Update app.py

app.py

@@ -1,95 +1,108 @@
 import streamlit as st
+from streamlit_webrtc import webrtc_streamer, WebRtcMode
 import numpy as np
-from io import BytesIO
-import tempfile
-import os
-from pydub import AudioSegment
+import av
+import queue
+import threading
+from typing import List, NamedTuple
 import time
 
+# Audio recording class
+class AudioFrame(NamedTuple):
+    data: np.ndarray
+    timestamp: float
+
+# Global variables for recording state
+AUDIO_BUFFER: queue.Queue = queue.Queue()
+RECORDING = False
+AUDIO_FRAMES: List[AudioFrame] = []
+
 def main():
     st.title("Audio Recorder App")
     
-    # Add a description
-    st.write("Click the microphone button below to start recording")
+    # Add description
+    st.write("Click 'Start' to begin recording audio")
     
     # Create a placeholder for status messages
     status_placeholder = st.empty()
     
-    # Add the audio recorder
-    audio_bytes = st.audio_recorder(
-        pause_threshold=2.0,  # Automatically stops recording after 2 seconds of silence
-        sample_rate=44100
-    )
+    # Recording controls
+    col1, col2 = st.columns(2)
+    with col1:
+        start_button = st.button("Start Recording")
+    with col2:
+        stop_button = st.button("Stop Recording")
     
-    if audio_bytes:
-        try:
-            # Show success message
-            status_placeholder.success("Recording captured successfully!")
-            
-            # Create a temporary file to store the audio
-            with tempfile.NamedTemporaryFile(delete=False, suffix='.wav') as tmp_file:
-                tmp_file.write(audio_bytes)
-                tmp_path = tmp_file.name
-            
-            # Display the audio player
-            st.audio(audio_bytes, format='audio/wav')
-            
-            # Add download button
-            st.download_button(
-                label="Download Recording",
-                data=audio_bytes,
-                file_name="recording.wav",
-                mime="audio/wav"
-            )
+    def audio_callback(frame: av.AudioFrame) -> av.AudioFrame:
+        global RECORDING, AUDIO_FRAMES
+        
+        if RECORDING:
+            # Convert audio frame to numpy array
+            audio_data = frame.to_ndarray()
+            timestamp = time.time()
+            AUDIO_FRAMES.append(AudioFrame(audio_data, timestamp))
             
-            # Convert to AudioSegment for analysis
-            audio = AudioSegment.from_wav(tmp_path)
-            
-            # Display audio information
-            st.subheader("Audio Information")
-            col1, col2, col3 = st.columns(3)
+            # Update status
+            status_placeholder.info(f"Recording... Duration: {len(AUDIO_FRAMES) * 0.02:.1f}s")
+        
+        return frame
+    
+    # Initialize WebRTC component
+    ctx = webrtc_streamer(
+        key="audio-recorder",
+        mode=WebRtcMode.SENDONLY,
+        audio_receiver_size=256,
+        rtc_configuration={
+            "iceServers": [{"urls": ["stun:stun.l.google.com:19302"]}]
+        },
+        media_stream_constraints={"video": False, "audio": True},
+        audio_callback=audio_callback,
+    )
+    
+    # Handle recording controls
+    if start_button:
+        RECORDING = True
+        AUDIO_FRAMES.clear()
+        status_placeholder.info("Starting recording...")
+    
+    if stop_button:
+        RECORDING = False
+        if len(AUDIO_FRAMES) > 0:
+            # Combine all audio frames
+            all_audio = np.concatenate([frame.data for frame in AUDIO_FRAMES])
             
-            with col1:
-                st.metric("Duration", f"{len(audio)/1000:.2f} seconds")
-            with col2:
-                st.metric("Channels", audio.channels)
-            with col3:
-                st.metric("Sample Rate", f"{audio.frame_rate} Hz")
+            # Save as numpy array
+            np.save("recorded_audio.npy", all_audio)
             
-            # Clean up temporary file
-            os.unlink(tmp_path)
+            # Show success message
+            status_placeholder.success(f"Recording saved! Duration: {len(AUDIO_FRAMES) * 0.02:.1f} seconds")
             
-            # Add processing options
-            st.subheader("Audio Processing Options")
-            if st.button("Normalize Audio"):
-                normalized_audio = audio.normalize()
-                normalized_bytes = normalized_audio.export(format='wav').read()
-                st.audio(normalized_bytes, format='audio/wav')
+            # Add download button for the numpy array
+            with open("recorded_audio.npy", "rb") as file:
                 st.download_button(
-                    label="Download Normalized Audio",
-                    data=normalized_bytes,
-                    file_name="normalized_recording.wav",
-                    mime="audio/wav"
+                    label="Download Recording",
+                    data=file,
+                    file_name="recorded_audio.npy",
+                    mime="application/octet-stream"
                 )
-            
-        except Exception as e:
-            status_placeholder.error(f"Error processing audio: {str(e)}")
-            st.stop()
+        else:
+            status_placeholder.warning("No audio recorded")
     
     # Add instructions
     with st.expander("How to use"):
         st.markdown("""
-        1. Click the microphone button to start recording
-        2. Speak into your microphone
-        3. Click the button again to stop recording
-        4. Your recording will appear above with playback controls
+        1. Allow microphone access when prompted
+        2. Click 'Start Recording' to begin
+        3. Speak into your microphone
+        4. Click 'Stop Recording' when finished
         5. Use the download button to save your recording
-        6. Optional: Use the Normalize Audio button to improve audio quality
+        
+        Note: You may need to click 'Start' in the WebRTC player above first.
         """)
     
-    # Add footer
+    # Add requirements info
     st.markdown("---")
-    st.caption("Note: Recording quality depends on your microphone and browser settings")
+    st.caption("Required packages: streamlit-webrtc, numpy, av")
 
 if __name__ == "__main__":
     main()