Search Mailing List Archives
[bioontology-support] Strange handling of ontology identifiers
Natasha F. Noy
noy at stanford.edu
Wed Jul 28 14:49:41 PDT 2010
I think it is up to the API to append or not append the namespace
prefix when there is a prefix defined for the default namespace. Note
that all the names defined in the default namespace are actually
implicitly starting with http://bhi.washington.edu/OPB#.
After expanding the prefix, the result is the same (and prefix is just
a shorthand anyway)
We use Jena API to process the OWL files.
In general, we just process the file that is available on the download
link. OPB, however, is one of only a handful of exceptions, where we
do some massaging before displaying the ontology in Bioportal (because
of some idiosyncrasies).
Does this answer your questions?
On Jul 28, 2010, at 9:30 AM, Camille Laibe wrote:
> Dear Natasha,
> Thank you for your answer, but please read below.
>> Actually, the OWL file does define an "OPB" prefix for the base
>> <rdf:RDF xmlns="http://bhi.washington.edu/OPB#"
> I agree with you, however:
> There is an "OPB" namespace which is defined (xmlns:OPB="http://bhi.washington.edu/OPB#
> ") but it is unused (there is no tag prefixed with "OPB:" in the OWL
> There is also a default namespace declared (xmlns="http://bhi.washington.edu/OPB#
> "), but it should be unused as well as all tags in the OWL file are
> prefixed (with 'owl:', 'rdfs:', 'rdf:', ...).
> Finally there is the xml:base declaration (xml:base="http://bhi.washington.edu/OPB
> "), that, I guess, you use for the generation of your "Full Id"
> field, which interestingly provides a correct value: http://bhi.washington.edu/OPB#OPB_00464
> (no addition of "OPB:" in the middle).
> Therefore, I am still puzzled by this weird display of
> "OPB:OPB_00464", and I would like to know if there is something
> wrong in the OWL file itself or if there is a bug in the library
> that you use to parse the OWL file.
> Which library do you use to parse the OWL files?
> The OWL files you provide for download, are they the same as the
> ones which were submitted to you or do you process them in any way?
> Best regards.
> Camille Laibe
> BioModels.net Coordinator
> European Bioinformatics Institute, Cambridge (UK)
More information about the bioontology-support