persistent memory to create_react_agent?

[mrctito]

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

No code.

Description

@dosu, How do I implement persistent memory (database and file system) to create_react_agent?

    memory = AsyncSqliteSaver.from_conn_string(":memory:")
    react_agent = create_react_agent(model=self.llm_with_tools, tools=self.tools, messages_modifier=self.prompt_str, checkpointer=memory, debug=True)

    inputs = {"messages": [HumanMessage(content=self.contexto["input"])], "chat_history:": "", "agent_scratchpad": ""}    
    config={"configurable": {"thread_id": "1"}}
    result = await react_agent.ainvoke(inputs, config=config, debug=True)

System Info

langchain==0.2.6
langchain-community==0.2.6
langchain-core==0.2.10
langchain-experimental==0.0.62
langchain-openai==0.1.13
langchain-qdrant==0.1.0
langchain-text-splitters==0.2.1
langchainhub==0.1.20

[dosubot[bot]]

To implement persistent memory using a database and file system for create_react_agent, you need to replace the in-memory storage with a persistent checkpointer. Here is an example using AsyncSqliteSaver with a file-based SQLite database:

from langchain_community.chat_message_histories import SQLChatMessageHistory
from langgraph.prebuilt import create_react_agent
from langchain.schema import HumanMessage

# Initialize with a file-based SQLite database
memory = SQLChatMessageHistory(
    session_id="test_session_id", 
    connection_string="sqlite:///sqlite.db"
)

# Create the react agent with persistent memory
react_agent = create_react_agent(
    model=self.llm_with_tools, 
    tools=self.tools, 
    messages_modifier=self.prompt_str, 
    checkpointer=memory, 
    debug=True
)

# Define inputs and configuration
inputs = {"messages": [HumanMessage(content=self.contexto["input"])], "chat_history": "", "agent_scratchpad": ""}
config = {"configurable": {"thread_id": "1"}}

# Invoke the agent
result = await react_agent.ainvoke(inputs, config=config, debug=True)

In this example:

• session_id is a unique identifier for the session.
• connection_string specifies the path to the SQLite database file (sqlite:///sqlite.db). If the file does not exist, it will be created.

This setup ensures that the agent's state is saved persistently across sessions ^[1][2][3].

To continue talking to Dosu, mention @dosu.