ask-form-sdk
TypeScript icon, indicating that this package has built-in type declarations

1.2.0 • Public • Published

Alexa Form SDK

Introduction

This Alexa Form SDK for Node.js provides support for creating skills for dynamic questions and answer forms.

Features

In order to support flexible and dynamic forms, this sample includes a number of advanced features.

  1. Conditional slot support (eg if answered 'yes' to slot one, then prompt for slot two).
  2. List disambiguation with support for custom prompts (eg I found 2 matches, did you mean "X")
  3. Decimal number support (eg "negative three point one four" as "-3.14") including units of measure (eg 5 percent).
  4. Numeric validate with custom prompts (eg require whole numbers, or min <= value <= max)
  5. Skipping field collection in a form (eg "skip" to move to next record, or "previous" to go back).
  6. Confirmation of form fields before delegating to save action.

Getting started

Using the template

Use the ASK CLI to Create a new skill with a template-url:

ask new --template-url https://github.com/aws-samples/ask-form-template.git

You are prompted to choose a method to host your skill's backend resources:

  • Alexa-hosted-skills (Do not choose this)
  • AWS with CloudFormation (Choose this to for AWS Lambda with persistence to Amazon DynamoDB)
  • AWS Lambda (Choose this if you want to deploy to AWS Lambda without AWS CloudFormation)

You will then be warned you about downloading an unofficial template.

Input Yes to continue provide a skill name eg. my-survey.

? Choose a method to host your skill's backend resources:  AWS with CloudFormation
[Warn]: CLI is about to download the skill template from unofficial template https://github.com/aws-samples/ask-form-template.git. Please make sure you understand the source code to best protect yourself from malicious usage.
? Would you like to continue download the skill template?  Yes
? Please type in your skill name: my-survey
? Please type in your folder name for the skill project (alphanumeric):  my-survey

Change into your project directory, and list the contents. You will see the template includes a lambda folder with a samples forms folder used to generate skill interaction model, as well as infrastructure to deploy cloud formation if required.

$ cd my-survey
$ tree
.
├── LICENSE.txt
├── README.md
├── ask-resources.json
├── infrastructure
│   └── cfn-deployer
│       └── skill-stack.yaml
├── lambda
│   ├── forms
│   │   └── index.js
│   ├── index.js
│   └── package.json
└── skill-package
    ├── assets
    │   ├── en-US_largeIcon.png
    │   └── en-US_smallIcon.png
    ├── interactionModels
    │   └── custom
    │       └── en-US.json
    └── skill.json

You will then be able to deploy this skill into your AWS account. By default this will be deployed to the us-east-1 region.

  • Update the .ask-resources.json to specify a different region.
$ ask deploy

The default invocation is "my survey". Once your skill is deployed, you can test your skill in the Alexa developer console.

Following is an example conversation.

ask-form-template

Troubleshooting

  • If you are getting the following error during ask deploy, then you already have the dynamo db table created with the name my-survey :
[Error]: The CloudFormation deploy failed for Alexa region "default": AlexaPersistentAttributesTable[AWS::DynamoDB::Table]  DELETE_COMPLETE(Resource deleted. For detailed reasoning, please check the "events" tab in the AWS console for CloudFormation stackId <STACK-ID>)

To resolve this issue you can fo to my-survey/infrastructure/cfn-deployer/skill-stack.yaml and change the name of the PersistentAttributeTableName default value from my-survey.

  • If you are getting the following CLI error during deployment, it might be due to the ask profile settings ;
[Error]: CliError: Skill package src is not found in ask-resources.json.

Go to my-survey/ask-resources.json and check the profile to see if the profile used for deployment is present in this. If not add the profile use for deployment to this file.

Adding to an existing skill.

If you already have an existing skill, you can extend it with the following three steps.

  1. Install library

Install the package as a dev dependency with:

$ npm install ask-form-sdk
  1. Update lambda

Add request handlers for DecimalIntentHandler and FormListHandler.

The FormListHandler class also has methods:

  • handlerNext - Move to next item in the form
  • handlePrevious - Move to previous item in the form
  • handleUserEvent - Handle APL touch event
  • handleFormReview - Display review screen

These can be called from other intents eg AMAZON.NextIntent as shown below.

// Import the decimal and form list handler
const { DecimalIntentHandler, FormListHandler } = require('ask-form-sdk');

// Import your form definitions
const { forms } = require('./forms');

// Add an handler to optionally load user from session or API
const ApiUserHandler = {
    async getUser(handlerInput) {
        const deviceId = handlerInput.requestEnvelope.context.System.device.deviceId;
        if (handlerInput.serviceClientFactory) {
            const upsServiceClient = handlerInput.serviceClientFactory.getUpsServiceClient();
            return {
                username: await upsServiceClient.getProfileName(),
                timeZone: await upsServiceClient.getSystemTimeZone(deviceId),
            };
        }
    },
    canHandle() {
        // Optionally add logical to check if there is a valid user for form
        return true;
    }
};

// Create the forms handler passing the api user handler
const formsHandler = new FormListHandler(forms, ApiUserHandler);

// Welcome intent
const LaunchRequestHandler = {
    canHandle(handlerInput) {
        return handlerInput.requestEnvelope.request.type === 'LaunchRequest';
    },
    handle(handlerInput) {
        const speakOutput = 'Welcome to sample form, to get started say "start form".';
        return handlerInput.responseBuilder
            .speak(speakOutput)
            .reprompt(speakOutput)
            .getResponse();
    }
};

// Next handler calls into forms handler
class NextHandler {
    constructor(formsHandler) {
        this.formsHandler = formsHandler;
    }
    canHandle(handlerInput) {
        return handlerInput.requestEnvelope.request.type === 'IntentRequest'
            && handlerInput.requestEnvelope.request.intent.name === 'AMAZON.NextIntent'
            && this.formsHandler.isCurrent(handlerInput);
    }
    handle(handlerInput) {
        return this.formsHandler.handleNext(handlerInput)
    }
};

// Add request handlers for (default intent handlers omitted for clarify)
const requestHandlers = Alexa.SkillBuilders.custom()
    .addRequestHandlers(
        // LaunchRequestHandler,
        new DecimalIntentHandler(formsHandler),
        formsHandler,
        new NextHandler(formsHandler),
        SaveFormIntentHandler, // Delete to SaveFormIntentHandler
        ErrorHandler,
    );

// Optionally add other API or persistence handlers, then return lambda
exports.handler = requestHandlers.lambda();
  1. Add forms

When creating the FormListHandler class you pass a list of forms, these include properties and optional validation as illustrated with this simple two slot form below.

const { WholeNumberValidation } = require('ask-form-sdk');

const temperature = {
    name: 'TemperatureIntent',
    title: 'Temperature Form',
    prompt: (slotNumber) => slotNumber === 0 ? 'Welcome to the form. ' : '',
    slots: [
        {
            index: 1,
            name: 'slot_three',
            type: 'AMAZON.NUMBER',
            prompt: 'What is the temperature (with units)?',
            reprompt: 'What temperature?',
            required: true,
            validation: [
                new WholeNumberValidation('Please specify temperature as a whole number.'),
            ],
        },
        {
            index: 2,
            name: 'slot_three_units',
            type: 'DECIMAL_UNITS', // Can we just use the same general UNITS type?
            prompt: 'For what units?',
            reprompt: 'Is that in "Celsius" or "Fahrenheit"?',
            required: false,
            options: {
                'C': [
                    'Celcius',
                    'degree celcius',
                    'degrees celcius',
                    'centigrade',
                ],
                'F': [
                    'Fahrenheit',
                    'degree fahrenheit',
                    'degrees fahrenheit',
                ],
            },
            conditional: [{
                name: 'slot_three',
                empty: false, // Conditional if the slot_three is not empty
            }],
        },
    ],
    confirmation: () => {
        return `Okay, would you like to save the temperature?`;
    },
    delegate: 'SaveFormIntent',
    samples: [
        'record temperature of {slot_three} {slot_three_units}',
        'record temperature of {slot_three}',
        'record temperature',
    ],
};

export const forms = [temperature];

Customising the Form and Slots definition

The form has a number of slots which you can configure.

  • name: unique name for the form.
  • title: unique title displayed on the card.
  • slots: the list of properties for each slot (see below)
  • prompt(slotNumber): a function to return the prompt to display when beginning the form.
  • confirmation(slots, slotValues): a function to return the prompt when completing the form.
  • delegate: the intent name to delegate to after confirmation eg "StartFormIntent".
  • samples: the list of samples heard to start form eg "start form".
  • reviewPrompt(aplSupported, slotValues): optional function to return a prompt after the review page is displayed.
  • aplProperties:
    • previousDisplayText: Optional text to prefix for previous item in list eg "Previous"
    • nextDisplayText: Optional text to prefix for next item in list eg "Next"
    • endOfFormText: Optional text to display for the end of the list eg "End of form".

Each slot has a series of properties.

  • name: unique name for the slot.
  • index: optional number for the slot.
  • type: the Alexa Slot Type
  • options: for custom slots, you can provide a list of options.
  • prompt: The prompt to ask user.
  • reprompt: The re-prompt message for user.
  • required: Specify boolean (true/false) if the field is required.
  • conditional: Optional list of conditional settings for this slot
    • name: slot name to match condition
    • empty: Optional boolean (true/false) to make on an empty or non-empty field.
    • value: Optional value to match for slot.

You can customize the form and add new fields and delete/modify existing fields as required by the skill. This can be done by editing the index.js file under lambda/forms folder.

Code Generating the model

Once you have made changes or additions to the form definition, you can code generate the model file(en-US). In order to do this, you need to install the ask-form-sdk. The following will install the SDKs globally for code generation, you can also access run locally via node_modules/.bin/ask-form-codegen command.

$ cd my-survey
$ npm install -g ask-form-sdk

The code generation tool can take in the following options:

  • -f : the path to forms
  • -s : the path to the source model json
  • -t : the path to the target model json
  • -i : invocation name for skill. Provide this argument when the model is created from scratch.

There are 2 options for code generation.

  • Have the new form changes applied on top of the existing model file. This can be done by providing the existing model file as the source file:
$ ask-form-codegen -f lambda/forms/ -t skill-package/interactionModels/custom/en-US.json -s skill-package/interactionModels/custom/en-US.json
  • Have the model code generated from scratch. To generate the model file form scratch, provide the invocation name and do not provide any source model files :
$ ask-form-codegen -f lambda/forms/ -t skill-package/interactionModels/custom/en-US.json -i 'my survey'

This will create a model file with the fields in the form along with few other intents required by the skill.

Tests

The examples folder defines a series of typescript test specifications using the ask-sdk-test framework that show cases some of the various features mentioned above.

Authors

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE.txt file.

Dependents (0)

Package Sidebar

Install

npm i ask-form-sdk

Weekly Downloads

2

Version

1.2.0

License

MIT

Unpacked Size

375 kB

Total Files

67

Last publish

Collaborators

  • brightsparc