jfromaniello / li

JavaScript utility to parse and generate the Link header according to RFC 5988.

Parse and format Link header according to RFC 5988.

Install

$ npm install li

Also works with bower, component.js, browserify, amd, etc.

Usage

Parse a Link header:

var li = require('li');
var someLinksHeader = '</api/users?page=0&per_page=2>; rel="first", ' +
                      '</api/users?page=1&per_page=2>; rel="next", ' +
                      '</api/users?page=3&per_page=2>; rel="last"';

console.log(li.parse(someLinksHeader));

// This will print:
// {
//   first: '/api/users?page=0&per_page=2',
//   next: '/api/users?page=1&per_page=2',
//   last: '/api/users?page=3&per_page=2'
// }

Generate a Link header as follow with stringify:

var linksObject = {
  first : '/api/users?page=0&per_page=2',
  next  : '/api/users?page=1&per_page=2',
  last  : '/api/users?page=3&per_page=2',
};

console.log(li.stringify(linksObject);

// This will print the string:
// </api/users?page=0&per_page=2>; rel="first",
// </api/users?page=1&per_page=2>; rel="next",
// </api/users?page=3&per_page=2>; rel="last"

Testing

$ npm test

License

MIT 2014 - JOSE F. ROMANIELLO

li's Issues

Add to bower

Doesn't parse link with no rel

Similarly to #6, the library fails to parse a link header where on link is without a rel parameter:

li.parse('</3>; rel="next", </2>; rel="prev", </home>, </void>; rel="ignored"')

Doesn't include the rel="ignored" link. It simply stops once it get's to </home>

When multiple links have the same rel, .parse should give all of them as an array.

Currently, the last link wins.

Doesn't parse when there are additional link parameters

li.parse doesn't return complete output when there are additional link parameters.

It seems that the library parses links until it encounters one with extra params and then ignores any that follow.

li.parse('</3>; rel="next", </2>; rel="prev", </home>; rel="up"; rev="home", </void>; rel="ignored"')

returns

{
  "next": "/3",
  "prev": "/2",
  "home": "/home"
}

What's more, if you reverse the params:

li.parse('</home>; rev="home"; rel="up"')

it will return an empty object.

Readme: Explain .parse behavior for false-y values

Hi! Thanks for this module.
Currently it seems that .parse() will simply return an empty object if given a false-y argument (including null, undefined, …). This robustness is good and desirable behavior. Would be nice if you could note that in the readme so that people can know even before testing that this module is very robust.

[Feature] Regroup links with same URI

First, thanks for your work. This little project makes my life much easier.

This is a feature suggestion.

For example:

li.parse('<my_uri>; rel="first last"')

returns:

{ first: 'my_uri', last: 'my_uri' }

but:

li.stringify({ first: 'my_uri', last: 'my_uri' })

returns:

'<my_uri>; rel="first", <my_uri>; rel="last"'

Regrouping links by URI is a best practice for Link headers and would result instead in:

'<my_uri>; rel="first last"'