Packaging, Distribution & Deployment

What you'll learn

A script on your laptop helps only you. Packaging turns it into something anyone can pip install and run — with a clear structure, declared dependencies, a command-line entry point, and, when needed, a container for deployment.

By the end of this lesson you'll be able to:

• Lay out a project with the recommended src structure
• Declare metadata and dependencies in pyproject.toml
• Build wheels and source distributions and publish to PyPI
• Expose your program as a command with a console entry point
• Containerize an app with a simple Dockerfile

Project structure

A predictable layout makes a project easy to build, test, and understand. The modern convention is the src layout: your importable package lives under src/, with tests and config alongside.

myapp/
├── src/
│   └── myapp/
│       ├── __init__.py
│       └── cli.py
├── tests/
│   └── test_cli.py
├── pyproject.toml       # project metadata & dependencies
└── README.md

pyproject.toml

pyproject.toml is the single, standard config file for a Python project. It declares your metadata, dependencies, entry points, and which build backend to use.

[project]
name = "myapp"
version = "0.1.0"
description = "A friendly greeter."
requires-python = ">=3.10"
dependencies = ["requests>=2.31"]

[project.scripts]
myapp = "myapp.cli:main"     # the 'myapp' command runs cli.main()

[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

Building & publishing

Building produces two artifacts: a wheel (.whl, a fast, pre-built install) and an sdist (.tar.gz, the source). twine uploads them to PyPI, where pip install myapp can find them.

$ pip install build twine
$ python -m build          # creates dist/*.whl and dist/*.tar.gz
$ twine upload dist/*      # publish to PyPI (test on TestPyPI first)

Console entry points

The [project.scripts] table turns a function into a terminal command. After install, running myapp calls myapp.cli:main — no python -m, no path needed.

# src/myapp/cli.py
import argparse

def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("name")
    args = parser.parse_args()
    print(f"Hello, {args.name}!")

if __name__ == "__main__":
    main()

$ pip install .
$ myapp Ada          # the entry point makes this a real command
Hello, Ada!

Docker & deployment

A container bundles your app and its exact environment so it runs identically everywhere — your machine, a teammate's, or a server. A minimal Dockerfile installs your package and sets the command to run:

FROM python:3.12-slim
WORKDIR /app
COPY . .
RUN pip install --no-cache-dir .
CMD ["myapp", "world"]

Recap & quick check

You can ship code now. The last professional skills tie it together: writing clean code and collaborating with Git. Next up: Module 40 — Clean Code, Git & Professional Workflow.