Wikidata:Property proposal/PyPI trove classifier

PyPI trove classifier

Originally proposed at Wikidata:Property proposal/Authority control

   Done: PyPI trove classifier (P12468) (Talk and documentation)

Description	standardized classification system for software, used and maintained by the Python Package Index
Represents	PyPI trove classifier (Q113568933)
Data type	String
Domain	item
Allowed values	Regex: ([\w\s]+)(::([\w\s]+)::([\w\s]+))? (a maximum of 3 strings (with or without whitespace), separated by colons.)
Example 1	MIT License (Q334661) → License :: OSI Approved :: MIT License
Example 2	Flask (Q289281) → Framework :: Flask
Example 3	macOS (Q14116) → Environment :: MacOS X and Operating System :: MacOS
Example 4	English (Q1860) → Natural Language :: English
Example 5	Rust (Q575650) → Programming Language :: Rust
Example 6	File Transfer Protocol (Q42283) Topic :: Internet :: File Transfer Protocol (FTP)
Source	https://pypi.org/classifiers/
Planned use	I have data which relates licenses' Trove Classifiers to Wikidata items, I plan to add those to Wikidata and use them for determining copyright licenses from PyPI projects.
Number of IDs in source	805, but more can be added
Expected completeness	eventually complete (Q21873974)
See also	PyPI project (P5568)
Wikidata project	None

Motivation

I proposed this so I can gain the ability to query for a specific trove classifier via SPARQL. Tomodachi94 (talk) 18:23, 21 August 2022 (UTC)

Discussion

•  Support —Ismael Olea (talk) 18:40, 22 August 2022 (UTC)
•  Support, can also help suggest other tags (like instance-of from the top-level colon separated tag part). Arlo Barnes (talk) 23:51, 22 August 2022 (UTC)
•  Note Trove classifiers can be programmatically accessed from Python via the trove-classifiers project on PyPI, or in plain-text at [1]. Tomodachi94 (talk) 20:28, 23 August 2022 (UTC)
• Dachary
  Metamorforme42
  NMaia
  Valerio Bozzolan
  MichaelSchoenitzer
  Jasc PL
  LiberatorG
  Dexxor
  Waldyrious
  Iwan.Aucamp
  Airon90
  Ainali
  Haansn08
  So9q
  Tomodachi94
  Zblace
  Labdajiwa
  Mind_Booster_Noori
  
  Notified participants of WikiProject Informatics/FLOSS as this might be relevant to them. Tomodachi94 (talk) 21:23, 23 August 2022 (UTC)
•  Comment tentative support, but I wonder if this would mean duplicating the maintenance of those classifiers? They are not fixed and can evolve throughout the lifetime of the software. These are most commonly related to adding or removing support for Python versions, but it also happens for other classifiers (examples: 1, 2, 3). Perhaps we should make sure there's a bot or automated process to update these (or at least a plan to have that) before we start adding this property. --Waldyrious (talk) 09:44, 24 August 2022 (UTC)
  •  Support I had misunderstood to which items this property was to be applied. My previous comment was made under the assumption that we'd be adding e.g. a statement PyPI trove classifierFramework :: Flask to items for software which relies on Flask (like Wembedder (Q63379435)), rather than as an external ID to Flask (Q289281) itself. Now that I have understood that the proposal is about doing the latter, my previous concern of duplicated maintenance of classifiers for each software is moot, so I changed my position to support. --Waldyrious (talk) 10:44, 11 February 2024 (UTC)
•  Comment I think that perhaps before we create a property, a Mix'n'Match (or other way to import the entire set) should be done first (with a way using either instance of (P31) or subclass of (P279)) so that we easily can create property constraints from day 1 to make sure not trash data is created in the beginning. Ainali (talk) 14:22, 24 August 2022 (UTC)

I have been working on creating 'pairs' for trove classifiers and their respective topics on Wikidata in OpenRefine. I hope to get a flood account and upload those soon. Tomodachi94 (talk) 18:10, 24 August 2022 (UTC)

•  Support, an important property for Python.--Arbnos (talk) 21:33, 17 January 2024 (UTC)
•  Oppose Datatype should be 'item' instead of 'external-id'. The items for each classifier should be created before. --Tinker Bell ★ ♥ 23:14, 9 February 2024 (UTC)
  •  Support --Tinker Bell ★ ♥ 18:46, 12 February 2024 (UTC)

  @Tomodachi94: could you please provide formatter URL, or change the datatype as Tinker Bell mentioned. Regards, ZI Jony ^(Talk) 05:38, 10 February 2024 (UTC)

• @ZI Jony: I've updated the property proposal to reflect what I believe I was thinking at the time. The theory is that this would be added as a string value to each equivalent item to enable detecting equivalent Wikidata statements from classifiers. Sorry for any confusion, this was my first property proposal so some things were confusing :) —Tomodachi94 (talk) 23:00, 10 February 2024 (UTC)

  Also pinging @Olea, Arlo Barnes, Waldyrious, Ainali, Arbnos, Tinker Bell: to re-vote due to substantial changes to the proposal.

   point of clarification: You simply changed the datatype and regex, right? My previous comment stays the same, if so. Arlo Barnes (talk) 23:05, 10 February 2024 (UTC)

  Yes. —Tomodachi94 (talk) 23:10, 10 February 2024 (UTC)

• @Tomodachi94, Olea, Arlo Barnes, Waldyrious, Arbnos, Tinker Bell:  Done as PyPI trove classifier (P12468). Regards, ZI Jony ^(Talk) 12:48, 17 February 2024 (UTC)