def clean_both_sets(self):
# Clean by directy changing the cnt variables...
clean_train_cnt, clean_test_cnt = clean_train_test_cnt(
self.cnt, self.test_cnt, self.cleaner, self.test_cleaner)
self.cnt = clean_train_cnt
self.test_cnt = clean_test_cnt
self.rejected_chan_names = [] # necessary since will be used
# in preprocess set...
self.rejected_trials = 'unknown'
self.clean_trials = 'unknown'
def clean_both_sets(self):
# Clean by directy changing the cnt variables...
clean_train_cnt, clean_test_cnt = clean_train_test_cnt(
self.cnt, self.test_cnt, self.cleaner,
self.test_cleaner)
self.cnt = clean_train_cnt
self.test_cnt = clean_test_cnt
self.rejected_chan_names = [] # necessary since will be used
# in preprocess set...
self.rejected_trials = 'unknown'
self.clean_trials = 'unknown'
def load(self):
# Loading both sets, cleaning cnts and finished
log.info("Load Training Set...")
self.train_set.signal_processor.load_signal_and_markers()
log.info("Load Test Set...")
self.test_set.signal_processor.load_signal_and_markers()
train_cnt = self.train_set.signal_processor.cnt
test_cnt = self.test_set.signal_processor.cnt
clean_train_cnt, clean_test_cnt = clean_train_test_cnt(train_cnt,
test_cnt,self.train_cleaner, self.test_cleaner)
self.train_set.signal_processor.cnt = clean_train_cnt
self.test_set.signal_processor.cnt = clean_test_cnt
assert np.array_equal(self.train_set.signal_processor.cnt.axes[1],
self.train_set.signal_processor.cnt.axes[1]), ("Sensor names should "
"be the same for train and test...")
log.info("Create sets from cleaned cnt...")
# in case of cnt signal matrix:
if isinstance(self.train_set, CntSignalMatrix):
self.train_set.load_from_cnt()
self.test_set.load_from_cnt()
# in case of raw set:
elif isinstance(self.train_set, CleanSignalMatrix):
for one_set in [self.train_set, self.test_set]:
# this is very fragile... as changes to
# the original class logic will affect this :(
one_set.load_from_cnt()
one_set.create_dense_design_matrix()
one_set.remove_signal_epo()
if one_set.unsupervised_preprocessor is not None:
one_set.apply_unsupervised_preprocessor()
one_set.y = np.argmax(one_set.y, axis=1).astype(np.int32)
else:
raise ValueError("Unknown type of train set {:s}".format(
self.train_set.__class__.__name__))
log.info("Loaded clean train data with shape {:s}.".format(
self.train_set.get_topological_view().shape))
log.info("Loaded clean test data with shape {:s}.".format(
self.test_set.get_topological_view().shape))
self.sets = [self.train_set, self.test_set]
self.y = self.sets[-1].y[0:1]
def load(self):
# Loading both sets, cleaning cnts and finished
log.info("Load Training Set...")
self.train_set.signal_processor.load_signal_and_markers()
log.info("Load Test Set...")
self.test_set.signal_processor.load_signal_and_markers()
train_cnt = self.train_set.signal_processor.cnt
test_cnt = self.test_set.signal_processor.cnt
clean_train_cnt, clean_test_cnt = clean_train_test_cnt(train_cnt,
test_cnt,self.train_cleaner, self.test_cleaner)
self.train_set.signal_processor.cnt = clean_train_cnt
self.test_set.signal_processor.cnt = clean_test_cnt
assert np.array_equal(self.train_set.signal_processor.cnt.axes[1],
self.train_set.signal_processor.cnt.axes[1]), ("Sensor names should "
"be the same for train and test...")
log.info("Create sets from cleaned cnt...")
# in case of cnt signal matrix:
if isinstance(self.train_set, CntSignalMatrix):
self.train_set.load_from_cnt()
self.test_set.load_from_cnt()
# in case of raw set:
elif isinstance(self.train_set, CleanSignalMatrix):
for one_set in [self.train_set, self.test_set]:
# this is very fragile... as changes to
# the original class logic will affect this :(
one_set.load_from_cnt()
one_set.create_dense_design_matrix()
one_set.remove_signal_epo()
if one_set.unsupervised_preprocessor is not None:
one_set.apply_unsupervised_preprocessor()
one_set.y = np.argmax(one_set.y, axis=1).astype(np.int32)
else:
raise ValueError("Unknown type of train set {:s}".format(
self.train_set.__class__.__name__))
log.info("Loaded clean train data with shape {:s}.".format(
self.train_set.get_topological_view().shape))
log.info("Loaded clean test data with shape {:s}.".format(
self.test_set.get_topological_view().shape))
self.sets = [self.train_set, self.test_set]
self.y = self.sets[-1].y[0:1]
