A pytest plugin that calculates CRAP scores and displays prioritized lists of high-risk functions to guide test writing.

CRAP stands for Change Risk Anti-Patterns (or, more colloquially, "Change Risk Analysis and Predictions"). The metric was introduced by Alberto Savoia and Bob Evans to help developers identify code that is both complex and poorly tested—a risky combination.

The CRAP score combines two factors:

• Cyclomatic Complexity (CC): A measure of how many independent paths exist through your code. More branches (if, for, while, try, etc.) means higher complexity.
• Code Coverage: The percentage of lines executed by your tests.

The formula is:

CRAP(m) = CC(m)² × (1 - cov(m))³ + CC(m)

Where CC(m) is the cyclomatic complexity and cov(m) is the code coverage (0.0 to 1.0) for method m.

• Prioritize testing efforts: Focus on functions that are both complex AND under-tested
• Identify risky code: High CRAP scores indicate code that's likely to harbor bugs and is difficult to change safely
• Track improvement: Monitor CRAP scores over time to ensure code quality improves

A function with CC=1 and 100% coverage has a CRAP score of 1 (perfect). A function with CC=10 and 0% coverage has a CRAP score of 110 (very high risk).

pip install pytest-crap

Or with Poetry:

poetry add pytest-crap

• Python 3.10+
• pytest 7.0+
• pytest-cov (for coverage data)

Run pytest with the --crap flag:

pytest --crap

This will run your tests with coverage enabled and display CRAP score tables at the end.

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃                            CRAP by Function                              ┃
┡━━━━━━━━┯━━━━━┯━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━┩
│   CRAP │  CC │ Coverage │ Function             │ File                    │
├────────┼─────┼──────────┼──────────────────────┼─────────────────────────┤
│  42.50 │   8 │    12.5% │ complex_parser       │ src/parser.py           │
│  31.00 │   5 │     0.0% │ validate_input       │ src/validator.py        │
│  12.25 │   3 │    50.0% │ process_data         │ src/processor.py        │
│   1.00 │   1 │   100.0% │ simple_helper        │ src/utils.py            │
└────────┴─────┴──────────┴──────────────────────┴─────────────────────────┘

Show top 10 functions with a stricter threshold:

pytest --crap --crap-threshold=15 --crap-top-n=10

Show all functions regardless of score:

pytest --crap --crap-top-n=0

Combine with other pytest options:

pytest --crap --cov-branch -v tests/

pytest-crap displays three summary tables:

1. CRAP by Function: Individual functions ranked by CRAP score
2. CRAP by File: Files ranked by their highest CRAP score, with count of functions above threshold
3. CRAP by Folder: Directories ranked by highest CRAP score

Add to your CI pipeline to track CRAP scores. Example GitHub Actions step:

- name: Run tests with CRAP reporting
  run: pytest --crap --crap-threshold=30

See contributing.md for development setup and guidelines.

MIT

• Original CRAP metric paper by Alberto Savoia
• Cyclomatic Complexity on Wikipedia
• radon - Python library for code metrics