app.py

import streamlit as st
import os
from audio_recorder_streamlit import audio_recorder
from streamlit_float import *
import base64
from openai import OpenAI

api_key = os.getenv("openapikey")

client = OpenAI(api_key=api_key)

def get_answer(messages):
    system_message = [{"role": "system", "content": "You are an helpful AI chatbot, that answers questions asked by User."}]
    messages = system_message + messages
    response = client.chat.completions.create(
        model="gpt-3.5-turbo-1106",
        messages=messages
    )
    return response.choices[0].message.content

def speech_to_text(audio_data):
    with open(audio_data, "rb") as audio_file:
        transcript = client.audio.transcriptions.create(
            model="whisper-1",
            response_format="text",
            file=audio_file
        )
    return transcript

def text_to_speech(input_text):
    response = client.audio.speech.create(
        model="tts-1",
        voice="nova",
        input=input_text
    )
    webm_file_path = "temp_audio_play.mp3"
    with open(webm_file_path, "wb") as f:
        response.stream_to_file(webm_file_path)
    return webm_file_path

def autoplay_audio(file_path: str):
    with open(file_path, "rb") as f:
        data = f.read()
    b64 = base64.b64encode(data).decode("utf-8")
    md = f"""
    <audio autoplay>
    <source src="data:audio/mp3;base64,{b64}" type="audio/mp3">
    </audio>
    """
    st.markdown(md, unsafe_allow_html=True)


# Initialize floating features for the interface
float_init()

# Initialize session state for managing chat messages
def initialize_session_state():
    if "messages" not in st.session_state:
        st.session_state.messages = [{"role": "assistant", "content": "Hi! How may I assist you today?"}]

initialize_session_state()

st.title("OpenAI Conversational Chatbot 🤖")

# Create a container for the microphone and audio recording
footer_container = st.container()
with footer_container:
    audio_bytes = audio_recorder()

for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.write(message["content"])

if audio_bytes:
    # Write the audio bytes to a file
    with st.spinner("Transcribing..."):
        webm_file_path = "temp_audio.mp3"
        with open(webm_file_path, "wb") as f:
            f.write(audio_bytes)

        transcript = speech_to_text(webm_file_path)
        if transcript:
            st.session_state.messages.append({"role": "user", "content": transcript})
            with st.chat_message("user"):
                st.write(transcript)
            os.remove(webm_file_path)

if st.session_state.messages[-1]["role"] != "assistant":
    with st.chat_message("assistant"):
        with st.spinner("Thinking🤔..."):
            final_response = get_answer(st.session_state.messages)
        with st.spinner("Generating audio response..."):    
            audio_file = text_to_speech(final_response)
            autoplay_audio(audio_file)
        st.write(final_response)
        st.session_state.messages.append({"role": "assistant", "content": final_response})
        os.remove(audio_file)

# Float the footer container and provide CSS to target it with
footer_container.float("bottom: 0rem;")