Machine Learning integration for Flutter using TensorFlow Lite and Firebase ML
日本語の概要は準備中です。原文の説明を表示しています。
Google Maps and Mapbox integration for Flutter applications
インストールする前に、エージェントに与えられる指示の中身を確認できます。
This skill enables map integration in Flutter applications using Google Maps and Mapbox. Includes features like markers, polylines, clustering, custom styles, and location tracking.
dependencies:
google_maps_flutter: ^2.5.0
mapbox_gl: ^0.16.0
geolocator: ^10.1.0
geocoding: ^2.1.1
flutter_polyline_points: ^2.0.0
maps_toolkit: ^3.0.0
Android:
<!-- AndroidManifest.xml -->
<manifest>
<application>
<meta-data
android:name="com.google.android.geo.API_KEY"
android:value="YOUR_API_KEY" />
</application>
</manifest>
iOS:
<!-- AppDelegate.swift -->
import GoogleMaps
GMSServices.provideAPIKey("YOUR_API_KEY")
import 'package:google_maps_flutter/google_maps_flutter.dart';
import 'package:flutter/material.dart';
class MapPage extends StatefulWidget {
const MapPage({super.key});
@override
State<MapPage> createState() => _MapPageState();
}
class _MapPageState extends State<MapPage> {
GoogleMapController? _mapController;
static const CameraPosition _initialPosition = CameraPosition(
target: LatLng(37.7749, -122.4194), // San Francisco
zoom: 12,
);
@override
Widget build(BuildContext context) {
return Scaffold(
body: GoogleMap(
mapType: MapType.normal,
initialCameraPosition: _initialPosition,
onMapCreated: (controller) {
_mapController = controller;
},
myLocationEnabled: true,
myLocationButtonEnabled: true,
),
);
}
@override
void dispose() {
_mapController?.dispose();
super.dispose();
}
}
class MapService {
GoogleMapController? _controller;
void setController(GoogleMapController controller) {
_controller = controller;
}
Future<void> animateToLocation(LatLng location, {double zoom = 15}) async {
await _controller?.animateCamera(
CameraUpdate.newCameraPosition(
CameraPosition(target: location, zoom: zoom),
),
);
}
Future<void> addMarker(Marker marker) async {
// Implementation depends on your state management
}
Future<void> drawPolyline(List<LatLng> points) async {
// Implementation
}
}
class MarkerManager {
final Set<Marker> _markers = {};
Set<Marker> get markers => _markers;
void addMarker({
required String id,
required LatLng position,
String? title,
String? snippet,
BitmapDescriptor? icon,
}) {
final marker = Marker(
markerId: MarkerId(id),
position: position,
infoWindow: InfoWindow(
title: title,
snippet: snippet,
),
icon: icon ?? BitmapDescriptor.defaultMarker,
);
_markers.add(marker);
}
void removeMarker(String id) {
_markers.removeWhere((marker) => marker.markerId.value == id);
}
void clearMarkers() {
_markers.clear();
}
}
import 'package:geolocator/geolocator.dart';
class LocationService {
Future<Position> getCurrentLocation() async {
bool serviceEnabled = await Geolocator.isLocationServiceEnabled();
if (!serviceEnabled) {
throw Exception('Location services are disabled');
}
LocationPermission permission = await Geolocator.checkPermission();
if (permission == LocationPermission.denied) {
permission = await Geolocator.requestPermission();
if (permission == LocationPermission.denied) {
throw Exception('Location permissions are denied');
}
}
return await Geolocator.getCurrentPosition();
}
Stream<Position> getLocationStream() {
return Geolocator.getPositionStream(
locationSettings: const LocationSettings(
accuracy: LocationAccuracy.high,
distanceFilter: 10, // meters
),
);
}
}
// Usage in a map
class TrackingMap extends ConsumerWidget {
const TrackingMap({super.key});
@override
Widget build(BuildContext context, WidgetRef ref) {
final locationService = LocationService();
return StreamBuilder<Position>(
stream: locationService.getLocationStream(),
builder: (context, snapshot) {
if (!snapshot.hasData) {
return const Center(child: CircularProgressIndicator());
}
final position = snapshot.data!;
return GoogleMap(
initialCameraPosition: CameraPosition(
target: LatLng(position.latitude, position.longitude),
zoom: 15,
),
markers: {
Marker(
markerId: const MarkerId('currentLocation'),
position: LatLng(position.latitude, position.longitude),
icon: BitmapDescriptor.defaultMarkerWithHue(
BitmapDescriptor.hueBlue,
),
),
},
);
},
);
}
}
import 'package:flutter_polyline_points/flutter_polyline_points.dart';
class RoutingService {
final String apiKey;
final PolylinePoints _polylinePoints = PolylinePoints();
RoutingService({required this.apiKey});
Future<List<LatLng>> getDirections(
LatLng origin,
LatLng destination,
) async {
final result = await _polylinePoints.getRouteBetweenCoordinates(
googleApiKey: apiKey,
request: PolylineRequest(
origin: PointLatLng(origin.latitude, origin.longitude),
destination: PointLatLng(destination.latitude, destination.longitude),
mode: TravelMode.driving,
),
);
if (result.points.isNotEmpty) {
return result.points
.map((point) => LatLng(point.latitude, point.longitude))
.toList();
}
return [];
}
}
class RescueMapPage extends ConsumerStatefulWidget {
const RescueMapPage({super.key});
@override
ConsumerState<RescueMapPage> createState() => _RescueMapPageState();
}
class _RescueMapPageState extends ConsumerState<RescueMapPage> {
GoogleMapController? _controller;
final Set<Marker> _markers = {};
List<LatLng> _routePoints = [];
@override
Widget build(BuildContext context) {
return Scaffold(
appBar: AppBar(title: const Text('Rescue Map')),
body: Stack(
children: [
GoogleMap(
mapType: MapType.normal,
initialCameraPosition: const CameraPosition(
target: LatLng(37.7749, -122.4194),
zoom: 12,
),
markers: _markers,
polylines: {
Polyline(
polylineId: const PolylineId('route'),
points: _routePoints,
color: Colors.blue,
width: 5,
),
},
onMapCreated: (controller) {
_controller = controller;
_loadRescueLocations();
},
onTap: _handleMapTap,
),
Positioned(
bottom: 16,
right: 16,
child: FloatingActionButton(
onPressed: _centerOnCurrentLocation,
child: const Icon(Icons.my_location),
),
),
],
),
);
}
void _loadRescueLocations() {
// Add rescue markers
setState(() {
_markers.addAll([
Marker(
markerId: const MarkerId('rescue1'),
position: const LatLng(37.7849, -122.4094),
infoWindow: const InfoWindow(
title: 'Dog Rescue',
snippet: 'Urgent - Needs transport',
),
icon: BitmapDescriptor.defaultMarkerWithHue(
BitmapDescriptor.hueRed,
),
),
]);
});
}
void _handleMapTap(LatLng position) {
// Add marker at tap location
setState(() {
_markers.add(
Marker(
markerId: MarkerId('tap_${position.latitude}_${position.longitude}'),
position: position,
),
);
});
}
Future<void> _centerOnCurrentLocation() async {
final position = await LocationService().getCurrentLocation();
_controller?.animateCamera(
CameraUpdate.newLatLng(
LatLng(position.latitude, position.longitude),
),
);
}
}
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Machine Learning integration for Flutter using TensorFlow Lite and Firebase ML
日本語の概要は準備中です。原文の説明を表示しています。
Implement tasks from an OpenSpec change. Use when the user wants to start implementing, continue implementation, or work through tasks.
日本語の概要は準備中です。原文の説明を表示しています。
Archive a completed change in the experimental workflow. Use when the user wants to finalize and archive a change after implementation is complete.
日本語の概要は準備中です。原文の説明を表示しています。
Enter explore mode - a thinking partner for exploring ideas, investigating problems, and clarifying requirements. Use when the user wants to think through something before or during a change.
日本語の概要は準備中です。原文の説明を表示しています。
Propose a new change with all artifacts generated in one step. Use when the user wants to quickly describe what they want to build and get a complete proposal with design, specs, and tasks ready for implementation.
日本語の概要は準備中です。原文の説明を表示しています。
Sync delta specs from a change to main specs. Use when the user wants to update main specs with changes from a delta spec, without archiving the change.
日本語の概要は準備中です。原文の説明を表示しています。