IoT Event Analytics

(c) Bosch.IO GmbH under MPL-2.0 licence

Introduction

IoT Event Analytics, is a complex and freely scalable event processing and agent network platform. The platform is basically build around so called Talents which form a distributed network. Each talent is a unit of work and can interact with other Talents via events. The platform itself orchestrates the flow of events and offers additional services, like an in-memory digital twin, calculation of statistical data based on given event histories and complex rule evaluation of when to inform a Talent, that something happened, that the Talent "is interested in". Thus, a talent is always an event consumer and may also be an event producer. The event-driven business logic can be implemented using one of our SDKs for JavaScript, Python or C++. There are more features which can be explored using various examples for the different SDKs.

Prerequisites

NodeJS

• Install Node.js >=12.13.0
• Install all dependencies using yarn package manager
  • If yarn is not available on your system, install it using npm install -g yarn
    • Please make sure, you use the latest version of yarn
  • Run yarn in the project root

Python

• If using Anaconda to manage your Python environments, create a new one
  • conda create --name <name-of-choice> python=3.7
    • Python needs to be at at version >=3.6.8
    • Pick a name of choice
  • conda activate <name-of-choice>
• Install all necessary packages using pip install -r requirements.dev.txt in the project root

SDKs

• Download or Build
  • Run yarn sdk.get to download the most recent prebuilt SDK releases from GitHub without needing to build them from scratch. You can continue with the installation. You will find the downloaded artifacts in their respective folder ./src/sdk/(javascript,python,vscode)/lib
  • JS SDK (mandatory)
    • Run yarn sdk.build in the project root
  • Python SDK (optional)
    • See README.md
  • VSCode Extension (optional, mandatory for tldr section)
    • Open ./src/sdk/vscode in your terminal
    • Run yarn package
• Manual Installation
  • JS SDK
    • Run yarn add file:*.tgz
  • Python SDK
    • See README.md
  • VSCode Extension
    • Directly install the *.vsix file

tldr - get it running - asap

• IMPORTANT The steps in the prerequisites section above need to be done in order to quickly getting started
• Install the latest IoT Event Analytics VSCode extension from src/sdk/vscode/lib/*.vsix
• Use Ctrl + Shift + P to bring up the command palette and select Bosch IoT Event Analytics: Create new JavaScript Talent project in an empty folder.
  • Follow the instructions to create a "ready to use" Talent project
• Bring the command palette up again and run Bosch IoT Event Analytics: Start platform using docker-compose.
  Select the generated .env file in the folder, which contains your newly created talent project.
• Run your talent using NodeJS by simple run node index.js in your newly created talent project.
• Send events to the platform by using the command Bosch IoT Event Analytics: Publish an MQTT Message
• Done

Project structure

src
| |- core      The platform itself
| |- sdk       The platform software development kits
| |- adapter   Integrations of other systems
|- resources   JSON schemas for object validations
L- docs        Documentation

Build and Run

The recommended way of quickstarting the platform is to use docker-compose where available. If not installed, all components can be built and run manually using the given build instructions under ./docker/README.md and ./docker/*/README.md For further information how to start the platform using docker-compose see here

Run an example/your own Talent on your machine (within the project)

• Simply create your first python or NodeJS talent following the examples given in src/sdk/(javascript|python|cpp)/examples
  
• Start your talent by connecting it to mqtt://localhost:1883
  
• There are examples, which start a platform instance by themselves. They only need an MQTT Broker running. To achieve this, simply run from within the ./docker-compose directory
  docker-compose -f docker-compose.mosquitto.yml up

Run an example/your own Talent on your machine (Using the JavaScript SDK)

• Create a new directory for your Talent

• Copy src/sdk/javascript/lib/boschio.iotea-<version>.tgz into this directory

• Run npm init to initialize your NodeJS project

• Install the SDK by typing yarn add boschio.iotea-<version>.tgz

  • If you have problems with the installation behind a proxy try the following:

    npm config set proxy ${HTTP_PROXY}
    npm config set https-proxy ${HTTPS_PROXY}
    npm config set strict-ssl false -g
    npm config set maxsockets 5 -g
    

• Create an index.js file with the following contents

  const iotea = require('boschio.iotea');
  
  const Logger = iotea.util.Logger;
  process.env.LOG_LEVEL = Logger.ENV_LOG_LEVEL.INFO;
  
  const {
      MqttProtocolAdapter
  } = iotea.util;
  
  const {
      Talent,
      OpConstraint,
      Rule,
      AndRules,
      ProtocolGateway,
      TalentInput
  } = iotea;
  
  const {
      VALUE_TYPE_RAW
  } = iotea.constants;
  
  class MyTalent extends Talent {
      constructor(protocolGatewayConfig) {
          super('some-unique-talent-id', protocolGatewayConfig);
      }
  
      getRules() {
          return new AndRules([
              new Rule(new OpConstraint('anyfeature', OpConstraint.OPS.ISSET,  null, 'anytype', VALUE_TYPE_RAW))
          ]);
      }
  
      async onEvent(ev, evtctx) {
          this.logger.info(`${TalentInput.getRawValue(ev)}`, evtctx);
      }
  }
  
  // Update mqttAdapterConfig.config.brokerUrl, if you specified a different one in your configuration !!!
  const mqttAdapterConfig = MqttProtocolAdapter.createDefaultConfiguration();
  
  new MyTalent(
      ProtocolGateway.createDefaultConfiguration([ mqttAdapterConfig ])
  ).start();

• Run node index.js to start the test talent.

Dependencies

• List all NodeJS dependencies on stdout yarn licenses list
• List all Python dependencies on stdout pip-licenses --from=mixed
  • To have a correct list of packages containing all subpackages needed by this project, make sure that you used a "fresh" environment of conda, where you install the dependencies using pip install -r requirements.dev.txt. Otherwise pip will list all packages you have installed in your environment.

Yarn Scripts

• The jsdoc of iot-event-analytics APIs will be generated in docs/jsdoc folder after running yarn or yarn docs.generate
• To run the unit tests, use yarn test