atharvaingle/smart-grow

Purpose and Runtime

Smart Grow/Farmassist is a Flutter mobile application for authenticated farm management, IoT telemetry monitoring, and on-device plant disease classification. Android and iOS hosts bootstrap Flutter; web scaffolding is also present. The TensorFlow Lite model is bundled at model.tflite.

Application Shell and Navigation

The UI is organized by feature under ui/, with the home shell and bottom navigation selecting major workflows. Key entry-facing paths are ui/home_page.dart, ui/bottom_navigation_bar/bottom_navigation_bar.dart, and platform launchers android/app/src/main/kotlin/com/example/farmassist/MainActivity.kt and ios/Runner/AppDelegate.swift.

Identity and User Boundary

Authentication uses Firebase Authentication with Bloc/Cubit state transitions for session, login, and signup. User profile records are handled separately through user repositories, forming the boundary between Firebase identity and app-owned user data in Cloud Firestore.

• bloc/authentication/bloc/authentication_bloc.dart

• bloc/login/login_cubit.dart

• data/authentication/repositories/authentication_repository.dart

• data/user/repositories/user_repository.dart

Farm Management and External Content

Planting and harvesting screens write/read user-scoped farm records through repositories, intended for Cloud Firestore. Farm views also compose weather forecasts and science news from external HTTP APIs; these are independent service boundaries rather than local domain state.

• ui/farm/farm_management_page.dart

• data/farm/repositories/storeData.dart

• data/farm/repositories/harvest_storeData.dart

• data/farm/api/openweathermap_weather_api.dart

• data/farm/resources/newsAPI.dart

IoT Telemetry Pipeline

The app reads per-user telemetry from Firebase Realtime Database via the telemetry repository, converts it into domain/chart models, and renders current readings and time-series charts. The upstream producer is an external IoT simulator/device. Alert evaluation and push delivery occur outside this repo: Realtime Database writes trigger Firebase Cloud Functions, which use Firebase Cloud Messaging.

• data/IoT/repositories/telemetry_data_repository.dart

• data/IoT/models/telemetry_data.dart

• ui/IoT/IoT_monitoring_page.dart

• ui/IoT/telemetry_data_chart.dart

Disease Detection Pipeline

Users capture/select plant imagery through the disease-detection UI. Classification code loads and executes the local TFLite model, returning one of the trained disease categories for presentation in diagnostic cards/views. Model inference is on-device; model training was externalized to Google Cloud AutoML Vision.

• ui/diseases/disease_detection_page.dart

• ui/diseases/diagnosis.dart

• data/diseases/classifier.dart

• data/diseases/disease_detection_model.dart

• model.tflite

State and Data Architecture

The repository mixes feature-local presentation state (Bloc/Cubit in authentication and news) with repository/service-based data access for Firebase and HTTP integrations. Domain models isolate UI from serialized backend/API data, notably weather, farm records, news, and telemetry.

• ui/farm/news/bloc/nBloc.dart

• data/farm/services/forecast_service.dart

• data/farm/models/Planting.dart

• data/farm/models/Weather.dart

• data/IoT/models/line_chart_spots.dart