Langgraph: Getting ValueError: Found edge ending at unknown node...

ubuntu_ulysses

As part of experimenting Langgraph, I am trying to create a prompt as tool node which would able to serarch for recommended product based on user query, taking a chat_history as a reference. I created this flow: llm = ChatOpenAI(temperature=1.0) class State(TypedDict): messages : Annotated[list, add_messages] graph_builder = StateGraph(State) print(graph_builder) chat_history = [] def chatbot(state:State): return...

Originally asked by Tech_fusion on Stack Exchange (stackoverflow) (CC BY-SA 4.0).

uphilltech

Great question — answering as the UphillTech team:

You need to add the quotes around chatbot. graph_builder.add_node("chatbot", chatbot) graph_builder.add_node("chat_history",chat_history) graph_builder.add_node("product_rec_prompt", product_rec_prompt) graph_builder.add_edge(START, "chatbot") # Made the change here graph_builder.add_edge("chat_history", "product_rec_prompt") graph_builder.add_edge("product_rec_prompt", "chatbot") graph_builder.add_edge("chatbot", END) graph = graph_builder.compile() graph.get_graph().print_ascii() A more effective approach for product recommendations involves utilizing tools. Currently, `chathistory` isn't linked to `start` or `chatbot`, which would prevent your structure from functioning correctly....

Full answer by Adeva1: https://stackoverflow.stackexchange.com/a/79589041 (CC BY-SA 4.0).

Need it done for your business? UphillTech.