GPT3, python and unity

Unity Engine
,
1

Hi
I am trying to use Speech to Text from within Unity to send the resulting text to a python file which will send to GPT3 and retrieve results.

I have the Speech to Text working and the python code which connects to GPT3 API but I am not sure how to link the two.

Can I embed the python code inside unity or do I need to host externally and somehow trigger from unity? What is the best (least lag) approach and how do I do it?

thanks

2

What value does Python add here?

If you just need to contact some API over the web, just do it directly with UnityWebRequest

If you absolutely must, you can integrate Python in Unity but it is a LOT of extra work. You can also get asset store packages that get you so much of the way there Python-wise.

3

thanks for the tip - if UnityWebRequest can support this then life becomes much easier. This is python code.

import os
import openai

def open_file(filepath):
    with open (filepath, 'r', encoding='utf-8') as infile:
        return infile.read()

def save_file(filepath, content):
    with open (filepath, 'w', encoding='utf-8') as outfile:
        outfile.write(content)

openai.api_key  = open_file('openapikey.txt')
def gpt3 (prompt):
    response = openai.Completion.create(
        model = "text-curie-001",
        prompt = prompt,
        temperature = 1.0,
        max_tokens = 500,
        top_p = 1,
        frequency_penalty = 0,
        presence_penalty = 0
        )
    text = response['choices'][0]['text'].strip()
    return text

prompt = open_file('prompt.txt')
prose = gpt3(prompt)
quesandanswer = prompt + prose
print ('\n\nQuestions:', prompt, prose)
save_file('questions.txt', prose)

prompt2 = open_file(prompt2.txt)

I want to replace the prompts so it is really just passing the text generated by the STT in unity that I need to worry about, and then how to retreive the result

The plugin for unity that i found seemed to suggest that I need to go via python. I did look at options in the assetstore but couldn’t find anything specifically for Speech to Text which is what I need