Skip to content

Refactoring 02 — Monolithic Report

Open messy.py. This is a monthly sales report generator. It reads a CSV file, filters data by month, computes statistics, formats a text report, and writes it to a file.

The entire pipeline is one 90-line function. It works, but it violates every principle of clean code.

Your mission

Split generate_report() into small, focused functions. Each function should do one thing. The tests in tests/test_messy.py must pass before and after every change.

Refactoring goals

  1. Decompose into functions. Extract at least these:
  2. read_csv(filepath) — reads and returns raw rows
  3. filter_by_month(rows, year, month) — returns matching rows
  4. compute_statistics(rows) — returns totals, averages, top products
  5. format_report(stats, year, month) — returns formatted text

  6. write_report(text, filepath) — writes to file

  7. Add logging. Replace silent processing with logging module calls at key stages (reading, filtering, computing, writing). Use logging.info(), not print().

  8. Make output format configurable. The report format is currently hardcoded. Refactor so the caller can choose between plain text (current format) and a simple CSV summary format.

  9. Add type hints and docstrings to all new functions.

Process

Do not attempt all changes at once. Follow this sequence:

1. Run tests (baseline)
2. Extract read_csv() -> tests pass
3. Extract filter_by_month() -> tests pass
4. Extract compute_statistics() -> tests pass
5. Extract format_report() -> tests pass
6. Extract write_report() -> tests pass
7. Add logging -> tests pass
8. Add format option -> add new tests -> all pass

What you are practicing

  • Decomposing a monolith into focused functions
  • Maintaining backward compatibility during refactoring
  • Adding logging to a pipeline
  • Making code configurable without over-engineering