Add MLFlow documentation

# MLFlow

Covering specifics of using MLFlow at PIC.

For generic help check the [official documentation](https://mlflow.org/docs/latest/index.html#)

## Running the UI

From a shell in `jupyter.pic.es`

Activate a conda|mamba environment with mlflow installed

    $  micromamba activate my_mlflow_environment

Start the MLflow UI in a free port of your choice, e.g. 50000

    (mlflow) $ mlflow ui --port 50000

    [2024-09-16 11:47:28 +0200] [962] [INFO] Starting gunicorn 22.0.0

    [2024-09-16 11:47:28 +0200] [962] [INFO] Listening at: http://127.0.0.1:50000 (962)

    [2024-09-16 11:47:28 +0200] [962] [INFO] Using worker: sync

    [2024-09-16 11:47:28 +0200] [971] [INFO] Booting worker with pid: 971

    [2024-09-16 11:47:28 +0200] [972] [INFO] Booting worker with pid: 972

    [2024-09-16 11:47:28 +0200] [973] [INFO] Booting worker with pid: 973

    [2024-09-16 11:47:28 +0200] [974] [INFO] Booting worker with pid: 974

Navigate to `jupyter.pic.es/user/{username}/proxy/{port}/` in your web browser to access the web UI, where `{username}` is the name of your user in `jupyter.pic.es` and `{port}` is the port number you chose to start the MLFlow UI on. The resulting url could be `jupyter.pic.es/user/torradeflot/proxy/50000/`.

Notice the trailing `/`, **this is important**, otherwise the static content won't be retrieved and the UI won't work.