Feature 870 create automated annotation for module identification

[jannahastings]

Opening a draft pull request so that I can add comments on the diff. Not to be merged yet.

[jannahastings]

Looks good, nearly there. I have made detailed comments in various places in the code.

A general point is that the robot.jar has snuck into the commit again so this will need to be reverted once more.

A bigger issue is that the robot template command is creating new classes where the ID referred not to a class but to something else, e.g. an individual. Thus, we will need to keep track of the OWL type associated with the ID, which we can get from the original export and then pass back to the template command later. Let's discuss and implement this after the rest of the pipeline is working correctly as per the other code comments.

[stage1407]

Ok, I've implemented your comments.
Now I think it is ready for the next release

But if anything is still open let me know.

Edited:
Except for overwriting the ontologies of course.
And I forgot to reintroduce the numeric ID for belongs_to_module after it was undone by the accidental reset.

[stage1407]

Should we also introduce numeric IDs for the representation of the oeo modules?
I'm unsure but I think that we should link with the IRI to the module directly. So without a numerical ID.

[jannahastings]

Should we also introduce numeric IDs for the representation of the oeo modules? I'm unsure but I think that we should link with the IRI to the module directly. So without a numerical ID.

No need to introduce numerical IDs for the oeo modules, we just use their IRIs directly.

[jannahastings]

Edited: Except for overwriting the ontologies of course. And I forgot to reintroduce the numeric ID for belongs_to_module after it was undone by the accidental reset.

OK, good, so there is just one more thing to add after these changes are done. The script needs to keep track of the OWL type associated with the ID, for example, individual, class, or object property. Robot provides headers both for extracting this in the first step and for specifying it in the template. We need this because otherwise the robot template step at the end assumes that all the IDs should be classes, when they are not.

[stage1407]

<http://www.geneontology.org/formats/oboInOwl#<http> "//openenergy-platform.org/ontology/oeo/OEO_00260001> <http://openenergy-platform.org/ontology/oeo/oeo-model>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#aggregation type>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#analytical approach>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#boolean variable>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#data format>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#methodical focus>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#software framework>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#time stamp alignment>"^^xsd:string,
<http://www.geneontology.org/formats/oboInOwl#<http> "//www.w3.org/2002/07/owl#uncertainty approach>"^^xsd:string

Ok the error arises from the fact that in the ontology e.g. 'http://www.geneontology.org/formats/oboInOwl#<http' is specified as IRI and such a thing can not be parsed, of course.

I am surprised at this point that something is imported from the geneontology at all and not owl: that is used directly in the script to specify types.

The main problem is that the http: of the following IRI ends up there.

@jannahastings
Do you have any guess, where could be the problem there?

Otherwise I have to look the next days, why this is happening.

[jannahastings]

@stage1407, I was just double checking the prefix command format in ROBOT:

--prefix "ex: http://example.com/"

so for our scenario that would be like this, I think:

--prefix "OEO: http://openenergy-platform.org/ontology/oeo/OEO_"

[stap-m]

@stage1407 @jannahastings How far is this PR?

[jannahastings]

Unfortunately, we did not yet make progress with solving the problem with the implementation of this issue :-(.

[stap-m]

This PR is not needed anymore, since 1) protege is now able to show the module itself, and 2) the restructuring in #1592. We can finally close it. 🎉