Tom Malaher's BrainScan

Java "wrapper" around SQLite. Also includes native binary libraries for Win, Mac, Linux x86, plus a pure-java implementation for fallback (e.g. sparc). Still not convinced that SQLite is the way to go, but lots of things seem to use it…

Looks like this does almost everything I need. Other than doing only tf-idf weighting. But it's a start. I wonder how hard it would be to add Document Index Graph weighting…