Skip to content

Fixes for most recent datasets from FTAG5

There are a few changes in here:

  • Removed a few input datasets from p3990, which correspond to DAODs with missing AODs (we weren't using these anyway)
  • Add some streamer operators for jets (to make debugging a bit easier)
  • Add an option to skip writing jets where some tracks are missing, and write defaults instead.
    • Also log these in an errors dataset.
  • Some rewriting of the submit-hbb.sh script, now I'm using xargs -P to submit multiple jobs in parallel.

The last "fix" is a bit troubling to me. I don't know why these jets are missing tracks. It's the third leading jet in the subjet, so it's possible that the ordering is being shuffled somehow.

In any case, it seems to be extremely rare: there's one jet, in all the datasets we typically run over, which seems to be have invalid track links. I need to track this down but we should be able to continue other studies in the meantime.

Edited by Dan Guest

Merge request reports