cbonoz / amznfire19

SideCook - A Chef sidekick on your FireTV while you cook.

SideCook

SideCook is your personal cooking assistant in the kitchen powered from your Amazon Alexa.

Inspiration

Too many of us watch Food Network on television, only to never actually make any of the recipes we watch.

There's a million cooking apps on the market today, but most of them are just recipe dictionaries, and don't interactively take you through the process of preparing the dish step by step. Rather they just throw a list of ingredients followed by all the steps in a single block.

We wanted to combine video and a more interactive/step-by-step process for cooking into a multimodal app. Simply find the recipe or food area you want to make, and we'll break it down into visual steps that you can follow along with on your television.

What it does

1. Recommends the most popular/simple recipes based on the search criteria.
2. Automatically breaks down the recipe into a step by step formula that can easily be stepped through on a visual interface.
3. Render a list of ingredients for the recipe that can be accessed at any time while proceeding through the recipe.

How we built it

• Alexa SDK
• APL (Alexa Presentation Language)
• TheMealDB API

Challenges we ran into

Integrating the Amazon APL library for voice. Learning how to integrate state management into the app to allow the user to navigate between different recipes and steps.

Accomplishments that we're proud of

It works.

What we learned

How to decompose a structured API response into step by step screens onto a visual display. Parse through descriptions and render descriptions and images.

What's next for SideCook.

Put SideCook's model and UI under more strenuous user testing to figure out the most user friendly way to step through recipes on the television while you're actually cooking them. Render additional content (such as videos) based on techniques being discussed in the current recipe step. For example, showing a video on proper dicing technique onto the display when the current step involves dicing.

Amazon Resources

• https://developer.amazon.com/docs/fire-tv/getting-started-with-web-apps.html#using-web-app-tester-and-devtools
• https://developer.amazon.com/docs/app-submission/submitting-apps-to-amazon-appstore.html

Packaging

  zip -r -X ../src.zip *

Testing

From the /text folder, run the intent input json test files.

  node index.js "filename"

where filename is the json file input to execute on. The output will be the rendered response from the function.

Other Resources

• https://themealdb.com/
• https://developer.amazon.com/blogs/alexa/post/2af6851b-0216-4e82-9aba-6fa2aec755d5/how-to-get-started-with-the-new-alexa-presentation-language-to-build-multimodal-alexa-skills
• https://www.talkingtocomputers.com/apl-alexa-presentation-language
• https://www.themealdb.com/

amznfire19's Issues

how to deploy?

Hi,

Random question. How do you deploy it? I am going through APL right now and the only way I know how to deploy skills, is to first create a new skill using:

ask new -n sidecook --lambda-name sidecook

then choose a random sample template like PetMatch, then:

ask deploy

then swap out the code in lambda/custom/ and models/ with your code, and add APL to the interfaces property to the skill.json then re-deploy and it works. It works, but is there an easier way?

I tested your skills in the simulator, a small bug (maybe u need JSON.stringify to fix?):