Wikidata:Requests for permissions/Bot/IngeniousBot 3

From Wikidata
Jump to navigation Jump to search

IngeniousBot (talkcontribsnew itemsnew lexemesSULBlock logUser rights logUser rightsxtools)
Operator: Premeditated (talkcontribslogs)

Task/s: Adding identifiers to album items, based on existing identifiers.

Code:

Function details: Adding Spotify album ID (P2205), Apple Music album ID (U.S. version) (P2281), YouTube playlist ID (P4300), SoundCloud ID (P3040), Pandora album ID (P10138), Amazon Standard Identification Number (P5749), Tidal album ID (P4577), Deezer album ID (P2723), Yandex Music release ID (P2819), Anghami album ID (P10972), Boomplay album ID (SOON), and Napster album ID (SOON). Based on previously mentioned properties. --Premeditated (talk) 16:29, 22 March 2023 (UTC)[reply]

can you go into more detail about how this lookup will be done? link to some test edits? BrokenSegue (talk) 16:36, 22 March 2023 (UTC)[reply]
@BrokenSegue: Test edits. Lookups are based on a given album identifier like for example, Spotify album ID (P2205). UPC, Spotify artist ID (P1902), artist name, number of tracks, name of tracks, ISRC (P1243), and more are compared and looked up on other streaming services API/scrapping to match "identical" relases. I have made a scoring system where only relases that score 80% or better are added by the bot. The matches that does not get published will be saved to a file for later to be added to Mix'n'match, maybe. - Premeditated (talk) 23:50, 22 March 2023 (UTC)[reply]
I believe you are misusing the inferred from (P3452) property. Look at the description of that property in English. Please go and fix all the test edits you made. Maybe you want stated in (P248) or similar.
I think you should add a based on heuristic (P887) statement in the reference? Maybe to record linkage (Q1266546) or similar. This whole workstream seems really similar to what is/was being done by User:Soweego bot. Can you explain how you are different/the same. Maybe we should get input from @Hjfocs:.
Can you go into more detail about what is creating these scores? How did you verify the scores are meaningful? What kind of model are you using? Is your source code available? What " looked up on other streaming services API/scrapping to match "identical" relases " are you using. Etc. BrokenSegue (talk) 16:59, 23 March 2023 (UTC)[reply]
Hey folks, happy to give my 2 cents. I second BrokenSegue's comments: (based on heuristic (P887), record linkage (Q1266546)) reference nodes sound good. @Premeditated: interesting project: it would be great if you could share the code and tell us something more about it. Cheers, Hjfocs (talk) 22:57, 25 March 2023 (UTC)[reply]
What is the situation here?--Ymblanter (talk) 19:04, 23 June 2023 (UTC)[reply]