Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Feature Request] Chat about all documents #259

Open
SHU-red opened this issue Jan 28, 2025 · 7 comments
Open

[Feature Request] Chat about all documents #259

SHU-red opened this issue Jan 28, 2025 · 7 comments

Comments

@SHU-red
Copy link

SHU-red commented Jan 28, 2025

Is your feature request related to a problem? Please describe.

  • I would like to only have documents edited by paperless-ai which have the tag "new" --> Works already
  • At the same time i would like to be able to ask AI about all of my documents questions like "How much did i pay for electricity since 2024"

Describe the solution you'd like
It would be cool to absolutely limit the documents which are edited to only a specific tag, but be able to chat about all documents no matter what tag

Describe alternatives you've considered
Process all documents but i do not want to edit my whole database by AI
Just the ones in my inbox

@clusterzx
Copy link
Owner

That will be hard for the tokenlimit on most LLMs or extremly costly when using a non selfhosted LLM.
Imagine you have 1000 Documents and you would push all of them to a LLM to find the right one -> Tokenlimit exceeded.
Imagine you would do it document by document until the LLM finds a matching document to your question -> Hell lot of tokens were used just for a single question (maybe millions of millions of tokens) -> Bill would be tremendous.

If you can point me in a direction to implement this feature without running in those problems I mentioned, then I can do it probally.

@MephistoJB
Copy link

Maybe just as an idea. This could be achieved by using a RAG. If openwebui could be used as an endpoint, the existing paperless plugin of it could be used to do the request then. This would of course only be possible for local machines. The FR would then be to create the possibility to have OpenWebui API as endpoint. Maybe this would also solve #217 (comment) (DeepSeek R1 integration) as well

@clusterzx
Copy link
Owner

Of course, but also implementing a check to see if the users environment is able to do RAG and only enable that feature AND rewrite a whole lot of the basic code structure. I dont think this is in my biggest interest right now (as speaking of time right now).

@SHU-red
Copy link
Author

SHU-red commented Jan 30, 2025

@clusterzx

  • yes i am also a local LLM user and fearlessly utilize my GPU 😄
  • if a first (test)implementation of this feature would be some kind of easy and only tokens are the problem there could be a warning/switch in the first step to make sure the guys paying for tokens are sufficently aware of what will happen using this function
  • (as you surely already guess) i did just brainstorm what could be done with your awesome application and came to this idea
  • i am very curious to see what would happen and how good it works 😅

@MephistoJB

  • thanks for your suggestion
  • Contribution! 😆 ... i am not able to do this

@clusterzx
Copy link
Owner

I will give it a try the coming days 👍 but time is rare at the moment as I am a dad since a couple of days 😆
Be patient my friends

@MephistoJB
Copy link

Hey congrats man!!!!

@clusterzx
Copy link
Owner

Thank you 👶

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants