Accept repositories which are both preferred and normal ranked

[davidhicks]

trafficstars

To cover the case where a software project has an official repository that is not GitHub, and this repository has a preferred rank, as well as an official GitHub mirror, which has a normal rank, I suggest a change to the query in free_software_items.rq to include GitHub repositories of either rank:

https://query.wikidata.org/#PREFIX%20schema%3A%20%3Chttp%3A%2F%2Fschema.org%2F%3E%0A%0ASELECT%20%3Fproject%20%3FprojectLabel%20%3Frepo%20WHERE%20%7B%0A%20%20%23%20Items%20with%20a%20repo%20on%20github%0A%20%20%3Fproject%20p%3AP1324%20%5Bps%3AP1324%20%3Frepo%3B%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20wikibase%3Arank%20%3Frank%3B%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%5D.%0A%0A%20%20%23%20Include%20both%20preferred%20and%20normal%20ranked%20repos%2C%20to%20capture%20mirrors%20of%20official%20repos%0A%20%20FILTER%28%3Frank%20IN%20%28wikibase%3APreferredRank%2C%20wikibase%3ANormalRank%29%29.%0A%0A%20%20%23%20Exclude%20projects%20not%20hosted%20on%20github%0A%20%20FILTER%28contains%28STR%28%3Frepo%29%2C%20%22%3A%2F%2Fgithub.com%2F%22%29%29.%0A%0A%20%20SERVICE%20wikibase%3Alabel%20%7B%20bd%3AserviceParam%20wikibase%3Alanguage%20%22en%22%20.%20%7D%0A%7D%0AORDER%20BY%20LCASE%28%3FprojectLabel%29

[davidhicks]

PREFIX schema: <http://schema.org/>

SELECT ?project ?projectLabel ?repo WHERE {
  # Items with a repo on github
  ?project p:P1324 [ps:P1324 ?repo;
                    wikibase:rank ?rank;
                   ].

  # Include both preferred and normal ranked repos, to capture mirrors of official repos
  FILTER(?rank IN (wikibase:PreferredRank, wikibase:NormalRank)).

  # Exclude projects not hosted on github
  FILTER(contains(STR(?repo), "://github.com/")).

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
ORDER BY LCASE(?projectLabel)

[konstin]

That's an interesting case! Do you have any projects specifically to which that would apply? Do those projects use github releases or would we only get tags?

[davidhicks]

@konstin One example is Chromium, see https://www.chromium.org/developers/github-collaboration where an official mirror exists at https://github.com/chromium/chromium

Both releases and tags show up in GitHub for this project, but I suspect the releases are just placeholders created automatically from tags--it's the same information in each.

[Ordoviz]

Performing the change would enable the github-wikidata-bot on around 800 items:

SELECT ?project ?projectLabel ?repo ?bestRepo WHERE {
  ?project p:P1324 [ps:P1324 ?repo;
                    wikibase:rank ?rank;
                   ].

  FILTER(?rank IN (wikibase:PreferredRank, wikibase:NormalRank)).

  # Only include projects with a repo on GitHub
  FILTER(contains(STR(?repo), "://github.com/")).
  
  # Exclude projects with main repo on GitHub
  ?project wdt:P1324 ?bestRepo.
  FILTER(!contains(STR(?bestRepo), "://github.com/")).

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}