if link_ids is None:
return None
return {par : [true_parents[par][link]] for par in true_parents \
for link in link_ids}
# TEST LINK GENERATION #########################################################
N_NODES = 4
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.5, 0.6, length=N_NODES), 1000, 42)
])
def a_sample(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Set the random seed
np.random.seed(seed_val)
# Generate the data
data, _ = pp.var_process(links_coeffs, T=time)
# Get the true parents
true_parents = _get_parent_graph(links_coeffs)
return pp.DataFrame(data), true_parents
# PCMCI CONSTRUCTION ###########################################################
@pytest.fixture()
"""
Select links given by from the true parents dictionary
"""
if link_ids is None:
return None
return {par : [true_parents[par][link]] for par in true_parents \
for link in link_ids}
# TEST LINK GENERATION #########################################################
N_NODES = 4
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.5, 0.6, length=N_NODES), 1000, 42)])
def a_sample(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Set the random seed
np.random.seed(seed_val)
# Generate the data
data, _ = pp.var_process(links_coeffs, T=time)
# Get the true parents
true_parents = _get_parent_graph(links_coeffs)
return pp.DataFrame(data), true_parents
# PCMCI CONSTRUCTION ###########################################################
@pytest.fixture()
# Fixture to build and return a parameterized PCMCI. Different selected
# variables can be defined here.
#seed, corr_val, T
(5, 0.3, 1000), # Default
(6, 0.3, 1000), # New Seed
(1, 0.9, 1000)]) # Strong Correlation
def data_sample_a(request):
# Unpack the parameters
seed, corr_val, T = request.param
# Return the data sample
return gen_data_sample(seed, corr_val, T)
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.1, 0.9), 1000, 2),
(a_chain(0.5, 0.6), 1000, 11),
(a_chain(0.5, 0.6, length=5), 1000, 42)])
def data_frame_a(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Generate the dataframe
return gen_data_frame(links_coeffs, time, seed_val)
def test_get_array_parcorr(par_corr, data_frame_a):
# Check the get_array function
check_get_array(par_corr, data_frame_a)
def test_run_test_parcorr(par_corr, data_frame_a):
# Check the run_test function
check_run_test(par_corr, data_frame_a)
from tigramite.models import Prediction
from tigramite.independence_tests import ParCorr
from test_pcmci_calculations import a_chain, gen_data_frame
# Pylint settings
# pylint: disable=redefined-outer-name
# TEST LINK GENERATION #########################################################
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.9, 0.7), 10000, 21),
(a_chain(0.9, 0.7), 10000, 42)
])
def data_frame_a(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Generate the dataframe
return gen_data_frame(links_coeffs, time, seed_val), links_coeffs
# TEST MODELS ##################################################################
def test_linear_mediation_coeffs(data_frame_a):
# Build the dataframe and the model
(dataframe, true_parents), links_coeffs = data_frame_a
med = LinearMediation(dataframe=dataframe, data_transform=None)
# Fit the model
import tigramite.data_processing as pp
from tigramite.models import Prediction
from tigramite.independence_tests import ParCorr
from test_pcmci_calculations import a_chain, gen_data_frame
# Pylint settings
# pylint: disable=redefined-outer-name
# TEST LINK GENERATION #########################################################
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.5, 0.2), 10000, 21),
(a_chain(0.5, 0.2), 10000, 42)])
def data_frame_a(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Generate the dataframe
return gen_data_frame(links_coeffs, time, seed_val), links_coeffs
# TEST MODELS ##################################################################
def test_linear_mediation_coeffs(data_frame_a):
# Build the dataframe and the model
(dataframe, true_parents), links_coeffs = data_frame_a
med = LinearMediation(dataframe=dataframe)
# Fit the model
med.fit_model(all_parents=true_parents, tau_max=3)
# Ensure the results make sense
#seed, corr_val, T
(5, 0.3, 1000), # Default
(6, 0.3, 1000), # New Seed
(1, 0.9, 1000)]) # Strong Correlation
def data_sample_a(request):
# Unpack the parameters
seed, corr_val, T = request.param
# Return the data sample
return gen_data_sample(seed, corr_val, T)
@pytest.fixture(params=[
# Generate a test data sample
# Parameterize the sample by setting the autocorrelation value, coefficient
# value, total time length, and random seed to different numbers
# links_coeffs, time, seed_val
(a_chain(0.1, 0.9), 1000, 2),
(a_chain(0.5, 0.6), 1000, 11),
(a_chain(0.5, 0.6, length=5), 1000, 42)])
def data_frame_a(request):
# Set the parameters
links_coeffs, time, seed_val = request.param
# Generate the dataframe
return gen_data_frame(links_coeffs, time, seed_val)
def test_get_array_parcorr(par_corr, data_frame_a):
# Check the get_array function
check_get_array(par_corr, data_frame_a)
def test_run_test_parcorr(par_corr, data_frame_a):
# Check the run_test function
check_run_test(par_corr, data_frame_a)