Agent Invoke method

[Harish-ilink]

@router.post("/invoke")
async def stream_run(
payload: CreateRunPayload, # for openapi docs
request: Request,
userid: OpengptsUserId,
tenantid: str = Header(None),
):
"""Create a invoke."""
input_request = payload.dict(exclude_unset=True)
input_, config = await _run_input_and_config(request, userid)
config = update_bitc_config(userid,tenantid,config)
return agent.invoke(input=input_request,config=config)

Im trying to call invoke method instead of streaming my bots response.But I was getting error .I have tried aiinvoke() method also.But Im ending up in same error.

File "/usr/local/lib/python3.11/site-packages/langgraph/channels/binop.py", line 58, in update
2024-07-16 10:55:18 self.value = self.operator(self.value, value)
2024-07-16 10:55:18 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2024-07-16 10:55:18 File "/usr/local/lib/python3.11/site-packages/langgraph/graph/message.py", line 22, in add_messages
2024-07-16 10:55:18 if m.id is None:
2024-07-16 10:55:18 ^^^^
2024-07-16 10:55:18 AttributeError: 'dict' object has no attribute 'id'

[vbarda]

@Harish-ilink what langgraph version are you using? this shouldn't be happening in the latest version

[Harish-ilink]

I was using lesser version (0.0.8). Now when i upgraded the version .One method is been missed from agent.py.
Error :
File "/backend/app/core/storage.py", line 6, in
2024-07-17 12:39:12 from app.core.agent import AgentType, get_agent_executor
2024-07-17 12:39:12 File "/backend/app/core/agent.py", line 9, in
2024-07-17 12:39:12 from langgraph.checkpoint import CheckpointAt
2024-07-17 12:39:12 ImportError: cannot import name 'CheckpointAt' from 'langgraph.checkpoint' (/usr/local/lib/python3.11/site-packages/langgraph/checkpoint/init.py)

Seems like this 'CheckpointAt' method is not availbale now.What can i use instead?

[vbarda]

i am not sure, it's hard to tell without seeing how you set up your graph / checkpointing. i would recommend checking out latest examples for adding persistence/checkpointing here https://langchain-ai.github.io/langgraph/how-tos/persistence/