Jupyter Lab¶

JupyterLab is a powerful, web-based interactive development environment for notebooks, code, and data. This instance runs in your browser for portability and ease of use.

This guide explains how to launch and use the JupyterLab app through Open OnDemand (OOD).

Creating a python kernel¶

The site-wide python installation is quite bare-bones by design. In most cases, users will want to prepare a python virtual environment (venv) and kernel for jupyter before launching:

Load your desired python version
```
ml load python/{version}
```
Create your venv in the desired location
```
python3 -m venv /path/to/venv
```
Activate your venv
```
/path/to/venv/bin/activate
```

Install desired packages

pip3 install numpy scipy ...
# OR
pip3 install -r requirements.txt

Install necessary ipython packages
```
pip3 install ipykernel
```

Add venv as a kernel in Jupyter

python -m ipykernel install --user --name=myenv --display-name "My Project Environment"

(Optional) Deactivate venv
```
deactivate
```

Launching JupyterLab¶

Log in to muscadine’s Open OnDemand portal.
Navigate to Interactive Apps → JupyterLab.
Fill in the requested parameters:
- Cluster: choose the compute resource you want.
- Number of cores / memory: select the appropriate resources.
- Wall time: specify how long you expect your session to run.
Click Launch.

Once your job starts, click Connect to JupyterLab to open the environment in a new browser tab.

Using the environment¶

JupyterLab provides an IDE-like interface for notebooks, terminals, and files.
You can:
- Create and edit Python or other supported notebooks.
- In the top right, select your desired kernel (see Creating a python kernel)
- Open a new Terminal from the Launcher.
- Manage files in your home or project directories.
- Install additional Python packages in your environment (if permitted).

Tip

All notebooks and files are saved directly to your remote filesystem — no upload or download needed.

Stopping your session¶

When finished:

Close the JupyterLab browser tab.
Return to the OOD dashboard.
Under My Interactive Sessions, click Delete or Stop to end the job and free up resources.

Troubleshooting¶

Jupyter launches but my custom kernel never launches
- You skipped module load python which caused your venv to include the system python. recreate your venv after loading the python module.
If JupyterLab fails to start, check:
- The requested resources (wall time, cores) are available.
- The Interactive App logs for errors (View Output Log).
If your session disconnects, simply reconnect through My Interactive Sessions.