When converting nq to ntriples with context, nabu assumes the context is shared

[C-Loftus]

Currently (comes from the previous upstream code) conversion is done in such a way that the context of the first quad is returned as the context for the full conversion. As noted by the TODO, this is error prone

        // loop on tr and make a set of triples
	triples := []rdf.Triple{}
	for _, quad := range quads {
		triples = append(triples, quad.Triple)
	}

	// Assume context of first triple sis context of all triples
	// TODO..   this is stupid if not dangers, at least return []string of all the contexts
	// that were in the graph.
	ctx := quads[0].Ctx
	quadContext := ctx.String()

[C-Loftus]

this seems to actually be ok. as of right now this function is only called once at nTriples, _, err = common.QuadsToTripleWithCtx(string(rawJsonldOrNqBytes)) and the context graph name is ignored. So for the time being, this behavior is not even being used.

rather instead when nabu upserts graph data it uses the URN that is generated from the prefix as it does here:

func (synchronizer *SynchronizerClient) upsertDataForGraph(rawJsonldOrNqBytes []byte, objectName string) error {

	graphResourceIdentifier, err := common.MakeURN(objectName)