New exported format "tabular data" on JSON, use case Wikimedia Commons

[fititnt]

• https://www.mediawiki.org/wiki/Help:Tabular_Data
• https://www.mediawiki.org/wiki/Extension:JsonConfig/Tabular
• https://www.mediawiki.org/wiki/Extension:Scribunto/Lua_reference_manual#mw.loadData
• https://specs.frictionlessdata.io/data-package/#metadata
  • https://specs.frictionlessdata.io/tabular-data-package/#why-csv

---

We need a forma which can be saved as parseable tabular data on Wikimedia Commons

[fititnt]

./999999999/0/1603_1.py --codex-de 1603_25_1 --codex-in-tabulam-json | jq

Not yet with real data, but is something

1603_1.py vs hxltmcli

• See https://github.com/EticaAI/hxltm

In theory, the ideal would be move this exporter to hxltm tools (such as other formats we document on https://hxltm.etica.ai/) but it would take a lot of time to move (aka generalize) the logistics of numerordinatio directly on hxltm. HXLTM is more general propuse, and numerordinatio is a very strict subset. Add to this that to generate the data packages to work on Mediawiki commons, it would need be aware of the 1603:1:51 (which have the language mappings) so... at least for now, this exporter is similar to the way we generate the book formats.

[fititnt]

Better output (but still fake data).

Anyway, I think we can have some MVP of this format soon. But it seems that the default way to show data tables with localized field on MediaWiki gives not hint for the user that it is showing a fallback language (not the target language) and this can lead to confusion).

We may even need to simulate a language that would never be target language from others to force the behavior. But it would still be better to do it via the programming, not on the data.