RedisChatMessageHistory

[snavid]

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

from langchain_openai import ChatOpenAI
from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.chat_history import BaseChatMessageHistory
from langchain_core.messages import BaseMessage, AIMessage
from typing import List
from langchain_core.pydantic_v1 import BaseModel, Field
from langchain_community.chat_message_histories import ChatMessageHistory
from dotenv import load_dotenv
import os
from .VectorSearch import add_context_to_vector, add_history_to_vector, add_url_to_vector, retrieve_context_from_vector, retrieve_history_from_vector, retrieve_url_from_vector, remove_context_from_vector, remove_url_from_vector
from langchain_redis import RedisChatMessageHistory


# Load environment variables from .env file
load_dotenv()

api_key = os.getenv('OPENAI_API_KEY')
REDIS_URL = os.getenv('REDIS_URL')

def enhance_chatbot():
        llm = ChatOpenAI(
            model= 'gpt-4o-mini',
            api_key= api_key,
            temperature= 0.2,
            verbose= True, #Allow us to debug the output usefull for long chains
        )


        prompt = ChatPromptTemplate.from_messages(
            [
                (
                    "system",
                    "formart the input begin with you are a customer service provider for {{company name / bussiness name / product name / platform name and others you have to answer questions according to this context Context goes here... }}.Then proceed keeping the meaning of the message. Do not do any attempt to reduce the contetxt of the message leaving usefull information. do not say Here is the formatted information: say here is the context:",
                ),
                 
                 
                ("human", "the input is : {input}")
            ]
        )
        
        chain = prompt | llm
        
        return chain
    




def get_redis_history(session_id: str) -> BaseChatMessageHistory:
    return RedisChatMessageHistory(session_id, redis_url="redis://localhost:6379", 
    ttl=3600, index_name='chat_history')




def chatbot(user, input:None):
            llm = ChatOpenAI(
            model= 'gpt-4o-mini',
            api_key= api_key,
            temperature= user.AgentTemp,
            verbose= True, #Allow us to debug the output usefull for long chains
        )
            if input:
                context = retrieve_context_from_vector(AgentKey=user.AgentKey, query=input)['documents'][0]
                print(context[0])
            else:
                context = ["", "", "", "", "", ""]

            prompt = ChatPromptTemplate.from_messages(
            [
                (
                    "system",
                    f"Your are in conversation your name is {user.AgentName}. Provide clear and short answers the best you can. your context is :" + context[0] + "when you receive a human message (start). You have to begin the conversion",
                ),
                MessagesPlaceholder(variable_name="chat_history"),
                ("{input}")
            ]
            )
            botchain = prompt | llm
            
            chain_with_message_history = RunnableWithMessageHistory(
            botchain,
            get_redis_history,
            input_messages_key="input",
            history_messages_key="chat_history",
        )

            return chain_with_message_history

Description

I'm working on storing conversation history in a Redis database, but I'm encountering an issue. Specifically, the stored history appears to be empty, and when attempting to access it, the application retrieves an empty list.

Here are some details about my setup and issue:

Redis Connection: The Redis database is connected correctly.

The conversation history stored in Redis is not appearing as expected.
When accessing the stored history, I receive an empty list.
Configuration and Code:

[FT.CREATE chat_history ON HASH PREFIX 1 chat_history: SCHEMA session_id TEXT SORTABLE message TEXT SORTABLE timestamp NUMERIC SORTABLE
.] on redis cli.
Questions:

Are there any specific Redis configurations or settings that I might be missing?

1. Could there be an issue with how I'm storing or retrieving the conversation history?
2. Are there common pitfalls or troubleshooting steps for this type of issue?
3. Any help or suggestions would be greatly appreciated. Thank you!

System Info

python 311
Docker image redis
windows 11

[dosubot[bot]]

Hey there, @snavid! How can I assist you today?

I found some similar issues and discussions that might help you resolve the problem with storing and retrieving conversation history in Redis using RedisChatMessageHistory:

1. Similar Solved Discussion:

   • create_retrieval_chain does not store RedisChatMessageHistory suggests manually logging the chat history using add_user_message and add_ai_message in RedisChatMessageHistory every time you invoke the create_retrieval_chain ^[1].

2. Similar Unsolved Discussions:

   • Integration Issues with LangGraph, RedisChatMessageHistory, and RunnableWithMessageHistory ^[2].
   • RunnableWithMessageHistory where input and output are Messages produces error ^[3].
   • ChatMessageHistory does not save subsequent HumanMessage ^[4].

3. Similar Open Issues:

   • Not able to store chat_history while Implementing Memory in ConversationalRetreival Chain ^[5].
   • ConversationBufferMemory+ConversationalRetrievalChain Can't remember history ^[6].
   • Conversationalbuffermemory to maintain chathistory ^[7].

To resolve the issue of storing and retrieving conversation history in Redis using RedisChatMessageHistory, ensure that you are correctly adding messages to the history. Here is a step-by-step guide:

1. Install Required Dependencies:
   Ensure you have the necessary packages installed and a Redis instance running.

   %pip install -qU langchain-redis langchain-openai redis

   You can start a Redis server using Docker:

   docker run -d -p 6379:6379 redis:latest

2. Instantiate the RedisChatMessageHistory object:

   from langchain_redis import RedisChatMessageHistory
   
   history = RedisChatMessageHistory(
       session_id="your-session-id",
       redis_url="redis://localhost:6379"
   )

3. Add messages to the history:

   from langchain_core.messages import AIMessage, HumanMessage
   
   # Add user message
   history.add_user_message("Hello, AI assistant!")
   
   # Add AI message
   history.add_ai_message("Hello! How can I assist you today?")

4. Retrieve messages from the history:

   print("Chat History:")
   for message in history.messages:
       print(f"{type(message).__name__}: {message.content}")

If the stored history appears empty and retrieves an empty list, ensure the following:

• Check Redis Connection: Verify that the Redis server is running and accessible at the specified URL.
• Correct Session ID: Ensure that the session_id used when adding messages is the same as the one used when retrieving messages.
• Message Format: Ensure that the messages being added are correctly formatted instances of AIMessage or HumanMessage.

Here is a complete example:

from langchain_redis import RedisChatMessageHistory
from langchain_core.messages import AIMessage, HumanMessage

# Instantiate the RedisChatMessageHistory
history = RedisChatMessageHistory(
    session_id="your-session-id",
    redis_url="redis://localhost:6379"
)

# Add messages
history.add_user_message("Hello, AI assistant!")
history.add_ai_message("Hello! How can I assist you today?")

# Retrieve messages
print("Chat History:")
for message in history.messages:
    print(f"{type(message).__name__}: {message.content}")

By following these steps, you should be able to store and retrieve conversation history in Redis effectively ^[8][9][10].

To continue talking to Dosu, mention @dosu.

[ignaceHelsen]

Hey, I just had the same problem. Turns out you shouldn't use any special characters like '-' in keys.