Launching Jupyter Labs in an easy way

For a much more simplified version of this repo, see jupyter-notebook-best-practice.

Usage

# Default Resolve
bin/lab

bin/lab duckdb
bin/lab http

How to define a Jupyter Lab

Three Highlights

• Free yourself of Python virtual environment
• Reproducible because of the dependencies lock mechanism
• As Easy as bin/lab xyz to launch a JupyterLab

Take duckdb for example

jupyter_lab(name="duckdb", requirements=["duckdb==0.5.1"])

• notebooks/duckdb is where we store the notebooks
• resolve duckdb is defined as duckdb = "3rdparty/python/duckdb.lock" in pants.toml
• run ./pants generate-lockfiles --resolve=duckdb to generate the lockfiles for dependencies

Name and Requirements

Name is used to:

• Locate where we store the notebooks: notebooks/{name}
• Set dependencies as the resolve name

The Jupyter Lab default is special, we kept all notebooks for default under notebooks.

Requirements are equiv to requirements.txt but we maintain it in a list here.

Everytime, we change the requirements, do not forget to run ./pants generate-lockfiles --resolve={name}.