How can I pass only the last chat message in history with StreamlitChatMessageHistory?

[KirtiJha]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

msgs = StreamlitChatMessageHistory(key="langchain_messages")
conversational_rag_chain = RunnableWithMessageHistory(
  rag_chain,
  lambda session_id: msgs,,
  input_messages_key="input",
  history_messages_key="chat_history",
  output_messages_key="answer",
)

Description

I am using StreamlitChatMessageHistory to pass chat history in my LLM application. It works fine. But it always pass all the session chat history. And as I continue with the conversation, it cross the LLM context limit soon.

I want to pass only the last message or last 2 messages as history, but I am not able to do that.

I tried to look into documentation and tried few things, but not able to do that.

System Info

╰─ pip freeze
aiohappyeyeballs==2.4.0
aiohttp==3.10.5
aiolimiter==1.1.0
aiosignal==1.3.1
aiostream==0.5.2
altair==5.4.1
annotated-types==0.7.0
anyio==4.4.0
asgiref==3.8.1
async-timeout==4.0.3
asyncio==3.4.3
attrs==24.2.0
backoff==2.2.1
bcrypt==4.2.0
blinker==1.8.2
build==1.2.1
cachetools==5.5.0
certifi==2024.7.4
charset-normalizer==3.3.2
chroma-hnswlib==0.7.6
chromadb==0.5.5
click==8.1.7
coloredlogs==15.0.1
dataclasses==0.6
dataclasses-json==0.6.7
Deprecated==1.2.14
exceptiongroup==1.2.2
fastapi==0.112.2
filelock==3.15.4
FlashRank==0.2.9
flatbuffers==24.3.25
frozenlist==1.4.1
fsspec==2024.6.1
gitdb==4.0.11
GitPython==3.1.43
google-auth==2.34.0
googleapis-common-protos==1.65.0
greenlet==3.0.3
grpcio==1.66.1
grpclib==0.4.7
h11==0.14.0
h2==4.1.0
hpack==4.0.0
httpcore==1.0.5
httptools==0.6.1
httpx==0.27.2
httpx-sse==0.4.0
huggingface-hub==0.24.6
humanfriendly==10.0
hyperframe==6.0.1
ibm-generative-ai==3.0.0
idna==3.8
importlib_metadata==8.4.0
importlib_resources==6.4.4
Jinja2==3.1.4
jsonpatch==1.33
jsonpointer==3.0.0
jsonschema==4.23.0
jsonschema-specifications==2023.12.1
kubernetes==30.1.0
langchain==0.2.15
langchain-community==0.2.14
langchain-core==0.2.36
langchain-text-splitters==0.2.2
langsmith==0.1.106
markdown-it-py==3.0.0
MarkupSafe==2.1.5
marshmallow==3.22.0
mdurl==0.1.2
mmh3==4.1.0
modal==0.64.61
monotonic==1.6
mpmath==1.3.0
multidict==6.0.5
mypy-extensions==1.0.0
narwhals==1.5.5
numpy==1.26.4
oauthlib==3.2.2
onnxruntime==1.16.3
opentelemetry-api==1.27.0
opentelemetry-exporter-otlp-proto-common==1.27.0
opentelemetry-exporter-otlp-proto-grpc==1.27.0
opentelemetry-instrumentation==0.48b0
opentelemetry-instrumentation-asgi==0.48b0
opentelemetry-instrumentation-fastapi==0.48b0
opentelemetry-proto==1.27.0
opentelemetry-sdk==1.27.0
opentelemetry-semantic-conventions==0.48b0
opentelemetry-util-http==0.48b0
orjson==3.10.7
overrides==7.7.0
packaging==24.1
pandas==2.2.2
pillow==10.4.0
posthog==3.6.0
protobuf==4.25.4
pyarrow==17.0.0
pyasn1==0.6.0
pyasn1_modules==0.4.0
pydantic==2.8.2
pydantic_core==2.20.1
pydeck==0.9.1
Pygments==2.18.0
PyPika==0.48.9
pyproject_hooks==1.1.0
python-dateutil==2.9.0.post0
python-dotenv==1.0.1
pytz==2024.1
PyYAML==6.0.2
referencing==0.35.1
requests==2.32.3
requests-oauthlib==2.0.0
rich==13.8.0
rpds-py==0.20.0
rsa==4.9
shellingham==1.5.4
sigtools==4.0.1
six==1.16.0
smmap==5.0.1
sniffio==1.3.1
SQLAlchemy==2.0.32
starlette==0.38.2
streamlit==1.38.0
sympy==1.13.2
synchronicity==0.7.5
tenacity==8.5.0
tokenizers==0.20.0
toml==0.10.2
tomli==2.0.1
tornado==6.4.1
tqdm==4.66.5
tree-sitter==0.21.3
typer==0.12.5
types-certifi==2021.10.8.3
types-toml==0.10.8.20240310
typing-inspect==0.9.0
typing_extensions==4.12.2
tzdata==2024.1
urllib3==2.2.2
uvicorn==0.30.6
uvloop==0.20.0
watchfiles==0.24.0
websocket-client==1.8.0
websockets==13.0.1
wrapt==1.16.0
yarl==1.9.4
zipp==3.20.1
Platform - MacOS
Python 3.9.13

[dosubot[bot]]

Hello, @KirtiJha! I'm here to help you with any bugs, questions, or contributions you may have. Let's tackle this issue together!

To pass only the last chat message in history with StreamlitChatMessageHistory, you can access the last message in the messages list. Here is how you can do it:

from langchain_community.chat_message_histories import StreamlitChatMessageHistory

# Initialize the message history
history = StreamlitChatMessageHistory(key="langchain_messages")

# Add some messages
history.add_user_message("hi!")
history.add_ai_message("whats up?")

# Access the last message
last_message = history.messages[-1]

# Print the last message content
print(last_message.content)

This code snippet initializes the StreamlitChatMessageHistory, adds some messages, and then accesses the last message in the history. The [-1] index is used to get the last item from the list of messages ^[1].

If you want to pass only the last two messages, you can slice the list accordingly:

# Access the last two messages
last_two_messages = history.messages[-2:]

# Print the last two messages content
for message in last_two_messages:
    print(message.content)

This will give you the last two messages in the history. You can then use these messages as needed in your application.

To continue talking to Dosu, mention @dosu.

[KirtiJha]

This is fine. But what I am facing challenge is to pass last chat message as history in RunnableWithMessageHistory. How can I do that?

msgs = StreamlitChatMessageHistory(key="langchain_messages")
conversational_rag_chain = RunnableWithMessageHistory(
  rag_chain,
  lambda session_id: msgs,,
  input_messages_key="input",
  history_messages_key="chat_history",
  output_messages_key="answer",
)

If I pass mags.messages[-1] in lambda function, its throwing error.

[dosubot[bot]]

To pass only the last chat message as history in RunnableWithMessageHistory using StreamlitChatMessageHistory, you need to modify the lambda function to return a new StreamlitChatMessageHistory instance containing only the last message. Here's how you can do it:

from langchain_community.chat_message_histories import StreamlitChatMessageHistory
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_openai import ChatOpenAI
import streamlit as st

# Initialize the message history
msgs = StreamlitChatMessageHistory(key="langchain_messages")

# Ensure there is an initial AI message if the history is empty
if len(msgs.messages) == 0:
    msgs.add_ai_message("How can I help you?")

# Define the prompt template
prompt = ChatPromptTemplate.from_messages(
    [
        ("system", "You are an AI chatbot having a conversation with a human."),
        MessagesPlaceholder(variable_name="history"),
        ("human", "{question}"),
    ]
)

# Create the chain with history
chain = prompt | ChatOpenAI()

# Define the function to get session history with only the last message
def get_session_history(session_id):
    last_message = msgs.messages[-1] if msgs.messages else None
    new_history = StreamlitChatMessageHistory(key="last_message")
    if last_message:
        new_history.add_message(last_message)
    return new_history

chain_with_history = RunnableWithMessageHistory(
    chain,
    get_session_history,  # Function to return the instance created earlier
    input_messages_key="question",
    history_messages_key="history",
)

# Display previous messages
for msg in msgs.messages:
    st.chat_message(msg.type).write(msg.content)

# Handle new user input
if prompt := st.chat_input():
    st.chat_message("human").write(prompt)

    # Invoke the chain with the new question and the existing history
    config = {"configurable": {"session_id": "any"}}
    response = chain_with_history.invoke({"question": prompt}, config)
    st.chat_message("ai").write(response.content)

In this code:

1. The get_session_history function creates a new StreamlitChatMessageHistory instance containing only the last message from the original history.
2. This new history is then used in the RunnableWithMessageHistory instance.

This approach ensures that only the last message is passed as history, preventing the context limit from being exceeded ^[1][2].

To continue talking to Dosu, mention @dosu.