Handle Data Layer events on Wear

When you make a call to the Data Layer API, you can receive the status of the call when it completes. You also can listen for data events resulting from data changes that your app makes anywhere on the Wear OS by Google network.

For an example of effectively working with the Data Layer API, see the Android DataLayer Sample app.

Wait for the status of Data Layer calls

Calls to the Data Layer API—such as a call using the putDataItem method of the DataClient class—sometimes return a Task<ResultType> object. As soon as the Task object is created, the operation is queued in the background. If you do nothing more after this, the operation eventually completes silently.

However, you usually want to do something with the result after the operation completes, so the Task object lets you wait for the result status, either asynchronously or synchronously.

Asynchronous calls

If your code is running on the main UI thread, don't make blocking calls to the Data Layer API. Run the calls asynchronously by adding a callback method to the Task object, which fires when the operation is completed:

// Using Kotlin function references
task.addOnSuccessListener(::handleDataItem)
task.addOnFailureListener(::handleDataItemError)
task.addOnCompleteListener(::handleTaskComplete)
...
fun handleDataItem(dataItem: DataItem) { ... }
fun handleDataItemError(exception: Exception) { ... }
fun handleTaskComplete(task: Task<DataItem>) { ... }

See the Task API for other possibilities, including chaining the execution of different tasks.

Synchronous calls

If your code is running on a separate handler thread in a background service, such as in a WearableListenerService, it's fine for the calls to block. In this case, you can call Tasks.await() on the Task object, which blocks until the request completes and returns a Result object. This is shown in the following example.

Note: Do not call this while on the main thread.

try {
    Tasks.await(dataItemTask).apply {
        Log.d(TAG, "Data item set: $uri")
    }
}
catch (e: ExecutionException) { ... }
catch (e: InterruptedException) { ... }

Listen for Data Layer events

Because the data layer synchronizes and sends data across the handheld and wearable devices, you usually need to listen for important events like data items being created and messages being received.

To listen for data layer events, you have two options:

With both of these options, you override the data event callback methods for the events you are interested in handling.

Note: Consider your app's battery usage when choosing a listener implementation. A WearableListenerService is registered in the app's manifest and can launch the app if it is not already running. If you only need to listen for events when your app is already running, which is often the case with interactive applications, then don't use a WearableListenerService. Instead, register a live listener. For example, use the addListener method of the DataClient class. This can reduce the load on the system and reduce battery usage.

Use a WearableListenerService

You typically create instances of WearableListenerService in both your wearable and handheld apps. However, if you are not interested in data events in one of the apps, then you don't need to implement the service in that app.

For example, you can have a handheld app that sets and gets data item objects and a wearable app that listens for these updates to update its UI. The wearable app never updates any of the data items, so the handheld app doesn't listen for any data events from the wearable app.

Some of the events you can listen for using WearableListenerService are the following:

  • onDataChanged(): whenever a data item object is created, deleted, or changed, the system triggers this callback on all connected nodes.
  • onMessageReceived(): a message sent from a node triggers this callback on the target node.
  • onCapabilityChanged(): when a capability that an instance of your app advertises becomes available on the network, that event triggers this callback. If you're looking for a nearby node, you can query the isNearby() method of the nodes provided in the callback.

You can also listen for events from ChannelClient.ChannelCallback, such as onChannelOpened().

All the preceding events are executed in a background thread, not on the main thread.

To create a WearableListenerService, follow these steps:

  1. Create a class that extends WearableListenerService.
  2. Listen for the events that you're interested in, such as onDataChanged().
  3. Declare an intent filter in your Android manifest to notify the system about your WearableListenerService. This declaration lets the system bind your service as needed.

The following example shows how to implement a WearableListenerService:

class DataLayerListenerService : WearableListenerService() {

    override fun onDataChanged(dataEvents: DataEventBuffer) {
        if (Log.isLoggable(TAG, Log.DEBUG)) {
            Log.d(TAG, "onDataChanged: $dataEvents")
        }

        // Loop through the events and send a message
        // to the node that created the data item.
        dataEvents
            .map { it.dataItem.uri }
            .forEach { uri ->
                // Get the node ID from the host value of the URI.
                val nodeId: String = uri.host!!
                // Set the data of the message to be the bytes of the URI.
                val payload: ByteArray = uri.toString().toByteArray()

                // Send the RPC.
                Wearable.getMessageClient(this)
                    .sendMessage(
                        nodeId,
                        DATA_ITEM_RECEIVED_PATH,
                        payload
                    )
            }
    }
}
DataLayerService.kt

The following section explains how to use an intent filter with this listener.

Use filters with WearableListenerService

An intent filter for the WearableListenerService example shown in the previous section might look like this:

<service
    android:name=".snippets.datalayer.DataLayerListenerService"
    android:exported="true"
    tools:ignore="ExportedService" >
    <intent-filter>
        <action android:name="com.google.android.gms.wearable.DATA_CHANGED" />
        <data
            android:scheme="wear"
            android:host="*"
            android:path="/start-activity" />
    </intent-filter>
</service>
AndroidManifest.xml

The DATA_CHANGED action filter tells the system your app is interested in data layer events.

In this example, the watch listens for the /start-activity data item, and the phone listens for the /data-item-received (DATA_ITEM_RECEIVED_PATH) message response.

Standard Android filter matching rules apply. You can specify multiple services per manifest, multiple intent filters per service, multiple actions per filter, and multiple data stanzas per filter. Filters can match on a wildcard host or on a specific one. To match on a wildcard host, use host="*". To match on a specific host, specify host=<node_id>.

You can also match a literal path or path prefix. To do this, you must specify a wildcard or specific host. Otherwise, the system ignores the path you specify.

For more information about the filter types that Wear OS supports, see the API reference documentation for WearableListenerService.

For more information on data filters and matching rules, see the API reference documentation for the <data> manifest element.

When matching intent filters, remember two important rules:

  • If no scheme is specified for the intent filter, the system ignores all the other URI attributes.
  • If no host is specified for the filter, the system ignores all the path attributes.

Use a live listener

If your app only cares about data-layer events when the user is interacting with the app, it may not need a long-running service to handle every data change. In such a case, you can listen for events in an activity by implementing one or more of the following interfaces:

To create an activity that listens for data events, follow these steps:

  1. Implement the required interfaces.
  2. In the onCreate() or onResume() method, call Wearable.getDataClient(this).addListener(), MessageClient.addListener(), CapabilityClient.addListener(), or ChannelClient.registerChannelCallback() to notify Google Play services that your activity is interested in data layer events.
  3. In onStop() or onPause(), unregister any listeners with DataClient.removeListener(), MessageClient.removeListener(), CapabilityClient.removeListener(), or ChannelClient.unregisterChannelCallback().
  4. If an activity only needs to receive events with a specific path prefix, add a listener with a prefix filter to only receive data relevant to the current application state.
  5. Implement onDataChanged(), onMessageReceived(), onCapabilityChanged(), or methods from ChannelClient.ChannelCallback, depending on the interfaces that you implemented. These methods are called on the main thread, or you can specify a custom Looper using WearableOptions.

Here's an example that implements DataClient.OnDataChangedListener:

class MainActivity : Activity(), DataClient.OnDataChangedListener {

    public override fun onResume() {
        super.onResume()
        Wearable.getDataClient(this).addListener(this)
    }

    override fun onPause() {
        super.onPause()
        Wearable.getDataClient(this).removeListener(this)
    }

    override fun onDataChanged(dataEvents: DataEventBuffer) {
        dataEvents.forEach { event ->
            if (event.type == DataEvent.TYPE_DELETED) {
                Log.d(TAG, "DataItem deleted: " + event.dataItem.uri)
            } else if (event.type == DataEvent.TYPE_CHANGED) {
                Log.d(TAG, "DataItem changed: " + event.dataItem.uri)
            }
        }
    }
}
DataLayerService.kt

Caution: Before using the Wearable Data Layer API, check that it's available on a device; otherwise, an exception occurs. Use the GoogleApiAvailability class, as implemented in Horologist.

Use filters with live listeners

As previously mentioned, just as you can specify intent filters for manifest-based WearableListenerService objects, you can use intent filters when registering a live listener through the Wearable API. The same rules apply to both API-based live listeners and manifest-based listeners.

A common pattern is to register a listener with a specific path or path prefix in an activity's onResume() method, and then to remove the listener in the activity's onPause() method. Implementing listeners in this fashion lets your app more selectively receive events, improving its design and efficiency.