Page MenuHomePhabricator

Run maintenance script to update imagelinks table on Wikidata.org
Closed, ResolvedPublic

Description

In https://bugzilla.wikimedia.org/show_bug.cgi?id=46358 tracking of image usage on Wikidata was introduced. This seems to work quite well, but the imagelinks table is still lagging.

According to https://www.wikidata.org/wiki/Special:MostLinkedPages , image (P18) has 574,947 links. We also have some other image properties.

When I look at the imagelinks table:

MariaDB [wikidatawiki_p]> SELECT COUNT(*) FROM imagelinks WHERE il_from_namespace=0 LIMIT 1;
+----------+

COUNT(*)

+----------+

217659

+----------+
1 row in set (0.22 sec)

That's quite a few missing links. Can someone please run some sort of maintenance script to get this table up to date?


Version: unspecified
Severity: normal

Details

Reference
bz72470

Related Objects

Mentioned Here
P18 my paste!

Event Timeline

bzimport raised the priority of this task from to Needs Triage.Nov 22 2014, 3:47 AM
bzimport set Reference to bz72470.
bzimport added a subscriber: Unknown Object (MLST).

First we should run that query against prod and make sure that it's really a problem that needs a maint script. (vs. a problem of labs being out of date/corrupt)

(see bug 72413 and its linked bugs)

it is a problem, since we introduced support for image links recently.

I am not sure which script or solution is best for this but think this is possible somehow.

i see it there now:

https://commons.wikimedia.org/w/index.php?title=Special:GlobalUsage&limit=50&from=Grote-Kerk-Haarlem.jpg%7Cukwiki%7C604749&target=Grote-Kerk-Haarlem.jpg

I think the edit triggers a "links update" job that updates the tables, so it might not update instantaneously.

I ran a bot to purge all the pages. That's done so no longer needed to run a server side script.

Multichill set Security to None.