mlflow-go server fails with AttributeError: module 'mlflow' has no attribute 'server'

[joelrobin18]

Fix #134

Actual behavior:
The CLI references mlflow.server.file in mlflow_go_backend/cli.py without ever importing the mlflow.server submodule, so Python raises an AttributeError and exits.

Expected behavior:
The server should import the mlflow.server module, locate the static folder, initialize the backend store, and start serving the MLflow UI without error.

Root cause:
mlflow.server is never imported, so it isn’t attached to the mlflow namespace. Accessing mlflow.server.file therefore fails.

Proposed fix:
Add an explicit import mlflow.server at the top of mlflow_go_backend/cli.py to ensure mlflow.server exists before accessing its attributes.

[nojaf]

Hello @joelrobin18,

Thank you for trying out Mlflow Go! We apologize that this has been your first experience.

I was able to reproduce the problem and verify your fix. This looks reasonable to me.

CI is failing on main, and we will look into that separately.

We can publish this fix shortly after it is merged.

[nojaf]

Turns out we need to fix CI before we can release another version.
We will try to do this this week.