View Source

New for 10.1.5 (draft preview). This page previews a feature shipping in FrameworX 10.1.5 (planned April 2026). The feature is not available in 10.1.4 or earlier. Content under review.

Bring an RDF/OWL ontology into FrameworX as UserTypes, Tags, and AssetTree folders.

Technical Reference → Solution → Designer → Settings and Tools → Import Tags → OWL/RDF Ontology

N-Triples as the mental model. The Unified Namespace is a triple store by construction. Every UserType, Tag, AssetTree folder, and member that this wizard writes corresponds to one or more RDF triples of the form <subject> <predicate> <object> . — the W3C N-Triples textual form. The import here is a serialization step onto that underlying triple model. See Industrial Ontology Integration How-to for the full standards map.

Prerequisites

A solution open in Designer (new or existing).
A source file: RDF/JSON (.rj). JSON-LD and Turtle support arrives in a future release.
(Optional) a policy template. Pick iof, isa88, or write your own JSON. Defaults work for generic OWL files.

Primary path: the Designer wizard

In Designer, go to Solution → Import Tags → Graph / RDF file.
Pick the source file. The wizard previews class count, individual count, and any warnings.
Pick a policy (iof / isa88 / custom) or leave blank for default.
Choose the mode. Auto is the default and inspects the file to decide schema-only, instances-only, or both.
Click Import. UserTypes and Tags appear in the UNS with full ontology metadata populated.

Alternative path: AI-driven with the MCP tool

If the AI assistant is connected to Designer, use the skill Skill Import Industrial Ontology. The assistant runs the same engine via the import_graph_model MCP tool.

import_graph_model(source="file", file_path="C:/.../biotech.rj", policy="iof", dry_run=true)
# review the summary, then re-run with dry_run=false

Always run dry_run=true once before committing. The preview reports class count, instance count, BaseUserType cycle warnings, and label collisions.

What the import writes

OWL construct	FrameworX column
`owl:Class`	New row in UnsUserTypes
`owl:NamedIndividual`	New row in UnsTags at path `<containment>/<entity>/Attr`
`owl:ObjectProperty` / `owl:DatatypeProperty`	Member row inside the UDT's member-def table
Primary `rdfs:label` / `skos:prefLabel`	`DisplayText`
`skos:altLabel` / `skos:hiddenLabel` / custom alt-names	`Labels` (semicolon-delimited)
`rdfs:comment` / `skos:definition`	`Description`
Class / individual / property IRI	`SourceIri` (full IRI, never truncated)
`rdfs:subClassOf <named class>`	`BaseUserType` (UDT only)
Everything else: `dcterms:*`, alt-language `rdfs:label`, `skos:note`, `owl:versionInfo`, custom predicates	`Attributes` JSON column, keyed as `prefix:localName` (with `@lang` suffix for language tags)

The `/Attr` convention

Every typed OWL individual imports as a Tag named <containmentPath>/<entityName>/Attr. AssetTree folders auto-create from the slashes. The importer never writes folder rows directly. The /Attr leaf collects the entity's typed members so one click reveals everything about the entity in one place. On re-export, the /Attr suffix is stripped so OWL entity IRIs reflect identity, not storage layout.

Common gotchas