[03:34:40] With regards to the pilot introduction of abstract articles in regular Wikipedias, here is how well each of the 5 test articles render in the 8 pilot languages, in terms of functions that make up the prose: : https://tools-static.wmflabs.org/bridgebot/f5bcc1ce/file_83453.jpg [03:58:34] Even those Igbo numbers are above the real level because I connected an early draft Z39155 (re @Earldridge Jazzed: With regards to the pilot introduction of abstract articles in regular Wikipedias, here is how well each of the 5 test articles ...) [03:59:33] The other thing I find strange about this test set is that the examples are so prominent that most languages already have articles about most of them. [04:20:07] I have another NLG function that supports Igbo (and 7-8 other languages), but the implementation and test cases have not been connected yet (re @u99of9: Even those Igbo numbers are above the real level because I connected an early draft Z39155) [04:53:00] I would like to show the abstract article for Q84 when rendered in Dagbani, which has only 297,867 labels on Wikidata : https://tools-static.wmflabs.org/bridgebot/0dd27f7c/file_83454.jpg [05:19:39] It looks like we need to turn two more sentences pink. I'll look into that sometime. (re @Earldridge Jazzed: I would like to show the abstract article for Q84 when rendered in Dagbani, where only 297,867 items on Wikidata have a Dagbani ...) [05:20:52] If it's ready I'm happy to connect it. Let us know which. (re @Earldridge Jazzed: I have another NLG function that supports Igbo (and 7-8 other languages), but the implementation and test cases have not been co...) [06:00:04] done (re @Al: Please consider connecting Z40327 as the equality function for Z96.) [09:22:24] Z40161, I mentioned this one a few days ago (re @u99of9: If it's ready I'm happy to connect it. Let us know which.) [09:30:00] Done. Sorry I didn't understand that you wanted it connected earlier. Feel free to apply for functioneer rights. (re @Earldridge Jazzed: Z40161, I mentioned this one a few days ago) [09:37:34] Also, the "See list of enabled languages" link (from e.g. here (https://abstract.wikipedia.org/wiki/Special:PreviewAbstract/ig/Q42)) points to "Abstract Wikipedia:Special:MyLanguage/Help:Abstract articles". Special pages are not supposed to be prefixed with the project namespace, which in this case is "Abstract Wikipedia:". For now, I created a redirect as [09:37:35] a workaround (from my t [09:37:35] esting, it's more of a soft redirect). [10:49:09] Anyways, thank you for connecting the implementation and tests in my function. I was doing this to expand language support for Z36193, but it seems like it is consistently reaching some sort of limit in the orchestrator (unless Z36197 was selected from the config). (re @u99of9: Done. Sorry I didn't understand that you wanted it connected earlier. Feel free [10:49:09] to apply for functioneer rights.) [14:52:26] I think perhaps you want the input to be an item reference (Z6091) rather than an item (Z6001). If you do want it to be an item, you probably don’t want it to default to the full item (via Z6821) but to be limited to statements with the relevant property type (ultimately via Z6820, perhaps using Z30120 or Z23762, as appropriate). (re @Earldridge Jazzed: Z40161, I [14:52:26] mentioned this [14:52:27] one a few days ago) [14:59:57] The global function Z28436 uses Wikidata item instead of item reference. (re @Al: I think perhaps you want the input to be an item reference (Z6091) rather than an item (Z6001). If you do want it to be an item,...) [15:15:02] Yes, but see Z30518, for example. Using a selective fetch in the call to Z28436 should be enough. (re @Earldridge Jazzed: The global function Z28436 uses Wikidata item instead of item reference.) [22:33:22] There are issues with object equality but I made Z40578 with the aid of Z40560. The performance isn’t great, mainly because all the function objects go into code, but Z40594 suggests it will typically be better than Z40013, but not for the top items in the list. I’m not sure what the default behaviour should be here. I currently return void and emit a message to debug [22:33:22] logs, b [22:33:23] ecause I don’t think an error is appropriate, but I guess that would become an error when it’s the target of function application. [22:33:24] One option is a function that is configured only for “mul” and defaults to no visible content in other languages. With “mul” as the view language, there could be visible error information, like in the debug logs, and fix-me links where useful. In this case, I don’t think other languages require visibility, because the absence of an infobox is already a strong signal. (r [22:33:24] [22:33:26] e @u99of9: If anyone is up for it, a code implementation would be very useful for Z40010)