Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.xpertai.cn/llms.txt

Use this file to discover all available pages before exploring further.

The semantic_model resource connects Xpert semantic models to the UOSE system. It syncs models, Cubes, metrics, measures, and dimensions as ontology objects and provides read-only analytical actions through semantic query interfaces.

Synced Objects

After sync, it usually contains:
  • semantic_model: top-level semantic model.
  • semantic_cube: analytical Cube under the model.
  • semantic_indicator: business metric.
  • semantic_measure: measure under the Cube.
  • semantic_dimension: dimension under the Cube.
Core relationships include:
  • semantic_model_has_cube
  • semantic_cube_has_indicator
  • semantic_cube_has_measure
  • semantic_cube_has_dimension
These relationships let Agents find the owning Cube, available measures, dimensions, and query contract from a metric.

Connection and Capabilities

Connections usually point to the xpert-pro API. Secrets support:
  • current_user_oidc: use the current user’s OIDC token.
  • static_bearer: use a static Bearer token.
  • api_key: use an API key.
Capabilities mainly include:
  • modelIds: local model filter after sync.
  • endpoints.modelList: semantic model list API.
  • endpoints.modelById: model detail API.
  • endpoints.query: semantic query API.
  • pagination.models: model list pagination configuration.
If the query endpoint is missing, query actions are rejected during the runtime readiness stage.

Analysis Contract

analysis_contract is the key for semantic model resources to execute fact queries. It usually appears in semantic_cube and semantic_indicator attributes and contains:
  • modelId.
  • cubeId.
  • cubeName.
  • metricRefs.
  • defaultTimeDimension.
  • supportedGranularities.
  • defaultDimensions.
  • filterableFields.
When analysis_contract is missing, the entity can still be searched and displayed, but query actions are rejected.

Actions

Common actions include:
  • semantic_model.query_metric_snapshot: query metric time series.
  • semantic_model.query_cube_slice: query Cube group slices.
  • semantic_model.query_calculated_expression: query calculated expressions.
  • semantic_model.query_statement: execute controlled semantic query statements.
  • semantic_model.query_native_dsl: execute native semantic query DSL.
These actions are low-risk read-only actions and usually do not require approval, but they still go through parameter validation, policy evaluation, and audit recording.
  1. Register the resource and complete full sync.
  2. In the ontology workspace, confirm that models, Cubes, metrics, measures, and dimensions have entered the graph.
  3. Use queryEntities to locate the target indicator or cube.
  4. Use getEntityNeighborhood to inspect the owning Cube and available dimensions.
  5. Use discoverActions to confirm executable query actions.
  6. Use simulateAction to validate time window, granularity, dimensions, and filters.
  7. Use executeAction to return trends, tables, or analytical summaries.