from flask import Flask, request, jsonify
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from flask_cors import CORS
import joblib
import logging
import requests
import datetime

app = Flask(__name__)
CORS(app)

# Setup logging
logging.basicConfig(level=logging.INFO)

# Load model dan scaler
try:
    model = load_model('lstm_model3.h5')
    scaler = joblib.load('scaler.pkl')
    logging.info("✅ Model dan scaler berhasil dimuat.")
except Exception as e:
    logging.error(f"❌ Gagal memuat model atau scaler: {e}")
    model = None
    scaler = None

SAMPLE_TRAINED = 30
LARAVEL_URL = "http://192.168.1.11:8000/api/predict"  # Sesuaikan dengan endpoint Laravel dan ip laptop

@app.route('/predict', methods=['POST'])
def predict():
    try:
        data = request.get_json()
        if not data or "sensor_data" not in data:
            return jsonify({"error": "Data tidak valid"}), 400

        sensor_data = data["sensor_data"]
        if len(sensor_data) != SAMPLE_TRAINED:
            return jsonify({"error": f"Jumlah data harus {SAMPLE_TRAINED}"}), 400
        
        # for sensor in data['sensor_data']:
        #     print(sensor) 

        # Pastikan semua data numerik
        try:
            input_data = np.array([[float(d["suhu"]), float(d["kelembaban"])] for d in sensor_data])
            timestamps = [datetime.datetime.strptime(d["datetime"], "%Y-%m-%d %H:%M:%S") for d in sensor_data]
        except (ValueError, KeyError) as e:
            logging.error(f"❌ Error dalam konversi data: {e}")
            return jsonify({"error": "Format data tidak valid"}), 400

        if model is None or scaler is None:
            return jsonify({"error": "Model atau scaler tidak tersedia"}), 500

        # Normalisasi data
        input_scaled = scaler.transform(input_data)
        input_scaled = np.reshape(input_scaled, (1, SAMPLE_TRAINED, 2))

        # Prediksi
        prediction_scaled = model.predict(input_scaled)
        prediction = scaler.inverse_transform(prediction_scaled)

        # Ambil waktu prediksi (5 menit setelah data terakhir)
        last_datetime = max(timestamps)
        prediction_datetime = last_datetime + datetime.timedelta(minutes=5)

        # Format hasil prediksi
        hasil_prediksi = {
            "predicted_suhu": round(float(prediction[0][0]), 2),
            "predicted_kelembaban": round(float(prediction[0][1]), 2),
            "prediction_datetime": prediction_datetime.strftime("%Y-%m-%d %H:%M:%S")
        }

        logging.info(f"✅ Hasil prediksi: {hasil_prediksi}")

        # Kirim ke Laravel
        try:
            response = requests.post(LARAVEL_URL, json=hasil_prediksi, timeout=30)
            if response.status_code == 201:
                logging.info("✅ Hasil prediksi berhasil dikirim ke Laravel")
            else:
                logging.error(f"❌ Gagal mengirim ke Laravel: {response.text}")
        except requests.RequestException as e:
            logging.error(f"❌ Error mengirim ke Laravel: {e}")
            return jsonify({"error": "Gagal mengirim ke Laravel"}), 500

        return jsonify({"message": "Data berhasil dikirim ke Laravel"}), 200  # Hanya mengirim status ke client

    except Exception as e:
        logging.error(f"❌ Terjadi error: {e}")
        return jsonify({"error": "Terjadi kesalahan pada server"}), 500


if __name__ == '__main__':
    app.run(host="0.0.0.0", port=5000, debug=True)