get_resource

Scope: read · Signature: get_resource(kind: {workspace|dataset|experiment|model|knowledgebase|connection|plan}, id: str) -> Any

Use when you have an id and need the full detail object.

kind ∈ {workspace, dataset, experiment, model, knowledgebase, connection, plan}. An experiment detail carries type (supervised | external | cluster | anomaly | forecast | recommendation | ltr | association), status, target, and dataset_id — match dataset_id in describe_catalog to find its INPUT table. A model detail carries its stage, lineage (source experiment, dataset, target, preset), versions and serving endpoints. A knowledge base carries its tables, relationships and recipes — each recipe's SQL, target, prediction moment and latest version's point-in-time check, not its version history. A plan carries its request, validity (valid, errors, warnings), resolved config and estimate. A dataset, experiment or knowledge base also carries its annotations: AI-written text (summaries, cluster names, descriptions) with the agent and model that wrote it. For a run's progress or failure use get_run.