VLIB - A College Library Management System built on Vector Database ??
A library management platform that simplifies the search experience in a library by allowing any user to use any basic natural query to find the most appropriate resource ?. Simply put, the goal is to design a system that improves the search experience for users and make it easier to find ? what they are looking for.
Technologies used :
- ReactJS
- DJango
- PostgreSQL
- pgvector
- Mixtral
Key Features:
- Use as a replacement for conventional library management systems by adding books/material of various types and thereby building a database.
- Search for books using very basic and descriptive questions rather than specifically including certain keywords like author name, topic name etc.
- Obtain AI generated answers for queries based on content from available top results reducing one step especially if youre in a lookout for quick answers
- Will take you to the exact page of the book/material that discusses the content requested in the query.
- LRU based semantic caching has been enabled for getting quicker results. It caches the most common search queries (eg: most beginner users search the same topics in the library) and loads it in when a queery of high semantic similarity is typed in.
Setting up the frontend:
- Install all packages with
npm install
- Add the backend API url in
src/pages/consts.js
- Run the server with
npm start
- Enjoy!!
