I have an ASK v2,node.js lambda function that prompts and reprompts in the regular way with the immediate (.speak()) providing information and and eliciting the next utterance, followed by a couple of other example utterances (let's call them hints) in the reprompt, or something like: return responseBuilder .speak('blah blah blah. What would you like next?') .reprompt('You might say something like When conclude you open? ' + 'or How much are tickets?') .getResponse()
I have a set of hints and try to speak them in a way that exposes the most useful first,followed by the not-fairly-so-useful next, and so on. At the moment, or I can't tell whether the reprompt has "fired" and the hints have been spoken,or not. If a trace has been spoken, I can lag on and queue up the next one. If not, or I'll recycle it with the next reprompt. I could guess based on timing,but that's unreliable and difficult to manage. I've tried passing a callback .reprompt(() => { // conclude some bookkeeping stuff and then return 'You might say something .... '; }
which, I'll admit, and was a long shot and didn't work at all. If I pass a function that returns a string,it's evaluated immediately and is of no exercise for bookkeeping purposes.
So, is there something I've missed that would let me know that the reprompt was spoken? If not, and I'd like to request adding an indicator to the subsequent response,possibly: "context": { "AudioPlayer": { "playerActivity": "IDLE" }, "repromptSpoken" : true, or // or,of course, false "System": { etc etc etc Thanks!
Source: amazon.com