Shifu advanced features: custom deviceShifu data handler

Custom deviceShifu data handler

Shifu allows users to customize deviceShifu to make digital twins more efficient.

default

By default, data from a device to deviceShifu will be provided in its original format.

custom processing

After adding custom logic, deviceShifu can process the data to make it more suitable for the application:

1. Convert the data into the format required by the user. 2. Filter out unwanted data. 3. Perform dynamic calculation and data analysis.

Add custom deviceShifu

Before we start running Shifu , there are three things to do first.

1. customized_handlers.pyWrite data processing logic in , the general structure of the processing method is as follows:

#使用指令/API名称作为方法名称
def humidity(raw_data): 
    new_data = process(raw_data)
    return new_data

2. customized_handlers.pyMove pkg/deviceshifu/pythoncustomizedhandlersto .

3. Generate deviceShifu .

Example: Moisture Detector

The example device used here is in examples/deviceshifu/customized/humidity_detectorthe directory.

This device is a virtual humidity detector that will consume humidity and temperature data in HTTPthe provided JSONformat, command/API only humidity.

0. Create a virtual device for the humidity detector

The first step is to use the docker image built humidity-detector.gowith it examples/deviceshifu/customized/humidity_detector's Dockerfile humidity-detector.

1. Add a custom data handler

In examples/deviceshifu/customized/humidity_detector/pythoncustomizedhandlers, there is a customized_hanlders.pyfile, copy that file into pkg/deviceshifu/pythoncustomizedhandlers.

2. Edit the Docker file of deviceShifu

will be examples/deviceshifu/customized/humidity_detector/sample_deviceshifu_dockerfiles/Dockerfile.deviceshifuHTTPcopied to dockerfiles.

3. Add the mapping of the data handler

In examples/deviceshifu/customized/humidity_detector/configuration中the deviceshifu-humidity-detector-configmap.yamlfile, map your device instructions with the funcName of the data handler.

(For example: instructions is '/123', funcName is 'humidity')

Then you need customInstructionsPythonto set up 123: humidityand set up the equipment instructions under instructions.instructionsand telemetries.telemetries.device_health.properties.instructionunder Instructions

As follows:

data:
  customInstructionsPython: |
    123: humidity 
    #123是instructions，humidity是处理程序funcName
  instructions: |
    instructions:
      123:
  telemetries: |
    telemetries:
      device_health:
        properties:
          instruction: 
            123

4. Create a docker image of deviceShifu

Build a new deviceShifu image to add custom data handlers.

In shifuthe root directory of , run the following command to build the docker image of deviceShifu .

make buildx-build-imag-deviceshifu-http-http

5. Roaring Shifu

This section is exactly the same as in the quickstart demo .

After Shifu is running, we need to import the newly created humidity detector virtual device into kindthe cluster.

kind load docker-image humidity-detector:v0.0.1

6. Check the processed data

Raw data from this virtual device should be handled by custom handlers defined customized_hanlders.pyin .

In the nginx shell, we should be able to curl the API humidityand check if the result was processed, the result is as follows:

[
  {
  "unit": "℃", 
  "code": "20990922009", 
  "exception": "temperature is too high", 
  "name": "atmosphere temperature", 
  "val": "37"
  }, 
  {
    "unit": "%RH", 
    "code": "20990922009", 
    "exception": "humidity is too high", 
    "name": "atmosphere humidity", 
    "val": "88"
  }
]

This article is published by Boundless Authorization