Update app.py

app.py

@@ -6,18 +6,17 @@ from tensorflow import keras
 import sys
 import io
 from ultralytics import YOLO
-import time
 from transformers import pipeline
 
+# Initialize image-to-text pipeline
 image_to_text_pipe = pipeline("image-to-text", model="naver-clova-ix/donut-base")
 
-
 # Set the default encoding to utf-8
 sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
 
 app = Flask(__name__)
 
-# Load your pre-trained vegetable classification model
+# Pre-defined vegetable and name classes
 vegetables = [
     "banana", "beans broad", "beans cluster", "beans haricot", "beetroot",
     "bitter guard", "bottle guard", "brinjal long", "brinjal[purple]", "cabbage",
@@ -27,8 +26,9 @@ vegetables = [
 ]
 names= ['Complan Classic Creme', 'Complan Kesar Badam', 'Complan Nutrigro Badam Kheer', 'Complan Pista Badam', 'Complan Royal Chocolate', 'Complan Royale Chocolate', 'EY AAAM TULSI TURMERIC FACEWASH50G', 'EY ADVANCED GOLDEN GLOW PEEL OFF M- 50G', 'EY ADVANCED GOLDEN GLOW PEEL OFF M- 90G', 'EY EXF WALNUT SCRUB AYR 200G', 'EY HALDICHANDAN FP HF POWDER 25G', 'EY HYD-EXF WALNT APR SCRUB AYR100G', 'EY HYDR - EXF WALNUT APRICOT SCRUB 50G', 'EY NAT GLOW ORANGE PEEL OFF AY 90G', 'EY NATURALS NEEM FACE WASH AY 50G', 'EY RJ CUCUMBER ALOEVERA FACEPAK50G', 'EY TAN CHOCO CHERRY PACK 50G', 'EY_SCR_PURIFYING_EXFOLTNG_NEEM_PAPAYA_50G', 'Everyuth Naturals Body Lotion Nourishing Cocoa 200ml', 'Everyuth Naturals Body Lotion Rejuvenating Flora 200ml', 'Everyuth Naturals Body Lotion Soothing Citrus 200ml', 'Everyuth Naturals Body Lotion Sun Care Berries SPF 15 200ml', 'Glucon D Nimbu Pani', 'Glucon D Nimbu Pani 1-KG', 'Glucon D Regular', 'Glucon D Regular 1-KG', 'Glucon D Regular 2-KG', 'Glucon D Tangy orange', 'Glucon D Tangy orange 1-KG', 'Nutralite ACHARI MAYO 300g-275g-25g-', 'Nutralite ACHARI MAYO 30g', 'Nutralite CHEESY GARLIC MAYO 300g-275g-25g-', 'Nutralite CHEESY GARLIC MAYO 30g', 'Nutralite CHOCO SPREAD CALCIUM 275g', 'Nutralite DOODHSHAKTHI PURE GHEE 1L', 'Nutralite TANDOORI MAYO 300g-275g-25g-', 'Nutralite TANDOORI MAYO 30g', 'Nutralite VEG MAYO 300g-275g-25g-', 'Nycil Prickly Heat Powder', 'SUGAR FREE GOLD 500 PELLET', 'SUGAR FREE GOLD POWDER 100GM', 'SUGAR FREE GOLD SACHET 50', 'SUGAR FREE GOLD SACHET 50 SUGAR FREE GOLD SACHET 50', 'SUGAR FREE GRN 300 PELLET', 'SUGAR FREE NATURA 500 PELLET', 'SUGAR FREE NATURA DIET SUGAR', 'SUGAR FREE NATURA DIET SUGAR 80GM', 'SUGAR FREE NATURA SACHET 50', 'SUGAR FREE NATURA SWEET DROPS', 'SUGAR FREE NATURA_ POWDER_CONC_100G', 'SUGAR FREE_GRN_ POWDER_CONC_100G', 'SUGARLITE POUCH 500G']
 
+
+# Load the YOLO models
 model = YOLO('fresh_model.pt')
-# Load the YOLO model
 yolo_model = YOLO('another_model.pt')  # Adjust the path as needed
 
 @app.route('/')
@@ -38,27 +38,20 @@ def index():
 @app.route('/predict', methods=['POST'])
 def predict():
     try:
-        # Extract and decode image data
         data = request.json
         image_data = data['image'].split(',')[1]  # Remove data:image/jpeg;base64,
 
-        # Decode the base64 string into a NumPy array
         nparr = np.frombuffer(base64.b64decode(image_data), np.uint8)
-
-        # Convert the NumPy array into an OpenCV image (BGR format)
         image = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
 
-        # Perform inference using the YOLO model
-        results = model(image)  # Pass the image to the YOLO model
-
-        # Extract the detected classes
+        results = model(image)  # Perform inference using the YOLO model
         class_indices = results[0].boxes.cls.cpu().numpy().astype(int)  # Get the class indices
-        unique_classes = np.unique(class_indices)  # Get unique classes detected
 
-        # Map class indices to labels if needed
-        # Assuming you have a mapping for YOLO classes
-        detected_classes = [str(cls) for cls in unique_classes]
-        detected_classes=vegetables[detected_classes[0]]
+        if len(class_indices) == 0:
+            return jsonify({'predictions': []})  # No classes detected
+
+        unique_classes = np.unique(class_indices)  # Get unique classes detected
+        detected_classes = [vegetables[cls] for cls in unique_classes if cls < len(vegetables)]  # Ensure valid index
 
         return jsonify({'predictions': detected_classes})
     except Exception as e:
@@ -67,57 +60,41 @@ def predict():
 @app.route('/image_to_text', methods=['POST'])
 def image_to_text():
     try:
-        # Extract and decode image data
         data = request.json
         image_data = data['image'].split(',')[1]  # Remove data:image/jpeg;base64,
 
-        # Decode the base64 string into a NumPy array
         nparr = np.frombuffer(base64.b64decode(image_data), np.uint8)
-
-        # Convert the NumPy array into an OpenCV image (BGR format)
         image = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
 
-        # Use PaddleOCR to detect and extract text
-        result = ocr.ocr(image, cls=True)  # Run OCR on the image
-
-        # Check if any text is detected
-        is_text_detected = any(word_info[1][0].strip() for line in result for word_info in line)
+        # Use your OCR model here (replace this with the actual call)
+        result = image_to_text_pipe(image)  # Replace with actual OCR processing
 
+        is_text_detected = bool(result)  # Check if any text was detected
         return jsonify({'text_detected': is_text_detected})
     except Exception as e:
         return jsonify({'error': str(e)}), 500
 
-
-
-
 @app.route('/detect', methods=['POST'])
 def detect():
     try:
-        # Extract and decode image data
         data = request.json
         image_data = data['image'].split(',')[1]  # Remove data:image/jpeg;base64,
 
-        # Decode the base64 string into a NumPy array
         nparr = np.frombuffer(base64.b64decode(image_data), np.uint8)
-
-        # Convert the NumPy array into an OpenCV image (BGR format)
         image = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
 
-        # Perform inference using the YOLO model
-        results = yolo_model(image)  # Pass the image to the YOLO model
+        results = yolo_model(image)  # Perform inference using the YOLO model
+        class_indices = results[0].boxes.cls.cpu().numpy().astype(int)
 
-        # Extract the detected classes
-        class_indices = results[0].boxes.cls.cpu().numpy().astype(int)  # Get the class indices
-        unique_classes = np.unique(class_indices)  # Get unique classes detected
+        if len(class_indices) == 0:
+            return jsonify({'predictions': []})  # No classes detected
 
-        # Map class indices to labels if needed
-        # Assuming you have a mapping for YOLO classes
-        detected_classes = [str(cls) for cls in unique_classes]
-        detected_classes=names[detected_classes[0]]
+        unique_classes = np.unique(class_indices)  # Get unique classes detected
+        detected_classes = [names[cls] for cls in unique_classes if cls < len(names)]  # Ensure valid index
 
         return jsonify({'predictions': detected_classes})
     except Exception as e:
         return jsonify({'error': str(e)}), 500
 
 if __name__ == '__main__':
-    app.run(host='0.0.0.0', port=7860, debug=True)
+    app.run(host='0.0.0.0', port=7860, debug=True)