import streamlit as st
import joblib
import numpy as np

# Load models
linear_model = joblib.load("linear_regression_model.pkl")
decision_tree_model = joblib.load("decision_tree_regression_model.pkl")
random_forest_model = joblib.load("random_forest_regress_model.pkl")

app, model_eval = st.tabs(["Application", "Model Evaluation"])

with app:
    # Model selection (exclusive to this tab)
    st.sidebar.title("Select Model")
    model_choice = st.sidebar.radio("Choose a model:", ("Linear Regression", "Decision Tree Regression", "Random Forest Regression"))

    if model_choice == "Linear Regression":
        model = linear_model
    elif model_choice == "Decision Tree Regression":
        model = decision_tree_model
    else:
        model = random_forest_model

    # User inputs
    st.title("House Price Prediction")
    area = st.number_input("Area (sq ft)", min_value=100, max_value=10000, step=10)
    bedrooms = st.slider("Bedrooms", 1, 5, 3)
    bathrooms = st.slider("Bathrooms", 1, 4, 2)
    floors = st.slider("Floors", 1, 3, 1)
    year_built = st.number_input("Year Built", min_value=1800, max_value=2025, step=1)

    location = st.radio("Location", ["Suburban", "Downtown", "Rural", "Urban"])
    condition = st.radio("Condition", ["Poor", "Good", "Excellent", "Fair"])
    garage = st.radio("Garage", ["No", "Yes"])

    # Encode categorical inputs
    location_dict = {"Suburban": 0, "Downtown": 1, "Rural": 2, "Urban": 3}
    condition_dict = {"Poor": 0, "Fair": 1, "Good": 2, "Excellent": 3}
    garage_dict = {"No": 0, "Yes": 1}

    location_encoded = location_dict[location]
    condition_encoded = condition_dict[condition]
    garage_encoded = garage_dict[garage]

    # Prediction button
    if st.button("Predict Price"):
        features = np.array([[area, bedrooms, bathrooms, floors, year_built, location_encoded, condition_encoded, garage_encoded]])
        predicted_price = model.predict(features)[0]
        st.write(f"### Predicted House Price: ${predicted_price:,.2f}")

with model_eval:
    st.title("Model Evaluation")
    
    st.subheader("Decision Tree Regression")
    st.image("DTR_metrics.png", caption="Decision Tree Regression Metrics")
    st.image("evalMetric_DTR.png", caption="Decision Tree Regression - Actual vs Predicted")
    
    st.subheader("Linear Regression")
    st.image("LR_metrics.png", caption="Linear Regression Metrics")
    st.image("evalMetric_LR.png", caption="Linear Regression - Actual vs Predicted")
    
    st.subheader("Random Forest Regression")
    st.image("RFR_metrics.png", caption="Random Forest Regression Metrics")
    st.image("evalMetric_RFR.png", caption="Random Forest Regression - Actual vs Predicted")