brianmacintosh / alexa-skill-test-framework Goto Github PK

I'm currently using version 1.2.4. It seems that when I try to set an existing session attribute using withSessionAttributes, it doesn't override it. If the session attribute is not an existing one then it gets properly set. I downloaded the latest changes from your master and this issue doesn't happen anymore, could you please release a new version?

Really like the alexa-skill-test-framework, i'd love to use it with our framework https://github.com/mediarain/voxa, which seems like we could right now as it is. However, there are some assertions we'd like to do, like making sure a transition results in a change to X state, that functionality looks like it would belong inside a plugin but the current mechanism in setExtraFeature is not extensible.

Would this be something you're interested in?

It would be really cool to assert certain words or phrases inside the output and reprompt speech instead of matching the whole string.

I really like this framework. One feature I'm missing is accessToken (account linking) and permissions. Any plans on adding them?

I am a noob in coding Alexa skills and I bump into your test framework. Cool tool btw.

I have a problem here when I try to run my tests. When I execute npm test it doesn't terminate

I just followed the code in one of your examples as an experiment but I still have the same behavior.

const alexaTest = require('alexa-skill-test-framework');

alexaTest.initialize(
	require('../index.js'),
	"skill-id",
  'user-id');

describe("Alexa skill using Staging configuration", function() {
  describe("LaunchRequest", function() {
    alexaTest.test([
      {
        request: alexaTest.getLaunchRequest(),
        says: "Welcome!", repromptsNothing: true, shouldEndSession: true
      }
    ]);
  });
  describe("CountIntent", function() {
    const slot = {'companyMetadata': 'employees', 'siteName': 'fort'};
		const requestWithSlot = alexaTest.getIntentRequest('CountIntent', slot);

		alexaTest.test([
			{
				request: requestWithSlot,
				says: "fort has 61 employees", shouldEndSession: true, repromptsNothing: true
			}
		]);
  });
});

My node version is v9.0.0

package.json is

"dependencies": {
"alexa-sdk": "^1.0.12",
"dotenv": "^4.0.0",
"pg": "^7.4.0"
},
"devDependencies": {
"alexa-skill-test-framework": "^1.1.2",
"chai": "^4.1.2",
"mocha": "^4.0.1"
}

Appreciate the help.

Hello everyone, wanted to know what to pass for index for

alexaTest.initialize(index, appId, userId, deviceId)`

I am using my custom webhook using NodeJS.
app.js -> router using alexa-verifer-middleware
app.use('/alexa', require('./routes/alexa'));

routes/alexa
router.post('/podcast', function(req, res) { const handlers = require('../handlers/alexaHandler'); if (!skill) { skill = Alexa.SkillBuilders.custom() .addRequestHandlers( handlers.LaunchRequestHandler, handlers.HelpHandler..... ).create();

handler/alexaHandler -> where all the functions are maintained like LaunchRequestHandlers and HelpIntent.

Please advice. Code is working fine but as it is becoming more complex planning to use a testing framework.

The code style of the index.js differs from the JavaScript standard but is not reflected in the editorconfig. So the automatic reformat of IDE changes the whole file.

I think there are two options:

1. Write an editorconfig to reflect the style of the index.js
2. Change the index.js to default JavaScript code style (Preferred)

Any comments on this topic?

Some intents require attributes from the session, so it would be useful to be able to set this before running a test.

There is currently withStoredAttributes that lets you set persistent attributes, but nothing for the session.

Allow switching off/on debugging tests by printing the response. Something based on
... console.log(response); var actualSay = response.response && response.response.outputSpeech ? response.response.outputSpeech.ssml : undefined; var actualReprompt = response.response && response.response.reprompt && response.response.reprompt.outputSpeech ? response.response.reprompt.outputSpeech.ssml : undefined; ...
in ../node_modules/alexa-skill-test-framework/index.js

This is a request for an enhancement.

I wrote a skill. Some of the skill's handlers modify attributes in the session object. Is there a way for me to test correctness of the session object?

'npm install alexa-skill-test-framework --save-dev' point to this branch:
https://github.com/BrianMacIntosh/alexa-skill-test-framework/blob/v1.1.1/index.js

Which doesn't contain the latest methods on master branch. Wonder if you can bump the version? Thxs

When using "it" in stead of "describe", test case always succed.

Referrs to:
alexaTest.getIntentRequest(intentName, [slots], [locale])

AMAZON.DATE returns "undefined" (contrary to other slot types returning "?"), if no or wrong slot value has been provided in the input. There is no way without hacks to simulate this behavior in the test framework.

The Session object of every request always has the new attribute set to true. This is a hardcoded value in https://github.com/BrianMacIntosh/alexa-skill-test-framework/blob/master/index.js#L400

Expected behavior would be to only have this be true for the first request in the sequence. The reference in https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/docs/alexa-skills-kit-interface-reference states that new "Returns true for a new session or false for an existing session."

I'm running into issues where the mock context times out after 3 seconds (which is the default). It would be great to be able to pass in options (such as timeout) to the aws-lambda-mock-context dependency. https://www.npmjs.com/package/aws-lambda-mock-context

While creating our Alexa skill, our team noticed that, if we forget to emit :responseReady when starting to play something in the audio player, the Alexa device will call a the Unhandled handler with the following request (see docs):

{ type: "System.ExceptionEncountered",
  requestId: "amzn1.echo-api.request.170b9121-04e5-4794-826d-1c00f51dcc8c",
  timestamp: "2017-10-02T11:01:02Z",
  locale: "en-GB",
  error:
   { type: "INVALID_RESPONSE",
     message: "An exception occurred while dispatching the request to the skill." },
  cause: { requestId: "amzn1.echo-api.request.cf8066e8-c7d8-4324-be51-f3c0d1ff9709" } }

So we changed our Unhandled handler to look a bit like this:

  Unhandled() {
    if (this.event.request.type === 'System.ExceptionEncountered') {
      // Log the system exception...
    }

    this.emit(':tell', 'Sorry, there was an error.');
  }

We would like to test how we handle system exceptions in our skill. However, as far as I know, there is no way within the testing framework to simulate an error in the request.

resolutions.forEach(resolution => {
^

TypeError: resolutions.forEach is not a function
at Object.addEntityResolutionsToRequest (C:\Users\Mohd Ashif Ali\Documents\GitHub\alexa-greet-my-friend\test\node_modules\alexa-skill-test-framework\index.js:357:15)
at Suite. (C:\Users\Mohd Ashif Ali\Documents\GitHub\alexa-greet-my-friend\test\greeter-test.js:36:23)
at Object.create (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\interfaces\common.js:112:19)
at context.describe.context.context (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\interfaces\bdd.js:40:27)
at Suite. (C:\Users\Mohd Ashif Ali\Documents\GitHub\alexa-greet-my-friend\test\greeter-test.js:30:2)
at Object.create (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\interfaces\common.js:112:19)
at context.describe.context.context (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\interfaces\bdd.js:40:27)
at Object. (C:\Users\Mohd Ashif Ali\Documents\GitHub\alexa-greet-my-friend\test\greeter-test.js:18:1)
at Module._compile (module.js:653:30)
at Object.Module._extensions..js (module.js:664:10)
at Module.load (module.js:566:32)
at tryModuleLoad (module.js:506:12)
at Function.Module._load (module.js:498:3)
at Module.require (module.js:597:17)
at require (internal/module.js:11:18)
at C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\mocha.js:250:27
at Array.forEach ()
at Mocha.loadFiles (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\mocha.js:247:14)
at Mocha.run (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\lib\mocha.js:576:10)
at Object. (C:\Users\Mohd Ashif Ali\AppData\Roaming\npm\node_modules\mocha\bin_mocha:637:18)
at Module._compile (module.js:653:30)
at Object.Module._extensions..js (module.js:664:10)
at Module.load (module.js:566:32)
at tryModuleLoad (module.js:506:12)
at Function.Module._load (module.js:498:3)
at Function.Module.runMain (module.js:694:10)
at startup (bootstrap_node.js:204:16)
at bootstrap_node.js:625:3

With a recent update, slots now support more than just a value but a match to a category, called resolutions.
https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/docs/custom-standard-request-types-reference#intentrequest

Would be nice to specify these parameters in the IntentRequest builder as well.

I'd like to be able to test synonym resolution logic in my Alexa skill but currently this framework doesn't allow any extension of the basic {name: "SlotName", value: "SlotValue"} structure.

I need to be able to add a resolutions object to a slot as documented here: https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/docs/entity-resolution-for-slot-types#intentrequest-changes

See below example JSON from the above link.

"resolutions": {
  "resolutionsPerAuthority": [
    {
      "authority": "amzn1.er-authority.echo-sdk.<skill_id>.MEDIA_TYPE",
      "status": {
        "code": "ER_SUCCESS_MATCH"
      },
      "values": [
        {
          "value": {
            "name": "song",
            "id": "SONG"
          }
        }
      ]
    }
  ]
}

Currently getIntentRequest does the following logic for slots:

if (!slots) {
  slots = {};
}
else {
  for (var key in slots) {
    slots[key] = {name: key, value: slots[key]};
  }
}

I would propose that instead we're allowed to inject the slots object as required. The Alexa SDKs will be changing quite frequently due to the constant shift in this new technology and frameworks like these will need to be able to keep up. So I would propose the following change:

intent": {"name": intentName, "slots": slots || {}}

And remove the if/else logic referenced above.

This allows users of the test framework to pass through whatever slot configurations they need. A user may pass through an invalid object but if they do their tests will fail and alert them to that.

I'm more than happy to raise a PR for this if the idea is accepted. We would really like to keep using this test framework.

Forgive me if I'm wrong, but I don't see a way to deal with different states.
I'm talking about this:

Alexa.CreateStateHandler(States.READY, { // handlers here })

It is strange. For example intent 'AMAZON.StopIntent' pass the test but every test I am expecting that alexa repeat the question fails :(

HelloWorldIntent:

module.exports = {
  'canHandle': (handlerInput) => {
    return handlerInput.requestEnvelope.request.type === 'IntentRequest'
      && handlerInput.requestEnvelope.request.intent.name === 'HelloWorldIntent';
  },
  'handle': (handlerInput) => {
    const requestAttributes = handlerInput.attributesManager.getRequestAttributes();
    const speakOutput = requestAttributes.t(('WELCOME'));
    const speakRepeat = requestAttributes.t(('REPEAT'));
    return handlerInput.responseBuilder
        .speak(speakOutput)
        .reprompt(speakRepeat)
        .getResponse();
  },
};

test:

alexaTest.initialize(
    require('../awsAdapter'),
    'amzn1.ask.skill.00000000-0000-0000-0000-000000000000',
    'amzn1.ask.account.VOID');
const textResources = require('../localization/translations.js');
alexaTest.initializeI18N(textResources);
const supportedLocales = ['en-US', 'de-DE'];
for (let i = 0; i < supportedLocales.length; i++) {
  const locale = supportedLocales[i];
  alexaTest.setLocale(locale);
  describe('skill tests', function() {
  // tests the behavior of the skill's HelloWorldIntent
    describe('CancelAndStopIntentIntent', () => {
      alexaTest.test([
        {
          request: alexaTest.getIntentRequest('AMAZON.CancelIntent'),
          says: alexaTest.t('STOP_ANSWER'), shouldEndSession: true,
        },
      ]);
    });
    describe('HelloWorldI ntent', function() {
      alexaTest.test([
        {
          request: alexaTest.getIntentRequest('HelloWorldIntent'),
          says: alexaTest.t('WELCOME'),
          reprompt: alexaTest.t('REPEAT'),
          repromptsNothing: false, shouldEndSession: false,
        },
      ]);
    });
  });
};

Output:

1)

 skill tests
       HelloWorldI ntent
         returns the correct responses:
     AssertionError: Request #1 (HelloWorldIntent): Possible Certification Problem: The response keeps the session open but does not contain a question mark.
      at Object._assert (/home/alexa/Templates/newAlexa/node_modules/alexa-skill-test-framework/index.js:845:9)
      at CallbackContext.assert (/home/alexa/Templates/newAlexa/node_modules/alexa-skill-test-framework/index.js:39:17)
      at CallbackContext._questionMarkCheck (/home/alexa/Templates/newAlexa/node_modules/alexa-skill-test-framework/index.js:64:8)
      at ctx.Promise.then.response (/home/alexa/Templates/newAlexa/node_modules/alexa-skill-test-framework/index.js:728:17)
      at <anonymous>

Could somebody help me please?

I am trying to keep session open after response so that user does not have to say "ask to " trigger next/new intent.
Currently response ends in period "." not in question mark "?". Is "?" necessary to end the response and keep session open.