Lite API: Gateway

Parent Document: ANT APIs

Gateway API comprises of M/L (machine learning) gateway offloading functions, sensor directory functions and sensor database functions.

var gatewayAPI = require('ant').gateway;

M/L Gateway Offloading

• M/L gateway offloading functions are strongly dependent on MLAPI.createMLFragmentElement() of Machine Learning APIs.

GatewayAPI.createImgClsImageNetElement()

• MLFragmentElement GatewayAPI.createImgClsImageNet(String modelPath, Integer numFragments, String targetUri)
  • Create an ImageNet-based image classification ML fragment element. The ML fragment element can be offloaded to the gateway device. When the gateway's load increases, the gateway also can offload some ML fragments to the user devices.
  • Argument 1: String modelPath
    • Target image classification model's absolute path. modelPath can be retrieved by MLAPI.downloadModel().
  • Argument 2: Integer numFragments
    • The number of ML model fragments.
  • Argument 3: String targetUri
    • Target user device's URI.
  • Return: MLFragmentElement
    • The resulting ML fragment element. This element can be added to the pipeline by Stream API.

Sensor Directory and Sensor Database

GatewayAPI.connect()

• void GatewayAPI.connect(void)

GatewayAPI.getSensor()

• void GatewayAPI.getSensor(String uri, Function handler)

GatewayAPI.findSensor()

• void GatewayAPI.findSensor(String type, Function handler)

GatewayAPI.getSensorData()

• void GatewayAPI.getSensorData(String uri, Function handler)