README.md

FloodRelief AI

Project Overview

This project is designed for flood disaster management. It integrates several components:

• A Flask-based frontend for user interaction.
• A FastAPI backend handling GIS operations, Postgres integration, and resource allocation.
• A CrewAI-based agent workflow communicating via Discord to analyze and process flood data from multiple agents.

Technologies Used

• Python
• Flask (Frontend)
• FastAPI (Backend)
• SQLAlchemy & GeoAlchemy2 (Database & GIS operations)
• PostgreSQL (Database)
• Or-Tools (Optimization for resource allocation)
• Geopy (Geocoding)
• CrewAI (Agent-based workflow and Discord communication)

Running the Project

1. Ensure you have all Python dependencies installed (use requirements.txt or pip install).
2. Set up environment variables including DATABASE_URL, OPENAI_API_KEY, and Discord configuration needed for CrewAI.
3. Start the backend server:
   • Navigate to flood-relief-ai/flood_relief/backend
   • Run: uvicorn main:app --reload
4. Start the Flask frontend:
   • Navigate to flood-relief-ai/flood_relief/frontend
   • Run: python app.py
5. For agent workflows, run the CrewAI agents (located under flood-relief-ai/flood_relief/crewai) to connect to Discord and execute tasks.

Experimenting with our Emergency Agent

1. You can join our Test Discord Server to recieve emergency alerts. These agents have been connected with our Geospatial OR Tools.