Paraminjector

This Python library allows you to define a set of parameters and their types which will be available to be passed into any (third party) user-defined function based on type annotations. This behavior is also referred to as "argument injection", "autowiring", "magic arguments", and many others. Most likely you will want to use paraminjector as part of a library that expects its users to define functions with variable parameters that are supposed to be provided automatically if present.

Installation

pip install paraminjector

Usage

Dynamic injection works as follows:

from typing import *


class Foo:
    """ Some class defined by your library """


class Bar:
    """ Some other class defined by your library """


# ...

# Your users know that they can use `Foo`, and `Bar` in their callbacks:
def user_defined_callback(i: int, a: Foo, b: Bar) -> None:
    assert i == 3
    assert type(a) is Foo
    assert type(b) is Bar


# Now, in library code:
from paraminjector import call_with_args

# Prepare the possible arguments we want to allow
f = Foo()
b = Bar()

available_args: Dict[Type, object] = {Foo: f, Bar: b}

# We know that the first argument to every callback is always the integer 3 (for sake of example)
fixed_positional_arguments: Tuple = (3,)

# Call the user-defined callback with all known parameters injected as arguments
call_with_args(user_defined_callback, available_args, fixed_positional_arguments)

Apart from call_with_args, there are also call_with_args_async for coroutines and call_with_args_maybe_async for cases where you want to allow either a regular function or a coroutine.