Non-compliant definitions

[mereolog]

Some definitions in FIBO seem do not follow the rules specified in https://github.com/edmcouncil/fibo/blob/master/ONTOLOGY_GUIDE.md#definitions: they are not sentence parts ("definientia"), but full sentences or even "sentence sequences".

Using some simple NLP technique one can identify some of them - see the attachments, which refer to Q1 2022 release.
incongruent_class_individuals_in_2022_Q1.xlsx
incongruent_class_definitions_in_2022_Q1.xlsx

I am thinking about incorporating the respective check to hygiene tests, although (i) it will not be a SPARQL query (ii) it may take some time before we implement it properly. So in the meantime we might fix at least some of the issues found.

Obviously, there are some false positives in the results, so the final check should by interpreted as a warning.

[ElisaKendall]

Most of the class definitions that are questionable are in provisional ontologies, though there are a few in production.

[ElisaKendall]

@mereolog Would you please re-run these tests on released ontologies only so we can see what's left? Some of these have been corrected I think.

[mereolog]

Here they are - from the latest commit:
possible_incongurent_definitions_for_individuals_20230208.xlsx
possible_incongurent_definitions_for_classes_20230208.xlsx

[ElisaKendall]

@mereolog Would you please re-run these tests again so that I can address anything remaining in production in Q2, and then maybe close this? Thanks!

[mereolog]

@mereolog Would you please re-run these tests again so that I can address anything remaining in production in Q2, and then maybe close this? Thanks!

Here you are:
possible_incongurent_definitions_for_classes_20240425.xlsx