Feature: Index caching

[mgalang]

Description

The process of re-indexing massive amount of entities are often quite slow because each entity needs to be loaded, recalculated and serialized. This process could be speed up significantly by caching the entity's serialized data when it is saved, and by loading the cached version of the entity's serialized data during re-indexing processes. The cached data is invalidated and re-created on each entity update to ensure that the cache remains up to date.

How it works

1. Caching could be enabled per index from the index's annotation (see example).
2. Entity is serialized when it is saved.
3. The serialized data is cached (the cache ID includes language, index-id, entity-type and id as a context).
4. The serialized cached data is loaded during re-indexing