required 0.3.2

A easy dependency validator

required is a simple Python package which allows you to validate function argument dependencies in situations where there are a large number of optional parameters. Places where you may have this are:

• In a API where you may have a number of optional query parameters, that are only valid under some permutations

• Functions which receive **kwargs but need to validate that it is correct

Installation

Install using pip

pip install required

Examples

Basic Dependencies

Say you have a function which takes two optional arguments, however, the first one x is only valid when y has been passed. You could eaily express this with:

from required import Requires, R, validate

x_requires_y = Requires("x", "y")
@validate(x_requires_y)
def fn1(x=None, y=None):
    return x,y

fn1(x=1)  # RequirementError: x requires 'y' to be present
fn1(x=1, y=1)  # (1,1)

Expression Dependencies

Thats a contrived example you say?! Ok, something a little bit more complex. You have a function which takes two arguments v and w, but v is only valid when the value of v is less than w.

v_less_than_w = Requires("v", R("v") <= R("w"))

@validate(v_less_than_w)
def fn2(v, w):
   return v,w

fn2(v=3, w=2)  # RequirementError: x requires x to be less than or equal to y
fn2(v=1, w=2)  # (1,2)

Combined Dependencies

Yes, but what if I have more than one constraint I hear you say? Well, yes, thats ok too. Imagine a function which takes x, y,z. x requires y to be greater than x and z requires y to be less than z

x_less_than_y = Requires("x", R("y") >= R("x"))
y_less_than_z = Requires("z", R("y") <= R("z"))

@validate(x_less_than_y + y_less_than_z)
def fn3(x,y,z):
    return x,y,z

fn3(x=1, y=2, z=1)  # RequirementError: z requires y to be less than or equal to z
fn3(x=1, y=2, z=3)  # (1,2,3)

Arithmetic with R objects

y_must_be_zero = Requires("x", R("y") + R("x") == R("x"))
@validate(y_must_be_zero)
def fn4(x,y):
    return x,y

fn4(x=1, y=2)  # RequirementError
fn4(x=1, y=0)  # (0, 1)

Caveats

• The validation is done through dictionary types. Therefore all parameters to your function must be passed as **kwargs, *args are unchecked.

TODO

• Add more expression operators

• Add support for more complex expressions