The Python script process the Draw.io library files and "extract" the images from within it. In my case, I want to save the images as SVG and PNG files.

extract-svg-from-drawio-library.py

import xml.etree.ElementTree as ET
import base64
import cairosvg
import os
import json

def process_mxlibrary(xml_file, output_dir="output"):
    """
    Processes a .drawio XML file containing an mxlibrary, extracts SVG/PNG images,
    and saves them to the specified output directory.

    Args:
        xml_file (str): Path to the .drawio XML file.
        output_dir (str):  (Optional) Base directory where output directories
                       will be created. Defaults to "output".
    """

    try:
        tree = ET.parse(xml_file)
        root = tree.getroot()

        if root.tag != "mxlibrary":
            print(f"Error: {xml_file} is not a valid mxlibrary file.")
            return

        #json_string = root.text.strip()  # Get the text and remove extra whitespace
        #print(f"JSON String: {json_string}")  # Print the JSON string

        try:
            json_array = json.loads(root.text)  # Load the JSON from the text content
        except json.JSONDecodeError as e:
            print(f"Error decoding JSON in {xml_file}: {e}")
            return

        # Create base output directory if it doesn't exist
        os.makedirs(output_dir, exist_ok=True)

        # Create subdirectories for each file
        file_name = os.path.basename(xml_file)
        directory_name = os.path.join(output_dir, os.path.splitext(file_name)[0])
        svg_dir = os.path.join(directory_name, "svg")
        png_dir = os.path.join(directory_name, "png")

        os.makedirs(svg_dir, exist_ok=True)
        os.makedirs(png_dir, exist_ok=True)

        for item in json_array:
            print("Processing item")  # Debugging line
            if "data" in item and "title" in item:
                data = item["data"]
                title = item["title"]

                if "data:image/svg+xml;base64," in data:
                    try:
                        svg_data = data.replace("data:image/svg+xml;base64,", "")
                        svg_bytes = base64.b64decode(svg_data)
                        svg_string = svg_bytes.decode("utf-8")

                        print(f"  Processing: {title}")
                        print(f"  First 50 chars of SVG: {svg_string[:50]}")  # Print a snippet

                        svg_filepath = os.path.join(svg_dir, f"{title}.svg")
                        png_filepath = os.path.join(png_dir, f"{title}.png")

                        cairosvg.svg2svg(bytestring=svg_string, write_to=svg_filepath)
                        print(f"    Saved SVG to: {svg_filepath}")

                        cairosvg.svg2png(bytestring=svg_string, write_to=png_filepath)
                        print(f"    Saved PNG to: {png_filepath}")

                    except Exception as e:
                        print(f"Error processing {title} from {xml_file}: {e}")
                elif "data:image/png;base64," in data:
                    try:
                        png_data = data.replace("data:image/png;base64,", "")
                        png_bytes = base64.b64decode(png_data)

                        print(f"  Processing: {title} (PNG)")

                        png_filepath = os.path.join(png_dir, f"{title}.png")
                        with open(png_filepath, "wb") as f:
                            f.write(png_bytes)
                        print(f"    Saved PNG to: {png_filepath}")

                    except Exception as e:
                        print(f"Error processing {title} (PNG) from {xml_file}: {e}")

    except FileNotFoundError:
        print(f"Error: XML file not found at {xml_file}")
    except ET.ParseError as e:
        print(f"Error parsing XML in {xml_file}: {e}")
    except Exception as e:
        print(f"An unexpected error occurred processing {xml_file}: {e}")


# --- Main Execution ---
if __name__ == "__main__":
    process_files = [
        'MuleSoftLibrary.xml'
    ]

    for file in process_files:
        process_mxlibrary(file)  # Use default output directory "output"