The schema is composed of 3 pieces of data, which are the parameters for the store() method:
redis-completion uses multiple sorted sets to perform searches for partial phrases. When indexing a term, it is split up into pieces of increasing length, so “test” becomes [te, tes, test]. Each of these partial phrases becomes the key to a sorted set. Each sorted set contains the IDs of data that matched that particular phrase. After finding the IDs, the associated data is pulled from a hash keyed by the object’s id.