Attempt 1: replacing DLExpressivity by DLExpressivityChecker

[matentzn]

This is try to sort out the problem that the previous solution shows gibberish expressivity levels:

"expressivity": "RRESTRCUCINTUNIVRESTREROIF(D)". But it does not work..

Resolves [#ISSUE, resolves #ISSUE]

• [ ] docs/ have been added/updated
• [ ] tests have been added/updated
• [ ] mvn verify says all tests pass
• [ ] mvn site says all JavaDocs correct
• [ ] CHANGELOG.md has been updated

[DESCRIPTION, mentioning relevant #ISSUE]

[ckindermann]

A workaround might be using Collection<Languages> languages = checker.expressibleInLanguages(); The documentation states that the returned languages are minimal w.r.t. isSubLanguageOf. So, if the set is not a singleton, we could impose an order and return the first result. Here is a snippet:

    Collection<Languages> languages = checker.expressibleInLanguages();

    if (languages == null || languages.isEmpty()) {
      System.err.println("No language found");
    } else {
      Languages[] array = languages.toArray(new Languages[languages.size()]);
      Arrays.sort(array);
      System.err.println(array[0]);
    }

[ckindermann]

Your commit message is not correct: expressibleInLanguages() does not return all languages. Each language in the collection

• allows for all necessary constructors (meaning the ontology can be expressed in the language),
• is minimal regarding sublanguages (meaning the set does not include more expressive languages than necessary).

However, there may be multiple languages satisfying these criteria. In this case, sorting just makes the output deterministic.

[matentzn]

Thank you @ckindermann - I will make sure to rectify the commit message when its time to create a squash merge.

[ckindermann]

Thanks! In any case, getDescriptionLogicName() is the correct function to call. The OWL API currently implements this incorrectly by concatenating the set of constructors used in the ontology (see here in version 4 and here in version 5).