Opening BRICK ontology with Protege

[Maryam Montazer]

Hi guys!

I am a PhD candidate and  I am new in ontology modelling domain. I would like to use BRICK ontology  to model the existing BMS system  of an existing building. I am in first step and I would like to know how I can open the BRICK ontology with Protege to see the BRICK schema.

Please  let me know If my question is not clear enough!

Best regards,

Maryam

[Gabe Fierro]

Hi Maryam:

Welcome and thanks for reaching out! You can download the latest Brick Turtle file at https://brickschema.org/resources onto your machine and open it in Protege using the file dialog (https://protegewiki.stanford.edu/wiki/Protege4GettingStarted). I also do most of my development using the Python RDFlib (https://rdflib.readthedocs.io/en/stable/) and the Brickschema packages (https://brickschema.readthedocs.io/ --- built over RDFlib). There is an example here (https://github.com/BrickSchema/Brick/blob/master/examples/simple_example_1.py) of using RDFlib to build a Brick model.

Hope those are some helpful pointers! Don't hesitate to reach out with more questions or feedback on your experience with Brick

Best,

Gabe

--
You received this message because you are subscribed to the Google Groups "Brick User Forum (Unified Building Metadata Schema)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to brickschema...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/brickschema/de417238-e8a3-4463-ab58-ad6c146b2fcfn%40googlegroups.com.

[majid.a...@gmail.com]

It's straightforward but I got a warning:

 WARN  11:14:00  Illegal redeclarations of entities: reuse of entity https://brickschema.org/schema/Brick#value in punning not allowed [Declaration(ObjectProperty(<https://brickschema.org/schema/Brick#value>)), Declaration(DataProperty(<https://brickschema.org/schema/Brick#value>))]

On that note, I've come across tools that don't like these redeclarations.

[Gabe Fierro]

This seems like a bit of a strange issue; I don't see anything in the OWL ontology that prevents a property from being both an ObjectProperty and a DatatypeProperty. Would you be able to file an issue on GitHub with the full report?

Best,

Gabe

To view this discussion on the web visit https://groups.google.com/d/msgid/brickschema/d757fbbc-e589-4b16-8060-08e2ba8f6eb5n%40googlegroups.com.

[Joel Bender]

According to https://www.w3.org/TR/2004/REC-owl-semantics-20040210/direct.html#3.3 the simplest distinction is that a DatatypeProperty cannot be Symmetric, InverseFunctional, or Transitive, which makes sense because the subject of a triple can only be an object.  The text for a DatatypeProperty has "EC(d1)×LV" which seems to reference the cross product of the domain and values, while for an ObjectProperty it has "EC(d1)×O" with the domain and objects.  I think this is strange because I would have expected the domains to be identical and the ranges to be different, I would have expected DatatypeProperties to formally only reference something in LV and ObjectProperties only something in O, not just in their descriptive text.  Since something can't be in LV and O at the same time, something can't be both a DatatypeProperty and an ObjectProperty, it would just be an rdf:Property.

In https://www.w3.org/TR/owl-ref/#Property there is a "NOTE: In OWL Full, object properties and datatype properties are not disjoint."  These warnings might be about having an ontology automatically trigger the dreaded OWL Full unintentionally, which is why it would be a warning and not an error.

[Gabe Fierro]

Oof, I'd really like to stay away from OWL Full. Does your investigation imply that the simplest fix is to make brick:value an rdf:Property rather than an ObjectProperty or DatatypeProperty? This feels like the simplest fix, even if it is a bit of a hack. I suppose the other fix would be to have two flavors of brick:value. One flavor ("brick:value") would be a datatype property and the other ('brick:entityValue ?') could be an object property.

Best,

Gabe

To view this discussion on the web visit https://groups.google.com/d/msgid/brickschema/28e66a44-10aa-4038-a5bd-d7f7701f9de2n%40googlegroups.com.

[Joel Bender]

Oof, I'd really like to stay away from OWL Full.

I had a similar issue with the first attempt at a "BACnet Ontology" and the ABSTRACT-SYNTAX.&Type references floating around turned it into a family of nested namespaces, one per object and PDU.  Frighteningly similar to translating IFC EXPRESS into RDF via UML and OWL.  Hideous!

 

Does your investigation imply that the simplest fix is to make brick:value an rdf:Property rather than an ObjectProperty or DatatypeProperty? This feels like the simplest fix, even if it is a bit of a hack.

Yes, that's the simplest, but you can't describe it's contents very well doing that.

 

I suppose the other fix would be to have two flavors of brick:value. One flavor ("brick:value") would be a datatype property and the other ('brick:entityValue ?') could be an object property.

The type safe purest in me says that's the right path because they really are different properties, but this is in the context of punning, so I really don't know.

Joel

[Gabe Fierro]

I've made a note of this issue on the Brick tracker: https://github.com/BrickSchema/Brick/issues/308

It doesn't strike me as super pressing at this stage, especially since most tools are not OWL Full-aware anyway, so the subtle implications of this redefinition can't actually be realized by most software. We can discuss in the Ontology WG and technical committee and land a fix in Brick v1.3

Best,

Gabe

--

You received this message because you are subscribed to the Google Groups "Brick User Forum (Unified Building Metadata Schema)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to brickschema...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/brickschema/209aaab2-098d-4fc1-88f1-f8a227019714n%40googlegroups.com.