I am looking to automate the process of setting the stress of Russian words. There are already tools that do this (RussianGram or https://morpher.ru/accentizer/), but nothing open source which can be executed offline as far as I know.
So what I am looking for is a dataset that contains Russian words with their stresses and correct forms of е or ё. I have had some success so far by using the dataset provided by https://kaikki.org/, which contains Russian words (and English translations) and many of their inflections. The only problem is that right now it does not support the short form of many verbs (like завершён). It is possible that the creator will parse the sites more so that the short form of verbs will be included in the dataset as well. (That was his plan I think). Edit: Now all this information is available, making this maybe the best resource.
Another option would be to attempt to use the https://github.com/tatuylonen/wiktextract project, which right now only works for the English Wiktionary, on the Russian Wiktionary or to try to parse the Russian Wiktionary dumps, but both options would probably be a lot of work.
I also found one project which attempted to parse the Russian Wiktionary: http://whinger.krc.karelia.ru/soft/wikokit/index.html Unfortunately the database dump contains the files misencoded in some form, which is not trivial to fix (at least for me).
Does anyone have any other ideas?
