Repeat After Me

Tinkering around with and learning the Alexa Skills Kit (ASK) so I can eventually take control of Alexa and train her to do everything for me.

Setup

Start with this link to get to the tutorial.

The main components in making a skill:

Voice User Interface (VUI): define which code should be exected when specific commands are uttered. Seems to be mainly the Interaction Model tab that matters here.
Lambda Functions: the actual code for the skill.

Fact Skill Tutorial

A condensed version of this tutorial for me to review when I forget stuff.

Step 1: Developer Portal (VUI)

developer.amazon.com. Go to/make your skill and you'll be at the screen with the following tabs on the left:

Skill Information
Interaction Model. Here's where you add new Intents.
Configuration
Test
Publishing Information
Privacy & Compliance

Step 2: Lambda Functions with AWS

Quick facts to remember:

Creating a new function: Click "Create a Lambda function". Choose a blueprint/template file (e.g. alexa-nodejs-template). Configure your trigger -- choose "Alexa Skills Kit".
Copy the ARN from the top right corner of the screen. You'll need it in the next step.

Links:

Step 3: Connect VUI to Lambda Function [Link]

Go to the "Configuration" tab of the developer console. Select "AWS Lambda ARN" option for your endpoint. Paste your Lambda's ARN into the textbox (should have from the end of step 2).

Step 4: Testing your Alexa Skill [Link]

The "Test" tab of the developer console. Here you'll find:

Voice simulator.
Service Simulator.

Step 5: Customization [Link]

The scripting part.

Interaction Model / Skill Builder

The fancy-looking UI where you add intents and stuff.

Intents

You can add sample utterances, which are actually training data for Alexa.

Links:

Youtube Tutorial -- Using intents, built-ins, and utterances

My Questions and Answers

How do I know which functions get which parameters?

You can actually define these via the "intent" JSON, with things like slots. A simple example from here:

// ... within "handlers" object.
// Corresponds to entry in the "intents" (json):
// {
//    "intent": "GetNewFactAboutIntent",
//    "slots": [
//          {
//            "name": "number",
//            "type": "AMAZON.NUMBER"
//          }
//        ]
//  },
'GetNewFactAboutIntent': function () {
        var theNumber = this.event.request.intent.slots.number.value;
        var speechOutput = "You said the number " + theNumber;
        this.emit(':tellWithCard', speechOutput, SKILL_NAME, speechOutput)
    },

How does Alexa handle multiple intent functions?

It looks like intents are activated by the user saying their name? Don't understand why only the "main" intent is emitted from LaunchRequest.
One way is assigning the samples key (within the intent) to an array of allowed utterance samples.

What SSML Do I Care About?

speak. Root element of an SSML doc.

<speak>
    This is what Alexa sounds like without any SSML.
</speak>

say-as. Describes how the text should be interpreted.

<say-as interpret-as='spell-out'>hello</say-as>

Useful Resources

Tutorials:

mckinziebrandon/RepeatAfterMe