doctodocx.py

import os
import time

import win32com.client as win32
from win32com.client import constants

### delete contais of C:\Users\David\AppData\Local\Temp\gen_py if error module 'win32com.gen_py.00020905-0000-0000-C000-000000000046x0x8x7' has no attribute 'MinorVersion' occurs


def convert_doc_to_docx(doc_path, docx_path=None):
    """
    Convert a .doc file to .docx format.

    Args:
        doc_path (str): Path to the input .doc file
        docx_path (str, optional): Path to save the output .docx file.
                                  If None, saves to same directory with .docx extension.
    Returns:
        str: Path to the converted file
    """
    # If output path not specified, create one with .docx extension
    if docx_path is None:
        base = os.path.splitext(doc_path)[0]
        docx_path = base + ".docx"

    # Make sure paths are absolute
    doc_path = os.path.abspath(doc_path)
    docx_path = os.path.abspath(docx_path)

    try:
        # Create Word application object
        word = win32.gencache.EnsureDispatch("Word.Application")
        word.Visible = False  # Run Word in background

        # Open the DOC file
        doc = word.Documents.Open(doc_path)

        # Save as DOCX (file format constant is 16 for docx)
        doc.SaveAs(docx_path, FileFormat=constants.wdFormatXMLDocument)

        # Close the document and quit Word
        doc.Close(False)
        word.Quit()

        print(f"Successfully converted: {doc_path} -> {docx_path}")
        return docx_path

    except Exception as e:
        print(f"Error converting {doc_path}: {str(e)}")
        return None


def convert_folder(input_folder, output_folder=None):
    """
    Convert all DOC files in a folder to DOCX format.

    Args:
        input_folder (str): Path to folder containing .doc files
        output_folder (str, optional): Folder to save converted files.
                                     If None, saves to same folder as input.
    """
    # Validate input folder
    if not os.path.isdir(input_folder):
        print(f"Error: Input folder does not exist: {input_folder}")
        return

    # Create output folder if specified
    if output_folder and not os.path.exists(output_folder):
        os.makedirs(output_folder)

    # Get all DOC files in input folder
    doc_files = [
        f
        for f in os.listdir(input_folder)
        if f.lower().endswith(".doc") and os.path.isfile(os.path.join(input_folder, f))
    ]

    if not doc_files:
        print("No DOC files found in the input folder.")
        return

    print(f"Found {len(doc_files)} DOC files to convert.")

    # Convert each file
    success_count = 0
    for doc_file in doc_files:
        input_path = os.path.join(input_folder, doc_file)

        if output_folder:
            output_path = os.path.join(
                output_folder, os.path.splitext(doc_file)[0] + ".docx"
            )

        else:
            output_path = None  # Let convert_doc_to_docx handle it

        if convert_doc_to_docx(input_path, output_path):
            success_count += 1
            # wait 1 second
            time.sleep(1)

    print(
        f"\nConversion complete. Successfully converted {success_count} of {len(doc_files)} files."
    )


# Example usage
if __name__ == "__main__":
    input_folder = r"C:\Users\David\Documents\PROJECTS\2025\ROLLOUT 2026\CIQ\FN4B\V2"  # Change this to your folder path
    output_folder = r"C:\Users\David\Documents\PROJECTS\2025\ROLLOUT 2026\CIQ\FN4B\V2\DCOX"  # Optional - set to None to save in same folder

    convert_folder(input_folder, output_folder)