def setUp(self):
super(TestDumpLoad, self).setUp()
self.lists = [[],
list(range(1000)),
list("Although that way may not be" +
"obvious at first" +
"unless you're Dutch.")]
self.arrays = [numpy.array([]),
numpy.random.random(1000),
numpy.arange(1000), ]
self.strings = ["",
"Although that way may not be" + \
"obvious at first" + \
"unless you're Dutch.",
"([0-9a-zA-Z]([-.\w]*[0-9a-zA-Z])"]
# Completly connnected graph
self.big_graph = []
for _i in range(250):
n = Node()
self.big_graph.append(n)
for n in self.big_graph:
n.connections = set(self.big_graph)
n.connections.remove(n)
# Protein
pdbfile = self.config.getTestFile('ake-xray-ensemble-ca.pdb')
self.protein = LegacyStructureParser(pdbfile).parse_models()[0]
self.objs = [self.lists, self.arrays, self.strings,
self.protein]
def setUp(self):
super(TestDumpLoad, self).setUp()
self.lists = [[],
list(range(1000)),
list("Although that way may not be" +
"obvious at first" + "unless you're Dutch.")]
self.arrays = [
numpy.array([]),
numpy.random.random(1000),
numpy.arange(1000),
]
self.strings = ["",
"Although that way may not be" + \
"obvious at first" + \
"unless you're Dutch.",
"([0-9a-zA-Z]([-.\w]*[0-9a-zA-Z])"]
# Completly connnected graph
self.big_graph = []
for _i in range(250):
n = Node()
self.big_graph.append(n)
for n in self.big_graph:
n.connections = set(self.big_graph)
n.connections.remove(n)
# Protein
pdbfile = self.config.getTestFile('ake-xray-ensemble-ca.pdb')
self.protein = LegacyStructureParser(pdbfile).parse_models()[0]
self.objs = [self.lists, self.arrays, self.strings, self.protein]
def __init__(self, alpha=1., beta=1., hyper_alpha=(4., 1.),
hyper_beta=(2., 1.)):
super(InvGammaPrior, self).__init__(alpha, beta)
self._hyper_alpha = Gamma(*hyper_alpha)
self._hyper_beta = Gamma(*hyper_beta)
self.estimator = InvGammaPosteriorSampler()
def testLogProb(self):
g = Gamma(1., 1.)
self.assertAlmostEqual(g(1.), numpy.exp(-1.), delta=1e-5)
self.assertAlmostEqual(g(2.), numpy.exp(-2.), delta=1e-5)
g = Gamma(2., 1.)
self.assertAlmostEqual(g(1.), 1. * numpy.exp(-1), delta=1e-5)
self.assertAlmostEqual(g(2.), 2. * numpy.exp(-2), delta=1e-5)
def testParameterEstimation(self):
alpha = 0.1
beta = 0.1
data = numpy.random.gamma(alpha, 1. / beta, 10000)
pdf = Gamma(1, 1)
pdf.estimate(data)
self.assertAlmostEqual(pdf.alpha, alpha, places=2)
self.assertAlmostEqual(pdf.beta, beta, places=1)
def testParameterEstimation(self):
alpha = 0.1
beta = 0.1
data = numpy.random.gamma(alpha, 1. / beta, 10000)
pdf = Gamma(1, 1)
pdf.estimate(data)
self.assertAlmostEqual(pdf.alpha, alpha, places=2)
self.assertAlmostEqual(pdf.beta, beta, places=1)
def calculate_KL((posterior_distances, prior_distances, bins)):
from csb.statistics.pdf import Gamma
from csb.numeric import log
g = Gamma()
g.estimate(prior_distances)
posterior_hist = np.histogram(posterior_distances, bins=bins,
normed=True)[0]
return np.trapz(posterior_hist * log(posterior_hist / g(bins[:-1])),
bins[:-1])
def testRandom(self):
gamma = Gamma(1., 1.)
samples = gamma.random(10000)
mu = numpy.mean(samples)
var = numpy.var(samples)
self.assertAlmostEqual(gamma.alpha / gamma.beta, mu, delta=1e-1)
self.assertAlmostEqual(gamma.alpha / gamma.beta**2, var, delta=1e-1)
def testParameters(self):
pdf = Gamma(1, 1)
pdf.alpha = 2
pdf.beta = 3
self.assertEqual(pdf.alpha, 2)
self.assertEqual(pdf.beta, 3)
self.assertEqual(pdf.beta, pdf['beta'])
def propertyAssignment():
pdf.beta = -1
def directAssignment(p):
pdf[p] = -1
self.assertRaises(ParameterValueError, propertyAssignment)
self.assertRaises(ParameterValueError, directAssignment, 'alpha')
self.assertRaises(ParameterValueError, directAssignment, 'beta')
def testParameters(self):
pdf = Gamma(1, 1)
pdf.alpha = 2
pdf.beta = 3
self.assertEqual(pdf.alpha, 2)
self.assertEqual(pdf.beta, 3)
self.assertEqual(pdf.beta, pdf['beta'])
def propertyAssignment():
pdf.beta = -1
def directAssignment(p):
pdf[p] = -1
self.assertRaises(ParameterValueError, propertyAssignment)
self.assertRaises(ParameterValueError, directAssignment, 'alpha')
self.assertRaises(ParameterValueError, directAssignment, 'beta')
