import tensorflow as tf
from tensorflow.keras.models import load_model
import streamlit as st
from PIL import Image
import numpy as np
import pandas as pd

# Load Sequential Model
model_cnn = load_model('car_model.h5')

# Load DataFrame
income_data = {
    'pekerjaan': ["Staff - Fresh Graduate", "Asisten Manajer", "Manajer", "General Manajer", "Kepala Divisi atau Direksi"],
    'rentang_usia': [(21, 25), (26, 29), (29, 35), (36, 42), (42, 100)],
    'rentang_gaji': [(2000000, 4000000), (5000000, 8000000), (10000000, 15000000), (20000000, 30000000), (50000000, 100000000)]
}

data = {
    'predicted_label': [0, 1, 2, 3],
    'jenis_mobil': ["Alphard", "Innova 2015-2022", "Innova Zennix", "New Veloz"],
    'merek_mobil': ["Toyota", "Toyota", "Toyota", "Toyota"],
    'harga_mobil_min': [1356100000, 215000000, 425000000, 288000000],
    'harga_mobil_max': [1950000000, 480000000, 610000000, 350000000]}

# Create DataFrame
df_car_info = pd.DataFrame(data)
df_income_info = pd.DataFrame(income_data)


def calculate_cicilan(predicted_label):
    # Get car info based on predicted_label
    car_info = df_car_info[df_car_info['predicted_label'] == predicted_label].iloc[0]

    # Calculate cicilan_per_bulan
    harga_mobil_avg = (car_info['harga_mobil_min'] + car_info['harga_mobil_max']) / 2
    cicilan_per_bulan = harga_mobil_avg / 60  # 5 tahun (60 bulan)

    # Find the matching gaji range based on cicilan_per_bulan
    matching_income = df_income_info[
    (cicilan_per_bulan / 0.3 >= df_income_info['rentang_gaji'].apply(lambda x: x[0])) &
    (cicilan_per_bulan / 0.3 <= df_income_info['rentang_gaji'].apply(lambda x: x[1]))]

    if not matching_income.empty:
        rentang_usia = matching_income['rentang_usia'].iloc[0]
        rentang_gaji_min = int(matching_income['rentang_gaji'].iloc[0][0])
        rentang_gaji_max = int(matching_income['rentang_gaji'].iloc[0][1])
        pekerjaan = matching_income['pekerjaan'].iloc[0]

        return {
            'Jenis Mobil': car_info['jenis_mobil'],
            'Merek Mobil': car_info['merek_mobil'],
            'Rentang Usia': f"{rentang_usia[0]}-{rentang_usia[1]} Tahun",
            'Rentang Gaji': f"Rp {rentang_gaji_min} - {rentang_gaji_max}",
            'Pekerjaan': pekerjaan
        }

    else:
        return None

def run():
    st.title('Toyota Car Prediction')
    st.subheader('This Prediction use for provide passanger information from his car')
    st.subheader('Only Uploaded Image New Alphard, Innova 2015 - 2022, Innova Zennix, New Veloz')
    st.image("Yow.jpeg")
    st.markdown('---')
    
    st.subheader('Car Prediction')
    upload = st.file_uploader("Please upload an only Toyota Car image for prediction", type=["jpg"])
    
    if upload is not None:
        img = Image.open(upload)
        img = img.resize((224, 224)).convert("RGB")
        img_array = np.array(img)
        img_array = img_array / 255.0
        img_array = np.expand_dims(img_array, axis=0)
    
        # Model inference
        images = np.vstack([img_array])
        classes = model_cnn.predict(images)
    
        # Memilih label kelas
        predicted_label = np.argmax(classes)
    
        # Assuming df_car_info and df_income_info are available
        result = calculate_cicilan(predicted_label)
    
        st.write("Hasil Prediksi:")
        if result is not None:
            for key, value in result.items():
                st.write(f"{key}: {value}")
        else:
            st.write("Maaf, tidak ditemukan.")

if __name__ == '__main__':
    run()