Disaster Bot

A real-time disaster response system using FastAPI, MongoDB Atlas, and Google AI.

Features

• Real-time earthquake data from USGS
• Emergency shelter information from OpenStreetMap
• Vector search for relevant shelters
• Natural language Q&A with Gemini AI
• Interactive web interface with live map
• Time series data storage for disasters

Setup Instructions

1. Prerequisites

• Python 3.10+
• MongoDB Atlas account
• Google Cloud Project with Vertex AI enabled
• Gemini API key

2. Environment Setup

# Clone the repository
git clone <your-repo-url>
cd disaster-bot

# Create virtual environment
python -m venv venv

# On Mac:
source venv/bin/activate

# On Windows:
. venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

3. Configuration

# Create .env with your credentials
MONGODB_URL=mongodb+srv://<username>:<password>@cluster.mongodb.net
GEMINI_API_KEY=your-gemini-api-key
GOOGLE_CLOUD_PROJECT=your-project-id
MONGODB_DATABASE="disaster_bot"

4. MongoDB Atlas Setup

1. Create a MongoDB Atlas cluster
2. Create a database called "disaster_bot"
3. Set up collections using the setup script:

python setup_mongodb_atlas.py

4. Create a vector search index in Atlas UI:
   • Go to Atlas → Search → Create Vector Search Index
   • Choose JSON Editor
   • Fill out the configuration:

{
  "fields": [
    {
      "type": "vector",
      "path": "embedding",
      "numDimensions": 768,
      "similarity": "cosine"
    }
  ]
}

5. Pre-process shelters data into MongoDB:

python fetch_shelters_to_db.py

5. Google Cloud Setup

1. Create a Google Cloud Project
2. Enable Vertex AI API
3. Create a service account and download JSON key
4. Set Google credentials in .env

6. Run the Application

# Development
uvicorn main:app --reload

The application will be available at http://localhost:8000

7. Deployment

Using the Cloud Console:

1. Go to: https://console.cloud.google.com/run

2. Click Create Service

3. Select:

   • Source: "Deploy from source"

   • Repository: Connect your GitHub repo (authorize if needed)

   • Branch: main

   • Build config: Dockerfile

4. Set the region (e.g., us-central1)

5. Allow unauthenticated access (so the public can access your API/UI)

6. Click Deploy

NOTE: If you're getting an SSL handshake error, you should check Network Access in Atlas. Go to MongoDB Atlas > Network Access > IP Access List, modify the IP address to 0.0.0.0/0