Daura clustering should return clusters as a ds-array with cluster ids
Is your feature request related to a problem? Please describe. Consistency between clustering algorithms.
Describe the solution you'd like Daura clustering should return clusters as a ds-array with cluster ids. Numbering clusters by discovery order and 0-based would be logic. It should return another ds-array or list, of length n_clusters_found, with the centers of each cluster.
Describe alternatives you've considered Leave it as it is, with a list of lists: a list of ids for each cluster. This is not consistent and does not use distributed arrays. We can consider this if the proposed solution is much slower. But we should try it first.
Additional context The current version has been used successfully for a scientific application, but all the outputing needs to reviewed.