Use Alex's input_files list to get a complete list of containers.
Added a datasets.py to grab the datasets.
Add a few running options so we can run on multiple sets of contrainers (--help below).
Build all the dask graphs and then run on them at once.
I've seen a few side effects:
For multiple datasets we have lots of log messages.
Sometimes when running with the xaod_small cut, we get files with no events in them, even with steps set to 1. This ticles the uproot bug. If you see that, until a new version of uproot is released, you'll need to pip install git+https://github.com/scikit-hep/uproot5.git@jpivarski/fix-the-bug-gordon-found-on-slack.
input_fileslist to get a complete list of containers.datasets.pyto grab the datasets.--helpbelow).daskgraphs and then run on them at once.I've seen a few side effects:
xaod_smallcut, we get files with no events in them, even withstepsset to 1. This ticles theuprootbug. If you see that, until a new version ofuprootis released, you'll need topip install git+https://github.com/scikit-hep/uproot5.git@jpivarski/fix-the-bug-gordon-found-on-slack.Fixes #90