def testFinalizeAndQuery(self):
""" Test the finalization and query functions. """
tsd = TimeStepDistribution()
# Set the member data.
histogram = numpy.array(numpy.random.rand(12) * 34, dtype=int)
normalized_histogram = histogram / float(numpy.sum(histogram))
binsize = 2.13
tsd._TimeStepDistribution__binsize = binsize
tsd._TimeStepDistribution__histogram = histogram
# Call the finalize function.
tsd.finalize()
# We now have the normalized histogram and timesteps on the class.
ret_histogram = tsd.histogram()
ret_norm_histogram = tsd.normalizedHistogram()
ret_time_steps = tsd.timeSteps()
# Check the histograms.
self.assertAlmostEqual(numpy.linalg.norm(ret_histogram - histogram),
0.0, 12)
self.assertAlmostEqual(
numpy.linalg.norm(ret_norm_histogram - normalized_histogram), 0.0,
12)
# Check the time steps.
time_steps = (numpy.arange(12) + 1) * binsize - binsize / 2.0
self.assertAlmostEqual(numpy.linalg.norm(ret_time_steps - time_steps),
0.0, 12)
def testSetup(self):
""" Test the setup interface function. """
# Construct.
tsd = TimeStepDistribution()
# Set the time.
time = numpy.random.rand()
# Call setup.
tsd.setup("step", time, "configuration")
# Check that the time was saved.
self.assertAlmostEqual(tsd._TimeStepDistribution__last_time, time, 12)
def testSetup(self):
""" Test the setup interface function. """
# Construct.
tsd = TimeStepDistribution()
# Set the time.
time = numpy.random.rand()
# Call setup.
tsd.setup("step", time, "configuration")
# Check that the time was saved.
self.assertAlmostEqual(tsd._TimeStepDistribution__last_time, time, 12)
def testFinalizeAndQuery(self):
""" Test the finalization and query functions. """
tsd = TimeStepDistribution()
# Set the member data.
histogram = numpy.array(numpy.random.rand(12)*34, dtype=int)
normalized_histogram = histogram / float(numpy.sum(histogram))
binsize = 2.13
tsd._TimeStepDistribution__binsize = binsize
tsd._TimeStepDistribution__histogram = histogram
# Call the finalize function.
tsd.finalize()
# We now have the normalized histogram and timesteps on the class.
ret_histogram = tsd.histogram()
ret_norm_histogram = tsd.normalizedHistogram()
ret_time_steps = tsd.timeSteps()
# Check the histograms.
self.assertAlmostEqual(numpy.linalg.norm(ret_histogram - histogram), 0.0, 12)
self.assertAlmostEqual(numpy.linalg.norm(ret_norm_histogram - normalized_histogram), 0.0, 12)
# Check the time steps.
time_steps = (numpy.arange(12)+1)*binsize - binsize/2.0
self.assertAlmostEqual(numpy.linalg.norm(ret_time_steps - time_steps), 0.0, 12)
def testConstruction(self):
""" Test the time step distribution analysis construction. """
# Check default member data.
tsd = TimeStepDistribution()
self.assertAlmostEqual(tsd._TimeStepDistribution__binsize, 1.0, 12)
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 10)
self.assertEqual(
numpy.linalg.norm(tsd._TimeStepDistribution__histogram -
numpy.zeros(10, dtype=int)), 0)
# Check construction with a given binsize.
tsd = TimeStepDistribution(binsize=2.3)
self.assertAlmostEqual(tsd._TimeStepDistribution__binsize, 2.3, 12)
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 10)
self.assertEqual(
numpy.linalg.norm(tsd._TimeStepDistribution__histogram -
numpy.zeros(10, dtype=int)), 0)
def testPrintResults(self):
""" Test the time step distribution print result function. """
tsd = TimeStepDistribution()
# Set the member data.
histogram = numpy.array(
[12, 41, 55, 23, 43, 12, 11, 19, 98, 97, 95, 93])
normalized_histogram = histogram / float(numpy.sum(histogram))
binsize = 2.13
tsd._TimeStepDistribution__binsize = binsize
tsd._TimeStepDistribution__histogram = histogram
# Call the finalize function.
tsd.finalize()
# Print the results to a stream.
stream = StringIO.StringIO()
tsd.printResults(stream)
ref_value = """ 1.06500 12 0.020033388981636
3.19500 41 0.068447412353923
5.32500 55 0.091819699499165
7.45500 23 0.038397328881469
9.58500 43 0.071786310517529
11.71500 12 0.020033388981636
13.84500 11 0.018363939899833
15.97500 19 0.031719532554257
18.10500 98 0.163606010016695
20.23500 97 0.161936560934891
22.36500 95 0.158597662771285
24.49500 93 0.155258764607679
"""
# Check the values.
if MPICommons.isMaster():
self.assertEqual(stream.getvalue(), ref_value)
else:
self.assertEqual(stream.getvalue(), "")
def testPrintResults(self):
""" Test the time step distribution print result function. """
tsd = TimeStepDistribution()
# Set the member data.
histogram = numpy.array([12,41,55,
23,43,12,
11,19,98,
97,95,93])
normalized_histogram = histogram / float(numpy.sum(histogram))
binsize = 2.13
tsd._TimeStepDistribution__binsize = binsize
tsd._TimeStepDistribution__histogram = histogram
# Call the finalize function.
tsd.finalize()
# Print the results to a stream.
stream = StringIO.StringIO()
tsd.printResults(stream)
ref_value = """ 1.06500 12 0.020033388981636
3.19500 41 0.068447412353923
5.32500 55 0.091819699499165
7.45500 23 0.038397328881469
9.58500 43 0.071786310517529
11.71500 12 0.020033388981636
13.84500 11 0.018363939899833
15.97500 19 0.031719532554257
18.10500 98 0.163606010016695
20.23500 97 0.161936560934891
22.36500 95 0.158597662771285
24.49500 93 0.155258764607679
"""
# Check the values.
if MPICommons.isMaster():
self.assertEqual(stream.getvalue(), ref_value)
else:
self.assertEqual(stream.getvalue(), "")
def testTimeStepDistribution(self):
""" Test the distribution for a system with only one moving particle. """
# Setup a system, a periodic 10x10x10 atoms long 3D block.
unit_cell = KMCUnitCell(cell_vectors=numpy.array([[1.0,0.0,0.0],
[0.0,1.0,0.0],
[0.0,0.0,1.0]]),
basis_points=[[0.0,0.0,0.0]])
# And a lattice.
lattice = KMCLattice(unit_cell=unit_cell,
repetitions=(10,10,10),
periodic=(True,True,True))
# Setup an initial configuration with one B in a sea of A:s.
types = ["A"]*10*10*10
types[5] = "B"
config = KMCConfiguration(lattice=lattice,
types=types,
possible_types=["A","B"])
# Setup a all diffusion processes.
coordinates_p0 = [[0.0, 0.0, 0.0],[-1.0, 0.0, 0.0]]
p0 = KMCProcess(coordinates=coordinates_p0,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.1)
coordinates_p1 = [[0.0, 0.0, 0.0],[1.0, 0.0, 0.0]]
p1 = KMCProcess(coordinates=coordinates_p1,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.2)
coordinates_p2 = [[0.0, 0.0, 0.0],[0.0,-1.0, 0.0]]
p2 = KMCProcess(coordinates=coordinates_p2,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.3)
coordinates_p3 = [[0.0, 0.0, 0.0],[0.0, 1.0, 0.0]]
p3 = KMCProcess(coordinates=coordinates_p3,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.4)
coordinates_p4 = [[0.0, 0.0, 0.0],[0.0, 0.0,-1.0]]
p4 = KMCProcess(coordinates=coordinates_p4,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.5)
coordinates_p5 = [[0.0, 0.0, 0.0],[0.0, 0.0, 1.0]]
p5 = KMCProcess(coordinates=coordinates_p5,
elements_before=["B","A"],
elements_after=["A","B"],
move_vectors=None,
basis_sites=[0],
rate_constant=0.6)
interactions = KMCInteractions(processes=[p0, p1, p2, p3, p4, p5],
implicit_wildcards=True)
model = KMCLatticeModel(configuration=config,
interactions=interactions)
# Setup the analysis.
tsd = TimeStepDistribution(binsize=0.001)
# Setup the control parameters.
control_parameters = KMCControlParameters(number_of_steps=1000000,
dump_interval=10000,
analysis_interval=1)
# Run the model.
model.run(control_parameters=control_parameters,
analysis=[tsd])
# Define the analytical reference function.
# The exponent is the total rate and the coefficient is
# chosen to normalize to one.
total_rate = 2.1
def ref(dt):
return total_rate*0.001*numpy.exp(-total_rate * dt)
hst = tsd.normalizedHistogram()
time = tsd.timeSteps()
sum_diff = 0.0
for t, h in zip(time, hst):
sum_diff += numpy.abs(ref(t) - h)
# Make sure the relative error is less than five percent.
self.assertTrue( sum_diff*100.0 < 5.0 )
def testRegisterStep(self):
""" Test the register step interface function. """
# Construct.
binsize = 1.23
tsd = TimeStepDistribution(binsize=binsize)
# Store the histogram size for reference.
h_size = len(tsd._TimeStepDistribution__histogram)
# Set the first time and call setup.
t0 = numpy.random.rand()
tsd.setup("step", t0, "configuration")
# Call the step function with a new time.
t1 = t0 + 5 * numpy.random.rand()
tsd.registerStep("step", t1, "configuration")
# Calculate the bin.
b = int((t1 - t0) / binsize)
# Check that this bin was incremented.
self.assertEqual(tsd._TimeStepDistribution__histogram[b], 1)
for i in range(10):
if i != b:
self.assertEqual(tsd._TimeStepDistribution__histogram[i], 0)
# Give a new value that requires the histogram to be extended.
t2 = 10
tsd.registerStep("step", t2, "configuration")
t3 = 23
tsd.registerStep("step", t3, "configuration")
# Check that the length has incresed to 11.
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 11)
self.assertEqual(tsd._TimeStepDistribution__histogram[10], 1)
# Increment again. This should end up with a length of 22.
t4 = 48
tsd.registerStep("step", t4, "configuration")
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 22)
#NiO process
import time
from KMCLib import *
from KMCLib.Analysis.TimeStepDistribution import TimeStepDistribution
from KMCLib.Analysis.ProcessStatistics import ProcessStatistics
from matplotlib import pyplot as plt
t_calstart = time.clock()
configuration = KMCConfigurationFromScript("config.py")
interactions = KMCInteractionsFromScript("interactions.py")
model = KMCLatticeModel(configuration=configuration, interactions=interactions)
control_parameters = KMCControlParameters(number_of_steps=10000000,
dump_interval=10000,
analysis_interval=1)
#set up the analysis
tsd = TimeStepDistribution(binsize=0.001)
pst = ProcessStatistics()
# Run the simulation - save trajectory to 'NiO POM.py'
model.run(control_parameters,
trajectory_filename="NiO POM.py",
analysis=[tsd, pst])
t_calend = time.clock()
print("sim complete time cost is :", t_calend - t_calstart)
#length of lattice
size = 100
def testRegisterStep(self):
""" Test the register step interface function. """
# Construct.
binsize = 1.23
tsd = TimeStepDistribution(binsize=binsize)
# Store the histogram size for reference.
h_size = len(tsd._TimeStepDistribution__histogram)
# Set the first time and call setup.
t0 = numpy.random.rand()
tsd.setup("step", t0, "configuration")
# Call the step function with a new time.
t1 = t0 + 5*numpy.random.rand()
tsd.registerStep("step", t1, "configuration")
# Calculate the bin.
b = int((t1 - t0) / binsize)
# Check that this bin was incremented.
self.assertEqual(tsd._TimeStepDistribution__histogram[b], 1)
for i in range(10):
if i != b:
self.assertEqual(tsd._TimeStepDistribution__histogram[i], 0)
# Give a new value that requires the histogram to be extended.
t2 = 10
tsd.registerStep("step", t2, "configuration")
t3 = 23
tsd.registerStep("step", t3, "configuration")
# Check that the length has incresed to 11.
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 11)
self.assertEqual(tsd._TimeStepDistribution__histogram[10], 1)
# Increment again. This should end up with a length of 22.
t4 = 48
tsd.registerStep("step", t4, "configuration")
self.assertEqual(len(tsd._TimeStepDistribution__histogram), 22)
