Thanks for the blog post @s-udhaya and @jiayi-wu . This has been very helpful to add filters for our RAG application. But I am having a hard time combining this chain with a chain that includes the conversation history.  I originally used the example notebook here 03-advanced-app, 02-advanced-chatbot-chain to build the chain that tracks user message history. Below is the chain I came up with that combines filters and message history. It works when I run it in the notebook with chain.invoke(model_config.get("input_example")) but fails when I try to deploy it to the review app. Do you have an example that combines both the filters and conversation history? Thanks.

# RAG Chain

chain = (
    {
        "question": itemgetter("messages") | RunnableLambda(extract_user_query_string),
        "chat_history": itemgetter("messages") | RunnableLambda(extract_chat_history),
        "formatted_chat_history": itemgetter("messages") | RunnableLambda(format_chat_history_for_prompt),
    }
    | RunnablePassthrough()
    | {
        "context": RunnableBranch(
            (
                lambda x: len(x["chat_history"]) > 0,
                query_rewrite_prompt | model | StrOutputParser(),
            ),
            itemgetter("question"),
        )
        | RunnableBranch(
            (
                # First path: Use configurable_vs_retriever with filters when applicable
                lambda input: "configurable" in input.lower(),
                RunnableLambda(
                    lambda input: configurable_vs_retriever.invoke(
                        input,
                        config=create_configurable_with_filters({"messages": input}, retriever_config),
                    )
                )
            ),
            # Second path: Default to vector_search_as_retriever
            vector_search_as_retriever,
        )
        | RunnableLambda(format_context),
        "formatted_chat_history": itemgetter("formatted_chat_history"),
        "question": itemgetter("question"),
    }
    | prompt
    | model
    | StrOutputParser()
)