This is an initial stab at fulltext and field-specific search.  It uses PyLucene/JCC,
which calls into the apache lucene implementation through the JNI.  The indexing is
done via a manage.py command "index" which can do either a full or incremental index
of the database.

There's still a few things left before this is perfect.  I need to figure out:
1. Review indexing.  I'm blocked on the fact that I can't import reviews now due
   to that bug that's being discussed on the list.
2. Filename indexing needs a better tokenizer.  I'll probably have to tokenize it
   myself.
3. It doesn't index dates yet.
4. This needs to be documented a lot -- especially how to build PyLucene, which is
   a total pain.

Used it a lot!  It's sweet!