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

memory #890

Open
1 task done
tlallema opened this issue Nov 20, 2024 · 4 comments
Open
1 task done

memory #890

tlallema opened this issue Nov 20, 2024 · 4 comments
Assignees
Labels
General Question ❓ Issue contains a general question

Comments

@tlallema
Copy link

Question

Hello,
I'm not sure but it seems that memory is not yet enabled in vizro_ai app.
Do you plan to add this feature soon ? It would be great to enable a user asking to change a chart already generated by AI without having to restart from scratch (e.g. change a filter, change the line color chosen by default, add an annotation on the Plotly graph...)
Thanks in advance for your support

Code/Examples

No response

Which package?

None

Code of Conduct

@tlallema tlallema added General Question ❓ Issue contains a general question Needs triage 🔍 Issue needs triaging labels Nov 20, 2024
@maxschulz-COL maxschulz-COL self-assigned this Nov 20, 2024
@maxschulz-COL maxschulz-COL removed the Needs triage 🔍 Issue needs triaging label Nov 20, 2024
@maxschulz-COL
Copy link
Contributor

maxschulz-COL commented Nov 20, 2024

Hey @tlallema ,

thanks for reaching out! Very good question - this is indeed something we are actively thinking about and it is on our roadmap to tackle very soon (but before the end of this year).

As you can imagine this is actually a tricky feature to get right, so we would like to take time to do it properly.

Maybe a question back if I may 😄 : if you could choose between the following two features, which one would you prefer and why?

  • vizro-ai remembers a history of prompts and answers it gave, and just uses this as additional context to rebuild an entirely new dashboard (which will likely be still overall consistent, but also slower to build)
  • vizro-ai output can be "locked", ie only a part of the dashboard is rebuilt while the rest is guaranteed to stay exactly the same (but that may potentially lead to inconsistencies that result in more errors, but this would also be faster)

@tlallema
Copy link
Author

Thanks for your prompt feedback.
Difficult to choose between the two options :-) as I'm not an expert yet with vizro...
I would say it depends on how the dashboard is built (and the cross-dependence between visuals).
Option 1 would probably offer higher consistency.
If you expect only one visual to be updated when new user-input is provided then option 2/ might be sufficient.

@tlallema
Copy link
Author

@maxschulz-COL did you have the chance to work on it ? could you please keep me informed when memory feature is available ? thank you

@maxschulz-COL
Copy link
Contributor

@maxschulz-COL did you have the chance to work on it ? could you please keep me informed when memory feature is available ? thank you

Hey! Thanks for coming back to this! I will definitely keep you informed - but it will be next year :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
General Question ❓ Issue contains a general question
Projects
None yet
Development

No branches or pull requests

2 participants