vsm / vsm-dictionary-cacher Goto Github PK

Maintain a cache for getRefTerms() too.

• Load all at first call for anything, then just look up any filter.str-requests in there.
• And catch and resolve concurrent calls.
• -> Then cacheEmpty-hits should work instantaneously in vsm-autocomplete.
  
  

Maintain cache a for getDictInfos() too.

• Piggy-back calls if an identical request is already running.
• Look if all elems of the given array are already in the cache.
  • If so, build a result with them and async-return it.
  • If not (if even only 1 of them isn't there yet), simply query them all.
  • Cache timeout/size management: only clear this when the master timeout fires.
• -> Then cache-hits should work instantaneously in autocomplete.
  
  

Enhance the cache management for loadFixedTerms() too.

• • Note: this function (of the VsmDictionary parent class) queries
    the VsmDictionary subclass's getEntries(), and puts the processed
    results in its own simple cache (in the VsmDictionary parent class).
  • But if a webpage contains multiple vsm-autocomplete components,
    then each of them will call the (shared) VsmDictionary's loadFixedTerms()
    and launch a query. So this queries and adds results to the VsmDictionary's
    fixedTermsCache, no matter whether these results were in there already.
• If a call with the same arguments has ever been made before,
  (?:) or if the requested idts are all already in the cache,
  then just async-return OK immediately.
• If a call with same arguments is still ongoing, then don't send a parallel
  same request, but wait until that one finishes, and then call the
  queued one's callback using the same results.