Concept Output Caching

You can cache concept outputs by enabling an in-session memory cache that provides fast reads of commonly accessed data. This eliminates redundant reads and accelerates rule evaluations, targeting sub-100ms access and high hit rates.

Cache invalidation works automatically by checking file modification times, or mtime. When a file changes, the system enforces automatic invalidation to ensure data freshness and maintain correctness across the session.

The best way to speed up sync workflows is using a seamless cache-aside integration. By reading concept outputs through the in-session cache, you optimize performance and accelerate repeated rule evaluations during the current session.

Persistent storage to survive the concept output cache across sessions is currently an optional future enhancement. Right now, the system enforces a multi-layer caching strategy primarily using an in-session memory cache for fast reads.

The main limitation of an in-memory session cache is that it does not persist data across sessions without optional persistent storage. It is designed specifically for workflows that repeatedly access concept outputs within a single session to accelerate rule evaluations.

Yes, concept output caching supports batch cache statistics for monitoring performance. This allows you to track high hit rates and verify that the in-session memory cache achieves sub-100ms access during concept evaluation tasks.