Closed ross-spencer closed 2 years ago
We have a better idea of the work that needs doing now, although a pattern for making those changes is yet to be decided upon:
The SPARQL we may read quite happily from the Siegfried ~/home/siegfried
folder. The URI changes are a bit more of a pain. Namely, the IRIs seemed nice enough, but are specific to Wikidata. The problem is, we need to know the label being used, e.g. "beginning of file" vs. relativity or the IRI itself. The IRI changes with every Wikibase instance, even with federated properties as some parts of the query refer to items.
The good/interesting/fascinating news is that I was able to demonstrate this working (unlisted YouTube video below):
As you can see the video above is primarily for TNA as we previously discussed, giving them a heads up seemed sensible. There's an outside chance it might impact PRONOM 2's alpha phase.
The demo server is here: http://54.36.119.4/wiki/Main_Page (feel free to create a login and play around)
NB. Penetration testing results of Wikibase from a consultancy blog: https://www.wikibase-solutions.com/developer-logs/wikisecuritypreparations
Implemented and released here! https://github.com/richardlehane/siegfried/commit/81f1f4cf49b227e04f246eed0868afdf8391e0fc ~(but more in another thread...)~
Description of problem
Parts of the Wikidata implementation have been developed with an open view to connecting things up to a generic Wikibase instance but this hasn't been proven yet. Siegfried and Wikiprov development has been somewhat blind in this regard and so it is unclear exactly what things are too hardcoded to work, and what changes are required. (If it doesn't work entirely out of the box :wink: ) . We have discussed potentially showing this capability at OPF next week.