LangChain4j

@kk-r sorry, no update so far. I will prioritize this

Duplicate of https://github.com/langchain4j/langchain4j/issues/199, please track it there

@andyflury how do you plan to select tools? Manually? On each call? At the moment using multiple AI services with the same memory/provider should work. Or you can use low-level...

@andyflury it seems that we can really simplify this, but I need to know a bit more about your use case. If I understood correctly, you are creating an AiService...

WIll the tools be selected deterministically by some condition? Or a human will be selection? How many tools do you have? Can you give a short descriptions what those tools...

Yes, this helps, thank you! Another thing I am considering is to have an automatic dynamic tool selection, based on the user intent. I am not sure if this will...

With both `EmbeddingModelTextClassifier` and RAG options I meant not sending all tools at once on each call to the LLM, but rather first get the set of tools relevant for...

Yes, although it will do up to 2 times more calls, the number of used tokens should be (much) less, at least for the cases where there are many tools...

Related: https://github.com/quarkiverse/quarkus-langchain4j/issues/341

Hi @czelabueno, please LMK when this PR is ready for review (no rush)