import streamlit as st
import pandas as pd
import joblib

st.set_page_config(page_title="Predictive Maintenance App V2", layout="centered")

@st.cache_resource
def load_model():
    return joblib.load("best_model.joblib")

st.title("Predictive Maintenance for Engine Health")
st.write("Enter the engine sensor values below to predict engine condition.")

engine_rpm = st.number_input("Engine RPM", min_value=0.0, value=850.0)
lub_oil_pressure = st.number_input("Lub Oil Pressure", min_value=0.0, value=3.5)
fuel_pressure = st.number_input("Fuel Pressure", min_value=0.0, value=6.8)
coolant_pressure = st.number_input("Coolant Pressure", min_value=0.0, value=2.4)
lub_oil_temp = st.number_input("Lub Oil Temperature", min_value=0.0, value=78.0)
coolant_temp = st.number_input("Coolant Temperature", min_value=0.0, value=80.5)

if st.button("Predict Engine Condition"):
    try:
        model = load_model()

        input_df = pd.DataFrame([{
            "Engine_rpm": engine_rpm,
            "Lub_oil_pressure": lub_oil_pressure,
            "Fuel_pressure": fuel_pressure,
            "Coolant_pressure": coolant_pressure,
            "lub_oil_temp": lub_oil_temp,
            "Coolant_temp": coolant_temp
        }])

        prediction = model.predict(input_df)[0]

        if prediction == 1:
            st.error("Prediction: Engine may require maintenance.")
        else:
            st.success("Prediction: Engine appears to be operating normally.")

        st.write("Input dataframe used for prediction:")
        st.dataframe(input_df)

    except Exception as e:
        st.error(f"Prediction failed: {e}")