Skip to main content
Ontology Workspace is the runtime inspection surface for published semantics. It groups ontology definitions, ontology-owned data, connected business objects, and partitioned project graphs so users can verify what is ready for search, Assistant reasoning, and plugin use. Use Ontology Studio to change a definition. Use Ontology Workspace to answer: what semantic version is active, which objects exist, how are they connected, and what exact snapshot supports this result?

Workspace layout

The page has three working areas: The resource and Assistant panels can be collapsed. The Assistant panel is resizable and remembers its width in the browser.

Summary and health

Summary cards aggregate the current organization view:
  • ontology definition count;
  • definition-owned data count;
  • business object count;
  • resources missing a snapshot;
  • resources with failed snapshots;
  • total nodes and RDF triples.
Each resource or partition has one of these health states:
A visible resource is not automatically ready for Agent or plugin use. Check the health badge and current snapshot before diagnosing search or Action discovery.

Resource grouping

The navigator groups resources by ontology definition when a definition reference is available. A group can contain:
  • the definition resource itself;
  • ontology-owned data published with the definition;
  • business objects or resources that reference that definition;
  • partitioned graphs, such as project-specific business-object subgraphs.
Resources without a definition reference appear under Not associated with an ontology definition. This distinction helps separate manually designed domain ontologies from adapter-generated or legacy resource graphs. Use the navigator search to match definition name, business-object type and key, description, resource ID, owner, version, resource type, or tag. You can also filter by resource type and health. Search active entities across resources by:
  • display label;
  • external key;
  • alias;
  • entity type.
You can optionally constrain the search to an entity type or resource type. Results include the matched field, resource and partition, entity type code, external key, and graph references. Selecting a result switches the graph to its resource and passes the entity to the Assistant context. Search is intentionally bounded. If a broad query reaches the applied limit, narrow the query or add an entity-type filter rather than treating the result set as an exhaustive export.

Instance graph

Select a ready resource or partition and open Instance graph. The graph shows:
  • entity instances as nodes;
  • relationship instances as edges;
  • typed properties and identity fields;
  • source and target direction;
  • resource, snapshot, graph version, ontology, and partition references;
  • resource-specific visualization when a registered visualization is available.
Selecting an entity exposes its details and makes it available to the ontology Assistant. The graph is a snapshot view: it describes published semantic state and does not edit the Studio draft or source system. For the underlying graph model, see Entity, Relationship, and Attribute Graph.

Snapshot history

Open Snapshot history for the selected resource or partition. Up to the latest 30 records show:
  • source data version;
  • whether the snapshot is current;
  • publication and projection health;
  • creation time;
  • graph version;
  • entity, relationship, and triple counts.
Snapshot history is useful for explaining count changes, reproducing an Agent or plugin decision, and distinguishing a stale consumer from a failed current publication. Downstream operations should carry snapshot ID and graph version when reproducibility matters.

Ontology Assistant

When the organization has an effective ontology Assistant, Ontology Workspace mounts it beside the graph. Every request can carry:
  • definition ID, definition resource ID, and semantic version;
  • selected resource and optional partition;
  • snapshot ID and graph version;
  • business-object type and key;
  • selected entity ID, type code, external key, label, and graph references;
  • the current search query.
This lets users ask questions such as:
  • Which objects and relationships explain the current business object?
  • Which constraints or Actions apply to the selected entity?
  • What changed between the current object and an earlier snapshot?
  • Which source or evidence supports this property?
The Assistant is scoped to the current organization and depends on administrator initialization. If the effective Assistant or ChatKit frame is unavailable, the page shows a configuration message rather than silently using another organization or tenant fallback.

Delete an ontology instance or partition

Ontology Workspace supports explicit deletion for runtime instances:
  • deleting a partition removes that project subgraph’s snapshot history, entities, relationships, query cache, and RDF graph;
  • deleting a resource instance archives the resource and cleans its runtime snapshots, instance projections, query cache, and RDF graph.
The unified ontology definition and other project partitions are not deleted. A confirmation dialog describes the exact target and consequence before the operation.
Runtime-instance deletion is different from deleting or archiving an ontology definition in Ontology Studio. Confirm the selected resource ID and partition key before continuing.

Verification checklist

After publishing or synchronizing a resource, verify:
  1. the resource appears in the expected ontology-definition group;
  2. health is ready and a current snapshot is present;
  3. node, relationship, and triple counts are plausible;
  4. a known object can be found by label or external key;
  5. its graph contains the expected typed relationships and properties;
  6. snapshot history shows the intended source version and graph version;
  7. the Assistant receives the selected object without requiring the user to restate its identity;
  8. downstream plugins pin and revalidate resource, entity type, snapshot, and graph context.
Next, see how the Valve Business Workbench uses published ontology context.