Adding list of actual "registered" domains in result data

[TytoCapensis]

Currently, when analyzing an email with eml_parser, the domains appearing in the body of the email are given in the output, like this:

"domain": [
	"b2b.parallels.com",
	"click.parallels.com",
	"coronavirus.data.gov.uk"
],

However, a possible issue here is that the full domains are listed, including the subdomains part. This can make the identification of entities and actors complicated if a lot of subdomains are present in the domain table.

This commit takes the opportunity to use publicsuffixlist (already used in eml_parser) to add a table named domain_registered in the data returned by an eml_parser analysis.

The domains in domain_registered are the true registered domains, i.e. the "closest" domains to the TLD. Thanks to publicsuffixlist, public suffixes like co.uk or co.jp can be taken into consideration.

Now, the output looks like this:

"domain": [
	"b2b.parallels.com",
	"click.parallels.com",
	"coronavirus.data.gov.uk"
],
"domain_registered": [
	"parallels.com",
	"data.gov.uk"
],

Do not hesitate to suggest any improvements (especially regarding the name of the table)