KRR-Oxford / RSAComb

Re-implementation of the combined approach for CQ answering over RSA ontologies
https://www.cs.ox.ac.uk/isg/tools/RSAComb
Apache License 2.0
2 stars 0 forks source link

Unsound cases in PAGOdA #12

Open dyamon opened 3 years ago

dyamon commented 3 years ago

The following queries return an empty set of answers in PAGOdA. RSAComb on the other hand is able to find the right set of answer. It might be possible that we are using an unsound version of PAGOdA.

Query 15 has 39 answers but none is found by PAGOdA:

https://github.com/KRR-Oxford/RSAComb/blob/2a3b5153955208dd4a9393fbaff57b3858efe271/tests/lubm/queries.sparql#L143-L151

Query 16 has 1 answer but it's not returned by PAGOdA:

https://github.com/KRR-Oxford/RSAComb/blob/2a3b5153955208dd4a9393fbaff57b3858efe271/tests/lubm/queries.sparql#L153-L162

dyamon commented 2 years ago

Another query that doesn't seem to return the right amount of answers is the following:

PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX ub: <http://www.lehigh.edu/~zhp2/2004/0401/univ-bench.owl#>
SELECT ?X ?Y
WHERE {
  ?X ub:memberOf ?Z .
  ?Y ub:memberOf ?Z .
  ?Z rdf:type ub:ResearchGroup .
  ?X ub:memberOf <http://www.Department0.University0.edu> .
  ?Y ub:memberOf <http://www.Department0.University0.edu> .
}

The number of answers is 39, but PAGOdA finds none.

dyamon commented 2 years ago

This is most likely due to which semantics we use to interpret the answers. Still PAGOdA should be able to guarantee certain answers semantics for internalisable queries.