This application is deployed to the Google Cloud Platform - Explore it here.
A full-stack project to search and rate movies using data from The Movie Database created as a take-home challenge for YearOne, a community I joined as a recent Front-End graduate of Turing School of Software and Design.
After solidifying my knowledge of front-end development over the last year, I viewed this challenge as an opportunity to solidify my full-stack skills, explore a new language, DevOps concepts and tools.
- Docker desktop minimum version 3.4.0
- GNU make - optional
In a new terminal window navigate to the parent directory you wish to clone the code into and run the following commands
- Clone code to local machine
# Downloads code to a new directory named yearone-rancid-tomatillos
git clone git@github.com:edmdc/yearone-rancid-tomatillos.git- Navigate into project directory and create a new directory for the Mongodb container volume
cd yearone-rancid-tomatillos
mkdir mongo
# The mongo directory will be used by Docker to persist database container data-
Make sure that Docker desktop is running in the background before running any of the following commands.
-
Build and start development containers as background process
# Web container with hot module reloading can be accessed on http://localhost:8080
# Api container with live reloading can be accessed on http://localhost:8081
make dev
# Docker command equivalent
docker compose up --build -d
# Use the docker command above without -d flag to output logs to current terminal window- End background processes
make end
# Docker command equivalent
docker compose down- Clear Docker build cache
make clean
# Docker command equivalent
docker builder prune
Docker, written in Go, is omnipresent when developing for the cloud. Its role cannot be understated, so learning the basics was a no-brainer choice and my main goal with this project.
In previous projects, I used a prebuilt Serverless Component to deploy a NextJS front-end to AWS. The simple configuration of this component is ideal for a simple, solo project. However, it presents some limitations as a project grows. It does not address the need for a consistent development environment across team members and into production. The added abstraction also complicates the creation of a build pipeline for automated deployments. I faced these issues building out a codebase during a fellowship at a young start-up where I was one of two front-end developers. While there, I took the responsibility of managing front-end deployments and quickly realized I lacked the knowledge to make it a frictionless process. Intent on filling this knowledge gap, I began to explore Docker.
Docker is an open-source project that simplifies development by packaging software into containers. Containers are a standardized unit of software based on the Linux kernel. They package application source code with operating system libraries and dependencies that allow the application to run the same in any environment. Docker did not invent containers, but they did help simplify their use, making it the industry standard for application containerization. Their comprehensive platform allows developers to speed up development, easily share code, and run applications securely.
Exploring Docker allowed me the flexibility to choose a cloud provider of my own. AWS provided a solid introduction to cloud technologies; however, I wanted to expand my knowledge by exploring a new provider. Both Google and Microsoft have done plenty to progress internet technologies and have made plenty of contributions to the open-source community. Given that their services are more or less equal, it became a question of company values. I chose GCP because its upfront about its sustainability goals and considers this commitment a feature worth mentioning.
NextJS is my preferred React implementation because it's production-ready out the box. Some of my favorite features include, but are not limited to, Typescript support, image optimization, hybrid Static Site Generation / Server Side Rendering, and File-system Routing.
I opted to use Emotion CSS-in-JS for its built-in theming API that allows for the straightforward implementation of a style system. The encapsulation of styles to individual components makes them easy to maintain. I've used TailwindCSS in the past, and though I like the concept of utility classnames, I have found that the long classname strings make code challenging to format, read, and maintain.
I chose to learn Go because great minds in computer science, like Ken Thompson, Rob Pike,
and Robert Griesemer, designed it for modern, networked machines. Go is easy to learn and maintain
because it has simple, idiomatic solutions to common problems creating consistent use cases.
It's statically typed and has built-in features that simplify development over Node.js. For example,
the compiler doubles as a code linter and go fmt formats project code, eliminating the need to
choose and configure packages like ESlint and prettier.
MongoDB is a document database that uses JSON to describe data rather than tables (NoSQL). JSON is an intuitive way to store and retrieve complex data that allows for a flexible schema. The flexibility and simplicity of storing data in MongoDB make it my go-to database choice for personal projects.
I'm a big fan of GraphQL as the new paradigm of API development. I've used the Apollo GraphQL platform in multiple projects and consider it one of my more advanced skills. I like that the client-side implementation has built-in state management that's intuitive to use. It's also helpful to have a single endpoint for back-end services as a project grows. However, this project has a limited amount of endpoints and is straightforward enough to accomplish without it. Choosing to omit GraphQL also allowed me to focus on learning the basics of Go without the added complexity on the server.