From 4f6c10c857dc6e8668987154d779072286fac37f Mon Sep 17 00:00:00 2001
From: cogwheel0 <172976095+cogwheel0@users.noreply.github.com>
Date: Wed, 8 Oct 2025 13:35:24 +0530
Subject: [PATCH] feat: enhance text-to-speech and voice call services

- Added volume, speech rate, and pitch settings to the TextToSpeechService for improved audio control.
- Reset the accumulated response in VoiceCallService before sending messages to ensure accurate response handling.
- Enhanced the handling of socket events in VoiceCallService to manage streaming content and completion more effectively.
- Improved logging for better debugging and tracking of TTS and voice call states.
---
 .../chat/services/text_to_speech_service.dart | 11 +++
 .../chat/services/voice_call_service.dart     | 69 ++++++++++++++-----
 2 files changed, 61 insertions(+), 19 deletions(-)

diff --git a/lib/features/chat/services/text_to_speech_service.dart b/lib/features/chat/services/text_to_speech_service.dart
index cef1c11..2df1701 100644
--- a/lib/features/chat/services/text_to_speech_service.dart
+++ b/lib/features/chat/services/text_to_speech_service.dart
@@ -54,6 +54,16 @@ class TextToSpeechService {
 
     try {
       await _tts.awaitSpeakCompletion(false);
+
+      // Set volume to maximum
+      await _tts.setVolume(1.0);
+
+      // Set speech rate (1.0 is normal)
+      await _tts.setSpeechRate(0.5);
+
+      // Set pitch (1.0 is normal)
+      await _tts.setPitch(1.0);
+
       if (!kIsWeb && Platform.isIOS) {
         await _tts.setSharedInstance(true);
         await _tts.setIosAudioCategory(IosTextToSpeechAudioCategory.playback, [
@@ -63,6 +73,7 @@ class TextToSpeechService {
           IosTextToSpeechAudioCategoryOptions.allowBluetoothA2DP,
         ]);
       }
+
       await _configurePreferredVoice();
       _available = true;
     } catch (e) {
diff --git a/lib/features/chat/services/voice_call_service.dart b/lib/features/chat/services/voice_call_service.dart
index f1a3b8d..35019ee 100644
--- a/lib/features/chat/services/voice_call_service.dart
+++ b/lib/features/chat/services/voice_call_service.dart
@@ -234,6 +234,7 @@ class VoiceCallService {
 
     try {
       _updateState(VoiceCallState.processing);
+      _accumulatedResponse = ''; // Reset response accumulator
 
       // Send message using the existing chat infrastructure
       sendMessageFromService(_ref, text, null);
@@ -243,34 +244,50 @@ class VoiceCallService {
     }
   }
 
+  String _accumulatedResponse = '';
+
   void _handleSocketEvent(
     Map<String, dynamic> event,
     void Function(dynamic response)? ack,
   ) {
     if (_isDisposed) return;
 
-    final type = event['type']?.toString();
-    final data = event['data'];
+    final outerData = event['data'];
 
-    if (data is Map<String, dynamic>) {
-      // Handle streaming response chunks
-      if (type == 'message' || type == 'delta') {
-        final content = data['content']?.toString() ?? '';
-        if (content.isNotEmpty) {
-          _responseController.add(content);
+    if (outerData is Map<String, dynamic>) {
+      final eventType = outerData['type']?.toString();
+      final innerData = outerData['data'];
+
+      if (eventType == 'chat:completion' && innerData is Map<String, dynamic>) {
+        // Handle streaming content chunks
+        if (innerData.containsKey('content')) {
+          final content = innerData['content']?.toString() ?? '';
+          if (content.isNotEmpty) {
+            _accumulatedResponse = content;
+            _responseController.add(content);
+          }
         }
-      }
 
-      // Handle completion
-      if (data['done'] == true || type == 'completion') {
-        final fullResponse = data['content']?.toString() ??
-            data['message']?.toString() ??
-            '';
-        if (fullResponse.isNotEmpty) {
-          _speakResponse(fullResponse);
-        } else {
-          // No response, restart listening
-          _startListening();
+        // Check for completion using choices[0].finish_reason
+        if (innerData.containsKey('choices')) {
+          final choices = innerData['choices'] as List?;
+          if (choices != null && choices.isNotEmpty) {
+            final firstChoice = choices[0] as Map<String, dynamic>?;
+            final finishReason = firstChoice?['finish_reason'];
+
+            if (finishReason == 'stop') {
+              // ignore: avoid_print
+              print('[VoiceCall] Response completed! Text: $_accumulatedResponse');
+
+              if (_accumulatedResponse.isNotEmpty) {
+                _speakResponse(_accumulatedResponse);
+                _accumulatedResponse = '';
+              } else {
+                // No response, restart listening
+                _startListening();
+              }
+            }
+          }
         }
       }
     }
@@ -280,10 +297,24 @@ class VoiceCallService {
     if (_isDisposed) return;
 
     try {
+      // ignore: avoid_print
+      print('[VoiceCall] _speakResponse called with: $response');
+
+      // Stop listening before speaking
+      await _voiceInput.stopListening();
+      await _transcriptSubscription?.cancel();
+      await _intensitySubscription?.cancel();
+
       _updateState(VoiceCallState.speaking);
+      // ignore: avoid_print
+      print('[VoiceCall] State updated to speaking, calling TTS...');
       await _tts.speak(response);
+      // ignore: avoid_print
+      print('[VoiceCall] TTS.speak() returned');
       // After speaking completes, _handleTtsComplete will restart listening
     } catch (e) {
+      // ignore: avoid_print
+      print('[VoiceCall] Error in _speakResponse: $e');
       _updateState(VoiceCallState.error);
       // Restart listening even if TTS fails
       await _startListening();