fix preprocessing steps

app.py

@@ -1,40 +1,35 @@
 import os
+import sys
+import numpy as np
 from flask import Flask, render_template, request, jsonify, send_from_directory
+
+import tensorflow as tf
 from tensorflow.keras.models import load_model
-from tensorflow.keras.preprocessing import image
-import numpy as np
-from PIL import Image
-import io
-import sys # Added for logging
+from tensorflow.keras.applications.resnet50 import preprocess_input
 
 app = Flask(__name__, template_folder='templates', static_folder='static')
 
 # 1. Load Model (Safe Mode)
 MODEL_PATH = 'model.h5'
+model = None
+
 try:
-    # compile=False prevents crashing if the model was trained on a different TF version
     model = load_model(MODEL_PATH, compile=False)
     print("SUCCESS: Model loaded!", file=sys.stderr)
 except Exception as e:
     print(f"CRITICAL ERROR: Failed to load model. {e}", file=sys.stderr)
 
-CLASS_NAMES = ['NonDemented', 'VeryMildDemented', 'MildDemented', 'ModerateDemented']
+CLASS_NAMES = ['Non Demented', 'Very Mild Demented', 'Mild Demented', 'Moderate Demented']
 
 def prepare_image(img_bytes):
     try:
-        img = Image.open(io.BytesIO(img_bytes))
+        image = tf.io.decode_image(img_bytes, channels=3, expand_animations=False)
+        image = tf.image.resize(image, [224, 224])
+        image = tf.cast(image, tf.float32)
+        image = tf.expand_dims(image, axis=0)
+        image = preprocess_input(image)
         
-        # FIX 1: Convert to RGB to remove Alpha channel (transparency) if present
-        if img.mode != 'RGB':
-            img = img.convert('RGB')
-            
-        # FIX 2: Resize
-        img = img.resize((224, 224))
-        
-        img_array = image.img_to_array(img)
-        img_array = np.expand_dims(img_array, axis=0)
-        #img_array = img_array / 255.0
-        return img_array
+        return image
     except Exception as e:
         print(f"Error processing image: {e}", file=sys.stderr)
         raise e
@@ -51,48 +46,46 @@ def serve_assets(filename):
 
 @app.route('/<path:filename>')
 def serve_root_files(filename):
-    if os.path.exists(os.path.join('static', filename)):
-        return send_from_directory('static', filename)
-    return "File not found", 404
+    return send_from_directory('static', filename)
 
-# Predict Route (Accepts any file key)
+# Predict Route
 @app.route('/api/classify', methods=['POST'])
 def predict():
+    if model is None:
+        return jsonify({'error': 'Model not loaded correctly'}), 500
+
     if not request.files:
         return jsonify({'error': 'No file uploaded'}), 400
     
-    # Grab the first file
     file = next(iter(request.files.values()))
     
     try:
-        # Debug print
-        print(f"Received file: {file.filename}", file=sys.stderr)
+        file_bytes = file.read()
         
-        processed_img = prepare_image(file.read())
+        processed_img = prepare_image(file_bytes)
         
-        print("Image processed. predicting...", file=sys.stderr)
-        prediction = model.predict(processed_img)
+        prediction_tensor = model(processed_img, training=False)
+        
+        prediction = prediction_tensor.numpy()
         
         class_index = np.argmax(prediction)
         confidence = float(np.max(prediction))
         
         result_class = CLASS_NAMES[class_index]
-        print(f"Prediction success: {result_class}", file=sys.stderr)
+        
+        print(f"Prediction: {result_class} ({confidence:.2%})", file=sys.stderr)
         
         result = {
-            # Frontend expects 'result', not 'class'
             'result': result_class,
-            
-            # Frontend expects a raw Number (float), NOT a string with '%'
-            'confidence': float(confidence) 
+            'confidence': confidence,
+            'index': int(class_index)
         }
         
         return jsonify(result)
         
     except Exception as e:
-        # This will show up in the Logs tab!
         print(f"PREDICTION CRASHED: {str(e)}", file=sys.stderr)
         return jsonify({'error': str(e)}), 500
 
 if __name__ == '__main__':
-    app.run(host='0.0.0.0', port=7860)
+    app.run(host='0.0.0.0', port=7860, debug=False)