I am unsure if this is technically feasible, but my ask is as follows: when displaying images from Wikimedia in image search results, it would be useful to be able to exclude the TLDs of image sources.
I was trying to find a map of Sri Lankan rail lines and the first image that came up was in Russian (this image was also shown high up in Google's rankings). That image was uploaded to Wikimedia from a source with a domain ending in .ru. I have blocked .ru in my personalised results, so to be able to filter these out secondarily in popular repositories like Wikimedia Commons would be good. See the below screen recording for reference.
I find this particularly relevant in my case because I find that TLDs correspond strongly with source language. I exclude .ru TLDs because I do not want Russian-language results, and I previously found that Kagi's use of Yandex raises those results disproportionally.
For reference, my region is set to the UK/GB.
