Skip to main content
Data Xpert organizes metadata and capabilities from external systems into a governed object space. Users register a Resource; Agents consume Entity, Relation, and Action instances projected from a published Ontology Snapshot.

​
Page actions and system objects

​
Object boundaries

​
Resource

Resource registers a connection and its capabilities in the control plane. It answers what the resource is, how it connects, and what it can sync or execute. Registration alone does not make a snapshot ready.

​
Sync Job

A Sync Job is one synchronization request and its queue state. full rebuilds the resource projection; incremental updates it using a cursor supported by the adapter. A job must complete before a new snapshot can be published.

​
Ontology Snapshot and Graph Version

An Ontology Snapshot is a published view of semantic facts, including sources, entities, relationships, actions, and statistics. A Graph Version identifies the runtime graph projection. Snapshots reproduce Agent context; graph versions locate graph-query results.

​
Ontology Definition and Runtime Projection

An Ontology Definition is an editable definition and its immutable published versions in Ontology Studio. A Runtime Projection turns a published version and resource snapshot into searchable entities, relationships, and actions. Editing a draft does not change runtime data until validation and publication succeed.

​
Entity, Relation, and Action

An Entity is an object instance with a stable type and external key. A Relation connects two entities with direction, endpoints, and properties. An Action is an executable semantic contract with a target entity, input schema, risk, and discovery mode. Defining an action does not grant permission or guarantee source-system execution.

​
Policy, Approval, and Audit

Policy decides whether an action is allowed, denied, or sent for approval. An Approval applies to a specific target, parameter set, and expected effect. Audit records discovery, simulation, policy decisions, and execution evidence. Together they form the governed execution loop.

​
semantic_model and business_ontology

semantic_model usually comes from an existing BI or OLAP semantic model and is suited to querying Cubes, dimensions, and metrics. business_ontology is a domain definition designed in Ontology Studio and is suited to stable business objects, relationships, owned instances, and domain actions. They can be related at runtime but serve different purposes. Register and sync an existing semantic model first; create a business_ontology definition when you need cross-resource semantics, owned instances, or a domain governance contract.