watson / throttling

Throttle a function and cache the result for x milliseconds

throttling

Throttle a function and cache the result for x milliseconds.

Installation

npm install throttling

Example

var fs = require('fs');
var throttle = require('throttling');

var getFile = throttle(1000*60, function (callback) {
  fs.readFile('/etc/passwd', callback);
});

getFile(function (err, data) {
  console.log(data);
});

Usage

Requireing throttling returns a generator function. The most simple way to use throttling is to call this generator function with the function you want to throttle:

var calls = 0;
var runner = throttle(function () {
  return ++calls;
});

The generator function returns a runner function. The first time you call the runner function, the throttled function will be called and its output will be cached and returned by the runner function. By default subsequent calls to the runner function will not call the throttled function, but just return the cached result:

runner(); // => 1
runner(); // => 1
runner(); // => 1

Cache timeout

You can also set a cache timeout. Parse an integer representing the cache timeout in milliseconds as the first argument to the generator function:

var calls = 0;
var runner = throttle(1000, function () {
  return ++calls;
});

runner(); // => 1
runner(); // => 1
setTimeout(runner, 1001); // => 2

Async

The throttled function is called with a callback as the first argument. If your function needs to do any async work, call this with the result when done:

var runner = throttle(1000, function (callback) {
  process.nextTick(function () {
    callback(Math.random());
  });
});

To get the result of your async function, use a callback when calling the runner function:

runner(function (result) {
  console.log('The random number is:', result);
});

The arguments parsed to the callback will be cached according to the same rules as described previously, so subsequent executions of runner will just call the supplied callback function with the previous arguments until the cache expires.

Delayed callback

When setting up the runner function, it's possible to specify that the callback shouldn't be called with the cached arguments, but instead wait for the cache to expire, and then be called with the new callback arguments from the throttled function:

var calls = 0;
var options = {
  timeout: 1000,
  wait: true
};
var runner = throttle(options, function (callback) {
  process.nextTick(function () {
    callback(++calls);
  });
});

runner(function (result) {
  console.log('1st call:', result);
});
runner(function (result) {
  console.log('2nd call:', result);
});
runner(function (result) {
  console.log('3rd call:', result);
});

The above code will first output 1st call: 1. Then it will wait aproximently 1 second and output:

2nd call: 2
3rd call: 2

Notice how the timeout is now supplied using the timeout property in the options hash.

Error handling

If the first argument parsed to the fn callback is an Error, the callback will not be cached.

var calls = 0;
var runner = throttle(1000, function (callback) {
  callback(new Error(), ++calls);
});

var fn = function (err, result) {
  console.log(result);
};

runner(fn); // calls `fn` with the error and the result 1
runner(fn); // calls `fn` with the error and the result 2
runner(fn); // calls `fn` with the error and the result 3

Note that if options.wait is true, the throttling will still be in effect and the fn function will only be called once for every options.timeout.

License

MIT