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

Web access toggle #1931

Open
bp3000bp opened this issue Aug 22, 2024 · 1 comment
Open

Web access toggle #1931

bp3000bp opened this issue Aug 22, 2024 · 1 comment
Labels
enhancement New feature or request rag Related to RAG flow/logic/approaches

Comments

@bp3000bp
Copy link

Please provide us with the following information:

This issue is for a: (mark with an x)

- [ ] bug report -> please search issues before submitting
- [x] feature request
- [ ] documentation issue or request
- [ ] regression (a behavior that used to work and stopped in a new release)

Minimal steps to reproduce

it would be sick if there was a button where you could switch on and off for web access. for example, toggled off would be answers based only off the knowledge base provided, and toggled on would allow it to function similar to GPT, where it can refer to both the docs and to all of the rest of its knowledge, with the security of this setup and the data not being used for training models. That to me would be the ultimate enterprise level custom internal chatbot. I don't know how difficult this would be to set up but that would be dope. thanks

Any log messages given by the failure

n/a

Expected/desired behavior

described earlier

OS and Version?

Windows 11

azd version?

n/a enhancement

Versions

n/a

Mention any other details that might be useful


Thanks! We'll be in touch soon.

@pamelafox
Copy link
Collaborator

To implement this sort of feature in our codebase, I'd take this approach:

  1. Add an additional function to the query rewriting tools definition like "search_web" so that the LLM had to choose between "search_web" and "search_sources".
  2. If "search_web" was suggested as the right function to use, then it'd use the Bing Search API to search the web for the suggested query.
  3. It would send the concatenated results to the LLM.

Alternatively, you could use the OpenAI Assistants API, which I believe already has a web searching tool. This repository does not use that API, as we are lower level.

It could be a neat enhancement. As usual, I can't make guarantees about whether any of the maintainers will have time to implement it. If you implement it yourself, please share in a branch/PR. Thanks for filing!

@pamelafox pamelafox added enhancement New feature or request rag Related to RAG flow/logic/approaches labels Aug 28, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request rag Related to RAG flow/logic/approaches
Projects
None yet
Development

No branches or pull requests

2 participants