Page MenuHomePhabricator

Add https://data.library.sh.cn/sparql to WDQS allowlist
Closed, ResolvedPublic

Description

We performed the analysis of external identifiers in Wikidata that have a SPARQL endpoint associated with them and discovered several candidates for the WD Allowlist (all already approved by @Lydia_Pintscher). This ticket nominates one of the discovered endpoints.

The Shanghai Library Open Data Platform is an initiative that publishes various datasets related to the library's digital humanities project, including knowledge bases (people, places, events, objects), literature (genealogy, archives, ancient books), and tools.
External IDs used over 400K times.

Details

Event Timeline

BTracy-WMF triaged this task as Medium priority.Oct 22 2025, 9:00 PM

Change #1199071 had a related patch set uploaded (by Bking; author: Bking):

[operations/puppet@production] wdqs: Add new endpoints to allowlist

https://gerrit.wikimedia.org/r/1199071

Change #1199071 merged by Bking:

[operations/puppet@production] wdqs: Add new endpoints to allowlist

https://gerrit.wikimedia.org/r/1199071

Hello Anton,

As requested, we added this endpoint to our WDQS allowlist. Please test a query and respond here with your sample query and its expected output if it is not working.

Thanks for your time,

Brian

@Lydia_Pintscher
Queries made on WDQS return an error
Unknown error: Status Code=403, Status Line=Forbidden, Response=Forbidden: Malicious request detected

I also can't seem to make SELECT queries on the web UI of this endpoint anymore (although a simple ASK {?s ?p ?o}) query returns true).

image.png (324×688 px, 20 KB)

I'm not sure what's going on, because when I suggested this endpoint, I was able to successfully get it to response with # of triples both from a Python notebook and from its own web UI.

I suggest we wait for a couple days to see if the issue on the endpoint is resolved.

@Lydia_Pintscher The endpoint is still not responding. I think we still need to discuss how we deal with endpoints that go offline

I suggest we resolve this ticket as Done, and deal with it in a separate initiative where we figure out what to do with the endpoints on the Allowlist that became non-operational (if we want to deal with them at all).

Closing per @Anton.Kokh request. Feel free to ping DPE SRE in a follow-up ticket if necessary.