How to Read 1D/2D Barcodes from PDF Files Using Dynamsoft Python Barcode SDK

#programming #python #barcode #pdf

The Dynamsoft Python Barcode SDK, now part of the Dynamsoft Capture Vision Bundle, provides APIs for reading 1D and 2D barcodes from various image and PDF files. This tutorial demonstrates how to build both a command-line tool and a GUI application to read barcodes from PDF documents using the Dynamsoft Python Barcode SDK and an AI agent:

Command-Line Tool: A CLI sample demonstrating how to read barcodes from PDF files using the Dynamsoft Python Barcode SDK. The sample also serves as a reference for an AI agent to build a more advanced GUI app.
GUI Application: A modern, feature-rich desktop app with drag-and-drop support, automatically built by an AI agent.

Demo: Python GUI Barcode Reader for PDF

Prerequisites

Install required dependencies:

pip install dynamsoft-capture-vision-bundle opencv-python Pillow tkinterdnd2 psutil

Get a 30-day trial license

Steps to Build a Python Command Line Barcode Reader for PDF

Initialize the license:

from dynamsoft_capture_vision_bundle import *

def initialize_license() -> Tuple[bool, Optional[str]]:
    error_code, error_message = LicenseManager.init_license(LICENSE_KEY)

    if error_code == EnumErrorCode.EC_OK or error_code == EnumErrorCode.EC_LICENSE_CACHE_USED:
        return True, None
    else:
        return False, f"License initialization failed: ErrorCode: {error_code}, ErrorString: {error_message}"

Create a custom intermediate receiver to fetch rasterized PDF pages:

class MyIntermediateResultReceiver(IntermediateResultReceiver):

    def __init__(self, im: IntermediateResultManager):
        self.images = {}
        self.im = im
        super().__init__()

    def on_localized_barcodes_received(self, result: "LocalizedBarcodesUnit", info: IntermediateResultExtraInfo) -> None:
        """Handle localized barcode results and capture images."""
        try:
            hash_id = result.get_original_image_hash_id()
            if hash_id and self.im:
                image = self.im.get_original_image(hash_id)

                if image is not None and self.images.get(hash_id) is None:
                    try:
                        image_io = ImageIO()
                        saved = image_io.save_to_numpy(image)
                        self.images[hash_id] = saved
                    except Exception as e:
                        print(f"{YELLOW}⚠ Warning: Could not save intermediate image: {e}{RESET}")
        except Exception as e:
            print(f"{YELLOW}⚠ Warning: Error in intermediate result receiver: {e}{RESET}")

Instantiate CaptureVisionReader and attach the receiver:

cvr_instance = CaptureVisionRouter()
intermediate_result_manager = cvr_instance.get_intermediate_result_manager()
custom_receiver = MyIntermediateResultReceiver(intermediate_result_manager)
intermediate_result_manager.add_result_receiver(custom_receiver)

Read barcodes from a PDF file:

results = cvr_instance.capture_multi_pages('PDF-File-Path', EnumPresetTemplate.PT_READ_BARCODES.value)
result_list = results.get_results()

Display each page and barcode results with OpenCV:

for page_index, result in enumerate(result_list):
    if result.get_error_code() != EnumErrorCode.EC_OK:
        print(f"{RED}✗ Error on page {page_index + 1}: {result.get_error_string()}{RESET}")
        continue

    hash_id = result.get_original_image_hash_id()
    items = result.get_items()
    page_number = page_index + 1 if len(result_list) > 1 else None

    if not items:
        page_info = f" (Page {page_number})" if page_number else ""
        print(f"{YELLOW}⚠ No barcodes found{page_info}{RESET}")
    else:
        print_barcode_details(items, page_number)
        total_barcodes += len(items)

    if hash_id in custom_receiver.images:
        try:
            image_data = custom_receiver.images[hash_id]

            if isinstance(image_data, np.ndarray):
                cv_image = image_data
            elif isinstance(image_data, tuple) and len(image_data) >= 3:
                cv_image = image_data[2]  
            else:
                print(f"{YELLOW}⚠ Unexpected image data format for page {page_number}{RESET}")
                continue

            if cv_image is not None and cv_image.size > 0:
                if items:
                    annotated_image = draw_barcode_annotations(cv_image, items)
                else:
                    annotated_image = cv_image.copy()

                display_image, _ = resize_image_for_display(annotated_image)

                window_name = f"Page {page_number} ({hash_id})" if page_number else f"Barcode Detection - {os.path.basename(image_path)} ({hash_id})"
                setup_window(window_name, display_image)
                cv2.imshow(window_name, display_image)

                processed_pages += 1

                if len(result_list) > 1:
                    print(f"\n{CYAN}📄 Displaying Page {page_number} ({hash_id}) - Press any key to continue to next page...{RESET}")
                    cv2.waitKey(0)
                    cv2.destroyWindow(window_name)

        except Exception as e:
            print(f"{RED}✗ Error displaying page {page_number}: {e}{RESET}")
            continue
    else:
        page_info = f" page {page_number}" if page_number else ""
        print(f"{YELLOW}⚠ No image data available for{page_info}{RESET}")

Using AI Agent to Build a Desktop GUI Barcode Reader for PDF

This section shows how to generate a professional GUI barcode reader using an AI coding assistant.

Key Features

Multi-format Support: JPG, PNG, BMP, TIFF, WEBP, and PDF files
Multi-page PDF Processing: Native PDF support without external dependencies
Real-time Visual Annotations: Draw bounding boxes and labels on detected barcodes
Multiple Display Modes: Individual windows for each PDF page
Professional UI: Modern interface with zoom, pan, and navigation controls
Export Capabilities: Save results in TXT, CSV, or JSON formats
Confidence Scoring: Display detection confidence levels

Generating the GUI App with Copilot AI Agent

In Visual Studio Code, start a Copilot chat in agent mode using Claude Sonnet4.

Describe the desired functionality in detail. For example, you can use the following prompt:

I want to create a Python GUI app that reads barcodes from PDF files. It should support drag-and-drop, show bounding boxes around barcodes, and allow saving results in multiple formats. You can use the command-line barcode reader code I provided as a reference.