in the UI and execute it. However, when building, it will be put into a zip file and will be uploaded to the Release server.

Step 5.2 — Refine Python Code to Create Custom Plugin​

Based on type-definitions.yaml file, the Python SDK will scan the src directory for Python classes with the same name as the type definition.

To develop a custom container plugin:

  1. In the hello.py file, rename the Hello class to Greet class.
  2. Rename hello.py file to greet.py file.
  3. Delete the following unused files:
  • sample_release_api_task.py
  • sample_server_task.py

You can now build the plugin.

Step 6 — Build and Run Container-based Plugin​

A Container-based integration plugin consists of two parts:

  • The metadata in the plugin zip file. You will install this file into Digital.ai Release.
  • The container image containing the code that will be executed when a task runs.

Learn how to package a plugin and publish the image.

  1. Configure the plugin and registry details in the project.properties file.
  2. Open the terminal and navigate to the root directory of project, and run the required command:

For Unix/mac OS run the following commands:

  • Builds the zip, image and pushes the image to the configured registry 
    sh build.sh
  • Builds the zip 
    sh build.sh --zip
  • Builds the image and pushes the image to the configured registry 
    sh build.sh --image

For Windows run the following commands:

  • Builds the zip, image and pushes the image to the configured registry 
    build.bat
  • Builds the zip 
    build.bat --zip
  • Builds the image and pushes the image to the configured registry 
    build.bat --image

Step 7 — Install Plugin zip into Release​

There are two ways to install the plugin into Release.

Install Plugin via CLI​

Update the Release server details in .xebialabs/config.yaml

Run the command for Unix / macOS:

sh build.sh --upload

Run the command for Windows:

build.bat --upload

The above command builds the zip and image and uploads the zip to the release server.

Install Plugin via UI​

In the Release UI, use the Plugin Manager interface to upload the zip from build. The zip takes the name of the project, for example release-integration-template-python-0.0.1.zip.

  1. On the top-right corner, click Settings > Manage plugins.
  2. In the left-navigation pane, click Installed plugins.
  3. Click Upload to upload the plugin file from your local machine.

Note: Browse the build directory of the acme-release-workshop-integration project in your local system. The plugin zip takes the name of the project specified in project.properties file. For example, acme-release-workshop-integration-0.0.1.zip.

  1. Select the required file and click Upload. The plugin is now available in your Release instance.

Step 8 — Run your First Custom Container Plugin​

  1. Create a template in Release.
  2. Select the newly built Greet container task and add to the template.
  3. Start a Release to run the task.

Note: If you change the task logic in Python files, just run the build script with --image flag to update the plugin image. If you change type-definitions.yaml, run the plugin script with --upload flag to update plugin types on Release. No restart is required.