regexgen

regexgen

Generates regular expressions that match a set of strings.

Installation

regexgen can be installed using npm:

npm install regexgen

Example

The simplest use is to simply pass an array of strings to regexgen:

const regexgen = require('regexgen');

regexgen(['foobar', 'foobaz', 'foozap', 'fooza']); // => /foo(?:zap?|ba[rz])/

You can also use the Trie class directly:

const {Trie} = require('regexgen');

let t = new Trie;
t.add('foobar');
t.add('foobaz');

t.toRegExp(); // => /fooba[rz]/

CLI

regexgen also has a simple CLI to generate regexes using inputs from the command line.

$ regexgen
Usage: regexgen [-gimuy] string1 string2 string3...

The optional first parameter is the flags to add to the regex (e.g. -i for a case insensitive match).

ES2015 and Unicode

By default regexgen will output a standard JavaScript regular expression, with Unicode codepoints converted into UCS-2 surrogate pairs.

If desired, you can request an ES2015-compatible Unicode regular expression by supplying the -u flag, which results in those codepoints being retained.

$ regexgen 👩 👩‍💻 👩🏻‍💻 👩🏼‍💻 👩🏽‍💻 👩🏾‍💻 👩🏿‍💻
/\uD83D\uDC69(?:(?:\uD83C[\uDFFB-\uDFFF])?\u200D\uD83D\uDCBB)?/

$ regexgen -u 👩 👩‍💻 👩🏻‍💻 👩🏼‍💻 👩🏽‍💻 👩🏾‍💻 👩🏿‍💻
/\u{1F469}(?:[\u{1F3FB}-\u{1F3FF}]?\u200D\u{1F4BB})?/u

Such regular expressions are compatible with current versions of Node, as well as the latest browsers, and may be more transferrable to other languages.

How does it work?

• Generate a Trie containing all of the input strings. This is a tree structure where each edge represents a single character. This removes redundancies at the start of the strings, but common branches further down are not merged.

• A trie can be seen as a tree-shaped deterministic finite automaton (DFA), so DFA algorithms can be applied. In this case, we apply Hopcroft's DFA minimization algorithm to merge the nondistinguishable states.

• Convert the resulting minimized DFA to a regular expression. This is done using Brzozowski's algebraic method, which is quite elegant. It expresses the DFA as a system of equations which can be solved for a resulting regex. Along the way, some additional optimizations are made, such as hoisting common substrings out of an alternation, and using character class ranges. This produces an an Abstract Syntax Tree (AST) for the regex, which is then converted to a string and compiled to a JavaScript RegExp object.

License

MIT