Docx file is broken after file.save() in a Flask upload

[guillermohs9]

Hi!
I'm using an upload function very similar to the one on Flask docs, just that I split the filename and do different things according to the uploaded file extension. Images and PDFs work fine, but when I upload a .docx or .doc file (any valid file, like this one), the file that gets saved on my filesystem is the same size but can't be opened by Word since its broken, and reports a different mimetype (application/zip)

Here's my upload function, there are other elifs with other extension but those are working fine.

def upload_file(file):
    if not file or not any(f for f in file) or not allowed_file(file.filename):
        raise ValidationError
    sfilename = secure_filename(file.filename)
    random_hex = secrets.token_hex(20)
    _, f_ext = os.path.splitext(sfilename)
    f_ext = f_ext.lower()
    if f_ext == ".doc" or f_ext == ".docx":
        filename = random_hex + f_ext
        file.save(os.path.join(current_app.root_path, 'static', 'uploads', filename))
        return filename

Of course the expected output would be a saved file that's identical to the uploaded one.

Environment:

• Python version: 3.10.2
• Werkzeug version: 2.3.6
• Flask version: 2.3.2

Any help would be appreciated.
Cheers

[guillermohs9]

Just to add some info:

I installed python-docx, and replaced the function with the following:

def upload_file(file):
    if not file or not any(f for f in file) or not allowed_file(file.filename):
        raise ValidationError
    sfilename = secure_filename(file.filename)
    random_hex = secrets.token_hex(20)
    _, f_ext = os.path.splitext(sfilename)
    f_ext = f_ext.lower()
    if f_ext == ".doc" or f_ext == ".docx":
        filename = random_hex + f_ext
        document = Document(file)
        document.save(os.path.join(current_app.root_path, 'static', 'uploads', filename))
        # file.save(os.path.join(current_app.root_path, 'static', 'uploads', filename))
        return filename

So now the save method is from the Document object, and now it saves just fine...