src/streamlit_app.py

import streamlit as st
import pickle
import numpy as np
from PIL import Image
import tensorflow as tf
from tensorflow.keras.applications.resnet50 import ResNet50, preprocess_input
from tensorflow.keras.preprocessing import image
from tensorflow.keras.layers import GlobalMaxPool2D
from sklearn.neighbors import NearestNeighbors
from numpy.linalg import norm
import os
import zipfile

# Sayfa başlığını ayarlıyoruz
st.title("Moda Ürün Öneri Sistemi")

BASE_DIR = "/app/src"

# images.zip dosyasını açıyoruz
if not os.path.exists(f"{BASE_DIR}/images"):
    os.makedirs(f"{BASE_DIR}/images", exist_ok=True)

    with zipfile.ZipFile(f"{BASE_DIR}/images.zip", "r") as zip_ref:
        zip_ref.extractall(f"{BASE_DIR}/images")

# Feature ve dosya yollarını yüklüyoruz
features = pickle.load(open(f"{BASE_DIR}/Images_features.pkl", "rb"))
filenames = pickle.load(open(f"{BASE_DIR}/filenames.pkl", "rb"))


# ResNet50 modelini yüklüyoruz
base_model = ResNet50(weights="imagenet", include_top=False, input_shape=(224, 224, 3))

base_model.trainable = False

model = tf.keras.Sequential([base_model, GlobalMaxPool2D()])

# Yüklenen resmi kaydetmek için fonksiyon yazıyoruz
def save_uploaded_file(uploaded_file):
    try:
        with open(os.path.join("uploads", uploaded_file.name), "wb") as f:
            f.write(uploaded_file.getbuffer())
        return 1
    except:
        return 0
    
# Yüklenen resmi vektöre çeviriyoruz
def feature_extraction(img_path, model):
    img = image.load_img(img_path, target_size=(224, 224))
    img_array = image.img_to_array(img)
    expanded_img = np.expand_dims(img_array, axis=0)
    preprocessed_img = preprocess_input(expanded_img)
    result = model.predict(preprocessed_img).flatten()
    normalized_result = result / norm(result)

    return normalized_result    

# En benzer ürünleri bulan fonksiyonu yazıyoruz
def recommend(features, feature_list):
    neighbors = NearestNeighbors(n_neighbors=6,algorithm="brute",metric="euclidean")

    neighbors.fit(feature_list)

    distances, indices = neighbors.kneighbors([features])

    return indices

# Upload klasörü yoksa oluşturuyoruz
if not os.path.exists("uploads"):
    os.makedirs("uploads")


# Kullanıcıdan ürün görseli alıyoruz
uploaded_file = st.file_uploader("Bir ürün görseli yükleyin",type=["jpg", "jpeg", "png"])


# Görsel yüklendiyse öneri sistemini çalıştırıyoruz
if uploaded_file is not None:
    if save_uploaded_file(uploaded_file):

        display_image = Image.open(uploaded_file)

        st.image(display_image, caption="Yüklenen Görsel", width=300)

        input_features = feature_extraction(
            os.path.join("uploads", uploaded_file.name),
            model
        )

        indices = recommend(input_features, features)

        st.subheader("Benzer Ürün Önerileri")

        cols = st.columns(5)

        for i, col in enumerate(cols):
            with col:
                img_path = os.path.join(BASE_DIR, filenames[indices[0][i + 1]])
        
                if os.path.exists(img_path):
                    st.image(img_path)
                else:
                    st.warning(f"Görsel bulunamadı: {img_path}")