Ts-regex-builder NPM

TS Regex Builder

Build maintainable regular expressions for TypeScript and JavaScript.

Goal

Regular expressions are a powerful tool for matching text patterns, yet they are notorious for their hard-to-parse syntax, especially in the case of more complex patterns.

This library allows users to create regular expressions in a structured way, making them easy to write and review. It provides a domain-specific langauge for defining regular expressions, which are finally turned into JavaScript-native RegExp objects for fast execution.

// Regular JS RegExp
const hexColor = /^#?([a-fA-F0-9]{6}|[a-fA-F0-9]{3})$/;

// TS Regex Builder DSL
const hexDigit = charClass(charRange('a', 'f'), charRange('A', 'F'), charRange('0', '9'));

const hexColor = buildRegExp([
  startOfString,
  optional('#'),
  capture(
    choiceOf(
      repeat(hexDigit, 6), // #rrggbb
      repeat(hexDigit, 3), // #rgb
    ),
  ),
  endOfString,
]);

Installation

npm install ts-regex-builder

yarn add ts-regex-builder

Basic usage

import { buildRegExp, capture, oneOrMore } from 'ts-regex-builder';

// /Hello (\w+)/
const regex = buildRegExp(['Hello ', capture(oneOrMore(word))]);

Regex domain-specific language

TS Regex Builder allows you to build complex regular expressions using domain-specific language.

Terminology:

regex construct (RegexConstruct) - common name for all regex constructs like character classes, quantifiers, and anchors.
regex element (RegexElement) - a fundamental building block of a regular expression, defined as either a regex construct, a string, or RegExp literal (/.../).
regex sequence (RegexSequence) - a sequence of regex elements forming a regular expression. For developer convenience, it also accepts a single element instead of an array.

Most of the regex constructs accept a regex sequence as their argument.

Examples of sequences:

single element (construct): capture('Hello')
single element (string): 'Hello'
single element (RegExp literal): /Hello/
array of elements: ['USD', oneOrMore(digit), /Hello/]

Regex constructs can be composed into a tree structure:

const currencyCode = repeat(charRange('A', 'Z'), 3);
const currencyAmount = buildRegExp([
  choiceOf('$', '€', currencyCode), // currency
  capture(
    oneOrMore(digit), // integer part
    optional(['.', repeat(digit, 2)]), // fractional part
  ),
]);

See Types API doc for more info.

Regex Builders

Builder	Regex Syntax	Description
`buildRegExp(...)`	`/.../`	Create `RegExp` instance
`buildRegExp(..., { ignoreCase: true })`	`/.../i`	Create `RegExp` instance with flags

See Builder API doc for more info.

Regex Constructs

Construct	Regex Syntax	Notes
`choiceOf(x, y, z)`	`x\\|y\\|z`	Match one of provided sequences
`capture(...)`	`(...)`	Create a capture group

See Constructs API doc for more info.

!NOTE TS Regex Builder does not have a construct for non-capturing groups. Such groups are implicitly added when required.

Quantifiers

Quantifier	Regex Syntax	Description
`zeroOrMore(x)`	`x*`	Zero or more occurence of a pattern
`oneOrMore(x)`	`x+`	One or more occurence of a pattern
`optional(x)`	`x?`	Zero or one occurence of a pattern
`repeat(x, n)`	`x{n}`	Pattern repeats exact number of times
`repeat(x, { min: n, })`	`x{n,}`	Pattern repeats at least given number of times
`repeat(x, { min: n, max: n2 })`	`x{n1,n2}`	Pattern repeats between n1 and n2 number of times

See Quantifiers API doc for more info.

Character classes

Character class	Regex Syntax	Description
`any`	`.`	Any character
`word`	`\w`	Word character: letter, digit, underscore
`digit`	`\d`	Digit character: 0 to 9
`whitespace`	`\s`	Whitespace character: space, tab, line break, ...
`anyOf('abc')`	`[abc]`	Any of provided characters
`charRange('a', 'z')`	`[a-z]`	Character in a range
`charClass(...)`	`[...]`	Union of multiple character classes
`negated(...)`	`[^...]`	Negation of a given character class

See Character Classes API doc for more info.

Assertions

Assertion	Regex Syntax	Description
`startOfString`	`^`	Match the start of the string (or the start of a line in multiline mode)
`endOfString`	`$`	Match the end of the string (or the end of a line in multiline mode)
`wordBoundary`	`\b`	Match the start or end of a word without consuming characters
`lookahead(...)`	`(?=...)`	Match subsequent text without consuming it
`negativeLookhead(...)`	`(?!...)`	Reject subsequent text without consuming it
`lookbehind(...)`	`(?<=...)`	Match preceding text without consuming it
`negativeLookbehind(...)`	`(?<!...)`	Reject preceding text without consuming it

See Assertions API doc for more info.

Examples

See Examples.

Performance

Regular expressions created with this library are executed at runtime, so you should avoid creating them in a context where they would need to be executed multiple times, e.g., inside loops or functions. We recommend that you create a top-level object for each required regex.

Contributing

See the contributing guide to learn how to contribute to the repository and the development workflow. See the project guidelines to understand our core principles.