def test_mi_local_values():
"""Test local MI estimation."""
n = 1000
cov = 0.4
source = [rn.normalvariate(0, 1) for r in range(n)] # correlated src
target = [0] + [sum(pair) for pair in zip(
[cov * y for y in source[0:n - 1]],
[(1 - cov) * y for y in
[rn.normalvariate(0, 1) for r in range(n - 1)]])]
source = np.expand_dims(np.array(source), axis=1)
target = np.expand_dims(np.array(target), axis=1)
analysis_opts = {
'kraskov_k': 4,
'normalise': False,
'theiler_t': 0,
'noise_level': 1e-8,
'local_values': True
}
mi_est = Estimator_mi('jidt_kraskov')
mi_res = mi_est.estimate(source, target, analysis_opts)
assert mi_res.shape[0] == n, 'Local MI estimator did not return an array.'
analysis_opts['local_values'] = False
mi_avg = mi_est.estimate(source, target, analysis_opts)
assert np.isclose(np.mean(mi_res), mi_avg, rtol=0.01), (
'Average local MI is not equal to estimated average MI.')
def test_lagged_mi_corr_gaussian():
"""Test MI estimation on correlated Gaussians.
Run MI estimation on two sets of random Gaussian data with a given
covariance. The second data set is shifted by one sample creating a source-
target delay of one sample. This example is modeled after the JIDT demo 4
for transfer entropy. The resulting MI can be compared to the analytical
result (but expect some error in the estimate).
Note:
This test runs considerably faster than other system tests.
"""
n = 1000
cov = 0.4
source = [rn.normalvariate(0, 1) for r in range(n)] # correlated src
target = [0] + [sum(pair) for pair in zip(
[cov * y for y in source[0:n - 1]],
[(1 - cov) * y for y in
[rn.normalvariate(0, 1) for r in range(n - 1)]])]
analysis_opts = {
'kraskov_k': 4,
'normalise': False,
'theiler_t': 0,
'noise_level': 1e-8,
'local_values': False,
'tau_target': 1,
'tau_source': 1,
'source_target_delay': 1,
'history_target': 1,
'history_source': 1,
'lag': 1
}
mi_est = Estimator_mi('jidt_kraskov')
mi_res = mi_est.estimate(np.expand_dims(np.array(source), axis=1),
np.expand_dims(np.array(target), axis=1),
analysis_opts)
mi_exp = np.log(1 / (1 - cov ** 2))
print('Correlated Gaussians: MI result {0:.4f} bits; expected to be '
'{1:0.4f} bit for the copy'
.format(mi_res, mi_exp))
np.testing.assert_approx_equal(
mi_res, mi_exp, significant=1,
err_msg='MI is not close to expected result.')
def test_jidt_kraskov_input():
"""Test handling of wrong inputs to the JIDT Kraskov MI-estimator."""
source_1 = np.empty((100, 1))
source_2 = np.empty((100, 1))
mi_est = Estimator_mi('jidt_kraskov')
# Wrong type for options dictinoary
with pytest.raises(TypeError):
mi_est.estimate(var1=source_1, var2=source_2, opts=1)
# Run without a options dict altogether
mi_est.estimate(var1=source_1, var2=source_2)
# Run analysis with all default vales
mi_est.estimate(var1=source_1, var2=source_2, opts={})
def test_mi_estimator_jidt_gaussian():
"""Test theoretical Gaussian estimation on random Gaussian time series.
source1 is a normally-distributed time series.
target is independent from source1.
The following MI value is estimated:
1) I(source1, target)
The expected MI value is the theoretical value:
I(source1, target) = - 1/2 * log(1 - corr ^ 2)
"""
# Set length of time series and correlation coefficient
n = 1000
cov = 0.9
# Set tolerance for assert
assert_tolerance = 0.001
# Generate random normally-distributed source time series
source = [rn.normalvariate(0, 1) for r in range(n)]
# Generate correlated target time series
target = [sum(pair) for pair in zip(
[cov * y for y in source[0:n]],
[(1 - cov) * y for y in
[rn.normalvariate(0, 1) for r in range(n)]])]
# Compute effective correlation from finite time series
corr_effective = np.corrcoef(source, target)[1, 0]
# Compute theoretical value for MI using the effective correlation
theoretical_res = - 0.5 * np.log(1 - corr_effective ** 2)
# Cast everything to numpy so the idtxl estimator understands it.
source = np.expand_dims(np.array(source), axis=1)
target = np.expand_dims(np.array(target), axis=1)
# Call JIDT to perform estimation
opts = {}
estimator_name = 'jidt_gaussian'
est = Estimator_mi(estimator_name)
res = est.estimate(var1=source, var2=target, opts=opts)
print('MI result: {0:.4f} bits'.format(res))
print('Expected theoretical result:'
'{0:.4f} bits'.format(theoretical_res))
assert np.abs(res - theoretical_res) < assert_tolerance, (
'MI calculation for Gaussian failed.')
def test_mi_local_values():
"""Test local MI estimation."""
n = 1000
cov = 0.4
source = [rn.normalvariate(0, 1) for r in range(n)] # correlated src
target = [0] + [sum(pair) for pair in zip(
[cov * y for y in source[0:n - 1]],
[(1 - cov) * y for y in
[rn.normalvariate(0, 1) for r in range(n - 1)]])]
source = np.expand_dims(np.array(source), axis=1)
target = np.expand_dims(np.array(target), axis=1)
analysis_opts = {
'kraskov_k': 4,
'normalise': 'false',
'theiler_t': 0,
'noise_level': 1e-8,
'local_values': True
}
mi_est = Estimator_mi('jidt_kraskov')
mi_res = mi_est.estimate(source, target, analysis_opts)
assert mi_res.shape[0] == n, 'Local MI estimator did not return an array.'
def test_mi_estimator_jidt_discrete():
"""Test MI estimation on two sets of discrete data.
"""
n = 1000
source_1 = np.zeros(n, np.int_)
source_2 = np.zeros(n, np.int_)
target = np.zeros(n, np.int_)
for t in range(n):
source_1[t] = (t >= n/2) * 1
source_2[t] = t % 2
target[t] = source_1[t]
opts = {'num_discrete_bins': 2, 'time_diff': 0, 'discretise_method' : 'none'}
calculator_name = 'jidt_discrete'
est = Estimator_mi(calculator_name)
res_1 = est.estimate(var1=source_1, var2=target, opts=opts)
res_2 = est.estimate(var1=source_2, var2=target, opts=opts)
print('Example 1: MI result {0:.4f} bits; expected to be 1 bit for the copy'
.format(res_1))
print('Example 2: MI result {0:.4f} bits; expected to be 0 bits for the '
'uncorrelated variable.'.format(res_2))
assert (res_1 == 1), ('MI calculation for copy failed.')
assert (res_2 == 0), ('MI calculation for no correlation failed.')
def test_estimator_change():
"""Test dynamic changing of estimators at runtime."""
estimator_name_1 = 'jidt_kraskov'
estimator_name_2 = 'opencl_kraskov'
cmi_estimator_1 = Estimator_cmi(estimator_name_1)
cmi_estimator_2 = Estimator_cmi(estimator_name_2)
mi_estimator_1 = Estimator_mi(estimator_name_1)
assert cmi_estimator_1.estimator_name == estimator_name_1, (
'The estimator was not set correctly')
assert cmi_estimator_2.estimator_name == estimator_name_2, (
'The estimator was not set correctly')
assert mi_estimator_1.estimator_name == estimator_name_1, (
'The estimator was not set correctly')
def test_mi_estimator_jidt_discrete():
"""Test MI estimation on two sets of discrete data."""
n = 1000
source_1 = np.zeros(n, np.int_)
source_2 = np.zeros(n, np.int_)
target = np.zeros(n, np.int_)
for t in range(n):
source_1[t] = (t >= n/2) * 1
source_2[t] = t % 2
target[t] = source_1[t]
opts = {'num_discrete_bins': 2,
'lag': 0,
'discretise_method': 'none'}
calculator_name = 'jidt_discrete'
est = Estimator_mi(calculator_name)
res_1 = est.estimate(var1=source_1, var2=target, opts=opts)
res_2 = est.estimate(var1=source_2, var2=target, opts=opts)
print('Example 1: MI result {0:.4f} bits; expected to be 1 bit for the '
'copy'.format(res_1))
print('Example 2: MI result {0:.4f} bits; expected to be 0 bits for the '
'uncorrelated variable.'.format(res_2))
assert (res_1 == 1), ('MI calculation for copy failed.')
assert (res_2 == 0), ('MI calculation for no correlation failed.')
def test_mi_estimator_jidt_discrete_discretisation():
"""Test MI estimation discretisation methods and error handling."""
opts = {'num_discrete_bins': 2,
'lag': 0,
'discretise_method': 'none'}
calculator_name = 'jidt_discrete'
est = Estimator_mi(calculator_name)
n = 1000 # Need this to be an even number for the test to work
cov = 0.4
# Test 1: input is not an integer array
source_1 = [rn.normalvariate(0, 1) for r in range(n)] # correlated src
source_2 = [rn.normalvariate(0, 1) for r in range(n)] # uncorrelated src
target = [sum(pair) for pair in zip(
[cov * y for y in source_1],
[(1 - cov) * y for y in [rn.normalvariate(0, 1) for r in range(n)]])]
# Cast everything to numpy so the idtxl estimator understands it.
source_1 = np.expand_dims(np.array(source_1), axis=1)
source_2 = np.expand_dims(np.array(source_2), axis=1)
target = np.expand_dims(np.array(target), axis=1)
with pytest.raises(AssertionError):
est.estimate(var1=source_1, var2=target, opts=opts)
# Test 2: alphabet sizes not specified correctly
source_1 = np.zeros(n, np.int_)
source_2 = np.zeros(n, np.int_)
source_3 = np.zeros(n, np.int_)
target = np.zeros(n, np.int_)
for t in range(n):
source_1[t] = (t >= n/2) * 1
source_2[t] = t % 2
source_3[t] = (t < n/2) * source_2[t] + (t >= n/2) * (1 - source_2[t])
target[t] = source_1[t]
opts = {'num_discrete_bins': 0,
'lag': 0,
'discretise_method': 'none'}
with pytest.raises(AssertionError):
est.estimate(var1=source_2, var2=target, opts=opts)
