Gila Software Test

This repository holds the technical test for Gila Software, based on specifications provided on this [file](./docs/Coding Challenge NT.pdf).

TL;DR;

Run the project with docker-compose up (Needs Docker and docker-compose installed)

Open http://localhost:3000 to use the application

Design Decisions

Backend

For the backend, I chose Nest JS framework to bootstrap the application for the following reasons:

• Fast
• Opinionated
• Works great with serverless and dedicated server architectures
• Out-of-the-box modularity

TypeORM was chosen as the ORM since it provides seamless connection, fast development time, and well-typed entities. It also provides migrations out-of-the-box (which were not used since Docker handled the database code) but is a nice feature if the project gets big.

Class interfaces and the Strategy pattern were used to support changes in the system, allowing for new notification services to be implemented fast without much maintenance.

Try yarn test or npm test to run unit tests.
Try yarn test:e2e or npm run test:e2e to run e2e tests.

Frontend

For the frontend, I chose ReactJS for simplicity. It is a well-known frontend framework used extensively on production in many applications. It manages state well, is easy to use, and is easy to maintain.

I'm using a couple of libraries to manage common needs; they are described as follows:

• Ant Design
  • Component library. A lot of well-styled resources working out of the box. It provides some flexibility while making you blazing fast on UI design. Also really easy to make things responsive.
• React-Redux Toolkit
  • State management library. The ability to manage states in a centralized way, provides easy connection to external APIs while also taking care of retry policies, caching, and a friendly interface for loading states and error handling.
• Typescript
  • Not much to say here. I like the safeguards that the type system provides. Sure, it can take more time to develop, so it's a design decision, nothing wrong with plain Javascript.

Try yarn test or npm test to run unit tests.

Database

PostgreSQL was chosen as the database for this application for the following reasons:

• Fast to develop
• Information was coupled enough to justify the SQL approach
• Plays well with TypeORM
• Easy to setup on Docker
• Well used in production on highly scalable applications
• Syntax could be reused in case of migrating to AWS Aurora, CockroachDB, or some similar services on the cloud

Infrastructure

Docker was used to glue all of this together. There is not much to say here other than that it makes the application easy to run in this case.

Screens

Home

Logs

Testing Scenarios

Automated tests

There are a couple of tests (unit and integration) on the frontend and the backend just to show that I had contact with those tools, but I didn't focused much on that.

Manual testing

The application has three notification categories: ['Sports', 'Finance', 'Movies'] You must select a category, fill in the notification content in the text area, and hit send.

The application also has 3 test users with the following configuration:

• Test User 1
  • Subscribed to categories: ['Sports']
  • Subscribed to channels: ['SMS']
• Test User 2
  • Subscribed to categories: ['Finance', 'Movies']
  • Subscribed to channels: ['E-Mail', 'Push Notification']
• Test User 3
  • Subscribed to categories: ['Sports', 'Finance', 'Movies']
  • Subscribed to channels: ['SMS', 'E-Mail', 'Push Notification']