this patch adds support for the hamshahri collection, approximately 160,000 persian documents.
There are two sets of corresponding queries and relevance judgements.
Also, I fixed two things in the readme instructions
1) I forgot to specify the class name of QueryDriver (the thing that actually runs the relevance tests)
I also forgot to specify -Dfile.encoding=UTF-8 for this QueryDriver (currently needed when running the command-line)
2) In the example .alg file, i forgot to specify content.source.encoding=UTF-8.