import gradio as gr
import pandas as pd
import joblib

# Load the trained model
model = joblib.load('random_forest_model.pkl')  # replace with your model path

# Define the function to make predictions
def predict_price(host_id, neighbourhood_group, room_type, number_of_reviews, calculated_host_listings_count, latitude, longitude):
    # Create a dictionary with default values
    data = {
        'host_id': host_id,
        'number_of_reviews': number_of_reviews,
        'calculated_host_listings_count': calculated_host_listings_count,
        'latitude': latitude,
        'longitude': longitude,
        # Set dummy values for categorical features
        'neighbourhood_group_Brooklyn': 0,
        'neighbourhood_group_Manhattan': 0,
        'neighbourhood_group_Queens': 0,
        'neighbourhood_group_Bronx': 0,
        'neighbourhood_group_Staten Island': 0,
        'room_type_Shared room': 0,
        'room_type_Private room': 0,
        'room_type_Entire home/apt': 0,
    }
    
    # Set appropriate values based on user input
    if neighbourhood_group == 'Brooklyn':
        data['neighbourhood_group_Brooklyn'] = 1
    elif neighbourhood_group == 'Manhattan':
        data['neighbourhood_group_Manhattan'] = 1
    elif neighbourhood_group == 'Queens':
        data['neighbourhood_group_Queens'] = 1
    elif neighbourhood_group == 'Bronx':
        data['neighbourhood_group_Bronx'] = 1
    elif neighbourhood_group == 'Staten Island':
        data['neighbourhood_group_Staten Island'] = 1
    
    if room_type == 'Shared room':
        data['room_type_Shared room'] = 1
    elif room_type == 'Private room':
        data['room_type_Private room'] = 1
    elif room_type == 'Entire home/apt':
        data['room_type_Entire home/apt'] = 1

    # Create DataFrame from the dictionary
    custom_data = pd.DataFrame([data])

    # Make prediction
    predicted_price = model.predict(custom_data)
    return predicted_price[0]

# Create the Gradio interface
iface = gr.Interface(
    fn=predict_price,
    inputs=[
        gr.Number(label="Host ID"),
        gr.Dropdown(label="Neighbourhood Group", choices=["Brooklyn", "Manhattan", "Queens", "Bronx", "Staten Island"]),
        gr.Dropdown(label="Room Type", choices=["Shared room", "Private room", "Entire home/apt"]),
        gr.Number(label="Number of Reviews"),
        gr.Number(label="Calculated Host Listings Count"),
        gr.Number(label="Latitude"),
        gr.Number(label="Longitude")
    ],
    outputs=gr.Number(label="Predicted Price"),
    title="Airbnb Price Prediction",
    description="Enter the details to predict the price of an Airbnb listing."
)

# Launch the interface
iface.launch(share=True)  # Set share=True to create a public link