progressive response is not progressive

[alrouen]

I'm submitting a...

[ ] Regression (a behavior that used to work and stopped working in a new release)
[X] Bug report 
[ ] Performance issue
[ ] Feature request
[ ] Documentation issue or request
[ ] Other... Please describe:

Expected Behavior

When sending a progressive response (https://developer.amazon.com/docs/custom-skills/send-the-user-a-progressive-response.html) is a lambda function (nodejs 8.x + ask-sdk-core 2.0.5), I'm expecting Alexa to send a response (i.e. : ok please wait...) before executing an async/long running action (i.e.: sleep, calling another ) and finally giving a final response.

Current Behavior

The progressive response is not sent before executing the next promise running task but:

• after the execution of all promises
• but the order of speech actions remain good

So if your handler behavior is the following:

• await progressive response : "please wait"
• await sleep 5s
• return final response "remote service answer is xxxx"

You expect to ear:

• "please wait"
• 5s of nothing
• "remote service answer is xxxx"

You will ear:

• 5s of nothing
• "please wait"
• "remote service answer is xxxx"

Steps to Reproduce (for bugs)

This bug can be reproduced by using the sample code from the Alexa cookbook.

• sample code available here : https://github.com/alexa/alexa-cookbook/tree/master/feature-demos/skill-demo-progressive-response
• Using the alexa console (https://developer.amazon.com/alexa/console/ask) to setup the skill
• create the lambda function with source code from: https://github.com/alexa/alexa-cookbook/tree/master/feature-demos/skill-demo-progressive-response/lambda/custom
• configure the skill to use the lambda
• Invoke the skill with the alexa console in test mode

Context

Building an Alexa skill that have to deal with async task. Big impact on user experience if not fixed.

Your Environment

• ASK SDK for Node.js used: 2.0.5 or ASK SDK Core / mode for Node.js used: 2.0.5 / 1.3.0
• AWS Lambda for Node.js 8.10

Node.js and NPM Info

• Node.js version used for development: 8.10

[alrouen]

I got a response from the Alexa support team explaining me that this is the expected behavior when testing the skill using the alexa console.

But I also get the same behavior from the device.

[sgmbh]

I have the same problem... After the spoken command the echo spot is silent for 3-4 seconds an then speaks the "please wait" & "finished successfully" just after each other. It's like one response.

[lucas-rudd]

Is there any update on this? We've run into this issue when testing our new Skill and it is a serious roadblock for us. In theory the progressive response API sounds like a game-changer; but, we've experienced the same issue reported above in both the simulator and on an echo dot.

[tianrenz]

Hi all,

Sorry kept you all waiting. I tested this with a sample hello world skill and have no issue with the response order. Below are the code sample i used.

const HelloWorldIntentHandler = {
  canHandle(handlerInput) {
    return handlerInput.requestEnvelope.request.type === 'IntentRequest'
      && handlerInput.requestEnvelope.request.intent.name === 'HelloWorldIntent';
  },
  async handle(handlerInput) {
    const speechText = 'Hello World!';

    await callDirectiveService(handlerInput);

    return new Promise((resolve, reject) => {
      setTimeout(() => {
        const response = handlerInput.responseBuilder
                        .speak(speechText)
                        .withSimpleCard('Hello World', speechText)
                        .getResponse();
        resolve(response)
      }, 3000);
    })
  },
};

function callDirectiveService(handlerInput) {
  // Call Alexa Directive Service.
  const requestEnvelope = handlerInput.requestEnvelope;
  const directiveServiceClient = handlerInput.serviceClientFactory.getDirectiveServiceClient();

  const requestId = requestEnvelope.request.requestId;

  // build the progressive response directive
  const directive = {
    header: {
      requestId,
    },
    directive: {
      type: 'VoicePlayer.Speak',
      speech: `this is a test speech`,
    },
  };

  // send directive
  return directiveServiceClient.enqueue(directive);
}

Please make sure you are testing this on an physical alexa device. The test console on developer portal doesn't support progressive response at the moment.

Regards

[artemkovalyov]

@tianrenz, your solution worked for me. Thank you very much for testing as it's lacking from the docs, how to use this part of SDK. By the way, in web testing tool it was fine as well.

[nikhilym]

Thanks for confirming @artemkovalyov !!

@alrouen , @sgmbh , @lucas-rudd did you guys try it out with @tianrenz's code-snippet? Please update the issue with your findings!!

[sgmbh]

My skill is based on ASK SDK v1. Is it possible that the progressive response only works with the v2 request handler interface? Or have I made a mistake in my code? I tested this with the online tool from amazon and with my real echo spot. Everything works, only the progressive response is to late. The behaviour is exactly like alrouen said:

• some seconds of nothing
• "please wait"
• "my ping message"

Thanks for your help!

'MyIntent': function () {
    .....
    callDirectiveService(this.event)
    .....
    apiRequestService.send(this.event.session.user.accessToken, '/api/devices/Ping')
        .then((success) => {
            const speechOutput = PING_MESSAGE;
            this.response.cardRenderer(SKILL_NAME, PING_MESSAGE);
            this.response.speak(speechOutput);
            this.emit(':responseReady');
        })
        .catch((err) => {
            const speechOutput = ERROR_MESSAGE;
            const reprompt = HELP_REPROMPT;
            this.response.cardRenderer(SKILL_NAME, ERROR_MESSAGE);
            this.response.speak(speechOutput).listen(reprompt);
            this.emit(':responseReady');
        });
},

.....

function callDirectiveService(event) {
// Call Alexa Directive Service.
const ds = new Alexa.services.DirectiveService();
const requestId = event.request.requestId;
const directive = new Alexa.directives.VoicePlayerSpeakDirective(requestId, WAIT_MESSAGE);
const endpoint = event.context.System.apiEndpoint;
const token = event.context.System.apiAccessToken;
return ds.enqueue(directive, endpoint, token);
};

[tianrenz]

Hi @sgmbh ,

callDirectiveService(event) returns a Promise. So you need to put your following code in the then((data) => {}) to ensure the correct execution order.

Regards

[tianrenz]

Hi all,

Closing this issue. Please reference the solution posted above.

Regards

[sgmbh]

Hi @tianrenz, I should have thought of that myself. Thanks a lot! It works now. Regards

[tianrenz]

Hi @artemkovalyov ,

We are aware of the indeterministic behavior. In our testing, the scenario described would happen randomly. Unfortunately, there's no clear reason as for what triggers it. We believe it could be a service issue, not SDK related. We would recommend to report this on Alexa developer forum.

Regards

[jimenaCabrejas]

Hi,

Do you know if this solution should work on the Amazon Alexa Android App? I tried @tianrenz example skill but I got the problem described above. It waits for 3 seconds then says this is a test speech and hello world just after each other.

Thanks and regards

[Junot-Jean]

hi? i'm having the same issue, i've been trying to solve it with @tianrenz code but nothing changes.