Add phyloref type to phyloreferences

[gaurav]

This PR adds a phyloreference type (phylorefType) to phyloreferences as defined in #6. The value of phylorefType must be one of the following classes from the Phyloref ontology:

• phyloref:PhyloreferenceUsingMinimumClade for minimum-clade phylorefs.
• phyloref:PhyloreferenceUsingMaximumClade for maximum-clade phylorefs.
• phyloref:PhyloreferenceUsingApomorphy for apomorphy-based phylorefs.

Separately, we calculate the phyloreference type while we are generating the logical expression (based on the internal and external specifiers), and make the phyloreference a subclass of that class. If the provided phylorefType differs from the computed phyloref type, we will throw an error and refuse to create this phyloreference.

Closes #6.

[hlapp]

I assume you know that the phyloref:Phyloreference subclass relationship is implied by the more specific types, and so enumerating them in OWL is redundant (and thus counter to good practice). I'm not sure whether you're accounting for this anyway in the conversion to OWL. I do, however, see the redundancy in the .nq files, and even RDF reasoners can do subclass reasoning, so I'm not sure.

[gaurav]

I assume you know that the phyloref:Phyloreference subclass relationship is implied by the more specific types, and so enumerating them in OWL is redundant (and thus counter to good practice). I'm not sure whether you're accounting for this anyway in the conversion to OWL. I do, however, see the redundancy in the .nq files, and even RDF reasoners can do subclass reasoning, so I'm not sure.

Exactly right. I tried removing this direct subclass relationship, but that caused a problem in JPhyloRef (phyloref/jphyloref#74). I'm not sure what the problem is there, but I suspect it's something to do with the way that we load ontologies. Once I fix that, I'll remove the direct subclass relationship here (#74). I don't think this is very urgent, so I'll work on it after we finish the Phyx.js manuscript.