Redoc

Document research and theme identification application with multi-level citation capabilities.

Overview

Redoc is a chatbot-style application that can:

1. Analyze 75+ documents
2. Identify common themes across documents
3. Provide cited responses to queries
4. Support multi-level citation (document, paragraph, sentence)

Tech Stack

• Frontend: React with TypeScript, Tailwind CSS
• Backend: FastAPI, Python
• Document Processing: Langchain, ChromaDB, Postgresql

Project Structure

• backend/: FastAPI backend service
• frontend/: React frontend application
• tests/: Test suite (backend unit/integration, frontend unit, E2E tests)
• docs/: Documentation (API docs, user guide)
• dataset/: Sample documents (PDF files)

Getting Started

Prerequisites

• Node.js 18+
• Python 3.10+
• npm or yarn

Installation

1. Clone the repository:

   git clone https://github.com/Abhinavexists/Redoc.git
   cd Redoc

2. Install backend dependencies:

   cd backend
   python -m venv .venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   pip install -r requirements.txt

   Recommended: Install with uv (faster, more reliable):

   cd backend
   uv venv .venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   uv pip install -e .

   This method leverages the pyproject.toml file for more consistent dependency management.

3. Install frontend dependencies:

   cd frontend
   npm install

Running the Application

Backend

# (http://localhost:8000)
cd backend 
python run.py

Frontend

# Frontend (http://localhost:3000)
npm run dev

Documentation

• API Documentation
• User Guide
• Architecture
• Test

Demo

Demo.mp4

License

MIT