Post date: Feb 19, 2013 4:31:14 PM
To prevent any potential issues with people not having their submission files sorted before passing them to evaluation.py, I have updated the script to do the sorting internally and check if the sets of IDs in the "actual" and "predicted" files are the same.
If your files were sorted previously, this will not change your results, but now you don't have to worry about doing that.