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: Keep history / embedding to allow query and report rewriting #1291

Open
ga-it opened this issue Mar 22, 2025 · 1 comment
Open

Comments

@ga-it
Copy link

ga-it commented Mar 22, 2025

Is your feature request related to a problem? Please describe.
When I go to the history, these reports are not available for further queries. When a report is immediately finished being written GPTR does allow query for a period (although this seems to be for a limited time?

Also, after a report has been produced, it seems like that is the perfect time to improve focus and style and ask for further research or for rewriting.

Describe the solution you'd like
Could embeddings per report be kept until a delete request or alternatively for a period defined as a parameter?

Then, either when accessed as a historical report or after the first iteration of the report, allow for further query or further research or for rewriting.

Describe alternatives you've considered
I'm not sure there is a good alternative other than feeding a finished report in to a hybrid search, but the loss of the original search and embeddings limits the usefulness of this.

@ElishaKay
Copy link
Collaborator

ElishaKay commented Mar 23, 2025

a) True - High Level, the documents are embedded into a Langchain Vector Store on every run

We'll probably want to add support for a Vector Store like Pinecone or Mongo to preload the data into the Vector Store and run a report with that Vector Store as Context

More on that here: Data Ingestion

b) Regarding chatting with previously saved reports

c) I like the idea of maintaining as much context for as long as possible within reason - I'd even go one step further to say that I would love for the project to have an advanced memory module.

For example, Langgraph put out this memory agent which is able to classify any chat inputs as:

  • About the User
  • Tracking To Dos of the user
  • How the user likes things done

That could be huge in terms of personalization for the app where the memory module keeps track of: what you do for a living, what's your areas of interest, why you're interested in it, how far along you are in your education of a specific topic...

d) For A-C above: Green light for the PR if someone would like to start investigating or start building it out

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

2 participants