Python Project Bootstrap

A lightweight starter kit to spin up reproducible Python environments for data science projects. With make, virtualenv, and Jupyter baked in. Plus a few data science libraries to get you started.

⸻

Features • Creates a .venv-myproject with essential DS libraries • Registers a Jupyter kernel for the venv • Generates standard project folders: notebooks/, data/, scripts/ • Includes Makefile targets for setup, cleanup, freezing, and launching notebooks • No poetry, no pipenv, just portable, dependency-free shell scripts

⸻

Usage

Full project setup

make init PROJECT=myproject

Creates this structure:

myproject/ ├── .venv-myproject/ ├── notebooks/ ├── data/ ├── scripts/ └── Makefile

Reset venv inside existing project

make venv PROJECT=myproject

(Re)creates .venv-myproject and reinstalls packages.

Freeze requirements

make freeze

Creates requirements.txt with pip freeze.

Clean cache + venv

make clean

Removes .venv-*, pycache, checkpoints, and .pyc files.

Launch Jupyter

make notebook

Starts Jupyter Notebook using your local venv.

Destroy project (IRREVERSIBLE!)

make destroy

Deletes venv, notebooks, scripts, data, and cached files.

⸻

Use Anywhere (Optional)

Set a global alias in your shell config:

newproj() {
  local project_name="$1"
  local current_dir="$(pwd)"
  make -C ~/PATH-TO-YOUR-PROJECT/python-bootstrap init PROJECT="$project_name" TARGET_DIR="$current_dir"
}

Then run:

newproj PROJECT=myproject - or anything else you want as the project name.

This will create a new project in the current directory with the specified name.