Bookworm Project

Bookworm is a Document-Based Q&A solution designed to work with Telegram bots. It leverages retrieval augmented generation (RAG) combined with llama-index to read documents and provide accurate answers to user queries. This bot helps users get the information they need from any document quickly and efficiently.

Progress

So far, the following features have been implemented:

1. Simple authentication system with Django.
2. Process to create custom bot instances.
3. Collection and data source creation process.
4. Endpoint to start the bot.

Prerequisites

• Python3

• Django

• pyTelegramBotAPI

• llama-index

How to run

• clone this repo

• get your bot api_key from botfather

• run : python manage.py runserver

• run celery : celery -A study_bot worker -l INFO --pool=eventlet --statedb=study_bot/worker/worker.state

• register via postman on http://localhost:8000/register

• login to get access token http://localhost:8000/login

• create collection http://localhost:8000/collection/

• add a datasource to collection http://localhost:8000/create_datasource/

• create a bot instance on db http://localhost:8000/bot/

• start bot http://localhost:8000/startbot/

• interact with bot on telegram