import streamlit as st
import tensorflow as tf
import numpy as np
from PIL import Image

from tensorflow.keras.applications import MobileNetV2
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, GlobalAveragePooling2D, Rescaling


st.title("Üzüm Hastalığı Sınıflandırma - Transfer Learning")


# Colab'da kullandığımız MobileNetV2 taban modelini kuruyoruz.

base_model = MobileNetV2(
    weights="imagenet",
    include_top=False,
    input_shape=(224, 224, 3)
)

base_model.trainable = False


tl_model = Sequential()

tl_model.add(Rescaling(1./255, input_shape=(224, 224, 3)))

tl_model.add(base_model)

tl_model.add(GlobalAveragePooling2D())

tl_model.add(Dense(128, activation="relu"))

tl_model.add(Dropout(0.3))

tl_model.add(Dense(4, activation="softmax"))


# Tam model yerine sadece ağırlıkları yüklüyoruz.

tl_model.load_weights("src/grape_transfer.weights.h5")


class_names = ['Black Rot', 'ESCA', 'Healthy', 'Leaf Blight']


uploaded_file = st.file_uploader(
    "Bir üzüm yaprağı resmi yükleyin",
    type=["jpg", "jpeg", "png"]
)


if uploaded_file is not None:

    img = Image.open(uploaded_file).convert("RGB")

    st.image(img, caption="Yüklenen Görsel", use_container_width=True)

    img = img.resize((224, 224))

    img_array = np.array(img)

    img_array = np.expand_dims(img_array, axis=0)

    prediction = tl_model.predict(img_array)

    predicted_class = class_names[np.argmax(prediction)]

    confidence = np.max(prediction) * 100

    st.subheader("Tahmin Sonucu")
    st.write("Sınıf:", predicted_class)
    st.write("Güven Oranı:", round(confidence, 2), "%")