Add the Earth Metabolome Initiative (EMI) ontology

[tarcisiotmf]

Please find below the filled request form:

EMI OLS ontology request.xlsx

Please, let me know if you need any further information from us.

Thank you very much!

Best regards,
Tarcisio Mendes de Farias

[haideriqbal]

Hi @tarcisiotmf! I've tried to load this ontology into OLS but at the moment one of the imports (http://www.linkedmodel.org/schema/dtype) is not been able to get processed by the OLS. The reason being JENA parser which OLS uses to parse ontology files can't determine the content type of the file.

When you open the file in the browser you see an XML file but for JENA to process the file, it need to have Content-Type HTTP header which this particular import doesn't have. See below:

curl -I http://www.linkedmodel.org/schema/dtype

HTTP/1.1 200 OK
Date: Tue, 10 Dec 2024 13:07:57 GMT
Server: Apache
Upgrade: h2,h2c
Connection: Upgrade
Last-Modified: Tue, 14 Nov 2023 22:20:42 GMT
ETag: "156120f-5d8f-60a24327cf6e0"
Accept-Ranges: bytes
Content-Length: 23951
Vary: Accept-Encoding
Access-Control-Allow-Origin: *

So my suggestion at this stage would be if possible you can get rid of this import. As I see from the linkedmodel.org website that it isn't being maintained actively, last update was in 2016 and dtype ontology itself was updated in 2011. Or you can try to reach out to linkedmodel people to get it sorted out.

In the meanwhile, I would try to see on our end if we can find a work around in OLS to process the ontology files not having HTTP Content-Type header in the request.

[tarcisiotmf]

Thank you very much for considering and looking at it. Actually, we are not importing directly the "http://www.linkedmodel.org/schema/dtype" ontology as you can see in the original file, but it is being imported by one of our imported ontologies, namely, https://qudt.org/schema/qudt/ . Therefore, it is an import of an imported ontology. For your information, qudt that imports dtype seems to be actively maintained as you can see in the bottom of its page (Generated 2024-09-20T17:35:35.094-07:00).

Meanwhile, may you consider the following purl to fetch the ontology? I've removed the qudt import. Thank you very much!

ontology_purl: https://purl.org/emi/ols

Remark: depending on the nature of the ontology and its application, it is indeed better to not change often over time. Particularly, the dtype ontology is to define data types that are not covered by xsd that is not being updated since 2004 and in use by the OWL 2 specification, from the "dtype" description:

The ontology "dtype" provides a specification of simple data types such as enumerations and codelists. These are needed in support of the conversion of XML Schemas and UML Models to OWL. The purpose of "dtype" is to provide a foundation for expressing rich data types. These data types are typically needed when dealing with OWL representations of data structures derived from XML Schema.

[henrietteharmse]

@tarcisiotmf Thanks for removing qudt!

We completely agree that some ontologies do not get updated for good reason. A good example of this in OLS is BFO, which is an
upper ontology.

I think the concern here is more that the http://www.linkedmodel.org/ website has a number of dead links. In particular, the link to "DTYPE 1.2", clearly indicating the intension for a later version of DTYPE. Same goes for "VAEM 2.0", "VOAG SCHEMA 2.0", etc., with no update wrt the status of any of these. Our intension was to reach out to them to help resolve this issue. But with no update to the website since 2016, that approach does not seem viable.

[tarcisiotmf]

Yes, you are right. I think this issue should be pointed out to the quadt maintainers (https://www.qudt.org/pages/QUDTcontactsPage.html) since it is an alive project that reuses "dtype 1.0" and "vaem". I will try to reach them to solve this issue.

In any case, for the EMI ontology it is not critical for us to not consider the "quadt" import at OLS. This is also because specially in our case we are only importing quadt to just be able to assign units with the controlled vocabulary (https://qudt.org/2.1/vocab/unit) via the property qudt:hasUnit that does not depend on dtype or veam vocabularies. Moreover, we are not (directly) importing "dtype" or "veam" that is part of the quadt schema that is not being reused by the EMI ontology.

Thanks again for your help! Please let me know if you have any issue to load the https://purl.org/emi/ols file with Jena.

[haideriqbal]

@tarcisiotmf new ontology file works fine with OLS. It will be available in production within 48 hours.

[tarcisiotmf]

Thanks!