Partition points into a chosen number of clusters by grouping each point with its nearest cluster centre.

Input parameters:

1. Pt → Points to group.
2. Ct → Number of clusters to form.
3. Sd → Optional starting centroid positions. When provided, used in place of k-means++ initial picks. If fewer seeds than Count are supplied, the remaining slots are filled via k-means++; extra seeds beyond Count are ignored.
4. Rn → Random engine used for distribution.

Output parameters:

1. Pt → Points in each group, one branch per group.
2. Ix → Indices of points in each group (relative to the input cloud), one branch per group.
3. Ce → Final cluster centres — the mean position of each group's members.