victor-rocha / TL-Pokedex

A "True" Pokedex using an API "Layer"

TL-Pokedex

A "True" Pokedex using an API "Layer" 🙃

Prerequisites

• .net5

Optional

• Docker desktop

How to run

Clone the repo

First of all, let's clone the repo. You can do it via Git or simply downloading the source code directly from GitHub

A few ways of running this service

1. Visual Studio / Rider

• Just simply navigate to TL-Pokedex and open Pokedex.sln file.
• Hit 'Run' or 'Debug'
• Then please navigate to https://localhost:5001/swagger/index.html to use the API via Swagger

2. Via dotnet cli

• Navigate to TL-Pokedex/src/Pokedex.API
• Run dotnet run --configuration Release --project Pokedex.API.csproj
• Open your browser and please navigate to http://localhost:5000/swagger/index.html to use the API via Swagger

3. Via Docker

• Navigate to TL-Pokedex
• Run docker build -t pokedex .
• Then run docker run -p 8080:80 pokedex
• Last but not least, please navigate to http://localhost:8080/swagger/index.html to use the API via Swagger

Development decisions

I usually like to try new things while doing home assignments like this. It allows me to try something new outside of my working hours and learning a bit more :)

This time around, I used Hexagonal Architecture as I was reading in some blogs. All the business logic are under Pokedex.Domain and the external queries (or outputs) are under Pokedex.Queries.

Moreover, another new stuff I've found was Refit for creating the HttpClient for the external providers, and decided to use as well.

Things I'd do for a production API / Things to be improved

I tried to keep as simple as possible due to time constraints and, of course, this being a home assignment a not a fully production-ready API. I've listed a few things that I would implement if we would deploy to production:

• Implement a caching strategy such as MemCached or redis for instance. I've pointed out in the code. This also would include changes by adding a docker-compose for running all the required containers all together.
• I'm filtering out English descriptions but I suppose we could make this more open.
• Add some circuit-break/retry policy for the third party APIs. Something with Polly would improve handling with errors, small timeouts, too many requests, etc.
• Probably add an API versioning too depending on use-case.
• Add some more logs throughout the application. Possibly implement a middleware that could help with that too.
• Add some tracing and observability (o11y) for incoming and outgoing requests and responses.
• For dev/staging environment we could create a mocking service using WireMock so we don't rely on third parties for the external API responses. That would help development and testing across.
• Improve and add more tests. Most of the unit tests cases were done but I would add some Integration Tests covering "real" API requests and responses.