Spaces:

Duplicated from  bhaskartripathi/pdfChatter

bhaskartripathi

/

pdfGPT_Turbo

like 14

Runtime error

I would like to propose refactoring the project.

1. We have multiple different functions doing the same thing (constructing the prompt)
2. GPT-4 model selection does not do anything, the query is answered by davinci, see this block:

 if model == "text-davinci-003" or model == "gpt-4" or model == "gpt-4-32k":
            answer = generate_answer_text_davinci_003(question, openAI_key)
        else:
            answer = generate_answer(question, openAI_key, model)

3. Everything is one file. It would be good to split it up into UI, Prompt Generation and Semantic Search