jupyterlab-spark-webui 0.1.0

JupyterLab extension to open the Spark Web UI via the notebook proxy

jupyterlab-spark-webui

A JupyterLab extension that adds a Spark UI menu entry to the menu bar. When clicked, it checks whether Apache Spark is running on port 4040 and opens the Spark jobs page as an embedded JupyterLab tab via jupyter-server-proxy.

Requirements

Requirement	Version
JupyterLab	≥ 4.0
jupyter-server-proxy	≥ 4.0
Python	≥ 3.8

Installation

pip install jupyterlab-spark-webui

No additional steps are needed. JupyterLab discovers the extension automatically after installation.

Verify installation

jupyter labextension list
# jupyterlab-spark-webui v0.1.0 enabled OK

Usage

1. Start a Spark session in a notebook (this starts the Spark UI on port 4040).
2. Click Spark UI in the JupyterLab menu bar.
3. Click Open Spark UI.
   • If Spark is running → the jobs page opens as a JupyterLab tab.
   • If Spark is not running → a dialog displays "No Spark found (port 4040)."

The extension resolves the proxy URL from the current page path, so it works in JupyterHub deployments at any base URL prefix (e.g. /notebook/<namespace>/<name>/lab).

Uninstallation

pip uninstall jupyterlab-spark-webui

---

Development

Prerequisites

• Python ≥ 3.8 with uv
• Node.js ≥ 18 with npm

Setup and build

git clone https://github.com/your-org/jupyterlab-spark-webui
cd jupyterlab-spark-webui

./build.sh          # build everything + produce dist/*.whl
./start-lab.sh      # launch JupyterLab with the extension loaded

Pass --no-wheel to skip the wheel build during development iteration:

./build.sh --no-wheel

Project structure

├── src/
│   └── index.ts          # Extension entry point (TypeScript)
├── jupyterlab_spark_webui/
│   ├── __init__.py       # Python package / labextension path declaration
│   └── labextension/     # Built JS bundle (generated by build.sh)
├── style/
│   └── index.css
├── pyproject.toml        # Python package metadata and build config
├── package.json          # npm package and TypeScript dependencies
├── tsconfig.json         # TypeScript compiler config
├── build.sh              # Full build + deploy + wheel script
└── start-lab.sh          # Launch JupyterLab using the uv environment

---

Publishing to PyPI

1. Prerequisites

• A PyPI account
• A PyPI API token (generate one at Account settings → API tokens)

2. Update metadata

Edit pyproject.toml and set the correct authors, [project.urls], and bump version before each release.

3. Build

./build.sh          # produces dist/jupyterlab_spark_webui-<version>-py3-none-any.whl

4. Publish

With uv (recommended):

uv publish --token pypi-<your-token>

Or with twine:

pip install twine
twine upload dist/* --username __token__ --password pypi-<your-token>

5. Test on TestPyPI first (optional but recommended)

# Publish to the test registry
uv publish --publish-url https://test.pypi.org/legacy/ --token pypi-<test-token>

# Install from the test registry
pip install --index-url https://test.pypi.org/simple/ jupyterlab-spark-webui

Checklist before publishing

• version bumped in pyproject.toml and package.json
• authors and [project.urls] filled in
• ./build.sh ran successfully and dist/*.whl exists
• Extension verified working locally via ./start-lab.sh
• git tag v<version> created and pushed