Re: CWM Bug: bNodes Mistakenly Converted to URIs

On 10/14/07, Yosi Scharf <syosi@mit.edu> wrote:

> That bug was caused by a missing line in query.py due to me
> being confused at some point. The cwm CVS has it fixed (I ran
> into the same thing about a month ago and just fixed it).

Using the CVS version of cwm and a different but similar test case,
I've found that the same bug(?) still appears:

$ cat lists-02.n3
# lists-02.n3 - Simple lists

@prefix : <http://example.org/#> .
@keywords a .

ThreeMemberList predicate (p q r) .
MixedThreeMemberList predicate ("p" _:q <http://example.org/path#r>) .

$ cwm lists-02.n3
#Processed by Id: cwm.py,v 1.195 2007-08-23 16:28:29 syosi Exp

#  Notation3 generation by
#       notation3.py,v 1.197 2007-09-09 22:49:43 timbl Exp

     @prefix : <http://example.org/#> .

    :MixedThreeMemberList     :predicate  (
        "p"
         [
             ]
        <http://example.org/path#r> ) .

    :ThreeMemberList     :predicate  (
        :p
        :q
        :r ) .

#ENDS

(This is fine.)

$ cat normalise.n3
@prefix log: <http://www.w3.org/2000/10/swap/log#> .

{ ?s ?p ?o . ?s log:rawType log:Formula .
  ?o log:rawType [ log:notEqualTo log:Formula ] }
 => { _:s ?p ?o } .

{ ?s ?p ?o . ?o log:rawType log:Formula .
  ?s log:rawType [ log:notEqualTo log:Formula ] }
 => { ?s ?p _:o } .

{ ?s ?p ?o .
  ?s log:rawType log:Formula .
  ?o log:rawType log:Formula }
 => { _:s ?p _:o } .

{ ?s ?p ?o .
  ?s log:rawType [ log:notEqualTo log:Formula ] .
  ?o log:rawType [ log:notEqualTo log:Formula ] }
 => { ?s ?p ?o } .

(Note: different to the last time you saw normalise.n3; fixed some bugs.)

$ cwm lists-02.n3 --filter=normalise.n3
#Processed by Id: cwm.py,v 1.195 2007-08-23 16:28:29 syosi Exp

#  Notation3 generation by
#       notation3.py,v 1.197 2007-09-09 22:49:43 timbl Exp

     @prefix : <http://example.org/#> .

    :MixedThreeMemberList     :predicate  (
        "p"
        <#_g0>
        <http://example.org/path#r> ) .

    :ThreeMemberList     :predicate  (
        :p
        :q
        :r ) .

#ENDS

(This is not fine: <#_g0> should be a bNode.)

> Thanks for the feedback. It would be nice to feel confident about
> cwm enough for me to make a release.

You're welcome; thanks for the fixes. It would indeed be great if we
could have a cutting-edge packaged release, though actually the CVS
cwm doesn't seem to be giving me any problems associated with not
having been through the python setup.py install procedure, so that's
not so bad.

-- 
Sean B. Palmer, http://inamidst.com/sbp/

Received on Sunday, 14 October 2007 12:42:27 UTC