Skip to content

A Spring Boot based implementation of a generic E-commerce Store

License

Notifications You must be signed in to change notification settings

prasadus92/ecommerce-store

Repository files navigation

E-commerce Application

Build Status codecov Maintainability

Problem Statement

Create a tiny REST/JSON web service in Java using Spring Boot (RestController) with an API that supports basic Products' CRUD:

  • Create a new Product

  • Get a list of all Products

  • Update a Product

The API should also support:

  • Placing an Order

  • Retrieving all orders within a given time period

A product should have a name and some representation of its price.

Each order should be recorded and have a list of products. It should also have the buyer’s e-mail, and the time the order was placed. The total value of the order should always be calculated, based on the prices of the products in it.

It should be possible to change the product’s price, but this shouldn’t affect the total value of orders which have already been placed.

Requirements

  • Implement your solution according to the above specification.

  • Provide unit tests.

  • Document your REST-API.

  • Provide a storage solution for persisting the web service’s state.

  • Have a way to run the service with its dependencies (database etc) locally. You can use either a simple script or docker or something else. It’s up to you.

Solution Overview

Tools Used

  • Java 8

  • Spring Boot v2.2.4.RELEASE

  • PostgreSQL v10

  • Maven

  • Docker

  • Travis (CI)

Building the Application

As the project uses Maven as the build system, just run the below command in the project directory -

mvn clean install

The name of the generated JAR is ecommerce-store-application.jar.

To build the Dockerfile as well (make sure that the Docker daemon is running) -

mvn clean install dockerfile:build

NOTE: Maven is configured with dockerfile-maven-plugin for building the Docker image

Running Tests

Run the below command -

mvn clean test

Running the Application

Using Maven

You could use Maven to run the Spring Boot application; here is the command to use -

mvn spring-boot:run

Using the generated artifact

Once you build the Application, the artifacts will be generated and copied to target directory. Hence, using the JAR, we could start the Application this way -

java -jar ecommerce-store-application.jar

Using Docker

If you have generated the Docker image in the build step, use the below command to run the Application on Docker -

docker-compose up

To tail the logs, just run -

docker-compose logs -f

NOTE: The Spring profile used while running with Docker is production. This requires PostgreSQL to be running, which is already configured in the Docker Compose file.

Accessing the API Documentation

Swagger is included in the project. Hence, the API documentation is available at http://localhost:8080/swagger-ui.html.

Questions on Extensibility

  1. You do not need to add authentication to your web service, but propose a protocol / method and justify your choice.

We could use OAuth2 and JWT based authentication which is quite straightforward to implement by adding appropriate Spring Security dependencies. Also, this is convenient as there might be multiple clients using this service in a real-world scenario (https://www.toptal.com/spring/spring-boot-oauth2-jwt-rest-protection). An alternate approach is to use an identity provider like Okta, Auth0 etc., if that's an allowed use-case.

  1. How can you make the service redundant? What considerations should you do?

The concurrency part is handled using Optimistic Locking, specifically for the update of the Products. It would be ideal to add additional mechanisms like an "ETag" - "If-Match" header support. Also, configuration (including passwords) is currently packaged along with the generated artifact; it would be better to use a separate Spring Cloud Config service in a scenario where the architecture is based on Microservices.

To Do

  1. Try finding an alternative to using wait-for-postgres.sh; it currently requires PostgreSQL client to be installed on the Docker container.
  2. Improve tests & coverage.
  3. Add pagination to "fetch" APIs.
  4. "User" is not included in the implemented model; that's something to be considered in a real-world scenario.