KMeans provides less than K clusters

[komzpa]

Clustering 25 distinct points into 25 clusters gets 24 clusters:

select count(distinct cid) from 
(WITH
points AS (
    SELECT ST_MakePoint(x,y) geom from generate_series(1,5) x, generate_series(1,5) y
)
SELECT ST_ClusterKMeans(geom, 25) over () AS cid, geom
FROM points) z;

The larger K is, the bigger losses are.

[komzpa]

fix: ​https://github.com/postgis/postgis/pull/179

[komzpa]

In 16212:

Fix KMeans initialization issue that lost clusters sometimes.

Closes #3965
Closes ​https://github.com/postgis/postgis/pull/179

[komzpa]

In 16213:

Fix KMeans initialization issue that lost clusters sometimes.

Closes #3965
Closes ​https://github.com/postgis/postgis/pull/179

[komzpa]

In 16214:

Fix KMeans initialization issue that lost clusters sometimes.

Closes #3965
Closes ​https://github.com/postgis/postgis/pull/179