simplest possible assistant ever (getting this right is incredibly hard)

assistant.py

from openai import OpenAI
import pyttsx3
import sounddevice as sd
import numpy as np
import io
from scipy.io.wavfile import write
from threading import Thread
import os
import sys
import queue

client = OpenAI(api_key=os.environ['OPENAI_API_KEY'])

def record_response(filename='response.wav'):
    q = queue.Queue()

    def callback(indata, frames, time, status):
        if status:
            print(status, file=sys.stderr)
        q.put(indata.copy())

    print("Recording... Press Enter to stop.")
    try:
        # Default values for sample rate and channels
        sample_rate = 44100
        channels = 1

        with sd.InputStream(samplerate=sample_rate, channels=channels, callback=callback):
            input()  # Wait for Enter key press
            audio_data = []
            while not q.empty():
                audio_data.append(q.get())
            audio_data = np.concatenate(audio_data, axis=0)

        # Write to file
        write(filename, sample_rate, audio_data)
    except Exception as e:
        print("An error occurred:", str(e))

    return filename

def transcribe_audio(filename):
    with open(filename, 'rb') as audio_file:
        transcript = client.audio.transcriptions.create(
            model="whisper-1", 
            file=audio_file,
            response_format="text"
        )
    print(transcript)
    return transcript['data']['text']

def generate_response(text):
    response = client.chat.completions.create(model="gpt-4",
    messages=[{"role": "system", "content": "You are a helpful assistant."},
              {"role": "user", "content": text}])
    return response.choices[0].message['content']

def text_to_speech(text):
    engine = pyttsx3.init()
    engine.say(text)
    engine.runAndWait()

if __name__ == "__main__":
    print("Assistant: How are you?")
    record_response()
    user_input = transcribe_audio('response.wav')
    assistant_response = generate_response(user_input)
    print(f"Assistant: {assistant_response}")
    text_to_speech(assistant_response)