def test_gini_values(income_arrays, inequality_results):
ad1 = ApodeData(income_arrays, income_column="y1")
ad2 = ApodeData(income_arrays, income_column="y2")
ad3 = ApodeData(income_arrays, income_column="y3")
ad4 = ApodeData(income_arrays, income_column="y4")
ad5 = ApodeData(income_arrays, income_column="y5")
ad6 = ApodeData(income_arrays, income_column="y6")
# Relative tets modified (avoid division by zero):
# test = ((val1+1)-(val2+1))/(val1+1)
np.testing.assert_allclose(ad1.inequality.gini() + 1,
inequality_results.gini[0] + 1)
np.testing.assert_allclose(ad2.inequality.gini(),
inequality_results.gini[1])
np.testing.assert_allclose(ad3.inequality.gini(),
inequality_results.gini[2])
np.testing.assert_allclose(ad4.inequality.gini(),
inequality_results.gini[3])
np.testing.assert_allclose(ad5.inequality.gini(),
inequality_results.gini[4])
np.testing.assert_allclose(ad6.inequality.gini(),
inequality_results.gini[5])
def make(*, seed, **kwargs):
random = np.random.RandomState(seed=seed)
x = random.uniform(**kwargs)
df1 = pd.DataFrame({"x": x})
return ApodeData(df1, varx="x")
def test_herfindahl_empty_array():
y = []
df = pd.DataFrame({"x": y})
data = ApodeData(df, income_column="x")
assert data.concentration.herfindahl() == 0
def test_gini_empty_array():
y = []
df = pd.DataFrame({"x": y})
data = ApodeData(df, income_column="x")
assert data.inequality.gini() == 0
def test_ratio_empty_array():
y = []
df = pd.DataFrame({"x": y})
data = ApodeData(df, income_column="x")
assert data.inequality.ratio(alpha=0.5) == 0
def test_kolm_empty_array():
y = []
df = pd.DataFrame({"x": y})
data = ApodeData(df, income_column="x")
assert data.inequality.kolm(alpha=1) == 0
def test_gini_size_one_array():
y = [2]
df = pd.DataFrame({"x": y})
ad = ApodeData(df, income_column="x")
assert ad.inequality.gini() == 0
def test_atkinson_size_one_array():
y = [2]
df = pd.DataFrame({"x": y})
ad = ApodeData(df, income_column="x")
assert ad.inequality.atkinson(alpha=1) == 0
def test_takayama_avoid_zero_div_error():
# u = 0
df = pd.DataFrame({"x": np.zeros(10)})
ad = ApodeData(df, income_column="x")
pline = 0.2
assert ad.poverty.takayama(pline=pline) == 0
def test_hagenaars_q_zero():
df = pd.DataFrame({"x": np.arange(1, 12)})
ad = ApodeData(df, income_column="x")
pline = np.min(ad.data.values) - 0.1
assert ad.poverty.hagenaars(pline=pline) == 0