Allow to configure blank node ids in toRDF

[nichtich]

Please support to configure how blank node ids are generated in toRDF (and possibly frameMergedOrDefault).

Function createMergedNodeMap supports to override default IdentifierIssuer (being new util.IdentifierIssuer('_:b')) with option issuer. Just change this line to

const issuer = options.issuer || new util.IdentifierIssuer('_:b');

or (more user-friendly) add an option bnode:

const issuer =  new util.IdentifierIssuer(`_:${options["bnode"]||"b"}`);

The problem

Blank node identifiers are repeated when serializing multiple RDF documents to N-Triples. For instance this two documents:

{"id":"x:1","x":{"title":"1"}}
{"id":"x:2","x":{"title":"2"}}

with additional @context

{
  "title": "http://purl.org/dc/terms/title",
  "id": "@id",
  "x": "http://example.org/x"
}

converted with

jsonld.toRDF(doc[0], {format: "application/n-quads"}),
jsonld.toRDF(doc[1], {format: "application/n-quads"}),

results in

<x:1> <http://example.org/x> _:b0 .                                                                                                                                                                                                              
_:b0 <http://purl.org/dc/terms/title> "1" .                                                                                                                                                                                                                 
<x:2> <http://example.org/x> _:b0 .                                                                                                                                                                                                              
_:b0 <http://purl.org/dc/terms/title> "2" .                                                                                                                                                                                                                 

Preferred solution

jsonld.toRDF(doc[0], {format: "application/n-quads", bnode: "a" }),
jsonld.toRDF(doc[1], {format: "application/n-quads", bnode: "b" }),

results in

<x:1> <http://example.org/x> _:a0 .                                                                                                                                                                                                              
_:a0 <http://purl.org/dc/terms/title> "1" .                                                                                                                                                                                                                 
<x:2> <http://example.org/x> _:b0 .                                                                                                                                                                                                              
_:b0 <http://purl.org/dc/terms/title> "2" .                                                                                                                                                                                                                 

[larsgw]

I (independently) ended up with this as well.

const issuer = options.issuer || new util.IdentifierIssuer('_:b');

I prefer this as I would not need to generate new prefixes, but can keep the auto-increment sequence going by reusing the issuer.

[nichtich]

I prefer this as I would not need to generate new prefixes

I don't mind as long as reuse of identifiers can be prevented. An additional option issuer should do.