regex-cache

What is regex-cache?

The regex-cache npm package is designed to cache the results of regular expression operations to improve performance. It is particularly useful when the same regular expressions are used repeatedly, as it avoids the overhead of recompiling the regex each time.

What are regex-cache's main functionalities?

Caching Regular Expressions

This feature allows you to cache a regular expression. The first time the regex is used, it is compiled and cached. Subsequent uses of the same regex will retrieve it from the cache, improving performance.

const regexCache = require('regex-cache');
const re = regexCache(/foo/);
console.log(re.test('foo')); // true

Using Cached Regular Expressions

This feature demonstrates that the same regular expression is retrieved from the cache, ensuring that the same instance is used, which can save memory and processing time.

const regexCache = require('regex-cache');
const re1 = regexCache(/foo/);
const re2 = regexCache(/foo/);
console.log(re1 === re2); // true

Other packages similar to regex-cache

lru-cache

The lru-cache package provides a least-recently-used (LRU) cache, which can be used to store any kind of data, including regular expressions. Unlike regex-cache, lru-cache is more general-purpose and can be used for caching various types of data.

memoizee

The memoizee package is a general-purpose memoization library that can cache the results of function calls. It can be used to cache the results of functions that create regular expressions, providing similar performance benefits to regex-cache but with more flexibility.

fast-memoize

The fast-memoize package is another memoization library focused on speed. It can be used to memoize functions that generate regular expressions, offering a high-performance alternative to regex-cache for caching regex results.

README

regex-cache

Memoize the results of a call to the RegExp constructor, avoiding repetitious runtime compilation of the same string and options, resulting in dramatic speed improvements.

• Read what this does.
• See the benchmarks

Install with npm

npm i regex-cache --save

Usage

Wrap a function like this:

var cache = require('regex-cache');
var someRegex = cache(require('some-regex-lib'));

Caching a regex

If you want to cache a regex after calling new RegExp(), or you're requiring a module that returns a regex, wrap it with a function first:

var cache = require('regex-cache');

function yourRegex(str, opts) {
  // do stuff to str and opts
  return new RegExp(str, opts.flags);
}

var regex = cache(yourRegex);

Recommendations

Use this when...

• No options are passed to the function that creates the regex. Regardless of how big or small the regex is, when zero options are passed, caching will be faster than not.
• A few options are passed, and the values are primitives. The limited benchmarks I did show that caching is beneficial when up to 8 or 9 options are passed.

Do not use this when...

• The values of options are not primitives. When non-primitives must be compared for equality, the time to compare the options is most likely as long or longer than the time to just create a new regex.

Example benchmarks

Performance results, with and without regex-cache:

# no args passed (defaults)
  with-cache x 8,699,231 ops/sec ±0.86% (93 runs sampled)
  without-cache x 2,777,551 ops/sec ±0.63% (95 runs sampled)

# string and six options passed
  with-cache x 1,885,934 ops/sec ±0.80% (93 runs sampled)
  without-cache x 1,256,893 ops/sec ±0.65% (97 runs sampled)

# string only
  with-cache x 7,723,256 ops/sec ±0.87% (92 runs sampled)
  without-cache x 2,303,060 ops/sec ±0.47% (99 runs sampled)

# one option passed
  with-cache x 4,179,877 ops/sec ±0.53% (100 runs sampled)
  without-cache x 2,198,422 ops/sec ±0.47% (95 runs sampled)

# two options passed
  with-cache x 3,256,222 ops/sec ±0.51% (99 runs sampled)
  without-cache x 2,121,401 ops/sec ±0.79% (97 runs sampled)

# six options passed
  with-cache x 1,816,018 ops/sec ±1.08% (96 runs sampled)
  without-cache x 1,157,176 ops/sec ±0.53% (100 runs sampled)

# 
# diminishing returns happen about here
# 

# ten options passed
  with-cache x 1,210,598 ops/sec ±0.56% (92 runs sampled)
  without-cache x 1,665,588 ops/sec ±1.07% (100 runs sampled)

# twelve options passed
  with-cache x 1,042,096 ops/sec ±0.68% (92 runs sampled)
  without-cache x 1,389,414 ops/sec ±0.68% (97 runs sampled)

# twenty options passed
  with-cache x 661,125 ops/sec ±0.80% (93 runs sampled)
  without-cache x 1,208,757 ops/sec ±0.65% (97 runs sampled)

# 
# when non-primitive values are compared
# 

# single value on the options is an object
  with-cache x 1,398,313 ops/sec ±1.05% (95 runs sampled)
  without-cache x 2,228,281 ops/sec ±0.56% (99 runs sampled)

Run benchmarks

Install dev dependencies:

npm i -d && npm run benchmarks

What this does

If you're using new RegExp('foo') instead of a regex literal, it's probably because you need to dyamically generate a regex based on user options or some other potentially changing factors.

When your function creates a string based on user inputs and passes it to the RegExp constructor, regex-cache caches the results. The next time the function is called if the key of a cached regex matches the user input (or no input was given), the cached regex is returned, avoiding unnecessary runtime compilation.

Using the RegExp constructor offers a lot of flexibility, but the runtime compilation comes at a price - it's slow. Not specifically because of the call to the RegExp constructor, but because you have to build up the string before new RegExp() is even called.

Run tests

Install dev dependencies:

npm i -d && npm test

Contributing

Pull requests and stars are always welcome. For bugs and feature requests, please create an issue

Author

Jon Schlinkert

• github/jonschlinkert
• twitter/jonschlinkert

License

Copyright (c) 2015 Jon Schlinkert
Released under the MIT license

---

This file was generated by verb-cli on March 25, 2015.