import streamlit as st
import cv2
from PIL import Image
import numpy as np
from ultralytics import YOLO

# Load YOLO models
model_sawit = YOLO("best_yolov8.pt")  # Ganti dengan path model YOLOv8 untuk pohon sawit
model_apel = YOLO("best_apple.pt")  # Ganti dengan path model YOLOv8 untuk apel

# Sidebar menu
menu = st.sidebar.selectbox("Pilih Aplikasi", ["Deteksi Pohon Sawit", "Deteksi Warna Apel"])

if menu == "Deteksi Pohon Sawit":
    st.title("Deteksi Pohon Sawit dengan YOLOv8")
    confidence_threshold = st.slider("Confidence Threshold", min_value=0.0, max_value=1.0, value=0.5, step=0.05)

    uploaded_file = st.file_uploader("Upload gambar", type=["jpg", "jpeg", "png"])

    if uploaded_file is not None:
        # Baca gambar
        image = Image.open(uploaded_file)
        img_array = np.array(image)

        # Jalankan deteksi dengan pengaturan confidence threshold
        results = model_sawit(img_array, conf=confidence_threshold)

        # Tambahkan bounding box dan nomor urut dengan background warna
        for i, box in enumerate(results[0].boxes):
            x1, y1, x2, y2 = map(int, box.xyxy[0])
            # Gambar bounding box
            cv2.rectangle(img_array, (x1, y1), (x2, y2), (0, 0, 255), 2)  # Gunakan warna merah

            # Tambahkan nomor urut dengan background
            text = f"{i+1}"
            font = cv2.FONT_HERSHEY_SIMPLEX
            font_scale = 0.9
            thickness = 2

            # Ukuran teks dan lokasi
            (text_width, text_height), baseline = cv2.getTextSize(text, font, font_scale, thickness)
            text_x, text_y = x1, y1 - 10  # Posisi teks di atas kotak
            rect_x1, rect_y1 = text_x, text_y - text_height - 5
            rect_x2, rect_y2 = text_x + text_width + 10, text_y + baseline - 5

            # Gambar background untuk teks
            cv2.rectangle(img_array, (rect_x1, rect_y1), (rect_x2, rect_y2), (0, 0, 255), -1)  # Background merah
            cv2.putText(img_array, text, (text_x, text_y), font, font_scale, (255, 255, 255), thickness)  # Teks putih

        # Tampilkan hasil
        st.image(img_array, caption=f"Total objek terdeteksi: {len(results[0].boxes)}", use_column_width=True)

elif menu == "Deteksi Warna Apel":
    st.title("Deteksi Warna Apel dengan YOLOv8")
    st.subheader("Unggah gambar apel untuk mendeteksi dan menampilkan hasil crop sesuai warna")

    # Upload gambar
    uploaded_file = st.file_uploader("Unggah gambar", type=["jpg", "jpeg", "png"])

    if uploaded_file is not None:
        # Baca gambar dari upload
        file_bytes = np.asarray(bytearray(uploaded_file.read()), dtype=np.uint8)
        image = cv2.imdecode(file_bytes, cv2.IMREAD_COLOR)

        # Deteksi objek menggunakan YOLOv8
        results = model_apel(image)[0]  # Ambil hasil prediksi pertama

        # Tampilkan gambar asli
        st.image(cv2.cvtColor(image, cv2.COLOR_BGR2RGB), caption="Gambar Asli", use_column_width=True)

        st.write("### Hasil Crop:")
        for i, box in enumerate(results.boxes):
            cls = int(box.cls)  # Indeks kelas
            confidence = box.conf.item()  # Tingkat kepercayaan
            class_name = model_apel.names[cls]  # Nama kelas (e.g., yellow, green, red)

            # Bounding box koordinat
            x1, y1, x2, y2 = map(int, box.xyxy[0].tolist())
            cropped_image = image[y1:y2, x1:x2]

            # Konversi ke RGB untuk ditampilkan di Streamlit
            cropped_image_rgb = cv2.cvtColor(cropped_image, cv2.COLOR_BGR2RGB)

            # Tampilkan hasil crop
            st.image(cropped_image_rgb, caption=f"{class_name} ({confidence:.2f})", use_column_width=False)