# Rank all mentities by the distance between `MEntity::name` and `name`
	#
	# Use the Levenshtein algorithm on all the indexed mentities `name`.
	# Warning: may not scale to large indexes.
	fun find_by_name_similarity(name: String, filter: nullable ModelFilter): IndexMatches do
		var results = new IndexMatches
		for match in name_distances.search(name) do
			var dist = match.distance
			var mname = match.key
			if not names.has_key(mname) then continue
			for mentity in names[mname] do
				if mentity isa MClassDef or mentity isa MPropDef then continue
				if filter != null and not filter.accept_mentity(mentity) then continue
				results.add new IndexMatch(mentity, dist)
			end
		end
		return results
	end