Skip to content

An Alexa Skill web API and Library developed live on-stream. Publically available as a Nuget package.

License

Notifications You must be signed in to change notification settings

essenbee2/Essenbee.Alexa

 
 

Repository files navigation

Essenbee.Alexa and Essenbee.Alexa.Lib

Build status Tests Release Build status

Twitter

We are building an Alexa Skill live on-stream at Codebase Alpha on Twitch. The skill, called DevStreams, is the voice interface for the DevStreams community website that is being developed on the DevChatter Twitch stream by Brendan Enrick and contributors. A fork of that repo can be found in this Github. Integration has been achieved through the introduction of a GraphQL endpoint into the DevStreams solution (also developed on-stream at Codebase Alpha). The Essenbee.Alexa code uses this endpoint to query DevStreams and thus provide users with the information about live coding streams that they have requested.

The primary reusable component created in this project is the Essenbee.Alexa.Lib NuGet package. Please see below for installation and usage instructions.

YouTube

The main phase of development for this code is encompassed by Episodes 4 - 11 of Codebase Alpha. The videos for these episodes are archived on YouTube here.

Installation

nuget Downloads

Install the Essenbee.Alexa.Lib NET Standard 2.0 library into your own ASP.NET Core (2.2+) Web App or Web API via Nuget with the command:

dotnet add package Essenbee.Alexa.Lib

Current Features

The library contains the following features at this stage:

  • Middleware and methods to satisfy Amazon's security requirements for Skills, see here
  • AlexaRequest and AlexaResponse classes for speech, cards and dialogs, but not yet for AudioPlayer requests at this time
  • A fluent ResponseBuilder that makes creating responses easy
  • A strongly-typed HTTP client (AlexaClient) that you can use to access Device Settings or User Address details (with the user's permission)

Further features will be added over time, with the intent being to provide library support for all of the features that Alexa skills can exploit. However, that takes time, especially doing the work live on Twitch, so please check back here often to see what is new.

Using the Library

Install the latest NuGet package into your project. In the Startup.cs class, add the following code to the Configure method:

app.UseWhen(context => context.Request.Path.StartsWithSegments("/api/alexa"), (appBuilder) =>
{
    appBuilder.UseAlexaRequestValidation();
});

To use the strongly-typed HTTP client, add this line to your ConfigureServices method:

services.AddHttpClient<IAlexaClient, AlexaClient>();

Make sure that you store your Alexa Skill Application ID (securely) in your Configuration. In the example code, it is held in Configuration["SkillId"]. This is needed for the AlexaRequestShouldProcessRequest()` static method call shown below. It is this method that, alongside the custom Middleware, ensures that your Skill staisfies Amazon's security requirements.

Create an AlexaController using the example below as a starting point:

[ApiController]
[Produces("application/json")]
public class AlexaController : ControllerBase
{
    private IConfiguration _config;
    private ILogger<AlexaController> _logger;
    private readonly IAlexaClient _client;

    public AlexaController(IConfiguration config, ILogger<AlexaController> logger, IAlexaClient client)
    {
        _config = config;
        _logger = logger;
        _client = client;
    }

    [HttpPost]
    [ProducesResponseType(200, Type = typeof(AlexaResponse))]
    [ProducesResponseType(400)]
    [Route("api/alexa/your-endpoint")]
    public async Task<ActionResult<AlexaResponse>> MySkill ([FromBody] AlexaRequest alexaRequest )
    {
        if (!AlexaRequest.ShouldProcessRequest(_config["SkillId"], alexaRequest))
        {
            _logger.LogError("Bad Request - application id did not match or timestamp tolerance exceeded!");
            return BadRequest();
        }
        
        // Your skill's code here
    }        
}

You use the ResponseBuilder like this:

Simple Speech Response

var response = new ResponseBuilder()
                .Say("To use this skill, ask me about the schedule of your favourite stream.")
                .Build(); 

Speech Response with Card

var response = new ResponseBuilder()
                .Say("Codebase Alpha is streaming now")
                .WriteSimpleCard("Streaming Now!", "Codebase Alpha")
                .Build();

SSML Speech Response

var ssml = @"<speak>Welcome to my skill</speak>";
var response = new ResponseBuilder()
                .SayWithSsml(ssml)
                .Build(); 

Ask Something, Expecting a User to Respond (with Reprompt or Not)

var response = new ResponseBuilder()
                .Ask("Try asking me who is streaming now",
                     "You can ask me who is streaming now")
                .Build();

Delegate Dialog to Alexa Manually

var response = new ResponseBuilder()
                .DelegateDialog()
                .Build();

Discord

If you have any questions, suggestions or bug reports relating to the Essenbee.Alexa.Lib library, head over to my Discord here. Why not join me for some live coding on Twitch? We work on a variety of projects, not just Alexa skills.

About

An Alexa Skill web API and Library developed live on-stream. Publically available as a Nuget package.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C# 95.9%
  • HTML 4.1%