Development cycle#

This section covers developing production-ready workflows for Union.ai.

Authentication	How to authenticate with Union.ai.
Project structure	Best practices in organizing a Union.ai workflow project repository.
Projects and domains	Understanding projects and domains in Union.ai.
Building workflows	Best practices in strucuring your workflows.
Setting up a project	Create a project on Union.ai and initialize a workflow directory on your local machine.
Local dependencies	Install the required dependencies locally.
Remote dependencies with ImageSpec	Use `ImageSpec` to specify the dependencies needed in the containers that will run your tasks on Union.ai.
Running your code	Use different deploy and run commands for different steps in the development cycle.
Overriding parameters	Use `with_overrides` to change parameters at execution time.
Details of union run	Programmatically perform Union.ai operations in Python.
Debugging with interactive tasks	Inspect and debug live task code directly in the Union.ai console.
Managing secrets	Create and manage secrets to connect to third-party services.
Managing apps	Create applications to allow external systems to run compute on Union.ai.
Task resource validation	How Union.ai handles workflows with unsatisfiable resource requests.
Running in a local cluster	Run your workflows in a local Kubernetes cluster on your machine.
CI/CD deployment	Automate workflow registration and execution.
UnionRemote	Programmatically perform Union.ai operations in Python.