marvel-xendit

README

Marvel Characters

Just a simple tool that fetches all characters from official marvel api. Developed with love by yours truly.

Requirements

• GNU Make (optional)
• Either:
  • Go 1.15
• Or:
  • Docker
  • Docker Compose

API Documentation

Our API is documented using Swagger OpenAPI. The documentation is served within the app. The root (/) route shall redirect you to the API documentation and teach you how to consume this endpoint.

Building and running the application

Clone our repository:

# Clones our source code into marvel folder
git clone https://github.com/ghabxph/marvel-xendit.git marvel

First, you need to be a marvel developer and get your app key and secret. Then create a config.yaml file in this directory.

# Marvel App Key
public_key: "<your-marvel-app-key>"

# Marvel Secret Key
private_key: "<your-marvel-secret-key"

You got two options: Use docker or to build the application by yourself. We utilize GNU Make to make things very convenient for you. Note that before doing the next steps below, be sure that config.yaml file is present or things will simply fall apart.

With Docker

Simply run make dock-build dock-up to build and run the docker container. If you don't have GNU Make installed on your machine, then simply execute the commands manually by yourself.

# Builds the docker container
docker-compose build

# Runs the docker container
docker-compose up

Without Docker

Simply run make clean build run to create a clean build and run your application at the same time. If you don't have GNU Make installed on your machine, then simply execute the commands manually by yourself.

# Removes existing bin folder
rm -fvr ./bin

# Builds the application
go build -o bin/marvel ./cmd/marvel

# Runs the application
./bin/marvel

(Shortcut: make clean run does the same thing, but I want to show you the full process)

Running the test

Simply run make test to run the test for the four important parts of our program, or if you don't have GNU Make installed on your machine, then you can run the test for yourself.

# Tests the HTTP endpoint
go test ./internal/gateway

# Tests our main business logic
go test ./internal/marvel

# Tests our in-memory cache
go test ./internal/memorydb

# Tests our scraper
go test ./internal/scraper

How things work

This is a one-binary application. The architechture of this design is defined in the block diagram below.

The application is composed of 4 major components: gateway, marvel, memorydb, and scraper. We have two main caching strategies: Cache-aside and scraping data in advanced behind the scenes.

Cache-aside Strategy

Our cache-aside strategy is triggered by characters/{id}. When client requests for specific character with given ID, marvel will check the data first on memorydb if there's a hit. If it's a miss, then marvel will request for live data on official marvel endpoint and then caches the data and serves the fresh data to user. And then if user requests the same item again, marvel will serve the item from the cache / memorydb.

Scraping

We implemented a scraper that runs on the background everyday same time when the program is executed. We implemented this in anticipation that Marvel will update the information about their characters from time to time, so it is rational to implement this.

Keeping things persistent

In an event where server might crash for some various reason, or the application was intentionally terminated, we implemented a persistence by creating json files in ./.characters folder in the current working directory. Persistence happens every time there's a call in memorydb.GetInstance().CreateCharacter(...) function. We read data from the persistence just once upon start of the program.

Application's flow of execution

1. Loads persistence data to MemoryDB
2. Starts the scraper in the background (go routine)
3. Serves the HTTP Endpoint (Gateway)

Contribution Guide

Feel free to fork this application and make your modifications. If you're generous to share it to us, feel free to open a ticket and send us a pull request.