src/src_backup/streamlit_app.py

import streamlit as st
import pandas as pd
import requests
import plotly.express as px

# -------------------------------
# Load Final Dataset
# -------------------------------
@st.cache_data
def load_data():
    url = 'https://raw.githubusercontent.com/xprafulx/car-predict/refs/heads/main/final_car_data.csv'
    df = pd.read_csv(url)
    return df

df = load_data()

# -------------------------------
# Sidebar: User Inputs
# -------------------------------
st.sidebar.header("Select Car Features")

# Numeric Features (already engineered in dataset)
car_age_choice = st.sidebar.select_slider(
    "Car Age",
    options=sorted(df['car_age'].unique())
)

mileage_choice = st.sidebar.slider(
    "Mileage per Year",
    float(df['mileage_per_year'].min()),
    float(df['mileage_per_year'].max()),
    float(df['mileage_per_year'].min())
)

levy_choice = st.sidebar.slider(
    "Levy",
    float(df['levy'].min()),
    float(df['levy'].max()),
    float(df['levy'].min())
)

engine_volume_choice = st.sidebar.select_slider(
    "Engine Volume",
    options=sorted(df['engine_volume'].unique())
)

cylinders_choice = st.sidebar.select_slider(
    "Cylinders",
    options=sorted(df['cylinders'].unique())
)

airbags_choice = st.sidebar.select_slider(
    "Airbags",
    options=sorted(df['airbags'].unique())
)

# Categorical Features
category_choice = st.sidebar.selectbox("Category", sorted(df['category'].unique()))
fuel_type_choice = st.sidebar.selectbox("Fuel Type", sorted(df['fuel_type'].unique()))
gear_box_choice = st.sidebar.selectbox("Gear Box Type", sorted(df['gear_box_type'].unique()))
drive_wheels_choice = st.sidebar.selectbox("Drive Wheels", sorted(df['drive_wheels'].unique()))
doors_choice = st.sidebar.selectbox("Doors", sorted(df['doors'].unique()))
wheel_choice = st.sidebar.selectbox("Wheel", sorted(df['wheel'].unique()))
color_choice = st.sidebar.selectbox("Color", sorted(df['color'].unique()))

# Boolean Feature
leather_interior_choice = st.sidebar.checkbox("Leather Interior", value=False)

# -------------------------------
# Prepare API Payload (fixed serialization)
# -------------------------------
payload = {
    "data": [
        [
            int(car_age_choice),
            float(engine_volume_choice),
            int(cylinders_choice),
            int(airbags_choice),
            float(mileage_choice),
            float(levy_choice),
            str(category_choice),
            str(fuel_type_choice),
            str(gear_box_choice),
            str(drive_wheels_choice),
            str(doors_choice),
            str(wheel_choice),
            str(color_choice),
            bool(leather_interior_choice)
        ]
    ]
}

# -------------------------------
# Call Hugging Face API
# -------------------------------
api_url = "https://appleballcay-car-price-api.hf.space/run/predict_batch"
response = requests.post(api_url, json=payload)

if response.status_code == 200:
    predicted_price = response.json()['data'][0]
    st.subheader("Predicted Car Price")
    st.write(f"${predicted_price:,.2f}")

    # -------------------------------
    # Scatter Plot (with predicted car)
    # -------------------------------
    fig = px.scatter(
        df, 
        x="engine_volume", 
        y="price",
        color="fuel_type",
        size='levy',
        hover_name="model",
        hover_data=[
            "manufacturer", "prod._year", "mileage_km", "leather_interior",
            'category','gear_box_type','drive_wheels','doors','wheel','color',
            'cylinders','airbags'
        ],
        title="Engine Volume vs. Price Colored by Fuel Type",
        labels={"engine_volume": "Engine Volume (L)", "price": "Price ($)"}
    )

    # Highlight the predicted car
    fig.add_scatter(
        x=[engine_volume_choice],
        y=[predicted_price],
        mode='markers',
        marker=dict(color='red', size=15),
        name='Predicted Car'
    )

    fig.update_layout(width=1200, height=700)
    st.plotly_chart(fig)

else:
    st.error("API request failed. Please check the Space.")