Skip to main content

Unit Testing Asp.NetCore Web API Using xUnit[.NET6]

In this article, we are going to write test cases to an Asp.NetCore Web API(.NET6) application using the xUnit.

xUnit For .NET:

The xUnit for .Net is a free, open-source, community-focused unit testing tool for .NET applications. By default .Net also provides a xUnit project template to implement test cases.

Unit test cases build upon the 'AAA' formula that means 'Arrange', 'Act' and 'Assert'
Arrange - Declaring variables, objects, instantiating mocks, etc.
Act - Calling or invoking the method that needs to be tested.
Assert - The assert ensures that code behaves as expected means yielding expected output.

Create An API And Unit Test Projects:

Let's create a .Net6 Web API and xUnit sample applications to accomplish our demo. We can use either Visual Studio 2022 or Visual Studio Code(using .NET CLI commands) to create any.Net6 application. For this demo, I'm using the 'Visual Studio Code'(using the .NET CLI command) editor.

Create a folder where we want to set up our applications. Then add a solution file by running the below command.
Command To Add Solution File:
dotnet new sln Name_of_your_solution_file

Now let's create a .NET6 Web API project.
dotnet new webapi -o Name_of_your_API_project

Now let's create xUnit project.
dotnet new xunit -o Name_of_your_API_project

Now add our API and xUnit project to our solution file
dotnet sln {your_solution}.sln add {Your_API_Project_Folder}/{Your_API_ProjectName}.csproj

dotnet sln {your_solution}.sln add {Your_xUnit_Project_Folder}/{Your_xUnit_ProjectName}.csproj

xUnit Project Folder Structure:

  • The 'System' folder represents the entire application it contains all testing files under subfolders like 'Controllers'(folder contains all testing files related to all our controllers), 'Services'(folder contains all service files(or business logic files) related to our application).
  • The 'MockData' contains files related to mocking data that can be used for response or request or input or output data object for testing.
  • The 'Helper' contains any common logic.

Install Required NuGet Package:

Now install the 'Moq' NuGet Package in the xUnit project.
Package Manager Command:
Install-Package Moq -Version 4.16.1

.NET CLI Command:
dotnet add package Moq --version 4.16.1

Now install the 'FluentAssertion' Package in the xUnit project.
Package Manager Command:
Install-Package FluentAssertions -Version 6.5.1

.NET CLI Command:
dotnet add package FluentAssertions --version 6.5.1

Now install the 'Microsoft.EntityFrameworkCore.InMemory' Package in the xUnit project
Package Manager Command:
Install-Package Microsoft.EntityFrameworkCore.InMemory -Version 6.0.2

.NET CLI Command:
dotnet add package Microsoft.EntityFrameworkCore.InMemory --version 6.0.2

Unit Test Cases On GetAction Method In Controller:

Let's try to write some unit test cases for the GetAction method of the API controller.

First, let's add your API Project assembly file reference into the xUnit project.

So our API action looks like as below:
ApiProject - Controllers/TodoController.cs:
[ApiController]
[Route("[controller]")]
public class TodoController : ControllerBase
{
    private readonly ITodoService _todoService;
    public TodoController(ITodoService todoService)
    {
        _todoService = todoService;
    }
    [Route("get-all")]
    [HttpGet]
    public async Task<IActionResult> GetAllAsync()
    {
        var result = await _todoService.GetAllAsync();
        if (result.Count == 0)
        {
            return NoContent();
        }
        return Ok(result);
    }
}
  • Here 'GetAllAsync()' method is an GetAction method that returns the collection some collection of 'TodoItems'
Let's first create some mock data for the 'Todo' items in the xUnit Project.
xUnitProject - MockData/TodoMockData.cs:
using System.Collections.Generic;
using Demo1.Api.Data.Entities;
namespace Demo1.TestApi.MockData;
public class TodoMockData
{
    public static List<Todo> GetTodos()
    {
        return new List<Todo>{
             new Todo{
                 Id = 1,
                 ItemName = "Need To Go Shopping",
                 IsCompleted = true
             },
             new Todo{
                 Id = 2,
                 ItemName = "Cook Food",
                 IsCompleted = true
             },
             new Todo{
                 Id = 3,
                 ItemName = "Play Games",
                 IsCompleted = false
             }
         };
    }

}

Now let's first create a test file like 'System/Controllers/TestTodoController.cs'(xUnit Project file)  for 'TodoController.cs'(API Project File).

Now let's add test cases for the 'GetAllAsync()' method(API Project method). So the first test case I want to implement is to check the '200 status code'.
xUnitProject - System/Controllers/TestTodoController.cs:
using System.Threading.Tasks;
using Demo1.Api.Controllers;
using Demo1.Api.Services;
using Demo1.TestApi.MockData;
using FluentAssertions;
using Microsoft.AspNetCore.Mvc;
using Moq;
using Xunit;

namespace Demo1.TestApi.System.Controllers;

public class TestController
{
    [Fact]
    public async Task GetAllAsync_ShouldReturn200Status()
    {
        /// Arrange
        var todoService = new Mock<ITodoService>();
        todoService.Setup(_ => _.GetAllAsync()).ReturnsAsync(TodoMockData.GetTodos());
        var sut = new TodoController(todoService.Object);

        /// Act
        var result = (OkObjectResult)await sut.GetAllAsync();


        // /// Assert
        result.StatusCode.Should().Be(200);
    }
}
  • (Line: 14) The method is decorated with an attribute like 'Fact'. So this attribute determines that the method should be run by the test runner.
  • (Line: 15) The recommended naming convention of the test method is 'Actual method name' + 'Name of the test'.
  • (Line: 17&22&26) Here we can observe code separated based on the 'AAA' formula.
  • (Line: 18-20) We want to test the controller 'GetAllAsync()' method and we know that our controller's action method depends on the 'TodoService.GetAllAsync()' method result which means our controller's action method doesn't care about the actual logical implementation of the 'TodoService.GetAllAsync()'.So we create a mock instance of 'ITodoService'(at line 18) and then mock the result of 'ITodoService.GetAllAsync()' method(at line 19). Finally creates the instance of 'TodoController'(at line 20). Here variable 'sut' means 'System Under Test' just a recommended naming convention.
  • (Line: 23) Here invoking our  controllers action method 'GetAllAsync()'. Since our action method returns 'OkObjectResult' for 200 status so here we explicitly typecasting the result.
  • (Line: 27) Checking our expected result that is 200 as the status code.

Now let's write one more test like 'TodoController.GetAllAsync()' returns 204 status when it has no data to return as a response.

Let's add an empty collection of mock data.
xUnitProject - MockData/TodoMockData.cs:
using System.Collections.Generic;
using Demo1.Api.Data.Entities;

namespace Demo1.TestApi.MockData;
public class TodoMockData
{
    // existing code hidden for display purpose
    public static List<Todo> GetEmptyTodos()
    {
        return new List<Todo>();
    }
}
Now let's write our test case like our expected status code should 204.
xUnitProject - System/Controllers/TestController.cs:
using System.Threading.Tasks;
using Demo1.Api.Controllers;
using Demo1.Api.Services;
using Demo1.TestApi.MockData;
using FluentAssertions;
using Microsoft.AspNetCore.Mvc;
using Moq;
using Xunit;

namespace Demo1.TestApi.System.Controllers;

public class TestController
{
    // existing cod hidden for display purpose
    [Fact]
    public async Task GetAllAsync_ShouldReturn204NoContentStatus()
    {
        /// Arrange
        var todoService = new Mock<ITodoService>();
        todoService.Setup(_ => _.GetAllAsync()).ReturnsAsync(TodoMockData.GetEmptyTodos());
        var sut = new TodoController(todoService.Object);

        /// Act
        var result = (NoContentResult)await sut.GetAllAsync();


        /// Assert
        result.StatusCode.Should().Be(204);
        todoService.Verify(_ => _.GetAllAsync(),Times.Exactly(1));
    }
}
  • (Line: 20) Mocking our 'ITodoSerivice.GetAllAsync()' method to return the empty collection of data.
  • (Line: 24) Invoking our 'TodoController.GetAllAsync()' method.
  • (Line: 28) Expecting our status code to be '204'
  • (Line: 29) Here verifying the 'ITodoService.GetAllAsync()' is called exactly one time inside of our 'TodoConroller.GetAllAsync()' method.

Unit Test Case On Get Method In Services:

Let's try to write a  unit test case for the 'Get' method in our services. So our 'TodoService.cs'(in API Project) get method looks as below:
ApiProject - Services/TodoService.cs:
using Demo1.Api.Data;
using Demo1.Api.Data.Entities;
using Microsoft.EntityFrameworkCore;
namespace Demo1.Api.Services;

public class TodoService: ITodoService
{
    private readonly MyWorldDbContext _context;
    public TodoService(MyWorldDbContext context)
    {
        _context = context;
    }

    public async Task<List<Todo>> GetAllAsync()
    {
       return await _context.Todo.ToListAsync();
    }
}
  • Here our 'GetAllAsync' method simply returns the collection of 'Todo' items. One more thing here we have to observe is that our 'GetAllAsync()' method directly depends on the 'MyWorldDbContext'.
Now let's create a test file like 'System/Services/TestTodoService.cs'. Now let's add unit test as follows:
xUnitProject - System/Services/TestTodoService.cs:
using System;
using System.Threading.Tasks;
using Demo1.Api.Data;
using Demo1.Api.Services;
using Demo1.TestApi.MockData;
using FluentAssertions;
using Microsoft.EntityFrameworkCore;
using Xunit;
namespace Demo1.TestApi.System.Services;

public class TestTodoService : IDisposable
{
    protected readonly MyWorldDbContext _context;
    public TestTodoService()
    {
        var options = new DbContextOptionsBuilder<MyWorldDbContext>()
        .UseInMemoryDatabase(databaseName: Guid.NewGuid().ToString())
        .Options;

        _context = new MyWorldDbContext(options);

        _context.Database.EnsureCreated();
    }

    [Fact]
    public async Task GetAllAsync_ReturnTodoCollection()
    {
        /// Arrange
        _context.Todo.AddRange(MockData.TodoMockData.GetTodos());
        _context.SaveChanges();

        var sut = new TodoService(_context);

        /// Act
        var result = await sut.GetAllAsync();

        /// Assert
        result.Should().HaveCount(TodoMockData.GetTodos().Count);
    }

    public void Dispose()
    {
        _context.Database.EnsureDeleted();
        _context.Dispose();
    }
}
  • (Line: 11) Inherited the 'IDisposable'.
  • (Line: 14-23) Instead of mocking the 'MyWorldDbContext', we are going to create the in-memory database, so that our testing will more easy and more flexible. Here constructor gets invoked for every execution of the test method. So we want to create a separate in-memory database for each uni test method, so we have to define the unique database name which we are going to achieve by using the 'Guid'.
  • (Line: 22) The 'EnsureCreated()' method to determine database for the context is exist.
  • (Line: 29-30) Here seeding some test data into the 'Todo' table of our in-memory database.
  • (Line: 32) Creating the object for our 'TodoService'.
  • (Line: 35) Invoking our test method that is 'TodoService.GetAllAsync()'.
  • (Line: 38) Verifying the output as expected or not.
  • (Line: 41-45) The 'Dispose' method gets executed on completion of test case method execution. So here we want to destroy the in-memory database so that every test case will have its own in-memory database.

Unit Test Case On PostAction Method In Controller:

Now let's try to write a unit test case for an HTTP POST action method of our controller. So our API PostAction method looks as follows:
ApiProject - Controllers/TodoController.cs:
using Demo1.Api.Data.Entities;
using Demo1.Api.Services;
using Microsoft.AspNetCore.Mvc;
namespace Demo1.Api.Controllers;

[ApiController]
[Route("[controller]")]
public class TodoController : ControllerBase
{
    private readonly ITodoService _todoService;
    public TodoController(ITodoService todoService)
    {
        _todoService = todoService;
    }
    
	// existing code hidden for display purpose
    [HttpPost]
    [Route("save")]
    public async Task<IActionResult> SaveAsync(Todo newTodo)
    {
        await _todoService.SaveAsync(newTodo);
        return Ok();
    }
}
  • Here 'SaveAsync()' method is the PostAction method, this method is for saving the new records into the database.
Let's create mock data that we want to use as payload to our test method.
xUnitProject - MockData/TodoMockData.cs:
using System.Collections.Generic;
using Demo1.Api.Data.Entities;

namespace Demo1.TestApi.MockData;
public class TodoMockData
{
	// existing code hidden for display purpose
    public static Todo NewTodo()
    {
        return new Todo
        {
            Id = 0,
            ItemName = "wash clothes",
            IsCompleted = false
        };
    }
}
Now let's write a test case for our 'TodoController.SaveAsync' method.
xUnitProject- Controllers/TestTodoController.cs:
using System.Threading.Tasks;
using Demo1.Api.Controllers;
using Demo1.Api.Services;
using Demo1.TestApi.MockData;
using FluentAssertions;
using Microsoft.AspNetCore.Mvc;
using Moq;
using Xunit;
namespace Demo1.TestApi.System.Controllers;

public class TestController
{
    [Fact]
    public async Task SaveAsync_ShouldCall_ITodoService_SaveAsync_AtleastOnce()
    {
        /// Arrange
        var todoService = new Mock<ITodoService>();
        var newTodo = TodoMockData.NewTodo();
        var sut = new TodoController(todoService.Object);

        /// Act
        var result = await sut.SaveAsync(newTodo);

        /// Assert
        todoService.Verify(_ => _.SaveAsync(newTodo),Times.Exactly(1));
    }
}
  • (Line: 18) Mocked payload object.
  • (Line: 22) Invoke the method that needs to be tested that is 'TodoController.SaveAsyn()'.
  • (Line: 25) Since our 'TodoController.SaveAsync()' method returns nothing so to accomplish our testing here we determine that the 'TodoService.SaveAsync()' method is executed exactly one time.

Unit Test Case On Save Method In Services:

Let's try to write a unit test case for the 'Save' method in our services. So our 'TodoService.cs'(in API Project) save method looks as follows: 
using Demo1.Api.Data;
using Demo1.Api.Data.Entities;
using Microsoft.EntityFrameworkCore;
namespace Demo1.Api.Services;

public class TodoService : ITodoService
{
    private readonly MyWorldDbContext _context;
    public TodoService(MyWorldDbContext context)
    {
        _context = context;
    }
	// existing code hidden for the display purpose
    public async Task SaveAsync(Todo newTodo)
    {
        _context.Todo.Add(newTodo);
        await _context.SaveChangesAsync();
    }
}
  • Here our 'SaveAsync()' method saves a new item into the database.
Now let's implement our test case for the 'TodoService.SaveAsync()' method.
xUnitProject - Services/TestTodoService.cs:
using System;
using System.Linq;
using System.Threading.Tasks;
using Demo1.Api.Data;
using Demo1.Api.Services;
using Demo1.TestApi.MockData;
using FluentAssertions;
using Microsoft.EntityFrameworkCore;
using Xunit;

namespace Demo1.TestApi.System.Services;


public class TestTodoService : IDisposable
{
    protected readonly MyWorldDbContext _context;
    public TestTodoService()
    {
        var options = new DbContextOptionsBuilder<MyWorldDbContext>()
        .UseInMemoryDatabase(databaseName: Guid.NewGuid().ToString())
        .Options;

        _context = new MyWorldDbContext(options);

        _context.Database.EnsureCreated();
    }

    [Fact]
    public async Task SaveAsync_AddNewTodo()
    {
        /// Arrange
        var newTodo = TodoMockData.NewTodo();
        _context.Todo.AddRange(MockData.TodoMockData.GetTodos());
        _context.SaveChanges();

        var sut = new TodoService(_context);

        /// Act
        await sut.SaveAsync(newTodo);

        ///Assert
        int expectedRecordCount = (TodoMockData.GetTodos().Count() + 1);
        _context.Todo.Count().Should().Be(expectedRecordCount);
    }

    public void Dispose()
    {
        _context.Database.EnsureDeleted();
        _context.Dispose();
    }
}
  • (Line: 32) The 'newTodo' variable holds the mock data that we want to insert as a new record.
  • (Line: 33-34) Seeding dummy data into the in-memory database.
  • (Line: 39) Invoking our test method that is 'ITodoService.SaveAsync()' method.
  • (Line: 42-43) Verifying the total records count in the database as our expected count after inserting the new record.

Support Me!
Buy Me A Coffee PayPal Me

Video Session:

Wrapping Up:

Hopefully, I think this article delivered some useful information on Unit Testing in Asp.NetCore Web API using xUnit. I love to have your feedback, suggestions, and better techniques in the comment section below.

Refer:

Follow Me:

Comments

  1. thank you for sharing ! really liked the way you organized the project structure and details

    very nice

    ReplyDelete
  2. can you please give the source code link ?

    ReplyDelete
    Replies
    1. https://github.com/Naveen512/Dotnet6-API-XUnit-Testing

      Delete
  3. you need to explain where you have used the different nuget packages

    ReplyDelete
  4. if you mock everything, why no just use regex to test that controller's source code contains line "await _todoService.GetAllAsync();"?

    ReplyDelete

Post a Comment

Popular posts from this blog

Angular 14 Reactive Forms Example

In this article, we will explore the Angular(14) reactive forms with an example. Reactive Forms: Angular reactive forms support model-driven techniques to handle the form's input values. The reactive forms state is immutable, any form filed change creates a new state for the form. Reactive forms are built around observable streams, where form inputs and values are provided as streams of input values, which can be accessed synchronously. Some key notations that involve in reactive forms are like: FormControl - each input element in the form is 'FormControl'. The 'FormControl' tracks the value and validation status of form fields. FormGroup - Track the value and validate the state of the group of 'FormControl'. FormBuilder - Angular service which can be used to create the 'FormGroup' or FormControl instance quickly. Form Array - That can hold infinite form control, this helps to create dynamic forms. Create An Angular(14) Application: Let'

.NET 7 Web API CRUD Using Entity Framework Core

In this article, we are going to implement a sample .NET 7 Web API CRUD using the Entity Framework Core. Web API: Web API is a framework for building HTTP services that can be accessed from any client like browser, mobile devices, and desktop apps. In simple terminology API(Application Programming Interface) means an interface module that contains programming functions that can be requested via HTTP calls either to fetch or update data for their respective clients. Some of the Key Characteristics of API: Supports HTTP verbs like 'GET', 'POST', 'PUT', 'DELETE', etc. Supports default responses like 'XML' and 'JSON'. Also can define custom responses. Supports self-hosting or individual hosting, so that all different kinds of apps can consume it. Authentication and Authorization are easy to implement. The ideal platform to build the REST full services. Install The SQL Server And SQL Management Studio: Let's install the SQL server on our l

ReactJS(v18) JWT Authentication Using HTTP Only Cookie

In this article, we will implement the ReactJS application authentication using the HTTP-only cookie. HTTP Only Cookie: In a SPA(Single Page Application) Authentication JWT token either can be stored in browser 'LocalStorage' or in 'Cookie'. Storing the JWT token inside of the cookie then the cookie should be HTTP Only. The HTTP-ONly cookie nature is that it will be only accessible by the server application. Client apps like javascript-based apps can't access the HTTP-Only cookie. So if we use the authentication with HTTP-only JWT cookie then we no need to implement the custom logic like adding authorization header or storing token data, etc at our client application. Because once the user authenticated cookie will be automatically sent to the server by the browser on every API call. Authentication API: To authenticate our client application with JWT HTTP-only cookie, I developed a NetJS(which is a node) Mock API. Check the GitHub link and read the document on G

.NET6 Web API CRUD Operation With Entity Framework Core

In this article, we are going to do a small demo on AspNetCore 6 Web API CRUD operations. What Is Web API: Web API is a framework for building HTTP services that can be accessed from any client like browser, mobile devices, desktop apps. In simple terminology API(Application Programming Interface) means an interface module that contains a programming function that can be requested via HTTP calls to save or fetch the data for their respective clients. Some of the key characteristics of API: Supports HTTP verbs like 'GET', 'POST', 'PUT', 'DELETE', etc. Supports default responses like 'XML' and 'JSON'. Also can define custom responses. Supports self-hosting or individual hosting, so that all different kinds of apps can consume it. Authentication and Authorization are easy to implement. The ideal platform to build REST full services. Create A .NET6 Web API Application: Let's create a .Net6 Web API sample application to accomplish our

Angular 14 State Management CRUD Example With NgRx(14)

In this article, we are going to implement the Angular(14) state management CRUD example with NgRx(14) NgRx Store For State Management: In an angular application to share consistent data between multiple components, we use NgRx state management. Using NgRx state helps to avoid unwanted API calls, easy to maintain consistent data, etc. The main building blocks for the NgRx store are: Actions - NgRx actions represents event to trigger the reducers to save the data into the stores. Reducer - Reducer's pure function, which is used to create a new state on data change. Store - The store is the model or entity that holds the data. Selector - Selector to fetch the slices of data from the store to angular components. Effects - Effects deals with external network calls like API. The effect gets executed based the action performed Ngrx State Management flow: The angular component needs data for binding.  So angular component calls an action that is responsible for invoking the API call.  Aft

Angular 14 Crud Example

In this article, we will implement CRUD operation in the Angular 14 application. Angular: Angular is a framework that can be used to build a single-page application. Angular applications are built with components that make our code simple and clean. Angular components compose of 3 files like TypeScript File(*.ts), Html File(*.html), CSS File(*.cs) Components typescript file and HTML file support 2-way binding which means data flow is bi-directional Component typescript file listens for all HTML events from the HTML file. Create Angular(14) Application: Let's create an Angular(14) application to begin our sample. Make sure to install the Angular CLI tool into our local machine because it provides easy CLI commands to play with the angular application. Command To Install Angular CLI npm install -g @angular/cli Run the below command to create the angular application. Command To Create Angular Application ng new name_of_your_app Note: While creating the app, you will see a noti

Part-1 Angular JWT Authentication Using HTTP Only Cookie[Angular V13]

In this article, we are going to implement a sample angular application authentication using HTTP only cookie that contains a JWT token. HTTP Only JWT Cookie: In a SPA(Single Page Application) Authentication JWT token either can be stored in browser 'LocalStorage' or in 'Cookie'. Storing JWT token inside of the cookie then the cookie should be HTTP Only. The HTTP-Only cookie nature is that it will be only accessible by the server application. Client apps like javascript-based apps can't access the HTTP-Only cookie. So if we use authentication with HTTP only JWT cookie then we no need to implement custom logic like adding authorization header or storing token data, etc at our client application. Because once the user authenticated cookie will be automatically sent to the server by the browser on every API call. Authentication API: To implement JWT cookie authentication we need to set up an API. For that, I had created a mock authentication API(Using the NestJS Se

ReactJS(v18) Authentication With JWT AccessToken And Refresh Token

In this article, we are going to do ReactJS(v18) application authentication using the JWT Access Token and Refresh Token. JSON Web Token(JWT): JSON Web Token is a digitally signed and secured token for user validation. The JWT is constructed with 3 important parts: Header Payload Signature Create ReactJS Application: Let's create a ReactJS application to accomplish our demo. npx create-react-app name-of-your-app Configure React Bootstrap Library: Let's install the React Bootstrap library npm install react-bootstrap bootstrap Now add the bootstrap CSS reference in 'index.js'. src/index.js: import 'bootstrap/dist/css/bootstrap.min.css' Create A React Component 'Layout': Let's add a React component like 'Layout' in 'components/shared' folders(new folders). src/components/shared/Layout.js: import Navbar from "react-bootstrap/Navbar"; import { Container } from "react-bootstrap"; import Nav from "react-boot

A Small Guide On NestJS Queues

NestJS Application Queues helps to deal with application scaling and performance challenges. When To Use Queues?: API request that mostly involves in time taking operations like CPU bound operation, doing them synchronously which will result in thread blocking. So to avoid these issues, it is an appropriate way to make the CPU-bound operation separate background job.  In nestjs one of the best solutions for these kinds of tasks is to implement the Queues. For queueing mechanism in the nestjs application most recommended library is '@nestjs/bull'(Bull is nodejs queue library). The 'Bull' depends on Redis cache for data storage like a job. So in this queueing technique, we will create services like 'Producer' and 'Consumer'. The 'Producer' is used to push our jobs into the Redis stores. The consumer will read those jobs(eg: CPU Bound Operations) and process them. So by using this queues technique user requests processed very fastly because actually