API with NestJS #88. Testing a project with raw SQL using integration tests

December 26, 2022

This entry is part 88 of 148 in the API with NestJS

Covering our project with tests can help us ensure that our application works as expected and is reliable. While unit tests play a significant role, they are not enough. Therefore, this article explains the significance of integration tests and implements them in a NestJS project that uses raw SQL queries.

The significance of integration tests

The role of a unit test is to cover an individual piece of code and verify if it works as expected. For example, a particular unit test suite can test a single function or a class.

When all our unit tests pass, it means that each part of our system works well on its own. However, that does not yet mean that all parts of the application interact with each other correctly. To verify the above, we need integration tests. Their job is to ensure that two or more pieces of our system integrate correctly.

Testing NestJS services

Writing integration tests does not mean we don’t mock any part of the application. For example, in our integration tests, we won’t be using an actual database.

Tests that verify the application from start to finish are referred to as end-to-end (E2E) tests. They should mimic a real system as close as possible.

posts.service.test.ts

import PostDto from './post.dto';

import { Test } from '@nestjs/testing';

import DatabaseService from '../database/database.service';

import { PostsService } from './posts.service';

import PostWithCategoryIdsModel, {

PostWithCategoryIdsModelData,

} from './postWithCategoryIds.model';

import PostsRepository from './posts.repository';

import PostsStatisticsRepository from './postsStatistics.repository';

import PostsSearchRepository from './postsSearch.repository';

describe('The PostsService', () => {

describe('when calling the create method with category ids', () => {

it('should return an instance of the PostWithCategoryIdsModel', async () => {

const postData: PostDto = {

title: 'Hello world!',

content: 'Lorem ipsum',

categoryIds: [1, 2, 3],

};

const sqlQueryResult: PostWithCategoryIdsModelData = {

id: 1,

author_id: 2,

title: postData.title,

post_content: postData.content,

category_ids: postData.categoryIds,

};

const runQueryMock = jest.fn();

runQueryMock.mockResolvedValue({

rows: [sqlQueryResult],

});

const module = await Test.createTestingModule({

providers: [

PostsService,

PostsRepository,

PostsStatisticsRepository,

PostsSearchRepository,

{

provide: DatabaseService,

useValue: {

runQuery: runQueryMock,

}).compile();

const postsService = await module.get(PostsService);

const result = await postsService.createPost(postData, 1);

expect(result instanceof PostWithCategoryIdsModel).toBe(true);

});

The PostsService
when calling the create method with category ids
✓ should return an instance of the PostWithCategoryIdsModel

Above, we test if the PostsService class works well with the PostsRepository.

posts.service.ts

import { Injectable } from '@nestjs/common';

import PostsRepository from './posts.repository';

import PostDto from './post.dto';

@Injectable()

export class PostsService {

constructor(

private readonly postsRepository: PostsRepository,

) {}

createPost(postData: PostDto, authorId: number) {

if (postData.categoryIds?.length) {

return this.postsRepository.createWithCategories(postData, authorId);

}

return this.postsRepository.create(postData, authorId);

}

// ...

}

In our test, we only mock the DatabaseService. Thanks to that, we test if the PostService integrates correctly with the PostsRepository.

posts.repository.ts

import {

BadRequestException,

Injectable,

} from '@nestjs/common';

import DatabaseService from '../database/database.service';

import PostModel from './post.model';

import PostDto from './post.dto';

import PostWithCategoryIdsModel from './postWithCategoryIds.model';

import PostgresErrorCode from '../database/postgresErrorCode.enum';

import { isDatabaseError } from '../types/databaseError';

@Injectable()

class PostsRepository {

constructor(private readonly databaseService: DatabaseService) {}

async create(postData: PostDto, authorId: number) {

try {

const databaseResponse = await this.databaseService.runQuery(

// ...

[postData.title, postData.content, authorId],

);

return new PostModel(databaseResponse.rows[0]);

} catch (error) {

if (

!isDatabaseError(error) ||

!['title', 'post_content'].includes(error.column)

) {

throw error;

}

if (error.code === PostgresErrorCode.NotNullViolation) {

throw new BadRequestException(

`A null value can't be set for the ${error.column} column`,

);

}

throw error;

}

async createWithCategories(postData: PostDto, authorId: number) {

const databaseResponse = await this.databaseService.runQuery(

// ...

[postData.title, postData.content, authorId, postData.categoryIds],

);

return new PostWithCategoryIdsModel(databaseResponse.rows[0]);

}

// ...

}

export default PostsRepository;

Mocking differently in each test

So far, in our test, we put all of our logic in a single it function. An alternative for that is moving the setup code into the beforeEach functions.

posts.service.test.ts

import PostDto from './post.dto';

import { Test } from '@nestjs/testing';

import DatabaseService from '../database/database.service';

import { PostsService } from './posts.service';

import PostWithCategoryIdsModel, { PostWithCategoryIdsModelData } from "./postWithCategoryIds.model";

import PostsRepository from './posts.repository';

import PostsStatisticsRepository from './postsStatistics.repository';

import PostsSearchRepository from './postsSearch.repository';

describe('The PostsService', () => {

let postData: PostDto;

let runQueryMock: jest.Mock;

let postsService: PostsService;

beforeEach(async () => {

runQueryMock = jest.fn();

const module = await Test.createTestingModule({

providers: [

PostsService,

PostsRepository,

PostsStatisticsRepository,

PostsSearchRepository,

{

provide: DatabaseService,

useValue: {

runQuery: runQueryMock,

}).compile();

postsService = await module.get(PostsService);

});

describe('when calling the create method with category ids', () => {

let sqlQueryResult: PostWithCategoryIdsModelData;

beforeEach(() => {

postData = {

title: 'Hello world!',

content: 'Lorem ipsum',

categoryIds: [1, 2, 3],

};

sqlQueryResult = {

id: 1,

author_id: 2,

title: postData.title,

post_content: postData.content,

category_ids: postData.categoryIds

}

runQueryMock.mockResolvedValue({

rows: [sqlQueryResult],

});

it('should return an instance of the PostWithCategoryIdsModel', async () => {

const result = await postsService.createPost(postData, 1);

expect(result instanceof PostWithCategoryIdsModel).toBe(true);

});

it('should return an object with the correct properties', async () => {

const result = await postsService.createPost(postData, 1) as PostWithCategoryIdsModel;

expect(result.id).toBe(sqlQueryResult.id);

expect(result.authorId).toBe(sqlQueryResult.author_id);

expect(result.title).toBe(sqlQueryResult.title);

expect(result.content).toBe(sqlQueryResult.post_content);

expect(result.categoryIds).toBe(sqlQueryResult.category_ids);

})

});

The PostsService
when calling the create method with category ids
✓ should return an instance of the PostWithCategoryIdsModel
✓ should return an object with the correct properties

Thanks to using beforEach, we can set up multiple tests using the same piece of code and avoid repeating it. We can take it further and change how we mock the DatabaseService at each test.

posts.service.test.ts

100

101

102

103

104

105

import PostDto from './post.dto';

import { Test } from '@nestjs/testing';

import DatabaseService from '../database/database.service';

import { PostsService } from './posts.service';

import PostWithCategoryIdsModel, {

PostWithCategoryIdsModelData,

} from './postWithCategoryIds.model';

import PostsRepository from './posts.repository';

import PostsStatisticsRepository from './postsStatistics.repository';

import PostsSearchRepository from './postsSearch.repository';

import PostModel, { PostModelData } from './post.model';

describe('The PostsService', () => {

let postData: PostDto;

let runQueryMock: jest.Mock;

let postsService: PostsService;

beforeEach(async () => {

runQueryMock = jest.fn();

const module = await Test.createTestingModule({

providers: [

PostsService,

PostsRepository,

PostsStatisticsRepository,

PostsSearchRepository,

{

provide: DatabaseService,

useValue: {

runQuery: runQueryMock,

}).compile();

postsService = await module.get(PostsService);

});

describe('when calling the create method with category ids', () => {

let sqlQueryResult: PostWithCategoryIdsModelData;

beforeEach(() => {

postData = {

title: 'Hello world!',

content: 'Lorem ipsum',

categoryIds: [1, 2, 3],

};

sqlQueryResult = {

id: 1,

author_id: 2,

title: postData.title,

post_content: postData.content,

category_ids: postData.categoryIds,

};

runQueryMock.mockResolvedValue({

rows: [sqlQueryResult],

});

it('should return an instance of the PostWithCategoryIdsModel', async () => {

const result = await postsService.createPost(postData, 1);

expect(result instanceof PostWithCategoryIdsModel).toBe(true);

});

it('should return an object with the correct properties', async () => {

const result = (await postsService.createPost(

postData,

)) as PostWithCategoryIdsModel;

expect(result.id).toBe(sqlQueryResult.id);

expect(result.authorId).toBe(sqlQueryResult.author_id);

expect(result.title).toBe(sqlQueryResult.title);

expect(result.content).toBe(sqlQueryResult.post_content);

expect(result.categoryIds).toBe(sqlQueryResult.category_ids);

});

describe('when calling the create method without category ids', () => {

let sqlQueryResult: PostModelData;

beforeEach(() => {

postData = {

title: 'Hello world!',

content: 'Lorem ipsum',

};

sqlQueryResult = {

id: 1,

author_id: 2,

title: postData.title,

post_content: postData.content,

};

runQueryMock.mockResolvedValue({

rows: [sqlQueryResult],

});

it('should return an instance of the PostModel', async () => {

const result = await postsService.createPost(postData, 1);

expect(result instanceof PostModel).toBe(true);

});

it('should return an object with the correct properties', async () => {

const result = await postsService.createPost(postData, 1);

expect(result.id).toBe(sqlQueryResult.id);

expect(result.authorId).toBe(sqlQueryResult.author_id);

expect(result.title).toBe(sqlQueryResult.title);

expect(result.content).toBe(sqlQueryResult.post_content);

});

The PostsService
when calling the create method with category ids
✓ should return an instance of the PostWithCategoryIdsModel
✓ should return an object with the correct properties
when calling the create method without category ids
✓ should return an instance of the PostMode
✓ should return an object with the correct properties

Testing controllers

Another approach to integration tests is to include our controllers by performing HTTP requests. By doing that, we can more closely mimic how our application works in a real environment.

1	npm install supertest @types/supertest

Since our application uses the ValidationPipe, we need to add it explicitly.

100

import { Test } from '@nestjs/testing';

import { AuthenticationController } from './authentication.controller';

import { AuthenticationService } from './authentication.service';

import DatabaseService from '../database/database.service';

import { INestApplication, ValidationPipe } from '@nestjs/common';

import * as request from 'supertest';

import UsersService from '../users/users.service';

import { ConfigModule } from '@nestjs/config';

import { JwtModule } from '@nestjs/jwt';

import UsersRepository from '../users/users.repository';

import RegisterDto from './dto/register.dto';

import { UserModelData } from '../users/user.model';

describe('The AuthenticationController', () => {

let runQueryMock: jest.Mock;

let app: INestApplication;

beforeEach(async () => {

runQueryMock = jest.fn();

const module = await Test.createTestingModule({

imports: [

ConfigModule.forRoot(),

JwtModule.register({

secretOrPrivateKey: 'Secret key',

}),

controllers: [AuthenticationController],

providers: [

AuthenticationService,

UsersRepository,

UsersService,

{

provide: DatabaseService,

useValue: {

runQuery: runQueryMock,

}).compile();

app = module.createNestApplication();

app.useGlobalPipes(new ValidationPipe());

await app.init();

});

describe('when making the /register POST request', () => {

describe('and using an incorrect email', () => {

it('should throw an error', () => {

return request(app.getHttpServer())

.post('/authentication/register')

.send({

email: 'not-an-email',

password: 'strongPassword',

})

.expect(400);

});

describe('and using the correct data', () => {

let registrationData: RegisterDto;

let userModelData: UserModelData;

beforeEach(() => {

registrationData = {

email: 'john@smith.com',

password: 'strongPassword',

};

userModelData = {

id: 1,

email: registrationData.email,

password: registrationData.password,

address_id: null,

address_country: null,

address_city: null,

address_street: null,

};

runQueryMock.mockResolvedValue({

rows: [userModelData],

});

it('should result with the 201 status', () => {

return request(app.getHttpServer())

.post('/authentication/register')

.send(registrationData)

.expect(201);

});

it('should respond with the data without the password', () => {

return request(app.getHttpServer())

.post('/authentication/register')

.send(registrationData)

.expect({

id: userModelData.id,

email: userModelData.email,

});

The AuthenticationController
when making the /register POST request
and using an incorrect email
✓ should throw an error
and using the correct data
✓ should result with the 201 status
✓ should respond with the data without the password

The SuperTest library is quite powerful and allows us to verify the response differently. For example, we can ensure the response headers are correct. For a complete list of the features, check out the documentation.

Summary

In this article, we’ve gone through the idea of integration tests. We’ve described why they’re important and how they can benefit our application. We’ve also implemented integration tests with two different approaches. One of them required us to install the supertest library to simulate actual HTTP requests. All of the above allowed us to test our API more thoroughly and increase the confidence that we’re creating a reliable product.

Series Navigation<< API with NestJS #87. Writing unit tests in a project with raw SQLAPI with NestJS #89. Replacing Express with Fastify >>