lru-cachify
v1.0.6
Published
Wrap functions in lru-cache
Downloads
2
Readme
lru-cachify
Purpose: easily wrap a function to provide caching. Inspired by memoizee, but simpler in some ways and with some extra / different functionality. Uses lru-cache for the caching functionality itself.
Usage
var cachify = require('lru-cachify');
var getThings = cachify(function () {
return { };
});
getThings('foo') === getThings('foo');
getThings('foo') !== getThings('bar');
Options
You can pass lru-cache options as an optional first argument
var getThings = cachify({
max: 1
}, function () {
return { };
});
var foo = getThings('foo'),
bar = getThings('bar');
getThings('bar') === bar;
getThings('foo') !== foo;
In addition to this, lru-cachify provides three options of its own:
- normalize: a function to normalize the options before constructing the cache key
- length: the number of arguments to use in constructing the cache key
- key: a function that returns the cache key, given the normalized (and possibly truncated) arguments
Example:
var getThings = cachify({
normalize: function (a, b, c) {
// allow passing of an array or separate arguments
if (Array.isArray(a)) { return a; }
return [a, b, c];
},
key: function (a, b) {
return a + b;
},
length: 2
}, function (a, b, c) {
return a * b + c;
});
getThings(2, 3, 4) === 10;
getThings([2, 3, 4]) === 10; // array normalized to arguments
getThings(3, 2, 4) === 10; // 3 + 2 === 2 + 3, so the key is the same
getThings(2, 3, 5) === 10; // length is 2, so the third argument is ignored when constructing the key
LRU Methods
Public methods of LRU instances are assigned to the returned function, so you can do things like peek or reset the cache:
var count = 0;
var getCount = cachify(function () { return ++count; });
getCount() === 1;
getCount() === 1;
getCount.reset();
getCount() === 2;
Promises
The promise itself is cached, but deleted if it rejects. Because of this, multiple equivalent synchronous calls will be combined into a single promise, reducing database or api overhead:
var getThings = cachify(function (id) {
return db.queryAsync('SELECT name FROM people WHERE id = $1', [ id ]);
});
var foo = getThings(1),
bar = getThings(1);
foo === bar;
foo.then(function (val) {
val === 'Bob'; // assuming people has a record with {id: 1, name: 'Bob'}
});
Rejections will be batched in the same manner as a result of this behavior:
var rejections = 0;
var getThings = cachify(function (id) {
return Promise.reject(++rejections);
});
var foo = getThings(1),
bar = getThings(1);
foo === bar;
foo.then(null, function (err) {
err === 1;
});
bar.then(null, function (err) {
err === 2;
});
The difference is shown here:
var succeed = cachify(function () {
return Promise.resolve('yes');
});
var fail = cachify(function () {
return Promise.reject('no');
});
var foo = succeed();
foo.then(function () {
succeed() === foo;
});
var bar = fail();
bar.then(function () {
fail() !== bar;
});
There is currently no similar handling for callback APIs.