Lmst

This week I presented to the Wikibase Stakeholder Group @wbstakeholders about the challenges of creating 5000+ lines of Shape Maps for RDF validation via rudof, and today the Query Shape Maps feature landed, with just documentation missing:

https://github.com/rudof-project/rudof/issues/387

Amazing work by José Emilio Labra and the WESO team!!

@thgie

#ArtBase

Pocky Day is coming! 🍫 I made a free art base you can use to draw your own cute couple or friends! 💖 Celebrate Pocky Day (11/11) with a new free base! FREE DOWNLOAD🔗 ko-fi.com/s/a765efcdc9 할로윈 다음의 오타쿠 명절 빼빼로데이가 다가옵니다. 늦지 말고 준비하세요 - 할로윈 데이 결국 준비 못한 인간 올림😭 #PockyDay #FreeBase #ArtBase #F2U

I was able to write Shape Expressions for most types of items in Rhizome's #ArtBase and validate them using rudof! https://rudof-project.github.io/

Boiling the ocean with RDF data validation
#ArtBase

Data in #ArtBase is quite incomplete. Displaying a minimum narration anyway is one of the challenges for the refresh.

This is a prototype of a timeline of an artwork that Rhizome doesn't even have a copy of. It was hosted on GeoCities (as per the URL the artist submitted). An events database is matching it with the date of GeoCities' closure. The Wayback Machine then finds a redirect to a new domain, but no other information about it, so it is drawn as "fading out" some time later.

The image is fully described in the post's body.

Current state of the #ArtBase data model (at least the class hierarchy) can be viewed at https://tinyurl.com/2c7uofkr — sorry for link shortener, that's how the Wikibase query service works 😅

With the current #ArtBase setup I need to be proficient in Lua, PHP, Python, HTML, Docker, SPARQL, and bash. <https://digipres.club/@despens/105508279785415448>

For the setup I'm working on, I'll need proficiency in SPARQL, Gotemplates, HTML/JS/CSS, Python, Docker, and Bash. Scratching Lua and PHP seems like such a relief!

For piloting the site locally, I'm using Qlever, a new, amazingly fast SPARQL service <https://qlever.cs.uni-freiburg.de>. Just like Wikidata, Wikibase is using Blazegraph for SPARQL, but that's kind of abandonware and requires so much configuration. I hope to replace the public Blazegraph server with Qlever as well.

#ArtBase

So the renewed Artbase will be rendered by @abbe98's snowman <https://github.com/glaciers-in-archives/snowman>, a static site generator for SPARQL backends. As with most SSGs, the templating is super flexible (using gotemplates).

Which means the data model doesn't need to take into account the crutches required to read anything with Mediawiki's built-in Lua interpreter. Awesome!!

#ArtBase

Mediawiki is designed to help a staggering amount of users to not accidentally screw up the site, which means most HTML tags being filtered out, changing CSS take lots of effort etc.

According to a very small sample size, most Wikibases aren't open for editing by the general public <https://wbstakeholder.group/projects/data-stewardship-policies>.

Anyway, I wrote an absolute monstrosity of a templating system in Lua and patched Mediawiki to make it display at least some sort of layout, but that's not going to last.

#ArtBase

In general, the data model needs a bit of an overhaul. It is very much designed with the severly limited capabilities of Wikibase for data retrieval in mind. Wikibase itself cannot run SPARQL queries. There's an extension that can do it to some extent <https://professional.wiki/en/extension/sparql> but then I still have to deal with the super-basic presentation capabilities of Mediawiki.

#ArtBase

I'm currently working on revamping Rhizome's Artbase public collection/archive. It is running on Wikibase, we offer a SPARQL endpoint, etc. There are several issues to tackle and I'll post about them from time to time. #ArtBase