SSWConsulting
🤔 Should we follow this Microsoft Blog Post?
cc: @jackreimers @MattParkerDev @Harry-Ross @adamcogan @calumjs Hi all,
I had a read through of this article https://devblogs.microsoft.com/dotnet/demystifying-retrieval-augmented-generation-with-dotnet/ (if tldr; read the last code snippet right at the bottom)
It seems that the guy basically built RulesGPT using Microsoft's .NET Semantic Kernel package. Using extensions to use Azure OpenAI using its Chat Completion & Embeddings.
I wonder if we should follow these instructions for the RulesGPT project (and as part of it migrate to persist the embeddings on Azure?)
- [ ] Have a read
- [ ] Decide if we should use the Semantic Kernel library
- [ ] Decide if we should use Azure OpenAI
@Hona from a quick read it looks like using Semantic Kernel could simplify and consolidate our infrastructure a lot. Given that we already have issues open such as #136 that will involve a lot changes to the backend I think it's worth some further investigation.
As per my conversation with @calumjs, we think using Semantic Kernel would be good if we were starting fresh but there isn't much value in doing it right now. We will add this to the backlog and revisit.