Update app.py

app.py

@@ -107,11 +107,12 @@ FILE_EMOJIS = {
 
 # 5. Voice Recognition Component
 def create_voice_component():
-    """Create auto-starting voice recognition component"""
+    """Create auto-searching voice recognition component"""
     return components.html(
         """
         <div style="padding: 20px; border-radius: 10px; background: #f0f2f6;">
-            <div id="status">Initializing voice recognition...</div>
+            <div id="status" style="margin-bottom: 10px; color: #666;">Starting voice recognition...</div>
+            <div id="interim" style="color: #666; min-height: 24px;"></div>
             <div id="output" style="margin-top: 10px; padding: 10px; min-height: 100px; 
                                   background: white; border-radius: 5px; white-space: pre-wrap;"></div>
             <script>
@@ -121,8 +122,11 @@ def create_voice_component():
                     recognition.interimResults = true;
                     
                     const status = document.getElementById('status');
+                    const interim = document.getElementById('interim');
                     const output = document.getElementById('output');
                     let fullTranscript = '';
+                    let lastPauseTime = Date.now();
+                    let pauseThreshold = 1500; // Time in ms to wait before triggering search
                     
                     // Auto-start on load
                     window.addEventListener('load', () => {
@@ -144,7 +148,8 @@ def create_voice_component():
                         for (let i = event.resultIndex; i < event.results.length; i++) {
                             const transcript = event.results[i][0].transcript;
                             if (event.results[i].isFinal) {
-                                finalTranscript += transcript + '\\n';
+                                finalTranscript += transcript + ' ';
+                                lastPauseTime = Date.now();
                             } else {
                                 interimTranscript += transcript;
                             }
@@ -152,17 +157,39 @@ def create_voice_component():
                         
                         if (finalTranscript) {
                             fullTranscript += finalTranscript;
+                            interim.textContent = '';
+                            output.textContent = fullTranscript;
+                            
+                            // Send to Streamlit for processing
                             window.parent.postMessage({
                                 type: 'streamlit:setComponentValue',
                                 value: fullTranscript,
                                 dataType: 'json',
                             }, '*');
+                        } else if (interimTranscript) {
+                            interim.textContent = '... ' + interimTranscript;
                         }
                         
-                        output.textContent = fullTranscript + (interimTranscript ? '... ' + interimTranscript : '');
                         output.scrollTop = output.scrollHeight;
                     };
                     
+                    // Check for pauses and trigger search
+                    setInterval(() => {
+                        if (fullTranscript && Date.now() - lastPauseTime > pauseThreshold) {
+                            if (output.dataset.lastProcessed !== fullTranscript) {
+                                output.dataset.lastProcessed = fullTranscript;
+                                window.parent.postMessage({
+                                    type: 'streamlit:setComponentValue',
+                                    value: {
+                                        text: fullTranscript,
+                                        trigger: 'pause'
+                                    },
+                                    dataType: 'json',
+                                }, '*');
+                            }
+                        }
+                    }, 500);
+                    
                     recognition.onend = () => {
                         try {
                             recognition.start();
@@ -186,6 +213,104 @@ def create_voice_component():
         height=200
     )
 
+# Available English voices
+ENGLISH_VOICES = [
+    "en-US-AriaNeural",      # Female, conversational
+    "en-US-JennyNeural",     # Female, customer service
+    "en-US-GuyNeural",       # Male, newscast
+    "en-US-RogerNeural",     # Male, calm
+    "en-GB-SoniaNeural",     # British female
+    "en-GB-RyanNeural",      # British male
+    "en-AU-NatashaNeural",   # Australian female
+    "en-AU-WilliamNeural",   # Australian male
+    "en-CA-ClaraNeural",     # Canadian female
+    "en-CA-LiamNeural",      # Canadian male
+    "en-IE-EmilyNeural",     # Irish female
+    "en-IE-ConnorNeural",    # Irish male
+    "en-IN-NeerjaNeural",    # Indian female
+    "en-IN-PrabhatNeural",   # Indian male
+]
+
+def render_search_interface():
+    """Render main search interface with auto-search voice component"""
+    st.header("🔍 Voice Search")
+    
+    # Voice settings
+    col1, col2 = st.columns([2, 1])
+    with col1:
+        selected_voice = st.selectbox(
+            "Select Voice",
+            ENGLISH_VOICES,
+            index=0,
+            help="Choose the voice for audio responses"
+        )
+    with col2:
+        auto_search = st.checkbox("Auto-Search on Pause", value=True)
+    
+    # Voice component
+    voice_result = create_voice_component()
+    
+    # Handle voice input
+    if voice_result and isinstance(voice_result, (str, dict)):
+        # Extract text and trigger info
+        if isinstance(voice_result, dict):
+            current_text = voice_result.get('text', '')
+            trigger = voice_result.get('trigger')
+        else:
+            current_text = voice_result
+            trigger = None
+            
+        # Process on pause trigger if enabled
+        if auto_search and trigger == 'pause' and current_text:
+            if current_text != st.session_state.get('last_processed_text', ''):
+                st.session_state.last_processed_text = current_text
+                
+                # Show the detected text
+                st.info(f"🎤 Detected: {current_text}")
+                
+                # Perform search
+                try:
+                    with st.spinner("Searching and generating audio response..."):
+                        response, audio_file = asyncio.run(
+                            process_voice_search(
+                                current_text,
+                                voice=selected_voice
+                            )
+                        )
+                        if response:
+                            st.markdown(response)
+                            if audio_file:
+                                render_audio_result(audio_file, "Search Results")
+                                
+                            # Save to history
+                            st.session_state.transcript_history.append({
+                                'timestamp': datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
+                                'query': current_text,
+                                'response': response,
+                                'audio': audio_file
+                            })
+                except Exception as e:
+                    st.error(f"Error processing search: {str(e)}")
+
+    # Manual search option
+    with st.expander("📝 Manual Search", expanded=False):
+        query = st.text_input("Search Query:", value=st.session_state.get('last_processed_text', ''))
+        if st.button("🔍 Search"):
+            try:
+                with st.spinner("Searching and generating audio..."):
+                    response, audio_file = asyncio.run(
+                        process_voice_search(
+                            query,
+                            voice=selected_voice
+                        )
+                    )
+                    if response:
+                        st.markdown(response)
+                        if audio_file:
+                            render_audio_result(audio_file)
+            except Exception as e:
+                st.error(f"Error processing search: {str(e)}")
+
 # 6. Audio Processing Functions
 def get_autoplay_audio_html(audio_path, width="100%"):
     """Create HTML for autoplaying audio with controls"""