I m prompting a private KB and I asked a

QA_PROMPT_TMPL = (
        "Context information is below. \n"
        "---------------------\n"
        "{context_str}"
        "\n---------------------\n"
        "Given the context information and not prior knowledge, "
        "answer the following question following a similar format to the context."
        "Add (un)ordered lists if applicable. If you're unsure of the answer, say \"Sorry, I don't know\": {query_str}\n"
    )

    QA_PROMPT_TMPL = (
        "Context information is below. \n"
        "---------------------\n"
        "{context_str}"
        "\n---------------------\n"
        "Given the context information and not prior knowledge, "
        "answer the following question following a similar format to the context."
        "Add (un)ordered lists if applicable. If you're unsure of the answer, say \"Sorry, I don't know\": {query_str}\n"
    )

    QA_PROMPT = QuestionAnswerPrompt(QA_PROMPT_TMPL)

    # configure response synthesizer
    response_synthesizer = ResponseSynthesizer.from_args(
        node_postprocessors=[
            SimilarityPostprocessor(similarity_cutoff=0.78)
        ],
        text_qa_template=QA_PROMPT
    )

DEFAULT_REFINE_PROMPT_TMPL = (
    "The original question is as follows: {query_str}\n"
    "We have provided an existing answer: {existing_answer}\n"
    "We have the opportunity to refine the existing answer "
    "(only if needed) with some more context below.\n"
    "------------\n"
    "{context_msg}\n"
    "------------\n"
    "Given the new context, refine the original answer to better "
    "answer the question. "
    "If the context isn't useful, return the original answer."
)
DEFAULT_REFINE_PROMPT = RefinePrompt(DEFAULT_REFINE_PROMPT_TMPL)


DEFAULT_TEXT_QA_PROMPT_TMPL = (
    "Context information is below. \n"
    "---------------------\n"
    "{context_str}"
    "\n---------------------\n"
    "Given the context information and not prior knowledge, "
    "answer the question: {query_str}\n"
)
DEFAULT_TEXT_QA_PROMPT = QuestionAnswerPrompt(DEFAULT_TEXT_QA_PROMPT_TMPL)

# configure response synthesizer
    response_synthesizer = ResponseSynthesizer.from_args(
        node_postprocessors=[
            SimilarityPostprocessor(similarity_cutoff=0.78)
        ],
        text_qa_template=DEFAULT_TEXT_QA_PROMPT 
        refine_template=DEFAULT_REFINE_PROMPT 
    )

# define prompt helper
# set maximum input size
max_input_size = 4096
# set number of output tokens
num_output = 512
# set maximum chunk overlap
max_chunk_overlap = 20
prompt_helper = PromptHelper(max_input_size, num_output, max_chunk_overlap)

# define LLM
llm_predictor = LLMPredictor(llm=OpenAI(temperature=0, model_name="text-davinci-003", max_tokens=num_output))

service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, prompt_helper=prompt_helper)

service_context = None

def construct_index():
    GoogleDriveReader = download_loader('GoogleDriveReader')
    loader = GoogleDriveReader()
    documents = loader.load_data(folder_id=folder_id)

    # define prompt helper

    # set maximum input size
    max_input_size = 4096
    # set number of output tokens
    num_output = 512
    # set maximum chunk overlap
    max_chunk_overlap = 20
    prompt_helper = PromptHelper(max_input_size, num_output, max_chunk_overlap)

    # define LLM
    #text-davinci-003
    #text-ada-001
    llm_predictor = LLMPredictor(llm=OpenAI(temperature=0, model_name="text-davinci-003", max_tokens=num_output))

    service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, prompt_helper=prompt_helper)

    #builds an index over the Google docs
    #index = GPTVectorStoreIndex.from_documents(documents)
    index = GPTVectorStoreIndex.from_documents(
        documents, service_context=service_context
    )

    #persists the index to disk (by default to ./storage) so that it can be used later
    index.storage_context.persist()

def ask_v():
    # rebuild storage context
    storage_context = StorageContext.from_defaults(persist_dir="./storage")

    # load index
    index = load_index_from_storage(storage_context, service_context=service_context)

    retriever = VectorIndexRetriever(index=index, similarity_top_k=1)

    DEFAULT_TEXT_QA_PROMPT_TMPL = (
        "Context information is below. \n"
        "---------------------\n"
        "{context_str}"
        "\n---------------------\n"
        "Given the context information and not prior knowledge, "
        "answer the question: {query_str}\nTreat this as a knowledge base article and don't end it out of nowhere.\n"
    )

    QA_PROMPT = QuestionAnswerPrompt(DEFAULT_TEXT_QA_PROMPT_TMPL)

    DEFAULT_REFINE_PROMPT_TMPL = (
        "The original question is as follows: {query_str}\n"
        "We have provided an existing answer: {existing_answer}\n"
        "We have the opportunity to refine the existing answer "
        "(only if needed) with some more context below.\n"
        "------------\n"
        "{context_msg}\n"
        "------------\n"
        "Given the new context, refine the original answer to better "
        "answer the question. Treat this as a knowledge base article."
        "If the context isn't useful, return the original answer."
    )

    DEFAULT_REFINE_PROMPT = RefinePrompt(DEFAULT_REFINE_PROMPT_TMPL)

    response_synthesizer = ResponseSynthesizer.from_args(
        node_postprocessors=[
            SimilarityPostprocessor(similarity_cutoff=0.78)
        ],
        text_qa_template=QA_PROMPT,
        refine_template=DEFAULT_REFINE_PROMPT 
    )

    # assemble query engine
    query_engine = RetrieverQueryEngine(
        retriever=retriever,
        response_synthesizer=response_synthesizer,
    )

    # query_engine = index.as_query_engine(similarity_top_k=1, retriever_mode="embedding") #return data from 1 node only

    response = query_engine.query("How do I install the company browser?")