PricesHousesModel.py

import numpy as np
import joblib
!pip install gradio
import gradio as gr
from huggingface_hub import hf_hub_download

# Download the model file if it doesn't exist locally
model_filename = "knn_house_model.pkl"
try:
    model_path = hf_hub_download(repo_id="Tahani1/Houses-Prices-Prediction", filename=model_filename)
except Exception as e:
    print(f"Error downloading '{model_filename}' from Hugging Face Hub: {e}")
    raise  # Re-raise the exception to stop execution

# Load the trained model and preprocessing tools
model = joblib.load(model_path) # Load the model from the downloaded path
scaler = joblib.load(hf_hub_download(repo_id="Tahani1/Houses-Prices-Prediction", filename="scaler.pkl"))
label_encoder = joblib.load(hf_hub_download(repo_id="Tahani1/Houses-Prices-Prediction", filename="label_encoder.pkl"))

# Function to predict house price
def predict_price(num_rooms, distance, country, build_quality):
    country_encoded = label_encoder.transform([country])[0]
    features = np.array([[num_rooms, distance, country_encoded, build_quality]])
    features_scaled = scaler.transform(features)
    predicted_price = model.predict(features_scaled)[0]
    return f"Predicted House Price: ${predicted_price:,.2f}"

# Gradio Interface
inputs = [
    gr.Number(label="Number of Rooms"),
    gr.Number(label="Distance to Center (km)"),
    gr.Dropdown(label="Country", choices=label_encoder.classes_.tolist()),
    gr.Slider(minimum=1, maximum=10, label="Build Quality")
]

outputs = gr.Textbox(label="Prediction Result")

# Create and launch Gradio app
app = gr.Interface(fn=predict_price, inputs=inputs, outputs=outputs, title="House Price Prediction")
app.launch()