Caches exist to make things faster (at the expense of taking up more memory and possibly outdated results). Our live search is a great use case for implementing a cache, so let's set up amapBroadcasterCacheoperator which can map a string to the result of a broadcaster.

export let mapBroadcasterCache = createBroadcaster => broadcaster => listener => {
  let cache 
 
= new Map()
  let cancel

  return broadcaster(value => {
    // abort previous 
    if (cancel) {
      cancel()
      return;
    } 

    if (cache.has(value)) {
      listener(cache.get(value))
      return
    }
    let newBroadcaster = createBroadcaster(value)
    cancel = newBroadcaster((newValue) => {
      
 
// Add to cache only if newValue isn't an error
      if (!(newValue instanceof Error)) {
        cache.set(value, newValue)
        }
      console.log(cache)
      listener(newValue)
    })
  })
}

export let ignoreError = broadcaster => listener => {
  return broadcaster((value) => {
    
 
if (value instanceof Error) {
      return
    }
    listener(value)
  })
}

Using:

  let inputToBooks = pipe(
    filter(name => name.length > 3),
    waitFor(150),
    pipe(
      map(name => `https://openlibrary.org/search.json?q=${name}`),
      mapBroadcasterCache(getUrl),
      ignoreError,
      map(json => json.docs)
    ))(inputValue)