Go Fetch! ( Live )
GoFetch is a web application written in NextJS / React, and styled with TailwindCSS; It allows users to view a stream of images from the Dog CEO API, being able to sort by breed. Images are converted to AVIF & WebP by NextJS, and cached locally, so as to be considerate and keep excessive calls to the 3rd party API to a minimum.
Additionally, it connects to the backend, written in Go, to "GoFetchBot", a github application that accepts user submitted images of their dogs; scans them with a Neural Network for compliance (images must contain a dog, and may not contain a human, for GDPR reasons). If the image is acceptable, it creates a pull request with the image as a commit, for final human approval.
To reiterate: that makes it a Pet Form as a Service, using Neural Networks.
- Golang, with 100% test coverage, tested with go test, using TDD & sticking to a "Given When Then" syntax
- Some limited End to End Testing
- using GoCV and Yolov3 neural network
- NextJS / React, styled with TailwindCSS
- Deployed to a personally administered docker server
I don't expect much attention on this repo; but this should be fairly easy to fork. The main requisite is either setting up openCV, or being comfortable using a docker container where openCV has been installed. The tests were written with a mock github repo (that has a LOT of closed PRs), so you will have to set up an application to get the same test coverage. Forker Beware: I am not proud of some of the React code, but I am pretty proud of the Go code, so...
Go ahead and shoot me an email or make an issue or reach out on my socials if this is interesting to you or you want to chat!
GPLv3