. init.sh
pip install -r requirements.txt

make test

Currently this will

• check syntax of shex file is correct
• ensure all in should_pass actually pass
• ensure all in should_fail actually fail

python ./gocam_validator.py ../test_ttl/go_cams/should_pass/test2.ttl

Test suite. just run CLI for now.

This does not rely on pre-injection of categories. Instead the pattern is:

<MolecularFunctionClass> @<OwlClass> AND {
  rdfs:subClassOf [ GoMolecularFunction: ] ;
}

<MolecularFunction> @<GoCamEntity> AND EXTRA a {
  a @<MolecularFunctionClass> {1};
  enabled_by:  ( @<Protein> OR @<Complex> ) {0,1};

The validator code will inject ?c rdfs:subClassOf ?a triples using a SPARQL subClassOf* query on the triplestore as a poor-man's reasoning step. This is cached using cachier (we may want a different solution for production but this is fine for now).

For every ?i rdf:type ?c we find the ancestors, and then assert that ?c is an ancestor (reflexive) if that ancestor is in the set of upper level classes mapped to a shape.

currently this is in the code header:

cmap = {
    "http://purl.obolibrary.org/obo/GO_0003674" : "MolecularFunction",
    "http://purl.obolibrary.org/obo/GO_0008150" : "BiologicalProcess",
    "http://purl.obolibrary.org/obo/GO_0005575" : "CellularComponent"
}

but should obviously be moved out somewhere more declarative.