NodeSet.js

A graphical modeler for OPC UA information models. Types and instances are drawn in the notation of OPC 10000-3 Annex C and stored as NodeSet2 files, with nothing in between: the file it opens and the file it writes are the NodeSets every other OPC UA tool reads. It runs in the browser on its own, and it embeds: the AutomationML Editor plugin AMLOpcUa hosts it and connects the models to AutomationML (OPC 10000-83 Annex A, AML-UA-XSLT rules) and to VDI 3682 process descriptions.

What it does: types (ObjectTypes, VariableTypes, DataTypes with fields, encodings, enumeration values, OptionSets and unions, ReferenceTypes), instance declarations with ModellingRules, values (built-in types and simple structures), method arguments, references, instances by ModellingRule, models built on DI or other loaded models with their RequiredModels kept, checks against OPC 10000-3, undo, editing from the canvas's context pad, and diagram positions kept in the NodeSet. Inside the AutomationML Editor it runs in the Modeler tab of AMLOpcUa, which opens a namespace of the document and imports the result back (AMLOpcUa docs/modeler.md).

Status: in development.

npm install
npm run dev        # http://localhost:3002
npm test
npm run typecheck
npm run lint       # the order of React hooks
npm run test:e2e   # the web build in Chromium (npx playwright install chromium once)

npm run test:e2e also opens twelve released companion specifications when UA_NODESET names a clone of OPCFoundation/UA-Nodeset with the folders in tests/e2e/corpus-folders.txt; CI checks them out at the commit named in .github/workflows/ci.yml.

On a page of its own the modeler keeps unsaved work in the browser (IndexedDB) and offers it at the next start; embedded in a host it does not.

Third-party content

assets/nodesets/ holds the OPC UA base NodeSet (1.05.07) and the OPC UA for Devices NodeSet (1.05.0) from UA-Nodeset, OPC Foundation MIT License 1.00.

License

MIT

As a library

npm run build       # dist/web (the app) and dist/lib (the library)

import { readNodeSet, writeNodeSet, Workspace, NodeSetModeler, App } from 'nodeset-js';

const ws = new Workspace();
await ws.open(xml);                         // loads the required UA and DI models
const pump = ws.editor!.addType('ObjectType', 'PumpType');
ws.editor!.addDeclaration(pump, 'Variable', 'Speed');
const nodeSet = ws.save();                  // NodeSet2 XML

App is the complete modeler as a React component; NodeSetModeler is the canvas alone. React 18 is a peer dependency.

For a script, in a page or in Node, there is a second entry without the canvas and without React:

const { Workspace, check } = require('nodeset-js/core');

const ws = new Workspace();
await ws.create('http://example.org/Pump/');            // brings the base model with it
const type = ws.editor.addStateMachineType('PumpStateMachineType');
const idle = ws.editor.addState(type, 'Idle', 1);
const running = ws.editor.addState(type, 'Running', 2);
ws.editor.addTransition(type, 'IdleToRunning', 1, idle, running);
console.log(check(ws.space, ws.editable));              // the rules on the model
require('fs').writeFileSync('Pump.NodeSet2.xml', ws.save());

The base NodeSet and DI are inside that build, so a script needs nothing else. npm run verify:core checks that it still runs outside a browser.