Apply a proposed clustering
apply_auto_clusteringDestructive
Apply the grouping a completed start_auto_clustering job proposed: creates the clusters it named and moves the tracked queries into them, as the job's mode said. In "full_regroup" mode a tracked query the proposal places also leaves every cluster the proposal did not put it in; "fill_gaps" and "add_on_top" only add. No cluster or tracked query is deleted. Cluster names are lower-cased like create_clusters, and a proposed name that matches an existing cluster reuses it instead of creating a second one. The response lists each tracked query's resulting clusters, not the memberships removed, so show the proposal (get_job) to the user first. Pass a fresh requestId and reuse it if you retry, so a retry never applies twice.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| jobId | Yes | ||
| projectId | Yes | ||
| requestId | No | Optional idempotency key, 8-255 printable characters. Reuse it only to retry this same call. | |
| organizationId | Yes |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| failed | No | Always present, empty list included: read it rather than inferring success from the status code. | |
| clusters | No | Every cluster the assignments referred to, and whether this call created it. | |
| successful | No | Tracked queries written, each with the cluster ids it ended up with - the resulting state, not the delta. | |
| unassigned | No | Tracked queries the job produced no assignment for. Nothing was written for them and they keep the clusters they already had, under every merge mode. | |
| skippedClusters | No | Proposed names the store cannot hold. A tracked query whose proposed cluster was skipped ends up with fewer clusters than the proposal showed. |