Delete old slices then split communities loaded from db

Use-cases, layers, clusters, timeslices, layerpairs

Empty cluster subsets are still stored as empty list ie. size=0

Added documentation and changed the Endpoints to match Rest conventions, added query attributes for use-case, tables  and layers. Also modified a bit of repository and calculate similarity code to work with the refactoring/adding of use-cases and tables

type_label is in {'high', 'medium', 'low', 'noise'}

Cluster sizes might get too large, so they will still be stored on disk.