runtime-memcache

A no dependency, high performance, near optimal javascript caching library

runtime-memcache is a caching library to store key-value cache store for small chunks of arbitrary data (strings, objects, numbers) from results of database calls, API calls, or etc. It is entirely written using Typescript and supports many commonly used caching policies.

When creating a new cache store, you can specify the policy to evict items from the store. The default policy is lru (Least Recently Used)

runtime-memcache provides flexible construction to create a cache with a combination of the following features:

• size-based eviction when a maximum is exceeded based on frequency and recency
• time-based expiration of entries, measured since last access or last write

Installation

npm install --save runtime-memcache
# or using yarn
yarn add runtime-memcache

Node Environment (ES6+ import/export)

import createStore from 'runtime-memcache';

Node Environment (CJS)

const createStore = require('runtime-memcache');

Browser (use as a script tag)

<script src="https://unpkg.com/[email protected]/dist/umd/index.js"></script>
<!-- OR JUST -->
<script src="https://unpkg.com/[email protected]"></script>
<script>
  // RMStore is globaly set
  const store = new RMStore();
</script>

API

Calling the createStore function returns an object with the following properties.

Property	Description
get(id)	Retrieves an item from the store
has(id)	Check if an item exists in the store
set(id, data)	Sets an item in the store
remove(id)	Removes an item from the store
size()	Returns the size of the item cache store
keys()	Returns all the keys of the cache store as array

Config

createStore takes an optional config object as an argument with the following properties.

Property	Description	Type	Default
timeToClear	Time in milliseconds for which the store will keep an item when the policy is timeout or tlru	Number	7200000
policy	A Policy to evict items from the store	timeout, lru, mru, tlru	lru
lruSize	Size of the cache store when the policy is lru or tlru	Number	500
mruSize	Size of the cache store when the policy is mru	Number	500

Caching Policies

Following caching policies are supported.

Policy	Name	Description
timeout	Timeout	The items in the cache store will be automatically evicted after a fixed amount of time has elapsed since that item was set
lru	Least Recently Used	This policy evicts the least recently used items when the store size is full
tlru	Time Aware Least Recently Used	This policy evicts the least recently used items when the store size is full and also evict untouched items after a fixed amount of time has elapsed since that item was set
mru	Most Recently Used	This policy evicts the most recently used items when the store size is full

Time Complexity

Policy	Method	Complexity
timeout	set, get, remove	O(1), O(1), O(1)
lru	set, get, remove	O(1), O(1), O(1)
tlru	set, get, remove	O(1), O(1), O(1)
mru	set, get, remove	O(1), O(1), O(1)

Usage

Typescript

import createStore, { Config } from 'runtime-memcache';

const config: Config = {
  policy: 'lru',
  lruSize: 300, // cache a maximum of 300 users at a given time
};

interface User {
  name: string;
}

const userCache = createStore<User>(config);

async function loginUser(userId: string) {
  if (userCache.has(userId)) {
    return userCache.get(userId);
  }

  const user: User = await UserService.getUser(userId);

  userCache.set(userId, user);

  return user;
}

Javascript

import createStore from 'runtime-memcache';

const config = {
  policy: 'timeout',
  timeToClear: 7200000, // 2 hours
};

type Keys = 'key1' | 'key2';

const store = createStore(config);

store.set('key1', { name: 'name' }); // store the object and associate it with the provided key

store.get('key1'); // retrieves the object associated with this key

store.has('key1'); // returns true

store.size(); // returns 1

store.keys(); // returns ['key1']

store.remove('key1'); // deletes the object associated with this key

NPM Script Commands

• npm run test -- Runs tests, lint and build.
• npm run lint -- Runs ESLint.
• npm run format -- Reformats all of the .ts and .tsx files with Prettier.
• npm run build -- Regenerates dist folder that gets included into NPM module.

Under The Hood

runtime-memcache uses a combination of modified doubly-linked lists and hashmap data structures to achieve O(1) search-time complexity for all the methods.

Todos

• Timeout Policy (TR)
• Least Recently Used Policy (LRU)
• Most Recently Used Policy (MRU)
• Least Frequently Used Policy (LFU)
• Time Aware Least Recently Used Policy (TLRU)
• Random Eviction Policy (RR)
• Add a warmup period for new items

For more information on caching policies read this