- From: Dan Brickley <danbri@google.com>
- Date: Thu, 21 Apr 2016 21:57:25 +0100
- To: Thad Guidry <thadguidry@gmail.com>, Richard Wallis <richard.wallis@dataliberate.com>
- Cc: Marc Twagirumukiza <marc.twagirumukiza@agfa.com>, "schema.org Mailing List" <public-schemaorg@w3.org>
On 21 April 2016 at 21:49, Thad Guidry <thadguidry@gmail.com> wrote: > So Dan, you or Richard will work on a bit of Python to catch this sorta > thing in extensions in the future, right ? Heh, I'd like to say "just contribute a new SPARQL-based test case, Thad!" but there are some subtleties that probably need investigating first. https://github.com/schemaorg/schemaorg/blob/sdo-deimos/tests/test_graphs.py is a collection of tests that work reasonably well against the core, and catch some similar issues. But yes we need to get this updated for the more complex environment we're in now with extensions, possibly using named graphs etc etc (the site is implemented using triples + contextual layer corresponding to a named extension, since you ask). Feel free to take a look though, maybe a pretty simple query would still catch some possible mistakes. For e.g. you could look for pairs of types ?c1 and ?c2 where ?c1 is subClassOf a broader type ?c2 and ?c1 isPartOf the core and ?c2 isn't. I'm not sure if we're loading all the extension data into the unit test environment yet though. Richard? :) Dan > Thad > +ThadGuidry >
Received on Thursday, 21 April 2016 20:57:54 UTC