column_to_dtype={
'file_path_raw_signals':
str,
'sampling_frequency_in_hz':
(int, float, np.int, np.float, np.int64,
np.float64),
})
signal_labels = [
dataset[column_name]
for column_name in state_annotation_signals
]
raw_signals = load_raw_signals(dataset['file_path_raw_signals'],
signal_labels)
plot_raw_signals(
raw_signals,
sampling_frequency=dataset['sampling_frequency_in_hz'],
ax=axes[0],
)
axes[0].set_ylabel("Raw signals")
# LDA tranformed signals
transformed_signals = annotator.transform(signal_arrays[ii])
plot_signals(transformed_signals, ax=axes[1])
axes[1].set_ylabel("Transformed signals")
predicted_state_vector = annotator.predict(signal_arrays[ii])
predicted_states, predicted_intervals = convert_state_vector_to_state_intervals(
predicted_state_vector, mapping=int_to_state)
plot_states(predicted_states, predicted_intervals, ax=axes[2])
axes[2].set_ylabel("Automated\nannotation")
psd_max = np.max(psd)
ax.set_ylim(0, psd_max)
psd_figure.canvas.draw_idle()
# initialise state annotation figure
fig = plt.figure()
gs = GridSpec(5, 1)
data_axis = fig.add_subplot(gs[:3, 0])
state_axis = fig.add_subplot(gs[3, 0], sharex=data_axis)
manual_state_axis = fig.add_subplot(gs[4, 0], sharex=data_axis)
fig.tight_layout(**{'rect': [0.05, 0, 1, 1], 'pad': 2., 'h_pad': 0.})
# plot signals
plot_raw_signals(
raw_signals,
sampling_frequency=dataset['sampling_frequency_in_hz'],
ax=data_axis,
linewidth=1.,
)
# plot manual annotations
plot_states(
manual_states,
manual_intervals,
ax=manual_state_axis,
linewidth=5.,
)
manual_state_axis.set_ylabel("Manual annotation")
# initialise annotator
annotator = TimeSeriesStateViewer(
data_axis,