Extracting MSTs

The estimator exposes the current HDBSCAN-style MST wrapper after fit(X, k=...).

clusterer.fit(X, k=10)
mst = clusterer.minimum_spanning_tree_

For advanced inspection, the internal object can also return a simple DataFrame:

mst_df = clusterer.core_sg_.extract_mst_from_core_sg(k=10, toDF=True)

The DataFrame has three columns:

to

First endpoint.

from

Second endpoint.

weight

Mutual-reachability edge weight used in the extracted MST.

With toDF=True, the internal Core-SG object returns typed columns named to, from, and weight.