dependency_injection.py

This Python library defines a helper for building a dependency injection framework.

Installation

dependency_injection is available on GitHub and on PyPI:

$ pip install dependency_injection

We test against Python 2.6, 2.7, 3.2, and 3.3.

dependency_injection is MIT-licensed.

What is Dependency Injection?

When you define a function you specify its parameters, and when you call the function you pass in arguments for those parameters. Dependency injection means dynamically passing arguments to a function based on the parameters it defines. So if you define a function:

>>> def foo(bar, baz):
...     pass

Then you are advertising to a dependency injection framework that your function wants to have the bar and baz objects passed into it. What bar and baz resolve to depends on the dependency injection framework. This library provides a helper, resolve_dependencies, for building your own dependency injection framework. It doesn’t provide such a framework itself, because that would take away all the fun.

API Reference

exception dependency_injection.CantUseThis
dependency_injection.get_signature(function)
Given a function object, return a namedtuple representing the function
signature.
Parameters:function – a function object or other callable
Returns:a namedtuple representing the function signature

This function is a helper for resolve_dependencies. It returns a namedtuple with these items:

  1. parameters - a tuple of all parameters, in the order they were defined
  2. required - a tuple of required parameters, in the order they were defined
  3. optional - a dict of optional parameters mapped to their defaults

For example, if you have this function:

>>> def foo(bar, baz=1):
...     pass
...

Then get_signature will return:

>>> get_signature(foo)
Signature(parameters=('bar', 'baz'), required=('bar',), optional={'baz': 1})

Here are the kinds of callable objects we support (in this resolution order):

  • functions
  • methods (both bound and unbound)
  • classes (both newstyle and oldstyle) with a user-defined __new__
  • classes (both newstyle and oldstyle) with a user-defined __init__
  • object instances with a __call__ method

So you can do:

>>> class Foo(object):
...     def __init__(self, bar, baz=1):
...         pass
...
>>> get_signature(Foo)
Signature(parameters=('self', 'bar', 'baz'), required=('self', 'bar'), optional={'baz': 1})
dependency_injection.resolve_dependencies(function, available)
Given a function object and a mapping of available dependencies, return a
namedtuple that has arguments to suit the function’s parameters.
Parameters:
  • function – a function object or other callable
  • available – a dict mapping arbitrary names to objects
Returns:

a namedtuple representing the arguments to use in calling the function

The return value of this function is a namedtuple with these attributes:

  1. as_args - a tuple of argument values
  2. as_kwargs - a dict of keyword arguments
  3. signature - a namedtuple as returned by get_signature

The as_args and as_kwargs arguments are functionally equivalent. You could call the function using either one and you’d get the same result, and which one you use depends on the needs of the dependency injection framework you’re writing, and your personal preference.

This is the main function you want to use from this library. The idea is that in your dependency injection framework, you call this function to resolve the dependencies for a function, and then call the function. So here’s a function:

>>> def foo(bar, baz):
...     return bar + baz

And here’s the basics of a dependency injection framework:

>>> def inject_dependencies(func):
...     my_state = {'bar': 1, 'baz': 2, 'bloo': 'blee'}
...     dependencies = resolve_dependencies(func, my_state)
...     return func(*dependencies.as_args)
...

And here’s what it looks like to call it:

>>> inject_dependencies(foo)
3

See get_signature for details on support for non-function callables.