app.py

import streamlit as st
from streamlit_webrtc import webrtc_streamer, WebRtcMode, ClientSettings
import numpy as np
import io
import wave
import requests
from audio_to_text import audio_to_text
import streamlit.components.v1 as components

# Initialize Streamlit app layout
st.title("Microphone Input in Streamlit")

# Load the custom component
audio_recorder_html = """
<script>
const recordButton = document.getElementById("recordButton");
const stopButton = document.getElementById("stopButton");
const audioChunks = [];

let mediaRecorder;

navigator.mediaDevices.getUserMedia({ audio: true })
.then(function(stream) {
  mediaRecorder = new MediaRecorder(stream);

  recordButton.onclick = function() {
    mediaRecorder.start();
    console.log("Recording started...");
  };

  stopButton.onclick = function() {
    mediaRecorder.stop();
    console.log("Recording stopped...");
  };

  mediaRecorder.ondataavailable = function(e) {
    audioChunks.push(e.data);
  };

  mediaRecorder.onstop = function(e) {
    const audioBlob = new Blob(audioChunks, { type: "audio/wav" });
    const reader = new FileReader();
    reader.readAsDataURL(audioBlob);
    reader.onloadend = function() {
      const base64data = reader.result.split(',')[1];
      Streamlit.setComponentValue(base64data);
    };
  };
});
</script>
<button id="recordButton">Start Recording</button>
<button id="stopButton">Stop Recording</button>
"""

# Embed the JavaScript and HTML in Streamlit
components.html(audio_recorder_html, height=300)

# Retrieve the audio data from the component (if available)
audio_data = st.query_params().get("value")


def audio_callback(frame):
    # Get raw audio data from the frame
    audio_data = frame.to_ndarray().astype(np.int16)

    # Convert audio to text
    transcription = audio_to_text(audio_data)
    
    # Display the transcription
    st.write("Transcription:", transcription)

    API_URL = "https://eaa0-34-74-179-199.ngrok-free.app/generate"
    # Optionally, send the transcription to an API
    headers = {
        "Content-Type": "application/json"
    }
    payload = {
        "prompt": transcription
    }
    response = requests.post(API_URL, json=payload, headers=headers)
    if response.status_code == 200:
        st.write("Assistant:", response.json())
    else:
        st.write("Error:", response.status_code, response.text)
            

webrtc_ctx = webrtc_streamer(
    key="audio-only",
    mode=WebRtcMode.SENDRECV,
    rtc_configuration={"iceServers": [{"urls": ["stun:stun1.l.google.com:19302"]}]},
    media_stream_constraints={
        "audio": True,
        "video": False
    },
    audio_frame_callback=audio_callback
)

# Placeholder for capturing audio
if webrtc_ctx.state.playing:
    st.write("Microphone is active. Speak into the microphone...")
else:
    st.write("Click to start microphone input.")