I just follow the README.html, so I do:
$ hadoop jar mirex-0.2.jar nl.utwente.mirex.TrecRun 'ClueWeb09_Anchors/*' TrecOut wt09-topics.txt
but TrecRun does not accept this:
$ hadoop jar mirex-0.2.jar nl.utwente.mirex.TrecRun 'ClueWeb09_Anchors/*' TrecOut wt09-topics.txt
Usage: TrecRun [inputFormat] inputFiles topicFile outputFile
inputFormat: either WARC or KEYVAL; default WARC
In fact, the whole argument reading seems to be broken. the attached patch fixes it.
While I'm at it, I also noticed the default inputFormat specified in the usage output is incorrect (it defaults to KEYVAL), fixed that as well.
although QueryTermCount uses WARC by default, so you should probably pick one default and use the same default in both...
if you have issues downloading the patchfile from sourceforge, you can also get it here:
http://pastie.org/3303313
Thanks for the patch! We will need a few days to fix this. In the meantime, don't give up. Best, Djoerd.