Wikibase Community User Group/Meetings/2021-02-11/Notes
Appearance
Archived notes from Etherpad of the 2021-02-11 meeting.
Schedule
[edit]- 16:00 UTC, 1 hour, Thursday 11th February 2021
- Online meeting on Google Meet: https://meet.google.com/nky-nwdx-tuf
- Join by phone: https://meet.google.com/tel/nky-nwdx-tuf?pin=4267848269474&hs=1
Agenda
[edit]- What have you been up to?
- Share out anything exciting or new that's been happening with your Wikibase instance
Participants (who is here)
[edit]- Mohammed (WMDE)
- Georgina (WMDE)
- Jay Gray
- Georgios
- Johannes
- Alexander
- Lyndsey
Notes
[edit]- Results of the poll asking "what would be abetter choice of platform for Wikibase community discussions?
- 63 percent want to keep existing telegram channel, and use parallel platform
- Suggested parallel platform: Mattermost, Stack Overflow, Discourse
- Professional.Wiki have volunteered to set up and maintain Mattermost channel
- Georgios: how representative is the pole when only four voted for Mattermost?
- Some prefer email, others suggest that it might be worth thinking about whether we want "an alternative/more searchable chat channel, vs something where a static or more permanent Q&A can be formulated"
- Jay: using amazon web services Neptune as a SPARQL end point. moving away from this. Want to move to Blazegraph, the tech behind wikibase and wikidata, and part of that reasoning is to be able to have federated queries between own triple store and wikidata. dont add triples straight to wikibase, but through a QA, then we promote it to a QA store. considering using wikibase as QA database that allows people to check their triples before they upload them. Does this make sense? Is this a good use case for wikibase?
- Background: Business mission is to create concept of an electronic label, anytime you see paper label, that information can be encoded, shared and reused via an electronic label. If youre attempting to determine what your food intake has been, you have no method of storing it systematically. but with an eletronic label, then over that period of time you could understand your intake and relate that to health outcomes
- Need an intermediate storage environment.
- Johannes: how many triples are we talking about?
- Jay: currently has about 300 million triples in it, but thats small compared to what production database would look like
- Georgios: have you got a wikibase instance already?
- Jay: no, more considering it atm rather than realising the idea.
- Johannes: running blazegraph in background, but doing it in a special way. MediaWiki software as the governer, other script that updates. Theres not been a solution where you put the informaiton into the blazegraph and there on from there. The major problem you encounter is that there is a very particular data model in the Wikibase blazegraph bc of the mediawiki software.
- Jay: comfortable with Wikidata RDF model. Less familiar with method by which you register data to wikibase.
- Georgios: WBStack, you can request and get a free wikibase sandbox. Is your data tabular?
- Jay: data is in JSON LD. Triples.
- Georgios: If tabular, good tool for it is open refine.
- Alexander: reconcilation service for Wikidata, is it easy to set up?
- Georgios: recently done by interns in the google summer of code. Need Wikibase manifest, set up your open refine to expose the wikibase manifest, then openrefine can reconcilate data.
- Lyndsey: did this fairly recently for Rhizomes wikibase instance: https://github.com/OpenRefine/OpenRefine/wiki/Write-a-Wikibase-manifest. Video that gives further instruction.
- Johannes: different versions, if he should wait and use the latest version, would it have any impact on his properties and the way it is laid out?