pyproteum 0.2.2

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.

V0.1.2

• Bug fixing: operator ccsr

V0.1.3

• Monthly review

• Bug fixing

• Color in the stdout

• Improved speed for mutant generation and execution

• Fix source files in different directories like: --S tests/test_foo.py

V0.1.4

• Bug fixing (mutants were generated with different numbers due to unordered sets)

V0.2

• Introduces the second argument for mutagen selection of mutant operatos.

V0.2.2

• Introduces argument --assert to enable mutation on assert statements.

---

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

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	Default MNOM
cccr	Constant by Constant Replacement	1
ccsr	Constant by Scalar Replacement	0
crcr	Required Constant Replacement	0
oaaa	Replace arithmetic assignment operator by other arithmetic assignment operator	5
oaan	Replace arithmetic operator by other arithmetic operator	5
oeap	Replace augmented assignment by plain assignment	0
oodl	Operator deletion	0
orrn	Replace relational operator (<, >, <=, >=, ==, !=) by other relational operator	0
sbrc	Replace each break statement by a continue statement	0
scrb	Replace each continue statement by a break statement	0
ssdl	Replace each statement by a pass statement	0

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

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

Add one or more unittest test files from a given location:

python3 -m pyproteum tcase --add --discover tests 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

By default assert statements are not considered to generate mutants. This can be changed with --assert argument.

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

Generate only orrn mutants, with 50% random sampling:

python -m pyproteum mutagen --create --D /home/user/proteum --orrn 50 0 foo_session

Generate only cccr mutants, with 100% random sampling but generate only 2 mutants on each mutation point. This is usefull, for instance, with cccr operator. A given occurrence will be replaced by only one another constant, since replacing it by several other constants might generate similar mutants that are all killed by the same test.

python -m pyproteum mutagen --create --D /home/user/proteum --cccr 50 2 foo_session

The third parameter of operator selection may be:

• a positive integer indicating the maximum number of mutants on that point

• 0 indicating all possible mutants on that point

• _ (underscore) indicating the default number for that operator

Defaults for Maximum Number of Mutants (MNOM) are mentioned in the operator table above.

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).
• Operators are scanned in the order they appear in the command line. --all 100 0 --ssdl 0 0 generate 100% for all operators, except for ssdl.

---

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

• If --x is not provided, all live mutants are set as equivalent.

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.
• csv still need some work

---

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

---

Color output

If you want to use color in pyproteum stdout, do:

export PYPROTEUMCOLOR=

Error messages and mutated code will appear in different colors.

---

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 .