def test_dist_cmr_record(data, patterns, param_def_dist, param_dist):
model = cmr.CMR()
states = model.record(data, param_dist, None, param_def_dist, patterns=patterns)
# test the series of context states
np.testing.assert_allclose(
states[0].c, np.array([0.5, 0.0, 0.0, 0.0, 0.0, 0.0, 0.8660254])
)
np.testing.assert_allclose(
states[1].c, np.array([0.4330127, 0.5, 0.0, 0.0, 0.0, 0.0, 0.75])
)
np.testing.assert_allclose(
states[2].c, np.array([0.375, 0.4330127, 0.5, 0.0, 0.0, 0.0, 0.64951905])
)
np.testing.assert_allclose(
states[3].c, np.array([0.375, 0.4330127, 0.5, 0.0, 0.0, 0.0, 0.64951905])
)
np.testing.assert_allclose(
states[4].c,
np.array([0.29319805, 0.80043616, 0.12426407, 0.0, 0.0, 0.0, 0.50783392]),
)
# test various network components
np.testing.assert_allclose(
states[5].c, np.array([0.0, 0.0, 0.0, 0.5, 0.0, 0.0, 0.8660254])
)
np.testing.assert_allclose(
states[5].c_in, np.array([0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0])
)
np.testing.assert_allclose(states[5].f, np.array([1.0, 0.0, 0.0, 0.0]))
np.testing.assert_allclose(states[5].f_in, np.array([0.0, 0.0, 0.0, 0.0]))
def test_dist_cmr_record_include(data, patterns, param_def_dist, param_dist):
model = cmr.CMR()
states = model.record(
data, param_dist, None, param_def_dist, patterns=patterns, include=['c', 'c_in']
)
assert hasattr(states[0], 'c') and hasattr(states[0], 'c_in')
assert not hasattr(states[0], 'f')
def test_dist_cmr(data, patterns, param_def_dist, param_dist):
"""Test localist CMR using the CMR-D implementation."""
model = cmr.CMR()
logl, n = model.likelihood(
data, param_dist, None, param_def_dist, patterns=patterns
)
np.testing.assert_allclose(logl, -5.936799964636842)
def test_sublayer_cmr(data, patterns, param_def_sublayer, param_dist):
"""Test evaluation of CMR-D with multiple context sublayers."""
param = param_dist.copy()
param['B_enc_loc'] = 0.5
param['B_enc_cat'] = 0.8
model = cmr.CMR()
logl, n = model.likelihood(data, param, None, param_def_sublayer, patterns=patterns)
np.testing.assert_allclose(logl, -5.8694368046085215)
def test_match_generate(data, patterns, param_def_dist, param_dist):
"""Test generating recalls using context match screening."""
param = param_dist.copy()
param['A1'] = -3
param['A2'] = 1
model = cmr.CMR()
param_def_dist.set_options(scope='pool', filter_recalls=True)
sim = model.generate(data, param, None, param_def_dist, patterns=patterns)
assert isinstance(sim, pd.DataFrame)
def test_dist_cmr_fit(data, patterns, param_def_dist):
"""Test fitted parameter values for CMR-D."""
param_def = param_def_dist.copy()
model = cmr.CMR()
param_def.set_fixed(w_loc=1)
param_def.set_free(B_enc=(0, 1))
results = model.fit_indiv(data, param_def, patterns=patterns, n_jobs=2)
np.testing.assert_allclose(
results['B_enc'].to_numpy(), np.array([0.72728744, 0.99883425]), atol=0.02
)
def test_dynamic_cmr_recall(data, patterns, param_def_dist, param_dist):
"""Test evaluation of a dynamic recall parameter."""
param = param_dist.copy()
param_def = param_def_dist.copy()
param['B_op'] = 0.2
param_def.set_dynamic('recall', B_rec='op * B_op')
model = cmr.CMR()
logl, n = model.likelihood(
data, param, None, param_def, patterns=patterns, recall_keys=['op']
)
np.testing.assert_allclose(logl, -5.919470385031945)
def test_dynamic_cmr(data, patterns, param_def_dist, param_dist):
"""Test evaluation of a dynamic study parameter."""
param = param_dist.copy()
param_def = param_def_dist.copy()
param['B_distract'] = 0.2
param_def.set_dynamic('study', B_enc='distract * B_distract')
model = cmr.CMR()
logl, n = model.likelihood(
data, param, None, param_def, patterns=patterns, study_keys=['distract']
)
np.testing.assert_allclose(logl, -5.9899248839454415)
def test_cmr_fit(data, param):
"""Test fit of CMR parameters to sample data."""
model = cmr.CMR()
n_item = data['item_index'].nunique()
param_def, patterns = cmr.config_loc_cmr(n_item)
param_def.set_fixed(param)
param_def.set_free(B_enc=(0, 1))
results = model.fit_indiv(data, param_def, patterns=patterns, n_jobs=2)
np.testing.assert_allclose(
results['B_enc'].to_numpy(), np.array([0.72728744, 0.99883425]), atol=0.02
)
np.testing.assert_array_equal(results['n'].to_numpy(), [3, 3])
def test_sublayer_generate(data, patterns, param_def_sublayer, param_dist):
"""Test using CMR-D with sublayers to generate data."""
param = param_dist.copy()
param['B_enc_loc'] = 0.5
param['B_enc_cat'] = 0.8
model = cmr.CMR()
# recall from list items only
param_def_sublayer.set_options(scope='list')
sim = model.generate(data, param, None, param_def_sublayer, patterns=patterns)
assert isinstance(sim, pd.DataFrame)
# recall from the pool
param_def_sublayer.set_options(scope='pool')
sim = model.generate(data, param, None, param_def_sublayer, patterns=patterns)
assert isinstance(sim, pd.DataFrame)
def setup(self):
self.data = fit.sample_data('Morton2013_mixed')
n_items = 768
study = self.data.query('trial_type == "study"')
items = study.groupby('item_index')['item'].first().to_numpy()
self.patterns = {'items': items, 'vector': {'loc': np.eye(n_items)}}
param_def = parameters.Parameters()
param_def.set_dependent(Dfc='1 - Lfc', Dcf='1 - Lcf')
param_def.set_sublayers(f=['task'], c=['task'])
param_def.set_weights(
'fc',
{
(('task', 'item'), ('task', 'item')): 'Dfc * loc',
},
)
param_def.set_weights(
'cf',
{
(('task', 'item'), ('task', 'item')): 'Dcf * loc',
},
)
self.param_def = param_def
param = {
'B_enc': 0.5,
'B_start': 0,
'B_rec': 0.8,
'Lfc': 0.8,
'Lcf': 0.8,
'P1': 1,
'P2': 1,
'T': 10,
'X1': 0.05,
'X2': 0.2,
}
self.param = param_def.eval_dependent(param)
self.model = cmr.CMR()
def test_cmr(data):
"""Test CMR likelihood evaluation."""
model = cmr.CMR()
param = {
'B_enc': 0.5,
'B_rec': 0.8,
'B_start': 0,
'Afc': 0,
'Dfc': 1,
'Acf': 0,
'Dcf': 1,
'Lfc': 1,
'Lcf': 1,
'P1': 0,
'P2': 1,
'T': 10,
'X1': 0.05,
'X2': 1,
}
param_def, patterns = cmr.config_loc_cmr(6)
logl, n = model.likelihood(data, param, param_def=param_def, patterns=patterns)
np.testing.assert_allclose(logl, -5.936799964636842)
assert n == 6
def test_dist_cmr_record_trim(data, patterns, param_def_dist, param_dist):
model = cmr.CMR()
states = model.record(
data, param_dist, None, param_def_dist, patterns=patterns, remove_blank=True
)
np.testing.assert_allclose(states[0].c, np.array([0.5, 0.0, 0.0, 0.8660254]))
def test_dist_cmr_generate(data, patterns, param_def_dist, param_dist):
"""Test that CMR-D generation runs."""
model = cmr.CMR()
sim = model.generate(data, param_dist, None, param_def_dist, patterns=patterns)
assert isinstance(sim, pd.DataFrame)