Design a strategy for keeping Freebase to Wikidata ID mappings up to date
Closed, InvalidPublic
Actions

Assigned To

None

Authored By

	• Mholloway
	Aug 23 2019, 5:51 PM

Description

Recent investigation in T230734 of the Freebase to Wikidata mappings revealed that only a small number of the referenced Wikidata items have been deleted. Great! However, the mappings are from 2013 and it's probably that we'll get a significant number of new concepts that weren't in the DB at that time.

We should figure out where to record suggestions for which no Wikidata item was found, along with the suggested label text, so that we can keep the mappings up to date. We should also consider whether/how they should be broadcast for interested third parties.

Related Objects

Mentioned In: T247460: [L] File:Kameldornbaum_Sossusvlei.jpg stuck in a user's personal uploads queue
Mentioned Here: T352884: Undeploy and archive the MachineVision extension
T230734: Create a script to identify Wikidata items from the Freebase mappings that no longer exist

Event Timeline

• Mholloway created this task.Aug 23 2019, 5:51 PM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptAug 23 2019, 5:51 PM

LGoto triaged this task as Medium priority.Aug 28 2019, 3:40 PM

LGoto moved this task from Needs triage to Needs investigation on the Product-Infrastructure-Team-Backlog-Deprecated board.

I notice that the Wikipedia/Wikidata processing flow in Google's SLING project includes a step to create an up-to-date mapping from the latest dumps. We could probably make use of that.

Restricted Application added a project: Structured-Data-Backlog. · View Herald TranscriptFeb 12 2020, 5:07 PM

• Mholloway edited projects, added SDC-Statements (Machine-vision-depicts); removed Product-Infrastructure-Team-Backlog-Deprecated.Feb 12 2020, 5:07 PM

Assigning to myself as a reminder to perhaps integrate this if we go with SLING for SDAW

In T231105#5877459, @Mholloway wrote:

I notice that the Wikipedia/Wikidata processing flow in Google's SLING project includes a step to create an up-to-date mapping from the latest dumps. We could probably make use of that.