def test_data_point():
"""obtain a point cloud based on the data
"""
# build world
world = tt.World(name="model test", ndim=3)
# test (x, Y)
x = np.linspace(0, 1, 100)
Y = np.zeros((100, 3)) # << dimension
Y[:, 0] = x
Y[:, 1] = x
Y[:, 2] = x
a = world._get_point_from_xY(x, Y, x1=0.0)
a_expected = np.array([0.0, 0.0, 0.0])
np.testing.assert_array_almost_equal_nulp(a, a_expected)
a = world._get_point_from_xY(x, Y, x1=1.0)
a_expected = np.array([1.0, 1.0, 1.0])
np.testing.assert_array_almost_equal_nulp(a, a_expected)
# test cluster_data
cluster_data = []
cluster_data.append((x, Y)) # 1
cluster_data.append((x, Y)) # 2
cluster_data.append((x, Y)) # 3
cluster_data.append((x, Y)) # 4
A = world._get_point_from_cluster_data(cluster_data, x1=0.0)
A_expected = np.zeros((4, 3))
np.testing.assert_array_almost_equal_nulp(A, A_expected)
A = world._get_point_from_cluster_data(cluster_data, x1=1.0)
A_expected = np.ones((4, 3))
np.testing.assert_array_almost_equal_nulp(A, A_expected)
# add trajectories
world.addCluster(cluster_data, "1")
world.addCluster(cluster_data, "2")
# obtain list of points
clusterP = world.getClusterPoints(x1=0.0)
# check values inside clusters
for i in [0, 1]:
np.testing.assert_array_almost_equal_nulp(clusterP[i], np.zeros(
(4, 3)))
# obtain list of points
clusterP = world.getClusterPoints(x1=1.0)
# check values
for i in [0, 1]:
np.testing.assert_array_almost_equal_nulp(clusterP[i], np.ones((4, 3)))
def test_model():
"""
tests the modelling functionality
"""
# build world
world_1 = tt.World(name="model test", ndim=3)
# add trajectories
for ntype in [0, 1]:
correct_cluster_name = "toy {0}".format(ntype)
correct_cluster_data = tt.helpers.get_trajectories(ntype,
ndim=3,
ntraj=20)
world_1.addCluster(correct_cluster_data, correct_cluster_name)
# model all trajectories
settings = {}
settings["model_type"] = "resampling"
settings["ngaus"] = 10
# build a model
world_1.buildModel(settings)
with pt.raises(TypeError) as testException:
world_1.buildModel(settings, "Hello World!")
with pt.raises(ValueError) as testException:
world_1.buildModel(settings, [-1])
# log-likelihood
(ss_list, [xx, yy, zz]) = world_1.getLogLikelihood([0, 1])
with pt.raises(TypeError) as testException:
(ss_list, [xx, yy, zz]) = world_1.getLogLikelihood("Hello World!")
with pt.raises(ValueError) as testException:
(ss_list, [xx, yy, zz]) = world_1.getLogLikelihood([-1])
# build tube (twice!)
for i in range(2):
(ss_list, [xx, yy, zz]) = world_1.getTube([0, 1])
# complexity map
ss, xx, yy, zz = world_1.getComplexityMap([0, 1])
# overview
world_1.overview()
# produce some numbers
world_1.getTubeStats()
# clear world
world_1.clear()
def test_visual_3d():
"""
can produce figures
"""
# build world
world_1 = tt.World(name="Example 3D", ndim=3, resolution=[10, 10, 10])
# extreme reduced resolution
# world_1.setResolution(xstep=3, ystep=3, zstep=3)
# add trajectories
for mtype in [0, 1]:
correct_cluster_name = "toy {0}".format(mtype)
correct_cluster_data = tt.helpers.get_trajectories(mtype,
ndim=3,
ntraj=20)
world_1.addCluster(correct_cluster_data, correct_cluster_name)
list_icluster = [0, 1]
# model all trajectories
settings = {}
settings["model_type"] = "resampling"
settings["ngaus"] = 10
world_1.buildModel(settings)
# this part is Python 2.7 [ONLY] due to Mayavi / VTK dependencies
# visuals by mayavi
visual = tt.visual_3d.Visual_3d(world_1)
# visualise trajectories
visual.plotTrajectories(list_icluster, ntraj=3)
visual.plotSamples(list_icluster, ntraj=3)
# visualise points
# visualise trajectories
visual.plotTrajectoriesPoints(x1=0.0, list_icluster=None, ntraj=3)
# visualise intersection
visual.plotLogLikelihood()
#
visual.plotComplexityMap()
# visualise tube
visual.plotTube()
# visualise difference
visual.plotTubeDifference([0, 1])
# visualise outline
visual.plotOutline()
# close
visual.close()
def test_init():
"""
<description>
"""
# test 1
# set values
name_1 = "Hello world!"
D_1 = 2
world_1 = tt.World(name=name_1, ndim=D_1)
assert (world_1._name == name_1)
assert (world_1._ndim == D_1)
# test 2
# default values
world_2 = tt.World()
assert (world_2._name == "")
assert (world_2._ndim == 3)
# test 3
# bad name
name_3 = 5
with pt.raises(TypeError) as testException:
world_3 = tt.World(name_3)
# bad TYPE dimension
name_4 = "Hello world!"
D_4 = "Hello World!"
with pt.raises(TypeError) as testException:
world_4 = tt.World(name_4, D_4)
# bad VALUE dimension
name_5 = "Hello World!"
D_5 = 1
with pt.raises(ValueError) as testException:
world_5 = tt.World(name_5, D_5)
def test_addCluster():
"""
<description>
"""
mdim = 3 # dimensionality
# build world
world_1 = tt.World(name="test", ndim=mdim)
# build a valid cluster
correct_cluster_name = "correct data"
# normal operation
for mtype in [0, 1]:
correct_cluster_data = tt.helpers.get_trajectories(mtype,
ndim=mdim,
ntraj=5)
world_1.addCluster(correct_cluster_data, correct_cluster_name)
#
these_clusters = world_1.getCluster()
assert (len(these_clusters) == 2)
#
wrong_cluster_name = 5
with pt.raises(TypeError) as testException:
world_1.addCluster(correct_cluster_data, wrong_cluster_name)
#
wrong_cluster_data = 5
with pt.raises(TypeError) as testException:
world_1.addCluster(wrong_cluster_data, correct_cluster_name)
#
wrong_trajectory_data = 5
wrong_cluster_data = correct_cluster_data
wrong_cluster_data.append(wrong_trajectory_data)
with pt.raises(TypeError) as testException:
world_1.addCluster(wrong_cluster_data, correct_cluster_name)
def test_visual_2d():
"""
produce figures
"""
mdim = 2
mtraj = 20
# build world
world_1 = tt.World(name="Example 3D", ndim=2, resolution=100)
# extreme reduced resolution
#world_1.setResolution(xstep=3, ystep=3, zstep=3)
# add trajectories
for mtype in [0, 1]:
correct_cluster_name = "toy {0}".format(mtype)
correct_cluster_data = tt.helpers.get_trajectories(mtype,
ndim=mdim,
ntraj=mtraj)
world_1.addCluster(correct_cluster_data, correct_cluster_name)
list_iclusters = [0, 1]
# model all trajectories
settings = {}
settings["model_type"] = "resampling"
settings["ngaus"] = 10
world_1.buildModel(settings)
# visuals by
visual = tt.visual_2d.Visual_2d(world_1)
# visualise trajectories
visual.plotTrajectories(list_iclusters)
# visualise intersection
visual.plotLogLikelihood()
# close
visual.close()
def test_visual_2d():
"""
produce figures
"""
mdim = 2
mtraj = 20
# build world
world_1 = tt.World(name="Example 3D", ndim=2, resolution=100)
# extreme reduced resolution
#world_1.setResolution(xstep=3, ystep=3, zstep=3)
# add trajectories
for mtype in [0, 1]:
correct_cluster_name = "toy {0}".format(mtype)
correct_cluster_data = tt.helpers.get_trajectories(mtype,
ndim=mdim,
ntraj=mtraj)
world_1.addCluster(correct_cluster_data, correct_cluster_name)
list_iclusters = [0, 1]
# model all trajectories
settings = {}
settings["model_type"] = "resampling"
settings["ngaus"] = 10
world_1.buildModel(settings)
# visuals by
visual = tt.visual_2d.Visual_2d(world_1)
# visualise intersection
visual.plotLogLikelihood()
# close
visual.close()
# visuals by
visual = tt.visual_2d.Visual_2d(world_1)
# visualise intersection
visual.plotComplexityMap()
# close
visual.close()
# visuals by
visual = tt.visual_2d.Visual_2d(world_1)
# visualise intersection
visual.plotTube()
# close
visual.close()
# visuals by
visual = tt.visual_2d.Visual_2d(world_1)
# visualise intersection
visual.plotTubeDifference()
# close
visual.close()
# new figure
visual = tt.visual_2d.Visual_2d(world_1)
# visualise trajectories
visual.plotTrajectories(list_iclusters)
# visualise points
visual.plotTrajectoriesPoints(x1=0.0, list_icluster=None)
# mean
visual.plotMean()
#
visual.plotSamples()
#
visual.plotBox(np.array([0.0, 0.0]), np.array([1.0, 1.0]))
# close
visual.close()
for ls in llsettings:
settings = {"model_type":ls[0],
"ngaus":ls[1],
"basis_type":ls[2],
"nbasis":ls[3]}
# build world
world_name = "[{0}] [{1}] [{2}] [{3}]".format(settings["model_type"],
settings["basis_type"],
settings["nbasis"],
ls[4])
# build world
new_world = tt.World(name=world_name, ndim=3, resolution=[10, 10, 10])
# add trajectories
for ntype in [0, 1]:
cluster_name = "toy {0}".format(ntype)
cluster_data = tt.helpers.get_trajectories(ntype,
ndim=3,
ntraj=50,
npoints=100,
noise_std=ls[4])
new_world.addCluster(cluster_data, cluster_name)
# overview
new_world.overview()
# model
for ls in llsettings:
settings = {
"model_type": ls[0],
"ngaus": ls[1],
"basis_type": ls[2],
"nbasis": ls[3]
}
# build world
world_name = "[{0}] [{1}] [{2}] [{3}]".format(settings["model_type"],
settings["basis_type"],
settings["nbasis"], ls[4])
# create a new world
new_world = tt.World(name=world_name, ndim=2, resolution=[50, 50, 50])
# add trajectories
for ntype in [0, 1]:
cluster_name = "toy {0}".format(ntype)
cluster_data = tt.helpers.get_trajectories(ntype,
ndim=2,
ntraj=50,
npoints=100,
noise_std=ls[4])
new_world.addCluster(cluster_data, cluster_name)
# output an overview
new_world.overview()
# build the model
