wenjin272 commented on issue #144: URL: https://github.com/apache/flink-agents/issues/144#issuecomment-6820190139
Thanks for offering to work on this @ChickenBenny! This issue was opened quite a while ago, and the framework has evolved since then. The proposed `RAGAgent` would essentially provide a fixed 2-step RAG flow: retrieve documents, then generate an answer using them. We already support this through the existing events and actions, as shown in the RAG example, so a dedicated agent would mainly reduce boilerplate. At this point, I think making retrieval easier to use as a tool would be more useful. A built-in vector store retrieval tool could let an existing `ReActAgent` search a knowledge base when needed, refine its query, or search again based on the results. It would also be reusable across different agents. Would you be interested in exploring that direction instead? We could start with a tool that references a registered vector store and returns document content and source metadata, while keeping the fixed 2-step flow covered by the example. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
