def setUp(self):
folder = os.path.dirname(os.path.realpath(__file__))
self.time_original = 8
spiketrain = np.array([1, 3, 5, 6])
self.values = [spiketrain, spiketrain, spiketrain, np.array([1])]
self.features = NetworkFeatures()
def test_preprocess(self):
self.features = NetworkFeatures()
time, spiketrains = self.features.preprocess(self.time_original, self.values)
self.assertEqual(time, self.time_original)
self.assertIsInstance(spiketrains[0], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[1], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[2], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[3], neo.core.SpikeTrain)
self.assertTrue(np.array_equal(spiketrains[0], self.values[0]))
self.assertTrue(np.array_equal(spiketrains[1], self.values[1]))
self.assertTrue(np.array_equal(spiketrains[2], self.values[2]))
self.assertTrue(np.array_equal(spiketrains[3], self.values[3]))
self.assertEqual(spiketrains[0].t_stop, self.time_original)
def setUp(self):
folder = os.path.dirname(os.path.realpath(__file__))
self.time_original = 8
spiketrain = np.array([1, 3, 5, 6])
self.values = [spiketrain, spiketrain, spiketrain, np.array([1])]
self.implemented_features = [
"cv", "mean_cv", "binned_isi", "mean_isi", "local_variation",
"mean_firing_rate", "fanofactor", "instantaneous_rate",
"van_rossum_dist", "victor_purpura_dist", "corrcoef", "covariance",
"mean_local_variation"
]
self.features = NetworkFeatures(instantaneous_rate_nr_samples=2)
self.time, self.spiketrains = self.features.preprocess(
self.time_original, self.values)
def test_init(self):
def feature(time, values):
return "t", "U"
features = NetworkFeatures(new_features=feature,
features_to_run=None,
adaptive=["cv"],
labels={"cv": ["test"]},
instantaneous_rate_nr_samples=-1.,
isi_bin_size=-1,
corrcoef_bin_size=-1,
covariance_bin_size=-1,
units=pq.ms)
self.assertIsInstance(features, NetworkFeatures)
self.assertEqual(features.features_to_run, [])
self.assertEqual(features.adaptive, ["cv"])
self.assertEqual(features.instantaneous_rate_nr_samples, -1)
self.assertEqual(features.isi_bin_size, -1)
self.assertEqual(features.corrcoef_bin_size, -1)
self.assertEqual(features.covariance_bin_size, -1)
self.assertEqual(features.units, pq.ms)
self.assertEqual(set(features.implemented_features()),
set(["feature"] + self.implemented_features))
def test_initNone(self):
self.features = NetworkFeatures()
self.assertIsInstance(self.features, NetworkFeatures)
class TestNetworkFeatures(unittest.TestCase):
def setUp(self):
folder = os.path.dirname(os.path.realpath(__file__))
self.time_original = 8
spiketrain = np.array([1, 3, 5, 6])
self.values = [spiketrain, spiketrain, spiketrain, np.array([1])]
self.implemented_features = [
"cv", "mean_cv", "binned_isi", "mean_isi", "local_variation",
"mean_firing_rate", "fanofactor", "instantaneous_rate",
"van_rossum_dist", "victor_purpura_dist", "corrcoef", "covariance",
"mean_local_variation"
]
self.features = NetworkFeatures(instantaneous_rate_nr_samples=2)
self.time, self.spiketrains = self.features.preprocess(
self.time_original, self.values)
def test_initNone(self):
self.features = NetworkFeatures()
self.assertIsInstance(self.features, NetworkFeatures)
def test_init(self):
def feature(time, values):
return "t", "U"
features = NetworkFeatures(new_features=feature,
features_to_run=None,
adaptive=["cv"],
labels={"cv": ["test"]},
instantaneous_rate_nr_samples=-1.,
isi_bin_size=-1,
corrcoef_bin_size=-1,
covariance_bin_size=-1,
units=pq.ms)
self.assertIsInstance(features, NetworkFeatures)
self.assertEqual(features.features_to_run, [])
self.assertEqual(features.adaptive, ["cv"])
self.assertEqual(features.instantaneous_rate_nr_samples, -1)
self.assertEqual(features.isi_bin_size, -1)
self.assertEqual(features.corrcoef_bin_size, -1)
self.assertEqual(features.covariance_bin_size, -1)
self.assertEqual(features.units, pq.ms)
self.assertEqual(set(features.implemented_features()),
set(["feature"] + self.implemented_features))
def test_preprocess(self):
self.features = NetworkFeatures()
time, spiketrains = self.features.preprocess(self.time_original,
self.values)
self.assertEqual(time, self.time_original)
self.assertIsInstance(spiketrains[0], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[1], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[2], neo.core.SpikeTrain)
self.assertIsInstance(spiketrains[3], neo.core.SpikeTrain)
self.assertTrue(np.array_equal(spiketrains[0], self.values[0]))
self.assertTrue(np.array_equal(spiketrains[1], self.values[1]))
self.assertTrue(np.array_equal(spiketrains[2], self.values[2]))
self.assertTrue(np.array_equal(spiketrains[3], self.values[3]))
self.assertEqual(spiketrains[0].t_stop, self.time_original)
def test_cv(self):
time, values = self.features.cv(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertTrue(
np.array_equal([
0.51207638319124049, 0.51207638319124049, 0.51207638319124049,
0
], values))
def test_mean_cv(self):
time, values = self.features.mean_cv(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(0.38405728739343037, values)
def test_binned_isi(self):
time, values = self.features.binned_isi(self.time, self.spiketrains)
centers = np.arange(0, self.time_original + 1)[1:] - 0.5
self.assertTrue(np.array_equal(centers, time))
self.assertTrue(
np.array_equal(
values, [[0, 1, 2, 0, 0, 0, 0, 0], [0, 1, 2, 0, 0, 0, 0, 0],
[0, 1, 2, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0, 0]]))
def test_mean_isi(self):
time, values = self.features.mean_isi(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(1.66666666666666667, values)
def test_lv(self):
time, values = self.features.local_variation(self.time,
self.spiketrains)
self.assertIsNone(time)
self.assertTrue(
np.array_equal([
0.16666666666666666, 0.16666666666666666, 0.16666666666666666,
None
], values))
def test_mean_lv(self):
time, values = self.features.mean_local_variation(
self.time, self.spiketrains)
self.assertIsNone(time)
mean = np.mean(
[0.16666666666666666, 0.16666666666666666, 0.16666666666666666])
self.assertEqual(mean, values)
def test_mean_firing_rate(self):
time, values = self.features.mean_firing_rate(self.time,
self.spiketrains)
self.assertIsNone(time)
self.assertTrue(np.array_equal([500, 500, 500, 125], values))
def test_instantaneous_rate(self):
time, values = self.features.instantaneous_rate(
self.time, self.spiketrains)
rates = [938.762478, 938.7583322]
correct_t = np.linspace(0, 8, 3)[:-1]
self.assertTrue(np.all(np.isclose(values[0], rates)))
self.assertTrue(np.all(np.isclose(values[1], rates)))
self.assertTrue(np.all(np.isclose(values[2], rates)))
self.assertIsNone(values[3])
self.assertTrue(np.array_equal(time, correct_t))
def test_fanofactor(self):
time, values = self.features.fanofactor(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(values, 0.51923076923076927)
def test_van_rossum_dist(self):
time, values = self.features.van_rossum_dist(self.time,
self.spiketrains)
self.assertIsNone(time)
self.assertEqual(values.shape, (4, 4))
# correct_values = [[0.0, 5.9604644775390625e-08, 5.9604644775390625e-08, 2.9980016657780828],
# [5.9604644775390625e-08, 0.0, 5.9604644775390625e-08, 2.9980016657780828],
# [5.9604644775390625e-08, 5.9604644775390625e-08, 0.0, 2.9980016657780828],
# [2.9980016657780828, 2.9980016657780828, 2.9980016657780828, 0.0]]
# self.assertTrue(np.array_equal(values, correct_values))
diag = np.diag_indices(4)
self.assertTrue(np.all(values[diag] == 0))
def test_victor_purpura_dist(self):
time, values = self.features.victor_purpura_dist(
self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(values.shape, (4, 4))
diag = np.diag_indices(4)
self.assertTrue(np.all(values[diag] == 0))
def test_corrcoef(self):
time, values = self.features.corrcoef(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(values.shape, (4, 4))
diag = np.diag_indices(4)
self.assertTrue(np.all(values[diag] == 1))
def test_covariance(self):
time, values = self.features.covariance(self.time, self.spiketrains)
self.assertIsNone(time)
self.assertEqual(values.shape, (4, 4))
def test_reference_feature(self):
time, values = self.features.reference_feature(1, 1)
self.assertIsNone(time)
self.assertIsNone(values)
def test_initNone(self):
self.features = NetworkFeatures(logger_level="error")
self.assertIsInstance(self.features, NetworkFeatures)
