regexparam

A tiny (394B) utility that converts route patterns into RegExp. Limited alternative to path-to-regexp 🙇

With regexparam, you may turn a pathing string (eg, /users/:id) into a regular expression.

An object with shape of { keys, pattern } is returned, where pattern is the RegExp and keys is an array of your parameter name(s) in the order that they appeared.

Unlike path-to-regexp, this module does not create a keys dictionary, nor mutate an existing variable. Also, this only ships a parser, which only accept strings. Similarly, and most importantly, regexparam only handles basic pathing operators:

• Static (/foo, /foo/bar)
• Parameter (/:title, /books/:title, /books/:genre/:title)
• Parameter w/ Suffix (/movies/:title.mp4, /movies/:title.(mp4|mov))
• Optional Parameters (/:title?, /books/:title?, /books/:genre/:title?)
• Wildcards (*, /books/*, /books/:genre/*)

This module exposes three module definitions:

• CommonJS: dist/index.js
• ESModule: dist/index.mjs
• UMD: dist/index.min.js

Install

$ npm install --save regexparam

Usage

1import { parse, inject } from 'regexparam';
2
3// Example param-assignment
4function exec(path, result) {
5  let i=0, out={};
6  let matches = result.pattern.exec(path);
7  while (i < result.keys.length) {
8    out[ result.keys[i] ] = matches[++i] || null;
9  }
10  return out;
11}
12
13
14// Parameter, with Optional Parameter
15// ---
16let foo = parse('/books/:genre/:title?')
17// foo.pattern => /^\/books\/([^\/]+?)(?:\/([^\/]+?))?\/?$/i
18// foo.keys => ['genre', 'title']
19
20foo.pattern.test('/books/horror'); //=> true
21foo.pattern.test('/books/horror/goosebumps'); //=> true
22
23exec('/books/horror', foo);
24//=> { genre: 'horror', title: null }
25
26exec('/books/horror/goosebumps', foo);
27//=> { genre: 'horror', title: 'goosebumps' }
28
29
30// Parameter, with suffix
31// ---
32let bar = parse('/movies/:title.(mp4|mov)');
33// bar.pattern => /^\/movies\/([^\/]+?)\.(mp4|mov)\/?$/i
34// bar.keys => ['title']
35
36bar.pattern.test('/movies/narnia'); //=> false
37bar.pattern.test('/movies/narnia.mp3'); //=> false
38bar.pattern.test('/movies/narnia.mp4'); //=> true
39
40exec('/movies/narnia.mp4', bar);
41//=> { title: 'narnia' }
42
43
44// Wildcard
45// ---
46let baz = parse('users/*');
47// baz.pattern => /^\/users\/(.*)\/?$/i
48// baz.keys => ['wild']
49
50baz.pattern.test('/users'); //=> false
51baz.pattern.test('/users/lukeed'); //=> true
52
53exec('/users/lukeed/repos/new', baz);
54//=> { wild: 'lukeed/repos/new' }
55
56
57// Injecting
58// ---
59
60inject('/users/:id', {
61  id: 'lukeed'
62}); //=> '/users/lukeed'
63
64inject('/movies/:title.mp4', {
65  title: 'narnia'
66}); //=> '/movies/narnia.mp4'
67
68inject('/:foo/:bar?/:baz?', {
69  foo: 'aaa'
70}); //=> '/aaa'
71
72inject('/:foo/:bar?/:baz?', {
73  foo: 'aaa',
74  baz: 'ccc'
75}); //=> '/aaa/ccc'
76
77inject('/posts/:slug/*', {
78  slug: 'hello',
79}); //=> '/posts/hello'
80
81inject('/posts/:slug/*', {
82  slug: 'hello',
83  wild: 'x/y/z',
84}); //=> '/posts/hello/x/y/z'
85
86// Missing non-optional value
87// ~> keeps the pattern in output
88inject('/hello/:world', {
89  abc: 123
90}); //=> '/hello/:world'

Important: When matching/testing against a generated RegExp, your path must begin with a leading slash ("/")!

Regular Expressions

For fine-tuned control, you may pass a RegExp value directly to regexparam as its only parameter.

In these situations, regexparam does not parse nor manipulate your pattern in any way! Because of this, regexparam has no "insight" on your route, and instead trusts your input fully. In code, this means that the return value's keys is always equal to false and the pattern is identical to your input value.

This also means that you must manage and parse your own keys~!
You may use named capture groups or traverse the matched segments manually the "old-fashioned" way:

Important: Please check your target browsers' and target Node.js runtimes' support!

1// Named capture group
2const named = regexparam.parse(/^\/posts[/](?<year>[0-9]{4})[/](?<month>[0-9]{2})[/](?<title>[^\/]+)/i);
3const { groups } = named.pattern.exec('/posts/2019/05/hello-world');
4console.log(groups);
5//=> { year: '2019', month: '05', title: 'hello-world' }
6
7// Widely supported / "Old-fashioned"
8const named = regexparam.parse(/^\/posts[/]([0-9]{4})[/]([0-9]{2})[/]([^\/]+)/i);
9const [url, year, month, title] = named.pattern.exec('/posts/2019/05/hello-world');
10console.log(year, month, title);
11//=> 2019 05 hello-world

API

regexparam.parse(input: RegExp)

regexparam.parse(input: string, loose?: boolean)

Returns: Object

Parse a route pattern into an equivalent RegExp pattern. Also collects the names of pattern's parameters as a keys array. An input that's already a RegExp is kept as is, and regexparam makes no additional insights.

Returns a { keys, pattern } object, where pattern is always a RegExp instance and keys is either false or a list of extracted parameter names.

Important: The keys will always be false when input is a RegExp and it will always be an Array when input is a string.

input

Type: string or RegExp

When input is a string, it's treated as a route pattern and an equivalent RegExp is generated.

Note: It does not matter if input strings begin with a / — it will be added if missing.

When input is a RegExp, it will be used as is – no modifications will be made.

loose

Type: boolean
Default: false

Should the RegExp match URLs that are longer than the str pattern itself?
By default, the generated RegExp will test that the URL begins and ends with the pattern.

Important: When input is a RegExp, the loose argument is ignored!

1const { parse } = require('regexparam');
2
3parse('/users').pattern.test('/users/lukeed'); //=> false
4parse('/users', true).pattern.test('/users/lukeed'); //=> true
5
6parse('/users/:name').pattern.test('/users/lukeed/repos'); //=> false
7parse('/users/:name', true).pattern.test('/users/lukeed/repos'); //=> true

regexparam.inject(pattern: string, values: object)

Returns: string

Returns a new string by replacing the pattern segments/parameters with their matching values.

Important: Named segments (eg, /:name) that do not have a values match will be kept in the output. This is true except for optional segments (eg, /:name?) and wildcard segments (eg, /*).

pattern

Type: string

The route pattern that to receive injections.

values

Type: Record<string, string>

The values to be injected. The keys within values must match the pattern's segments in order to be replaced.

Note: To replace a wildcard segment (eg, /*), define a values.wild key.

Deno

As of version 1.3.0, you may use regexparam with Deno. These options are all valid:

1// The official Deno registry:
2import regexparam from 'https://deno.land/x/regexparam/src/index.js';
3// Third-party CDNs with ESM support:
4import regexparam from 'https://cdn.skypack.dev/regexparam';
5import regexparam from 'https://esm.sh/regexparam';

Note: All registries support versioned URLs, if desired.
The above examples always resolve to the latest published version.

Related

• trouter - A server-side HTTP router that extends from this module.
• matchit - Similar (650B) library, but relies on String comparison instead of RegExps.

License

MIT © Luke Edwards