Python / Cython bindings & further upgrades

[Evgeny Blokhin]

@wrobell let's switch to this thread? So the other one is too general.

[Evgeny Blokhin]

I would use FHKB ontology as a "gold standard" for testing, although there are such statements as equivalentClass, ObjectProperty, equivalentProperty, FunctionalProperty, propertyChainAxiom, inverseOf, unionOf, intersectionOf, someValuesFrom, onProperty etc. And so they all should be implemented then, right?

[Evgeny Blokhin]

For a nice comparison rdflib provides naive Python-only OWL2 RL reasoner by Ivan Herman:

DeductiveClosure(OWLRL_Extension).expand(graph)

which is suitable for all above, although of course tremendously slow.

[wrobell]

This discussion is probably out of scope for FaCT++. The Coras project (https://bitbucket.org/wrobell/coras) parses OWL 2 statements with RDFLib Python library and uses FaCT++ via the factpp Python bindings for reasoning. Both, Coras project and the Python bindings are maintained in their own separate repositories.

[Evgeny Blokhin]

For anyone visiting this thread: this work was published and continues at https://github.com/tilde-lab/pyfactxx

[Evgeny Blokhin]

@Dmitry Tsarkov please let me know, if you would like to accept pyfactxx back into this repo. This is basically a new folder FaCT++.Python plus some enhancements to the core.