How to pass a Excel sheet as part of a prompt to OpenAI LLM?

Question

I have a set of texts ("descriptions") for various news items in a csv/xlsx file which I want to pass to Azure OpenAI LLM to categorize. Is there a way to pass this file in the prompt? For now, I send 1-2 descriptions in the prompt as text, not as a list of arguments.

My code is this:

 response = openai.Completion.create(

   engine="text-davinci-003",

   prompt="<command here> \n\n <description 1:...\n\n description 2:...>",

   temperature =0.1)

display(response)

Answer 1

First Read the Excel sheet (you can use a library in Python to read the contents of the Excel sheet into a DataFrame. Assuming you have the sheet data in a variable called df, you can extract the descriptions column using descriptions = df['column_name'], where 'column_name' is the name of the column containing the descriptions.

Convert the descriptions into a format suitable for including in the prompt. You can create a string that concatenates all the descriptions, separating them with appropriate delimiters (e.g., newlines or commas) to distinguish between different descriptions.

descriptions_text = '\n\n'.join(descriptions)

1. Construct the prompt

prompt = f"<command here> \n\n{descriptions_text}"

1. Pass the prompt to OpenAI API

response = openai.Completion.create(
    engine="text-davinci-003",
    prompt=prompt,
    temperature=0.1
)