USER_GUIDE.md

🌿 Farm Intelligence Platform (Pune AgriAI) - User Guide

Welcome to the Pune AgriAI Farm Intelligence Platform, a state-of-the-art AI-driven ecosystem designed to transform traditional farming into high-precision, data-backed agriculture. This guide provides comprehensive instructions for deploying and utilizing the platform's advanced features.

---

🚀 Quick Start Guide

1. Prerequisites

• Python: 3.9 or 3.10 (3.11+ may require minor dependency adjustments)
• Database: MongoDB (Local or Atlas)
• API Keys: Required for AI reasoning and market data.

2. Installation

Clone the repository and install the required packages:

pip install -r requirements.txt

3. Environment Configuration

Create/Update your .env file in the root directory:

MONGO_CONNECTION_URL="your_mongodb_url"
GEMINI_API_KEY="your_google_gemini_key"
NVIDIA_API_KEY="your_nvidia_ai_key"
GROQ_API_KEY="your_groq_key"

4. Running the Application

Start the Flask server:

python app.py

Access the dashboard at: http://127.0.0.1:5000

---

🛠️ The Intelligence Modules

📋 1. Strategic Planner (The Brain)

The Strategic Planner generates a comprehensive 12-month Agricultural Roadmap. It doesn't just suggest crops; it builds an entire business plan for your land.

• How to Initiate: Navigate to the "Plan" section and click "Run AI Agent".
• Calculations: It processes land size, capital, and soil data to determine:
  • Crop Roster: Optimal main and inter-cropping sequences.
  • Cashflow Projections: Predicted monthly revenue vs. expenses (CapEx).
  • Strategic Timing: Precise sowing and harvest windows to catch high market prices.
• AI Logic: Primarily uses NVIDIA/Groq for rapid reasoning, with an auto-switch to Gemini 1.5 Flash if endpoints are busy.

🌱 2. Real-Time Growth Monitor

This dashboard provides a "Command Center" view of your farm's health.

• Glassmorphic Interface: Visualizes soil pH, Nitrogen, and Moisture via real-time (or simulated IoT) sensors.
• Dynamic Insights: AI analyzes current sensor data and weather forecasts to provide immediate recommendations.
• Health Indicators: Categorizes farm state (Optimal, Good, Needs Care) using ML thresholds.

🐛 3. Pest & Disease Intelligence

A predictive engine that forecasts biological risks before they manifest.

• Predictive Modeling: Analyzes humidity, temperature, and crop stage to predict bug outbreaks.
• Actionable Alerts: If a "High Risk" is detected, the system provides a specific protection protocol (spraying schedule, biological controls).

🍎 4. Market & Harvest Intelligence

Maximizes profit by connecting your farm to real-world market trends (via Data.gov.in).

• Mandi Insights: Live tracking of modal prices across Maharashtra districts (Pune, Nashik, Nagpur).
• Selling Strategy: High-level AI advice on whether to sell immediately or store harvest based on price trends.
• Geo-Restricted Search: Prioritizes local markets (Pune/MH) first, then suggests high-value national hubs.

💰 5. Financial Flow & ROI

Track your farm's performance like a modern enterprise.

• Net Yield Tracker: Real-time calculation of revenue minus expenses.
• ROI Analysis: Visualizes Return on Investment for each crop cycle.
• Monthly Fiscal Bar Charts: Identifies months of high surplus or potential deficit to help with financial planning.

---

🔍 System Architecture & Troubleshooting

Bulletproof AI Pipeline

The platform uses a Multi-Model Intelligence Layer:

• Layer 1 (Performance): Groq/NVIDIA Llama-3.1-405B for strategy.
• Layer 2 (Reliability): Gemini 1.5 Flash acts as a fallback to ensure the system never crashes during API outages.

Common Issues

Note: "AI Pipeline Initializing": This occurs during the first strategic run as the reasoning agent warms up. Please wait 10-15 seconds.

MongoDB Errors: Ensure your IP address is whitelisted in MongoDB Atlas or your local service is running.

---

🎓 For Developers & Researchers

• Logic Engines: Explore modules/pest_logic.py, soil_logic.py, and irrigation_logic.py for the core agricultural intelligence.
• Agentic Pipeline: See the agentic/ directory for the complex reasoning chains that power the 12-month planner.
• UI Design: The frontend is built with high-fidelity glassmorphism, using custom CSS variables (Emerald/Dark Mode) located in index2.html and templates/.

---

Tip: Use the "Refresh Report" button in the dashboard to force-sync the latest market data and sensor readings.

Developed for the Pune AgriAI Precision Agriculture Initiative 2026