Computation on similarity/prediction is very time-consuming. Rather than rebuild all the tables each time, the APIs should have incremental update that only make changes to the updated info.

Comments

danithaca’s picture

danithaca commented
Version: » 6.x-1.1
Status: Active » Postponed
danithaca’s picture

danithaca commented
Status: Postponed » Closed (duplicate)

merged to the umbrella issue #480300: Implement different modes of indexing computation: rebuild/incremental/on-the-fly