Extrapolate or replace the HTML metadata triplifier

[luigi-asprino]

The HTML metadata triplifier is based on a mirror of Any23 (see #392).
Many dependency conflicts originate from this, and the mirrored code (see #590) has never integrated well with the rest of the project (for example, SA is based on Jena, whereas Any23 relies on RDF4J).
Furthermore, the HTML module is the heaviest in the project due to the long list of dependencies that SA inherited from Any23 (see #484).
One solution would be to replace it with a new HTML metadata triplifier, but I am not an HTML microdata expert and I am unaware of any alternatives to Any23 in Java.
Another solution would be to let the metadata triplifier live in a spin-off project and replacing/upgrading it separately from the main branch.
I would prefer the second solution, but I am open to discussion.

[luigi-asprino]

Better to convert it to a discussion for now, and then put it back as an issue once there is a clear plan.

[VladimirAlexiev]

@luigi-asprino do you mean RDF ( microdata and rdfa) extraction?
Then I'd add JSONLD and Turtle extraction from embedded "script" elements. This is imho a modern way