ConversationChain

Chain to have a conversation and load context from memory.

This class is deprecated in favor of RunnableWithMessageHistory. Please refer to this tutorial for more detail: https://python.langchain.com/docs/tutorials/chatbot/

RunnableWithMessageHistory offers several benefits, including:

• Stream, batch, and async support;
• More flexible memory handling, including the ability to manage memory outside the chain;
• Support for multiple threads.

Below is a minimal implementation, analogous to using ConversationChain with the default ConversationBufferMemory:

from langchain_core.chat_history import InMemoryChatMessageHistory
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_openai import ChatOpenAI

store = {}  # memory is maintained outside the chain

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

model = ChatOpenAI(model="gpt-3.5-turbo-0125")

chain = RunnableWithMessageHistory(model, get_session_history)
chain.invoke(
    "Hi I'm Bob.",
    config={"configurable": {"session_id": "1"}},
)  # session_id determines thread

Memory objects can also be incorporated into the get_session_history callable:

from langchain_classic.memory import ConversationBufferWindowMemory
from langchain_core.chat_history import InMemoryChatMessageHistory
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_openai import ChatOpenAI

store = {}  # memory is maintained outside the chain

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

    memory = ConversationBufferWindowMemory(
        chat_memory=store[session_id],
        k=3,
        return_messages=True,
    )
    assert len(memory.memory_variables) == 1
    key = memory.memory_variables[0]
    messages = memory.load_memory_variables({})[key]
    store[session_id] = InMemoryChatMessageHistory(messages=messages)
    return store[session_id]

model = ChatOpenAI(model="gpt-3.5-turbo-0125")

chain = RunnableWithMessageHistory(model, get_session_history)
chain.invoke(
    "Hi I'm Bob.",
    config={"configurable": {"session_id": "1"}},
)  # session_id determines thread