def test_calc_split_by_sex(mocker, fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
variants = Counter({
"f4.p1": 1,
"f5.p1": 1,
"f7.p1": 1,
"f10.p1": 1,
"f12.p1": 1,
"f16.p1": 1,
"f24.p1": 1,
"f25.p1": 1,
"f30.p1": 1,
"f32.p1": 1,
})
merged_df = PhenoTool.join_pheno_df_with_variants(pheno_tool.pheno_df,
variants)
mocker.spy(PhenoTool, "join_pheno_df_with_variants")
mocker.spy(pheno_tool, "_calc_stats")
pheno_tool.calc(variants, sex_split=True)
assert pheno_tool._calc_stats.call_count == 2
assert PhenoTool.join_pheno_df_with_variants.call_count == 1
call_arg_df, call_arg_sex_split = pheno_tool._calc_stats.call_args_list[0][
0]
assert merged_df.equals(call_arg_df)
assert call_arg_sex_split is Sex.M
call_arg_df, call_arg_sex_split = pheno_tool._calc_stats.call_args_list[1][
0]
assert merged_df.equals(call_arg_df)
assert call_arg_sex_split is Sex.F
def prepare_pheno_tool_adapter(self, data):
study_wrapper = self.gpf_instance.get_wdae_wrapper(data["datasetId"])
if not (study_wrapper and study_wrapper.phenotype_data.has_measure(
data["measureId"])):
return None
if study_wrapper.is_remote:
return RemotePhenoToolAdapter(study_wrapper.rest_client,
study_wrapper._remote_study_id)
helper = PhenoToolHelper(study_wrapper)
pheno_filter_family_ids = helper.pheno_filter_persons(
data.get("familyFilters"))
study_persons = helper.genotype_data_persons(data.get("familyIds", []))
person_ids = set(study_persons)
tool = PhenoTool(
helper.genotype_data.phenotype_data,
measure_id=data["measureId"],
person_ids=person_ids,
family_ids=pheno_filter_family_ids,
normalize_by=data["normalizeBy"],
)
return PhenoToolAdapter(tool, helper)
def test_calc_empty_variants(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
variants = Counter({})
res = pheno_tool.calc(variants)
assert isinstance(res, PhenoResult)
assert res.positive_count == 0
assert res.positive_mean == 0
assert res.positive_deviation == 0
assert res.negative_count == len(pheno_tool.pheno_df)
assert res.negative_mean
assert res.negative_deviation
assert res.pvalue == "NA"
resM, resF = pheno_tool.calc(variants, sex_split=True).values()
assert isinstance(resM, PhenoResult)
assert isinstance(resF, PhenoResult)
assert resM.positive_count == 0
assert resM.positive_mean == 0
assert resM.positive_deviation == 0
assert resM.negative_count == res.negative_count - resF.negative_count
assert resM.negative_mean
assert resM.negative_deviation
assert resM.pvalue == "NA"
assert resF.positive_count == 0
assert resF.positive_mean == 0
assert resF.positive_deviation == 0
assert resF.negative_count == res.negative_count - resM.negative_count
assert resF.negative_mean
assert resF.negative_deviation
assert resF.pvalue == "NA"
def test_get_normalize_measure_id_same_measure(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
measure_id = pheno_tool._get_normalize_measure_id({
"measure_name": "m1",
"instrument_name": "i1"
})
assert measure_id is None
def test_calc(mocker, fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
variants = Counter({
"f4.p1": 1,
"f5.p1": 1,
"f7.p1": 1,
"f10.p1": 1,
"f12.p1": 1,
"f16.p1": 1,
"f24.p1": 1,
"f25.p1": 1,
"f30.p1": 1,
"f32.p1": 1,
})
merged_df = PhenoTool.join_pheno_df_with_variants(pheno_tool.pheno_df,
variants)
mocker.spy(PhenoTool, "join_pheno_df_with_variants")
mocker.spy(pheno_tool, "_calc_stats")
pheno_tool.calc(variants, sex_split=False)
pheno_tool._calc_stats.assert_called_once()
PhenoTool.join_pheno_df_with_variants.assert_called_once()
call_arg_df, call_arg_sex_split = pheno_tool._calc_stats.call_args_list[0][
0]
assert merged_df.equals(call_arg_df)
assert call_arg_sex_split is None
def test_get_normalize_measure_id_no_instrument_name(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
measure_id = pheno_tool._get_normalize_measure_id({
"measure_name": "m3",
"instrument_name": None
})
assert measure_id == "i1.m3"
def test_get_normalize_measure_id_non_dict_measure(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
with pytest.raises(AssertionError):
pheno_tool._get_normalize_measure_id(["measure"])
with pytest.raises(AssertionError):
pheno_tool._get_normalize_measure_id("measure")
with pytest.raises(AssertionError):
pheno_tool._get_normalize_measure_id(None)
def test_init_with_person_ids(fake_phenotype_data):
pheno_tool = PhenoTool(
fake_phenotype_data,
"i1.m1",
person_ids=["f1.p1", "f3.p1", "f5.p1", "f7.p1"],
)
assert set(pheno_tool.pheno_df["person_id"]) == set(
["f1.p1", "f3.p1", "f5.p1", "f7.p1"])
def test_get_normalize_measure_id_non_existent(fake_phenotype_data,
measure_name, instrument_name):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
measure_id = pheno_tool._get_normalize_measure_id({
"measure_name":
measure_name,
"instrument_name":
instrument_name
})
assert measure_id is None
def test_init_empty_person_ids_normalize(fake_phenotype_data):
pheno_tool = PhenoTool(
fake_phenotype_data,
"i1.m1",
person_ids=[],
normalize_by=[{
"instrument_name": "i1",
"measure_name": "m2"
}],
)
assert pheno_tool
def test_calc_empty_pheno_df(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1", person_ids=[])
variants = Counter({
"f4.p1": 1,
"f5.p1": 1,
"f7.p1": 1,
"f10.p1": 1,
"f12.p1": 1,
"f16.p1": 1,
"f24.p1": 1,
"f25.p1": 1,
"f30.p1": 1,
"f32.p1": 1,
})
res = pheno_tool.calc(variants)
assert isinstance(res, PhenoResult)
assert res.positive_count == 0
assert res.positive_mean == 0
assert res.positive_deviation == 0
assert res.negative_count == 0
assert res.negative_mean == 0
assert res.negative_deviation == 0
assert res.pvalue == "NA"
resM, resF = pheno_tool.calc(variants, sex_split=True).values()
assert isinstance(resM, PhenoResult)
assert isinstance(resF, PhenoResult)
assert resM.positive_count == 0
assert resM.positive_mean == 0
assert resM.positive_deviation == 0
assert resM.negative_count == 0
assert resM.negative_mean == 0
assert resM.negative_deviation == 0
assert resM.pvalue == "NA"
assert resF.positive_count == 0
assert resF.positive_mean == 0
assert resF.positive_deviation == 0
assert resF.negative_count == 0
assert resF.negative_mean == 0
assert resF.negative_deviation == 0
assert resF.pvalue == "NA"
def test_init_normalize_measures_non_continuous(fake_phenotype_data,
measure_name):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
norm_measures = [
{
"measure_name": "??",
"instrument_name": ""
},
{
"measure_name": "m5",
"instrument_name": ""
},
{
"measure_name": "m7",
"instrument_name": "i1"
},
]
with pytest.raises(AssertionError):
pheno_tool._init_normalize_measures(norm_measures)
def test_init_normalize_measures(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
norm_measures = [
{
"measure_name": "??",
"instrument_name": ""
},
{
"measure_name": "m3",
"instrument_name": ""
},
{
"measure_name": "m7",
"instrument_name": "i1"
},
]
norm_measures = pheno_tool._init_normalize_measures(norm_measures)
assert len(norm_measures) == 2
assert set(norm_measures) == {"i1.m3", "i1.m7"}
for measure_id in norm_measures:
assert fake_phenotype_data.has_measure(measure_id)
def test_init_pheno_df(fake_phenotype_data):
pheno_tool = PhenoTool(
fake_phenotype_data,
"i1.m1",
normalize_by=[{
"instrument_name": "i1",
"measure_name": "m2"
}],
)
assert pheno_tool.pheno_df is not None
assert not pheno_tool.pheno_df.empty
assert set(pheno_tool.pheno_df) == {
"person_id",
"family_id",
"role",
"status",
"sex",
"i1.m1",
"i1.m2",
"normalized",
}
def test_init_non_continuous_or_ordinal_measure(fake_phenotype_data):
with pytest.raises(AssertionError):
PhenoTool(fake_phenotype_data, "i1.m5") # categorical
with pytest.raises(AssertionError):
PhenoTool(fake_phenotype_data, "i1.m9") # raw
def test_init_nonexistent_measure(fake_phenotype_data):
with pytest.raises(AssertionError):
PhenoTool(fake_phenotype_data, "i1.??")
def test_init_empty_person_ids(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1", person_ids=[])
assert pheno_tool
def test_get_normalize_measure_id_measure_dict_no_keys(fake_phenotype_data):
pheno_tool = PhenoTool(fake_phenotype_data, "i1.m1")
with pytest.raises(AssertionError):
pheno_tool._get_normalize_measure_id({"measure_name": "something"})
with pytest.raises(AssertionError):
pheno_tool._get_normalize_measure_id({"instrument_name": "something"})