Closed DrCuriosity closed 5 days ago
Reordering the tests in the file breaks it so yes, there's definite interaction there where there shouldn't be.
Aha: I think the lru_cache is causing the misbehaviour, because the tests have identical calling parameters, so it's giving us cached results from the first test called when we call matched_rupture_sections_gdf
in the subsequent tests.
Adding a fix for this to #54.
Description
When working on some other refactoring for #51, I tried to execute the following test:
Which failed with:
This leads me to believe that, when run individually, it's not using the internal location resolution. It's possible that the patching on the previous
test_cached_ruptures_with_store
function is not being cleared down properly.In
cached.py
, this makes me think that the ground assumptions for testing have changed since this was originally written:The
DEPLOYMENT_STAGE
value isn't imported or referenced here anymore.I'm going to investigate and see if I can get the tests in this file to work in a proper unitary fashion.