Skip to main content

Caching

Ligero ships a small Cache<K,V> abstraction with a load-through helper and TTLs. The in-process InMemoryCache lives in ligero-core; a distributed RedisCache ships in ligero-redis. Program against the interface so you can swap between them.

Install with the CLI
ligero add cache          # in-process (ligero-core)
ligero add redis          # + distributed RedisCache

InMemoryCache needs no extra dependency (it's in ligero-core).

Quick start​

Cache<String, User> users = new InMemoryCache<>();

// load-through: compute + cache on miss, with a 10-minute TTL
User u = users.get(id, Duration.ofMinutes(10), db::findUser);

users.put("token", value, Duration.ofSeconds(30));
users.evict(id);

API​

MethodDescription
get(key)Optional<V> — empty when absent or expired
put(key, value) / put(key, value, ttl)store, optionally with a TTL
get(key, loader) / get(key, ttl, loader)return cached value, else compute + cache
evict(key) / clear()remove one / all entries

Distributed cache (Redis)​

Cache<String, String> cache = RedisCache.usingJedis(pool);
String json = cache.get(id, Duration.ofMinutes(10), this::loadUserJson);

RedisCache is a string cache — serialize structured data (e.g. to JSON) before caching it. Keys are namespaced under cache: by default. clear() is intentionally unsupported (flushing a shared Redis is unsafe) — evict specific keys instead.

Notes​

  • InMemoryCache expires entries lazily (on access). Pair it with a Scheduler task if you want periodic sweeping of entries that are never read again.
  • Load-through uses an atomic compute, so concurrent callers for the same key load at most once.