improve the app layout a little

app.py

@@ -16,6 +16,8 @@ from sample_utils.turn import get_ice_servers
 import json
 from typing import List
 
+st.set_page_config(layout="wide")
+
 from vosk import SetLogLevel, Model, KaldiRecognizer
 SetLogLevel(-1) # mutes vosk verbosity
 
@@ -40,15 +42,18 @@ async def main():
 
     system_one_audio_status = st.empty()
 
-    playing = st.checkbox("Playing", value=True)
 
     system_one_audio_status.write("Initializing streaming")
-    system_one_audio_output = st.empty()
 
-    system_one_video_output = st.empty()
+    # system_one_audio_output = st.empty()
+    # system_one_video_output = st.empty()
+    # system_one_audio_history = []
+
+    col1, col2 = st.columns(2)
 
-    system_one_audio_history = []
-    system_one_audio_history_output = st.empty()
+    with col1:
+        listening = st.checkbox("Listen", value=True)
+        system_one_audio_history_output = st.empty()
 
     # Initialize resources if not already done
     system_one_audio_status.write("Initializing streaming")
@@ -58,28 +63,28 @@ async def main():
 
     system_one_audio_status.write("resources referecned")
 
-
-
     system_one_audio_status.write("Initializing webrtc_streamer")
-    webrtc_ctx = webrtc_streamer(
-        key="charles",
-        desired_playing_state=playing,
-        queued_audio_frames_callback=st.session_state.streamlit_av_queue.queued_audio_frames_callback,
-        queued_video_frames_callback=st.session_state.streamlit_av_queue.queued_video_frames_callback,
-        mode=WebRtcMode.SENDRECV,
-        media_stream_constraints={
-            "video": True, 
-            "audio": {
-                "sampleRate": 48000,
-                "sampleSize": 16,
-                "noiseSuppression": True,
-                "echoCancellation": True,
-                "channelCount": 1,
-                }
-            },
-        rtc_configuration={"iceServers": get_ice_servers()},
-        async_processing=True,
-    )
+    with col2:
+        playing = st.checkbox("Playing", value=True)
+        webrtc_ctx = webrtc_streamer(
+            key="charles",
+            desired_playing_state=playing,
+            queued_audio_frames_callback=st.session_state.streamlit_av_queue.queued_audio_frames_callback,
+            queued_video_frames_callback=st.session_state.streamlit_av_queue.queued_video_frames_callback,
+            mode=WebRtcMode.SENDRECV,
+            media_stream_constraints={
+                "video": True, 
+                "audio": {
+                    "sampleRate": 48000,
+                    "sampleSize": 16,
+                    "noiseSuppression": True,
+                    "echoCancellation": True,
+                    "channelCount": 1,
+                    }
+                },
+            rtc_configuration={"iceServers": get_ice_servers()},
+            async_processing=True,
+        )
 
     if not webrtc_ctx.state.playing:
         exit
@@ -92,6 +97,8 @@ async def main():
 
     try:
         while True:
+            if "streamlit_av_queue" in st.session_state:
+                st.session_state.streamlit_av_queue.set_listening(listening)
             if not webrtc_ctx.state.playing:
                 system_one_audio_status.write("Stopped.")
                 await asyncio.sleep(0.1)

streamlit_av_queue.py

@@ -14,10 +14,16 @@ class StreamlitAVQueue:
     def __init__(self, audio_bit_rate=16000):
         self._output_channels = 2
         self._audio_bit_rate = audio_bit_rate
+        self._listening = True
+        self._lock = threading.Lock()
         self.queue_actor = WebRtcAVQueueActor.options(
             name="WebRtcAVQueueActor", 
             get_if_exists=True,
             ).remote() 
+        
+    def set_listening(self, listening: bool):
+        with self._lock:
+            self._listening = listening
             
     async def queued_video_frames_callback(
                 self,
@@ -37,8 +43,10 @@ class StreamlitAVQueue:
                 frames: List[av.AudioFrame],
             ) -> av.AudioFrame:
         try:
+            with self._lock:
+                should_listed = self._listening
             sound_chunk = pydub.AudioSegment.empty()
-            if len(frames) > 0:
+            if len(frames) > 0 and should_listed:
                 for frame in frames:
                     sound = pydub.AudioSegment(
                         data=frame.to_ndarray().tobytes(),