#qlever 420ms, #wikidata #query service Timeout nach 60 Sek. , also mindestens Faktor 143. Keine Ahnung, was da die technischen Hintergründe sind; für die produktive Arbeit mit Wikidata ist #Blazegraph aber ein enorm limitierender Faktor.

c'est acté. #Wikidata va se doter de plusieurs graphes. Les articles scientifiques vont être stockés dans un autre graphe, mais ça ne change rien aux données Wikidata. Restera à trouver une alternative à #Blazegraph qui est en fin de vie. #GraphSplit https://wikimania.eventyay.com/2024/talk/SYV8AP/

This is a cached query result: https://results.orbopengraph.com/#%23%20Vaccine%20safety%3A%20vaccine%20types%0A%23%20defaultView%3ATable%0Aselect%20%3Fitem%20%3Farticle%20where%20%7B%0A%20%20%3Fitem%20wdt%3AP31%2A%20wd%3AQ105967696.%0A%20%20%3Farticle%20schema%3Aabout%20%3Fitem%3B%0A%20%20%20%20schema%3AisPartOf%20%3Chttps%3A//en.wikipedia.org/%3E.%0A%7D

It updates every 24 hours, as it is configured to. A backend process submits the query to a private, load-balanced query service, and makes it visible through that frontend. If you try to submit a query through that frontend, it won't work.

The next step is setting up the app that lets you enter queries into the database.

#wikidata #wikibase #sparql #rdf #query #blazegraph #graphdatabase

Les cathédrales ont un plan de sauvegarde des oeuvres d'art pour déterminer les oeuvres à sauver. #Wikidata a un document listant ce qu'il faut supprimer en cas de plantage complet de #Blazegraph. Ce sera les articles scientifiques 😃 https://m.wikidata.org/wiki/Wikidata:SPARQL_query_service/WDQS_backend_update/Blazegraph_failure_playbook

However, when #Wikidata emerged, I wanted to connect my database with it. Before two years, I tried to store SMW in #Blazegraph but I made a bad experience with it. When the power went off, it led to fatal error and problem to boot.

Two months ago, I tried a second attempt, with #Virtuoso for this time. It runs without problems.

Now, I can make #sparql queries, however it is quite complicated, e.g. to query properties and pages with longer string names.

Now available for download, the 2023-04-23 #Blazegraph dump for #Wikidata Query Service. Use to bootstrap your own local copy of WDQS. https://datasets.scatter.red/orb/

#Wikidata Query Service #Blazegraph dump now available to download, free of charge (344 GB compressed; over 1 TB decompressed).

https://datasets.scatter.red/orb/

Wrapping up this thread with a new #blog post: "Deploying #Wikidata to different graph databases and what works best" #blazegraph #sparql #wikibase

https://harej.co/posts/2023/01/loading-wikidata-into-different-graph-databases-blazegraph-qlever/

The challenge here isn't going to be initial setup, as far as I can tell. The index build is in progress. It is going to be taking updater tech built for #Blazegraph and adapting it to #QLever. Which I suspect is possible, if both use #SPARQL, but I haven't actually tried it yet.

After successfully experimenting with and deploying #Blazegraph for #Wikidata querying I am now experimenting with #QLever which is like a breath of fresh air.

Eventually I would like to offer QLever as an experimental service alongside Blazegraph, eventually retiring Blazegraph.

Rebuilding #Wikidata in #Blazegraph was successful, as was uploading the 1 TB database file to cloud storage.

Now I am syncing the database with present day. Unlike last time, I am pretty sure I am actually pulling from Wikidata this time.

Rebuilding #Wikidata on #Blazegraph ...again... after a misconfigured updater caused data corruption. In parallel I am working on re-attempting Wikidata in #QEndpoint

Recommended system configuration to run a #Blazegraph instance loaded with #Wikidata –

* 2x 1 TB NVMe SSDs, formatted together as a single 2 TB volume
* 128 GB of RAM
* If your CPU can address 128 GB of RAM it's probably good enough

Successfully rebuilt #Blazegraph data.jnl from dump; now backing up my effort. (I may make such dumps available for download, since building #Wikidata from dump takes multiple days even on good hardware.)

[tool] couchdb-blazegraph-sync: keep a #CouchDB database and a #BlazeGraph namespace in sync https://github.com/maxlath/couchdb-blazegraph-sync