rahulop11/appumangly

1. Purpose and Runtime

Umangly is a cross-platform Flutter/Dart mental-wellness client combining mood/diary tracking, mindfulness content, AI companions, and camera-based emotion/smile features. The primary application entry point is lib/main.dart; it builds for Android, iOS, web, Linux, macOS, and Windows.

2. Client Navigation and Feature Surfaces

Flutter screens form the feature-oriented presentation layer, including dynamic/home flows, authentication, mood selection, diary, meditation, smile camera, and companion chat. Key user-facing orchestration paths include:

• lib/screens/dynamic_home_screen.dart

• lib/screens/emotion_check_screen.dart

• lib/screens/virtual_friend_chat_screen.dart

Authentication-specific screens are under lib/screens/auth/login_screen.dart and lib/screens/auth/signup_screen.dart.

3. Firebase Identity and User Data Boundary

Firebase is the cloud system of record for authenticated user data, with generated app initialization options and Firestore access controls. App services should isolate Firebase reads/writes from screen code; the rules file defines the authorization boundary.

• lib/firebase_options.dart

• lib/services/diary_service.dart

• firestore.rules

The committed Android configuration indicates Firebase services are provisioned for Android: android/app/google-services.json.

4. Mood, Diary, and Analytics Pipeline

Users select or derive emotional state, persist diary/mood-related records, then consume aggregation and insight views. Typed domain models separate diary, mood, meditation, smile, and companion concepts from UI/service implementations.

• lib/models/mood_models.dart

• lib/models/diary_models.dart

• lib/services/mood_analytics_service.dart

Primary presentation endpoints are lib/screens/mood_selection_screen.dart, lib/screens/diary_screen.dart, and lib/screens/diary_analytics_screen.dart.

5. AI Companion Integration

The virtual-friend layer supports distinct AI providers/personas: Gemini for Meera and Dialogflow-oriented flows for Arav, with an offline fallback path. Screen code should call the service abstraction rather than provider APIs directly.

• lib/services/virtual_friend_service.dart

• lib/services/gemini_ai_service.dart

• lib/services/dialogflow_service.dart

Offline continuity is represented by lib/services/offline_virtual_friend_service.dart; companion data contracts live in lib/models/virtual_friend_models.dart.

6. Emotion and Smile Detection Boundary

Emotion recognition is separated from the Flutter UI through detection services and camera/webview screens. The repository also contains independently runnable Python/OpenCV-oriented implementations, making this a client-to-local/HTTP detector integration boundary rather than a purely Dart model.

• lib/services/emotion_detection_service.dart

• lib/services/offline_emotion_detection_service.dart

• web_emotion_detection/emotion_api.py

Relevant UI routes are lib/screens/smile_camera_screen.dart and lib/screens/emotion_webview_screen.dart. Model assets include assets/models/mock_emotion_model.tflite.

7. Mindfulness Media and Rewards

Meditation, breathing, affirmations, ambient audio, TTS, and smile rewards are local client workflows backed by service abstractions and typed meditation/smile models. Audio/content providers are external-service boundaries where configured.

• lib/services/audio_service.dart

• lib/services/freesound_api_service.dart

• lib/services/smile_rewards_service.dart

Core feature routes include lib/screens/meditation_screen.dart, lib/screens/breathing_exercise_screen.dart, and lib/screens/affirmation_screen.dart.

8. Platform Integration

Flutter owns most behavior, while Android adds WebView navigation and permission handling for embedded emotion-detection experiences:

• android/app/src/main/kotlin/com/example/umangly/MainActivity.kt

• android/app/src/main/kotlin/com/example/umangly/CustomWebViewClient.kt

• android/app/src/main/kotlin/com/example/umangly/WebViewPermissionHandler.kt

Native runners in ios/, web/, linux/, macos/, and windows/ are platform shells rather than separate application architectures.