app.py

# app.py (Final Version: SAHAR AUTO, Custom Design, and Manufacturer Dropdown)

import streamlit as st
import pickle
import pandas as pd
import numpy as np
import os
import sys

# --- CONFIGURATION ---
PKL_FILE_PATH = "model.pkl"
LOGO_PATH = r"F:\Car_price\sahar_logo.png" 
COMPANY_NAME = "SAHAR AUTO"
# --- CUSTOM CSS FOR STYLING ---
st.markdown(
    """
    <style>
    /* 1. General App Background and Text Color */
    .stApp {
        background-color: #1a1a1a; 
        color: white; 
    }
    
    /* 2. Headers and General Text (White) */
    h1, h2, h3, h4, h5, h6, p, label, .stMarkdown, .streamlit-expanderHeader {
        color: white !important;
    }
    
    /* 3. Button Styling (Red) */
    .stButton > button {
        background-color: #ff4b4b; 
        color: white; 
        border-radius: 5px;
        padding: 10px 20px;
        font-size: 1.1em;
        font-weight: bold;
        border: none;
        transition: background-color 0.3s ease;
    }
    .stButton > button:hover {
        background-color: #e03a3a; 
    }
    
    /* 4. Custom Banner Text */
    .banner-text-large {
        font-size: 3.5em; 
        font-weight: bold;
        color: white; 
        margin-bottom: 0px;
        line-height: 1.2;
    }
    .banner-text-red {
        font-size: 3.5em;
        font-weight: bold;
        color: #ff4b4b; 
        margin-top: 0px;
        line-height: 1.2;
    }
    
    /* 5. Input Fields Styling */
    .stSelectbox > div > div,
    .stNumberInput > div > div {
        background-color: #333333; 
        color: white; 
    }
    
    /* 💡 تغيير لون النص داخل حقول إدخال الأرقام إلى الأحمر */
    div[data-testid*="stNumberInput"] input {
        color: #ff4b4b !important; 
        font-weight: bold;
        background-color: #333333; 
    }
    
    /* Other input text remains white */
    div[data-testid*="stTextInput"] input,
    div[data-testid*="stSelectbox"] div[data-testid*="stInput"],
    div[data-testid*="stRadio"] label,
    .stSlider label {
        color: white !important;
    }

    /* 6. Slider Coloring */
    .stSlider > div > div > div > div {
        background-color: #ff4b4b; 
    }
    .stSlider > div > div > div > div > div {
        background-color: #ff4b4b; 
    }
    
    .st-emotion-cache-eq8hrv {
        max-width: 1000px;
        padding-top: 2rem;
        padding-bottom: 2rem;
    }
    </style>
    """,
    unsafe_allow_html=True
)


# --- Load Model and Columns ---
@st.cache_resource
def load_model_data():
    """Loads model, all columns, and extracts manufacturer names from columns."""
    if not os.path.exists(PKL_FILE_PATH):
        st.error(f"Error: Model file not found at: {PKL_FILE_PATH}")
        sys.exit(1)
        
    try:
        with open(PKL_FILE_PATH, "rb") as f:
            saved = pickle.load(f)
            
        all_columns = saved["columns"]
        
        # 💡 يتم استخراج جميع الشركات المصنعة، لا حاجة لإضافة toyota يدوياً الآن
        manufacturer_cols = [col.replace('Maker_', '') for col in all_columns if col.startswith('Maker_')]
             
        return saved["model"], all_columns, manufacturer_cols
        
    except Exception as e:
        st.error(f"Error loading model: {e}")
        sys.exit(1)

model, training_columns, manufacturer_options = load_model_data()

# --- Prediction Function ---
def predict_price(input_data):
    """Processes input and predicts the price."""
    full_data = {col: 0 for col in training_columns}
    for key, value in input_data.items():
        if key in full_data:
            full_data[key] = value
    
    df = pd.DataFrame([full_data])
    df = df.reindex(columns=training_columns, fill_value=0)
    
    prediction = model.predict(df)[0]
    return np.round(prediction, 2)


# --- Streamlit APP LAYOUT ---
st.set_page_config(page_title=f"{COMPANY_NAME} Predictor", layout="wide")

# Logo and Headline 
col_logo, col_title = st.columns([1, 4])

with col_logo:
    if os.path.exists(LOGO_PATH):
        st.image(LOGO_PATH, width=150) 
    else:
        st.markdown(f"<h3 style='color:white;'>{COMPANY_NAME}</h3>", unsafe_allow_html=True)
        st.markdown(f"<p style='color:white; font-size:0.8em; margin-top:-10px;'>AUTO</p>", unsafe_allow_html=True)

with col_title:
    st.markdown("<p class='banner-text-large'>LET'S GET YOU</p>", unsafe_allow_html=True)
    st.markdown("<p class='banner-text-red'>ON THE ROAD</p>", unsafe_allow_html=True)

st.markdown("---") 

st.header("Enter Car Specifications")

# Input Fields (main numeric)
col1, col2 = st.columns(2)

with col1:
    horsepower = st.slider("1. Horsepower", 50, 200, 100)
    enginesize = st.slider("2. Engine Size", 70, 300, 120)
    
with col2:
    curbweight = st.number_input("3. Curbweight", 1500, 4500, 2500)
    carwidth = st.number_input("4. Car Width", 60.0, 75.0, 68.0)
    
st.markdown("---")
st.subheader("Categorical Details")

# Input Fields (categorical)
carbody_options = ['sedan', 'hatchback', 'wagon', 'hardtop', 'convertible']
fueltype_options = ['gas', 'diesel']
drivewheel_options = ['fwd', 'rwd', '4wd'] 

# Create 4 columns for categorical inputs including Maker
col_cat0, col_cat1, col_cat2, col_cat3 = st.columns(4)

with col_cat0:
    selected_maker = st.selectbox("5. Manufacturer", sorted(manufacturer_options)) 
    
with col_cat1:
    selected_carbody = st.selectbox("6. Car Body Type", carbody_options)
    
with col_cat2:
    selected_fueltype = st.radio("7. Fuel Type", fueltype_options)
    
with col_cat3:
    selected_drivewheel = st.radio("8. Drive Wheel", drivewheel_options)


# --- Prepare Input Data for Prediction ---
input_data_for_prediction = {
    'horsepower': horsepower,
    'enginesize': enginesize,
    'curbweight': curbweight,
    'carwidth': carwidth,
    
    # --- One-Hot Encoded Features ---
    **{col: 0 for col in training_columns if '_' in col}
}

# Set selected One-Hot features to 1
# 💡 Manufacturer: Set the selected maker column to 1 
maker_column_name = f'Maker_{selected_maker}'
if maker_column_name in training_columns:
    input_data_for_prediction[maker_column_name] = 1

# Other One-Hot features
if f'carbody_{selected_carbody}' in training_columns:
    input_data_for_prediction[f'carbody_{selected_carbody}'] = 1

if f'fueltype_{selected_fueltype}' in training_columns:
    input_data_for_prediction[f'fueltype_{selected_fueltype}'] = 1

if f'drivewheel_{selected_drivewheel}' in training_columns:
    input_data_for_prediction[f'drivewheel_{selected_drivewheel}'] = 1

# --- Prediction Button ---
st.markdown("---")
if st.button("REQUEST A QUOTE"):
    try:
        input_data_for_prediction = {k: v for k, v in input_data_for_prediction.items() if k is not None}
        
        predicted_price = predict_price(input_data_for_prediction)
        st.success(f"## Your Estimated Car Price: **${predicted_price:,.2f}**")
    except Exception as e:
        st.error("Please ensure all fields are filled correctly.")
        st.error(f"Detailed Error: {e}")