EE596 Lab 1 -- Build Alexa Skills with the Alexa Skill Kit

Course Webpage: EE596 -- Conversational Artificial Intelligence

Task 1: Build your first Alexa Skill

• Please follow the walkthrough.

Task 2: Create an AWS Lambda function from scratch

Requirements:

• Node.js
• NPM

Steps:

• This time we will use the Alexa Skills Kit SDK for Node.js
  • We will still use the same Alexa Skill you created for Color Picker.
• Create another Alexa Lambda function. This time please choose "Author from scratch".
  • Name: ee596_spr2018_lab1
  • Runtime: Nodejs 6.10 (default)
  • Role: you can use the previously created role, or you can create a new role as you did in Step 2.10 in the walkthrough.
  • Click "Create function".
• Configuration
  • Add triggers under the Designer section: choose "Alexa Skill Kit".
  • Configure triggers: enter your skill ID (check "End Point" page in your Amazon Developer account) and click "Add".
  • Click "Save" at the top-right corner.
• On your laptop
  • Clone the repository

  $ git clone https://github.com/hao-fang/ee596_spr2018_lab1.git
  

  • Install Alexa SDK for Node.js. You will see a folder node_modules under the lambdaFunc folder.

  $ cd lambdaFunc
  $ npm install --save alexa-sdk
  

  • Edit the lambdaFunc/index.js file, replace the APP_ID in line 8 with your Skill ID. See Step 2.22 in the walkthrough.
  • Zip the lambdaFunc folder

  $ cd lambdaFunc
  $ zip -r ../lambdaFunc.zip *
  

• Upload the zip file to your Lambda Function.
  • Go back to your Lambda function webpage.
  • Go to the Function code section. (See Step 1.14 in the walkthrough).
  • If you cannot find Function code section, please click the center block ee596_spr2018_lab1 in the section Designer first.
  • Use the drop-down menu Code entry type.
  • Choose Upload a ZIP file.
  • Click Save at the top-right corner.
• Change your Alexa Skill's Endpoint ARN to this new Lambda Function.
  • See Step 1.13 and Step 2.21 in the walkthrough.
• Repeat Step 3 in the walkthrough.

Task 3: (Optional) Deploy the AWS Lambda function using command-line

Requirements:

• Python
• virtualenv

Systems:

• Linux
• MacOS
• Windows 10 WSL (Windows Subsystem for Linux)

Steps

• Create a virtualenv.

$ virtualenv ee596
$ source ee596/bin/activate

• Install AWS-CLI.

(ee596)
$ pip install awscli
$ aws --version
aws-cli/1.11.53 Python/2.7.6 Linux/3.13.0-105-generic botocore/1.5.16

• Setup your AWS credential ~/.aws/config.
  • Manage Access Keys for your AWS Account
  • Configuration and Credential Files
• Run sync.sh

$ ./sync.sh YOUR_LAMBDA_FUNC_NAME

Lab Checkoff

• Task 1:
  • Illustrate the sample dialog of Step 3.3 in the walkthrough.
  • In Step 3.4 of in the walkthrough, what is the detected intent when you say how are you?
  • Do you know when you are talking to your bot vs. Alexa?
• Task 2:
  • Show the configuration of your Alexa Skill Endpoint and the Lambda function ee596_spr2018_lab1.
  • Repeat the sample dialog as you have done in Task 1.
• Task 3 (Optional):
  • Show the output of your sync.sh command.

Lab Report

• Please attach the JSON object of your Interaction Model definition. See Step 2.19 in the walkthrough. Describe individual fields in the JSON object.
• Please attach the JSON input and output in Alexa Simulator in the Developer Console. See Step 3.4 in the walkthrough. Describe individual fields in these JSON objects.
• Find the functions in the Python codes lambda_function.py in Task 1 that correspond the following handlers in NodeJS codes lambdaFunc/index.js in Task 2.
  • NewSession
  • LaunchRequest
  • SessionEndedRequest
  • MyColorIsIntent
  • WhatsMyColorIntent
  • AMAZON.HelpIntent
  • AMAZON.CancelIntent
  • AMAZON.StopIntent
  • Unhandled
• Discuss other findings and issues.

References

• Alexa Skill Kit Docs
• Alexa Skill Kit SDK for NodeJs
• AWS Command Line Interface
• AWS Command Line Interface Docs
• AWS Security Credentials