Update app.py

app.py

@@ -1,53 +1,45 @@
 import streamlit as st
 from streamlit_webrtc import webrtc_streamer, WebRtcMode
-import requests
 import av
-import numpy as np
 import wave
+import requests
 import io
 
 st.title("Sai Vahini AI Voice Assistant 🕉️")
 
-# Replace with your actual Render API endpoint
-api_url = "https://saivahini.onrender.com/process_audio"
-
-frames = []
-
-# Function to save audio as WAV file in memory
-def save_audio(frames, sample_rate=16000):
-    audio_bytes = io.BytesIO()
-    wf = wave.open(audio_bytes, 'wb')
-    wf.setnchannels(1)
-    wf.setsampwidth(2)
-    wf.setframerate(sample_rate)
-    wf.writeframes(b''.join(frames))
-    wf.close()
-    audio_bytes.seek(0)
-    return audio_bytes
+api_url = "https://saivahini.onrender.com/process_audio"  # Replace if needed
 
-# Callback to handle audio frames
 def audio_frame_callback(frame):
     audio = frame.to_ndarray(format="s16le")
-    frames.append(audio.tobytes())
+    audio_bytes = audio.tobytes()
+    st.session_state.frames.append(audio_bytes)
     return av.AudioFrame.from_ndarray(audio, format="s16", layout="mono")
 
-# Button to start recording audio
-if st.button("🎙️ Record your voice"):
-    frames.clear()
-    webrtc_streamer(
-        key="audio-recorder",
-        mode=WebRtcMode.SENDRECV,
-        audio_frame_callback=audio_frame_callback,
-        media_stream_constraints={"audio": True, "video": False}
-    )
+if "frames" not in st.session_state:
+    st.session_state.frames = []
 
-# Button to stop recording and process audio
-if st.button("✅ Stop & Process Audio"):
-    if frames:
+# WebRTC streamer for automatic audio capture
+webrtc_streamer(
+    key="audio-recorder",
+    mode=WebRtcMode.SENDRECV,
+    audio_frame_callback=audio_frame_callback,
+    media_stream_constraints={"audio": True, "video": False},
+)
+
+if st.button("✅ Process Recorded Audio"):
+    if st.session_state.frames:
         with st.spinner("🔄 Processing your voice..."):
-            audio_file = save_audio(frames)
+            audio_bytes = io.BytesIO()
+            with wave.open(audio_bytes, "wb") as wf:
+                wf.setnchannels(1)
+                wf.setsampwidth(2)
+                wf.setframerate(16000)
+                wf.writeframes(b''.join(st.session_state.frames))
+
+            audio_bytes.seek(0)
 
-            response = requests.post(api_url, files={"file": ("audio.wav", audio_file, "audio/wav")})
+            # Send to your API
+            response = requests.post(api_url, files={"file": ("audio.wav", audio_bytes, "audio/wav")})
 
             if response.status_code == 200:
                 result = response.json()
@@ -55,12 +47,15 @@ if st.button("✅ Stop & Process Audio"):
                 st.write("**Transcription:**", result["transcription"])
                 st.write("**Answer:**", result["response"])
 
-                audio_response_url = result.get("audio")
+                audio_response = result["audio"]
+
+                # Fetch and play audio response
+                audio_content = requests.get(audio_response).content
+                st.audio(audio_content, format="audio/wav")
 
-                # If the API returns a URL, fetch and play the audio directly
-                audio_content = requests.get(audio_response_url).content
-                st.audio(audio_content, format='audio/wav')
+                # Clear session state for new recording
+                st.session_state.frames = []
             else:
                 st.error(f"❌ API Error: {response.status_code}")
     else:
-        st.error("⚠️ No audio captured. Please record again.")
+        st.error("⚠️ No audio captured. Please record first!")