Documentation: Fcrepo-transform is dead

[whikloj]

Need to refactor or redesign the work done in #185.

Based on the MVP's statement about an ID Mapper service, perhaps this will not be the triplestore.

If we want to maintain the current workflow I would dump all the work I did to get the fcrepo-transform to load and instead retrieve the full RDF as JSON-LD, parse out the UUID and store it.

[acoburn]

@whikloj you may also want to consider using https://github.com/fcrepo4-exts/fcrepo-camel-toolbox/tree/master/fcrepo-ldpath, which is the replacement for fcrepo-transform.

[whikloj]

@acoburn I did think about that, and I will look into it. I had such a fun time getting the original ldpath document to work I thought for what we need out of Fedora, which is just the path and the Islandora UUID and only during transactions. Perhaps I was over thinking it.

[whikloj]

Okay so my understanding is that to use the fcrepo-ldpath service I have to make my LDPath web-accessible. So I could add an endpoint to the Crayfish service what just replies to a GET request with this ldpath document. Then I can remove all of this.

Does that sounds like a good idea? Or should I go to simply with requesting the document and parsing it for the UUID.

[DiegoPino]

@whikloj it sounds like a good idea: but my pragmatic side says: "request the document, parse it for the UUID". Once we have all working again, we can make it as elegant as it was with the defunct transform.

[acoburn]

request the document, parse it for the UUID

isn't that what fcrepo-ldpath does with the transform here: https://github.com/Islandora-CLAW/Crayfish/blob/master/src/TransactionService/resources/islandora_uuid.txt?

[whikloj]

@acoburn it does, but the question is should we built Crayfish to require fcrepo-ldpath to get this information or should I just GET the resource as (say) JSON-LD and parse out the part myself?

I'm torn, the work is already there to use the transformed data, so pivoting from fcrepo-transform to fcrepo-ldpath makes the most sense. Unless we stop using the triplestore as our main F4 path -> UUID key store.

[acoburn]

curl http://repo.example.org/rest/path/to/resource -H"Accept: application/ld+json; profile=\"http://www.w3.org/ns/json-ld#compacted\"" is pretty easy too.

[whikloj]

Agreed. I could go either way. The only reason not to use fcrepo-ldpath is if we want to avoid the dependency. But we are already relying on fcrepo-camel-indexing-triplestore anyways.

[acoburn]

The big win with fcrepo-ldpath is for dereferencing external URIs. Since that's not what you're doing here, I think you'd be better off just parsing the JSON-LD yourself.

[whikloj]

Resolved with https://github.com/Islandora-CLAW/Crayfish/commit/bbd0dfd0aa1a6a0ca0bf6f743c3e07ec29daf4f7