Implement one iteration of Lloyd's algorithm.
assign(X, centroids) — return an integer array of length n with the index of the nearest centroid (Euclidean) for each point.update(X, labels, k) — return a (k, d) array of the mean of each cluster's points. If a cluster is empty, its centroid should be all zeros.kmeans(X, init_centroids, iters) — run iters iterations and return the final centroids.