that works by just tarring up the current hour's worth of files. If you need to narrow this down further just the inputs to the particular analysis would be needed as shown in section 3 of the README. Also, things like the 'cdl', 'time' and 'static' subdirectories should be ...