You are viewing the RapidMiner Developers documentation for version 9.6 - Check here for latest version

JupyterHub on the RapidMiner Platform

Starting from version 9.6, we provide a JupyterHub instance as part of the RapidMiner platform deployment. We built this integration with collaboration in mind between coders and non-coders across the enterprise.

On this page, you will learn about the features of this environment, how to deploy it, and how to interact with data and processes stored in the platform deployment.

The integrated JupyterHub component is only available as part of the containerized RapidMiner platform deployment. If you haven’t adopted this deployment method yet, contact our support team if you need help getting started.

Deploy JupyterHub

The integrated JupyterHub component is shipped as a set of containers. As it is most often used as a non-linear, exploratory coding environment, we recommend deploying it as part of a RapidMiner platform deployment instance used for development and testing purposes.

Visit our deployment templates to quickly get started with it. You can get it by creating a new deployment or adding it to an existing one.

You can find a detailed list of configuration possibilities on the related Docker image reference page.

Log in to JupyterHub

After navigating to your RapidMiner Server instance and logging in, you will see a link to JupyterHub on the Server web interface.

Alternatively, you can find your JupyterHub instance by pointing your browser to http(s)://your.deployment.address/jupyter (this URL is configurable, see the image reference for more details). If not logged in yet, you will be redirected to the RapidMiner Server login page.

If you log out from either your Jupyter or your RapidMiner Server, you will be logged out from both components.

We only support an interactive usage of the Jupyter environment. Don’t expect your notebook code to continue running after you have logged off.

Use JupyterHub

By default, we have enabled JupyterLab as a user interface. If you prefer the classic Jupyter notebook environment, click Help -> Launch Classic Notebook.

Each user's JupyterLab instance inside JupyterHub is running in a separate Docker container. A default Python environment is provisioned, including the most popular Python packages used for data science. The working directory for each user's instance is separated, and currently there's no filesystem level access to the contents of the RapidMiner Server repository.

Data stored in the RapidMiner Server repository can be accessed by using the pre-installed python-rapidminer library. To see examples, just create a new notebook in your JupyterLab instance and follow the instructions there on how to read and write data to the RapidMiner Server repository, and how to run processes and get their results.

You can use the RapidMiner Server repository in your deployment without interactive authentication in your notebook.