Parse DNS domain names using the public suffix list from https://publicsuffix.org

The public suffix list distinguishes registrable domains from the "suffix" of the domain. It was originally defined by Mozilla to make sure that websites could not set cookies on suffixes. It is also useful for normalizing domains into registrable name, suffix and any subdomains.

For simple TLDs like .com, it's clear which part is the name, but there are many compound suffixes like .ac.uk, and it becomes difficult to tell what part is the name.

Various countries have special subdomains that act like top-level domains, to express geographic hierarchy, e.g., shibuya.tokyo.jp.

This library reads the standard public_suffix_list.dat and generates a parse/1 function which matches domains against it, splitting them into host/subdomains, name, and suffix:

iex(1)> PublicSuffixList.parse("www.ic.ac.uk")
{:ok, {["www"], "ic", "ac.uk"}}

It also defines utility functions, normalize/1 which removes the subdomains and name/1 which gets just the domain, and is_suffix?/1 which checks if a domain is a suffix:

iex(2)> PublicSuffixList.normalize("www.ic.ac.uk")
{:ok, "ic.ac.uk"}

iex(3)> PublicSuffixList.name("www.ic.ac.uk")
{:ok, "ic"}

iex(4)> PublicSuffixList.is_suffix("ac.uk")
true

Add public_suffix_list to the list of dependencies in mix.exs:

def deps do
  [
    {:public_suffix_list, "~> 0.7.1"}
  ]
end

The docs can be found at https://hexdocs.pm/public_suffix_list.