It should be possible to deploy the Serbian analysis chain with the Serbian stemmer (or some variant) to the Serbo-Croatian (sh/442K Wikipedia articles), Croatian (hr/185K Wikipedia articles), and Bosnian (bs/77K Wikipedia articles) wikis. (See the English Wikipedia article on Serbo-Croatian; the standard varieties of the language differ, but have the same basic grammar, so the stemmer should do as good a job on them as on Serbian.)
The task here is to review the results of the analysis chain and make sure there aren't any surprises. Then, we can re-index wikis in these three languages.