OpenAI's ChatGPT + Whisper speech recognition + MacOS say

jarvis.py

import speech_recognition as sr
"""Make some requests to OpenAI's chatbot"""

import time
import os

from playwright.sync_api import sync_playwright

PLAY = sync_playwright().start()
BROWSER = PLAY.chromium.launch_persistent_context(
    user_data_dir="/tmp/playwright",
    headless=False,
)
PAGE = BROWSER.new_page()

# Set up the speech recognition
r = sr.Recognizer()

def get_input_box():
    """Get the child textarea of `PromptTextarea__TextareaWrapper`"""
    return PAGE.query_selector("textarea")

def is_logged_in():
    # See if we have a textarea with data-id="root"
    return get_input_box() is not None

def send_message(message):
    # Send the message
    box = get_input_box()
    box.click()
    box.fill(message)
    box.press("Enter")

def get_last_message():
    """Get the latest message"""
    page_elements = PAGE.query_selector_all("div[class*='ConversationItem__Message']")
    last_element = page_elements[-1]
    return last_element.inner_text()

def start_browser():
    PAGE.goto("https://chat.openai.com/")
    if not is_logged_in():
        print("Please log in to OpenAI Chat")
        print("Press enter when you're done")
        input()
    else:
        print("Logged in")

if __name__ == "__main__":
    start_browser()

    previous_response = ""
    with sr.Microphone() as source:
        print("Adjusting for ambient noise...")
        r.adjust_for_ambient_noise(source)

        while True:
            # Listen for the user's input
            print("Listening...")
            audio = r.listen(source)

            # Try to recognize the user's speech
            try:
                text = r.recognize_whisper(audio, language="english")
                print("I think you said: %s" % text)
                send_message(text)
                time.sleep(5) # TODO: there are about ten million ways to be smarter than this
                response = get_last_message()
                if response != previous_response:
                    previous_response = response
                    os.system("say \"%s\"" % response)
            except:
                print("I'm sorry, I didn't catch that. Could you please repeat it?")

Credit goes to @taranjeet and @danielgross for the ChatGPT integration: https://github.com/taranjeet/chatgpt-api

The dependencies are kinda fiddly to get installed on an Apple Silicon mac.

Had to manually install portaudio to get pyaudio to recognize it as seen here:
https://stackoverflow.com/a/69293153

You'll need to install speech_recognition and whisper from git in order for r.recognize_whisper to work:
pip install git+https://github.com/openai/whisper.git
pip install git+https://github.com/Uberi/speech_recognition