pyproteum 0.1

A tool for mutation testing in Python. Implemented as a Python module.

pyproteum

A mutation testing toolkit for Python. Implemented as a Python module.

It is based on Proteum (Program Testing Using Mutants) , the first widely used mutation tool, developed for C.

https://link.springer.com/book/10.1007/978-1-4757-5939-6

Status: early preview — core mutation operators are available; more coming soon.

---

Features

• Parse Python source code and generate mutants using configurable operators
• Run your test suite against each mutant (kill/survive reporting)
• Simple CLI

Requirements

• Python >= 3.10
• Dependencies: timeout-decorator, peewee
  (Note: pickle is part of the Python standard library and does not need to be installed.)

Installation

From PyPI (once published)

pip install pyproteum

pyproteum — Command Line Guide

pyproteum uses the concept of a test session.
To use the tool, follow these steps:

1. Create a test session specifying which source files will be tested (mutated).
2. Add test cases from unittest test files.
3. Generate mutants from the selected source files.
4. Execute the generated mutants against the test cases.

This cycle can be repeated multiple times — you can remove existing test cases, add new ones, and generate additional mutants as needed.

Mutation operators (currently implemented)

Name	Description
cccr	Constant by Constant Replacement
ccsr	Constant by Scalar Replacement
crcr	Required Constant Replacement
oaaa	Replace arithmetic assignment operator by other arithmetic assignment operator
oaan	Replace arithmetic operator by other arithmetic operator
oeap	Replace augmented assignment by plain assignment
oepa	Replace plain assignment by augmented assignment
oodl	Operator deletion
orrn	Replace relational operator (<, >, <=, >=, ==, !=) by other relational operator
sbrc	Replace each break statement by a continue statement
scrb	Replace each continue statement by a break statement
ssdl	Replace each statement by a pass statement

Notes
crcr replaces literals with a small, meaningful set of alternative constants (e.g., -1, 0, 1, for integers).

oepa will produce a lot of useless mutants for the assignments that initialize a variable.

Configuration tips

• Start with a small set of operators (e.g., orrn,crcr) to get quick feedback.
• Scope your target to the most critical modules first to keep the run time manageable.

Command-line tools

pyproteum provides 5 main subcommands:

1. testnew — Create a new test session

python -m pyproteum testnew --D /home/user/proteum --S foo.py foo_session

• Creates a test session in the specified directory (--D) for the file foo.py, naming the session foo_session.

• --D: base directory where the session will be created. All file names are relative to this directory.
  Example: the source file in this case would be /home/user/proteum/foo.py.

• --S: specifies the source file(s) to be mutated; can be repeated multiple times.
  If not provided, defaults to <session_name>.py.

• --test or --research: exactly one must be used.

  --test: each mutant is executed until the first test case kills it.
  --research: each mutant is executed against all test cases in the session.
  If omitted, --test is assumed.

---

2. tcase — Manage test case files in a session

Add a unittest test file:

python -m pyproteum tcase --add --D /home/user/proteum --S test_foo.py foo_session

Remove a test file:

python -m pyproteum tcase --del --D /home/user/proteum --S test_foo.py foo_session

List all test files and test cases in a session:

python -m pyproteum tcase --list --D /home/user/proteum foo_session

---

3. mutagen — Generate mutants

Generate mutants for all operators:

python -m pyproteum mutagen --create --D /home/user/proteum foo_session

Generate only ssdl mutants, with 50% random sampling:

python -m pyproteum mutagen --create --D /home/user/proteum --ssdl 50 foo_session

Notes:

• You can specify multiple operators in the same command.
• You can use --all to sample all operators.
• You can specify a seed for sampling, Using the same seed generates the same mutants. Use: --seed <integer number>.
• Partial operator names are allowed. For example, --c selects all constant-related operators (those starting with c).

---

4. exemuta — Execute mutants

Run generated mutants against the current test set:

python -m pyproteum exemuta --exec --D /home/user/proteum foo_session

Set mutants as equivalents:

python -m pyproteum exemuta --equiv --D /home/user/proteum --x "1 4 44" foo_session

Generate a CSV table of results:

python -m pyproteum exemuta --csv --D /home/user/proteum --O foo.csv foo_session

• Each row corresponds to a mutant; each column corresponds to a test case.
• --O sets the CSV output filename. If omitted, defaults to <session_name>.csv.

---

5. mutaview — View mutants

Show original vs. mutated code in the terminal:

python -m pyproteum mutaview --view --D /home/user/proteum --x 1 foo_session

• --x specifies the mutant number to display. It can be a list of numbers like --x "1 2 3 4 5".

Generate a .py file with the mutant's code:

python -m pyproteum mutaview --build --D /home/user/proteum --x 1 foo_session

• The output file will be named foo_session_mutant0001.py in this example.

List all mutants and their current status

python -m pyproteum mutaview --list --D /home/user/proteum foo_session

Show mutants in a GUI interface. Allows to compare with original code and set mutant as equivalent.

python -m pyproteum mutaview --gui --D /home/user/proteum foo_session

---

Workflow summary

1. testnew → create a session
2. tcase → add/remove/list test files
3. mutagen → generate mutants
4. exemuta → execute mutants & get results
5. mutaview → inspect mutants

Contributing

Contributions are welcome! Please open an issue describing the change you’d like to make or the bug you found. If you’re adding a new operator, include:

1. Operator name, short key, and rationale
2. AST node types affected and transformation rules
3. Unit tests that both generate the mutant and exercise its effects

License

MIT

Citation

If this toolkit helps your research or teaching, consider citing the project as:

Delamaro, M. (2025). pyproteum: a mutation testing toolkit for Python. Version 0.1.

---

Questions, ideas, or bugs? Open an issue or ping us .