iocframework 1.0.0

Read the latest Real Python tutorials

Documentation for IOCFramework

This project is a dependency injection frameework for python Object Oriented projects.

Installation

pip install iocframework

Basic uses

• Registering a project's class dependency maps.
• Resolving a project's dependencies from registered dependency maps.
• Resolving a project's class' nested dependency, leveraging typing to instantiate dependencies that have no args constructors.

Example use case

from IOCFramework import get_app

class BarService:
	def __init__(self):
    	pass

class FooService:
	bar_service: BarService
    def __init__(self, bar_service: BarService):
    	self.bar_service = bar_service

class FooController:  
	foo_service: FooService
	def __init__(self, foo_service: FooService):
		self.foo_service = foo_service

app = get_app()
foo_controller = app.make(FooController) # makes an instance of fooController

In the example above the dependency tree goes from FooController down to BarService, in real life it could be deeper, this package smartly resolves and instantiates FooController by going down the dependency graph. It recognizes BarService as a dependency of FooService which on its own is a dependency of the required FooController.

Another use case is when you want to resolve a class but with some default values.

Example use case

from IOCFramework import get_app

class BarService:
	def __init__(self):
    	pass

class FooService:
	bar_service: BarService
    def __init__(self, bar_service: BarService):
    	self.bar_service = bar_service


app = get_app()
custom_bar_service = BarService() # custom instance of barService
foo_service = app.make_with(FooService, {"bar_service": custom_bar_service})

In the example above, we created an instance of FooService to which we we manually injected an already created instance of its dependency BarService by manually creating an instance of BarService and passing it by parameter name to the container. This is very helpful for passing constants to class instances or when passing an object created by another class to the new class.

Another use case is binding classes or interfaces to their implementation in the container.

Example use case

from IOCFramework import get_app

class BarService:
	def __init__(self):
    	pass

class AbstractFooService:
	def say_foo(self):
    	raise MethodNotImplemented

class FooService(AbstractFooService):
	bar_service: BarService
    def __init__(self, bar_service: BarService):
    	self.bar_service = bar_service

    def say_foo(self):
    	return "foo"

app = get_app()
bindings = {
	AbstractFooService: FooService
}
app.add_bindings(bindings)
foo_service = app.make(AbstractFooService)
foo_val = foo_service.say_foo() # returns foo

In the example above, we registered AbstractFooService to return an instance of FOOService, this pattern can be useful when a single base class has multiple different implementations for differenct use cases.

The use case above can be extended for singletons, suppose we want our entire application to use a single instance of a very expensive class, we can use the singleton method of app; here is an example

Example use case

from IOCFramework import get_app

class BarService:
	def __init__(self):
    	pass

class FooService:
	bar_service: BarService
    def __init__(self, bar_service: BarService):
    	self.bar_service = bar_service

class FizzService:'
	bar_service: BarService
    def __init__(self, bar_service: BarService):
    	self.bar_service = bar_service

app = get_app()
singletons = {
	BarService: BarService
}
app.add_singletons(singletons)
foo_service = app.make(FooService)
fizz_service = app.make(FizzService)
fizz_service.bar_service == foo_service.bar_service # True because barService is a singleton 

There is yet another use case which adds power to the use cases above, imagine for you have a base class with multiple subclasses, with each of the subclasses serving a single service class. To resolve this, you could type the dependency on the services with their specific class implementation, locking yourself to an implementation or you could type them all with their base class and resolve them by need.

Example use case

from IOCFramework import get_app

class BaseBarService:
	def __init__(self):
    	pass

class FooBarService(BaseBarService):
	def __init__(self):
    	pass

class FizzBarService(BaseBarService):
	def __init__(self):
    	pass

class FooService:
	bar_service: BaseBarService
    def __init__(self, bar_service: BaseBarService):
    	self.bar_service = bar_service

class FizzService:'
	bar_service: BaseBarService
    def __init__(self, bar_service: BaseBarService):
    	self.bar_service = bar_service

app = get_app()
app.when(FizzService, needs=BaseBarService, give=FizzBarService) # give can also be a callable which returns an instance of the service in question
app.when(FooService, needs=BaseBarService, give=FooBarService)
fizz_service = app.make(FizzService)
foo_service = app.make(FooService)

isinstance(fizz_service.bar_service, FizzBarService) # True

isinstance(foo_service.bar_service, FooBarService) # True

This project is based on Laravel's IOC container.