Hello everybody!

I would like to detect two or more words, that are one to each other during the indexing process. I achieved that by implementing my own "stopword" filter, which always return true on accept function, and on the behind it checks if there is appearance of given syntagma. Problem is that when I detect syntagma appearance, I don't know which entity is being indexed (I have only token stream, not whole object).

So my questions are:
Is there a way to get object which is indexed and whose token stream is being passed to the filter?
and..
Is there some other way to detect syntagma?

tnx