labelmeWithAI/labelme/ai/barcode_decode.py

import imgviz
from qtpy import QtCore
from qtpy import QtGui
from qtpy import QtWidgets
import labelme.ai
import labelme.utils
from labelme import QT5
from labelme.logger import logger
from labelme.shape import Shape
import collections
import threading
import numpy as np
import openvino as ov
import os.path as osp
import cv2
import os
import time
from labelme.utils import img_qt_to_arr
from labelme.utils import load_barcode_dict

class CodeSet:
    NONE = 0
    A = 1
    B = 2
    C = 3

# class Normalize:
#     def __init__(self, mean=(0.45, 0.45, 0.45), std=(0.25, 0.25, 0.25)):
#         if not (isinstance(mean, (list, tuple)) and isinstance(std, (list, tuple))):
#             raise ValueError("mean and std should be of type list or tuple.")
#         self.mean = np.array(mean, dtype=np.float32)
#         self.std = np.array(std, dtype=np.float32)

#         # Reshape for broadcasting to apply mean and std across the spatial dimensions of an image
#         self.mean = self.mean.reshape((1, 1, 3))
#         self.std = self.std.reshape((1, 1, 3))

#     def __call__(self, img):
#         img = img.astype(np.float32) / 255.0  # Scale pixel values to [0, 1]
#         img = (img - self.mean) / self.std  # Normalize
#         return img

class BarcodeDecodeModel:
    def __init__(self, decoding_model_path=None):
        self.ie = ov.Core()
        self.pixmap = QtGui.QPixmap()
        #Load Decoding model if provided
        self.decoding_net = None
        self.decoding_sess = None
        self._characters = load_barcode_dict() 
        if decoding_model_path:
            self.decoding_net = self.ie.read_model(model=decoding_model_path)
            self.decoding_sess = self.ie.compile_model(model=self.decoding_net, device_name="CPU")

        self.decoding_input_shape = (1, 3, 32, 512)
        # self.normalize = Normalize()  # Normalization instance
        self._lock = threading.Lock()
        self._image_embedding_cache = collections.OrderedDict()
        self._max_cache_size = 10
        self.pixmap = QtGui.QPixmap()

    # def set_pixmap(self, pixmap: QtGui.QPixmap):
    #     """
    #     Set the QPixmap object for decoding.
    #     Args:
    #         pixmap (QtGui.QPixmap): The QPixmap object containing the image.
    #     """
    #     if pixmap is None or pixmap.isNull():
    #         raise ValueError("Invalid QPixmap provided.")
    #     self.pixmap = pixmap
    #     logger.debug("Pixmap set successfully in BarcodeDecodeModel.")

    def preprocess_image(self, img):
        logger.debug(f"Preprocessing image for decoding: {img.shape}")
        img = cv2.resize(
            img,
            (self.decoding_input_shape[3], self.decoding_input_shape[2]),
            interpolation=cv2.INTER_NEAREST
        )
        img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
        img = img.astype("float32") / 255.0
        mean = np.array([0.45, 0.45, 0.45], dtype=np.float32)
        std = np.array([0.25, 0.25, 0.25], dtype=np.float32)
        img = (img - mean) / std
        img = img.transpose(2, 0, 1)
        img = img.astype("float32")
        input_tensor = np.expand_dims(img, 0)
        logger.debug(f"Processed image shape: {input_tensor.shape}")

        return input_tensor
    
    def decode_from_points(self, points, detection_idx, original_image):
        """
        Decodes the cropped image based on points and returns the decoded text.
        Args:
            points (list): List of points defining the bounding box.
            pixmap (QPixmap): Original image pixmap to crop from.
        Returns:
            str: Decoded text from the decoding model.
        """
        try:
            # Convert scaled_points to a numpy array
            polygon = np.array(points, dtype=np.int32)

            # Create a mask of the same size as the original image
            # original_image = labelme.utils.img_qt_to_arr(self.pixmap.toImage())
            # cv2.imwrite(f"original_image{detection_idx + 1}.png", original_image)

            # mask = np.zeros(original_image.shape[:2], dtype=np.uint8)
            # cv2.fillPoly(mask, [polygon], 255)  # Fill the polygon with white

            # Apply the mask to the original image
            # masked_image = cv2.bitwise_and(original_image, original_image, mask=mask)
            # cv2.imwrite(f"masked_image{detection_idx + 1}.png", masked_image)

            # Get the bounding rectangle of the polygon to crop the ROI
            x, y, w, h = cv2.boundingRect(polygon)
            # cropped_image_dec = masked_image[y:y+h, x:x+w]
            # cv2.imwrite(f"cropped_exact_{detection_idx + 1}.png", cropped_image_dec)
            # logger.debug(f"cropped_exact image saved at  {detection_idx + 1}.")

            src_points = np.float32(points)
            

            # Calculate the width and height of the barcode based on scaled_points
            width = int(np.ceil(np.linalg.norm(src_points[0] - src_points[1])))
            print(f" width:{ width}")
            height = int(np.ceil(np.linalg.norm(src_points[1] - src_points[2])))
            print(f" height:{ height}")

            # Correct width/height if needed
            if width < height:
                width, height = height, width
                # Reorder src_points to ensure the transformation aligns the longer side to the width
                src_points = np.float32([
                    src_points[1],  # Top-left becomes top-right
                    src_points[2],  # Top-right becomes bottom-right
                    src_points[3],  # Bottom-right becomes bottom-left
                    src_points[0]   # Bottom-left becomes top-left
                ])
            print(f"src_points:{src_points}")

            # Define destination points for the flattened barcode
            dst_points = np.float32([
                [0, 0],
                [width, 0],
                [width, height],
                [0, height]
            ])

            # Calculate the perspective transformation matrix
            M = cv2.getPerspectiveTransform(src_points, dst_points)            
            print(f" M:{ M}")
            # t_start = time.perf_counter()
            
            # Apply the perspective transformation
            aligned_barcode = cv2.warpPerspective(original_image, M, (width, height), flags=cv2.INTER_LANCZOS4)
            
            # t_end = time.perf_counter()
            # print(
            #     f"[TIMING] Total crop+align time: {(t_end - t_start) * 1000:.3f} ms | "
            #     f"aligned_size=({height},{width})"
            # )

            # Save the aligned barcode image
            # cv2.imwrite(f"old_aligned_barcode_decoding_barcode_{detection_idx + 1}.png", aligned_barcode)
            # logger.debug(f"Aligned barcode saved at  {detection_idx + 1}.")

            # Normalize the image to scale pixel intensities to the range [0, 255]
            normalized_img = np.zeros(aligned_barcode.shape, aligned_barcode.dtype)
            cv2.normalize(aligned_barcode, normalized_img, 0, 255, cv2.NORM_MINMAX)
            logger.debug("Image normalized.")

            # # Save the cropped image
            cv2.imwrite(f"cropped_image__normalized{detection_idx + 1}.png",normalized_img)
            # logger.debug(f"Saved normalized image for decoding : {detection_idx + 1}")

            # Run decoding model
            # confidence = None 
            # Run decoding on the original image
            is_valid, decoded, decoded_text, msg, avg_conf, detection_idx = self.run_decoding(normalized_img, detection_idx)
            print(f"Valid: {is_valid}, detection_idx:{detection_idx + 1}, decoded:{decoded}, decoded_text:{decoded_text}, msg:{msg}, avg_conf:{avg_conf}")
            
            if not is_valid:
                logger.warning(f"Decoding failed for detection {detection_idx + 1}: {msg}. Retrying with 180° rotation")
                # Rotate image 180 degrees and retry
                rotated_image = cv2.rotate(normalized_img, cv2.ROTATE_180)
                is_valid, decoded, decoded_text, msg, avg_conf, detection_idx = self.run_decoding(rotated_image, detection_idx)
                
                if is_valid:
                    logger.debug(f"Valid: {is_valid}, detection_idx:{detection_idx + 1}, decoded:{decoded}, decoded_text:{decoded_text}, msg:{msg}, avg_conf:{avg_conf}")
                    return decoded
                else:
                    logger.warning(f"Decoding still failed after rotation for detection {detection_idx + 1}: {msg}")
                    return ""
            return decoded
        except Exception as e:
            logger.error(f"Error in decode_from_points: {e}")
            return "Error: Decoding failed"
        
    def run_decoding(self, normalized_img, detection_idx):
        """Helper to run decoding on the given image."""
        preprocessed_img = self.preprocess_image(
            normalized_img
        )

        # cv2.imwrite(f"cropped_image_decoding_normalized{detection_idx + 1}.png",normalized_img)

        decode_result = self.decoding_sess.infer_new_request({'x': preprocessed_img})
        output_tensor = decode_result['save_infer_model/scale_0.tmp_0']
        logger.debug(f"Output tensor shape: {output_tensor.shape}")
        # print(f"decode_result: {decode_result}")
        output_indices_batch = np.argmax(output_tensor, axis=2)
        output_probs_batch = np.max(output_tensor, axis=2)
        # print(f"output_probs_batch:{output_probs_batch}")
        # Decode text from indices

        def preprocess_output_indices(output_indices_batch, output_probs_batch):
            # Ensure it's a proper 2D numpy array
            if output_indices_batch is None or len(output_indices_batch) == 0:
                return False, "Empty output indices batch", None

            # print(f"output_indices_batch: {output_indices_batch}")
            first_row = output_indices_batch[0]
            first_row_probs = output_probs_batch[0]
            if first_row is None or len(first_row) == 0:
                return False, "Empty output indices", None

            sequence = first_row.tolist()
            probs = first_row_probs.tolist()

            # Step 1: Trim at first 0
            if 0 in sequence:
                zero_index = sequence.index(0)
                cropped_indices = sequence[:zero_index]
                cropped_probs = probs[:zero_index]
            else:
                cropped_indices = sequence
                cropped_probs = probs
            
            index_prob_pairs = list(zip(cropped_indices, cropped_probs))


            cropped = [x - 1 for x in cropped_indices]
            # cropped = cropped_indices
            # print(f"cropped: {cropped}")
            # Step 2: Check for invalid trailing 107/108 after 0
            if not any(val in (106, 107) for val in cropped):
                return False, "Invalid: missin stop code (106 or 107) before first 0"

            # Step 3: Truncate at second 108 if two 108s appear (EAN start/stop)
            if cropped.count(108) >= 2:
                print("got here")
                first_108 = cropped.index(108)
                # print(f"first_108:{first_108}")
                second_108 = cropped.index(108, first_108 + 1)
                # print(f"second_108:{second_108}")
                cropped = cropped[:second_108 + 1]

                # print(f"cropped: {cropped}")
                index_prob_pairs = index_prob_pairs[:second_108 + 1]
                # print(f": {index_prob_pairs}")
            # Step 4: Check start code validity
            start_code = cropped[0] if cropped else None
            if start_code not in [103, 104, 105, 107]:
                return False, f"Invalid start code: {start_code}"
            return True, (cropped, index_prob_pairs)
        decoded_text = ""
        status, result_or_error = preprocess_output_indices(output_indices_batch, output_probs_batch)
        # print(f"Raw barcode: {result_or_error}, status: {status}")
        
        if status == False:
            print(f"msg: {result_or_error}")
            decoded = "Decoding failed"
            decoded_text = ""
            msg = result_or_error
            avg_conf = 0.0
            return status, decoded, decoded_text, msg, avg_conf, detection_idx
        else:
            mapped_indices, conf_pairs = result_or_error
            avg_conf = round(np.mean([conf for (_, conf) in conf_pairs]), 2)
            # print(f"✅ Average confidence: {avg_conf:.3f}")
            is_valid, barcode_value, msg, predicted_ean_digits_print = self.validate_checksum(mapped_indices)
            # print(f"barcode_value: {barcode_value}, msg: {msg}, predicted_ean_digits_print: {predicted_ean_digits_print}, is_valid: {is_valid}")
            if not is_valid:
                decoded = "Decoding failed"
                decoded_text = ""
                msg = msg
                avg_conf = 0.0
                is_valid = is_valid
                # logger.warning(f"{is_valid}, readable: {decoded}, raw: {decoded_text}, conf: {avg_conf}, format: {msg}")
                return is_valid, decoded, decoded_text, msg, avg_conf, detection_idx
            else:
                if msg == "code128":
                    decoded_text = ''.join([self._characters[idx] for idx in barcode_value])
                    decoded = self.decode_code128(barcode_value)
                    logger.debug(f"✅ {is_valid}, readable: {decoded}, raw: {decoded_text}, conf: {avg_conf}, format: {msg}")
                else:
                    # print(predicted_ean_digits_print)
                    decoded_text = ''.join([self._characters[idx] for idx in predicted_ean_digits_print])
                    decoded = ''.join(str(d) for d in barcode_value)                        
                    logger.debug(f"✅ {is_valid}, readable: {decoded}, raw: {decoded_text}, conf: {avg_conf}, format: {msg}, detection_idx: {detection_idx}")   
        return is_valid, decoded, decoded_text, msg, avg_conf, detection_idx
    
    def decode_code128(self, values):
        # print(f"values:{values}")
        values = values[:]  # avoid modifying original
        start_code = values.pop(0)
        checksum = values.pop(-1)
        stop_code = values.pop(-1)

        if start_code == 103:
            current_set = 'A'
        elif start_code == 104:
            current_set = 'B'
        elif start_code == 105:
            current_set = 'C'
        else:
            return "Invalid start code"

        decoded_chars = []
        i = 0
        while i < len(values):
            val = values[i]

            # Handle switch codes
            if val == 99:
                current_set = 'C'
                i += 1
                continue
            elif val == 100:
                current_set = 'B'
                i += 1
                continue
            elif val == 101:
                current_set = 'A'
                i += 1
                continue

            # Decode values
            if current_set == 'C':
                decoded_chars.append(f"{val:02d}")
            else:
                if 0 <= val < len(self._characters):
                    decoded_chars.append(self._characters[val])
                else:
                    decoded_chars.append('?')

            i += 1

        return ''.join(decoded_chars)

    def validate_checksum(self, raw_values):
        # raw_values = [characters.index(char, 1) - 1 if char in characters[1:] else -1 for char in decoded_text]
        logger.debug(f"raw_values passed to validate_checksum: {raw_values}")
        if raw_values[0] == 107:
            logger.debug("Ean Barcode detected")
            
            code_map_predicted = {
                'C1': '107', 'C2': '108', '0A': '109', '1A': '110', '2A': '111', '3A': '112', '4A': '113',
                '5A': '114', '6A': '115', '7A': '116', '8A': '117', '9A': '118', '0B': '119', '1B': '120',
                '2B': '121', '3B': '122', '4B': '123', '5B': '124', '6B': '125', '7B': '126', '8B': '127',
                '9B': '128', '0C': '129', '1C': '130', '2C': '131', '3C': '132', '4C': '133', '5C': '134',
                '6C': '135', '7C': '136', '8C': '137', '9C': '138'
            }
            
            LEFT_PATTERN = (
                "AAAAAA", "AABABB", "AABBAB", "AABBBA", "ABAABB",
                "ABBAAB", "ABBBAA", "ABABAB", "ABABBA", "ABBABA"
            )
            reverse_code_map = {v: k for k, v in code_map_predicted.items()}
            predicted_digits = [val for val in raw_values[1:-1] if val != 108]
            # print(f"predicted_digits: {predicted_digits}")
            mapped_keys = [reverse_code_map.get(str(val), f"UNK({val})") for val in predicted_digits]
            logger.debug(f"Mapped keys: {mapped_keys}")
            detected_pattern = ''.join(k[-1] if len(k) == 2 else '?' for k in mapped_keys[:6])
            # print(f"Detected_pattern: {detected_pattern}")
            if detected_pattern in LEFT_PATTERN:
                pattern_index = LEFT_PATTERN.index(detected_pattern)
                # print(f"pattern_index: {pattern_index}")
            else:
                
                return False, None, "wrong pattern", None
            predicted_ean_value = ''.join(k[0] if len(k) == 2 and k[0].isdigit() else '?' for k in mapped_keys)
            # print(f"predicted_ean_value:{predicted_ean_value}")
            ean13 = str(pattern_index) + predicted_ean_value[:12]
            logger.debug(f"ean13 base:{ean13}")
            if len(ean13) != 13 or not ean13.isdigit():
                logger.warning(f"Invalid Ean value needs to be 13 digits")
                return False, None, "Invalid EAN-13 base", None
            else:
                def calculate_ean13_checksum(ean13):
                    digits = [int(d) for d in ean13]
                    even_sum = sum(digits[i] for i in range(0, 12, 2))
                    odd_sum = sum(digits[i] for i in range(1, 12, 2))
                    total = even_sum + odd_sum * 3
                    # print(f"total:{total}")
                    return (10 - (total % 10)) % 10
            
                calculated_ean_checksum = calculate_ean13_checksum(ean13)
                # print(f"calculated_ean_checksum:{calculated_ean_checksum}")
                predicted_ean_checksum = predicted_ean_value[-1]
                # print(f"predicted_ean_checksum:{predicted_ean_checksum}")
                if str(predicted_ean_checksum) != str(calculated_ean_checksum):
                    logger.warning(f"Invalid ean 13 checksum value, supposed to be {predicted_ean_checksum} but got {calculated_ean_checksum}")
                    return False, None, f"Invalid ean 13 checksum: expected {predicted_ean_checksum}", None
                else:
                    ean_list = [int(char) for char in ean13]
                    predicted_ean_digits_print = raw_values
                return True, ean_list, "ean13", predicted_ean_digits_print
        else:
            logger.debug("Code128 Barcode detected")
            # dict_converted_to_code128_dict = [x - 1 for x in raw_values]
            # print("dict_converted_to_code128_dict", dict_converted_to_code128_dict)
            code128 = raw_values
            # print("code128code128", code128)
            start_code = code128[0]
            
            predicted_code128_checksum = code128[-2]
            # print("predicted_code128_checksum", predicted_code128_checksum)
            code128_base = code128[1:-2]
            # print(code128_base)
            weighted_sum = sum(val * (i +1) for i, val in enumerate(code128_base))
            calculated_128_checksum =(start_code + weighted_sum) % 103
            logger.debug(f"predicted_code128_checksum: {predicted_code128_checksum}, calculated_128_checksum: {calculated_128_checksum}")
            if predicted_code128_checksum != calculated_128_checksum:
                logger.warning(f"Invalid checksum value, supposed to be {calculated_128_checksum} but model predicted {predicted_code128_checksum}")
                return False, None, "Invalid checksum value", None
            
            return True, code128, "code128", None