Closed fcbr closed 5 years ago
It looks like it was something related to encoding; I changed the server encoding to UTF8 and it showed the explanation. I'll close the issue.
Oops, it wasn't the encoding, but UNA. With UNA set to true I get an explanation; otherwise I get NIL.
Hi Fabricio,
If you send me you abox (or a subset of it that shows the error) and the respective calls that produce the error, I can certain try to fix it.
Best regards, Ralf
No sure I understood this one:
CL-USER> (check-abox-coherence) (NIL NIL) Shouldn't the second argument be one explanation for the inconsistency? Is this due to some out of memory condition? (My ontology contains a couple of million triples.) Or am I missing some initialization steps?
It looks like it was something related to encoding; I changed the server encoding to UTF8 and it showed the explanation. I'll close the issue.
On 23.08.2015, at 15:31, Fabricio Chalub notifications@github.com wrote:
Oops, it wasn't the encoding, but UNA. With UNA set to true I get an explanation; otherwise I get NIL.
— Reply to this email directly or view it on GitHub https://github.com/ha-mo-we/Racer/issues/7#issuecomment-133832811.
I'll try to come up with a smaller data set and also a reliable way of reproducing it (so far it has been a bit inconsistent in the results, so I don't even thing it has anything to do with UNA anymore). Thanks.
Anyway, the data that we are working on is https://github.com/own-pt/openWordnet-PT. The TBOX is the wn30.ttl file. The ABOX is composed by files wordnet-en.nt.gz and own-pt.nt.gz.
I also get (NIL NIL), I thought that was normal. When I had a satisfiability issue, it was shown there instead of (NIL NIL). Anyway, I also wrote my own satisfiability checker:
(defun check-satisfiable (&optional (concepts (all-atomic-concepts)))
(cond
((eq concepts nil) nil)
(t ((lambda (concept)
(cond
((or (eq concept 'bottom)
(concept-satisfiable-p concept 'default))
(check-satisfiable (cdr concepts)))
(t (cons (car concepts)
(check-satisfiable (cdr concepts)))))) (car concepts)))))
Oh and to be sure, I am working directly in SBCL (in-package "RACER-USER"), not in some server or limited racer-toplevel.
Not sure I understood this one:
Shouldn't the second argument be one explanation for the inconsistency? Is this due to some out of memory condition? (My ontology contains a couple of million triples.) Or am I missing some initialization steps?