Voice Assistant Pipeline

from openai import OpenAI
import sounddevice as sd
import soundfile as sf
import numpy as np
import time

client = OpenAI()

def record_audio(duration=5, sample_rate=16000):
    """Record audio from microphone."""
    print("Listening...")
    audio = sd.rec(
        int(duration * sample_rate),
        samplerate=sample_rate,
        channels=1,
        dtype='float32'
    )
    sd.wait()  # Wait until recording is finished
    return audio, sample_rate

def transcribe_audio(audio, sample_rate):
    """Convert audio to text using Whisper."""
    # Save to temporary file (Whisper API requires file)
    sf.write("temp_input.wav", audio, sample_rate)

    with open("temp_input.wav", "rb") as f:
        transcript = client.audio.transcriptions.create(
            model="whisper-1",
            file=f
        )
    return transcript.text

def get_llm_response(user_text, conversation_history):
    """Get response from LLM."""
    conversation_history.append({
        "role": "user",
        "content": user_text
    })

    response = client.chat.completions.create(
        model="gpt-4",
        messages=conversation_history,
        max_tokens=150  # Keep responses concise for voice
    )

    assistant_message = response.choices[0].message.content
    conversation_history.append({
        "role": "assistant",
        "content": assistant_message
    })

    return assistant_message

def text_to_speech(text):
    """Convert text to speech using OpenAI TTS."""
    response = client.audio.speech.create(
        model="tts-1",
        voice="alloy",  # Options: alloy, echo, fable, onyx, nova, shimmer
        input=text
    )

    # Save and play the audio
    response.stream_to_file("temp_response.mp3")
    return "temp_response.mp3"

def play_audio(file_path):
    """Play audio file."""
    import pygame
    pygame.mixer.init()
    pygame.mixer.music.load(file_path)
    pygame.mixer.music.play()
    while pygame.mixer.music.get_busy():
        time.sleep(0.1)

def voice_assistant():
    """Main voice assistant loop."""
    print("Voice Assistant Ready")
    print("=" * 40)

    conversation_history = [{
        "role": "system",
        "content": "You are a helpful voice assistant. Keep responses concise and conversational, under 2-3 sentences."
    }]

    while True:
        try:
            # 1. Record user speech
            audio, sr = record_audio(duration=5)

            # 2. Transcribe with Whisper
            start = time.time()
            user_text = transcribe_audio(audio, sr)
            transcribe_time = time.time() - start
            print(f"You said: {user_text}")

            if not user_text.strip():
                print("(No speech detected)")
                continue

            # Check for exit commands
            if any(word in user_text.lower() for word in ["goodbye", "exit", "quit"]):
                print("Goodbye!")
                break

            # 3. Get LLM response
            start = time.time()
            response_text = get_llm_response(user_text, conversation_history)
            llm_time = time.time() - start
            print(f"Assistant: {response_text}")

            # 4. Text-to-speech
            start = time.time()
            audio_file = text_to_speech(response_text)
            tts_time = time.time() - start

            # 5. Play response
            play_audio(audio_file)

            # Print timing breakdown
            total_time = transcribe_time + llm_time + tts_time
            print(f"[Whisper: {transcribe_time:.2f}s | LLM: {llm_time:.2f}s | TTS: {tts_time:.2f}s | Total: {total_time:.2f}s]")
            print("-" * 40)

        except KeyboardInterrupt:
            print("\nExiting...")
            break

if __name__ == "__main__":
    voice_assistant()