Production Disruption Early Warning System - Backend API

This is the backend server for the AI-Enabled Production Disruption Early Warning System. It simulates real-time manufacturing telemetry, runs an Isolation Forest model to detect anomalies on the fly, and provides a GenAI Decision Support layer for factory operators.

🚀 Setup & Execution

1. Install Dependencies

pip install fastapi uvicorn scikit-learn numpy pandas httpx

2. Run the Backend Server

uvicorn main:app --reload

• Server URL: http://127.0.0.1:8000
• Interactive API Docs: http://127.0.0.1:8000/docs

---

⚡ API Endpoints Reference

Core Telemetry & Status

Endpoint	Method	Description
/api/status	GET	Returns current system state ("NORMAL" / "DISRUPTED") & latest telemetry.
/api/telemetry/history	GET	Returns the last 20 telemetry records (chronological order for line charts).
/api/alerts	GET	Returns the 10 most recent disruption alerts.
/api/telemetry	POST	Manually pushes telemetry to evaluate against Isolation Forest.

Demo Simulation & Scenarios

Endpoint	Method	Description
/api/inject-failure	POST	Forces live simulator into generating disruption data.
/api/reset-demo	POST	Clears all alerts and restores simulator to normal operation.
/api/scenarios	GET	Returns the 5 predefined factory disruption scenarios.
/api/scenarios/{name}/activate	POST	Instantly loads and activates one of the 5 scenarios into the database (normal_operation, high_processing_delay, high_error_rate, low_throughput, combined_disruption).

GenAI Decision Support & Chat Assistant

Endpoint	Method	Description
/api/ai/decision-support	GET	Returns the structured 5-part AI intelligence (what_happened, possible_causes, prediction, risk_score, action_required with dynamic role filtering).
/api/ai/decision-support/scenario/{name}	POST	Returns 5-part decision intelligence for any specific scenario.
/api/ai/chat	POST	Conversational assistant endpoint for chat widgets (e.g. {"query": "What should maintenance do?"}).

---

🔒 CORS

CORS is enabled for all origins (*), so any frontend application running on localhost:3000, localhost:5173, etc., can communicate freely with this API.