ConjunctiveGraph.parse with BNode as publicID

[uholzer]

This is weird:

from rdflib import ConjunctiveGraph, BNode, URIRef
g = ConjunctiveGraph()
b = BNode()
data = '<d:d> <e:e> <f:f> .'
print ("Parsing '{}' into {}".format(data, repr(b)))
g.parse(data=data, format='turtle', publicID=b)

triples = list( g.get_context(b).triples((None,None,None)) )
if triples:
    print("Good: Found in graph {}: {}".format(repr(b), triples))
else:
    print("Bad: No triples found in graph {}".format(repr(b)))

u = URIRef(b)

triples = list( g.get_context(u).triples((None,None,None)) )
if triples:
    print("Bad: Found in graph {}: {}".format(repr(u), triples))
else:
    print("Good: No triples found in graph {}".format(repr(u)))

output:

Parsing '<d:d> <e:e> <f:f> .' into rdflib.term.BNode('N9f934fec82174ee2bd5555be5eadcde7')
Bad: No triples found in graph rdflib.term.BNode('N9f934fec82174ee2bd5555be5eadcde7')
Bad: Found in graph rdflib.term.URIRef(u'N9f934fec82174ee2bd5555be5eadcde7'): [(rdflib.term.URIRef(u'd:d'), rdflib.term.URIRef(u'e:e'), rdflib.term.URIRef(u'f:f'))]

The problem here is that ConjunctiveGraph.parse always makes a URIRef out of publicID:

    def parse(self, source=None, publicID=None, format="xml",
              location=None, file=None, data=None, **args):
        source = create_input_source(
            source=source, publicID=publicID, location=location,
            file=file, data=data, format=format)
        g_id = URIRef(publicID and publicID or source.getPublicId())
        ...

[joernhees]

Interestingly enough the default graph has a BNode as identifier ;)

I guess you're right g_id should be publicID if that is a URIRef or BNode.
If neither should it try to parse n3?

[uholzer]

I think it should parse n3 if and only if Graph.__init__ does so with identifier. I am not aware of any part of the API that does parse n3 and I think it would add another source of bugs.