Skip to main content

Dependency injection library

Project description

Opyoid

CI codecov

Dependency injection library using typings, to easily manage large applications.

This project is inspired from Guice.

Installation

Run pip install opyoid to install from PyPI.

Run pip install . to install from sources.

This project follows the Semantic Versioning Specification. All breaking changes are described in the Changelog.

Usage

Simple Injection

from opyoid import Module, Injector


class MyClass:
    pass


class MyParentClass:
    def __init__(self, my_param: MyClass):
        self.my_param = my_param


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass)
        self.bind(MyParentClass)


injector = Injector([MyModule()])
my_instance = injector.inject(MyParentClass)
assert isinstance(my_instance, MyParentClass)
assert isinstance(my_instance.my_param, MyClass)

If they are multiple bindings for the same class, the latest will be used.

Module

The module is used to group bindings related to a feature. You can include a module in another with install:

from opyoid import Module, Injector


class MyClass:
    pass


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass)


class MyParentClass:
    def __init__(self, my_param: MyClass):
        self.my_param = my_param


class MyParentModule(Module):
    def configure(self) -> None:
        self.install(MyModule())
        self.bind(MyParentClass)


injector = Injector([MyParentModule()])
my_instance = injector.inject(MyParentClass)
assert isinstance(my_instance, MyParentClass)
assert isinstance(my_instance.my_param, MyClass)

Binding Subclasses

from opyoid import Module, Injector


class MyClass:
    pass


class MySubClass(MyClass):
    pass


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass, MySubClass)

injector = Injector([MyModule()])
my_instance = injector.inject(MyClass)
assert isinstance(my_instance, MySubClass)

Binding Instances

from opyoid import Module, Injector


class MyClass:
    def __init__(self, my_param: str):
        self.my_param = my_param

my_instance = MyClass("hello")


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass, to_instance=my_instance)


injector = Injector([MyModule()])
injected_instance = injector.inject(MyClass)
assert my_instance is injected_instance

Binding scopes

When binding a class, you can choose the scope in which it will be instantiated. This will only have an effect when binding classes, not instances.

Singleton Scope

By default, all classes are instantiated in a Singleton scope. This means that only one instance of each class will be created, and it will be shared between all classes requiring it.

from opyoid import Module, Injector, SingletonScope


class MyClass:
    pass


class MyParentClass:
    def __init__(self, my_param: MyClass):
        self.my_param = my_param


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass, scope=SingletonScope)
        self.bind(MyParentClass, scope=SingletonScope)

injector = Injector([MyModule()])
instance_1 = injector.inject(MyClass)
instance_2 = injector.inject(MyClass)
parent_instance = injector.inject(MyParentClass)
assert instance_1 is instance_2
assert instance_1 is parent_instance.my_param

PerLookup Scope

If you use the per lookup scope, a new instance will be created every time each class is injected.

from opyoid import Module, Injector, PerLookupScope


class MyClass:
    pass


class MyParentClass:
    def __init__(self, my_param: MyClass):
        self.my_param = my_param


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass, scope=PerLookupScope)
        self.bind(MyParentClass)

injector = Injector([MyModule()])
instance_1 = injector.inject(MyClass)
instance_2 = injector.inject(MyClass)
parent_instance = injector.inject(MyParentClass)
assert instance_1 is not instance_2
assert instance_1 is not parent_instance.my_param

Thread Scope

This scope only creates a new instance the first time that the class is injected in the current thread. There will only be one instance of each class in each thread, and two instances injected from different threads will be different objects.

from threading import Thread

from opyoid import Module, Injector, ThreadScope


class MyClass:
    pass


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass, scope=ThreadScope)

injector = Injector([MyModule()])
instance_1 = injector.inject(MyClass)
instance_2 = injector.inject(MyClass)

def thread_target():
    instance_3 = injector.inject(MyClass)
    assert instance_1 is not instance_3

Thread(target=thread_target).start()

assert instance_1 is instance_2

Bindings without Module

If you prefer, you can add bindings to your injector without creating a Module class (or using both).

from opyoid import Module, Injector, SelfBinding


class MyClass:
    pass


class MyParentClass:
    def __init__(self, my_param: MyClass):
        self.my_param = my_param


class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass)


injector = Injector([MyModule()], [SelfBinding(MyParentClass)])
my_instance = injector.inject(MyParentClass)
assert isinstance(my_instance, MyParentClass)
assert isinstance(my_instance.my_param, MyClass)

The same options of Module.bind are available when using bindings:

from opyoid import ClassBinding, InstanceBinding, PerLookupScope, SelfBinding


class MyClass:
    pass


class MySubClass(MyClass):
    pass

my_instance = MyClass()

SelfBinding(MyClass)  # binding a class to itself
ClassBinding(MyClass, MySubClass)  # binding a class to a subclass
SelfBinding(MyClass, scope=PerLookupScope)  # specifying scope
InstanceBinding(MyClass, my_instance)  # binding an instance
SelfBinding(MyClass, named="my_name")  # binding a class to itself with a specific name
InstanceBinding(MyClass, my_instance, named="my_name")  # binding an instance with a specific name

Injecting Type

If no explicit binding is defined, the last class binding will be used to inject a type:

from typing import Type

from opyoid import Module, Injector

class MyClass:
    pass

class SubClass(MyClass):
    pass

class MyParentClass:
    def __init__(self, my_param: Type[MyClass]):
        self.my_param = my_param

my_instance = MyClass()

class MyModule(Module):
    def configure(self) -> None:
        self.bind(MyClass)
        self.bind(MyClass, to_instance=my_instance)
        self.bind(MyClass, SubClass)
        self.bind(MyParentClass)


injector = Injector([MyModule()])
parent_instance = injector.inject(MyParentClass)
assert isinstance(parent_instance, MyParentClass)
assert parent_instance.my_param is SubClass

Dataclasses

opyoid can inject classes and parameters defined with the attrs library and python data classes.

Notes about Generics

  • The only supported generic types are List, Set, Tuple, Optional and Type (and any combination of them). Other generics must be bound explicitly (e.g. you must bind a dict to Dict[str, MyClass] if you want to inject it).
  • Be careful when using generics, the bindings will only be used if the type matches exactly. For example, you cannot implicitly bind MyClass[T] to inject MyClass, or MyClass[str] to inject MyClass[T]. You need to bind something to MyClass[str] to be able to inject it.

Advanced usage

More advanced features and examples are available in the ./docs folder.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

opyoid-0.10.0.tar.gz (21.0 kB view details)

Uploaded Source

Built Distribution

opyoid-0.10.0-py3-none-any.whl (42.2 kB view details)

Uploaded Python 3

File details

Details for the file opyoid-0.10.0.tar.gz.

File metadata

  • Download URL: opyoid-0.10.0.tar.gz
  • Upload date:
  • Size: 21.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.3.0 pkginfo/1.6.1 requests/2.25.1 setuptools/49.2.1 requests-toolbelt/0.9.1 tqdm/4.55.1 CPython/3.9.1

File hashes

Hashes for opyoid-0.10.0.tar.gz
Algorithm Hash digest
SHA256 5cea74bd13aef3c8e4bf2d4d5a7d511bb2ab3a07188be3f7d3df36aba48bf69c
MD5 12ef6dde755db47c3e232dfd65e8b55f
BLAKE2b-256 2c98491d1129ee71f7c26638806b2b16766bff43e653bed9d2aa9923aed69573

See more details on using hashes here.

File details

Details for the file opyoid-0.10.0-py3-none-any.whl.

File metadata

  • Download URL: opyoid-0.10.0-py3-none-any.whl
  • Upload date:
  • Size: 42.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.3.0 pkginfo/1.6.1 requests/2.25.1 setuptools/49.2.1 requests-toolbelt/0.9.1 tqdm/4.55.1 CPython/3.9.1

File hashes

Hashes for opyoid-0.10.0-py3-none-any.whl
Algorithm Hash digest
SHA256 1f325f63134444d9a892bdd4cb686360c073a573353f9051144c4a478e7bc517
MD5 f8aaf2e4333ea516963237c86707547f
BLAKE2b-256 44d606bb218bab3d86f34087e0c8fcaa573648fa40416aa1d5c43c6b536429c4

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page