Cliqz Url Parser

A Fast implementation of url parsing, mostly API-compatible with the standard URL class while being on average 2-3 times faster. Evaluation of URL components is lazy, so this implementation should be fast for all read use-cases.

Known differences to standard URL:

• Parameters returned via URL.searchParams.entries() are decoded only with decodeURIComponent. This differs to standards parsing in some subtle ways.
• You can iterate a URL parameters array directly via URL.searchParams.params. This is around 20% faster than using an iterator.
• Parameter strings (; sepearated key/value pairs) are parsed, and accessible via URL.parameters.
• Domain parsing with tldts is built in. The URL.domainInfo attribute returns output from tldts' parseHost method.
• Hostname validation is not done on initial parse. The isValidHost() method is provided for this purpose.
• All URLs with a valid authority are given an origin, regardless of the protocol scheme. This differs from the standard that only does so for a set of known schemes.

Install

npm install @cliqz/url-parser

Usage

const parsed = new URL('https://www.example.com');
parsed.hostname // == 'www.example.com'

Performance

We benchmark against a list of 250,000 URLs collected from popular sites, as previously used in our adblocker benchmark. We compare two use-cases:

1. URL object creation: new URL(url)
2. Query string parsing: new URL(url).searchParams.entries()

We compare to a reference implementation on each platform:

• Node: URL class from the url library
• Firefox: window.URL
• Chrome: window.URL
• Safari: window.URL

All benchmarks were run on a Mid 2015 Macbook Pro, 2.5 GHz Intel Core i7, 16GB.

License

Mozilla Public License 2.0