Home Automation demo

This repo contains a sample project based on Kogito and Camel-K.

Folders content

• camel-k: this folder contains a set of Camel-K routes used to integrate with different services
• knownImages: set of images used as knowledge base. All known member of the family should provide one or two pictures
• sample-kogito: Kogito application that implement process orchestration and decision logic
• testImages: folder of pictures used to mock a camera service. Camel-K take-picture route emulate a camera just picking randomly one of the file from that folder

How to run

This project can run with OpenShift 4 with or without Serverless (knative). It should work in the same way on K8s but it has not been tested

Deploy face recognition service

In this demo we use a face recognition service based on deep learning (dlib library) wrapped in a Python web service to have a REST API.

You can find the code in this repo.

It is enough to build it

docker build -t tag_name:1.0 .

Push it to your registry

docker push tag_name:1.0

Deploy it with a YAML like this (oc apply -f file.yaml)

apiVersion: apps/v1
kind: Deployment
metadata:
  name: face-recognition
  namespace: cloud-sweet-cloud
spec:
  selector:
    matchLabels:
      app: face-recognition
  replicas: 1
  template:
    metadata:
      labels:
        app: face-recognition
    spec:
      containers:
        - name: face-recognition
          image: yourrepo/your-image
          ports:
            - containerPort: 8080

Finally expose it

oc expose deployment face-recognition

oc expose service face-recognition

Deploy Camel-K services

Use Camel-K installation guide to install kamel CLI and to deploy the camel-k-operator to your cluster.

After that you should:

• create a (telegram bot)[https://core.telegram.org/bots]
• define a secret named telegram containing:

telegram.token   = ... # the tocken
telegram.chat-id = ... # the chat id

camel.component.telegram.authorization-token = {{telegram.token}}

• deploy the integrations

kamel kit create notifier-q --image quay.io/lburgazzoli/home-automation-poc-camel-k-notifier
kamel run camel-k/notifier.yaml --kit notifier-q --trait jvm.enabled=false --secret telegram
kamel run camel-k/notifier-playlist.yaml --kit notifier-q --trait jvm.enabled=false --secret telegram
kamel run camel-k/sensor-gw.groovy
kamel run camel-k/take-picture.groovy

Deploy sample-kogito

This module is a Quarkus application and you need to decide if you want to run it in JVM mode or in native mode.

Just compile it with mvn package for JVM mode or with mvn package -Pnative for native mode (you need GraalVM 20 for this).

After than build your container image with the Dockerfile you need:

• Dockerfile.jvm for JVM mode
• Dockerfile.native for native mode

Then push your image to your registry and finally deploy it on OpenShift Serverless with this YAML file

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: sample-kogito
  namespace: cloud-sweet-cloud
spec:
  template:
    spec:
      containers:
        - name: sample-kogito
          image: quay.io/danielezonca/home-automation-poc:1.0
          ports:
            - containerPort: 8080
          env:
            - name: CAMERA_HOST
              value: "take-picture"
            - name: CAMERA_PORT
              value: "80"
            - name: CAMERA_SSL
              value: "false"
            - name: CAMERA_ENDPOINT
              value: "/"
            - name: PLAYER_HOST
              value: "notifier-playlist"
            - name: PLAYER_PORT
              value: "80"
            - name: PLAYER_SSL
              value: "false"
            - name: PLAYER_ENDPOINT
              value: "/"
            - name: RECOGNITION_SERVICE_HOST
              value: "face-recognition"
            - name: RECOGNITION_SERVICE_PORT
              value: "80"
            - name: RECOGNITION_SERVICE_SSL
              value: "false"
            - name: RECOGNITION_SERVICE_ENDPOINT
              value: "/"
            - name: TELEGRAM_HOST
              value: "notifier"
            - name: TELEGRAM_PORT
              value: "80"
            - name: TELEGRAM_SSL
              value: "false"
            - name: TELEGRAM_ENDPOINT
              value: "/"

If you want to deploy without knative just change apiVersion to apps/v1 and kind to Deployment.

NOTE: service and route are created automatically using knative while you still need to create them if you are using Deployment