Add test for NCHS national & proportional data #1942
Conversation
nchs_mortality/tests/test_run.py
Outdated
| gmpr = GeoMapper() | ||
| national_pop = gmpr.get_crosswalk("nation", "pop") | ||
| us_pop = national_pop.loc[national_pop["nation"] == "us"]["pop"][0] |
There was a problem hiding this comment.
This is just repeating the same logic added in #1938 ... I think its better to hardcode the US population value here in case GeoMapper starts behaving differently.
nchs_mortality/tests/test_run.py
Outdated
| prop_value = pytest.approx(prop.iloc[0]["val"], 0.0001) | ||
| INCIDENCE_BASE = 100000 | ||
| assert(prop_value == num.iloc[0]["val"] / us_pop * INCIDENCE_BASE) |
There was a problem hiding this comment.
Is this approx comparison necessary?? These calculations should be deterministic since the test and reference data were generated on the same machine using virtually (if not exactly) the same operations.
There was a problem hiding this comment.
There was a problem hiding this comment.
well then, it might not be a bad idea to hardcode this too
|
@melange396 fixed the test, re-requesting review! I've also realized the test would technically pass if both national and proportional values were 0s, so added a check for that as well. |
Description
Adds an extra test for #1912 + #1938, making sure that national data is generated, and prop(ortional) signals are correctly calculated.