DeepScriptum

Convert any file into it's MarkDown source

Table of Contents

1. Current state and updates
2. About The Project
   • Built With
3. Getting Started
   • Installation
4. Acknowledgments

Current state and updates

This branch contains demo DeepScriptum implementation and map-reduce approach to perfom OCR in parallel. Service provides fast transformation of any file to .md The program allows you to quickly transform files Easy to use interface

December 2024:

Implementation with GPT-4o. Includes FastAPI backend (Celery used for task managing), PostgreSQL and HTML/CSS frontend.

October 2024:

Deep dive into OCR, understanding how it work and how to use it for the project.

About The Project

DeepScriptum is an advanced project designed to seamlessly convert any type of document into Markdown (.md) format. Utilizing cutting-edge AI technology, it ensures accurate and efficient translation while preserving the original structure and content. Ideal for developers, writers, and businesses, DeepScriptum streamlines documentation processes and enhances collaboration.

(back to top)

Architecture

Installation

1. Git clone or download
2. Create .env file in root due to .env.example
3. sudo docker-compose build && sudo docker-compose up

Congrats! You can access your service on localhost

(back to top)

Built With

• PostgreSQL
• RabbitMQ
• FastAPI
• Celery
• OpenAI GPT

(back to top)

Authors

• Vyaznikov Pavel
• Busko Nikita
• Rychkov Pavel
• Orlova Angelina

(back to top)