5. Register and use it

Bundle the files

One library file ties the parts together. part files share imports, and the .g.dart file holds the generated JSON code.

import 'dart:async';
import 'package:json_annotation/json_annotation.dart';
import 'package:carp_serializable/carp_serializable.dart';
import 'package:carp_core/carp_core.dart' hide Smartphone;
import 'package:carp_mobile_sensing/carp_mobile_sensing.dart';

part 'mug_data.dart';
part 'mug_probe.dart';
part 'mug_device_manager.dart';
part 'mug_package.dart';
part 'carp_mug_package.g.dart';

Register and use it

Register before you create the study. Then the mug is an ordinary device in the protocol:

SamplingPackageRegistry().register(MugSamplingPackage());

final phone = Smartphone();
final mug = MugDevice();

final protocol = SmartphoneStudyProtocol(
  name: 'Coffee and sleep',
  dataEndPoint: SQLiteDataEndPoint(),
)
  ..addPrimaryDevice(phone)
  ..addConnectedDevice(mug, phone)
  ..addTaskControl(
    ImmediateTrigger(),
    BackgroundTask(name: 'Coffee', measures: [Measure(type: MugSamplingPackage.SIP)]),
    mug,
  );

Add the same lines to the Sleep & mood protocol, next to the morning survey. The study now stores a sip stream and a mood answer each morning, and “coffee after 16:00 vs. mood next morning” becomes a query.

What the protocol looks like

These are the relevant parts of the real serialized protocol. The mug is a connected device, tied to the phone, and its task starts on an immediate trigger:

"connectedDevices": [
  {
    "__type": "dk.carp.cams.devices.MugDevice",
    "roleName": "Mug",
    "isOptional": true,
    "serviceUuids": [],
    "allowDuplicates": true
  }
],
"connections": [{ "roleName": "Mug", "connectedToRoleName": "Smartphone" }],
"tasks": [
  {
    "__type": "dk.cachet.carp.common.application.tasks.BackgroundTask",
    "name": "Coffee",
    "measures": [
      {
        "__type": "dk.cachet.carp.common.application.tasks.Measure.DataStream",
        "type": "dk.cachet.carp.mug.sip"
      }
    ]
  }
],
"taskControls": [
  {
    "triggerId": 2,
    "taskName": "Coffee",
    "destinationDeviceRoleName": "Mug",
    "control": "Start"
  }
]

CAMS also added a Monitoring Mug task (with a NoOpTrigger). Every device gets one, for errors and task events.

Test it without hardware

The fake Mug lets you test the chain in plain flutter test:

final manager = MugDeviceManager(MugDevice.DEVICE_TYPE);
await manager.onConnect();
manager.status = DeviceStatus.connected;

final probe = SipProbe()..deviceManager = manager;
final m = await probe.stream!.first;
expect(m.data, isA<MugSip>());

A second test can build the protocol, encode it with jsonEncode, and decode it with SmartphoneStudyProtocol.fromJson. If onRegister() misses a class, that test fails.

Release

Publish it as a Dart package on pub.dev. Users add it and the carp_mobile_sensing package, then register it as above. The existing packages are listed under Available packages.