README.rst

Takes

Decorator for run-time type checks.

• Free software: MIT license
• Documentation: https://takes.readthedocs.io.

Features

• Convert undefined dictionaries accepted by functions to strong data types, without changing calling code.
• Decorated functions can be called with dictionaries as before, or with instances of the desired type.

Example

This function takes an undocumented dictionary. Maintainers must read the function body to determine the expected shape of the data.

def my_function(data):
    x, y = data["x"], data["y"]
    return f"x={x}, y={y}"

>>> my_function({"x": 1, "y": 1})
"x=1, y=1"

Takes lets you redefine this function to accept a well-defined data type, without needing to immediately change all of the calling code:

from takes import takes

@dataclass
class Point:
    x: int
    y: int

@takes(Point)
def my_function(point):
    x, y = point.x, point.y
    return f"x={x}, y={y}"

# Can still call my_function with a dictionary,
# Takes will convert it to a point by instantiating
# a Point using the given data as kwargs:
>>> my_function({"x": 1, "y": 1})
"x=1, y=1"

# Of course, you can also call the function with the proper
# type:
>>> my_function(Point(x=1, y=1))
"x=1, y=1"

Further, takes performs type checks at run-time.

Credits

This package was created with Cookiecutter and the pymetrics/cookiecutter-python-library project template.