mlflow-token

Obtain an access token for an MLFlow instance deployed behind OAuth2-proxy and keycloak.

This script will use your current setting of MLFLOW_TRACKING_URI to look for the keycloak redirect from it's OAuth2-proxy. From there it will start an OAuth device flow to allow you to obtain a valid access token. You can use this to update your MLFLOW_TRACKING_TOKEN by executing the command as

% export $(mlflow-token)

and following the prompt.

Usage In Jupyter Notebook

If you want to authenticate to an MLFlow instance from within a Jupyter notebook you can add the following lines to a cell:

import mlflow_token
mlflow_token.setup_mlflow_environment("https://mlflow-demo.software-dev.ncsa.illinois.edu/")

This will update the notebook's os.environ so you can immediately use the mlflow SDK. The token will eventually expire, so you may need to occasionally re-run the cell.

The cell will print the url for the user to visit and wait for the device flow to complete.