The repo is just a simple project to categorize tweets based on the text entered.
the pipeline code is written in javascript.
- The simpler, faster, and cheaper the architecture, the better. And the more interesting the categories, the better. It should not be simple keyword extraction.
- It will work like this.
- the backend will receive an array of tweets via an API call
- check if the tweet exists on the DB
- if not, use chatGPT to generate categories
- send back categorized tweets