Twake mail extension: watermarking

[chibenwa]

Why

This is a feature idea taken from BPI review of the Wimi project. This issue records the idea...

Bob sends documents via email to journalists.

Given the transmitted document I need to be able to trace back the guy who sent the document.

How

The idea is to "water-mark" the document prior to delivery so that we would be able to track the message metadata.

A naive version of it would be to use metadata (eg on PDF)

Example:

• add X-LINAGORA-DOC-ID: {UUID} on first document processing, then preserve it
• add X-LINAGORA-WATERMARKING: {UUID} on each document processing then override it

Then store :

Watermarking UUID	Root document Id	Document name	Mailet identifier	Date	Sender	Recipient
xxx-xxx	yyy-zzz	configential-instruction.pdf	LOCAL-DELIVERY	2025-03-23 14:04:34	[email protected]	[email protected]
aaa-bbb	yyy-zzz	configential-instruction.pdf	LOCAL-DELIVERY	2025-03-25 17:19:34	[email protected]	[email protected]

Components

• A mailet to split email: one for each recipient
• A mailet to mark PDF:
  • modifies the email body
  • record the watermarking info
• A puggable storage extension: Cassandra, postgres

Reading the watermarking info can be done directly on the DB

Downside

If enabled it will not play well with dedup