Do I have to close SQLChatMessageHistory() session after adding messages?

[byrocuy]

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

def save_to_history(message, response, config):
    user_id = config.get("configurable").get("user_id")
    chat_history = SQLChatMessageHistory(session_id=user_id, connection_string=MYSQL_CONN_STRING, table_name="message_store")

    chat_history.add_messages(
        [
            HumanMessage(content=message),
            AIMessage(content=response)
        ]
    )

Description

I am trying to use SQLChatMessageHistory() in chatbot to store user message and AI response. The mechanisms is pretty simple, I just add them using add_messages(). What I want to ask is, is my implementation is enough? Shouldn't I have to close the connection? But I don't see any implementation in the SQLChatMessageHistory module.

Thanks

System Info

pip freeze | grep langchain:

langchain-cli==0.0.24
langchain-community==0.2.1
langchain-core==0.2.2
langchain-openai==0.1.8
langchain-postgres==0.0.8
langchain-text-splitters==0.2.0

platform: windows using WSL
python version: 3.9

[dosubot[bot]]

Hey @byrocuy! I'm here to help you with any bugs, questions, or contributions. Let's tackle this together!

Your implementation of SQLChatMessageHistory does not explicitly close the connection after adding messages. It is generally a good practice to ensure that database connections are properly closed to avoid potential resource leaks.

You can modify your implementation to include explicit connection closing. Here is an example:

def save_to_history(message, response, config):
    user_id = config.get("configurable").get("user_id")
    chat_history = SQLChatMessageHistory(session_id=user_id, connection_string=MYSQL_CONN_STRING, table_name="message_store")

    try:
        chat_history.add_messages(
            [
                HumanMessage(content=message),
                AIMessage(content=response)
            ]
        )
    finally:
        chat_history.session_maker.remove()  # Explicitly close the session

This ensures that the session is properly closed after the messages are added ^[1][2].

To continue talking to Dosu, mention @dosu.

[byrocuy]

SQLChatMessageHistory does not have session_maker() method. It throws error AttributeError: 'SQLChatMessageHistory' object has no attribute 'session_maker'

[dosubot[bot]]

To properly close the connection in SQLChatMessageHistory after adding messages, you should use the clear method for synchronous operations and the aclear method for asynchronous operations. These methods ensure that the connection is properly closed and resources are released.

Here is how you can do it:

For synchronous operations:

def save_to_history(message, response, config):
    user_id = config.get("configurable").get("user_id")
    chat_history = SQLChatMessageHistory(session_id=user_id, connection_string=MYSQL_CONN_STRING, table_name="message_store")

    try:
        chat_history.add_messages(
            [
                HumanMessage(content=message),
                AIMessage(content=response)
            ]
        )
    finally:
        chat_history.clear()  # Properly close the connection

For asynchronous operations:

async def save_to_history_async(message, response, config):
    user_id = config.get("configurable").get("user_id")
    chat_history = SQLChatMessageHistory(session_id=user_id, connection_string=MYSQL_CONN_STRING, table_name="message_store")

    try:
        await chat_history.aadd_messages(
            [
                HumanMessage(content=message),
                AIMessage(content=response)
            ]
        )
    finally:
        await chat_history.aclear()  # Properly close the connection

These methods ensure that the connection is properly closed after the messages are added ^[1][2][3].

To continue talking to Dosu, mention @dosu.