@platformatic/undici-cache-memory

Memory cache store implementation for the undici cache interceptor, providing efficient in-memory caching of HTTP responses.

Installation

npm install @platformatic/undici-cache-memory

Usage

const { Client, interceptors } = require('undici')
const MemoryCacheStore = require('@platformatic/undici-cache-memory')

// Create a client with the cache interceptor
const client = new Client('https://example.com')
  .compose(interceptors.cache({
    store: new MemoryCacheStore()
  }))

// Make a request - first request will hit the origin
const response1 = await client.request({
  method: 'GET',
  path: '/'
})

// Second request to the same URL will be served from cache if cacheable
const response2 = await client.request({
  method: 'GET',
  path: '/'
})

Configuration

The MemoryCacheStore constructor accepts an options object with the following properties:

const store = new MemoryCacheStore({
  // Maximum number of entries to store (default: 1024)
  maxCount: 1000,
  
  // Maximum total size in bytes (default: 100MB)
  maxSize: 1024 * 1024 * 10, // 10MB
  
  // Maximum size of a single entry in bytes (default: 5MB)
  maxEntrySize: 1024 * 1024, // 1MB
  
  // Header name to parse for cache tags (default: undefined)
  cacheTagsHeader: 'cache-tag'
})

Cache Invalidation

By Cache Tag

// Create the store with a header to parse for cache tags
const store = new MemoryCacheStore({
  cacheTagsHeader: 'cache-tag'
})

// Add the store to the client
const client = new Client('https://example.com')
  .compose(interceptors.cache({ store }))

// Later, invalidate all cached responses with specific tags
store.deleteTags(['product-123', 'category-456'])

By Request Key

// Invalidate specific paths
store.deleteKeys([
  { origin: 'example.com', path: '/products/123', method: 'GET' },
  { origin: 'example.com', path: '/categories', method: 'GET' }
])

// Or delete a specific entry
store.delete({ origin: 'example.com', path: '/products/123', method: 'GET' })

Advanced Features

This implementation supports:

• Caching based on standard HTTP cache semantics
• Respect for Vary headers for content negotiation
• Support for stale-while-revalidate and stale-if-error directives
• Memory management with configurable limits
• Cache tag-based invalidation for efficient cache purging

License

Licensed under Apache-2.0.