def applyClassifiers(scst_pkl_path: str,
keras_pkl_path: str,
data_path: Optional[str] = None,
truth_path: Optional[str] = None,
start_second: Optional[float] = None,
end_second: Optional[float] = None) -> None:
'''
Apply both a ``SCSTClassifier`` and a ``TransientKerasClassifier`` to the data sequence
specified by the inputs, and plot the results.
:param scst_pkl_path: Full path to a pickle file containing a saved ``SCSTClassifier`` object.
:param keras_pkl_path: Full path to a pickle file containing a saved
``TransientKerasClassifier`` object.
:param data_path, truth_path, start_second, end_second: See ``data_utils.getPlotDataTuple``.
'''
# Define SCST classification parameters
SIG_THRESH = 100
NOISE_THRESH = 100
# Load data and classifiers
test_matrix, truth_array, start_second, title = getPlotDataTuple(
truth_path, data_path, start_second, end_second)
scst_classifier = SCSTClassifier.load(scst_pkl_path)
keras_classifier = TransientKerasClassifier.load(keras_pkl_path)
# Apply classifiers
num_obs = test_matrix.shape[1]
for test_idx in range(num_obs):
scst_classifier.classify(test_matrix[:, test_idx], SIG_THRESH,
NOISE_THRESH)
keras_classifier.classify(test_matrix[:, test_idx])
_printProgress(test_idx / float(num_obs), title)
_printProgress(1.0, title)
sys.stdout.write("\n")
# Display results
id_array_list = [
scst_classifier.class_labels, keras_classifier.class_labels
]
id_tag_list = ["SCST IDs", "Keras IDs"]
if truth_array is not None:
for classifier, name in zip([scst_classifier, keras_classifier],
["SCST", "Keras"]):
conf_matrix = ConfusionMatrix(classifier.class_labels, truth_array,
True, name)
conf_matrix.display()
id_array_list.append(truth_array)
id_tag_list.append("Truth IDs")
plotSequence(test_matrix, start_second, title, id_array_list, id_tag_list)
plt.show()
