def read_dataset(input_file_path, target_file_path, read_dataset_info):
"""Read dataset from pickle files and preprocess it.
Parameters
----------
input_file_path : str, os.PathLike or pathlib.Path
The path of pickle file.
target_file_path : str, os.PathLike or pathlib.Path
The path of pickle file.
read_dataset_info : dict
Returns
-------
resistance : numpy.ndarray
The input data of the neural network.
resistivity_log10 : numpy.ndarray
The target data of the neural network.
"""
# read data and assign
# data = read_pkl(file_path.numpy().decode('utf-8'))
# resistance = data['resistance']
# resistivity_log10 = data['resistivity_log10']
data = read_pkl(input_file_path.numpy().decode('utf-8'))
resistance = data
data = read_pkl(target_file_path.numpy().decode('utf-8'))
resistivity_log10 = data
# parse read_dataset_info dictionary
preprocess = read_dataset_info['preprocess']
Tx_locations = read_dataset_info['Tx_locations']
Rx_locations = read_dataset_info['Rx_locations']
nCx = read_dataset_info['nCx']
nCy = read_dataset_info['nCy']
# preprocess
for k, v in preprocess.items():
if k == 'add_noise' and v.get('perform'):
add_noise(resistance, **v.get('kwargs'))
elif k == 'log_transform' and v.get('perform'):
log_transform(resistance, **v.get('kwargs'))
elif k == 'to_midpoint' and v.get('perform'):
resistance = to_midpoint(resistance, Tx_locations, Rx_locations)
elif k == 'to_txrx' and v.get('perform'):
resistance = to_txrx(resistance, Tx_locations, Rx_locations)
elif k == 'to_section' and v.get('perform'):
resistivity_log10 = to_section(resistivity_log10, nCx, nCy)
return resistance, resistivity_log10
def get_data(self, temp_file_list):
resistance = np.empty((len(temp_file_list), *self.input_shape))
for i, file in enumerate(temp_file_list):
data = read_pkl(file)
if self.preprocess['to_midpoint']['perform']:
resistance[i, ] = to_midpoint(data['resistance'],
self.Tx_locations,
self.Rx_locations)
elif self.preprocess['to_txrx']['perform']:
resistance[i, ] = to_txrx(data['resistance'],
self.Tx_locations, self.Rx_locations)
else:
resistance[i, ] = data['resistance'].reshape(self.input_shape)
for k, v in self.preprocess.items():
if k == 'add_noise' and v.get('perform'):
add_noise(resistance, **v.get('kwargs'))
elif k == 'log_transform' and v.get('perform'):
log_transform(resistance, **v.get('kwargs'))
return resistance
def plot_data(iterator, simulator, num_figs):
SRCLOC = simulator.urf.abmn_locations[:, :4]
RECLOC = simulator.urf.abmn_locations[:, 4:]
active_idx = simulator.active_idx
nCx = simulator.mesh.nCx
nCy = simulator.mesh.nCy
vectorCCx = simulator.mesh.vectorCCx
vectorCCy = simulator.mesh.vectorCCy
num_figs = 1 if num_figs < 1 else num_figs
i = 1
for file in iterator:
data = read_pkl(file.path)
print(data['resistance'].shape,
data['resistivity_log10'].shape)
resistance = data['resistance']
resistivity = data['resistivity_log10']
# plot resistance
# txrx version
fig, ax = plt.subplots(figsize=(16, 9))
im = ax.imshow(
to_txrx(
resistance,
SRCLOC,
RECLOC,
value=np.nan
)[:, :, 0],
origin='lower'
)
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = fig.colorbar(im, cax=cax)
ax.set_xlabel('Rx_pair')
ax.set_ylabel('Tx_pair')
cbar.set_label(r'$\Delta V/I$')
# midpoint version
fig, ax = plt.subplots(figsize=(4, 3))
im = ax.imshow(
to_midpoint(
resistance,
SRCLOC,
RECLOC,
value=np.nan
)[:, :, 0]
)
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = fig.colorbar(im, cax=cax)
ax.set_xlabel('common midpoint')
ax.set_ylabel('count')
cbar.set_label(r'$\Delta V/I$')
ax.set_aspect('auto', adjustable='box')
# plot resistivity
# imshow version
fig, ax = plt.subplots()
im = simulator.mesh.plotImage(resistivity[active_idx], ax=ax)
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = fig.colorbar(im[0], cax=cax)
ax.set_xlabel('m')
ax.set_ylabel('m')
cbar.set_label(r'$\Omega \bullet m (log_{10})$')
# contourf version
fig, ax = plt.subplots()
simulator.mesh.plotImage(resistivity[active_idx], ax=ax)
im = ax.contourf(vectorCCx, vectorCCy,
resistivity[active_idx].reshape((nCy, nCx)))
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="5%", pad=0.05)
cbar = fig.colorbar(im, cax=cax)
ax.set_xlabel('m')
ax.set_ylabel('m')
cbar.set_label(r'$\Omega \bullet m (log_{10})$')
plt.show()
if i == num_figs:
break
else:
i += 1