[ERROR] ValueError: Unsupported chat history format: <class 'str'>. Full chat history:

[vinayburam]

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

import json
from langchain_community.retrievers import AmazonKendraRetriever
from langchain.chains import ConversationalRetrievalChain
from langchain.prompts import PromptTemplate
from langchain_openai import AzureChatOpenAI
import os
from prompts import prompt_template1

def build_chain():
  region = os.environ["REGION"]
  kendra_index_id = os.environ["KENDRA_INDEX"]

  #llm = OpenAI(batch_size=5, temperature=0, max_tokens=300)
  # The base URL for your Azure OpenAI resource. e.g. "https://<your resource name>.openai.azure.com"
  #azure_endpoint=""
  #gpt 4 config
  azure_endpoint =  ""  
  # API version e.g. ""
  #openai_api_version=""
  openai_api_version="2"
  
  # The name of your Azure OpenAI deployment chat model. e.g. "gpt-35-turbo-0613"
  #deployment_name="gpt-35-turbo"
  deployment_name="gpt-4o"
  # The API key for your Azure OpenAI resource.
  #openai_api_key = ""
  
  # This is set to `azure`
  openai_api_type="azure"
  
  
  llm = AzureChatOpenAI(
      azure_endpoint=azure_endpoint,
      openai_api_version=openai_api_version,
      deployment_name=deployment_name,
      openai_api_key=openai_api_key,
      openai_api_type=openai_api_type,
  )
      
  #retriever = AmazonKendraRetriever(index_id=kendra_index_id, region_name=region)
  data_source_id = 'sourceid'
  retriever = AmazonKendraRetriever(index_id=kendra_index_id, top_k=3, 
    attribute_filter={
        'EqualsTo': {
            'Key': '_language_code',
            'Value': {'StringValue': 'en'}
        },
        'EqualsTo': {
            'Key': '_data_source_id',
            'Value': {'StringValue': data_source_id}
        }
    }) 

  '''prompt_template = """
  The following is a friendly conversation between a human and an AI. 
  The chatbot is talkative and provides lots of specific details from its context.
  If the chatbot does not know the answer to a question, it suggests some user some products based on the given user input and context
  {context}
  
  Instruction:  Based on the above documents, provide a below answer for, {question}
                Do not add additional information in the response.
                
                Only suggest some topics to search from the fetched content to the user if there is no relavant result. 
                Do not suggest from internet. Always show this information in text
                column.
  Solution:"""
  '''
  prompt_template = prompt_template1
  PROMPT = PromptTemplate(
      template=prompt_template, input_variables=["context", "question"]
  )

  condense_qa_template = """
  Given the following conversation and a follow up question, rephrase the follow up question 
  to be a standalone question.

  Chat History:
  {chat_history}
  Follow Up Input: {question}
  Standalone question:"""
  standalone_question_prompt = PromptTemplate.from_template(condense_qa_template)

  qa = ConversationalRetrievalChain.from_llm(
        llm=llm, 
        retriever=retriever, 
        condense_question_prompt=standalone_question_prompt, 
        return_source_documents=True, 
        combine_docs_chain_kwargs={"prompt":PROMPT})
    
  return qa
  
def run_chain(chain, prompt: str, history=[]):
  return chain({"question": prompt, "chat_history": history})

Description

when i m trying to pass the chat history getting unsupported [ERROR] ValueError: Unsupported chat history format: <class 'str'>. Full chat history:
can you please help which format is supported

System Info

ok

[dosubot[bot]]

@vinayburam Hello there! I'm here to assist you with any issues you might be experiencing. If you need help solving bugs, answering questions, or becoming a contributor, I'm here to help while we wait for a human maintainer.

The error "ValueError: Unsupported chat history format: <class 'str'>" occurs because the chat history is not in the expected format. The supported format for the chat history is a list of tuples, where each tuple contains two strings, or a list of BaseMessage objects.

Here is how you can modify your run_chain function to use the correct format for the chat history:

1. Convert the chat history to a list of tuples:

   def run_chain(chain, prompt: str, history=[]):
       # Convert chat history to list of tuples if it's not already in the correct format
       if isinstance(history, list) and all(isinstance(item, str) for item in history):
           history = [(history[i], history[i + 1]) for i in range(0, len(history), 2)]
       return chain({"question": prompt, "chat_history": history})

2. Ensure the chat history is a list of BaseMessage objects:

   from langchain_core.messages import BaseMessage, HumanMessage, AIMessage
   
   def run_chain(chain, prompt: str, history=[]):
       # Convert chat history to list of BaseMessage objects if it's not already in the correct format
       if isinstance(history, list) and all(isinstance(item, str) for item in history):
           history = [HumanMessage(content=history[i]) if i % 2 == 0 else AIMessage(content=history[i]) for i in range(len(history))]
       return chain({"question": prompt, "chat_history": history})

Choose one of the above methods based on your preference. The first method converts the chat history to a list of tuples, while the second method converts it to a list of BaseMessage objects.

This should resolve the error and allow your ConversationalRetrievalChain to process the chat history correctly ^[1][2].

To continue talking to Dosu, mention @dosu.

[vinayburam]

Thanks it worked. but when i am using history with langchain getting the same resopnse again and again for diff quries>

History from run chain function [('User: who is the suplier for Distributor liquid and gas handling, seals, MRO\nBot: \nUser: who is the suplier for Distributor liquid and gas handling, seals, MRO\nBot: Category: Electrical & Mechanical OEM|Fluid and Gas Distribution (OEM)|Fluid and Gas Static Components|o-rings, seals & gaskets \nSupplierName: eeeee \nDescription: Distributor liquid and gas handling, seals, MRO \nPrimary Contact: Mmmmm l \nAlternate Contact: Dddd \nGlobal Contact: Miiii \nUser: Distributor liquid and gas handling, seals, MRO give me primary contact\nBot: \nUser: Distributor liquid and gas handling, seals, MRO give me primary contact\nBot: Solution:\n\n### Summary Response\n\nCategory: 661 Electrical & Mechanical OEM|Fluid and Gas Distribution (OEM)|Fluid and Gas Static Components\n\nSupplierName: eeee\n\nDescription: Distributor liquid and gas handling, seals, MRO\n\nPrimary Contact: Macintyre, Paul\n\nAlternate Contact: Ddddd\n\nGlobal Contact: Miiii\n\n---\n\n### Handling No Results\n- If no relevant result was found, consider looking into:\n - Supplier contact information for other categories within Electrical & Mechanical OEM.\n - Additional details on Fluid and Gas Distribution components.\n - Contact information for other suppliers listed in the provided excerpts.\nUser: what is category for this\nBot: ', '')]

Here in the above resopnse you can see i am getting same resopnse for each query

[dosubot[bot]]

The issue of receiving the same response for different queries in the ConversationalRetrievalChain can be due to several factors, including how the chat history is being processed and how the chain is configured. Here are a few steps to troubleshoot and resolve this issue:

1. Ensure Proper Chat History Formatting:
   Make sure the chat history is correctly formatted as a list of tuples or BaseMessage objects. It seems you have already addressed this, but double-check to ensure consistency.

2. Check the Configuration of the Chain:
   Ensure that the ConversationalRetrievalChain is properly configured to handle different queries. Specifically, check the combine_docs_chain_kwargs and condense_question_prompt to ensure they are set up correctly.

3. Review the Prompt Templates:
   Ensure that the prompt templates used for generating standalone questions and combining documents are correctly defined and do not inadvertently lead to repetitive responses.

4. Debug the Chain Execution:
   Add debug statements to inspect the intermediate outputs, such as the generated standalone question and the retrieved documents, to identify where the repetition might be occurring.

Here is an updated version of your run_chain function with additional debug statements:

from langchain_core.messages import BaseMessage, HumanMessage, AIMessage

def run_chain(chain, prompt: str, history=[]):
    # Convert chat history to list of BaseMessage objects if it's not already in the correct format
    if isinstance(history, list) and all(isinstance(item, str) for item in history):
        history = [HumanMessage(content=history[i]) if i % 2 == 0 else AIMessage(content=history[i]) for i in range(len(history))]
    
    # Debug: Print the chat history
    print("Chat History:", history)
    
    # Run the chain and capture the output
    output = chain({"question": prompt, "chat_history": history})
    
    # Debug: Print the output
    print("Output:", output)
    
    return output

Additionally, ensure that the ConversationalRetrievalChain is correctly set up to handle different queries. Here is an example configuration:

from langchain.chains import ConversationalRetrievalChain
from langchain.prompts import PromptTemplate

condense_qa_template = """
Given the following conversation and a follow up question, rephrase the follow up question 
to be a standalone question.

Chat History:
{chat_history}
Follow Up Input: {question}
Standalone question:"""
standalone_question_prompt = PromptTemplate.from_template(condense_qa_template)

qa_template = """
You are an assistant for question-answering tasks.
Use the following pieces of retrieved context to answer
the question. If you don't know the answer, say that you
don't know. Use three sentences maximum and keep the
answer concise.

Chat History:
{chat_history}

Other context:
{context}

Question: {question}
"""
qa_prompt = PromptTemplate.from_template(qa_template)

convo_qa_chain = ConversationalRetrievalChain.from_llm(
    llm=llm, 
    retriever=retriever, 
    condense_question_prompt=standalone_question_prompt, 
    return_source_documents=True, 
    combine_docs_chain_kwargs={"prompt": qa_prompt}
)

By ensuring the chat history is correctly formatted and the chain is properly configured, you should be able to resolve the issue of repetitive responses ^[1][2].

To continue talking to Dosu, mention @dosu.