ValueError: Got unexpected message type: AIMessageChunk #804

1nnovat1on · 2023-10-16T19:10:11Z

Policy and info

Maintainers will close issues that have been stale for 14 days if they contain relevant answers.
Adding the label "sweep" will automatically turn the issue into a coded pull request. Works best for mechanical tasks. More info/syntax at: https://docs.sweep.dev/

Expected Behavior

I typed in a prompt after setting it up. When it responds with clarifying questions, I expect to be able to respond.

Current Behavior

It fails with the error message in the title.

Failure Information

Windows, GPT-4, CMD,

Steps to Reproduce

If possible, provide detailed steps for reproducing the issue.

pip install gpt-engineer
Created folder
enter gpt-engineer "C:\python\test_project" "gpt-4"
Get back some feedback, respond to a few
then error

Failure Logs

│ s.py:336 in _message_from_dict │
│ │
│ 333 │ elif _type == "function": │
│ 334 │ │ return FunctionMessage(**message["data"]) │
│ 335 │ else: │
│ ❱ 336 │ │ raise ValueError(f"Got unexpected message type: {_type}") │
│ 337 │
│ 338 │
│ 339 def messages_from_dict(messages: List[dict]) -> List[BaseMessage]: │
│ │
│ ╭─────────────────────────────────────────── locals ───────────────────────────────────────────╮ │
│ │ _type = 'AIMessageChunk' │ │
│ │ message = { │ │
│ │ │ 'type': 'AIMessageChunk', │ │
│ │ │ 'data': { │ │
│ │ │ │ 'content': 'Summary of areas that need clarification:\n\n1. The specific │ │
│ │ type of reinforcement'+377, │ │
│ │ │ │ 'additional_kwargs': {}, │ │
│ │ │ │ 'type': 'AIMessageChunk', │ │
│ │ │ │ 'example': False │ │
│ │ │ } │ │
│ │ } │ │
│ ╰──────────────────────────────────────────────────────────────────────────────────────────────╯ │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ValueError: Got unexpected message type: AIMessageChunk

sirpeebs · 2023-10-17T12:34:07Z

I Just had the same error and found this via Google search. Running the most recent version of gpt-engineer and have previous had other versions work in my WSL setup on this same box.

/home/pb/venvs/v1/lib/python3.11/site-packages/langchain/schema/messages.py:336 in │
│ _message_from_dict │
│ │
│ 333 │ elif _type == "function": │
│ 334 │ │ return FunctionMessage(**message["data"]) │
│ 335 │ else: │
│ ❱ 336 │ │ raise ValueError(f"Got unexpected message type: {_type}") │
│ 337 │
│ 338 │
│ 339 def messages_from_dict(messages: List[dict]) -> List[BaseMessage]: │
│ │
│ ╭─────────────────────────────────────────── locals ───────────────────────────────────────────╮ │
│ │ _type = 'AIMessageChunk' │ │
│ │ message = { │ │
│ │ │ 'type': 'AIMessageChunk', │ │
│ │ │ 'data': { │ │
│ │ │ │ 'content': 'Summary of areas that need clarification:\n\n1. The specific │ │
│ │ functionalities and c'+626, │ │
│ │ │ │ 'additional_kwargs': {}, │ │
│ │ │ │ 'type': 'AIMessageChunk', │ │
│ │ │ │ 'example': False │ │
│ │ │ } │ │
│ │ } │ │
│ ╰──────────────────────────────────────────────────────────────────────────────────────────────╯ │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ValueError: Got unexpected message type: AIMessageChunk

wwencel · 2023-10-18T09:00:05Z

Traceback (most recent call last):

  File "/Users/wlodek/dev/venv-gpt-eng/bin/gpt-engineer", line 8, in <module>
    sys.exit(app())

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/gpt_engineer/main.py", line 96, in main
    messages = step(ai, dbs)

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/gpt_engineer/steps.py", line 192, in gen_clarified_code
    messages = AI.deserialize_messages(dbs.logs[clarify.__name__])

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/gpt_engineer/ai.py", line 216, in deserialize_messages
    return list(messages_from_dict(json.loads(jsondictstr)))  # type: ignore

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/langchain/schema/messages.py", line 348, in messages_from_dict
    return [_message_from_dict(m) for m in messages]

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/langchain/schema/messages.py", line 348, in <listcomp>
    return [_message_from_dict(m) for m in messages]

  File "/Users/wlodek/dev/venv-gpt-eng/lib/python3.10/site-packages/langchain/schema/messages.py", line 336, in _message_from_dict
    raise ValueError(f"Got unexpected message type: {_type}")

ValueError: Got unexpected message type: AIMessageChunk

same error I got, but it looks like langchain issue

0x1pikachu · 2023-10-18T14:28:25Z

Traceback (most recent call last):

  File "<frozen runpy>", line 198, in _run_module_as_main

  File "<frozen runpy>", line 88, in _run_code

  File "E:\VSC Workspaces\venv\Scripts\gpt-engineer.exe\__main__.py", line 7, in <module>
    sys.exit(app())
             ^^^^^

               ^^^^^^^^^^^^^

  File "E:\VSC Workspaces\venv\Lib\site-packages\gpt_engineer\steps.py", line 192, in gen_clarified_code
    messages = AI.deserialize_messages(dbs.logs[clarify.__name__])
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "E:\VSC Workspaces\venv\Lib\site-packages\gpt_engineer\ai.py", line 216, in deserialize_messages
    return list(messages_from_dict(json.loads(jsondictstr)))  # type: ignore
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "E:\VSC Workspaces\venv\Lib\site-packages\langchain\schema\messages.py", line 348, in messages_from_dict
    return [_message_from_dict(m) for m in messages]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "E:\VSC Workspaces\venv\Lib\site-packages\langchain\schema\messages.py", line 348, in <listcomp>
    return [_message_from_dict(m) for m in messages]
            ^^^^^^^^^^^^^^^^^^^^^

  File "E:\VSC Workspaces\venv\Lib\site-packages\langchain\schema\messages.py", line 336, in _message_from_dict
    raise ValueError(f"Got unexpected message type: {_type}")

ValueError: Got unexpected message type: AIMessageChunk

Same here. Anything I can do ?

ATheorell · 2023-10-18T18:06:54Z

I'm a little confused, is this the same problem as in #802 ?
What confuses me is that, without passing the --clarify (not in the reproduction steps), in the latest version (0.1.0), gpt-engineer is no longer asking clarifying questions. In #802, it is explicitly stated that the --clarify flag is passed.

paulternate · 2023-10-19T04:01:11Z

I traced the error back a little bit to a missing handler in C:\Python\lib\site-packages\langchain\schema\messages.py and was able to get past it by adding this to the _message_from_dict function right before the else statement.

elif _type == "AIMessageChunk":
        return AIMessageChunk(**message["data"])

The whole function now looks like this:

def _message_from_dict(message: dict) -> BaseMessage:
    _type = message["type"]
    if _type == "human":
        return HumanMessage(**message["data"])
    elif _type == "ai":
        return AIMessage(**message["data"])
    elif _type == "system":
        return SystemMessage(**message["data"])
    elif _type == "chat":
        return ChatMessage(**message["data"])
    elif _type == "function":
        return FunctionMessage(**message["data"])
    elif _type == "AIMessageChunk":
        return AIMessageChunk(**message["data"])
    else:
        raise ValueError(f"Got unexpected message type: {_type}")

pralad-p · 2023-10-21T09:28:52Z

I think how gpt-engineer deals with deserialization of user-entered messages (because the OP says it didn't happen on the prompt file input, but rather on the subsequent messages) could indicate that the implicit properties added during deserialization (or this AIMessageChunk type cast) has a strong dependency with how Langchain modifies its schema as per its requirements. What I mean is previously, Langchain might have been quite loose with the type of messages it recieved, but now (I assume due to extra features) needed the type separation. Maybe it's a good idea to brainstorm ways on how to loosen the coupling of this dependency. Additional tests in this dept. would be the first step to know the scope of the problem.

ibrahim-sowunmi · 2023-10-21T17:35:15Z

elif _type == "AIMessageChunk":
        return AIMessageChunk(**message["data"])

Fixed it for me on mac.

python3 -c "import site; print(site.getsitepackages())"
Then open the schema with open . and edited the file

ATheorell · 2023-10-21T17:59:33Z

@wwencel @0x1pikachu
Can you rerun on the latest main commit or at tag 0.1.0?

The stack trace
File "E:\VSC Workspaces\venv\Lib\site-packages\gpt_engineer\ai.py", line 216, in deserialize_messages
return list(messages_from_dict(json.loads(jsondictstr))) # type: ignore
is not up to date.

ATheorell · 2023-10-21T18:01:06Z

@pralad-p agreed.
I think the first thing that needs to be done is to switch to better dependency management so we get less unpleasant surprises when some dependency is updated.

MWals · 2023-10-22T20:30:08Z

elif _type == "AIMessageChunk":
        return AIMessageChunk(**message["data"])
Fixed it for me on mac.

python3 -c "import site; print(site.getsitepackages())" Then open the schema with open . and edited the file

This also fixed the problem for me on Mac. Please consider integrating this into the code.

ATheorell · 2023-10-23T08:19:49Z

@MWals @ibrahim-sowunmi @paulternate
The code you are referring to is inside langchain and not gpt-engineer. Thus, we cannot modify it. It would be awesome if you could request this change at https://github.com/langchain-ai/langchain

ATheorell · 2023-10-23T08:21:15Z

A direct fix here, instead of modifying langchain, is probably to downgrade the langchain version in your environment.
So far, nobody has provided an example that I can reproduce, which would help me in debugging this.

ATheorell · 2023-10-31T09:10:41Z

This is now addressed in PR #833 by restricting the langchain version. Reinstalling from main branch source should solve it.

@baskaryan

… breaks message history runnable. (#15327) - **Description:** fix parse issue for AIMessageChunk when using - **Issue:** #14511 - **Dependencies:** none - **Twitter handle:** none Taken from this fix: gpt-engineer-org/gpt-engineer#804 (comment) Please make sure your PR is passing linting and testing before submitting. Run `make format`, `make lint` and `make test` from the root of the package you've modified to check this locally. See contribution guidelines for more information on how to write/run tests, lint, etc: https://python.langchain.com/docs/contributing/ If you're adding a new integration, please include: 1. a test for the integration, preferably unit tests that do not rely on network access, 2. an example notebook showing its use. It lives in `docs/docs/integrations` directory. If no one reviews your PR within a few days, please @-mention one of @baskaryan, @eyurtsev, @hwchase17. --------- Co-authored-by: Harrison Chase <hw.chase.17@gmail.com>

jonnolen · 2023-12-29T21:45:44Z

the update from this issue landed in upstream langchain really quickly... langchain-ai/langchain#15327. credit to @paulternate

1nnovat1on added bug Something isn't working triage Interesting but stale issue. Will be close if inactive for 3 more days after label added. labels Oct 16, 2023

ATheorell removed the triage Interesting but stale issue. Will be close if inactive for 3 more days after label added. label Oct 18, 2023

ATheorell closed this as completed Oct 31, 2023

ATheorell mentioned this issue Oct 31, 2023

restricting langchain version to <=0.0.312 #833

Merged

This was referenced Dec 29, 2023

RunnableWithMessageHistory streaming bug langchain-ai/langchain#14511

Closed

core: Update messages/__init__.py to account for AIMessageChunk which breaks message history runnable. langchain-ai/langchain#15327

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ValueError: Got unexpected message type: AIMessageChunk #804

ValueError: Got unexpected message type: AIMessageChunk #804

1nnovat1on commented Oct 16, 2023

sirpeebs commented Oct 17, 2023

wwencel commented Oct 18, 2023

0x1pikachu commented Oct 18, 2023

ATheorell commented Oct 18, 2023

paulternate commented Oct 19, 2023 •

edited

Loading

pralad-p commented Oct 21, 2023 •

edited

Loading

ibrahim-sowunmi commented Oct 21, 2023

ATheorell commented Oct 21, 2023

ATheorell commented Oct 21, 2023

MWals commented Oct 22, 2023

ATheorell commented Oct 23, 2023

ATheorell commented Oct 23, 2023

ATheorell commented Oct 31, 2023

jonnolen commented Dec 29, 2023

ValueError: Got unexpected message type: AIMessageChunk #804

ValueError: Got unexpected message type: AIMessageChunk #804

Comments

1nnovat1on commented Oct 16, 2023

Policy and info

Expected Behavior

Current Behavior

Failure Information

Steps to Reproduce

Failure Logs

sirpeebs commented Oct 17, 2023

wwencel commented Oct 18, 2023

0x1pikachu commented Oct 18, 2023

ATheorell commented Oct 18, 2023

paulternate commented Oct 19, 2023 • edited Loading

pralad-p commented Oct 21, 2023 • edited Loading

ibrahim-sowunmi commented Oct 21, 2023

ATheorell commented Oct 21, 2023

ATheorell commented Oct 21, 2023

MWals commented Oct 22, 2023

ATheorell commented Oct 23, 2023

ATheorell commented Oct 23, 2023

ATheorell commented Oct 31, 2023

jonnolen commented Dec 29, 2023

paulternate commented Oct 19, 2023 •

edited

Loading

pralad-p commented Oct 21, 2023 •

edited

Loading