E-Commerce API with Smart Recommendations

A backend REST API using MongoDB Atlas, Neo4j, and Redis Cloud running in Docker.

---

Prerequisites

• Docker Desktop installed and running
• MongoDB Atlas account
• Neo4j AuraDB account
• Redis Cloud account

---

Setup

1. Clone the repo

git clone https://github.com/yourteam/ecommerce-api.git
cd ecommerce-api

2. Configure environment

cp .env.example .env

Open .env and fill in your credentials:

PORT=3000
MONGO_URI=mongodb+srv://<user>:<password>@cluster.mongodb.net/ecommerce?retryWrites=true&w=majority
NEO4J_URI=neo4j+s://<your-instance>.databases.neo4j.io
NEO4J_USER=<your-username>
NEO4J_PASSWORD=<your-password>
REDIS_URL=redis://:<password>@<host>:<port>

3. Seed MongoDB

Open MongoDB Compass → connect to your Atlas cluster → open the MongoShell tab at the bottom and paste the contents of scripts/seed.mongo.js, then press Enter.

4. Generate product embeddings

npm install
node ./scripts/generateEmbeddings.js

5. Seed Neo4j and Redis

node ./scripts/seed.js

6. Start the API

docker compose up --build

API runs at http://localhost:3000

---

Endpoints

Method	Endpoint	Description
GET	/products/:id	Get product (Redis cache then MongoDB)
GET	/products/top-selling	Top 5 products by units sold
GET	/products/revenue-by-category	Revenue grouped by category
GET	/home	Home page data, cached 5 mins in Redis
POST	/orders	Place an order
GET	/recommendations/:userId	Neo4j graph recommendations
GET	/similar/:productId	Similar products by vector search
POST	/similar/by-image	Similar products from a text description
GET	/cart/:userId	Get cart
POST	/cart/:userId/add	Add item to cart
DELETE	/cart/:userId	Clear cart

---

Example Requests

Place an order

POST /orders
{
  "userId": "user1",
  "items": [
    { "productId": "<productId>", "quantity": 1, "price": 79.99 }
  ]
}

Add to cart

POST /cart/user1/add
{
  "productId": "<productId>",
  "quantity": 2,
  "price": 39.99
}

Search by image description

POST /similar/by-image
{
  "description": "wireless audio device for music"
}

---

Stop the API

docker compose down