[Important] Unknown error about docx document

[cokacoda]

First of all, thank you very much for contributing such a great translator that uses the DeepL API. However, I found a problem in using it: the software can't provide translations for some docx documents. The software says that it can't count the number of words, probably because it is being opened by another software (e.g. word) and refuses to translate.
After many tests I found that the error is triggered for docx files that have been converted from PDF files directly by Adobe Acrobat. However if I open the file in Microsoft Word and save a copy directly then the software will work fine.
I'm not sure what the saving of the copy is causing, but it's safe to assume that the software has a design flaw that should allow the user to upload the file directly for translation even if the software can't read the word count of the document.
Finally, I would like to briefly recount my reasons for choosing to convert a PDF file to a docx document before translating it. Indeed, DeepL can directly translate PDF documents, but may be limited by DeepL's OCR capabilities, direct translation of the layout is not very good. In addition, this allows me to get an editable version of the translation and make changes.

I hope you can fix the problem, thank you very much.