This is a MediaWiki bot that grabs a list of lexemes off of Wikidata's JSON dumps that have senses in a specific language (configured by the bot operator), checks if a specific Wiktionary has any entry about it already and makes a new one that jus…
This is a MediaWiki bot that grabs a list of lexemes off of Wikidata’s JSON dumps that have senses in a specific language (configured by the bot operator), checks if a specific Wiktionary has any entry about it already and makes a new one that just uses the lexeme’s data via a template if not.
Operating the bot requires credentials such as a Wikimedia bot password unless you comment out lines 8 to 11 of bot.py which can be done safely if you do not use a Wikimedia account. But if you do not want to do that, you can extract the lexeme IDs (along with other information the bot needs) using the extract_lemmas.py script after downloading https://dumps.wikimedia.org/other/wikibase/wikidatawiki/latest-lexemes.json.gz (~570 MB).
The demo link shows entries made by the bot on Bengali Wiktionary.
There is no GUI because it is intended to be used from CLIs (preferably on Wikimedia’s Toolforge cloud environment where the dumps are readily available but it can be run anywhere Python runs).
I used Gemini and Claude Sonnet (Thinking) to generate the script that extracts lexeme ids from Wikidata’s JSON-GZIPped dumps. As the project neared completion, I also had Claude Sonnet review my code for bugs (and it found three that I subsequently fixed on my own).
