chapter05

Chapter05

Optimize response more contextual

In this chapter, you learn techniques to build response which is optimized to the context.

Check suggestedResponseStyle in ControlInput

suggestedResponseStyle is the property in ControlInput class. It would be VOICE, SCREEN or INDETERMINATE. Recommend to just showing the screen without voice when SCREEN, respond with the voice and show screen optionaly if VOICE, INDETERMINATE means response type is not yet known.

You can check it in suggestedResponseStyle property of the ControlInput when build response in the Act. Example is as follows (from ./src/acts/RootControlActs.ts). Checking suggestedResponseStyle and add voice response if modality of it is VOICE.

export class ChangeColorAct extends ContentAct {
    render(input: ControlInput, responseBuilder: ControlResponseBuilder): void {
        if (input.suggestedResponseStyle.modality === OutputModality.VOICE) {
            responseBuilder.addPromptFragment('Okay, this is the new color');
        }

        const colors = ['black', 'white', 'red', 'blue', 'green'];

        responseBuilder.addAPLExecuteCommandsDirective('BackgroundColorToken', [
            {
                type: 'SetValue',
                componentId: 'BackgroundColorFrame',
                property: 'colorName',
                value: colors[getRandomInt(colors.length)],
            },
        ]);
    }
}

Test

As this test case shows, Alexa wouldn't say anything if sends request by touch interaction (from ./tests/index.test.ts).

describe('index', () => {
    test('Change color by voice', async () => {
        const invoker = new SkillInvoker(
            new ControlHandler(new BackgroundColorControlManager()),
        );

        await testTurn(
            invoker,
            '',
            TestInput.launchRequest(),
            "a: Welcome to the color generator, I'll generate the color randomly. Please say 'generate new color' if you want to generate new one",
        );

        await testTurn(
            invoker,
            'u: Change color',
            TestInput.intent('ChangeColorIntent'),
            'a: Okay, this is the new color',
        );
    });

    test('Change color by touch', async () => {
        const invoker = new SkillInvoker(
            new ControlHandler(new BackgroundColorControlManager()),
        );

        await testTurn(
            invoker,
            '',
            TestInput.launchRequest(),
            "a: Welcome to the color generator, I'll generate the color randomly. Please say 'generate new color' if you want to generate new one",
        );

        await testTurn(
            invoker,
            'u: Change color',
            TestInput.simpleUserEvent(['RootControl']),
            'a:',
        );
    });
});

Let's deploy and run it.

See chapter01.

And I prepared the recognition model to interactionModel.json. You can build interaction model by using it.