Dependency injection

From Infogalactic: the planetary knowledge core
(Redirected from Dependency Injection)
Jump to: navigation, search

In software engineering, dependency injection is a software design pattern that implements inversion of control for resolving dependencies. A dependency is an object that can be used (a service). An injection is the passing of a dependency to a dependent object (a client) that would use it. The service is made part of the client's state.[1] Passing the service to the client, rather than allowing a client to build or find the service, is the fundamental requirement of the pattern.

Dependency injection allows a program design to follow the dependency inversion principle. The client delegates to external code (the injector) the responsibility of providing its dependencies. The client is not allowed to call the injector code.[2] It is the injecting code that constructs the services and calls the client to inject them. This means the client code does not need to know about the injecting code. The client does not need to know how to construct the services. The client does not need to know which actual services it is using. The client only needs to know about the intrinsic interfaces of the services because these define how the client may use the services. This separates the responsibilities of use and construction.

There are three common means for a client to accept a dependency injection: setter-, interface- and constructor-based injection. Setter and constructor injection differ mainly by when they can be used. Interface injection differs in that the dependency is given a chance to control its own injection. All require that separate construction code (the injector) take responsibility for introducing a client and its dependencies to each other.[3]

Overview

<templatestyles src="Template:Quote_box/styles.css" />

Dependency injection for five-year-olds

When you go and get things out of the refrigerator for yourself, you can cause problems. You might leave the door open, you might get something Mommy or Daddy doesn't want you to have. You might even be looking for something we don't even have or which has expired.

What you should be doing is stating a need, "I need something to drink with lunch," and then we will make sure you have something when you sit down to eat.

John Munsch, 28 October 2009.[4][5][6]

Dependency injection separates the creation of a client's dependencies from the client's behavior, which allows program designs to be loosely coupled[7] and to follow the dependency inversion and single responsibility principles.[4][8] It directly contrasts with the service locator pattern, which allows clients to know about the system they use to find dependencies.

An injection, the basic unit of dependency injection, is not a new or a custom mechanism. It works the same way that "parameter passing" works.[9] Referring to "parameter passing" as an injection carries the added implication that it's being done to isolate the client from details.

An injection is also about what is in control of the passing (never the client) and is independent of how the passing is accomplished, whether by passing a reference or a pointer.

Dependency injection involves four roles:

  • the service object(s) to be used
  • the client object that is depending on the services it uses
  • the interfaces that define how the client may use the services
  • the injector, which is responsible for constructing the services and injecting them into the client

Any object that may be used can be considered a service. Any object that uses other objects can be considered a client. The names have nothing to do with what the objects are for and everything to do with the role the objects play in any one injection.

The interfaces are the types the client expects its dependencies to be. At issue is what they make accessible. They may truly be interface types implemented by the services but also may be abstract classes or even the concrete services themselves, though this last would violate DIP[10] and sacrifice the dynamic decoupling that enables testing. It's only required that the client does not know which they are and therefore never treats them as concrete, say by constructing or extending them.

Given that the client has no concrete knowledge, so long as the what the client uses, the interfaces name and API then the client won't need to change even if what is behind the interface changes. However, if the interface is refactored from being a class to an interface type (or vice versa) the client will need to be recompiled.[11] This is significant if the client and services are published separately. This unfortunate coupling is one that dependency injection cannot resolve.

The injector introduces the services into the client. Often, it also constructs the client. An injector may connect together a very complex object graph by treating an object like a client and later as a service for another client. The injector may actually be many objects working together but may not be the client. The injector may be referred to by other names such as: assembler, provider, container, factory, builder, spring, construction code, or main.

Dependency injection can be applied as a discipline. One that asks that all objects separate construction and behavior. Relying on a DI framework to perform construction can lead to forbidding the use of the new keyword, or, less strictly, only allow direct construction of value objects.[12][13][14][15]

Taxonomy

Inversion of Control (IoC) is more general than DI. Put simply, IoC means letting other code call you rather than insisting on doing the calling. An example of IoC without DI is the template pattern. Here polymorphism is achieved through subclassing, that is, inheritance.[16]

Dependency injection implements IoC through composition so is often identical to that of the strategy pattern, but while the strategy pattern is intended for dependencies to be interchangeable throughout an object's lifetime, in dependency injection it may be that only a single instance of a dependency is used.[17] This still achieves polymorphism, but through delegation and composition.

Dependency injection frameworks

Application frameworks such as Spring, Guice, Play framework, Salta, Glassfish HK2, and Microsoft Managed Extensibility Framework (MEF) support dependency injection but are not required to do dependency injection.[18][19]

Advantages

  • Dependency injection allows a client the flexibility of being configurable. Only the client's behavior is fixed. The client may act on anything that supports the intrinsic interface the client expects.
  • Dependency injection can be used to externalize a system's configuration details into configuration files allowing the system to be reconfigured without recompilation. Separate configurations can be written for different situations that require different implementations of components. This includes, but is not limited to, testing.
  • Because dependency injection doesn't require any change in code behavior it can be applied to legacy code as a refactoring. The result is clients that are more independent and that are easier to unit test in isolation using stubs or mock objects that simulate other objects not under test. This ease of testing is often the first benefit noticed when using dependency injection.
  • Dependency injection allows a client to remove all knowledge of a concrete implementation that it needs to use. This helps isolate the client from the impact of design changes and defects. It promotes reusability, testability and maintainability.[20]
  • Reduction of boilerplate code in the application objects since all work to initialize or set up dependencies is handled by a provider component.[20]
  • Dependency injection allows concurrent or independent development. Two developers can independently develop classes that use each other, while only needing to know the interface the classes will communicate through. Plugins are often developed by third party shops that never even talk to the developers who created the product that uses the plugins.
  • Dependency Injection decreases coupling between a class and its dependency.[21][22]

Disadvantages

  • Dependency injection creates clients that demand configuration details be supplied by construction code. This can be onerous when obvious defaults are available.
  • Dependency injection can make code difficult to trace (read) because it separates behavior from construction. This means developers must refer to more files to follow how a system performs.
  • Dependency injection typically requires more upfront development effort since one can not summon into being something right when and where it is needed but must ask that it be injected and then ensure that it is injected.
  • Dependency injection can cause an explosion of types, especially in languages that have explicit interface types, like Java and C# [23]
  • Dependency injection forces complexity to move out of classes and into the linkages between classes which might not always be desirable or easily managed.[24]
  • Ironically, dependency injection can encourage dependence on a dependency injection framework.[24][25][26]

Examples

Without dependency injection

In the following Java example, the Client class contains a Service member variable that is initialized by the Client constructor. The client controls which implementation of service is used and controls its construction. In this situation, the client is said to have a hard-coded dependency on ServiceExample.

// An example without dependency injection
public class Client {
    // Internal reference to the service used by this client
    private Service service;

    // Constructor
    Client() {
        // Specify a specific implementation in the constructor instead of using dependency injection
        this.service = new ServiceExample();
    }

    // Method within this client that uses the services
    public String greet() {
        return "Hello " + service.getName();
    }
}

Dependency injection is an alternative technique to initialize the member variable rather than explicitly creating a service object as shown above.

Three types of dependency injection

There are at least three ways an object can receive a reference to an external module:[27]

  • constructor injection: the dependencies are provided through a class constructor.
  • setter injection: the client exposes a setter method that the injector uses to inject the dependency.
  • interface injection: the dependency provides an injector method that will inject the dependency into any client passed to it. Clients must implement an interface that exposes a setter method that accepts the dependency.

Other types

It is possible for DI frameworks to have other types of injection beyond those presented above.[28]

Testing frameworks may also use other types. Some modern testing frameworks do not even require that clients actively accept dependency injection thus making legacy code testable. In particular, in the Java language it is possible to use reflection to make private attributes public when testing and thus accept injections by assignment.[29]

Some attempts at Inversion of Control do not provide full removal of dependency but instead simply substitute one form of dependency for another. As a rule of thumb, if a programmer can look at nothing but the client code and tell what framework is being used, then the client has a hard-coded dependency on the framework.

Constructor injection

This method requires the client to provide a parameter in a constructor for the dependency.

// Constructor
Client(Service service) {
    // Save the reference to the passed-in service inside this client
    this.service = service;
}

Setter injection

This method requires the client to provide a setter method for each dependency.

// Setter method
public void setService(Service service) {
    // Save the reference to the passed-in service inside this client
    this.service = service;
}

Interface injection

This is simply the client publishing a role interface to the setter methods of the client's dependencies. It can be used to establish how the injector should talk to the client when injecting dependencies.

// Service setter interface.
public interface ServiceSetter {
    public void setService(Service service);
}

// Client class
public class Client implements ServiceSetter {
    // Internal reference to the service used by this client.
    private Service service;

    // Set the service that this client is to use.
    @Override
    public void setService(Service service) {
        this.service = service;
    }
}

Constructor injection comparison

Preferred when all dependencies can be constructed first because it can be used to ensure the client object is always in a valid state, as opposed to having some of its dependency references be null (not be set). However, on its own, it lacks the flexibility to have its dependencies changed later. This can be a first step towards making the client immutable and therefore thread safe.

// Constructor
Client(Service service, Service otherService) {
    if (service == null) {
        throw new InvalidParameterException("service must not be null");
    }
    if (otherService == null) {
        throw new InvalidParameterException("otherService must not be null");
    }

    // Save the service references inside this client
    this.service = service;
    this.otherService = otherService;
}

Setter injection comparison

Requires the client to provide a setter method for each dependency. This gives the freedom to manipulate the state of the dependency references at any time. This offers flexibility, but if there is more than one dependency to be injected, it is difficult for the client to ensure that all dependencies are injected before the client could be provided for use.

// Set the service to be used by this client
public void setService(Service service) {
    if (service == null) {
        throw new InvalidParameterException("service must not be null");
    }
    this.service = service;
}

// Set the other service to be used by this client
public void setOtherService(Service otherService) {
    if (otherService == null) {
        throw new InvalidParameterException("otherService must not be null");
    }
    this.otherService = otherService;
}

Since these injections happen independently there is no way to tell when the injector is finished wiring the client. A dependency can be left null simply by the injector failing to call its setter. This forces the check that injection was completed from when the client is assembled to whenever it is used.

// Set the service to be used by this client
public void setService(Service service) {
    this.service = service;
}

// Set the other service to be used by this client
public void setOtherService(Service otherService) {
    this.otherService = otherService;
}

// Check the service references of this client
private void validateState() {
    if (service == null) {
        throw new IllegalStateException("service must not be null");
    }
    if (otherService == null) {
        throw new IllegalStateException("otherService must not be null");
    }
}

// Method that uses the service references
public void doSomething() {
    validateState();
    service.doYourThing();
    otherService.doYourThing();
}

Interface injection comparison

The advantage of interface injection is that dependencies can be completely ignorant of their clients yet can still receive a reference to a new client and, using it, send a reference-to-self back to the client. In this way, the dependencies become injectors. The key is that the injecting method (which could just be a classic setter method) is provided through an interface.

An assembler is still needed to introduce the client and its dependencies. The assembler would take a reference to the client, cast it to the setter interface that sets that dependency, and pass it to that dependency object which would turn around and pass a reference-to-self back to the client.

For interface injection to have value, the dependency must do something in addition to simply passing back a reference to itself. This could be acting as a factory or sub-assembler to resolve other dependencies, thus abstracting some details from the main assembler. It could be reference-counting so that the dependency knows how many clients are using it. If the dependency maintains a collection of clients, it could later inject them all with a different instance of itself.

Assembling examples

Manually assembling in main by hand is one way of implementing dependency injection.

public class Injector {
    public static void main(String[] args) {
        // Build the dependencies first
        Service service = new ServiceExample();

        // Inject the service, constructor style
        Client client = new Client(service);

        // Use the objects
        System.out.println(client.greet());
    }	
}

The example above constructs the object graph manually and then invokes it at one point to start it working. Important to note is that this injector is not pure. It uses one of the objects it constructs. It has a purely construction-only relationship with ServiceExample but mixes construction and using of Client. This should not be common. It is, however, unavoidable. Just like object oriented software needs a non-object oriented static method like main() to get started, a dependency injected object graph needs at least one (preferably only one) entry point to get the using started.

Manual construction in the main method may not be this straight forward and may involve calling builders, factories, or other construction patterns as well. This can be fairly advanced and abstract. The line is crossed from manual dependency injection to framework dependency injection once the constructing code is no longer custom to the application and is instead universal.[30]

Frameworks like Spring can construct these same objects and wire them together before returning a reference to client. Notice that all mention of ServiceExample has been removed from the code.

import org.springframework.beans.factory.BeanFactory;
import org.springframework.context.ApplicationContext;
import org.springframework.context.support.ClassPathXmlApplicationContext;

public class Injector {

	public static void main(String[] args) {

		// -- Assembling objects -- //

        	BeanFactory beanfactory=new ClassPathXmlApplicationContext("Beans.xml");

	        Client client=(Client)beanfactory.getBean("client");

		// -- Using objects -- //
		
		System.out.println( client.greet() );
	}
}

Frameworks like Spring allow assembly details to be externalized in configuration files. This code (above) constructs objects and wires them together according to Beans.xml (below). ServiceExample is still constructed even though it's only mentioned below. A long and complex object graph can be defined this way and the only class mentioned in code would be the one with the entry point method, which in this case is greet().

 <?xml version="1.0" encoding="UTF-8"?>
 <beans xmlns="http://www.springframework.org/schema/beans"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://www.springframework.org/schema/beans
  http://www.springframework.org/schema/beans/spring-beans-3.0.xsd">

    <bean id="service" class="ServiceExample">
    </bean>

    <bean id="client" class="Client">
        <constructor-arg value="service" />        
    </bean>
</beans>

In the example above Client and Service have not had to undergo any changes to be provided by spring. They are allowed to remain simple POJOs.[31][32][33] This shows how spring can connect services and clients that are completely ignorant of its existence. This could not be said if spring annotations are added to the classes. By keeping spring specific annotations and calls from spreading out among many classes, the system stays only loosely dependent on spring.[25] This can be important if the system intends to outlive spring.

The choice to keep POJOs pure doesn't come without cost. Rather than spending the effort to develop and maintain complex configuration files it is possible to simply use annotations to mark classes and let spring do the rest of the work. Resolving dependencies can be simple if they follow a convention such as matching by type or by name. This is choosing convention over configuration.[34] It is also arguable that, when refactoring to another framework, removing framework specific annotations would be a trivial part of the task[35] and many injection annotations are now standardized.[36][37]

import org.springframework.beans.factory.BeanFactory;
import org.springframework.context.ApplicationContext;
import org.springframework.context.annotation.AnnotationConfigApplicationContext;

public class Injector {
    public static void main(String[] args) {
        // Assemble the objects
        BeanFactory beanfactory = new AnnotationConfigApplicationContext(MyConfiguration.class);
        Client client = beanfactory.getBean(Client.class);

        // Use the objects
        System.out.println(client.greet());
    }
}
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.ComponentScan;
import org.springframework.context.annotation.Configuration;

@ComponentScan
static class MyConfiguration {
    @Bean
    public Client client(ServiceExample service) {
        return new Client(service);
    }
}
@Component
public class ServiceExample {
    public String getName() {
        return "World!";
    }
}

Assembly comparison

The different injector implementations (factories, service locators, and dependency injection containers) are not that different as far as dependency injection is concerned. What makes all the difference is where they are allowed to be used. Move calls to a factory or a service locator out of the client and into main and suddenly main makes a fairly good dependency injection container.

By moving all knowledge of the injector out, a clean client, free of knowledge of the outside world, is left behind. However, any object that uses other objects can be considered a client. The object that contains main is no exception. This main object is not using dependency injection. It's actually using the service locator pattern. This can't be avoided because the choice of service implementations must be made somewhere.

Externalizing the dependencies into configuration files doesn't change this fact. What makes this reality part of a good design is that the service locator is not spread throughout the code base. It's confined to one place per application. This leaves the rest of the code base free to use dependency injection to make clean clients.

AngularJS example

In the AngularJS framework, there are only three ways a component (object or function) can directly access its dependencies:

  1. The component can create the dependency, typically using the new operator.
  2. The component can look up the dependency, by referring to a global variable.
  3. The component can have the dependency passed to it where it is needed.

The first two options of creating or looking up dependencies are not optimal because they hard code the dependency to the component. This makes it difficult, if not impossible, to modify the dependencies. This is especially problematic in tests, where it is often desirable to provide mock dependencies for test isolation.

The third option is the most viable, since it removes the responsibility of locating the dependency from the component. The dependency is simply handed to the component.

function SomeClass(greeter) {
  this.greeter = greeter;
}

SomeClass.prototype.doSomething = function(name) {
  this.greeter.greet(name);
}

In the above example SomeClass is not concerned with creating or locating the greeter dependency, it is simply handed the greeter when it is instantiated.

This is desirable, but it puts the responsibility of getting hold of the dependency on the code that constructs SomeClass.

To manage the responsibility of dependency creation, each AngularJS application has an injector. The injector is a service locator that is responsible for construction and look-up of dependencies.

Here is an example of using the injector service:

// Provide the wiring information in a module
var myModule = angular.module('myModule', []);

// Teach the injector how to build a greeter service. 
// Notice that greeter is dependent on the $window service. 
// The greeter service is an object that
// contains a greet method.
myModule.factory('greeter', function($window) {
  return {
    greet: function(text) {
      $window.alert(text);
    }
  };
});

Create a new injector that can provide components defined in the myModule module and request our greeter service from the injector. (This is usually done automatically by the AngularJS bootstrap).

var injector = angular.injector(['myModule', 'ng']);
var greeter = injector.get('greeter');

Asking for dependencies solves the issue of hard coding, but it also means that the injector needs to be passed throughout the application. Passing the injector breaks the Law of Demeter. To remedy this, we use a declarative notation in our HTML templates, to hand the responsibility of creating components over to the injector, as in this example:

<div ng-controller="MyController">
  <button ng-click="sayHello()">Hello</button>
</div>
function MyController($scope, greeter) {
  $scope.sayHello = function() {
    greeter.greet('Hello World');
  };
}

When AngularJS compiles the HTML, it processes the ng-controller directive, which in turn asks the injector to create an instance of the controller and its dependencies.

injector.instantiate(MyController);

This is all done behind the scenes. Notice that by having the ng-controller ask the injector to instantiate the class, it can satisfy all of the dependencies of MyController without the controller ever knowing about the injector. This is the best outcome. The application code simply declares the dependencies it needs, without having to deal with the injector. This setup does not break the Law of Demeter.

See also

References

  1. Lua error in package.lua at line 80: module 'strict' not found.
  2. Lua error in package.lua at line 80: module 'strict' not found.
  3. Lua error in package.lua at line 80: module 'strict' not found.
  4. 4.0 4.1 Lua error in package.lua at line 80: module 'strict' not found.
  5. Lua error in package.lua at line 80: module 'strict' not found.
  6. Lua error in package.lua at line 80: module 'strict' not found.
  7. Lua error in package.lua at line 80: module 'strict' not found.
  8. Niko Schwarz, Mircea Lungu, Oscar Nierstrasz, “Seuss: Decoupling responsibilities from static methods for fine-grained configurability”, Journal of Object Technology, Volume 11, no. 1 (April 2012), pp. 3:1-23
  9. Lua error in package.lua at line 80: module 'strict' not found.
  10. Lua error in package.lua at line 80: module 'strict' not found.
  11. Lua error in package.lua at line 80: module 'strict' not found.
  12. Lua error in package.lua at line 80: module 'strict' not found.
  13. Lua error in package.lua at line 80: module 'strict' not found.
  14. Lua error in package.lua at line 80: module 'strict' not found.
  15. Lua error in package.lua at line 80: module 'strict' not found.
  16. Lua error in package.lua at line 80: module 'strict' not found.
  17. Lua error in package.lua at line 80: module 'strict' not found.
  18. Lua error in package.lua at line 80: module 'strict' not found.
  19. Lua error in package.lua at line 80: module 'strict' not found.
  20. 20.0 20.1 Lua error in package.lua at line 80: module 'strict' not found.
  21. Lua error in package.lua at line 80: module 'strict' not found.
  22. Lua error in package.lua at line 80: module 'strict' not found.
  23. Lua error in package.lua at line 80: module 'strict' not found.
  24. 24.0 24.1 Lua error in package.lua at line 80: module 'strict' not found.
  25. 25.0 25.1 Lua error in package.lua at line 80: module 'strict' not found.
  26. Lua error in package.lua at line 80: module 'strict' not found.
  27. Lua error in package.lua at line 80: module 'strict' not found.
  28. Lua error in package.lua at line 80: module 'strict' not found.
  29. Lua error in package.lua at line 80: module 'strict' not found.
  30. Lua error in package.lua at line 80: module 'strict' not found.
  31. Lua error in package.lua at line 80: module 'strict' not found.
  32. Lua error in package.lua at line 80: module 'strict' not found.
  33. Lua error in package.lua at line 80: module 'strict' not found.
  34. Lua error in package.lua at line 80: module 'strict' not found.
  35. Lua error in package.lua at line 80: module 'strict' not found.
  36. Lua error in package.lua at line 80: module 'strict' not found.
  37. Lua error in package.lua at line 80: module 'strict' not found.

External links