Based on pywikibot, see example of how a simple version works -> https://byabbe.se/2020/09/15/writing-structured-data-on-commons-with-python

The most reasonable structure would be for it to be as small and content-independent as possible and to able to take a spreadsheet/csv file as input.

See also: the code used by Nordiska museet to add ID's -> https://github.com/Ambrosiani/batch-SDC (forked)

Until T262750 is solved it should also be able to filter out files by who uploaded them, so as not to send a lot of notification to users who have uploaded many files. It's not crucial for the Fataburen upload since it will only affect files uploaded by us.