unaddressable URLs about bao

[sschurer]

Hi Leon, Sorry for the late reply. BAO is built modular with various files directly and indirectly references in bao_complete. For you example if you open bao_complete in Protege it will bring in all the other needed vocabulary and module files. If you require a resolution for all URIs we can point to the bao_complete_merged file. We are considering the pros and cons internally. All the best, Stephan

…

On Fri, Mar 6, 2020 at 7:58 PM leonqli ***@***.***> wrote: Hi, there several terms are not addressable by using the following URLs. Pl http://www.bioassayontology.org/bao#BAO_0001094 http://www.bioassayontology.org/bao#BAO_0001067 http://www.bioassayontology.org/bao#BAO_0000540 After digging it, I found they are in this file, bao_complete_merged.owl, but not in bao_complete.owl. But the default url point to bao_complete.owl. Is it a bug? Best, Leon — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#35?email_source=notifications&email_token=AAQF7WHCITQW7IVJCGJEOULRGGL2HA5CNFSM4LDKUEKKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4ITIDHAA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAQF7WGSZ4WQIZ6QHY464Y3RGGL2HANCNFSM4LDKUEKA> .

[sschurer]

Leon, We discussed this internally some more. Can you share your use case where you are getting a problem. Briefly, bao_complete imports various other files and some of these import others; we opted for that modular design to streamline development and maintenance. If you use bao in a triple store or Protege or another tool that works with ontologies, you should be able to answer requests for all URIs. You would need to do that anyway to use the ontology for example for reasoning (typically only the asserted ontology would not be distributed). Would be helpful to understand your use case to determine if we need to change anything.

[leonqli]

Dear Stephan,

Thanks for your reply! Our use case is PubChem RDF (https://pubchemdocs.ncbi.nlm.nih.gov/rdf). We have those links in our RDF project. When clicking those links, it will be directed to bao_complete.owl as default, which has no those terms in the file. For example, BAO_0001094, http://www.bioassayontology.org/bao/bao_complete.owl#BAO_0001094

Thanks for your help! Look forward to hearing from you.

Best,
Leon

[sschurer]

The URI would be http://www.bioassayontology.org/bao#BAO_0001094. The reason that resolves to bao_complete is because BAO complete is the upper level file that directly or indirectly imports various other files with classes and relations. In order to answer a semantic query you need a system that can interpret the request such as protege or some other system that uses the OWL API. Typically to integrate and query semantic data you would use a RDF store, such as Virtuoso? That should resolve any query. It would also give you answers using inferred axioms, which you could not get from the asserted hierarchy, even if they were all one file.
We don't have PURL like resolution system at the moment. But again it should be no problem if BAO is loaded in a triple store or protege for example.