VERITAS (Versatile Environmental Risk Intelligence & Transparency Analysis System)

Overview

VERITAS is a state-of-the-art environmental risk assessment system which is designed to monitor, analyze, and provide actionable intelligence on indoor air quality. It leverages advanced machine learning models and a holistic synergy engine to detect complex, co-occurring environmental hazards.

This repository contains the modernized architecture of the system, migrated from a legacy Flask application to a high-performance FastAPI backend with a futuristic, responsive frontend.

Key Features

• Real-time Risk Intelligence: Instant analysis of sensor data (Temperature, Humidity, CO2, PM2.5, etc.).
• Synergistic Risk Detection: Identifies critical combinations of factors (e.g., High Humidity + Poor Ventilation).
• Holistic Health Score: A dynamic 0-100 score reflecting the overall safety of the environment.
• AI Agent Insights: Provides actionable, localized advice (Hinglish support included) for remediation.
• Modern Tech Stack: Built with FastAPI, Pydantic, Tailwind CSS, and Scikit-Learn/XGBoost.

System Architecture

Backend (main.py)

• Framework: FastAPI (Asynchronous, High Performance).
• Validation: Pydantic models ensure strict data integrity for all sensor inputs.
• ML Integration: Secure loading of veritas_model_bundle.pkl with fallback mechanisms.
• Logic:
  • prepare_input_features: Handles complex feature engineering (lag features, interaction terms).
  • generate_veritas_advice: Rule-based engine for actionable recommendations.

Configuration (config.py)

• Modularized configuration separating business logic from application code.
• Contains PROFESSIONAL_RISK_GUIDE and SYNERGISTIC_RISK_RULES.

Frontend (templates/index.html)

• Design: Glassmorphism aesthetic with Neon accents using Tailwind CSS.
• Interactivity: Vanilla JavaScript for lightweight, fast interactions with the API.
• Visualization: SVG-based Health Score gauge and dynamic risk cards.

Setup Instructions

Prerequisites

• Python 3.9+
• pip

Installation

1. Clone the repository:

   git clone <repository-url>
   cd veritas-system

2. Install dependencies:

   pip install -r requirements.txt

3. Verify Model File: Ensure veritas_model_bundle.pkl is present in the root directory.

Running the Application

Start the server using Uvicorn:

uvicorn main:app --reload

The application will be available at:

• Dashboard: http://localhost:8000
• API Docs: http://localhost:8000/docs

Environment Variables

To enable the email alert system, you must configure the following environment variables. You can create a .env file in the project root to store these secrets.

• EMAIL_USER: The Gmail address used to send alert emails (e.g., your.email@gmail.com).
• EMAIL_PASS: The Google App Password for the EMAIL_USER account. This is not your regular password. You must generate it from your Google Account security settings.

Example .env file:

EMAIL_USER="your.email@gmail.com"
EMAIL_PASS="yourgogleappassword"

API Endpoints

POST /api/analyze

Analyzes environmental data and returns a comprehensive risk assessment.

Request Body (JSON):

{
  "temperature": 25.0,
  "humidity": 50.0,
  "co2": 800,
  "pm2.5": 10.0,
  "pm10": 15.0,
  "tvoc": 50,
  "co": 0.5,
  "occupancy_count": 2,
  "motion_detected": 1,
  "light_intensity": 500
}

Response:

• primary_risks: List of detected single-source risks.
• synergistic_risks: List of combined hazard scenarios.
• holistic_assessment: Overall status text and severity level.
• health_score: Numeric score (0-100).
• veritas_advice: List of recommended actions.

Deployment Guidelines

• Production: Use a process manager like Gunicorn with Uvicorn workers.

  gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app

• Environment: Ensure the server has read access to veritas_model_bundle.pkl.
• Security: Place behind a reverse proxy (Nginx/Apache) with SSL termination.

Migration Notes

• Renamed project identity to VERITAS.
• Refactored app_old.py to main.py and config.py.
• Updated index.html to use modern Tailwind CSS classes.
• Preserved all feature engineering logic for model compatibility.