Skip to content

UL: change of sample list handling (matching to json files -> using DAS keys)

Sebastien Wertz requested to merge ulDAS into master

Mostly concerns the sample handling for now - lots of changes still needed in SF loading etc., so don't try to run things from this.

Instead of matching the samples with the JSON files from the skimmer, use directly DAS datasets in the template (bamboo uses dasgoclient to retrieve the file lists). We'll still use a template still to avoid having 4 entries (one per era) for each process (not to mention sub-processes). @umolinat now you have an example for filling in the datasets.

A local caching of the file lists is done, to avoid querying DAS every time. If for some reason the file lists are expected to change (e.g. change nanoAOD v8 -> v9), or change from remote (xrootd) reading (as done now) to using local copies, the cache should be removed!

Edited by Sebastien Wertz

Merge request reports