Is it possible to use Kor along with ConversationalRetrievalChain or ConversationBufferMemory?

[RamkumarManavalan]

For a while, I have been using ConversationalRetrievalChain and PydanticOutputParser to extract data from a document (say PDF or CSV) in to an Object. I used ConversationalRetrievalChain in order to leverage the chat_history feature.

Recently, I have been trying my hand at Kor to extract data from user's prompt/text in to Objects. It works fine. However, I have a use case where the user may not have entered all data required for me to do an action. For example, see below:

My Object has fields such as form name and assignee.

If the user types just "I need to assign the ABC form", I can not do anything until I know who to assign to. Instead of letting the user type "I need to assign the ABC form to Mary", I would like the conversation to leverage the chat_history feature, such as below.

User: "I need to assign the ABC form"
Bot: "Sure, who do you want to assign to?"
User: Mary
Now, Kor should get me the object with both the form name and the assignee name.

Is the above possible with Kor? Are there better alternatives in langchain (OpenAI)? Thanks in advance for any help/ideas.

[dosubot[bot]]

🤖

Hi there,

Thanks for reaching out with your question. It's great to hear that you're exploring the capabilities of Kor and the ConversationalRetrievalChain in LangChain.

To better assist you, I need a bit more information about your current implementation. Could you please provide some details about how you're currently using Kor and ConversationalRetrievalChain in your application? Specifically, I'm interested in how you're handling user input and how you're integrating it with the chat_history feature.

Additionally, could you share some code snippets of your current implementation? This would help me understand your setup better and provide a more accurate solution. Please remember to remove or replace any sensitive information in the code before sharing.

Looking forward to your response.

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.