Jupyter Notebook

Gitpod supports Jupyter Notebooks through support of the VS Code Jupyter extension. Providing a specific editing experience for those working in data science, data analysis, machine learning, artificial inelligence (AI), scientific research or education.

Jupyter Extension in VS Code

Gitpod supports Jupyter Notebooks by use of the Jupyter extension in VS Code. The following instructions apply to both VS Code Desktop and VS Code Browser. Whilst you can run a Jupyter notebook server or Jupyter lab server in a Gitpod workspace, exposed on a running port these approaches are not compatible with Gitpod Workspace timeouts as they do not actively tell Gitpod when work (e.g. keystrokes) are being executed in the running server.

Configuring your .gitpod.yml for the Jupyter extension

Note: Configuring a Jupyter Notebook extension is not “global” to all Organization users. You must include the extension definition in each repo that requires it.

Step 1: Create a .gitpod.yml in your repository

To install the VS Code extension, you’ll need a committed .gitpod.yml in the repository that your Gitpod workspace is opened for.

Tip: Use gp init to generate a scaffolded .gitpod.yml.

Step 2: Define your installation tasks

You must configure your Gitpod workspace with the VS Code extension, and the installation of any related packages. See the following, which:

1. Installs requirements.txt to install Python packages with pip.
2. Installs both the Jupyter extension and Python VS Code extension(s).
3. Opens the quickstart-demo.ipynb file when the workspace starts.

tasks:
    - name: Install required packages & Open Jupyter Notebook
      init: pip install -r requirements.txt
      command: gp open quickstart-demo.ipynb

vscode:
    extensions:
        - ms-toolsai.jupyter
        - ms-python.python

Caption: An example .gitpod.yml with Python and the Jupyter extension.

Step 3: Commit and test your configuration

Changes to the .gitpod.yml only apply to new workspaces. Commit your .gitpod.yml to git and restart a new worksapce to test your changes.

Tip: You can use the gp validate command to test configuration changes to the .gitpod.yml without committing.

Accessing a Jupyter Notebook from VS Code

1. Launch a new workspace from your repository. Gitpod will automatically initiate the Jupyter Notebook as per your configuration.
2. Create and manage new notebooks, write and execute code, and persist your work, all within a Gitpod workspace.

Advantages of the Jupyter Notebook VS Code extension

1. An editor and a notebook - You have the full power of the VS Code editor, making it easy to transition between basic notebook work, and data tasks requiring a full editor.
2. Collaboration between data science and data engineers - Using notebooks in VS Code means that you don’t have one setup for notebooks, and another for VS Code, making it simpler to maintain a single setup, encouraging collaboration between data scientists and engineers.
3. Advanced editor features - VS Code has many personalisation options for themes, etc.

Examples

For a practical reference, explore our Jupyter Notebook template on GitHub.