Replace Enola "hack" with "proper" (full) RDF Dataset Canonicalization

[vorburger]

Replying to @filip26 on filip26/ld-cli#71 today reminded me of (my) RdfCanonicalizer, and how that was a "quick hack", but not yet a "real" (full) implementation of https://github.com/w3c/rdf-canon/.

One can still already use it as-is via Enola's CLI as https://docs.enola.dev/use/canonicalize/ though. I originally wrote it, and currently actively use it, in the ModelSubject utility (which is a sort of "Matcher" for unit testing, for Truth).

But perhaps one day, especially if in the future it might also be used for (H)MAC hashing for "security" related ideas, or in the context of some things I'm dreaming about (but unfortunately don't really have much time for exploring, at my current day job workload...) e.g. in the #777 (IPFS IPLD & Co.) context, it would make sense to replace that original initial first quick stab with a "real" (full) implementation of https://github.com/w3c/rdf-canon/, one day.

Looking at https://w3c.github.io/rdf-canon/reports/, the candidate (Java) libraries seem to be:

1. https://project.inria.fr/corese/ - probably much too huge?
2. https://github.com/setl/rdf-urdna - looks better and somewhat more self contained already
3. https://github.com/filip26/titanium-rdfc - not listed on rdf-canon/reports (yet?), but maybe the way to go?
4. Isn't there something alreaddy directly using the RDF4j API? I've created rdf-canon (RDF Dataset Canonicalization) support eclipse-rdf4j/rdf4j#5255 about it.

[vorburger]

Looking at https://w3c.github.io/rdf-canon/reports/, the candidate (Java) libraries seem to be:

Next step could be to explore, one day (not a high priority for me right now), just locally downstream in Enola, before further engaging upstream, if RdfCanonicalizer really needs the full https://github.com/filip26/titanium-json-ld/, or could be reduced to only require com.apicatalog.rdf.api.RdfQuadConsumer (as-is?), and then engage on apache/jena#2216 to see if proposing a separately released com.apicatalog:rdf-api sort of artefact would make any sense.

[filip26]

if RdfCanonicalizer really needs the full https://github.com/filip26/titanium-json-ld/, or could be reduced to only require com.apicatalog.rdf.api.RdfQuadConsumer (as-is?)

The next version, 2.0, which I hope will be released in a week or two, should depend only on RDF artifacts that are now part of Titanium (they will be removed from the next Titanium release). It is primarily used to obtain an N-Quads serialized representation, which is part of the canonicalization process. The long-term goal is to have it as self-contained as possible with minimal dependencies. However, further separation of artifacts beyond RdfQuadConsumer is more complex. The separation process will be incremental.