Café Employee Manager — Backend

Node.js REST API for managing cafes and employees, built with Express and PostgreSQL.

---

Tech Stack

Category	Technology
Runtime	Node.js 22
Framework	Express.js
Database	PostgreSQL 14
Query Builder	Knex.js (migrations + queries)
Validation	Joi
File Uploads	Multer (disk storage)
Image Storage	Cloudinary
Dependency Injection	Awilix
Security	Helmet, CORS, express-rate-limit
Logging	Morgan
Containerisation	Docker

---

Architecture & Design Patterns

Mediator Pattern

Controllers do not call repositories directly. Every operation is encapsulated as a Command or Query and dispatched through a central Mediator, which routes it to the correct handler. This decouples the HTTP layer from business logic.

Controller → Mediator → Handler → Repository → Database

CQRS (Command Query Responsibility Segregation)

Read operations use Query objects (GetCafesQuery, GetEmployeesQuery) and write operations use Command objects (CreateCafeCommand, UpdateCafeCommand, etc.), keeping reads and writes clearly separated.

---

Security Features

Feature	Implementation
Security headers	helmet sets X-Content-Type, X-Frame-Options, HSTS, etc.
CORS	Whitelist-based origin validation via CORS_ORIGIN env var
Rate limiting	Global: 100 req/15 min · Write endpoints: 30 req/15 min
Input validation	Joi schemas on all POST/PUT request bodies, Multer enforces image-only MIME types and 2MB size cap
Error sanitisation	Raw database/stack errors are never exposed to clients, instead all errors return clean JSON messages

---

Database Design

cafes
├── id           UUID, primary key (auto-generated)
├── name         VARCHAR(255), not null
├── description  TEXT, not null
├── logo         VARCHAR(500) -- Cloudinary URL
├── location     VARCHAR(255), not null
├── created_at   TIMESTAMP
└── updated_at   TIMESTAMP

employees
├── id           VARCHAR(10), primary key -- format: UIXXXXXXX
├── name         VARCHAR(255), not null
├── email_address VARCHAR(255), not null, unique
├── phone_number VARCHAR(8), not null
├── gender       VARCHAR(10), not null
├── cafe_id      UUID, foreign key → cafes.id (CASCADE DELETE)
├── start_date   DATE, not null
├── created_at   TIMESTAMP
└── updated_at   TIMESTAMP

Relationships:

• One cafe has many employees (one-to-many)
• One employee can only work in one cafe
• Deleting a cafe automatically deletes all its employees (CASCADE)
• days_worked is derived at query time: current date − start_date

---

API Endpoints

Backend API URL

Cafes

Method	Endpoint	Description
GET	/cafes	List all cafes, sorted by employee count desc
GET	/cafes?location=<value>	Filter cafes by location
POST	/cafes	Create a new cafe
PUT	/cafes/:id	Update an existing cafe
DELETE	/cafes/:id	Delete a cafe and all its employees

Employees

Method	Endpoint	Description
GET	/employees	List all employees, sorted by days worked desc
GET	/employees?cafe=<name>	Filter employees by cafe name
POST	/employees	Create a new employee
PUT	/employees/:id	Update an employee
DELETE	/employees/:id	Delete an employee

Other

Method	Endpoint	Description
GET	/health	Health check

---

Local Setup

Prerequisites

• Node.js 22+
• PostgreSQL 14+

1. Clone and install dependencies

git clone <your-repo-url>
cd platform-backend
npm install

2. Create the database

CREATE DATABASE cafe_db;

3. Configure environment

Create a .env file in the root:

PORT=4000
NODE_ENV=development

DB_HOST=localhost
DB_PORT=5432
DB_NAME=cafe_db
DB_USER=postgres
DB_PASSWORD=your_password

CLOUDINARY_CLOUD_NAME=your_cloud_name
CLOUDINARY_API_KEY=your_api_key
CLOUDINARY_API_SECRET=your_api_secret

UPLOAD_DIR=uploads
MAX_FILE_SIZE=2097152
CORS_ORIGIN=http://localhost:3000

4. Run migrations and seed data

npm run migrate   # creates tables
npm run seed      # inserts sample data

5. Start the server

npm run dev    # development (nodemon, auto-restart)
npm start      # production

API runs at http://localhost:4000.

---

Docker (Full Stack)

Runs backend and PostgreSQL together:

cd platform-backend
docker-compose up --build

Migrations and seeds run automatically on startup.

To stop:

docker-compose down

To stop and wipe the database:

docker-compose down -v