cached_resource 1.0.5

Cached resource implementation based on `NetworkBoundResource` to follow the single source of truth principle.

Cached Resource

Cached resource implementation based on the NetworkBoundResource approach to follow the single source of truth principle.

Define a cached resource, subscribe for updates in multiple places, and be sure that a single network request will be called and all listeners receive a new value.

Usage

To use this plugin, add cached_resource as a dependency in your pubspec.yaml file.

Resource Storage

From the box there is only In-Memory storage shipped with the package.

Other storages should be added as new dependencies:

1. resource_storage_hive - simple persistent storage based on hive with simple JSON decoder.
2. resource_storage_secure - secure persistent storage based flutter_secure_storage with simple JSON decoder.

Configuration

In any place before usage of CachedResource, call ResourceConfig.setup and provide factories for persistent and/or secure storage.

Note: This step is required if you want to use CachedResource.persistent and CachedResource.secure. But it is optional if you will use CachedResource.inMemory or CachedResource.new.

void main() {
  // Configuration for cached_resource.
  ResourceConfig.setup(
    //inMemoryStorageFactory: const CustomMemoryResourceStorageProvider(),
    persistentStorageFactory: const HiveResourceStorageProvider(),
    secureStorageFactory: const FlutterSecureResourceStorageProvider(),
    logger: CustomLogger(),
  );

  runApp(const MyApp());
}

Define a resource/repository

There are a few ways to create a resource depending on used storage.

With In-Memory cache

See example.

class AccountBalanceRepository extends CachedResource<String, AccountBalance> {
  AccountBalanceRepository(AccountBalanceApi api)
      : super.inMemory(
          'account_balance',
          fetch: api.getAccountBalance,
          cacheDuration: const CacheDuration(minutes: 15),
        );
}

//or

final accountBalanceResource = CachedResource<String, AccountBalance>.inMemory(
  'account_balance',
  fetch: api.getAccountBalance,
  decode: AccountBalance.fromJson,
  cacheDuration: const CacheDuration(minutes: 15),
);

With persistent cache

The persistentStorageFactory should be already set by ResourceConfig.setup. See example.

class CategoryRepository {
  CategoryRepository(CategoryApi api)
      : _categoryResource = CachedResource.persistent(
          'categories',
          fetch: (_) => api.getCategories(),
          cacheDuration: const CacheDuration(days: 15),
          decode: Category.listFromJson,
          // Use executor only if [decode] callback does really heavy work,
          // for example if it parses a large json list with hundreds of heavy items
          executor: IsolatePoolExecutor().execute,
        );

  final CachedResource<String, List<Category>> _categoryResource;

  // We can use any constant key here, as the category list does not require any identifier.
  // But in some cases, you may need a unique key; for example, if you need to separate lists
  // by the currently authenticated user, you can use currentUserId as a key.
  final _key = 'key';

  Stream<Resource<List<Category>>> watchCategories() =>
      _categoryResource.asStream(_key);

  Future<void> removeCategoryFromCache(String categoryId) {
    return _categoryResource.updateCachedValue(
        _key,
            (categories) =>
            categorys?.where((category) => category.id != categoryId).toList());
  }

  Future<void> invalidate() => _categoryResource.invalidate(_key);
}

With secure cache

The secureStorageFactory should be already set by ResourceConfig.setup. See example.

class ProductSecretCodeRepository extends CachedResource<String, String> {
  ProductSecretCodeRepository(ProductApi api)
      : super.secure(
          'secret_code',
          fetch: api.getProductSecretCode,
          cacheDuration: const CacheDuration.newerStale(),
        );
}

With custom resource storage

You can create custom resource storage by extending ResourceStorage from resource_storage.

class UserRepository extends CachedResource<String, User> {
  UserRepository(UserApi api)
      : super(
          'users',
          fetch: api.getUserById,
          cacheDuration: const CacheDuration(days: 15),
          storage: YourCustomStorage(),
        );
}

Pageable resource

You can create a resource that can load a list of items page by page. See example.

class TransactionHistoryRepository
        extends OffsetPageableResource<String, TransactionItem> {
   TransactionHistoryRepository(TransactionHistoryApi api)
           : super.persistent(
      'transaction_history',
      loadPage: (filter, offset, limit) => api.getTransactionHistoryPage(filter, offset, limit),
      cacheDuration: const CacheDuration(minutes: 15),
      decode: TransactionItem.fromJson,
      pageSize: 15,
   );
//or
// : super.inMemory(
//     'transaction_history',
//     loadPage: api.getTransactionHistoryPage,
//     cacheDuration: const CacheDuration(minutes: 15),
//     pageSize: 15,
//   );
}

void init() {
// Then listen for a data
  transactionHistoryRepository.asStream(filter).listen(_showListOfItems);
}
// on need to load new page
void loadMore() => transactionHistoryRepository.loadNextPage(filter);

Listen for the resource stream or get a value

Call cachedResource.get(key) to get a single value. If the cache is not stale, it returns the cached value; otherwise, it triggers a new fetch request and returns the received value.

void foo() async {
  final resource = await resource.get(productId);
  if (resource.hasData) {
    final product = resource.data!;
    // do some work with product
  } else if (resource.isError) {
    final error = resource.error;
    final Product? productFromCache = resource.data;
    // show an error or use cached data
  }
}

To listen for resource updates, call cachedResource.asStream(key). It will emit Resource that can be one of 3 states:

• Resoorce.loading(data) - fetch request triggered. Resource.data may contain old cached value.
• Resoorce.success(data) - fetch request completed with fresh data or cache is not stale yet. Resource.data contains non null fresh value.
• Resoorce.error(data, error) - fetch request completed with error. Resource.data may contain old cached value.

void startListening() async {
   _subscription = _categoryRepository.watchCategories().listen((resource) {
      if (resource.hasData) {
         final categories = resource.data!;
         // show categories
      } else if (resource.isError) {
         final error = resource.error!;
         final cachedCategories = resource.data;
         // handle error
      } else /*if (resource.isLoading)*/ { 
         // show loading state
      }
   });
}

// On need to reload categories from the server
void refresh() => _categoryRepository.invalidate();

void deleteCategory(String categoryId) async {
  await _api.deleteCategory(categoryId);
  // We don't want to reload the list from the server,
  // so we just delete the item from the list in cache.
  // Each observer will receive an updated list immediately.
  _categoryRepository.removeCategoryFromCache(categoryId);
}

Contribution

Any contribution is welcome!