The pipeline (since introduction of models) has no way to generate first-time historic data in a cohort - this was being silently skipped
We also have no way to generate that historic data from a results file where it wasn't generated first time around.
Proposed Changes
Edits the date-filtering logic to satisfy the condition "if a location for historic data exists in config, but no prior file is found - generate one"
adds a helper script which can generate a historic data file from real results (limited purpose)
also adjusts some helper files to prevent the automatic copying of data into a test bucket (that whole process needs a think, some projects do not have a test bucket)
adds a helper script to stratify the list of saved cases we are given by collaborators, translate the IDs to CPG internal IDs, and sort into separate project groups
Fixes
Proposed Changes
Checklist