Extract Structured Fields from Manufacturing Invoices with PaddleOCR and Docling

                      
                      import os
import paddleocr
from typing import Dict, Any

# Configuration
OcrModel = paddleocr.OCR()

# Function to extract fields from invoice
def extract_invoice_fields(image_path: str) -> Dict[str, Any]:
    try:
        # Perform OCR on the image
        results = OcrModel.ocr(image_path, cls=True)
        # Parse results to extract structured fields
        extracted_fields = parse_results(results)
        return extracted_fields
    except Exception as e:
        print(f"Error during OCR processing: {str(e)}")
        return {}

# Helper function to parse OCR results
def parse_results(results: Any) -> Dict[str, str]:
    fields = {}
    for result in results:
        for line in result:
            text = line[1][0]
            # Implement logic to categorize fields based on text
            if "Invoice Number" in text:
                fields["invoice_number"] = text.split(':')[-1].strip()
            elif "Date" in text:
                fields["date"] = text.split(':')[-1].strip()
            # Add more fields as necessary
    return fields

if __name__ == '__main__':
    # Example usage
    image_file = os.getenv('INVOICE_IMAGE_PATH', 'invoice.jpg')
    fields = extract_invoice_fields(image_file)
    print(fields)