AllenInstitute / ipfx

computes intrinsic cell features from intracellular electrophysiology data
https://ipfx.readthedocs.io/en/latest/
Other
26 stars 36 forks source link

MIES NWB integration tests should not argue global ontology filepath to run_pipeline #351

Open NileGraddis opened 4 years ago

NileGraddis commented 4 years ago

Passing a stimulus ontology filepath to run_pipeline queries LIMS in order to get up-to-date stimulus ontology info, then writes it to that path. This test reads in an input json, which might (in practice: does) contain an ontology path, then argues that to run_pipeline. This:

  1. overrides the existing file at that path, if the test runner is so permitted
  2. fails, if not

Instead, we should make two changes:

  1. (must do) This test should make sure that any paths it writes to are within its tmpdir
  2. (should do) We should deprecate the existing stimulus_ontology_file parameter of run_sweep_extraction and replace it with an interface whose default behavior is more sensible.
wbwakeman commented 4 years ago

1 (must part) was completed in PR #352

2 (should part) remains to be done