You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This will allow us to build a true agentic workflow as opposed to the static chain implementation that we have right now.
This should help reduce hallucination and generation of incorretly-syntaxed JSON
Definition of Done
I can define an agent using Langchain's react_agent class, and I can declare nlq-to-vds as an available tool. I can pass a natural language query and Tableau Published Data Source metadata to the ReAct agent. The agent will invoke the nlq-to-vds tool and output syntactically correct structured JSON that could be passed through the query-datasource endpoint exposed by VizQL Data Service.
The text was updated successfully, but these errors were encountered:
Convert the NLQ-to-VDS prompt into a langchain tool - https://python.langchain.com/v0.1/docs/modules/model_io/chat/function_calling/
why?
Definition of Done
I can define an agent using Langchain's react_agent class, and I can declare nlq-to-vds as an available tool. I can pass a natural language query and Tableau Published Data Source metadata to the ReAct agent. The agent will invoke the nlq-to-vds tool and output syntactically correct structured JSON that could be passed through the query-datasource endpoint exposed by VizQL Data Service.
The text was updated successfully, but these errors were encountered: