How do I use create_react_agent in langServe?

[Jimmy-L99]

Server test code as follows:

llm_end_point_url = "http://172.16.21.155:8000/v1/"
model = ChatOpenAI(model="glm4v-9b",base_url=llm_end_point_url, api_key="api_key")

### embedding ###
embedding_model = HuggingFaceEmbeddings(model_name='/root/ljm/bge/bge-large-zh-v1.5')

### milvus ###
milvus_host = "***"
milvus_port = ***
collection_name = "***"

vector_store = Milvus(
    embedding_function=embedding_model,
    collection_name="langchain_lichi_txt",
    connection_args={"host": milvus_host, "port": milvus_port, "db_name": "***"}, 
)

retriever = vector_store.as_retriever(search_type="similarity", search_kwargs={"k": 3})

RAG = create_retriever_tool(
    retriever=retriever,
    name="***",
    description="***",
)
tools = [RAG]

app = FastAPI(
    title="GLM4 LangChain Server",
    version="1.0",
    description="A simple api server using Langchain's Runnable interfaces",
)

agent_executor = create_react_agent(model, tools, debug=True)
rag_chain = ConversationalRetrievalChain.from_llm(
    llm=model,
    retriever=retriever
)
add_routes(
    app,
    agent_executor,
    path="/Litchi_RAG",
)

if __name__ == "__main__":
    import uvicorn

    uvicorn.run(app, host="172.16.21.155", port=8010)

Client test code as follows:

rag_chain = RemoteRunnable("http://172.16.21.155:8010/RAG")
query = "********?"
response = rag_chain.invoke({"question": query})
print(response)

Server error:

  File "/root/anaconda3/envs/ljm_glm4_conda/lib/python3.11/site-packages/langchain_core/messages/utils.py", line 301, in _convert_to_message
    raise NotImplementedError(f"Unsupported message type: {type(message)}")
NotImplementedError: Unsupported message type: <class 'NoneType'>

Q:

If I use create_react_agent, always have some error, but it will work if I use ConversationalRetrievalChain in add_routes. How should I fix it if I want to use create_react_agent?

[eyurtsev]

https://github.com/langchain-ai/langserve?tab=readme-ov-file#%EF%B8%8F-langgraph-compatibility

[Jimmy-L99]

@eyurtsev Thank you for your answer, I saw something about agent from the example, but most of them did not talk about create_retriever_tool and the create_react_agent of langgraph.prebuilt,
I still couldn't get an answer with the agent tool according to my code, I would appreciate it if you could give me more specific examples! TvT