shm-lru-cache
v0.1.9
Published
A shared memory cache. This is a wrapper for shm-typed-lru that provides a familiar cache interface.
Downloads
19
Readme
shm-lru-cache
A cache object that deletes the least-recently-used items.
Installation:
npm install shm-lru-cache --save
Usage:
In general a cache will be created for some object using it.
Here is one that creates a section of memory for the LRU:
this.cache = new LRU({
"token_path" : "./ml_relay.conf",
"am_initializer" : true,
"seed" : 95982749,
"record_size" : 384,
"el_count" : 20000,
"stop_child" : true
})
Here is one that attaches to a section that has already been created:
this.cache = new LRU({
"token_path" : "./ml_relay.conf",
"am_initializer" : false,
"seed" : 95982749, // same seed
"record_size" : 384, // same
"el_count" : 20000, // same
"stop_child" : true
})
In the following a library is making use of the created cache:
// ----
check_hash(hh_unid,value) {
let sdata = ( typeof value !== 'string' ) ? JSON.stringify(value) : value
let hh_unidentified = this.cache.hasher(sdata)
return hh_unidentified === hh_unid
}
// ----
delete(key) { // token must have been returned by set () -> augmented_hash_token
let augmented_hash_token = this.cache.hash(key)
this.cache.del(augmented_hash_token)
}
// ----
async get(key) { // token must have been returned by set () -> augmented_hash_token
let augmented_hash_token = this.cache.hash(key)
let value = await this.cache.get(augmented_hash_token)
if ( typeof value !== 'string' ) {
return false
}
return value
}
// ----
async get_with_token(token) {
let value = await this.cache.get(token)
if ( typeof value !== 'string' ) {
return false
}
return value
}
// ----
run_evictions(time_shift,reduced_max) { // one process will run this
let evict_list = this.cache.immediate_mapped_evictions(time_shift,reduced_max)
return evict_list
}
// ...
// ----
disconnect(opt) {
if ( this.ev_interval ) {
clearInterval(this.ev_interval)
this.ev_interval = false
}
this.in_operation = false
return this.cache.disconnect(opt)
}
Options
am_initializer
: whether or not this is a creator.max_evictions
:el_count
: number of elements that will be stored in the cachetoken_path
: a path to an actual file on disk. The path name is used to create a token for the memory region.record_size
: the size of an elementseed
: this is the seed for the hash function.sessions_length
:aged_out_secs
:
API
hash(value)
: returns an augmented hash, includes and index,hyphen,pure hashpure_hash(value)
: returns a hash of the value returned by the default hasher (which may be an xxhash)augment_hash(hash)
: given a pure hash, return the augmented versionasync set(hash_augmented,value)
: put the value into the LRUasync get(hash_augmented) => value
: retrieve the valueBoth of these will update the "recently used"-ness of the key. They do what you think.
If the key is not found,
get()
will returnfalse
.The key and val can be any value.
async del(hash_augmented)
alsoasync delete(hash_augmented)
Deletes a key out of the cache.
The following methods have to do with removing old keys from the LRU:
immediate_evictions(age_reduction,e_max)
: Returns an array of keys that were removed.
For the next two, the map of pure hashes to values can be passed upstream to other LRUs. These methods return the map.
immediate_mapped_evictions(age_reduction,e_max)
: Returns a map of keys to valuesimmediate_targeted_evictions(hash_augmented,age_reduction,e_max)
Returns a map of keys to values. The evictions are mostly from the bucket related to the hash parameter.disconnect(opt)
: Cleans up the library assets when shutting down.app_handle_evicted(evict_list)
: If a descendant is written, this will be passed a list of hashes. Use by immediate_evictions.