Class RunClusterAnalysisInput

Clustering algorithm: 'kmeans' or 'dbscan'.

Legend mode: color points by 'cluster' (default) or 'entity'.

Number of dimensions for the projected layout (2 or 3). Defaults to 2.

Distance metric: 'cosine', 'euclidean' or 'dotproduct'.

Multi-entity clustering: one or more Entity Document IDs whose vectors are merged into a single analysis. Takes precedence over EntityDocumentID when non-empty. All selected docs must share the same embedding model (the engine hard-blocks mismatches).

Epsilon neighbourhood radius for DBSCAN.

Optional SQL-style filter applied by the vector source.

Number of clusters for K-Means.

Maximum number of records to fetch for clustering.

Minimum points to form a dense region in DBSCAN.

When true, runs the optional LLM cluster-naming step.

When supplied, the result is persisted as an MJ: Cluster Analysis row with this Name and the returned AnalysisID is populated.