It could kind of work. I'm looking into Apache Tika to parse text from document and Apache Lucene to index contents for full text search support.