In the docs concerning results - https://recast-docs.web.cern.ch/recast-docs/workflowauthoring/running/#defining-recast-results - I see that you save json files as results, presumably because you want to use them for making reinterpretation. But should I also be able to save more “verbose” things like plots (i.e. PDFs and PNGs)? Or is this an abuse of RECAST results feature? If so, how should we save such outputs?
Hey @meehan, sorry for delayed reply here - another option is to save additional output (eg. .pdfs and .pngs) to the workdir and then download the workdir after the workflow is finished running (or access it directly if you’re running locally). See eg. monojet example:
Debugging plots and info copied to a directory outputdir in the steps.yml file: https://gitlab.cern.ch/recast-atlas/exotics/ANA-EXOT-2018-006/-/blob/master/specs/steps.yml#L327
This ouptutdir containing debugging files is then saved to the workdir under {workdir}/Interpretation_output: https://gitlab.cern.ch/recast-atlas/exotics/ANA-EXOT-2018-006/-/blob/master/specs/workflow.yml#L144
If the workflow is then run as recast run [workflow_name] --tag [tag_name]
then the output dir is accessible from the top level of the workflow specs dir (i.e. the level containing recast.yml at: [tag_name]/[stage_name]/Interpretation_output
But what about binary files like .root? Should I expect the same behavior with those as well?