This Declarative Automation Bundles example demonstrates how to use a private wheel package from a job.

If you are using notebooks, you can use the approach documented in Notebook-scoped Python libraries to install wheels from a private repository in a notebook. You can use the workaround documented here if you are not using notebooks.

• Databricks CLI v0.235.0 or above
• Python 3.10 or above

You can refer to private wheel files from job libraries or serverless environments by downloading the wheel and making it part of your bundle deployment.

To emulate this for this example, we will download a wheel from PyPI, include it in deployment, and refer to it from job configuration.

First, download the wheel to the dist directory:

pip download -d dist cowsay==6.1

Next, update the host field under workspace in databricks.yml to the Databricks workspace you wish to deploy to.

Run databricks bundle deploy to upload the wheel and deploy the jobs.

Run databricks bundle run to run either job.

Example output:

$ databricks bundle run
Run URL: https://...

2024-11-27 13:23:01 "[dev pieter_noordhuis] Example to demonstrate using a private wheel package on serverless" TERMINATED SUCCESS
  _____________
| Hello, world! |
  =============
             \
              \
                ^__^
                (oo)\_______
                (__)\       )\/\
                    ||----w |
                    ||     ||