langflow icon indicating copy to clipboard operation
langflow copied to clipboard
verified publisher icon langflow-ai

Session ID Error When Using Agent Within a Sub Flow

Open jmelm93 opened this issue 1 year ago • 3 comments

Bug Description

Description
When creating a “Sub Flow” component that references a flow containing an Agent, an error is repeatedly thrown:
"It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields."

Despite investigating the source code, I haven’t yet located the exact cause or solution.

Steps to Reproduce

  1. Create the first flow (e.g., “Main Flow”) with the following structure:
    • Chat Message → Sub Flow
      • The Sub Flow points to the second flow.
  2. Create the second flow (e.g., “Basic Agent Flow”) with:
    • Chat Message (input) → Agent (input)
    • Calculator (Tool) → Agent (tool)
    • The output goes back to Chat Message.
  3. Send a message in the first flow (e.g., “2x2”).

Basic Agent Flow.json Main Flow.json

Expected Result
The flow should process the conversation seamlessly, including the Agent’s response from the sub flow.

Actual Result
The conversation fails with:
"It looks like we're missing some important information: session_id (unique conversation identifier)..."

Full stack trace

Error building Component Sub Flow: 

Error running graph: Error building Component Agent: 

It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields.

Traceback (most recent call last):
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 837, in build_results
    return await self._build_with_tracing()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 819, in _build_with_tracing
    _results, _artifacts = await self._build_results()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 883, in _build_results
    result = await method()
  File "<string>", line 88, in message_response
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/base/agents/agent.py", line 158, in run_agent
    result = await process_agent_events(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/base/agents/events.py", line 253, in process_agent_events
    agent_message = await asyncio.to_thread(send_message_method, message=agent_message)
  File "/Users/jasonmelman/.pyenv/versions/3.10.14/lib/python3.10/asyncio/threads.py", line 25, in to_thread
    return await loop.run_in_executor(None, func_call)
  File "/Users/jasonmelman/.pyenv/versions/3.10.14/lib/python3.10/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 1009, in send_message
    stored_message = self._store_message(message)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 1034, in _store_message
    messages = store_message(message, flow_id=flow_id)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/memory.py", line 179, in store_message
    raise ValueError(msg)
ValueError: It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/vertex/base.py", line 709, in _build_results
    result = await initialize.loading.get_instance_results(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/interface/initialize/loading.py", line 68, in get_instance_results
    return await build_component(params=custom_params, custom_component=custom_component)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/interface/initialize/loading.py", line 145, in build_component
    build_results, artifacts = await custom_component.build_results()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 848, in build_results
    self.send_error(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 1135, in send_error
    self.send_message(error_message)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 1009, in send_message
    stored_message = self._store_message(message)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 1034, in _store_message
    messages = store_message(message, flow_id=flow_id)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/memory.py", line 179, in store_message
    raise ValueError(msg)
ValueError: It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 771, in _run
    await self.process(start_component_id=start_component_id, fallback_to_env_vars=fallback_to_env_vars)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 1518, in process
    next_runnable_vertices = await self._execute_tasks(tasks, lock=lock)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 1570, in _execute_tasks
    raise result
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 1419, in build_vertex
    await vertex.build(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/vertex/base.py", line 819, in build
    await step(user_id=user_id, event_manager=event_manager, **kwargs)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/vertex/base.py", line 497, in _build
    await self._build_results(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/vertex/base.py", line 724, in _build_results
    raise ComponentBuildError(msg, tb) from exc
langflow.exceptions.component.ComponentBuildError: Error building Component Agent: 

It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/vertex/base.py", line 709, in _build_results
    result = await initialize.loading.get_instance_results(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/interface/initialize/loading.py", line 68, in get_instance_results
    return await build_component(params=custom_params, custom_component=custom_component)
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/interface/initialize/loading.py", line 145, in build_component
    build_results, artifacts = await custom_component.build_results()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 837, in build_results
    return await self._build_with_tracing()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 819, in _build_with_tracing
    _results, _artifacts = await self._build_results()
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/component.py", line 883, in _build_results
    result = await method()
  File "<string>", line 100, in generate_results
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/custom/custom_component/custom_component.py", line 490, in run_flow
    return await run_flow(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/helpers/flow.py", line 123, in run_flow
    return await graph.arun(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 840, in arun
    run_outputs = await self._run(
  File "/Users/jasonmelman/dev/langflow_testing/.venv/lib/python3.10/site-packages/langflow/graph/graph/base.py", line 776, in _run
    raise ValueError(msg) from exc
ValueError: Error running graph: Error building Component Agent: 

It looks like we're missing some important information: session_id (unique conversation identifier). Please ensure that your message includes all the required fields.

Reproduction

  1. Create the first flow (e.g., “Main Flow”) with the following structure:
    • Chat Message → Sub Flow
      • The Sub Flow points to the second flow.
  2. Create the second flow (e.g., “Basic Agent Flow”) with:
    • Chat Message (input) → Agent (input)
    • Calculator (Tool) → Agent (tool)
    • The output goes back to Chat Message.
  3. Send a message in the "Main Flow" (e.g., “2x2+3”).

Expected behavior

The flow should process the conversation seamlessly, including the Agent’s response from the sub flow.

Who can help?

No response

Operating System

MacOS Sequoia 15.1.1 (24B91)

Langflow Version

1.1.1

Python Version

3.10

Screenshot

No response

Flow File

Basic Agent Flow.json Main Flow.json

jmelm93 avatar Dec 24 '24 23:12 jmelm93

Hello, @jmelm93! I'm here to help you with any bugs, questions, or contributions you have. Let's squash those bugs together! 😄

To resolve the "session_id (unique conversation identifier) missing" error when using an Agent within a Sub Flow in Langflow, ensure that the session_id is properly set in your agent's context. The session_id is typically retrieved from either self.graph.session_id or self._session_id. If neither is available, session_id will be None, which could lead to the error.

Here's what you can do:

  1. Check if self.graph is properly initialized: Ensure that self.graph is set and has a session_id attribute before running the agent.

  2. Set _session_id if self.graph is not used: If self.graph is not used or does not have a session_id, make sure to set self._session_id with a unique identifier before running the agent.

  3. Debugging: Add debugging statements to check if self.graph and self._session_id are set correctly before the agent is executed.

Example code to ensure session_id is set:

if hasattr(self, "graph") and hasattr(self.graph, "session_id"):
    session_id = self.graph.session_id
elif hasattr(self, "_session_id"):
    session_id = self._session_id
else:
    # Set a default session_id or raise an error
    session_id = "default_session_id"  # or raise an error

Ensure that the session_id is correctly set in your flow to avoid the error [1].

To continue talking to Dosu, mention @dosu.

Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other

dosubot[bot] avatar Dec 24 '24 23:12 dosubot[bot]

The above bot answer does not resolve the problem. When logging self.graph.session_id within the "Sub Flow" component, the session_id is present, as shown in the log output below:

Log output:
2024-12-24 15:42:10 - INFO - <string>:92 <string> - self.graph.session_id: f053ea08-de8c-4844-a24e-fe5d0fcaaf80

However, this context does not appear to propagate to the components being built by the "Sub Flow." This seems to be an issue with the default behavior of the out-of-the-box "Sub Flow" component.

jmelm93 avatar Dec 24 '24 23:12 jmelm93

having the same issue. any solution?

Amakua avatar Dec 29 '24 21:12 Amakua

No, I’ve kind of given up on this after trying to resolve it for several days. Hopefully, this issue will get a response at some point down the line.

jmelm93 avatar Jan 07 '25 18:01 jmelm93

Hi @Amakua and @jmelm93

Thank you for your feedback. We are currently developing a new component called RunFlow, which will function similarly to both Sub Flow and Flow as Tool.

Please test your use case with the new component.

Here is the PR: https://github.com/langflow-ai/langflow/pull/5518.

If you continue to experience issues after testing, please reopen this issue. I look forward to hearing your feedback.

edwinjosechittilappilly avatar Jan 13 '25 23:01 edwinjosechittilappilly