get and set Sensitivity Labels in Docx file

[msuguino]

Companies often require that sensitivity labels are set for Microsoft Office files such as docx and xlsx.

I've used python-docx combined with win32com (pywin32) to get the label from a file which had the label properly set, and then, set the same label in a new document to be read by python-docx.

To get the sensitivity label:
myWord.ActiveDocument.SensitivityLabel.GetLabel().LabelId

To get the sensitivity label:
myWord.ActiveDocument.SensitivityLabel.GetLabel().LabelName

To get the assignment method:
myWord.ActiveDocument.SensitivityLabel.GetLabel().AssignmentMethod

To get the justification:
myWord.ActiveDocument.SensitivityLabel.GetLabel().Justification

---

I used the code below to set the label in a new document:

# Opening a MS Word file in pywin32 to set sensitivity label
myWord = Dispatch('Word.Application', pythoncom.CoInitialize())
myWord.Visible = 0
myWord.Documents.Open(src_absolute)

# Set Label
label_info = myWord.ActiveDocument.SensitivityLabel.CreateLabelInfo()
label_info.AssignmentMethod = 2
label_info.Justification = 'init'
label_info.LabelId = '123123123-123123123-ajskldfj123'
label_info.LabelName = 'Confidential'
myWord.ActiveDocument.SensitivityLabel.SetLabel(label_info, label_info)

# Close Word Application
myWord.ActiveDocument.Save()
myWord.Quit()

---

The problem lies in that linux servers often do not support microsoft office. The win32com solution is also slow, as it has to open the docx file, change settings, save it, and then close it.

By checking the Docx file and unzipping it, I found out that inside the folder "docProps" there is a file called "custom.xml" which stores the sensitivity label information.

Is it possible to support getting and setting sensitivity labels in python-docx?

Best regards

Marcos

[scanny]

It could potentially be supported but isn't at the moment. We would need to have some sort of a spec for that before even assessing the effort required. If you have or can find further information on that we could discuss further.

From the little I've read it seems like those are applied and read by some external system like Office365 or Purview and are accompanied by some sort of encryption. So I'm not exactly sure how far we could get with this but certainly open to further conversation if you can dig up some technical details.

Closing for now, but feel free to reopen if you are able to find more details.