def test_normalize_df_no_normalize_by():
pheno_df = pd.DataFrame(
[
{
"person_id": 112233,
"i1.m1": 10
},
{
"person_id": 445566,
"i1.m1": 20
},
],
columns=["person_id", "i1.m1"],
)
expected = pd.DataFrame(
[
{
"person_id": 112233,
"i1.m1": 10,
"normalized": 10
},
{
"person_id": 445566,
"i1.m1": 20,
"normalized": 20
},
],
columns=["person_id", "i1.m1", "normalized"],
)
normalized = PhenoTool._normalize_df(pheno_df, "i1.m1")
assert pd.DataFrame.equals(normalized, expected)
def test_normalize_df():
pheno_df = pd.DataFrame(
[
{
"person_id": 112233,
"i1.m1": 1e6,
"i1.m2": 1e3
},
{
"person_id": 445566,
"i1.m1": 2e12,
"i1.m2": 1e-3
},
],
columns=["person_id", "i1.m1", "i1.m2"],
)
normalized = PhenoTool._normalize_df(pheno_df,
"i1.m1",
normalize_by=["i1.m2"])
assert list(normalized) == ["person_id", "i1.m1", "i1.m2", "normalized"]
assert normalized["person_id"][0] == pytest.approx(112233)
assert normalized["person_id"][1] == pytest.approx(445566)
assert normalized["i1.m1"][0] == pytest.approx(1e6)
assert normalized["i1.m1"][1] == pytest.approx(2e12)
assert normalized["i1.m2"][0] == pytest.approx(1e3)
assert normalized["i1.m2"][1] == pytest.approx(1e-3)
assert normalized["normalized"][0] == pytest.approx(-0.000299,
abs=1e-4) # FIXME:
assert normalized["normalized"][1] == pytest.approx(0.000488,
abs=1e-2) # FIXME:
def test_normalize_df_does_not_contain_normalize_measure_id():
pheno_df = pd.DataFrame(
[
{
"person_id": 112233,
"i1.m1": 1e6,
"i1.m2": 1e3
},
{
"person_id": 445566,
"i1.m1": 2e12,
"i1.m2": 1e-3
},
],
columns=["person_id", "i1.m1", "i1.m2"],
)
with pytest.raises(AssertionError):
PhenoTool._normalize_df(pheno_df, "i1.m1", normalize_by=["i1.m3"])
def test_normalize_df_by_empty_df(fake_phenotype_data):
pheno_df = fake_phenotype_data.get_persons_values_df(["i1.m1", "i1.m2"],
person_ids=[])
with pytest.raises(AssertionError):
PhenoTool._normalize_df(pheno_df, "i1.m1", "i1.m2")