check SPARQL query for Pathway DataNodes with WikiPathways Xrefs

[egonw]

@egonw In https://www.wikipathways.org/index.php/Pathway:WP3172, the pathway gastric pepsin release links to https://www.wikipathways.org/instance/WP2601, and Gastric Gastrin Release links to https://www.wikipathways.org/instance/WP2599, but these pathways don't exist. Also, the pathway is from Bos taurus, but Parietal Cell production links out to a human pathway. Is this intended? The remaining two Pathways in here look clickable, but are not.

Interesting. Some time ago I created a test to detect links to deleted pathways. Apparently, these instances are not picked up properly.

Originally posted by @egonw in wikipathways/wikipathways-development#55 (comment)

[egonw]

Thought: may have to do with the ^^xsd:string situation with Apache Jena/VOS.

[egonw]

No, it seems not. To be continued.

[egonw]

Okay, I remember now. The current tests runs in batches, and because of that, it uses https://github.com/bridgedb/tiwid/blob/main/data/wikipathways.csv but that file is intrinsically behind.

What I will do, and this too is not a perfect solution, is write a query to be run (and only there) for the Jenkins job that runs on all pathways (including the uncurated/new ones).

[egonw]

Okay, I remember now. The current tests runs in batches, and because of that, it uses https://github.com/bridgedb/tiwid/blob/main/data/wikipathways.csv but that file is intrinsically behind.