Bixby Developer Center


This capsule demonstrates how to use continuations in your training.

Download Capsule


Because you cannot submit a capsule with the example namespace, in order to test a sample capsule on a device, you must change the id in the capsule.bxb file from example to your organization's namespace before making a private submission.

For example, if your namespace is acme, change example.continuation to acme.continuation.

A continuation is an utterance that relies on a previous utterance for context. For instance, the user might ask Bixby: "What's the weather in Boston?"

Bixby would respond with the weather forecast for Boston. Then, the user might ask: "How about Chicago?"

A human would understand that this was a request for the weather forecast in Chicago by remembering the context of the previous question. For Bixby to understand this, the training for the utterance "How about Chicago?" needs to be a continuation. This lets Bixby know that this utterance continues the previous one--that is, it relies on data from the previous utterance.

The continuation example capsule lets you ask Bixby for information about planets in our solar system. Try the utterances in the training, and examine the way the training examples are constructed. For instance, the utterance "how far is it to Saturn" has the following Aligned NL:

[g:CompareDistanceResult] how far is it to
{[g:ThatPlanet] (Saturn)[v:PlanetName]}

Continuing that with "and to Jupiter" has the following Aligned NL:

[g:CompareDistanceResult:continue] and to
{[g:ThatPlanet] (Jupiter)[v:PlanetName]}

This goal is a continuation of CompareDistanceResult.

A JSON data structure with information about the planets is contained in code/lib/planets.js.