DIY IoT lock - application

This package represents application layer for "smart lock" application. It contains all application logic that will be reused in packages that are responsible for UI.

Installing

npm i -s @diy-iot-lock/app

Usage

// 1. import library
import {DIYIoTlockApp} from "@diy-iot-lock/app";

// 2. create an instance of DIYIoTlockApp
const app = new DIYIoTlockApp();

// 3. configure Azure Cognitive Services Face API
const faceUrl = "https://centralus.api.cognitive.microsoft.com/";
const faceKey = "<your face api key>";
app.setFaceConfig(faceUrl, faceKey);

// 4.a. configure Azure Blob Storage with Key
const blobName = "<your azure storage account name>";
const blobKey  = "<your azure storage account key>";
app.setBlobConfig(blobName, blobKey);

// 4.b. configure Azure Blob Storage with SAS
const blobName = "<your azure storage account name>";
const blobSAS  = "<your azure storage account SAS>";
app.setBlobConfigSAS(blobName, blobSAS);

// 5. initialize app
await app.initializeAsync();

// 6. now you can start using the library

Logging

Builtin

By default all logs are disabled. If you want to enable them you can use builtin console logger like this:

// create an instance of builtin console logger
const log = new ConsoleLogService();

// create an instance of DIYIoTlockApp with builtin console logger 
const app = new DIYIoTlockApp(log);

Custom

For custom implementation of logger you can implement your own class based on ILogService interface:

// implement custom logger
class MyCustomLogger implements ILogService {
    public error(message: string): any {
    }

    public info(message: string): any {
    }
}

// create an instance of custom logger
const log = new MyCustomLogger();

// create an instance of DIYIoTlockApp with custom logger 
const app = new DIYIoTlockApp(log);

Training

await app.train.addPersonAsync(person: PersonModel): Promise<PersonModel>
await app.train.addPersonFaceAsync(personId: string, photo: Readable | Blob | ArrayBuffer, rectangle: RectangleModel): Promise<void>

Predicting

await app.predict.detectFacesAsync(photo: Readable | Blob | ArrayBuffer): Promise<DetectFaceModel[]>
await app.predict.identifyFacesAsync(photo: Readable | Blob | ArrayBuffer): Promise<IdentifyExtendedModel[]>

Known issues

Using setBlobConfigSAS and initializeAsync fails with AuthorizationFailure error

Description: In case you are configuring your storage account via SAS url you may experience an issue with initializing your app.
Reason: initializeAsync checks that container faces is present and public access level is set to either container or blob. For now there is no way to generate SAS url that will allow you to modify public access level of the container. So, if a container is present, but public access level is set to none, the initializeAsync method will fail.
How to fix: There are 2 options:

1. Create a container called faces with public access level set to either container or blob.
2. Delete a container called faces, it will be recreated with required access level during next call of initializeAsync.

Facing ContainerBeingDeleted error upon calling initializeAsync

Description: initializeAsync failed with error message ContainerBeingDeleted.
Reason: You had recently deleted a faces container. Azure Storage API may work with a small delay and container deletion may not be instantaneous. If you delete and try to recreate a container using the same name without delay, you may face this error.
How to fix: Just wait a bit (usually from 1 to 5 minutes max is enough) and everything will be fixed.