we use language-specifyc analyzer.
in my understanding, language-specific analyzer is a combination of
standard tokenizer plus customized stop words plus language-specific
stemming (I don't know if that last thing is done through snowball, though).
I want to add synonyms feature, but I'm guessing I can not just add the
synonym filter to the language-specific analyzer, since the tokens that the
synonym filter would receive would be 'stemmed' versions.
I was thinking if there is anywhere I can see what exact combination of
filters the language analyzers are made of, I could plug the synonym filter
before the stemming.