Remove Testcase ontology

[gaurav]

We use some RDF properties in the testcase: namespace to tie the JSON-LD file together (since keys in JSON-LD files without an associated RDF property will not be read as RDF). However, these properties don't actually need to be defined anywhere, certainly not in the Phyloref Testcase Ontology that we created ages ago as a repository for these terms. This PR removes the Phyloref Testcase Ontology from the OWL imports in generated phyloreferences.

It also removes many of the uses of the testcase: namespace, including the no-longer-used taxonomic unit matching system. Three of these terms could be replaced with CDAO terms, so I've replaced these in this PR:

• testcase:PhyloreferenceTestPhylogeny corresponds to a CDAO:RootedTree.
• testcase:has_root_node corresponds to CDAO:has_Root.
• testcase:has_node corresponds to CDAO:has_Element (confusingly, it does NOT correspond to CDAO:has_Node, whose domain is Edges, not Trees).

As of #78 (comment), there are five testcase: terms remaining in Phyx.js:

• Needed to tie the generated OWL/JSON-LD file together:
  • testcase:has_phylogeny, in the form <ontology> testcase:has_phylogeny <phylogeny>.
  • testcase:has_phyloreference, in the form <ontology> testcase:has_phyloreference <phyloref>.
  • testcase:has_component_class, in the form <phyloref> testcase:has_component_class <component class>.
• Optional; only needed if we want to keep the verbatim specifiers in the generated OWL/JSON-LD file:
  • testcase:internal_specifier, in the form <phyloref> testcase:internal_specifier <internal specifier>.
  • testcase:external_specifier, in the form <phyloref> testcase:external_specifier <external specifier>.

We could figure out what to do with these terms in this PR, or leave that for a separate PR.

[hlapp]

See inline comment. I'm also wondering whether CDAO doesn't simply need to be fixed so as not to assert an edge as the domain of has_Node.

[gaurav]

I'm also wondering whether CDAO doesn't simply need to be fixed so as not to assert an edge as the domain of has_Node.

I don't think so, since has_Node (and its subproperties, has_Parent_Node and has_Child_Node) are designed specifically for use with edges, rather than the node-based has_Parent/has_Child. But I've opened an issue to discuss this on CDAO: evoinfo/cdao#17.

[hlapp]

👍🏻