Skip to content

Latest commit

 

History

History
240 lines (182 loc) · 9.05 KB

TestCaseIssues.md

File metadata and controls

240 lines (182 loc) · 9.05 KB
Raw

Issues and problems with the OwL 2 test cases

General Issues

  1. Should we emit owl:Thing a owl:Class? The spec doesn't appear to say we shouldn't, but the test cases themselves are inconsistent. Bnode2somevaluesfrom.func shows it as not being emitted, yet other tests do. For the time being, identifiers.py#43 does not emit any declarations about XSD, RDF, RDFS or OWL, and our tests remove them (with warnings) from the target files.

  2. Reified annotations: the Finland parser and, possibly, the test case itself replicates BNodes used as annotatedSource or annotatedTarget elements. Using FS2RDF/propertychain/2/annotated/ar.func as an example, we can either point at a copy of the subject and object in the annotation Axiom:

    <http://example.org/p> a :ObjectProperty ;
    :propertyChainAxiom ( <http://example.org/q> <http://example.org/r> ) .

[] a :Axiom ;
    rdfs:comment "I hereby annotate this" ;
    :annotatedProperty :propertyChainAxiom ;
    :annotatedSource <http://example.org/p> ;
    :annotatedTarget ( <http://example.org/q> <http://example.org/r> ) .

    or we can point at the items themselves:

    <http://example.org/p> a :ObjectProperty ;
    :propertyChainAxiom _:fb1b444ef721943fd96f22d4d6c53d9b1b1 .

[] a :Axiom ;
    rdfs:comment "I hereby annotate this" ;
    :annotatedProperty :propertyChainAxiom ;
    :annotatedSource <http://example.org/p> ;
    :annotatedTarget _:fb1b444ef721943fd96f22d4d6c53d9b1b1 .

_:fb1b444ef721943fd96f22d4d6c53d9b1b2 rdf:first <http://example.org/r> ;
    rdf:rest () .

_:fb1b444ef721943fd96f22d4d6c53d9b1b1 rdf:first <http://example.org/q> ;
    rdf:rest _:fb1b444ef721943fd96f22d4d6c53d9b1b2 .

    The first form is obviously more readable, but it is inconsistent but will be a real challenge to process, as one has to go on the assumption that _:b1 a rdf:Thing . _:b2 a rdf:Thing implies _:b1 == _:b2, something that I suspect runs counter to the RDF philosophy.

    Since the test cases use the first form, we have added an option USE_BNODE_COPIES to funowl/base/clone_subgraph.py to allow us to have it either way.

  3. sameIndividuals annotations The RDF spec seems to have a gap when it comes to how annotations apply to `sameIndividuals(Annotation(rdfs:comment "annot") :a :b :c :d). The Finnish converter seems to interpret this as

    [] a :Axiom ;
 rdfs:comment "annot" ;
 :annotatedProperty :sameAs ;
 :annotatedSource :a, :b, :c ;
 :annotatedTarget :b, :c, :d .

    Not sure quite what to think of this, but for the time being we're going to go with it.

  4. ObjectPropertyChain Does ObjectPropertyChain(:p1 :p2) imply :p1 a owl:ObjectProperty and :p2 a owl:ObjectProperyt? The generator currently emits them, but the test code (Rdfbased/sem/chain/def.func) does not show this as happening. For the moment we are going to assume that we are right and change the test targets:

    • Rdfbased/sem/chain/def.ttl
    • Owl2/rl/rules/ifp/askey.ttl
    • TestCase-WebOnt/I5.1/001.ttl
    • Rdfbased/sem/chain/def.ttl
    • Rdfbased/sem/key/def.ttl
    • Rdfbased/sem/ndis/alldisjointproperties/fw.ttl

  5. HasKey Does HasKey(ex:c () (ex:p1 ex:p2)) imply ex:p1 a owl:DatatypeProperyt? Same as ObjectPropertyChain.
    see: Rdfbased/sem/key/def.func and Rdfbased/sem/ndis/alldisjointproperties/fw.func

  6. Representation of quoted strings inside triple-quoted strings:

    • TestCase-WebOnt/I5.8/013.func
    • TestCase-WebOnt/I5.8/015.func

    Expected:

    <http://example.org/user/data#type> rdfs:comment """
This type maps the string "foo" to the number 3.
All other strings are not in the lexical space.
""" .

    FunOWL

    <http://example.org/user/data#type> rdfs:comment """
This type maps the string \\"foo\\" to the number 3.
All other strings are not in the lexical space.

""" .

As a short-term solution, we have added the escapes to the expected side of things

## Specific file issues

**Direct_Semantics_Literal_disjoint_from_Thing.func**

There is a conversion error on the Finland site, where `EquivalentClasses(owl:Thing ObjectOneOf(:b :a))` gets converted
into:
```text
owl:Thing owl:equivalentClass [ a owl:Class ;
          owl:oneOf ( example:a example:b ) ] .

FS2RDF/different/individuals/2/annotation/ar.func

The test case has two errors:

  1. DifferentIndividuals(Annotation(rdfs:comment "I hereby annotate this") :a :b)) should map to: 
    <http://example.org/a> a :NamedIndividual ;
    :differentFrom <http://example.org/b> .
    instead of allDifferent and
  2. The annotations were dropped.

FS2RDF/different/individuals/2/ar.func DifferentIndividuals(:a :b)) should map to: text <http://example.org/a> a :NamedIndividual ; :differentFrom <http://example.org/b> . instead of allDifferent

FS2RDF/disjoint/classes/2/annotation/ar.func

The annotation axiom was repeated twice. Error in Finland converter.

FS2RDF/domain/range/expression/ar.func

    Declaration(DataProperty(:p5))
    Declaration(Class(:c4))
    ObjectPropertyDomain(ObjectInverseOf(:p5) :c4)

InverseObjectProperty := 'ObjectInverseOf' '(' ObjectProperty ')'

The results of this test indicate that :p5 should be declared as both an ObjectProperty and a DataProperty. This is an outstanding issue in the parser

FS2RDF/equivalent/classes/2/annotation/ar.func

The annotation axiom was repeated twice. Error in Finland converter.

FS2RDF/equivalent/classes/3/annotation/ar.func

The annotation axiom was repeated twice. Error in Finland converter.

FS2RDF/literals/ar.func

Finland converter failed to return anything. Probably choked on some of the embedded XML

Python tooling is unable to differentiate between "1.0"^^<http://www.w3.org/2001/XMLSchema#decimal> and "1"^^<http://www.w3.org/2001/XMLSchema#decimal> . The test has temporarily been added to the skip list until we can determine

FS2RDF/same/individual/2/annotation/ar.func

The annotation axiom was repeated twice. Error in Finland converter.

FS2RDF/same/individual/3/annotation/ar.func

The annotation axiom was repeated multiple times. Error in Finland converter.

New/Feature/DisjointDataProperties/001.func TestCase-WebOnt/disjointWith/008.func TestCase-WebOnt/disjointWith/009.func

The Finland converter got the properties backward -- sb hasName disjointWith hasAddress Same with 008

New/Feature/Keys/002.func Validating New/Feature/ObjectQCR/001.func Validating New/Feature/ObjectQCR/002.func Owl2/rl/rules/fp/differentFrom.fun Owl2/rl/rules/ifp/differentFrom.func Rdfbased/sem/eqdis/different/sameas.func TestCase-WebOnt/description/logic/501.func TestCase-WebOnt/maxCardinality/001.func TestCase-WebOnt/differentFrom/001.func TestCase-WebOnt/description/logic/502.func

The Finland converter records a pair as allDifferent rather than differentFrom

Note that maxCardinality/001.func was missing the ontology declaration as well

New/Feature/AnnotationAnnotations/001.func

The annotation on the annotation wasn't emitted by the Finland processor

Rdfbased/sem/eqdis/disclass/irrflxv.func

Failed the Finland conversion completely. ttl added on our side

Rdfbased/sem/npa/dat/fw.func

This test uses a non-standard OWL property -- owl:targetValue . Our current implementation does not emit declarations about the OWL, RDF or other known namespaces, so this doesn't get emitted. We are ignoring this test for the time being.

TestCase-WebOnt/I5.26/004.func TestCase-WebOnt/I5.26/005.func

The Finnish converter reversed the order of the declaration going from XML to functional and then switched it again going to turtle. Edited the functional declaration to match the XML.

TestCase-WebOnt/I5.3/009.func Finnish conversion failed -- loaded expected with output of generator.

TestCase-WebOnt/miscellaneous/002.func Several implied declarations were missing -- not sure why.

TestCase-WebOnt/I5.8/017.func Uses document local URI's -- funowl processor doesn't work with fragments (e.g. "01"^^<#bar> .)

TestCase-WebOnt/I6.1/001.func The BNode elements of the test were missing completely.

TestCase-WebOnt/imports/001.func TestCase-WebOnt/imports/003.func TestCase-WebOnt/imports/006.func TestCase-WebOnt/imports/007.func TestCase-WebOnt/imports/011.func All of these curiously miss one triple -- either <http://www.w3.org/2002/03owlt/imports/support001-A#Man> a <http://www.w3.org/2002/07/owl#Class> . <http://www.w3.org/2002/03owlt/imports/imports006#c> a <http://www.w3.org/2002/07/owl#Class> . or <http://www.w3.org/2002/03owlt/imports/imports007#p> a <http://www.w3.org/2002/07/owl#ObjectProperty> .

TestCase-WebOnt/disjointWith/010.func Finnish conversion failed -- loaded with output of generator

FS2RDF/ontology/annotation/annotation/ar.func Finnish loader didn't get the annotations correct. Added our output