Closed nag-24 closed 4 months ago
also check out this Langmsith trace using Conversational Retrieval QA chain. https://smith.langchain.com/public/dc94321b-f83d-49bd-8fca-4d308daf52ee/r, the retrieval is again correct but sent as "context": "
also check out this Langmsith trace using Conversation agent https://smith.langchain.com/public/b54e85e9-3313-4954-ba94-f0947aa2e32c/r , Received tool input did not match expected schema
Error: Received tool input did not match expected schema at DynamicStructuredTool.call (/usr/src/node_modules/.pnpm/@langchain+core@0.1.57/node_modules/@langchain/core/dist/tools.cjs:68:19) at process.processTicksAndRejections (node:internal/process/task_queues:95:5) at async /usr/src/packages/components/dist/src/agents.js:315:39 at async Promise.all (index 0) at async AgentExecutor.call (/usr/src/packages/components/dist/src/agents.js:302:30) at async AgentExecutor.invoke (/usr/src/node_modules/.pnpm/langchain@0.1.33@aws-crypto+sha256-js@5.2.0@aws-sdk+client-bedrock-runtime@3.422.0@aws-sdk_4nagnmezyqrtxortipronoidfi/node_modules/langchain/dist/chains/base.cjs:58:28) at async ConversationalAgent_Agents.run (/usr/src/packages/components/dist/nodes/agents/ConversationalAgent/ConversationalAgent.js:119:19) at async utilBuildChatflow (/usr/src/packages/server/dist/utils/buildChatflow.js:243:15) at async createInternalPrediction (/usr/src/packages/server/dist/controllers/internal-predictions/index.js:7:29)
It worked now when I changed the metadata field in my pinecone index from "content" to "text". See this langsmith trace. https://smith.langchain.com/public/078984df-3237-4275-845a-cc633291fc2d/r. I think the issue is fixed now.
Describe the bug A clear and concise description of what the bug is. Retriever Tool fetches the matching docs correctly from Pinecone's database but the retrievals are not sent correctly to LLM .
To Reproduce Steps to reproduce the behavior: See Langsmith trace where the correct document is retrieved but the tooloutput is sent to LLM as "toolOutput": "\n\n\n\n\n\n" https://smith.langchain.com/public/6595062e-ac4f-4d30-85aa-e36b809c8d57/r
Expected behavior A clear and concise description of what you expected to happen.
Screenshots If applicable, add screenshots to help explain your problem.
Flow If applicable, add exported flow in order to help replicating the problem.
Setup
npx flowise start
,pnpm start
]pnpm start
Additional context Add any other context about the problem here. The LLM responses works fine when accessed through Jupyter notebook using Langchain Python code. See screenshot.