import gradio as gr
import openai
from decouple import config
from gtts import gTTS
import io
import pydub
import config

openai.api_key = config.API_KEYS['openai']

# The Models Job or role
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
]

# Main method goes here
def decipher(audio):
    global messages

    # Using openAI's speech to text model
    transcript = openai.Audio.transcribe("whisper-1", audio)

    messages.append({"role": "user", "content": transcript["text"]})

    response = openai.Completion.create(
        model="text-davinci-002",
        prompt="Conversation:\n" + "\n".join([f"{m['role']}: {m['content']}" for m in messages]),
        temperature=0.7,
        max_tokens=1024,
    )

    system_message = response.choices[0].text
    messages.append({"role": "assistant", "content": system_message})

    # Convert the text to audio using gTTS
    tts = gTTS(text=system_message, lang='en')
    audio_data = io.BytesIO()
    tts.write_to_fp(audio_data)

    # Convert the audio to a playable format using pydub
    audio_data.seek(0)
    audio = pydub.AudioSegment.from_file(io.BytesIO(audio_data.getvalue()))

    # Play the audio using VLC
    player = pydub.playback.play
    player(audio)

    chat_transcript = ""
    for message in messages:
        if message['role'] != 'system':
            chat_transcript += message['role'] + ": " + message['content'] + "\n\n"

    return chat_transcript

# Using Gradio's audio Interface 
interface = gr.Interface(fn=decipher, inputs=gr.inputs.Audio(
    type="filepath", label="Record a voice message"), 
    outputs="text",
    title="AI Assistant",
    description="An AI assistant that can transcribe and respond to voice messages",
    theme="compact")

if __name__ == "__main__":
    interface.launch()