quarkus-langchain4j

quarkus-langchain4j copied to clipboard

It should be possible to control access to the RAG store with roles or permissions

4

It may already work but some tests should be added

sberyozkin

User input and output to the prompt should be sanitized

4

By default, basic escaping of `` and a few other characters should be done, but users should also be able to register per/post handlers for sanitizing if they would like...

sberyozkin

Add an OpenId Connect demo where a user is required to approve the `LLM` scope

2

I propose to demo and also recommend the following set up: a user logs in to a frontend Quarkus application (say `Quarkus LLM`) which talks to a microservice which uses...

sberyozkin

Advise users in the docs to be careful with using RAG with sensitive/personal information

3

As it may end up propagating into the LLM model outside of the trust boundaries or a response based on this augmentation can be logged by the agent

sberyozkin

jmartisk

Support a user specific RAG augmentation

1

Consider how Quarkus Authentication and LLM API Keys will work together. I'm assuming, if I start a demo with an API key, in prod, it will be more like a...

sberyozkin

how to load local llm in quarkus

3

hello，how to load local llm in quarkus

oneanime

Look into supporting langchain4j-vertex-ai-gemini

5

This module depends on gRPC so we'll have to see what the best way of supporting it is - Depends on: #185

geoand

Cover the whole Ollama API

3

I can see that Quarkus extension only covers the `/api/generate` Ollama endpoint but do we plan to cover everything like here https://github.com/langchain4j/langchain4j/blob/main/langchain4j-ollama/src/main/java/dev/langchain4j/model/ollama/OllamaApi.java ? I'm mostly interested in the `embeddings` endpoint....

sebastienblanc

OpenSearch as Embedding store

1

I started working on that but then I remember that @sboeckelmann said he was also working on it. @sboeckelmann are you still planning on submitting a PR ?

sebastienblanc