app.py

# -*- coding: utf-8 -*-
import gradio as gr
from model import Model

# Initialize model
model = Model("sgd_model_pipeline.joblib")


def predict_price(
    squere,
    dist_1,
    dist_2,
    dist_3,
    category_encoded,
    floor_,
    offer_type,
    series,
    condition,
    building_type,
    year
):
    """
    Make apartment price prediction based on input features.
    """
    # Create features dictionary
    features = {
        'squere': int(squere),
        'dist_1': float(dist_1),
        'dist_2': float(dist_2),
        'dist_3': float(dist_3),
        'CategoryEncoded': float(category_encoded),
        'floor_': str(floor_),
        'Тип предложения': offer_type,
        'Серия': series,
        'Состояние': condition,
        'dom': building_type,
        'year': int(year)
    }

    # Get prediction
    try:
        predicted_price = model.predict(features)
        return f"Predicted Price: ${predicted_price:,.2f} USD"
    except Exception as e:
        return f"Error: {str(e)}"


# Define dropdown options based on schema
offer_types = ['от агента', 'от собственника']

building_series = [
    'индивид. планировка',
    'элитка',
    '106 серия',
    'хрущевка',
    '106 серия улучшенная',
    '104 серия',
    '108 серия',
    '105 серия',
    '104 серия улучшенная',
    '105 серия улучшенная',
    'пентхаус',
    'малосемейка',
    'сталинка',
    '107 серия'
]

conditions = [
    'евроремонт',
    'под самоотделку (псо)',
    'хорошее',
    'среднее',
    'не достроено'
]

building_types = ['кирпичный', 'монолитный', 'панельный']


# Create Gradio interface
with gr.Blocks(theme=gr.themes.Ocean()) as demo:
    gr.Markdown("# Bishkek Apartment Price Prediction")
    gr.Markdown("Enter apartment details to get a price prediction")

    with gr.Row():
        with gr.Column():
            gr.Markdown("### Basic Information")
            squere = gr.Number(
                label="Area (square meters)",
                value=50,
                minimum=10,
                maximum=500
            )
            floor_ = gr.Textbox(
                label="Floor Number",
                value="1",
                placeholder="e.g., 1, 3, 6"
            )
            year = gr.Number(
                label="Building Year",
                value=2010,
                minimum=1950,
                maximum=2025
            )

            gr.Markdown("### Location Features")
            category_encoded = gr.Number(
                label="Category Encoded (mean square for location)",
                value=60.0
            )
            dist_1 = gr.Number(
                label="Distance 1 (km)",
                value=1.0,
                minimum=0
            )
            dist_2 = gr.Number(
                label="Distance 2 (km)",
                value=2.0,
                minimum=0
            )
            dist_3 = gr.Number(
                label="Distance 3 (km)",
                value=3.0,
                minimum=0
            )

        with gr.Column():
            gr.Markdown("### Building Characteristics")
            offer_type = gr.Dropdown(
                choices=offer_types,
                label="Offer Type",
                value=offer_types[0]
            )
            series = gr.Dropdown(
                choices=building_series,
                label="Building Series",
                value=building_series[0]
            )
            condition = gr.Dropdown(
                choices=conditions,
                label="Condition",
                value=conditions[0]
            )
            building_type = gr.Dropdown(
                choices=building_types,
                label="Building Type",
                value=building_types[0]
            )

    # Predict button and output
    with gr.Row():
        predict_btn = gr.Button("Predict Price", variant="primary", size="lg")

    with gr.Row():
        output = gr.Textbox(label="Prediction Result", scale=2)

    # Set up the prediction function
    predict_btn.click(
        fn=predict_price,
        inputs=[
            squere,
            dist_1,
            dist_2,
            dist_3,
            category_encoded,
            floor_,
            offer_type,
            series,
            condition,
            building_type,
            year
        ],
        outputs=output
    )

    # Example section
    gr.Markdown("---")
    gr.Markdown("### Example Values")
    gr.Markdown("""
    - **Area**: 60-80 sq meters (typical 2-room apartment)
    - **Floor**: 1-16 (depending on building)
    - **Year**: 1960-2024
    - **Distances**: 0.5-5 km to points of interest
    """)


if __name__ == "__main__":
    demo.launch(share=True)