Ensure correct information about Wikimedia sites in the Sites facility on the Wikimedia cluster.
Open, MediumPublic
Actions

Assigned To

None

Authored By

	daniel
	Jun 10 2016, 10:42 AM

Description

Currently, SiteLookup is backed by the information in the sites table. The information in this table is incorrect for many wikis in the Wikimedia cluster. In particular:

the site language is set to always be the same as the interwiki prefix. This is not correct for sites like simple.wikipedia.org, which should have the language code "en". This could be solved by applying the $wgDummyLanguageCode mapping when importing (or when reading, as proposed by T137534: Map dummy language codes in sites)
the language code interwiki prefixes (e.g. "fr" or "en") are set to always point to wikipedia; this is incorrect for sites like wiktionary, where "fr" should point to fr.wiktionary.org, not fr.wikipedia.org.

Possible solutions:

fix the import script
manually fix the data in sites table
switch to a different system for backing the Sites facility alltogether, see T113034: RFC: Overhaul Interwiki map, unify with Sites and WikiMap.

Related Objects
Search...

Status	Assigned	Task
Open	None	T124286 [Epic] Wikidata language support
Open	None	T39459 [Task] Don't try to add labels in non-existing languages: restrict to Language::isKnownLanguageTag
Open	None	T145535 language code disparity (Wikidata tracking)
Open	None	T66649 'Add link' function in Wikipedia creates items with wrong language code
Open	None	T137536 Use correct term language when creating items from the client
Open	None	T137534 Map dummy language codes in sites
Open	None	T137538 Use interwiki code from Sites facility to create language links.
Open	None	T137537 Ensure correct information about Wikimedia sites in the Sites facility on the Wikimedia cluster.
Open	None	T135149 Implement a SiteLookup based on a nested array structure.

Event Timeline

daniel renamed this task from Ensure currect information about Wikimedia sites in the Sites facility on the Wikimedia cluster. to Ensure correct information about Wikimedia sites in the Sites facility on the Wikimedia cluster..Jun 10 2016, 10:42 AM

daniel created this task.

daniel added a subtask: T135149: Implement a SiteLookup based on a nested array structure..

daniel added a project: MediaWiki-Site-system.Jun 10 2016, 10:44 AM

daniel removed a project: MediaWiki-extensions-WikibaseRepository.

daniel mentioned this in T137538: Use interwiki code from Sites facility to create language links..Jun 10 2016, 10:49 AM

daniel added a parent task: T137538: Use interwiki code from Sites facility to create language links..

daniel updated the task description. (Show Details)Jun 10 2016, 10:56 AM

Currently, SiteLinksRdfBuilder and OtherProjectsSitesGenerator depend on the wrong language codes.

Addshore moved this task from incoming to in progress on the Wikidata board.Jun 22 2016, 1:01 PM

Smalyshev subscribed.Jul 6 2016, 5:42 PM

thiemowmde triaged this task as Medium priority.Sep 6 2016, 10:41 AM

thiemowmde moved this task from in progress to hold on the Wikidata board.