Data Batcher

Data batcher batches and de-dupes data fetched in the same task of the event loop.

final batcher = DataBatcher<String>(
  execute: (ids) async {
    print(ids) // ['1', '2', '3', '4']
    return Api.fetch(ids);
  },
);

batcher.add('1');
batcher.add('2');

batcher.addMany(['2', '3']);

await batcher.add('4');

When IDs are added to a batcher, they are grouped by the current cycle of the event loop and scheduled to be executed on a micro-task. Once executed successfully, the batcher resolves each caller's future with its data as shown below:

final batcher = DataBatcher<String>(
  execute: (ids) async {
    return ['a', 'b', 'c'];
  },
);

batcher.add('1').then(((resp) => print(resp)); // 'a'
batcher.add('2').then(((resp) => print(resp)); // 'b'

batcher.addMany(['2', '3']).then(((resp) => print(resp)) // ['b', 'c']

Data added to a batcher across different ticks of the event loop is broken into separate batches:

final batcher = DataBatcher<String>(
  execute: (ids) async {
    print(ids);
    // ['1', '2']
    // ['3']
    return Api.fetch(ids);
  },
);

batcher.add('1');
await batcher.add('2');

batcher.add('3');

By default, data IDs that are still in-flight from a previous batch which are requested again are not re-fetched:

final batcher = DataBatcher<String>(
  execute: (ids) async {
    print(ids);
    // ['1', '2']
    return Future.delayed(Duration(seconds: 5));
  },
);

batcher.add('1');
batcher.add('2');

await Future.delayed(Duration(seconds: 1));

batcher.add('1');

The second attempt to request data with ID 1 is de-duped, since it is called while the in-flight request for '1' and '2' has not resolved. The Future returned by the second call to add ID 1 will resolve when the first original batch succeeds and with its returned value for ID 1.

If de-duping of in-flight data is not preferred, the dedupeInFlight flag can be set to false:

final batcher = DataBatcher<String>(
  dedupeInFlight: false,
  execute: (ids) async {
    print(ids);
    // ['1', '2']
    // ['1']
    return Future.delayed(Duration(seconds: 5), () {...});
  },
);

batcher.add('1');
batcher.add('2');

await Future.delayed(Duration(seconds: 1));

batcher.add('1');

IDs are mapped to data responses using the order of the returned data. If needed, an idExtractor can be specified instead in order to associated response data with its matching input ID:

final batcher = DataBatcher<DataModel>(
  idExtractor: (dataModel) => dataModel.id,
  execute: (ids) async {
    print(ids);
    // ['1', '2']
    return [DataModel('1'), DataModel('2')]
  },
);

batcher.add('1');
batcher.add('2');