Loon은 Dart 및 Flutter 애플리케이션을 위한 반응형 컬렉션 데이터 저장소로, NoSQL 및 SQL 기능을 갖춘 지속성 저장소를 제공합니다. 데이터가 변경될 때 자동으로 UI를 업데이트할 수 있도록 반응형으로 데이터 컬렉션을 관리할 수 있습니다. 크로스플랫폼 사용을 위해 설계되어 Android, iOS, Web, macOS를 지원하며, 오프라인 우선 기능과 실시간 데이터 동기화가 필요한 앱에 적합합니다. 가벼운 설계로 Flutter 워크플로우에 원활하게 통합됩니다.
사용 사례
오프라인 우선 모바일 애플리케이션
실시간 데이터 동기화
크로스 플랫폼 데이터 지속성
반응형 데이터를 갖춘 동적 UI
웹 앱에서의 로컬 데이터 관리
주요 기능
반응형 데이터 바인딩
지속 가능한 NoSQL 및 SQL 저장소
크로스 플랫폼 호환성
데이터 변경 시 자동 UI 업데이트
경량이며 쉽게 통합 가능
적합한 대상
로컬 데이터 지속성을 찾는 플러터 개발자
오프라인 기능이 필요한 앱
반응형 상태 관리가 필요한 프로젝트
다중 플랫폼 앱을 개발하는 팀
간단하면서도 강력한 데이터 저장소를 원하는 개발자
고려 사항
SQL 기능을 익히기 위해 학습 곡선이 필요할 수 있습니다
성숙한 대안에 비해 문서가 제한적
성능은 데이터 크기와 복잡성에 따라 달라집니다
대규모 서버 측 데이터베이스에 적합하지 않습니다
백엔드 서비스보다 클라이언트 측 데이터에 더 적합
원본 README
연결된 소스 저장소의 README 원문 스냅샷입니다. 최신 내용은 저장소에서 확인하세요.
Loon
Android
iOS
Web
macOS
✅
✅
✅
✅
Loon is a reactive document data store for Flutter.
Features
Synchronous reading, writing and querying of documents.
Streaming of changes to documents and queries.
Out of the box persistence and encryption.
Install
flutter pub add loon
➕ Creating documents
Loon makes it easy to work with collections of documents.
If persistence is enabled, then a typed collection needs to specify a fromJson/toJson serialization pair. In order to avoid having to specify types or serializers whenever a collection is accessed, it can be helpful to store the collection in a variable or as an index on the data model:
You can then use Flutter's built-in StreamBuilder or the library's DocumentStreamBuilder widget to access data from widgets:
classMyWidgetextendsStatelessWidget{
@overridebuild(context) {
returnDocumentStreamBuilder(
doc: UserModel.store.doc('1'),
builder: (context, snap) {
final user = snap?.data;
if (user == null) {
returnText('Missing user');
}
returnText('Found user ${user.name}');
}
)
}
}
𖢞 Subcollections
Documents can be nested under subcollections. Documents in subcollections are uniquely identified by the path to their collection and
their document ID.
final friendsCollection = UserModel.store.doc('1').subcollection('friends');
friendsCollection.doc('2').create(UserModel(name: 'Jack', age: 17));
friendsCollection.doc('3').create(UserModel(name: 'Brenda', age: 40));
friendsCollection.doc('4').create(UserModel(name: 'Bill', age: 70));
final snaps = friendsCollection.get();
for (final snap in snaps) {
print("${snap.data.name}: ${snap.path}");
// Jack: users__1__friends__2// Brenda: users__1__friends__3// Bill: users__1__friends__4
}
🔎 Queries
Documents can be filtered using queries:
final snapshots = friendsCollection.where((snap) => snap.data.name.startsWith('B')).get();
for (final snap in snapshots) {
print(snap.data.name);
// Brenda// Bill
}
Queries can also be streamed, optionally using the QueryStreamBuilder:
Deleting a document removes it and all of its subcollections from the store.
UserModel.store.doc('1').delete();
🌊 Streaming changes
Documents and queries can be streamed for changes which provides the previous and current document data as well as the event type of the change:
UserModel.store.streamChanges().listen((changes) {
for (final changeSnap in changes) {
switch(changeSnap.event) {
caseBroadcastEvents.added:
print('New document ${changeSnap.id} was added to the collection.');
break;
caseBroadcastEvents.modified:
print('The document ${changeSnap.id} was modified from ${changeSnap.prevData} to ${changeSnap.data}.');
break;
caseBroadcastEvents.removed:
print('${changeSnap.id} was removed from the collection.');
break;
caseBroadcastEvents.hydrated:
print('${changeSnap.id} was hydrated from the persisted data.');
break;
}
}
});
🔁 Data Dependencies
Data relationships in the store can be established using the data dependencies builder.
In this example, whenever a post's associated user is updated, the post will also be rebroadcast to its active listeners.
Additionally, whenever a document is updated, it will rebuild its set of dependencies, allowing documents to support dynamic dependencies
that can change in response to updated document data.
🪴 Root collection
Not all documents necessarily make sense to be grouped together under any particular collection. In this scenario, any one-off documents can be stored
on the root collection:
Loon.doc('current_user_id').create('1');
🗄️ Data Persistence
Persistence is supported on both web and native platforms. It works out of the box and can be configured on app start.
Native platforms (iOS, Android, macOS) use a default file-based persistence implementation, while web persists data to IndexedDB.
The currently available persistence options are broken down by platform:
Native
SqlitePersistor: A SQLite persistence implementation using sqflite. Documents are distributed in rows based on their persistence configuration.
FilePersistor: A file-based persistence implementation for native platforms. Documents are stored in one or more files based on the persistence configuration.
Web
IndexedDBPersistor: The default persistence implementation for web platforms. Documents are stored in IndexedDB, a low-level API for web-based client-side storage.
Note: Encryption on web is experimental through flutter_secure_storage. It is important to enable the correct headers in order to ensure security of encryption on web.
⚙️ Configuration
Specify your preferred persistor on startup or use Persistor.current() to dynamically select the default persistence implementation for the current platform (FilePersistor on Android/iOS, IndexedDBPersistor on web).
No persisted data is available on startup until it is manually requested by calling hydrate.
The call to hydrate returns a Future that resolves when the data has been hydrated from the persistence layer. By default, calling hydrate() will hydrate all persisted data. If only certain data should be hydrated, then it can be called with a list of documents and collections to hydrate. All subcollections of the specified paths are also hydrated.
Persistence options can be specified both globally as well as on a per-collection basis.
// main.dartvoidmain() {
// Globally enable encryption.Loon.configure(persistor: FilePersistor(settings: PersistorSettings(encrypted: true));
Loon.hydrate().then(() {
print('Hydration complete');
});
}
// models/user.dartclassUserModel{
finalString name;
finalint age;
UserModel({
requiredthis.name,
requiredthis.age,
});
staticCollection<UserModel> get store {
returnLoon.collection(
'users',
fromJson: UserModel.fromJson,
toJson: (user) => user.toJson(),
// Disable encryption specifically for this collection and its subcollections.
settings: PersistorSettings(encrypted: false),
)
}
}
In this example, file encryption is enabled globally for all collections, but disabled
specifically for the users collection and its subcollections in the store.
By default all data in a single __store__.json persistence file.
loon>
__store__.json
If data needs to be persisted differently, either by merging data across collections into a single store or by breaking down a collection
into multiple stores, then a custom persistence key can be specified on the collection:
Now instead of storing all users in the users.json file, they will be distributed across multiple files based on the user's age. The key is recalculated
whenever a document's data changes and if its associated key is updated, then the document is moved from its previous file to its updated location.
🎨 Custom persistence
If you would prefer to persist data using an alternative implementation than the default persistors, you can implement the persistence interface:
import'package:loon/loon.dart';
classMyPersistorextendsPersistor{
/// Initialization function called when the persistor is instantiated to execute and setup work.Future<void> init();
/// Persist function called with the bath of documents that have changed (including been deleted) within the last throttle window/// specified by the [Persistor.persistenceThrottle] duration.Future<void> persist(Set<Document> docs);
/// Hydration function called to read data from persistence. If no references are specified,/// then it hydrates all persisted data. if refs are specified, it hydrates only the data at/// and under those paths.Future<Json> hydrate([Set<StoreReference>? refs]);
/// Clear function used to clear all documents under the given store paths.Future<void> clear(Set<StoreReference> refs);
/// Clears all documents and removes all persisted data.Future<void> clearAll();
}