def test_get_cond_prob_triplet():
"""
Tests getting conditional probability between two variables and a third set of variables.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(
0.6666666666666667,
data.__get_cond_prob_triplet__('x1', 't', 'x2', 't', ['x3'], ['t']))
assert_almost_equal(
0.0, data.__get_cond_prob_triplet__('x1', 't', 'x2', 't', ['x3'],
['f']))
assert_almost_equal(
0.16666666666666669,
data.__get_cond_prob_triplet__('x1', 't', 'x2', 'f', ['x3'], ['t']))
assert_almost_equal(
0.0, data.__get_cond_prob_triplet__('x1', 't', 'x2', 'f', ['x3'],
['f']))
assert_almost_equal(
0.16666666666666669,
data.__get_cond_prob_triplet__('x1', 'f', 'x2', 't', ['x3'], ['t']))
assert_almost_equal(
0.0, data.__get_cond_prob_triplet__('x1', 'f', 'x2', 't', ['x3'],
['f']))
assert_almost_equal(
0.0, data.__get_cond_prob_triplet__('x1', 'f', 'x2', 'f', ['x3'],
['t']))
assert_almost_equal(
1.0, data.__get_cond_prob_triplet__('x1', 'f', 'x2', 'f', ['x3'],
['f']))
def test_get_cond_prob_set():
"""
Tests getting conditional probability between a variable and a set of variables.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(
0.8, data.__get_cond_prob_set__('x1', 't', ['x2', 'x3'], ['t', 't']))
assert_almost_equal(
0.2, data.__get_cond_prob_set__('x1', 'f', ['x2', 'x3'], ['t', 't']))
assert_almost_equal(
0.0, data.__get_cond_prob_set__('x1', 't', ['x2', 'x3'], ['f', 'f']))
assert_almost_equal(
1.0, data.__get_cond_prob_set__('x1', 'f', ['x2', 'x3'], ['f', 'f']))
assert_almost_equal(
0.0, data.__get_cond_prob_set__('x1', 't', ['x2', 'x3'], ['t', 'f']))
assert_almost_equal(
0.0, data.__get_cond_prob_set__('x1', 'f', ['x2', 'x3'], ['t', 'f']))
assert_almost_equal(
1.0, data.__get_cond_prob_set__('x1', 't', ['x2', 'x3'], ['f', 't']))
assert_almost_equal(
0.0, data.__get_cond_prob_set__('x1', 'f', ['x2', 'x3'], ['f', 't']))
def test_get_joint_prob_triplet():
"""
Tests getting joint probability between two variables and a set of variables.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(
0.4,
data.__get_joint_prob_triplet__('x1', 't', 'x2', 't', ['x3'], ['t']))
assert_almost_equal(
0.0,
data.__get_joint_prob_triplet__('x1', 't', 'x2', 't', ['x3'], ['f']))
assert_almost_equal(
0.1,
data.__get_joint_prob_triplet__('x1', 't', 'x2', 'f', ['x3'], ['t']))
assert_almost_equal(
0.0,
data.__get_joint_prob_triplet__('x1', 't', 'x2', 'f', ['x3'], ['f']))
assert_almost_equal(
0.1,
data.__get_joint_prob_triplet__('x1', 'f', 'x2', 't', ['x3'], ['t']))
assert_almost_equal(
0.0,
data.__get_joint_prob_triplet__('x1', 'f', 'x2', 't', ['x3'], ['f']))
assert_almost_equal(
0.0,
data.__get_joint_prob_triplet__('x1', 'f', 'x2', 'f', ['x3'], ['t']))
assert_almost_equal(
0.4,
data.__get_joint_prob_triplet__('x1', 'f', 'x2', 'f', ['x3'], ['f']))
def test_get_joint_prob():
"""
Tests getting joint probability between two variables.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(0.4, data.__get_joint_prob__('x1', 't', 'x2', 't'))
assert_almost_equal(0.1, data.__get_joint_prob__('x1', 't', 'x2', 'f'))
assert_almost_equal(0.1, data.__get_joint_prob__('x1', 'f', 'x2', 't'))
assert_almost_equal(0.4, data.__get_joint_prob__('x1', 'f', 'x2', 'f'))
def test_to_query():
"""
Tests generating query.
:return: None.
"""
data = DiscreteData(get_good_df())
names = ['a', 'c', 'b', 'd']
vals = ['t', 'f', 't', 'f']
observed = data.__to_query__(names, vals)
expected = 'a == "t" and b == "t" and c == "f" and d == "f"'
assert observed == expected
def test_count_parents_child():
"""
Tests counting parent-child values.
:return: None.
"""
data = DiscreteData(get_good_df())
assert 4 == data.__count_parents_child__('x1', 't', ['x2'], ['t'])
assert 1 == data.__count_parents_child__('x1', 't', ['x2'], ['f'])
assert 1 == data.__count_parents_child__('x1', 'f', ['x2'], ['t'])
assert 4 == data.__count_parents_child__('x1', 'f', ['x2'], ['f'])
def test_get_cond_prob():
"""
Tests getting conditional probability between two variables.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(0.8, data.__get_cond_prob__('x1', 't', 'x2', 't'))
assert_almost_equal(0.2, data.__get_cond_prob__('x1', 't', 'x2', 'f'))
assert_almost_equal(0.2, data.__get_cond_prob__('x1', 'f', 'x2', 't'))
assert_almost_equal(0.8, data.__get_cond_prob__('x1', 'f', 'x2', 'f'))
def test_sort():
"""
Tests sorting names and values by names ascending.
:return: None.
"""
data = DiscreteData(get_good_df())
observed = data.__sort__(['a', 'c', 'b', 'd'], ['t', 'f', 't', 'f'])
expected = [('a', 't'), ('b', 't'), ('c', 'f'), ('d', 'f')]
assert len(observed) == len(expected)
for o, e in zip(observed, expected):
assert o[0] == e[0]
assert o[1] == e[1]
def test_get_local_kutato():
"""
Tests computing local Kutato score.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(-1.540639,
data.get_local_kutato('x1', ['x2']),
places=5)
assert_almost_equal(-1.540639,
data.get_local_kutato('x2', ['x1']),
places=5)
assert_almost_equal(-1.082607,
data.get_local_kutato('x1', ['x3']),
places=5)
assert_almost_equal(-1.124863,
data.get_local_kutato('x3', ['x1']),
places=5)
assert_almost_equal(-0.709086,
data.get_local_kutato('x1', ['x4']),
places=5)
assert_almost_equal(-0.709086,
data.get_local_kutato('x4', ['x1']),
places=5)
def test_get_mwst_skeleton():
"""
Tests getting the MWST skeleton structure.
:return: None.
"""
data = DiscreteData(get_good_df())
g = get_mwst_skeleton(data)
assert len(g.nodes) == 4
def test_get_variable_profiles():
"""
Tests getting variable profiles.
:return: None.
"""
data = DiscreteData(get_good_df())
profiles = data.get_variable_profiles()
print(data)
assert 4 == len(profiles)
for i in range(4):
name = 'x{}'.format(i + 1)
assert name in profiles
profile = profiles[name]
assert 't' in profile
assert 'f' in profile
def test_get_pairwise_mutual_information():
"""
Tests get pairwise mutual information.
:return: List of pairwise mutual information.
"""
data = DiscreteData(get_good_df())
observed = sorted(data.get_pairwise_mutual_information(),
key=lambda tup: (-tup[2], tup[0], tup[1]))
expected = [('x1', 'x4'), ('x1', 'x3'), ('x2', 'x3'), ('x3', 'x4'),
('x1', 'x2'), ('x2', 'x4')]
print(observed)
print('--')
print(expected)
assert len(observed) == len(expected)
for o, e in zip(observed, expected):
assert o[0] == e[0]
assert o[1] == e[1]
def test_is_valid_bad():
"""
Tests if a dataset is valid with a bad dataframe.
:return: None.
"""
try:
df = get_bad_df()
data = DiscreteData(df)
assert 1 == 2
except Exception:
assert 1 == 1
def test_fit():
"""
Tests fitting data.
:return: None.
"""
data = DiscreteData(get_good_df())
mwst = MwstAlgo()
mwst.fit(data)
bbn = mwst.bbn
assert 4 == len(bbn.nodes)
assert 3 == len(bbn.edges)
def test_get_mi():
"""
Tests getting mutual information.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(0.19274475702175753, data.__get_mi__('x1', 'x2'))
assert_almost_equal(0.4228104552401625, data.__get_mi__('x1', 'x3'))
assert_almost_equal(0.6931471805599453, data.__get_mi__('x1', 'x4'))
assert_almost_equal(0.4228104552401625, data.__get_mi__('x2', 'x3'))
assert_almost_equal(0.19274475702175753, data.__get_mi__('x2', 'x4'))
assert_almost_equal(0.4228104552401625, data.__get_mi__('x3', 'x4'))
def test_get_prob():
"""
Tests getting simple probability.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(0.5, data.__get_prob__('x1', 't'))
assert_almost_equal(0.5, data.__get_prob__('x1', 'f'))
assert_almost_equal(0.5, data.__get_prob__('x2', 't'))
assert_almost_equal(0.5, data.__get_prob__('x2', 'f'))
assert_almost_equal(0.6, data.__get_prob__('x3', 't'))
assert_almost_equal(0.4, data.__get_prob__('x3', 'f'))
assert_almost_equal(0.5, data.__get_prob__('x4', 't'))
assert_almost_equal(0.5, data.__get_prob__('x4', 'f'))
def test_count():
"""
Tests simple counting.
:return: None.
"""
data = DiscreteData(get_good_df())
assert 5 == data.__count__(['x1'], ['t'])
assert 5 == data.__count__(['x1'], ['f'])
assert 5 == data.__count__(['x2'], ['t'])
assert 5 == data.__count__(['x2'], ['f'])
assert 6 == data.__count__(['x3'], ['t'])
assert 4 == data.__count__(['x3'], ['f'])
assert 5 == data.__count__(['x4'], ['t'])
assert 5 == data.__count__(['x4'], ['f'])
def test_get_cond_mi():
"""
Tests getting conditional mutual information.
:return: None.
"""
data = DiscreteData(get_good_df())
assert_almost_equal(0.0, data.__get_cond_mi__('x1', 'x2', ['x3', 'x4']))
assert_almost_equal(0.0, data.__get_cond_mi__('x1', 'x3', ['x2', 'x4']))
assert_almost_equal(0.25020121176909393,
data.__get_cond_mi__('x1', 'x4', ['x2', 'x3']))
assert_almost_equal(0.0, data.__get_cond_mi__('x2', 'x1', ['x3', 'x4']))
assert_almost_equal(0.25020121176909393,
data.__get_cond_mi__('x2', 'x3', ['x1', 'x4']))
assert_almost_equal(0.0, data.__get_cond_mi__('x2', 'x4', ['x1', 'x3']))
assert_almost_equal(0.0, data.__get_cond_mi__('x3', 'x1', ['x2', 'x4']))
assert_almost_equal(0.25020121176909393,
data.__get_cond_mi__('x3', 'x2', ['x1', 'x4']))
assert_almost_equal(0.0, data.__get_cond_mi__('x3', 'x4', ['x1', 'x2']))
assert_almost_equal(0.25020121176909393,
data.__get_cond_mi__('x4', 'x1', ['x2', 'x3']))
assert_almost_equal(0.0, data.__get_cond_mi__('x4', 'x2', ['x1', 'x3']))
assert_almost_equal(0.0, data.__get_cond_mi__('x4', 'x3', ['x1', 'x2']))
