Jump to content

Wikibase/Wikibase.cloud

From Meta, a Wikimedia project coordination wiki

Project

[edit]

Wikibase.cloud is a cloud-based platform developed by Wikimedia Deutschland (Wikimedia.de) to host instances of Wikibase. Wikibase is an open-source software suite for creating collaborative knowledge bases. Originally developed for Wikidata, the Wikibase project is being actively developed and maintained by the team at Wikimedia Deutschland. With Wikibase.cloud, we want to make it easier than ever to create, connect and grow a linked open database. In so doing, we aim to liberate data from silos and increase access to knowledge around the world.

Wikibase.cloud is now in Open Beta -- feel free to sign up now! Consider adding your instance to wikibase.world, an invaluable community-run site for Wikibase and the various instances you can find across the net.

We'd like to say a special thanks to Adam Shorland for building WBStack on which Wikibase Cloud is based, and Rhizome, for their support, innovation and leadership.

Resources

[edit]

Connect with the Wikibase.cloud community:

Documentation:

Contact us

[edit]

As with all Wikimedia projects, we use the public ticketing system Phabricator (phabricator.wikimedia.org). You can keep up with our progress via the Wikibase.cloud Phabricator board. Otherwise, reach out to us at swe_partnerships@wikimedia.de.

Updates

[edit]

Q2

  • Achieved in this quarter:
    • Owners of existing Wikibase Cloud instances can now specify the intended purposes, lifespan and target audience for them. This helps us understand the use cases behind individual instances to make strategic decisions around the acceptable use of the platform.
    • We designed and sent out the Data Governance Research activity to our users as part of the initiative to craft the acceptable use policy for Wikibase Cloud. In this activity, our users and stakeholders assign hypothetical use cases to different products in our ecosystem.
    • We have been focusing on establishing high-level metrics to measure the volume of query federation happening in the Wikibase ecosystem. Wikidata Query Service is the only place in the ecosystem that currently has some kind of data infrastructure, so we started by looking at it and improved on our capability to get meaningful insights from it.
      • We made sure that appropriate user agents are set from Wikibase Cloud, so we could differentiate queries based on their provenance (for example, whether they are coming from QS UI or from a QS of another Wikibase Cloud instance).
      • We can also now recognize queries that were executed on the main Wikidata graph vs. the scholarly articles graph.
    • We started a series of publications focusing on the stories of holders of marginalized knowledge using Wikibase Cloud: Post 1 , Post 2.
    • We made a number of small fixes in the Query Service UI (https://phabricator.wikimedia.org/T358602, https://phabricator.wikimedia.org/T399300).
    • We have established the missing metrics for the platform to allow us to monitor its health and growth and do the necessary reporting.
    • OAuth2 consumers bug fix: https://phabricator.wikimedia.org/T336937
  • Plans for next quarter:
    • We’re fighting off the attack of bad traffic and exploring long-term solutions for this.: https://wmde.github.io/wikibase-cloud-status/issues/2025-07-15-scrapers-again/
    • We want to improve on the user’s ability to federate queries (focusing on opportunities related to the Allowlist). This also includes establishing a way to measure our impact with meaningful metrics.
    • In order to make the platform compliant with the Digital Services Act, we will introduce a way for the users to report illegal content. We will also update our Terms of Use.
    • We resumed our work on Mediawiki 1.43 update.
    • We want to complete & synthesize the data governance research. This will become the foundation for the draft of the Acceptable Use Policy that we will later request feedback for.
    • We want to continue the series of publications focusing on the stories of holders of marginalized knowledge using Wikibase Cloud.

April 8, 2025

  • What we’ve done:
    • When one creates a new Wikibase instance, we ask about its intended purpose, lifespan and target audience. We will later start collecting the same information about pre-existing instances as well, to better understand the use cases of individual instances.
    • Addressed several incidents:
      • Fixed the broken status page.
      • We were impacted by a significant increase in traffic that we attribute to scraping activity. This brought the sites down multiple times and we have worked to mitigate this.
      • Solved the problem with database connections failing.
    • Added robots.txt to MediaWiki behind Cloud instances to reduce the performance load created by engine crawlers.
    • Updated Cradle and Quick Statements
    • Added support of a shapes expression validator to simplify validation of entity schemas (like on Wikidata)
    • Small UI fixes: 1, 2
  • What we’re working on:
    • Introducing new approach to metric tracking to evaluate health of the ecosystem
    • Introducing the workflow to collect information on the intended purpose, lifespan and target audience of pre-existing instances
    • Updating to MediaWiki 1.43

December 17, 2024

  • What we’ve done:
    • Solved the problem of user sessions randomly disappearing making the service and surrounding tools unusable: 1, 2, 3, 4
    • Temporarily disabled the Graph Builder functionality in Query Service, as it is not functioning correctly. We will re-enable it when the responsible team fixes the functionality upstream.
    • Made some copy improvements in our automated emails and fixed some links that were leading to the wrong places
    • Some copy and UX/UI improvements on wikibase.cloud and Status page. Also check out the new default logo on the Wikibase instances!:)
    • Started Cloud’s own FAQ page, since the one for Wikibase was mostly useless to our users.
  • What we’re working on:
    • Investigating how to enabled Query Builder and Quality Constraints
    • Planning some long-overdue software updates
    • An announcement of our plans for 2025 will follow after the official WMDE strategy announcement.
    • The Wikibase Cloud team wishes you happy holidays! 🎄⭐ ☃️

November 21st 2024

  • What we’ve done:
    • Visual Editor is now enabled for everyone. It allows editing pages as rich content rather than as source markup code.
    • Enabled personal scripts and personal stylesheets. All Cloud users can now customize their UI experience by editing their own Special:MyPage/common.js, Special:MyPage/common.css and other personal scripts and stylesheets.
      • Setting up global scripts (like MediaWiki:common.js) remains unavailable even to admins.
      • Setting up global stylesheets (like MediaWiki:common.css) remains available to admins as before.
    • Previously, Query Service on Wikibase Cloud only allowed federated queries to the same endpoints that are allowed for Wikidata. Last month, we also whitelisted other Wikibase Cloud instances that already existed at that point. From now on, all new Wikibase Cloud instances will be automatically whitelisted for query federation from other Cloud instances every Monday.
    • Resolved a couple related bugs with QuickStatements: 1, 2
    • Resolved a couple bugs for Query Service: 1, 2
    • We disabled the Graph extension for everyone because of a major security risk.
    • Sped up our deployment process
    • Performed some software updates (Minikube and Kubernetes)
    • Fixed the misconfiguration that prevented using Lexemes in Wikitext via Lua modules
    • Fixed the link from the Contact page to FAQs
  • What we are working on:
    • We want to address the problem with user sessions randomly disappearing
    • We want to enable Query Builder and Quality Constraints
    • We are investing 1 week to fix a large amount of several low-hanging copy and UI problems all around the place

October 16th 2024

  • Status of cloud access
    • 1502 wikis in total, incl. 418 deleted ones
  • News from the product team:
    • We enabled federated queries across already existing instances. We plan to automatically include emerging instances into the scope of federation later.
    • The usability research to identify opportunities that will improve experience for novice users is complete.
    • We established an approach to measuring health metrics for the product and tried it out. We will build on this to settle on a sustainable solution in the future.
    • We performed some important software updates and addressed a security problem.
  • Here is what we are currently focusing on:
    • Including emerging instances into the scope of federation.
    • Addressing low-hanging bug fixes, features familiar from Wikidata, quick copy and UI improvements.

September 5th 2024

  • Status of cloud access
    • 1415 in total, incl. 390 that have already been deleted
  • News from the product team:
    • First important milestones achieved in the initiatives related to automation of testing and deployment
    • A new dialog introduced to learn more about the reasons why managers delete their Wikibase Cloud instances
    • First iteration of the feature to import entities from Wikidata implemented. Managers can import a fixed set of entities curated by the Wikibase Cloud team once per instance. The list of entities currently includes several properties (Mother, Father, Child, Image, Instance of, Subclass of) and items (Universe, World, Human, Berlin, Douglas Adams, The Hitchhiker's Guide to the Galaxy). In the future we might let the users choose the entities for import. Due to incompatibility with the Wikidata’s default label feature, the entities are currently imported in a state from July 28, 2024. The problem will be resolved when Wikibase Cloud switches to the same long term support version of Wikibase that Wikidata is running on.
    • Each new Wikibase now gets a new default main page that guides the users in kicking off their instance.
    • Users can now illustrate their entities with images from Wikimedia Commons.
    • We optimized how the search engine (Elasticsearch) is shared across instances. This reduces costs for maintenance, complexity, stability of the system and makes it easier to manage. We expect that this will also significantly increase our capacity to work on new features.
  • Here is what we are currently focusing on:
    • Synthesizing the results of the recent usability research to identify opportunities to improve experience for novice users
    • Revising our metrics and improving data instrumentation to be able to track our progress more reliably
    • Identifying the useful features familiar from Wikidata that could easily be enabled on all instances to improve feature parity between different Wikimedia products and facilitate transition of users between them
    • Identifying the confusing features that would benefit from better documentation
    • Exploring ways to differentiate between instances with valuable knowledge that is intended for re-use and instances used for experimentation and data preparation
  • Anything else worth mentioning:
    • We completed our first research of knowledge equity in Linked Open Data. The results are published here.

February 9th 2024

January 10th 2024

  • Status of cloud access
    • 824 Wikibases on cloud total
  • News from engineering:
  • Anything else worth mentioning:
    • Next Wikibase Live Session on 25th of January
    • End of the year survey share-out coming soon!
    • Wikidata & Wikibase office hours on 17th of January

More at /Updates/2023 and /Updates/2022