/pvv

Minimal Python decorator to enforce type validation from type hints

Primary LanguagePythonMIT LicenseMIT

pvv

Minimal Python decorator to enforce type validation from type hints

Installation

This library has been created to have zero dependencies, and work with native Python. Use your favorite package manager to install pvv from PyPI.

pip install pvv

Usage

There are two ways to use the validate decorator from pvv:

Validate all parameters

Just use the decorator. A TypeError will be raised if the function is called with parameters of incorrect type.

Note

pvv will check if the given parameters to a function are instances of the class annotated as a parameter. If one parameter has no type hints, it will be ignored in the validation.

Warning

pvv will only work with type hints that can be used with class and instance checks.

from pvv import validate

@validate
def function(a: str, b: int, c):
    pass
>>> function("a", 3, 2)
>>> function(c=2, a="a", b=2)
>>> function(3, "a", 2)
TypeError: Incorrect type of function arguments: 'a' must be of type 'str', 'b' must be of type 'int'

Validate some parameters

from pvv import validate

@validate('a', 'c')
def function(a: str, b: int, c: bool | None):
    pass
>>> function("a", 3, True)
>>> function(3, "b", 0)
TypeError: Incorrect type of function arguments: 'a' must be of type 'str', 'c' must be of type 'bool | None'

An important note

The parameters of the validate decorator must all be of type str, otherwise a ValidatorError will be raised:

>>> @validate('a', 1)
... def func(*args, **kwargs):
...     pass
pvv.exceptions.ValidatorError: All arguments of decorator must be of type 'str'