isit-a-cat is an application to predict whether an uploaded picture shows a cat or not.

The core components are a tensorflow machine learning model and a web application to upload pictures and trigger prediction.

There also is a telegram bot.

• frontend: Vue.js application where you can upload a picture and see if it's a cat or not
• backend: golang service that handles RESTful HTTP requests from the frontend
• bot: golang telegram bot as an alternative to the web application
• learn: the python machine learning code

• min.io: Kubernetes-native object storage. Used for storing the machine learning model and uploaded images
• Keras: Python deep learning API. Used with tensorflow backend to train a deep learning model

Prerequisites

• Docker
• docker-compose

Deploy:

Clone this repo to a folder of your choice.

Rename the env_template to .env and set the two environment variables within to random values.

E.g. with:

openssl rand --base64 10
openssl rand --base64 40

Then build and run everything with:

docker-compose up --build -d --remove-orphans