Using Jupyter Lab as a Lightweight Remote Desktop

Jupyter Lab isn't just for data science - it's a powerful all-in-one environment offering:

• Programming (Python, R, etc.)
• Shell Access
• File Browser & Editor
• Plugins

All within your browser, without heavy windowing systems or VNC, making it a great alternative to traditional remote desktops!

Setup Instructions

Server Side

First, install Conda and set up a Conda environment.

Install Jupyter Lab

conda install -c conda-forge jupyterlab

Install kernels for Jupyter Lab:

Python (IPyKernel):

conda install -c conda-forge ipykernel

R (IRKernel):

conda install -c conda-forge r-irkernel

Launch Jupyter Lab and Make it accessible

Start a tmux session.

Start JupyterLab in it, listening on localhost:8888, with no password/token:

jupyter lab \
    --ip=127.0.0.1 --port 8888 \
    --allow-root \
    --no-browser \
    --IdentityProvider.token='' \
    --ServerApp.password='' \
    --ServerApp.allow_remote_access=True

Then open another tmux pane, set up push-pull-port, and push port 8888 to port $PRIVATE_PORT (localhost only) of remote public server $PUBLIC_HOST:

sh push-local-port.sh -l 8888 -r $PRIVATE_PORT -u $PUBLIC_USER -h $PUBLIC_HOST -x

Then detach from the tmux session.

Client Side

Set up push-pull-port.

Pull the remote port and forward it to your local 8888 port:

sh pull-remote-port.sh -r $PRIVATE_PORT -l 8888 -u $PUBLIC_USER -h $PUBLIC_HOST

Access in Browser: Open http://localhost:8888/lab.

Exporting Jupyter Notebooks

You can export a Jupyter Notebook to a Python script or Markdown file by using nbconvert to convert notebooks:

Export to Python

jupyter nbconvert --to script your_notebook.ipynb

Export to Markdown

jupyter nbconvert --to markdown your_notebook.ipynb

This will create your_notebook.py or your_notebook.md in the current directory.