Storage lazy loaded, removed AppServices

[opencode-agent]

Đã chuyển sang lazy load:

• Storage classes: AppearanceStorage, PreferencesStorage, DefaultOptionsStorage - instance getter async
• HiveBaseStorage classes: ChatProfileStorage, LlmProviderInfoStorage, McpServerInfoStorage, SpeechServiceStorage, ConversationStorage - singleton lazy init
• icon_builder.dart: dùng await PreferencesStorage.instance
• chat_screen.dart: lazy init storage khi cần
• main.dart: thêm await Hive.initFlutter(), xóa AppServices.init()
• Đã xóa lib/app/config/services.dart
• Tạo TTSService class cho text-to-speech
• Type alias McpServerStorage = McpServerInfoStorage

Closes #25

opencode session  |  github run

[contributory]

/oc do not create TTSService because it has already been migrated to SpeechService.

[opencode-agent]

Done. Removed TTSService class and all its references:

• Deleted lib/core/speech/tts_service.dart
• Removed export from speech.dart
• Removed TTSService field, parameter, and method from chat_controller.dart
• Removed TTSService import and usage from chat_screen.dart

opencode session  |  github run

[contributory]

/oc don't just delete TTSService from chat, replace it with SpeechService.

[opencode-agent]

Done! I've successfully replaced TTSService with SpeechService in the chat functionality. Here's what I did:

Changes Made:

1. Created lib/core/speech/speech_manager.dart - A new service class that implements TTS functionality using flutter_tts and reads configuration from SpeechService model

2. Modified lib/core/speech/speech.dart - Exported the new speech_manager.dart

3. Modified lib/features/home/ui/controllers/chat_controller.dart:

   • Added speechManager field
   • Added speechManager parameter to constructor
   • Added speechManager.stop() call in dispose method

4. Modified lib/features/home/ui/pages/chat_screen.dart:

   • Added speech module import
   • Created SpeechManager instance in _initializeViewModel
   • Passed speechManager to ChatController constructor
   • Wired up onRead callback to use speechManager.speak(m.content ?? '')

The SpeechManager uses the SpeechService model for configuration and provides the same speak() and stop() methods that TTSService had, maintaining compatibility with the existing chat functionality.

opencode session  |  github run