文件
gprMax/testing/test_models.py
2023-02-17 10:48:29 +00:00

255 行
11 KiB
Python

# Copyright (C) 2015-2022: The University of Edinburgh, United Kingdom
# Authors: Craig Warren, Antonis Giannopoulos, and John Hartley
#
# This file is part of gprMax.
#
# gprMax is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# gprMax is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with gprMax. If not, see <http://www.gnu.org/licenses/>.
import logging
import sys
from pathlib import Path
import gprMax
import h5py
import matplotlib.pyplot as plt
import numpy as np
from testing.analytical_solutions import hertzian_dipole_fs
logger = logging.getLogger(__name__)
if sys.platform == 'linux':
plt.switch_backend('agg')
"""Compare field outputs
Usage:
cd gprMax
python -m tests.test_models
"""
# Specify directoty with set of models to test
modelset = 'models_basic'
# modelset += 'models_advanced'
# modelset += 'models_pmls'
basepath = Path(__file__).parents[0] / modelset
# List of available basic test models
testmodels = ['hertzian_dipole_fs_analytical', '2D_ExHyHz', '2D_EyHxHz', '2D_EzHxHy',
'cylinder_Ascan_2D', 'hertzian_dipole_fs', 'hertzian_dipole_hs',
'hertzian_dipole_dispersive', 'magnetic_dipole_fs']
# List of available advanced test models
# testmodels = ['antenna_GSSI_1500_fs', 'antenna_MALA_1200_fs']
# List of available PML models
# testmodels = ['pml_x0', 'pml_y0', 'pml_z0', 'pml_xmax', 'pml_ymax', 'pml_zmax', 'pml_3D_pec_plate']
# Select a specific model if desired
# testmodels = testmodels[:-1]
# testmodels = [testmodels[6]]
testresults = dict.fromkeys(testmodels)
path = '/rxs/rx1/'
# Minimum value of difference to plot (dB)
plotmin = -160
for i, model in enumerate(testmodels):
testresults[model] = {}
# Run model
file = basepath / model / model
gprMax.run(inputfile=file.with_suffix('.in'), gpu=None)
# Special case for analytical comparison
if model == 'hertzian_dipole_fs_analytical':
# Get output for model file
filetest = h5py.File(file.with_suffix('.h5'), 'r')
testresults[model]['Test version'] = filetest.attrs['gprMax']
# Get available field output component names
outputstest = list(filetest[path].keys())
# Arrays for storing time
float_or_double = filetest[path + outputstest[0]].dtype
timetest = np.linspace(0, (filetest.attrs['Iterations'] - 1) * filetest.attrs['dt'],
num=filetest.attrs['Iterations']) / 1e-9
timeref = timetest
# Arrays for storing field data
datatest = np.zeros((filetest.attrs['Iterations'], len(outputstest)),
dtype=float_or_double)
for ID, name in enumerate(outputstest):
datatest[:, ID] = filetest[path + str(name)][:]
if np.any(np.isnan(datatest[:, ID])):
logger.exception('Test data contains NaNs')
raise ValueError
# Tx/Rx position to feed to analytical solution
rxpos = filetest[path].attrs['Position']
txpos = filetest['/srcs/src1/'].attrs['Position']
rxposrelative = ((rxpos[0] - txpos[0]),
(rxpos[1] - txpos[1]),
(rxpos[2] - txpos[2]))
# Analytical solution of a dipole in free space
dataref = hertzian_dipole_fs(filetest.attrs['Iterations'],
filetest.attrs['dt'],
filetest.attrs['dx_dy_dz'], rxposrelative)
filetest.close()
else:
# Get output for model and reference files
fileref = file.stem + '_ref'
fileref = file.parent / Path(fileref)
fileref = h5py.File(fileref.with_suffix('.h5'), 'r')
filetest = h5py.File(file.with_suffix('.h5'), 'r')
testresults[model]['Ref version'] = fileref.attrs['gprMax']
testresults[model]['Test version'] = filetest.attrs['gprMax']
# Get available field output component names
outputsref = list(fileref[path].keys())
outputstest = list(filetest[path].keys())
if outputsref != outputstest:
logger.exception('Field output components do not match reference solution')
raise ValueError
# Check that type of float used to store fields matches
if filetest[path + outputstest[0]].dtype != fileref[path + outputsref[0]].dtype:
logger.warning(f'Type of floating point number in test model ' +
f'({filetest[path + outputstest[0]].dtype}) does not ' +
f'match type in reference solution ({fileref[path + outputsref[0]].dtype})\n')
float_or_doubleref = fileref[path + outputsref[0]].dtype
float_or_doubletest = filetest[path + outputstest[0]].dtype
# Arrays for storing time
timeref = np.zeros((fileref.attrs['Iterations']), dtype=float_or_doubleref)
timeref = np.linspace(0, (fileref.attrs['Iterations'] - 1) * fileref.attrs['dt'],
num=fileref.attrs['Iterations']) / 1e-9
timetest = np.zeros((filetest.attrs['Iterations']), dtype=float_or_doubletest)
timetest = np.linspace(0, (filetest.attrs['Iterations'] - 1) * filetest.attrs['dt'],
num=filetest.attrs['Iterations']) / 1e-9
# Arrays for storing field data
dataref = np.zeros((fileref.attrs['Iterations'], len(outputsref)),
dtype=float_or_doubleref)
datatest = np.zeros((filetest.attrs['Iterations'], len(outputstest)),
dtype=float_or_doubletest)
for ID, name in enumerate(outputsref):
dataref[:, ID] = fileref[path + str(name)][:]
datatest[:, ID] = filetest[path + str(name)][:]
if np.any(np.isnan(datatest[:, ID])):
logger.exception('Test data contains NaNs')
raise ValueError
fileref.close()
filetest.close()
# Diffs
datadiffs = np.zeros(datatest.shape, dtype=np.float64)
for i in range(len(outputstest)):
max = np.amax(np.abs(dataref[:, i]))
datadiffs[:, i] = np.divide(np.abs(dataref[:, i] - datatest[:, i]), max,
out=np.zeros_like(dataref[:, i]),
where=max != 0) # Replace any division by zero with zero
# Calculate power (ignore warning from taking a log of any zero values)
with np.errstate(divide='ignore'):
datadiffs[:, i] = 20 * np.log10(datadiffs[:, i])
# Replace any NaNs or Infs from zero division
datadiffs[:, i][np.invert(np.isfinite(datadiffs[:, i]))] = 0
# Store max difference
maxdiff = np.amax(np.amax(datadiffs))
testresults[model]['Max diff'] = maxdiff
# Plot datasets
fig1, ((ex1, hx1), (ey1, hy1), (ez1, hz1)) = plt.subplots(nrows=3, ncols=2,
sharex=False, sharey='col',
subplot_kw=dict(xlabel='Time [ns]'),
num=model + '.in',
figsize=(20, 10),
facecolor='w',
edgecolor='w')
ex1.plot(timetest, datatest[:, 0], 'r', lw=2, label=model)
ex1.plot(timeref, dataref[:, 0], 'g', lw=2, ls='--', label=model + '(Ref)')
ey1.plot(timetest, datatest[:, 1], 'r', lw=2, label=model)
ey1.plot(timeref, dataref[:, 1], 'g', lw=2, ls='--', label=model + '(Ref)')
ez1.plot(timetest, datatest[:, 2], 'r', lw=2, label=model)
ez1.plot(timeref, dataref[:, 2], 'g', lw=2, ls='--', label=model + '(Ref)')
hx1.plot(timetest, datatest[:, 3], 'r', lw=2, label=model)
hx1.plot(timeref, dataref[:, 3], 'g', lw=2, ls='--', label=model + '(Ref)')
hy1.plot(timetest, datatest[:, 4], 'r', lw=2, label=model)
hy1.plot(timeref, dataref[:, 4], 'g', lw=2, ls='--', label=model + '(Ref)')
hz1.plot(timetest, datatest[:, 5], 'r', lw=2, label=model)
hz1.plot(timeref, dataref[:, 5], 'g', lw=2, ls='--', label=model + '(Ref)')
ylabels = ['$E_x$, field strength [V/m]', '$H_x$, field strength [A/m]',
'$E_y$, field strength [V/m]', '$H_y$, field strength [A/m]',
'$E_z$, field strength [V/m]', '$H_z$, field strength [A/m]']
for i, ax in enumerate(fig1.axes):
ax.set_ylabel(ylabels[i])
ax.set_xlim(0, np.amax(timetest))
ax.grid()
ax.legend()
# Plot diffs
fig2, ((ex2, hx2), (ey2, hy2), (ez2, hz2)) = plt.subplots(nrows=3, ncols=2,
sharex=False, sharey='col',
subplot_kw=dict(xlabel='Time [ns]'),
num='Diffs: ' + model + '.in',
figsize=(20, 10),
facecolor='w',
edgecolor='w')
ex2.plot(timeref, datadiffs[:, 0], 'r', lw=2, label='Ex')
ey2.plot(timeref, datadiffs[:, 1], 'r', lw=2, label='Ey')
ez2.plot(timeref, datadiffs[:, 2], 'r', lw=2, label='Ez')
hx2.plot(timeref, datadiffs[:, 3], 'r', lw=2, label='Hx')
hy2.plot(timeref, datadiffs[:, 4], 'r', lw=2, label='Hy')
hz2.plot(timeref, datadiffs[:, 5], 'r', lw=2, label='Hz')
ylabels = ['$E_x$, difference [dB]', '$H_x$, difference [dB]',
'$E_y$, difference [dB]', '$H_y$, difference [dB]',
'$E_z$, difference [dB]', '$H_z$, difference [dB]']
for i, ax in enumerate(fig2.axes):
ax.set_ylabel(ylabels[i])
ax.set_xlim(0, np.amax(timetest))
ax.set_ylim([plotmin, np.amax(np.amax(datadiffs))])
ax.grid()
# Save a PDF/PNG of the figure
filediffs = file.stem + '_diffs'
filediffs = file.parent / Path(filediffs)
# fig1.savefig(file.with_suffix('.pdf'), dpi=None, format='pdf',
# bbox_inches='tight', pad_inches=0.1)
# fig2.savefig(savediffs.with_suffix('.pdf'), dpi=None, format='pdf',
# bbox_inches='tight', pad_inches=0.1)
fig1.savefig(file.with_suffix('.png'), dpi=150, format='png',
bbox_inches='tight', pad_inches=0.1)
fig2.savefig(filediffs.with_suffix('.png'), dpi=150, format='png',
bbox_inches='tight', pad_inches=0.1)
# Summary of results
for name, data in sorted(testresults.items()):
if 'analytical' in name:
logger.info(f"Test '{name}.in' using v.{data['Test version']} compared " +
f"to analytical solution. Max difference {data['Max diff']:.2f}dB.")
else:
logger.info(f"Test '{name}.in' using v.{data['Test version']} compared to " +
f"reference solution using v.{data['Ref version']}. Max difference " +
f"{data['Max diff']:.2f}dB.")