def test_population_rset(self):
import pyhmf as pynn
pynn.setup()
p = pynn.Population(123, pynn.EIF_cond_exp_isfa_ista)
r1, r2 = pynn.NativeRNG(0), pynn.NativeRNG(0)
d1 = pynn.RandomDistribution("uniform", [-55.0, -50.0])
d2 = pynn.RandomDistribution("uniform", [-55.0, -50.0])
p.rset("v_reset", d1)
assert_array_equal(p.get("v_reset"), d2.next(len(p)))
def test_binomial_distribution(self):
import pyhmf as pynn
pynn.setup()
rng = pynn.NativeRNG()
self.assertEqual(rng.next(
distribution = "binomial", parameters = [100, 1.]), 100)
def test_distribution_clip(self):
import pyhmf as pynn
pynn.setup()
r = pynn.NativeRNG(0)
d = pynn.RandomDistribution("uniform", [0.0, 25.0], r, (12.5, 12.5), "clip")
assert_array_equal(d.next(100), numpy.ones(100) * 12.5)
def build(self):
weights = self._create_nn_unit_weights(self.linearsize, self.dimension)
self.neurons = [
pynn.Population(1, self.model)
for _ in range(self.linearsize**self.dimension)
]
self.exsources = pynn.Population(self.nsources,
pynn.SpikeSourcePoisson,
{'rate': self.sourcerate})
self.insources = pynn.Population(self.nsources,
pynn.SpikeSourcePoisson,
{'rate': self.sourcerate})
self.biasneurons = pynn.Population(self.nbiasneurons, self.model)
connector = pynn.FixedNumberPreConnector(n=self.ksources, weights=0.3)
proj = pynn.Projection(self.exsources,
self.neurons,
connector,
target='excitatory',
rng=pynn.NativeRNG(42))
proj = pynn.Projection(self.exsources,
self.neurons,
connector,
target='inhibitory',
rng=pynn.NativeRNG(42))
connector = pynn.FixedNumberPreConnector(n=self.nbiasneurons,
weights=0.4)
proj = pynn.Projection(self.biasneurons,
self.neurons,
connector,
target='inhibitory',
rng=pynn.NativeRNG(42))
for ipre, ipost, w in weights:
connector = pynn.AllToAllConnector(weights=1)
proj = pynn.Projection(self.neurons[ipre],
self.neurons[ipost],
connector,
target="excitatory")
def build(self):
weights = self._create_nn_unit_weights(self.linearsize, self.dimension)
self.neurons = [
pynn.Population(1, self.model)
for _ in range(self.linearsize**self.dimension)
]
self.noise = pynn.Population(self.nsources, pynn.IF_cond_exp)
self.biasneurons = pynn.Population(self.nbiasneurons, self.model)
connector = pynn.FixedNumberPreConnector(n=30,
weights=0.3,
allow_self_connections=False)
pynn.Projection(self.noise, self.noise, connector, target='inhibitory')
connector = pynn.FixedNumberPreConnector(n=self.ksources, weights=0.3)
pynn.Projection(self.noise,
self.neurons,
connector,
target='excitatory',
rng=pynn.NativeRNG(42))
pynn.Projection(self.noise,
self.neurons,
connector,
target='inhibitory',
rng=pynn.NativeRNG(43))
connector = pynn.FixedNumberPreConnector(n=self.kbiasneurons,
weights=0.4)
pynn.Projection(self.biasneurons,
self.neurons,
connector,
target='inhibitory',
rng=pynn.NativeRNG(44))
for ipre, ipost, w in weights:
connector = pynn.AllToAllConnector(weights=1)
pynn.Projection(self.neurons[ipre],
self.neurons[ipost],
connector,
target="excitatory")
def build(self):
self.neurons = pynn.Population(self.N, self.model)
connector = pynn.FixedProbabilityConnector(p_connect=self.prob,
allow_self_connections=True,
weights=0.003)
pynn.Projection(self.neurons,
self.neurons,
connector,
target='excitatory',
rng=pynn.NativeRNG(42))
def build(self):
self.neurons = pynn.Population(self.N, self.model)
connector = pynn.FixedNumberPreConnector(self.K,
weights=1,
allow_self_connections=False)
pynn.Projection(self.neurons,
self.neurons,
connector,
target='excitatory',
rng=pynn.NativeRNG(42))
def test_normaldistribution(self):
import pyhmf as pynn
pynn.setup()
mean = 42.0
std = 3.3
r = pynn.NativeRNG(10)
d = pynn.RandomDistribution("normal", [mean, std])
data = d.next(100000)
self.assertAlmostEqual(numpy.mean(data), mean, places=1)
self.assertAlmostEqual(numpy.std(data), std, places=1)
def test_distribution_redraw(self):
import pyhmf as pynn
pynn.setup()
r = pynn.NativeRNG(0)
d = pynn.RandomDistribution("uniform", [0.0, 25.0], r, (5.0, 20.0), "redraw")
data = d.next(260)
self.assertLessEqual(max(data), 20.0)
self.assertGreaterEqual(min(data), 5.0)
data2 = data[data <= 5.0]
data2 = data2[data2 >= 20]
self.assertEqual(len(data2), 0)
def test_init_random(self):
size = random.randint(2, 10)**2
seed = 1337
pynn_dist = pynn.RandomDistribution(distribution='uniform',
rng=NativeRNGProxy(seed))
pyhmf_dist = pyhmf.RandomDistribution(distribution='uniform',
rng=pyhmf.NativeRNG(seed))
numpy.testing.assert_almost_equal(
self.construct_with_backend(pynn, size, pynn_dist),
self.construct_with_backend(pyhmf, size, pyhmf_dist))
def build(self):
self.neurons = []
for i in range(self.num_layers):
self.neurons.append(
pynn.Population(self.neurons_per_layer, self.model))
connector = pynn.FixedProbabilityConnector(
p_connect=self.conn_prob,
allow_self_connections=False,
weights=0.003)
proj = []
for i in range(1, self.num_layers):
proj = pynn.Projection(self.neurons[i - 1],
self.neurons[i],
connector,
target='excitatory',
rng=pynn.NativeRNG(42))
def test_projection(self):
path = tempfile.mkstemp()[1]
size_a = random.randint(1, 100)
size_b = random.randint(1, 100)
dist = pyhmf.RandomDistribution(rng=pyhmf.NativeRNG(1337))
conn_pyhmf = pyhmf.AllToAllConnector(weights=dist, delays=42)
proj_pyhmf = pyhmf.Projection(
pyhmf.Population(size_a, pyhmf.IF_cond_exp),
pyhmf.Population(size_b, pyhmf.IF_cond_exp),
conn_pyhmf)
proj_pyhmf.saveConnections(getattr(pyhmf, self.file_type)(path, 'w'))
conn_pynn = pynn.FromFileConnector(getattr(pynn.recording.files, self.file_type)(path))
proj_pynn = pynn.Projection(
pynn.Population(size_a, pynn.IF_cond_exp),
pynn.Population(size_b, pynn.IF_cond_exp),
conn_pynn)
numpy.testing.assert_equal(proj_pyhmf.getWeights(format='array'), proj_pynn.getWeights(format='array'))
def __init__(self, *args, **kwargs):
self.random = pyhmf.NativeRNG()
def __init__(self, *args, **kwargs):
self.random = pyhmf.NativeRNG()
unittest.TestCase.__init__(self, *args, **kwargs)
def __init__(self, *args, **kwargs):
AbstractRNG.__init__(self)
self._impl = pyhmf.NativeRNG(*args, **kwargs)
