U.S. City Population API

A Rest API written in Node.JS. Said API is meant to serve population data by city for a given state in the United States. I wrote this API for a assignment for a company I am interviewing with. The backend for this API is meant to be light on dependecies in order to achieve the best performance and throughtput possible.

Objectives

In order for this assignment to be considered complete the following requirements for the project must be met:

• Must use Node.JS version 18
• Written in plain Javascript (no typescript)
• Service must run on port 5555
• Service data must persist
• The service must include the API endpoint /api/population/state/:state/city/:city that utilizes both the GET, and the PUT method for HTTP requests
• The API endpoint above must take in the City and State for both GET and PUT requests as URL parameters
• The State and City parameters should also not be case-sensitive
• For HTTP requests using the GET method a response with the 400 status code should be returned. Along with a appropriate error message if the City / State combination cannot be found
• For HTTP requests using the GET method a response with 200 status code should returned along with the population number for the City / State combo provided
• For HTTP requests using the PUT method the body should be provided as the population number in a plain text format for the City / State combination provided
• For HTTP requests using the PUT method a response with the status code 200 should be returned if the city population is updated for a City / State combo that already exists
• For HTTP requests using the PUT method a response with the status code 201 should be returned if the city population is created for a City / State combo that does not exist

Dependecies Used

For this project I opted not to use a web framework for implementing the backend for this REST API. For web server I am using the http module provided by Node.JS along with a handful of libraries that I would consider to be perfomant. The reason I decided to go down this path is because when compared to different Javascript web frameworks such as Fastify, Express.JS, and HAPI, the generic HTTP module that Node.JS provides still has the better performance metrics when compared to said web frameworks. Benchmarks from the Fastify benchmarks Repository, TechEmpower Web Framwork benchmarks, and the benchmarks performed by Mark Choubey support these claims.

This projects implementes the following libraries:

• find-my-way (HTTP router)
• dotenv (Environment variable loading)
• Ajv (Input validation)
• node-cache (Data caching)

Test Cases

Even though providing unit tests for this assignment was not required, I have been trying to practice Test-Driven Development whenever working on projects that use plain Javascript or Typescript. I felt like this project would provide a decent challenge for implementing TDD. The Unit tests are created using Mocha and Chai. Furthermore, I have also added the script command npm build to package.json in order to provide a convenient way of running unit tests included in this project.

The current unit tests test for two things in project code:

• Reading/Writing JSON files for data persistance
• HTTP responses for different HTTP requests sent to the /api/population/state/:state/city/:city endpoint

Items that could be imporoved

While I feel pretty good about how this project turned out, there are some things I would have done differently if I were to deploy this API for production. As this project is for a interview assignment and meant to light on dependecies there are things I did differently compared to how I would normally implement certain pieces of functionality in Node.JS. Some of these things include:

• Using Redis for caching instead of using node-cache
• Persist the City / State data in a proper SQL database instead of storing said data in a JSON file.
• Spent extra time to for finding a solution for implementing middleware in order to simplify input validation (find-my-way does not have a way of implementing route middleware beyond middleware for bad urls and the default route)
• Implement a queue for write operations to where ever data ends up being persisted in order to minimize execution blocks during said write operations.