Inconsistent behaviour from RunnableWithMessageHistory documentation examples with runnables of different signatures

Checked other resources

[X] I added a very descriptive title to this issue.
[X] I searched the LangChain documentation with the integrated search.
[X] I used the GitHub search to find a similar question and didn't find it.
[X] I am sure that this is a bug in LangChain rather than my code.
[X] The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

Error Message and Stack Trace (if applicable)

No response

Description

Examples of runnables with different signatures:

In the RunnableWithMessageHistory documentation there are three main examples of runnables with different signatures:

Takes a dict as input and returns a BaseMessage.


from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_openai.chat_models import ChatOpenAI
from langchain_community.chat_message_histories import ChatMessageHistory
from langchain_core.chat_history import BaseChatMessageHistory
from langchain_core.runnables.history import RunnableWithMessageHistory

model = ChatOpenAI(api_key=OPENAI_API_KEY) prompt = ChatPromptTemplate.from_messages( [ ( "system", "You're an assistant who's good at {ability}. Respond in 20 words or fewer", ), MessagesPlaceholder(variable_name="history"), ("human", "{input}"), ] ) runnable = prompt | model

store = {}

def get_session_history(session_id: str) -> BaseChatMessageHistory: if session_id not in store: store[session_id] = ChatMessageHistory() return store[session_id]

with_message_history = RunnableWithMessageHistory( runnable, get_session_history, input_messages_key="input", history_messages_key="history", )


2. Messages input, dict output
```python
from langchain_core.messages import HumanMessage
from langchain_core.runnables import RunnableParallel

runnable = RunnableParallel({"output_message": ChatOpenAI()})

def get_session_history(session_id: str) -> BaseChatMessageHistory:
    if session_id not in store:
        store[session_id] = ChatMessageHistory()
    return store[session_id]

with_message_history = RunnableWithMessageHistory(
    runnable,
    get_session_history,
    output_messages_key="output_message",
)

with_message_history.invoke(
    [HumanMessage(content="What did Simone de Beauvoir believe about free will")],
    config={"configurable": {"session_id": "baz"}},
)

Messages input, messages output

runnable = ChatOpenAI(api_key=OPENAI_API_KEY)
with_message_history = RunnableWithMessageHistory(
runnable,
get_session_history,
)

Comparing input schema generated for each

If I run the following lines for each example:

print(runnable.get_input_jsonschema())
print(issubclass(chain.get_input_schema(), BaseModel))
print(issubclass(chain.get_input_schema(),RootModel))

I get the following outputs respectively:

dict input, message output

True --> input schema is subclass of pydantic BaseModel False --> input schema is not a subclass of pydantic RootModel

Messages input, dict output True --> input schema is subclass of pydantic BaseModel False --> input schema is not a subclass of pydantic RootModel
Messages input, messages output

True --> input schema is a subclass of pydantic BaseModel True --> input schema is a subclass of pydantic RootModel

Inconsistencies

Even though both the second and the third examples expect Messages input... the underlying runnable generated a completely different schema type for each...

the schema for the second example contains a required field "root" which is only used with RootModel subclass, yet the field is defined on a BaseModel subclass which I believe is unintended, and is the result of wrapping the runnable --> ChatOpenAI() that has a messages input segnature in a RunnableParallel

runnable = RunnableParallel({"output_message": ChatOpenAI()})

I would expect the second and the third example underlying runnables:

runnable = RunnableParallel({"output_message": ChatOpenAI()})

and

runnable = ChatOpenAI()

to produce similar input schema when i call

runnable.get_input_schema()

Screenshot 2024-10-11 at 9 12 49 PM

I would love to get it clear if the current RunnableParallel get_input_schema behaviour is intended.

System Info

System Information

OS: Darwin OS Version: Darwin Kernel Version 22.1.0: Sun Oct 9 20:14:54 PDT 2022; root:xnu-8792.41.9~2/RELEASE_X86_64 Python Version: 3.11.7 (v3.11.7:fa7a6f2303, Dec 4 2023, 15:22:56) [Clang 13.0.0 (clang-1300.0.29.30)]

Package Information

langchain_core: 0.3.10 langchain: 0.3.3 langchain_community: 0.3.2 langsmith: 0.1.132 langchain_cli: 0.0.24 langchain_openai: 0.2.2 langchain_text_splitters: 0.3.0 langserve: 0.2.1

Optional packages not installed

langgraph

Other Dependencies

aiohttp: 3.9.5 async-timeout: Installed. No version info available. dataclasses-json: 0.6.6 fastapi: 0.110.3 gitpython: 3.1.43 httpx: 0.27.0 jsonpatch: 1.33 langserve[all]: Installed. No version info available. libcst: 1.4.0 numpy: 1.26.4 openai: 1.51.2 orjson: 3.10.3 packaging: 23.2 pydantic: 2.9.2 pydantic-settings: 2.5.2 pyproject-toml: 0.0.10 PyYAML: 6.0.1 requests: 2.32.3 requests-toolbelt: 1.0.0 SQLAlchemy: 2.0.30 sse-starlette: 1.8.2 tenacity: 8.3.0 tiktoken: 0.7.0 tomlkit: 0.12.5 typer[all]: Installed. No version info available. typing-extensions: 4.12.0 uvicorn: 0.23.2

langchain-ai / langchain