NodeJS Ecommerce API

An E-commerce API

Technology Stack

This API is built on NodeJS and MySQL. Some of the modules used include:

1. Express JS for handling HTTP requests.
2. MySQL database for storing application data.
3. Sequelize.js ORM for mapping database tables to objects.
4. Passport and jsonwebtoken for API Authentication and token generation.
5. Helmet for security.
6. Dotenv to handle configurations.
7. ApiCache to handle caching

Installation

Follow the following steps to install

1. Clone the repo
2. Move to project directory
3. Run the following to install dependences

   npm install
   

4. Set up database. Modified sql file can be found here. You need to copy the file contents unto a MySQL console.
5. Rename the .env-example file to .env and fill in the correct settings and credentials
6. Run the following code to start up the app

   npm run start:dev
   

Tests

You can test the app as a whole or test the various modules seperately.

Note You need to have mocha installed globally to run tests. run the following command to install mocha globally:

npm install -g mocha

To test the whole app run the following command inside the project directory.

npm run test

To test the various modules seperately, run one of the following commands

npm run test:attributes
npm run test:categories
npm run test:customers
npm run test:departments
npm run test:orders
npm run test:products
npm run test:shipping
npm run test:shopping_cart
npm run test:taxes

In order to test the email feature, the email address in test/order.js can be changed to the email the notification should sent to.

API Documentation

The API Documentation can be found here.

Change Log

The following changes were made during the course of development

1. Database changes: I changed the length of the password column in the customer table from 50 to 60 in order to accommodate password hashes.

I also modified some stored procedures. The modified procedures include

i.   shopping_cart_add_product

ii.  shopping_cart_update

iii. shopping_cart_get_products

2. Error message changes: Some error cases (such as invalid phone number, invalid credit card number, etc) were not accounted for in the error document; I created them.

3. Stripe key changes: In the backend challenge document, the Stripe secret key was provided for use, however the publishable key was not provided. Although the pub-key is not required to complete the task, it is required to generate a matching token. So I used the test secret key and pub-key provided on Stripe website to complete the task.

However, this can be set in the config (.env) file.

4. Authentication header key: The API challenge document says to use USER-KEY as the header while the Swagger documentation uses API-KEY. I used USER-KEY in this application.

Advanced Requirements

App Scaling

The following steps can be taken to make the ap scale better.

1. Implement a more robust caching mechanism using Redis. To add Redis to the current cache mechanism, the following adaptations can be made to the code:

   const express = require('express')
   const redis = require('redis')
   const apicache = require('apicache')
   
   const app = express()
   
   const redisCache = apicache.options({
       redisClient: redis.createClient()
   }).middleware
   
   //...
   
   app.get('/route-to-cache', redisCache('1 hour'), (req, res) => {
       res.send({ message: 'I am cached' })
   })
   
   //...
   
   

2. Using a connection pool in MySQL. To support connection pool in the application, the following adaptations can be made to the code:

   const sequelize = new Sequelize(/* database credentials */, {
       pool: {
           max: 7,
           min: 0,
           acquire: 30000,
           idle: 10000
       }
   })
   
   //...
   

3. To further boost performance, configure enable parallel processing using the cluster module. The following code adaptations can be made to configure parallel processing. i. Create a file named cluster.js in the root folder and put the following code in the

   const cluster = require('cluster')
   const os = require('os)
   const CPUS = os.cpus()
   
   if (cluster.isMaster) {
       CPUS.forEach(() => {
           cluster.fork()
       })
       cluster.on('listening', (worker) => {
           console.log("Cluster %d connected", worker.process.pid)
       })
       cluster.on('disconnect', (worker) => {
           console.log("Cluster %d connected", worker.process.pid)
       })
       cluster.on('exit', (worker) => {
           console.log("Cluster %d is dead", worker.process.pid)
           //Ensure a new cluster starts when an old one dies
           cluster.fork()
       })
   } else {
       require('./www')
   }
   

   ii. Modify the package.json as follows: start:dev": "nodemon ./cluster.js

4. Aside from code adaptations, the server infrastructures can be scaled.