kata-python-refactor

To safely refactor Python code, you must validate current behavior by running test coverage and adding missing characterization tests first. You then apply incremental transformations while verifying Ruff, mypy, and pytest between commits to ensure no unintended changes occur.

Characterization tests capture the current behavior of your Python code before you refactor. You need them when existing test coverage is insufficient, ensuring you have a safety net to detect unintended changes when reorganizing entangled modules like god classes.

You split Python refactoring tasks into small, independently committable transformations that can each be validated by running the test suite. This risk-minimized planning ensures behavior and interface changes never occur in the same commit, keeping endpoints stable.

Yes, you can and should use Ruff for linting and formatting, and mypy for type checking to verify Python refactoring steps. Running these alongside full pytest checks after each commit enforces a tight feedback loop to catch issues early.

The best way to refactor a Python god class is to progressively rename, extract, and reorganize the entangled infrastructure and domain logic. This uses continuous verification through tests and linting to keep endpoints and outputs unchanged throughout the process.

Behavior and interface changes must not occur in the same commit to maintain a verifiable safety net. Separating them ensures that each incremental transformation is purely structural, allowing test suite validation to isolate whether the refactor preserved existing outputs.