Module lobe_cache 

Lobe output cache — skip re-execution when input hasn’t changed.

If a lobe receives the same input twice (common in multi-iteration cognitive graphs), the cache returns the previous result without calling the lobe’s process() method. TTL-based expiry ensures stale results are evicted.

Structs

LobeCache

Cache for lobe outputs, keyed by (lobe_name, input_hash).