DEV Community

Cover image for Create a multi tenant application in Nest.js Part 2 (database setup using Sequelize & mongoose)
ismaeil-shajar
ismaeil-shajar

Posted on • Edited on

Create a multi tenant application in Nest.js Part 2 (database setup using Sequelize & mongoose)

In part 1, we setup nestjs framework, configured and tested the microservice application using nest.js.

Database

Nest gives us all the tools to work with any SQL and NoSQL database. You have a lot of options, you can also use almost all ORM and libraries in nodejs and typescript, like Sequelize, TypeORM, Prisma, and of course mongoose.

In this application, we will work with MySQL and MongoDB. We will also use the most popular js libraries; Sequelize as ORM for MySQL, and mongoose for MongoDB.

Database integration

Sequelize

To start using sequelize; we first need to install the required dependencies which include @nestjs/sequelize, mysql2 because we will connect to MySQL database and other needed dependencies.

$ npm install --save @nestjs/sequelize sequelize sequelize-typescript mysql2
$ npm install --save-dev @types/sequelize
Enter fullscreen mode Exit fullscreen mode

In the services, we will import SequelizeModule in the main modules to set connection configuration:

ex: app.module.ts

@Module({
  imports: [
    SequelizeModule.forRoot({
      dialect: 'mysql',
      host: 'localhost',
      port: 3306,
      username: 'root',
      password: 'root',
      database: 'test',
      models: [],
    }),
  ],
})
Enter fullscreen mode Exit fullscreen mode

The forRoot() method will include all the configuration properties. You can read more details here.

After configuring the connection, we need to create a table entity.
For example, we can set a user model in the user service (will add the connection in the service too) by creating user.model.ts which will look like:

user.model.ts

/// imports
@Table({tableName:'Users'})
export class Users extends Model<Users> {
    @Column( {allowNull: false })
    firstName: string;

    @Column( {allowNull: false })
    lastName: string;

    @Column( { allowNull: false,unique: true })
    email: string;

    @Column( {allowNull: false})
    password: string;    

    @Column( { allowNull: false})
    type: string;
}
Enter fullscreen mode Exit fullscreen mode

We should also add the dto:

create-user-dto.ts

export class CreateUserDto{
    readonly firstName:string
    readonly lastName:string
   readonly   email:string
   readonly password:string
   readonly type:string
}
Enter fullscreen mode Exit fullscreen mode

And don't forget to add Users in models array in forRoot()

Now let's complete the setup and configuration.
If you don't have a database you need to create an empty table and change Sequelize configuration by adding: autoLoadModels: true,
synchronize: true
.
Then in the module, you will add the repository by adding
SequelizeModule.forFeature([Users]) in the imports array.
In our case, we use the main module so it will be:

user-service.module.ts

@Module({
  imports: [SequelizeModule.forRoot({
    dialect: 'mysql',
    host: 'localhost',
    port: 3306,
    username: 'ismaeil',
    password: 'root',
    database: 'test',
    autoLoadModels: true,
    synchronize: true,
    models: [Users],
  }),SequelizeModule.forFeature([Users])],
  controllers: [UserServiceController],
  providers: [UserServiceService],
})
Enter fullscreen mode Exit fullscreen mode

And we will edit main service to add findall and create method:

user-service.service.ts

@Injectable()
export class UserServiceService {
  constructor(
    @InjectModel(Users)
  private readonly userModel: typeof Users){}
  async findAll(): Promise<Users[]> {
    return this.userModel.findAll() ;
  }

  async create( createUserDto:CreateUserDto):Promise<Users> {
    return this.userModel.create(<Users>createUserDto)
  }
}
Enter fullscreen mode Exit fullscreen mode

Lastly, edit the controller to enable the use of REST requests to access and edit the database:

user-service.controller.ts

@Controller('users')
export class UserServiceController {
  constructor(private readonly userServiceService: UserServiceService) {}

  @Get()
  async findAll(){
      return this.userServiceService.findAll();
  }

  @Post()
  async createUser(@Body() createUserDto:CreateUserDto){
    return  this.userServiceService.create(createUserDto)
  }

}
Enter fullscreen mode Exit fullscreen mode

Now run the browser and test http://127.0.0.1:3003/users. This should access the database and create a table for the first time and return an empty array.

We can add data using a POST request:
Image description

Tips

If we have an existing database and need to import tables with types without a lot of work we can use sequelize-typescript-generator.

You can search about it to see how it works. But here are some simple steps:

1- Download and install npx

2- Create a folder to save output typescript models mkdir models

3- Install sequelize-typescript-generator in your machine: npm install -g sequelize-typescript-generator

4- Install mysql driver: npm install -g mysql2

5- Run the command: npx stg -D mysql -h localhost -p 3306 -d <databaseName> -u <username> -x <password> --indices --case camel --out-dir models --clean

Source code available in git branch database-connection

mongoose

Just like the previous one, we need to install dependencies to use MongoDB in Nest:

$ npm install --save @nestjs/mongoose mongoose
Enter fullscreen mode Exit fullscreen mode

Import MongooseModule into the root Module

@Module({
  imports: [MongooseModule.forRoot('mongodb://localhost:27017/test')],
})
Enter fullscreen mode Exit fullscreen mode

forRoot() accepts the same configuration as mongoose.connect() from the Mongoose package.

We will use the MongoDB database in the notification service. First, we will add forRoot() in the root module and will create a child module called a message to serve notification messages.

The root module will look like this:

notification.module.ts


@Module({
  imports: [MongooseModule.forRoot('mongodb://localhost:27017/test'),
  MessageModule],
  controllers: [NotificationController],
  providers: [NotificationService],
})
Enter fullscreen mode Exit fullscreen mode

The message module files will be as follows:

Image description

Because we are using mongoose, we need to create a schema and after that import the repository in a module.

In src/message/schemes we will create message.schema.ts file which will look like this:

export type MessageSchemaDocument = MessageSchema & Document;

@Schema()
export class MessageSchema{
    @Prop()
    name: string    

    @Prop()
    createdAt: Date

    @Prop({type:mongoose.Schema.Types.Mixed})
    data: Record<string, any>
}

export const MessageSchemaSchema = SchemaFactory.createForClass(MessageSchema);
Enter fullscreen mode Exit fullscreen mode

Put the following code in message.module:

message.module.ts

@Module({
  imports: [MongooseModule.forFeature([{name:MessageSchema.name,schema:MessageSchemaSchema}])],
  controllers: [MessageController],
  providers: [MessageService],
})
Enter fullscreen mode Exit fullscreen mode

And put the following methods in the message service:

message.service.ts

@Injectable()
export class MessageService {
    constructor(@InjectModel(MessageSchema.name) private readonly messageModel: Model<MessageSchemaDocument>) {}
    async findAll () {
        return await this.messageModel.find().exec()
    }    
    async create (messageDto:MessageDto) {
        return await this.messageModel.create(messageDto)
    }
}
Enter fullscreen mode Exit fullscreen mode

Create MessageDto:

export class MessageDto {
    readonly name: string    
    readonly createdAt:Date = new Date();
    readonly data?: any
}
Enter fullscreen mode Exit fullscreen mode

For Request mapping:

message.controller.ts

@Controller('message')
export class MessageController {
  constructor(private readonly messagenService: MessageService) {}

  @Get()
  async findAll(){
    return this.messagenService.findAll();
  }

  @Post()
  @UsePipes(new ValidationPipe({ transform: true }))
  async create(@Body() messageDto:MessageDto){
    return this.messagenService.create(messageDto);
  }
}
Enter fullscreen mode Exit fullscreen mode

*Note: Pipes are used in transforming and validating input data, in our case we can use @UsePipes(new ValidationPipe({ transform: true })) to set the empty properties in the Dto with default values. For more details refer to Pipes and validation.

Now you can test using a Post request to the URL http://127.0.0.1:3002/message with body:

    {
        "name":"validation",
        "data":{"message":"testing validation message if it success","status":"valid"}
    }
Enter fullscreen mode Exit fullscreen mode

To retrieve all the records, use the Get request http://127.0.0.1:3002/message

Source code available in git branch mongodb-connection

In part 3, we will complete the database setup to use multiple databases depending on the request header.

Top comments (0)