Installation

lely is a Python tool (3.11 or newer), on PyPI.

In a project

With uv, as a development dependency of the project that has the bundle:

uv add --dev lely
uv run lely --version

A plugin that is a package of yours is found this way, because lely is installed next to it.

On its own

uvx lely --version        # run it once, installing nothing
uv tool install lely      # or keep the command

Installed this way lely has the plugins it comes with and a class in your repository (./ops/steps.py:Class), and not a plugin that is a package of its own.

From a checkout

git clone https://github.com/kostavo-oss/lely && cd lely
uv sync
uv run lely --version

What else it needs

• The Databricks CLI, for the bundle plugin — a version with the direct engine (generally available since v1.3.0). lely has been run with v1.19.0.
• Credentials for a workspace, the way the Databricks CLI and SDK already read them: a profile in ~/.databrickscfg (--profile), or the DATABRICKS_* variables.
• git, if plans are saved to files: a plan file is held to the git tree it was made on.

lely doctor

says which of these it finds, which workspace it reaches and as whom, and whether each program a step runs is there.

For your editor

lely schema -o lely.schema.json

writes a JSON Schema of the config, built from the plugins this project uses. Make

# yaml-language-server: $schema=lely.schema.json

the first line of lely.yml, and the editor completes uses: and checks each step's with: against its plugin's own options as you type.