Predix Machine and how to configure them

In September, waiting for a contract proposal from the company where I currently work, I wrote a tool for the (more or less) automatic configuration of Predix© Machine. Predix©¹ is the platform created by GE for Industry 4.0, powered by CloudFoundry², to securely connect machines, data, and analytics to improve operational efficiency, help you develop, deploy, and operate industrial apps at the edge and in the cloud. As part of my internship I worked with the predix machine and I shared my work to the GE platform. GE mantains an open repository with predix tool

I wrote an Angular 4 + Python based Predix© configurator tool (cli) to help me and the others who are using predix machine to transfer data from respective factories, managing an edge connectivity stack, from machines in your factory to Predix© microservices in your cloud space: the idea is to provide a web application tool to create a JSON seed, and use the seed to create configuration files. In this article I will talk about my tool - available for free here. If you’re looking for a boilerplate for multistage form, this could be a good project to start from!

Factory stack

The tool is intended for a factory infrastructure like the one below:

Use of a machine-protocol gateway able to handle several industrial communication protocols used by different machines. For our purpose we used IGS by GE Automation;
Use of a Predix Machine (you can find pre-configured version of Predix Machine here and other useful starting template for your application in Predix Dev’s Github page);
Available cloud space in Predix Platform;

How it works

The tool is intended for both Predix dev / common users. I prepared two different step-by-step guidelines:

Old-school way: prepare your JSON schema manually => run the configuration tool => enjoy.
Third Millennium way: prepare your JSON schema through the use of a stunning serverless web app XD => run the configuration tool => enjoy.

How it works: old-school way

The only script you need is the python script under the folder cli/ named predix-configurator.py. The script has only one dependency to pretty print xml to file. You can run it from your command line simply following this step:

git clone https://github.com/made2591/predix-configurator
cd predix-configurator
cd cli
#[optional create venv] virtualenv .venv & source .venv/bin/activate
pip install requirements.txt
python predix-configurator.py

When predix-configurator starts, it looks for a configuration file under the folder schema called your-factory.json (you can change the name of the file and update the const value in the header of the script with the new name / path as well).

A your-factory.json schema is already available in schema folder: for obvious reasons, I can’t provide more details using real parameters instances XD.

JSON Schema details

GLOBAL_CONFIG

The GLOBAL_CONFIG key in the root of JSON specify the output path of python-configurator. When you run the predix-configurator.py, generated configuration files will be created in the GLOBAL_CONFIG.OUTPUT_DIR folder.

{
	"GLOBAL_CONFIG" : {

		"OUTPUT_DIR" : "./config/"

	},
}

GLOBAL_IGS_CONFIG

The GLOBAL_IGS_CONFIG key in the root of JSON specify IGS parameters. The most important parameter is SERVER_URI, that specify opcua endpoint exposed by your IGS instance to Predix Machine.

{
	"IGS_CONFIG" : {

		"SERVER_URI" : "opc.tcp://localhost:49310",
		"APP_URI" : "urn:localhost:Intellution.IntellutionGatewayOPCServer:UA Client Driver",
		"PRODUCT_URI" : "urn:localhost:UA:PredixMachine_OPCUAAdapter",
		"NAMESPACE_INDEX" : "your.namespace.index"
	},
}

PREDIX_GLOBAL_CONFIG

The PREDIX_GLOBAL_CONFIG key in the root of JSON specify Predix Machine most-important configuration file names and extensions. This field is intended to provide more flexibility to future version of Predix Machine.

	"PREDIX_GLOBAL_CONFIG" : {

		"GLOBAL_CONFIG_SUFFIX" : ".config",
		"DATATYPE_CONFIG_SUFFIX" : ".xml",
		"OPCUA_DATANODES" : "com.ge.dspmicro.machineadapter.opcua-0",
		"HOOVER_SPILLWAY" : "com.ge.dspmicro.hoover.spillway-0",
		"PREDIX_CLOUD_IDENTITY" : "com.ge.dspmicro.predixcloud.identity",
		"PREDIX_WEBSOCKER_RIVER" : "com.ge.dspmicro.websocketriver.send-0",
		"PROXY_CONFIGURATION" : "org.apache.http.proxyconfigurator-0"

	},

PREDIX_CLOUD_CONFIG

The PREDIX_CLOUD_CONFIG key in the root of JSON specify Predix Cloud most-important configuration parameters. Here you can define your timeseries Zone-ID (env vars), UAA parameters and proxy (if needed by your organization) with exceptions. The files involved by these keys are the one pointed by PREDIX_GLOBAL_CONFIG.PREDIX_WEBSOCKER_RIVER, PREDIX_GLOBAL_CONFIG.PREDIX_CLOUD_IDENTITY and PREDIX_GLOBAL_CONFIG.PROXY_CONFIGURATION.

	"PREDIX_CLOUD_CONFIG": {

		"TIMESERIES_ZONE_ID": "f37*****-****-****-****-************",
		"URL_OAUTH_TOKEN": "https://********-****-****-****-************.*****************/oauth/token",
		"APP_CLIENT_ID": "**********",
		"APP_CLIENT_SECRET": "**********",
		"PROXY_HOST": "**********",
		"PROXY_EXCEPTIONS": [
			"**********"
		]

	},

You can add more then one exception from Webgui (see below).

TAG_MAPPING_SCHEMA

The TAG_MAPPING_SCHEMA key in the root of JSON specify Tag Mapping between your Predix Cloud application and IGS (or your collector): it is a dict, with key defining Tag Mapping schema names, and values, the respective content of the schema. The file involved by these keys are the one pointed by PREDIX_GLOBAL_CONFIG.OPCUA_DATANODES.

NOTE: each Tag Schema have to be named: if you procede without giving a name to a Tag Schema, a default name will be assigned to the schema. Further, each Tag Couple has to be “headed” with a destination tags: if you procede without giving a name to a Tag Couple (pairing), a default head tag will be assigned. This logic is intended to prevent losing of partial setup during step.

	"TAG_MAPPING_SCHEMA": {

		"tag_mapping_schema_1": {

			"CHANNEL_PREFIX": "channel_prefix_1",
			"MAPPING": {
				"PREDIX_TAG_1": "MACHINE_TAG_1",
				"PREDIX_TAG_2": "MACHINE_TAG_2",
				"PREDIX_TAG_3": "MACHINE_TAG_3",
				"PREDIX_TAG_4": "MACHINE_TAG_4",
				"PREDIX_TAG_5": "MACHINE_TAG_5",
				"PREDIX_TAG_6": "MACHINE_TAG_6",
				"PREDIX_TAG_7": "MACHINE_TAG_7",
				"PREDIX_TAG_8": "MACHINE_TAG_8",
				"PREDIX_TAG_9": "MACHINE_TAG_9"
			}

		},

		"tag_mapping_schema_2": { ... },
		... // other schema go here

	}

You can easly add new Tag Schema using web app.

SITES

The SITES key in the root of JSON specify Sites and Groups of machine: it is a dict, with key defining Site names, that maps into another dict, with key defining Group name, the maps into another dict with values defining the specific Group. In the last step before result you can define machines for each Group in each Site. The key are clear: GLOBAL_ENABLE_SETUP enable setup for each machine, GLOBAL_TAG_MAPPING_SCHEMA schema will define a general tag mapping schema (you could define a local one for each machine in each group, and enable/disable setup for specific machine), and so on… The file involved by these keys are the one pointed by PREDIX_GLOBAL_CONFIG.OPCUA_DATANODES.

NOTE: each Site and each Group of machines in each Site have to be named: if you procede without giving a name to a Site Schema, a default name will be assigned to the schema. Further, each Group has to be “headed” with a name: if you procede without giving a name to a Group, a default head name will be assigned. This logic is intended to prevent losing of partial setup during step.

  "SITES": {
    "site_1_name": {
      "group_1_name": {
        "GLOBAL_TAG_MAPPING_SCHEMA": "tag_mapping_schema_1",
        "GLOBAL_ENABLE_SETUP": true,
        "GLOBAL_PREFIX": "predix.group.prefix.1_",
        "MACHINES": {
          "machine_plate_1": {
            "PCU": "pcu_name_1",
            "IP": "machine_ip",
            "NAT": "machine_natted_ip",
            "INFO": "machine_description",
            "PROTOCOL": "machine_protocol",
            "LOCAL_TAG_MAPPING_SCHEMA": "tag_mapping_schema_1",
            "LOCAL_ENABLE_SETUP": true
          },
          "machine_plate_2": {

        "tag_mapping_schema_2": { ... },
        ... // other schema go here

	}

MACHINES

The MACHINES key in the body of a Group specify machines: it is a dict, with key defining properties. They are not all necessary: the most important are the machine plate and the PCU name for IGS. The other one are only descriptive information. The file involved by these keys are the one pointed by PREDIX_GLOBAL_CONFIG.OPCUA_DATANODES.

NOTE: each machine plate must be different from the others in the same Group.

  "SITES": {
    "site_1_name": {
      "group_1_name": {
        "GLOBAL_TAG_MAPPING_SCHEMA": "tag_mapping_schema_1",
        "GLOBAL_ENABLE_SETUP": true,
        "GLOBAL_PREFIX": "predix.group.prefix.1_",
        "MACHINES": {
          "machine_plate_1": {
            "PCU": "pcu_name_1",
            "IP": "machine_ip",
            "NAT": "machine_natted_ip",
            "INFO": "machine_description",
            "PROTOCOL": "machine_protocol",
            "LOCAL_TAG_MAPPING_SCHEMA": "tag_mapping_schema_1",
            "LOCAL_ENABLE_SETUP": true
          },
          "machine_plate_2": {

        "tag_mapping_schema_2": { ... },
        ... // other schema go here

	}

How it works: Third Millennium way

I create an angular4-based webapp to help you create your your-factory.json configuration file. The app will guide you through various step, but you can always see in real time the preview of your json configuration changing over time. You can setup your app to run locally or use the one host on TODO

Setup Webapp to run locally

Make sure you have NPM and AngularCLI installed on your machine. Then:

git clone https://github.com/made2591/predix-configurator
cd predix-configurator/webgui
npm install
ng serve

And point your browser to http://localhost:4200/

Thank you everybody for reading!

more at http://predix.io ↩︎
more at http://pivotal.io ↩︎

Subscribe to my newsletter to be informed about my new blog posts, talks and activities.