main.py

# main.py
# Continuous AI service for SmartFlood (DEMO MODE)
# Waits 10s → reads data → processes AI → waits 5s → saves → repeat

import time
from datetime import datetime, UTC

from db import sensor_readings_col, weather_readings_col, ai_predictions_col
from predictor import predict_flood_risk


print("🤖 SmartFlood AI Service started (DEMO MODE: 10s READ → 5s SAVE)...")

while True:
    # 1️⃣ WAIT BEFORE READING
    print("\n⏳ Waiting 5 seconds before reading data...")
    time.sleep(5)

    print("🔍 Reading latest data for all sensors...")

    # 2️⃣ Get all sensor IDs
    sensor_ids = sensor_readings_col.distinct("sensorId")

    if not sensor_ids:
        print("⚠️ No sensors found.")
        continue

    for sensor_id in sensor_ids:
        print(f"\n📡 Processing sensor: {sensor_id}")

        # Latest sensor reading
        sensor = sensor_readings_col.find_one(
            {"sensorId": sensor_id},
            sort=[("createdAt", -1)]
        )

        # Latest weather reading
        weather = weather_readings_col.find_one(
            {"sensorId": sensor_id},
            sort=[("createdAt", -1)]
        )

        if not sensor or not weather:
            print(f"⚠️ Missing data for {sensor_id}. Skipping...")
            continue

        # 3️⃣ Extract features
        water_level = sensor.get("waterLevel", 0)
        rain_1h = weather.get("rainfall_1h", 0)
        rain_3h = weather.get("rainfall_3h", 0)

        created_at = sensor.get("createdAt", datetime.now(UTC))

        hour = created_at.hour
        day_of_week = created_at.weekday()
        month = created_at.month

        # 4️⃣ AI prediction
        risk_level, risk_score = predict_flood_risk(
            water_level, rain_1h, rain_3h
        )

        print(f"🧠 AI result for {sensor_id}: {risk_level.upper()}")

        # 5️⃣ WAIT BEFORE SAVING
        print("⏳ Waiting 5 seconds before saving...")
        time.sleep(5)

        # 6️⃣ Save to MongoDB
        prediction = {
            "sensorId": sensor_id,
            "inputs": {
                "waterLevel": water_level,
                "rainfall_1h": rain_1h,
                "rainfall_3h": rain_3h,
                "hour": hour,
                "dayOfWeek": day_of_week,
                "month": month
            },
            "prediction": {
                "riskLevel": risk_level,
                "riskScore": risk_score
            },
            "createdAt": datetime.now(UTC)
        }

        ai_predictions_col.insert_one(prediction)

        print(f"✅ Prediction saved for {sensor_id}")

    print("\n🔁 Cycle complete (≈15 seconds total).")