import streamlit as st
import numpy as np
import pandas as pd
import pickle
from sklearn.linear_model import LinearRegression
import warnings

# Ignore InconsistentVersionWarning from scikit-learn
warnings.filterwarnings("ignore", category=UserWarning, message=".*InconsistentVersionWarning.*")

# Example data for training - include 'year_built' as a feature
X_train = np.array([[1000, 3, 2, 1, 2000], [1500, 4, 3, 2, 1995], [2000, 3, 2, 2, 2010]])  # Include year_built
y_train = np.array([300000, 400000, 500000])  # Example target data

# Train the model with the example data
model = LinearRegression()
model.fit(X_train, y_train)

# Save the model with the current version of scikit-learn (after training)
with open("elite27_new.pkl", "wb") as f:
    pickle.dump(model, f)

# Title for the app
st.title("🏡 House Price Prediction App")

# User input fields (including 'year_built')
square_feet = st.number_input("Enter Square Feet:", min_value=500, max_value=10000)
bedrooms = st.number_input("Enter Number of Bedrooms:", min_value=1, max_value=10)
bathrooms = st.number_input("Enter Number of Bathrooms:", min_value=1, max_value=10)
neighborhood = st.selectbox("Select Neighborhood: 0:Rural, 1:Semi Urban, 2:Urban", [0, 1, 2])
year_built = st.number_input("Enter Year Built:", min_value=1900, max_value=2025)

# Define feature names (with 'YearBuilt')
feature_names = ['SquareFeet', 'Bedrooms', 'Bathrooms', 'Neighborhood', 'YearBuilt']

# Predict price when the button is clicked
if st.button("Predict Price 💰"):
    # Reshape user input into a 2D array (including 'year_built')
    user_data = np.array([[square_feet, bedrooms, bathrooms, neighborhood, year_built]])
    
    # Convert user data into a DataFrame and set proper column names
    user_data_df = pd.DataFrame(user_data, columns=feature_names)
    
    # Predict price
    prediction = model.predict(user_data_df)
    
    # Display result
    st.success(f"🏠 Estimated House Price: ${prediction[0]:,.2f}")