I have a bit of confusion. The documentation lists some file purposes, as shown in the link, but “user_data” is not included. However, when uploading files in the dashboard, “user_data” is listed as an option for the purpose.
I’m wondering if “user_data” is a valid file purpose for both uploading and listing files, and what exactly is the purpose of this purpose?
I noted this purpose at the start of July.
Such “user_data” uploaded files can be attached to vector stores.
VectorStoreFile(id='file-5gomYFChuzS...', created_at=1727502500, last_error=None, object='vector_store.file', status='completed', usage_bytes=8077, vector_store_id='vs_py4b48rNm...', chunking_strategy=StaticFileChunkingStrategyObject(static=StaticFileChunkingStrategy(chunk_overlap_tokens=400, max_chunk_size_tokens=800), type='static'))
And then used:
Document extraction is also performed.
I would conclude that “user_data” was created with the intent of more robust project scoping, containing a user’s or sub-organization’s files and other personal information to within a project.
However, in the intervening two months, there is little more than a hint that any data would be firewalled between projects, so I would not trust it for any specific purpose without documentation and clear announcements that “scoping” is vetted and production-ready.
