LOKE cache-kit

Examples

Basic usage

For the most common cases use a simple 2 layer LRU + Redis cache is sufficient.

import { Cache, LruMemoryCacheStore, RedisCacheStore } from "@loke/cache-kit";
import Redis from "ioredis";
import { create as createLogger } from "@loke/logger";

const logger = createLogger();
const redisClient = new Redis();

const cache = new Cache(
  // Keys should be prefixed with the project name to make them unique across
  // all of LOKE
  "foo-service:users:fetch:{userId}",
  // Cache stores will be checked in order, and the first cache to return a
  // value will be used
  [new LruMemoryCacheStore({ max: 1000 }), new RedisCacheStore(redisClient)],
  logger,
);

const userId = 123;

const user = cache.apply({ userId }, async () => {
  // This function will only be called if the cache is empty
  const value = await fetch(`http://example.com/user/${userId}`);

  // the ttl is returned as part of the result so it can vary per request
  return {
    value,
    ttl: 60 * 60 * 1000, // 1 hour in milliseconds
  };
});

Single flight cache

In cases where you don't want the overhead of storing the value in a cache, you can use the cache for single flighting requests. Concurrent requests will piggyback on the result of the first request.

import { Cache } from "@loke/cache-kit";
import { create as createLogger } from "@loke/logger";

const logger = createLogger();

const cache = new Cache(
  "users:{userId}",
  // An empty list of caches will result in a single flight only cache, where
  // the value is only cached for the duration of the request
  [],
  logger,
);

TODO

• JSON reviver support