mypy-pure

Enforce functional purity in Python with static type checking.

A mypy plugin that helps you write safer, more predictable code by detecting side effects in functions marked as @pure.

Why mypy-pure?

Pure functions are:

• Easier to test - No mocks needed, same inputs always give same outputs
• Easier to reason about - No hidden state changes or side effects
• Easier to refactor - Can be moved, renamed, or reordered safely
• Easier to parallelize - No race conditions or shared state issues
• Easier to cache - Results can be memoized safely

But enforcing purity manually is error-prone. mypy-pure catches impure code at type-check time, before it reaches production.

What is a Pure Function?

A pure function:

1. Always returns the same output for the same inputs (deterministic)
2. Has no side effects (no I/O, no mutations, no external state changes)

# ✅ Pure - deterministic, no side effects
def add(x: int, y: int) -> int:
    return x + y

# ❌ Impure - side effect (I/O)
def add_and_log(x: int, y: int) -> int:
    print(f"Adding {x} + {y}")  # Side effect!
    return x + y

Installation

pip install mypy-pure

Enable the plugin in your mypy.ini or pyproject.toml:

[mypy]
plugins = mypy_pure.plugin

Quick Start

Mark functions as pure with the @pure decorator:

from mypy_pure import pure

@pure
def calculate_total(prices: list[float], tax_rate: float) -> float:
    subtotal = sum(prices)
    return subtotal * (1 + tax_rate)

Run mypy to check for purity violations:

mypy your_code.py

Examples

✅ Valid Pure Functions

from mypy_pure import pure

@pure
def fibonacci(n: int) -> int:
    """Pure recursive function."""
    if n <= 1:
        return n
    return fibonacci(n - 1) + fibonacci(n - 2)

@pure
def process_data(items: list[dict]) -> list[str]:
    """Pure data transformation."""
    return [item['name'].upper() for item in items if item['active']]

@pure
def calculate_discount(price: float, discount_pct: float) -> float:
    """Pure business logic."""
    return price * (1 - discount_pct / 100)

❌ Detected Impurity Violations

from mypy_pure import pure
import os

@pure
def read_config() -> dict:
    # Error: Function 'read_config' is annotated as pure but calls impure functions.
    with open('config.json') as f:  # I/O is impure!
        return json.load(f)

@pure
def delete_temp_files(directory: str) -> None:
    # Error: Function 'delete_temp_files' is annotated as pure but calls impure functions.
    for file in os.listdir(directory):
        os.remove(file)  # File system modification is impure!

@pure
def log_and_calculate(x: int, y: int) -> int:
    # Error: Function 'log_and_calculate' is annotated as pure but calls impure functions.
    print(f"Calculating {x} + {y}")  # Logging is impure!
    return x + y

More examples

More examples can be found in the mypy-pure-examples repository.

Advanced Usage

Configuration Options

mypy-pure supports two configuration options in the [mypy-pure] section of your mypy.ini:

1. impure_functions (Blacklist)

Add custom impure functions that should be flagged as side-effecting:

[mypy-pure]
impure_functions = my_module.send_email, analytics.track_event, cache.set

Use cases:

• Your own functions that have side effects
• Third-party library functions not in the built-in blacklist
• Project-specific impure operations

Example:

# my_module.py
def send_email(to: str, subject: str) -> None:
    # Sends an email (side effect)
    ...

# main.py
from mypy_pure import pure
from my_module import send_email

@pure
def process_user(user: dict) -> dict:
    send_email(user['email'], 'Welcome')  # ❌ Error: calls impure function
    return user

2. pure_functions (Whitelist)

Mark third-party library functions as pure, overriding the default assumption:

[mypy-pure]
pure_functions = requests.utils.quote, pandas.DataFrame.copy, my_lib.helper

Use cases:

• Pure utility functions from third-party libraries
• Functions you've verified have no side effects
• Overriding false positives

Example:

from mypy_pure import pure
import requests.utils

@pure
def sanitize_url(url: str) -> str:
    # OK because requests.utils.quote is in pure_functions config
    return requests.utils.quote(url)

Combining Both

You can use both options together:

[mypy-pure]
# Blacklist your impure functions
impure_functions = my_module.send_email, my_module.log_event

# Whitelist pure third-party functions
pure_functions = requests.utils.quote, requests.utils.unquote

Priority: pure_functions (whitelist) takes precedence over impure_functions (blacklist).

Library Authors: Auto-Discovery with __mypy_pure__

If you're a library author, you can declare your pure functions using the __mypy_pure__ module-level list. This enables zero-configuration purity checking for your users.

Declaring Pure Functions

Add a __mypy_pure__ list to your module:

# my_library.py
__mypy_pure__ = [
    'pure_helper',
    'utils.calculate',
    'ClassName.method_name',
]

def pure_helper(x: int) -> int:
    """A pure utility function."""
    return x * 2

class utils:
    @staticmethod
    def calculate(a: int, b: int) -> int:
        """A pure calculation."""
        return a + b

def impure_logger(msg: str) -> None:
    """Not in __mypy_pure__, so treated as impure."""
    print(msg)

User Experience

Users of your library automatically benefit without any configuration:

from mypy_pure import pure
import my_library

@pure
def process(x: int) -> int:
    # ✅ OK - pure_helper is in __mypy_pure__
    return my_library.pure_helper(x)

@pure
def log_process(x: int) -> int:
    # ❌ Error: Function 'log_process' is impure because it calls 'my_library.impure_logger'
    my_library.impure_logger(f"Processing {x}")
    return x

Benefits

• Zero configuration for library users
• Self-documenting API - pure functions are explicitly declared
• Compile-time guarantees - purity violations caught during type checking
• Better IDE support - users see which functions are safe to use in pure contexts

Cross-Module References

Reference functions from other modules using fully qualified names:

[mypy-pure]
impure_functions = external_lib.impure_function

Supported Function Types

mypy-pure works with all Python function and method types:

• ✅ Regular functions
• ✅ Instance methods
• ✅ Class methods (@classmethod)
• ✅ Static methods (@staticmethod)
• ✅ Async functions (async def)
• ✅ Async methods
• ✅ Property methods (@property)
• ✅ Nested/inner functions

Built-in Impurity Detection

mypy-pure includes a comprehensive blacklist of 200+ impure functions from Python's standard library:

• File I/O: open(), pathlib.Path.write_text(), etc.
• System operations: os.remove(), subprocess.run(), etc.
• Network: socket.socket(), urllib.request.urlopen(), etc.
• Logging: logging.info(), print(), etc.
• State modification: random.seed(), sys.exit(), etc.
• Databases: sqlite3.connect(), etc.
• And many more...

See full blacklist

Limitations

mypy-pure performs static analysis and has some limitations:

What it CAN detect:

• ✅ Direct calls to known impure functions
• ✅ Indirect calls through pure functions calling impure functions
• ✅ Deeply nested impure calls

What it CANNOT detect:

• ❌ Mutations of mutable arguments (e.g., list.append())
• ❌ Global variable modifications
• ❌ Impure functions not in the blacklist
• ❌ Dynamic function calls (e.g., getattr(), eval())
• ❌ Side effects in third-party libraries (unless configured)

Recommendation: Use mypy-pure as a helpful guard rail, not a guarantee of purity. Combine it with code reviews and testing.

Real-World Use Cases

Data Processing Pipelines

@pure
def transform_user_data(raw_data: dict) -> dict:
    """Pure transformation - easy to test and parallelize."""
    return {
        'id': raw_data['user_id'],
        'name': raw_data['full_name'].title(),
        'age': calculate_age(raw_data['birth_date']),
    }

Business Logic

@pure
def calculate_shipping_cost(
    weight_kg: float,
    distance_km: float,
    is_express: bool
) -> float:
    """Pure business logic - deterministic and testable."""
    base_cost = weight_kg * 0.5 + distance_km * 0.1
    return base_cost * 1.5 if is_express else base_cost

Configuration Processing

@pure
def merge_configs(default: dict, user: dict) -> dict:
    """Pure config merging - no file I/O."""
    return {**default, **user}

Contributing

Contributions are welcome! Here's how you can help:

1. Expand the blacklist - Add more impure functions from stdlib or popular libraries
2. Report bugs - Open an issue if you find incorrect behavior
3. Suggest features - Ideas for improving purity detection
4. Improve documentation - Help make the docs clearer

See CONTRIBUTING.md for details.

Development

# Clone the repository
git clone https://github.com/diegojromerolopez/mypy-pure.git
cd mypy-pure

# Install dependencies
uv sync

# Run tests
pytest

# Run mypy
mypy mypy_pure

License

MIT License - see LICENSE for details.

Acknowledgments

This project was created with the assistance of AI tools (ChatGPT and Antigravity/Gemini) and manually reviewed and refined.

Related Projects

• mypy - Optional static typing for Python
• mypy-raise - A mypy plugin that enforces exception declarations in function signatures, ensuring functions explicitly declare all exceptions they may raise.
• mypy-plugins-examples - A project that contains some examples for my mypy-pure and mypy-raise plugins.

---

Made with ❤️ for the Python community