Description

This package contains the Kubernetes implementation of the deployer interface of pypz. It enables the deployment of the pipelines to a Kubernetes cluster.

Check the documentation for more details.

Install

The python artifact is hosted on https://pypi.org/, so you can install it via pip:

pip install pypz-k8s-deployer

If you want to work on it locally, then you should install in editable mode:

pip install -e ./deployers/k8s

Test

You can run tests against an existing Kubernetes cluster, however, it is more convenient and safe to use Kind for this purpose.

1. Follow the instructions to install Kind. Note that Kind prepares the required config file ~/.kube/config. If you work with your own cluster, make sure, you have this file already in place.
2. Install the subproject in editable more.

pip install -e ./deployers/k8s

3. Start the cluster

kind create cluster

4. Build the test image from the project root

docker build -t pypz-test-base -f deployers/k8s/test/resources/docker/Dockerfile .

5. Load the test image onto the Kind node

kind load docker-image pypz-test-base

6. To run the tests locally, you need to execute the following command:

python -m unittest discover .\deployers\k8s\test\ -p "*.py"

7. Delete cluster

kind delete cluster

Build

Before you build, you will need to install the "build" package:

pip install build

Optionally, you might want to set a version for local testing. Using setuptools-scm, you can pretend a version via setting the following environment variable:

# Powershell
$env:SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYPZ_K8S_DEPLOYER="v1.2.3"

# Linux
export SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYPZ_K8S_DEPLOYER="v1.2.3"

Then from the subproject's root, you will need to execute the following command:

python -m build

It will create the source distribution and the wheel file in the "dist" folder.