Currently this script runs against all datasets (or all with schema = ['image']). We need t be able to run against single datasets, specified in Command Line arg, as well as against all. It should be easy to do this by refactoring to runner a script (with argparse) + a separate file of functions.
Logging -> reports is a bit of an ad hoc mess right now. Should be refactored to nicely structured JSON with standard keys for test name, number of failing inds + list of failing inds.
With these improvements in place, we should look at running in Jenkins - coupled to curation loading (dataset names can be extracted from curation record YAML files)
Currently this script runs against all datasets (or all with schema = ['image']). We need t be able to run against single datasets, specified in Command Line arg, as well as against all. It should be easy to do this by refactoring to runner a script (with argparse) + a separate file of functions.
Logging -> reports is a bit of an ad hoc mess right now. Should be refactored to nicely structured JSON with standard keys for test name, number of failing inds + list of failing inds.
With these improvements in place, we should look at running in Jenkins - coupled to curation loading (dataset names can be extracted from curation record YAML files)