import gradio as gr
import joblib
import pandas as pd
from datetime import datetime
import numpy as np

# Load the model and scaler
model = joblib.load('crash_detection_model.pkl')
scaler = joblib.load('scaler.pkl')

# Define the prediction function
def predict_crash(accel_x, accel_y, accel_z, gyro_x, gyro_y, gyro_z, timestamp):
    # Convert timestamp
    timestamp = pd.to_datetime(timestamp)
    hour_of_day = timestamp.hour
    day_of_week = timestamp.dayofweek

    # Create DataFrame
    sensor_data = pd.DataFrame([[accel_x, accel_y, accel_z, gyro_x, gyro_y, gyro_z, hour_of_day, day_of_week]],
                               columns=['accel_x', 'accel_y', 'accel_z', 'gyro_x', 'gyro_y', 'gyro_z', 'hour_of_day', 'day_of_week'])

    # Scale the input data
    sensor_data_scaled = scaler.transform(sensor_data)

    # Predict crash
    prediction = model.predict(sensor_data_scaled)
    
    return "Crash Detected" if prediction[0] == 1 else "No Crash Detected"

# Create Gradio interface
iface = gr.Interface(
    fn=predict_crash,
    inputs=[
        gr.Number(label="Accelerometer X"),
        gr.Number(label="Accelerometer Y"),
        gr.Number(label="Accelerometer Z"),
        gr.Number(label="Gyroscope X"),
        gr.Number(label="Gyroscope Y"),
        gr.Number(label="Gyroscope Z"),
        gr.Textbox(label="Timestamp (YYYY-MM-DD HH:MM:SS)")
    ],
    outputs="text",
    title="Crash Detection Model",
    description="Predicts if a crash has occurred based on accelerometer and gyroscope data."
)

# Launch the Gradio app
if __name__ == "__main__":
    iface.launch()