Takehiro Suzuki
Takehiro Suzuki
@msalheb @axelpina Hi, thank you for reporting this issue. Could you describe the detail to reproduce the problem? e.g. region, bot configuration, etc.
@mohanraj-u Thank you for your contribution and implementing this feature request! I appreciate the work you've put into supporting local WebSocket development and debugging. At the moment, our team is...
Could you provide a log? Is there any error?
#862 merged. Please check if the problem resolved. Thank you
@nagineko お返事が遅くなりすみません。UIコンポーネントに関する詳細なフィードバックをありがとうございます。 現在、私たちのチームではv3リリースとボット機能の強化を最優先で進めております。また、UIデザイン (css) についても、アプリケーションの性質や用途を考慮した上で現在の形としております。 そのため、申し訳ございませんが、デザインの改善提案については一旦保留とさせていただきたく存じます。v3リリース以降、リソースの状況を見ながら、必要に応じて検討させていただければと思います (本Issueはopenのままにしておきます)。 ご理解いただけますと幸いですmm
@zcqqq Thank you for fixing bug! Could you fix the eslint error?
The details of the tool usage and the final generation should inform the user accordingly (indicating that they should wait). I don't think special handling is necessary. If admin want...
Have you tried to create a new conversation after Aurora waked up? In the case of multi-turn conversation, LLM would not take tool use actions to retrieve the chunks because...
Understood. It would be better to implement retry logic in the following files: - https://github.com/aws-samples/bedrock-chat/blob/v3/backend/app/agents/tools/knowledge.py - https://github.com/aws-samples/bedrock-chat/blob/v3/backend/app/vector_search.py
Have you tried importing KnowledgeBase ID?