Build a Python CLI that another person can install

A script that works from its author’s folder is not yet a distributable command. An installable CLI declares metadata, Python compatibility, dependencies and an entry point.

Check the metadata contract

The engineering foundations lab builds a metadata fixture.

from engineering_cases import cli_package_case

result = cli_package_case()
assert result["installable_fields"] is True
assert result["entry_point_target"] == "fixture_cli:main"
assert result["metadata"]["requires_python"] == ">=3.11"
assert result["package_built"] is False

The check proves the record is complete; no wheel was built or published.

Use a src/ package, pyproject.toml, README, license and tests. Keep parsing/output in the CLI layer and business logic in importable functions. Return meaningful exit codes, write normal results to stdout and errors to stderr. Add [project.scripts] so installation creates the command.

Verify in a new virtual environment: build the wheel, install it, run --help, exercise a success and invalid-input path, then uninstall. A colleague should reproduce this from the README without modifying source or PYTHONPATH.

The FDE for Freshers course uses a tested CLI as its first shippable engineering artifact.

Exercise

Package a CSV-summary command with three exit codes and golden output. Save build/install/test evidence from a clean environment; publish only after review.

Continue learning

This article is part of the FDE engineering foundations sequence. Use the neighbouring tasks when you need the prerequisite or the next application.

• Continue with Design a small package with explicit dependencies.
• Then apply it in Use type hints to document an integration boundary.

References: PyPA command-line tool guide and entry-point specification.