def test_thresh_isolation_distances():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
K=10,
seed=0)
s_threshold = 200
iso = compute_isolation_distances(sort, rec, apply_filter=False, seed=0)
sort_iso = threshold_isolation_distances(sort,
rec,
s_threshold,
'less',
apply_filter=False,
seed=0)
original_ids = sort.get_unit_ids()
new_iso = []
for unit in sort_iso.get_unit_ids():
new_iso.append(iso[original_ids.index(unit)])
new_iso = np.array(new_iso)
assert np.all(new_iso >= s_threshold)
check_dumping(sort_iso)
shutil.rmtree('test')
def test_thresh_l_ratios():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
K=10,
seed=0)
l_ratios_thresh = 0
l_ratios = compute_l_ratios(sort, rec, apply_filter=False, seed=0)
sort_l_ratios = threshold_l_ratios(sort,
rec,
l_ratios_thresh,
"less",
apply_filter=False,
seed=0)
original_ids = sort.get_unit_ids()
new_l_ratios = []
for unit in sort_l_ratios.get_unit_ids():
new_l_ratios.append(l_ratios[original_ids.index(unit)])
new_l_ratios = np.array(new_l_ratios)
assert np.all(new_l_ratios >= l_ratios_thresh)
check_dumping(sort_l_ratios)
shutil.rmtree('test')
def test_remove_artifacts():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
seed=0)
triggers = [30000, 90000]
ms = 10
ms_frames = int(ms * rec.get_sampling_frequency() / 1000)
rec_rmart = remove_artifacts(rec, triggers, ms_before=10, ms_after=10)
traces_all_0 = rec_rmart.get_traces(start_frame=triggers[0] - ms_frames,
end_frame=triggers[0] + ms_frames)
traces_short_0 = rec_rmart.get_traces(start_frame=triggers[0] - 10,
end_frame=triggers[0] + 10)
traces_all_1 = rec_rmart.get_traces(start_frame=triggers[1] - ms_frames,
end_frame=triggers[1] + ms_frames)
traces_short_1 = rec_rmart.get_traces(start_frame=triggers[1] - 10,
end_frame=triggers[1] + 10)
assert not np.any(traces_all_0)
assert not np.any(traces_all_1)
assert not np.any(traces_short_0)
assert not np.any(traces_short_1)
check_dumping(rec_rmart)
shutil.rmtree('test')
def test_thresh_nn_metrics():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
s_threshold_hit = 0.9
s_threshold_miss = 0.002
nn_hit, nn_miss = compute_nn_metrics(sort, rec, apply_filter=False, seed=0)
sort_hit = threshold_nn_metrics(sort, rec, s_threshold_hit, 'less', metric_name="nn_hit_rate",
apply_filter=False, seed=0)
sort_miss = threshold_nn_metrics(sort, rec, s_threshold_miss, 'greater', metric_name="nn_miss_rate",
apply_filter=False, seed=0)
original_ids = sort.get_unit_ids()
new_nn_hit = []
for unit in sort_hit.get_unit_ids():
new_nn_hit.append(nn_hit[original_ids.index(unit)])
new_nn_miss = []
for unit in sort_miss.get_unit_ids():
new_nn_miss.append(nn_miss[original_ids.index(unit)])
new_nn_hit = np.array(new_nn_hit)
new_nn_miss = np.array(new_nn_miss)
assert np.all(new_nn_hit >= s_threshold_hit)
assert np.all(new_nn_miss <= s_threshold_miss)
check_dumping(sort_hit)
check_dumping(sort_miss)
shutil.rmtree('test')
def test_transform():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
seed=0)
scalar = 3
offset = 50
rec_t = transform(rec, scalar=scalar, offset=offset)
assert np.allclose(rec_t.get_traces(),
scalar * rec.get_traces() + offset,
atol=0.001)
scalars = np.random.randn(4)
offsets = np.random.randn(4)
rec_t_arr = transform(rec, scalar=scalars, offset=offsets)
for (tt, to, s, o) in zip(rec_t_arr.get_traces(), rec.get_traces(),
scalars, offsets):
assert np.allclose(tt, s * to + o, atol=0.001)
check_dumping(rec_t)
check_dumping(rec_t_arr)
shutil.rmtree('test')
def test_thresh_silhouettes():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
K=10,
seed=0)
silhouette_thresh = .5
silhouette = compute_silhouette_scores(sort,
rec,
apply_filter=False,
seed=0)
sort_silhouette = threshold_silhouette_scores(sort,
rec,
silhouette_thresh,
"less",
apply_filter=False,
seed=0)
original_ids = sort.get_unit_ids()
new_silhouette = []
for unit in sort_silhouette.get_unit_ids():
new_silhouette.append(silhouette[original_ids.index(unit)])
new_silhouette = np.array(new_silhouette)
assert np.all(new_silhouette >= silhouette_thresh)
check_dumping(sort_silhouette)
shutil.rmtree('test')
def test_thresh_frs():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
fr_thresh = 2
sort_fr = threshold_firing_rates(sort, fr_thresh, 'less', rec.get_num_frames())
new_fr = compute_firing_rates(sort_fr, rec.get_num_frames())
assert np.all(new_fr >= fr_thresh)
check_dumping(sort_fr)
shutil.rmtree('test')
def test_rectify():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
rec_rect = rectify(rec)
assert np.allclose(rec_rect.get_traces(), np.abs(rec.get_traces()))
check_dumping(rec_rect)
shutil.rmtree('test')
def test_thresh_num_spikes():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
s_threshold = 25
sort_ns = threshold_num_spikes(sort, s_threshold, 'less')
new_ns = compute_num_spikes(sort_ns, sort.get_sampling_frequency())
assert np.all(new_ns >= s_threshold)
check_dumping(sort_ns)
shutil.rmtree('test')
def test_thresh_isi_violations():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
s_threshold = 0.01
sort_isi = threshold_isi_violations(sort, s_threshold, 'greater', rec.get_num_frames())
new_isi = compute_isi_violations(sort_isi, rec.get_num_frames(), sort.get_sampling_frequency())
assert np.all(new_isi <= s_threshold)
check_dumping(sort_isi)
shutil.rmtree('test')
def test_thresh_presence_ratios():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
s_threshold = 0.18
sort_pr = threshold_presence_ratios(sort, s_threshold, 'less', rec.get_num_frames())
new_pr = compute_presence_ratios(sort_pr, rec.get_num_frames(), sort.get_sampling_frequency())
assert np.all(new_pr >= s_threshold)
check_dumping(sort_pr)
shutil.rmtree('test')
def test_curation_sorting_extractor():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
K=3,
seed=0)
# Dummy features for testing merging and splitting of features
sort.set_unit_spike_features(
1, 'f_int', range(0 + 1,
len(sort.get_unit_spike_train(1)) + 1))
sort.set_unit_spike_features(2, 'f_int',
range(0, len(sort.get_unit_spike_train(2))))
sort.set_unit_spike_features(
2, 'bad_features', np.repeat(1, len(sort.get_unit_spike_train(2))))
sort.set_unit_spike_features(3, 'f_int',
range(0, len(sort.get_unit_spike_train(3))))
CSX = st.curation.CurationSortingExtractor(parent_sorting=sort)
merged_unit_id = CSX.merge_units(unit_ids=[1, 2])
assert np.allclose(merged_unit_id, 4)
original_spike_train = np.concatenate(
(sort.get_unit_spike_train(1), sort.get_unit_spike_train(2)))
indices_sort = np.argsort(original_spike_train)
original_spike_train = original_spike_train[indices_sort]
original_features = np.concatenate(
(sort.get_unit_spike_features(1, 'f_int'),
sort.get_unit_spike_features(2, 'f_int')))
original_features = original_features[indices_sort]
assert np.allclose(CSX.get_unit_spike_train(4), original_spike_train)
assert np.allclose(CSX.get_unit_spike_features(4, 'f_int'),
original_features)
assert CSX.get_unit_spike_feature_names(4) == ['f_int']
assert np.allclose(CSX.get_sampling_frequency(),
sort.get_sampling_frequency())
unit_ids_split = CSX.split_unit(unit_id=3,
indices=[0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
assert np.allclose(unit_ids_split[0], 5)
assert np.allclose(unit_ids_split[1], 6)
original_spike_train = sort.get_unit_spike_train(3)
original_features = sort.get_unit_spike_features(3, 'f_int')
split_spike_train_1 = CSX.get_unit_spike_train(5)
split_spike_train_2 = CSX.get_unit_spike_train(6)
split_features_1 = CSX.get_unit_spike_features(5, 'f_int')
split_features_2 = CSX.get_unit_spike_features(6, 'f_int')
assert np.allclose(original_spike_train[:10], split_spike_train_1)
assert np.allclose(original_spike_train[10:], split_spike_train_2)
assert np.allclose(original_features[:10], split_features_1)
assert np.allclose(original_features[10:], split_features_2)
check_dumping(CSX)
def test_thresh_snrs():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
snr_thresh = 4
sort_snr = threshold_snrs(sort, rec, snr_thresh, 'less', apply_filter=False, seed=0)
new_snr = compute_snrs(sort_snr, rec, apply_filter=False, seed=0)
assert np.all(new_snr >= snr_thresh)
check_dumping(sort_snr)
shutil.rmtree('test')
def test_rectify():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
seed=0)
rec_rect = rectify(rec)
assert np.allclose(rec_rect.get_traces(), np.abs(rec.get_traces()))
check_dumping(rec_rect)
shutil.rmtree('test')
def test_blank_saturation():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
threshold = 2
rec_bs = blank_saturation(rec, threshold=threshold)
index_below_threshold = np.where(rec.get_traces() < threshold)
assert np.all(rec_bs.get_traces()[index_below_threshold] < threshold)
check_dumping(rec_bs)
shutil.rmtree('test')
def test_thresh_amplitude_cutoffs():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
amplitude_cutoff_thresh = 0
sort_amplitude_cutoff = threshold_amplitude_cutoffs(sort, rec, amplitude_cutoff_thresh, "less",
apply_filter=False, seed=0)
new_amplitude_cutoff = compute_amplitude_cutoffs(sort_amplitude_cutoff, rec, apply_filter=False, seed=0)
assert np.all(new_amplitude_cutoff >= amplitude_cutoff_thresh)
check_dumping(sort_amplitude_cutoff)
shutil.rmtree('test')
def test_center():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
rec_c = center(rec, mode='mean')
assert np.allclose(np.mean(rec_c.get_traces(), axis=1), 0, atol=0.001)
check_dumping(rec_c)
rec_c = center(rec, mode='median')
assert np.allclose(np.median(rec_c.get_traces(), axis=1), 0, atol=0.001)
check_dumping(rec_c)
shutil.rmtree('test')
def test_thresh_presence_ratios():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
K=10,
seed=0,
folder='test')
s_threshold = 0.18
sort_pr = threshold_presence_ratios(sort, s_threshold, 'less')
new_pr = compute_presence_ratios(sort_pr, sort.get_sampling_frequency())[0]
assert np.all(new_pr >= s_threshold)
check_dumping(sort_pr)
shutil.rmtree('test')
def test_thresh_frs():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
K=10,
seed=0,
folder='test')
fr_thresh = 2
sort_fr = threshold_firing_rates(sort, fr_thresh, 'less')
new_fr = compute_firing_rates(sort_fr)[0]
assert np.all(new_fr >= fr_thresh)
check_dumping(sort_fr)
shutil.rmtree('test')
def test_thresh_isi_violations():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
K=10,
seed=0,
folder='test')
s_threshold = 0.01
sort_isi = threshold_isi_violations(sort, s_threshold, 'greater')
new_isi = compute_isi_violations(sort_isi,
sort.get_sampling_frequency())[0]
assert np.all(new_isi <= s_threshold)
check_dumping(sort_isi)
shutil.rmtree('test')
def test_clip():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
threshold = 5
rec_clip = clip(rec, a_min=-threshold, a_max=threshold)
index_below_threshold = np.where(rec.get_traces() < -threshold)
index_above_threshold = np.where(rec.get_traces() > threshold)
assert np.all(rec_clip.get_traces()[index_below_threshold] == -threshold)
assert np.all(rec_clip.get_traces()[index_above_threshold] == threshold)
check_dumping(rec_clip)
shutil.rmtree('test')
def test_notch_filter():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
rec_n = notch_filter(rec, 3000, q=10)
assert check_signal_power_signal1_below_signal2(
rec_n.get_traces(),
rec.get_traces(),
freq_range=[2900, 3100],
fs=rec.get_sampling_frequency())
check_dumping(rec_n)
shutil.rmtree('test')
def test_blank_saturation():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=2,
num_channels=4,
seed=0)
threshold = 2
rec_bs = blank_saturation(rec, threshold=threshold)
index_below_threshold = np.where(rec.get_traces() < threshold)
assert np.all(rec_bs.get_traces()[index_below_threshold] < threshold)
check_dumping(rec_bs)
shutil.rmtree('test')
def test_center():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=2,
num_channels=4,
seed=0)
rec_c = center(rec, mode='mean')
assert np.allclose(np.mean(rec_c.get_traces(), axis=1), 0, atol=0.001)
check_dumping(rec_c)
rec_c = center(rec, mode='median')
assert np.allclose(np.median(rec_c.get_traces(), axis=1), 0, atol=0.001)
check_dumping(rec_c)
shutil.rmtree('test')
def test_resample():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test')
resample_rate_low = 0.1 * rec.get_sampling_frequency()
resample_rate_high = 2 * rec.get_sampling_frequency()
rec_rsl = resample(rec, resample_rate_low)
rec_rsh = resample(rec, resample_rate_high)
assert rec_rsl.get_num_frames() == int(rec.get_num_frames() * 0.1)
assert rec_rsh.get_num_frames() == int(rec.get_num_frames() * 2)
check_dumping(rec_rsl)
check_dumping(rec_rsh)
shutil.rmtree('test')
def test_whiten():
rec, sort = se.example_datasets.create_dumpable_extractors(duration=10,
num_channels=4,
folder='test',
seed=0)
rec_w = whiten(rec)
cov_w = np.cov(rec_w.get_traces())
assert np.allclose(cov_w, np.eye(4), atol=0.3)
# should size should not affect
rec_w2 = whiten(rec, chunk_size=30000)
assert np.array_equal(rec_w.get_traces(), rec_w2.get_traces())
check_dumping(rec_w)
shutil.rmtree('test')
def test_thresh_threshold_drift_metrics():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
s_threshold = 1
sort_max = threshold_drift_metrics(sort, rec, s_threshold, 'greater', metric_name="max_drift",
apply_filter=False, seed=0)
sort_cum = threshold_drift_metrics(sort, rec, s_threshold, 'greater', metric_name="cumulative_drift",
apply_filter=False, seed=0)
new_max_drift, _ = compute_drift_metrics(sort_max, rec, apply_filter=False, seed=0)
_, new_cum_drift = compute_drift_metrics(sort_cum, rec, apply_filter=False, seed=0)
assert np.all(new_max_drift <= s_threshold)
assert np.all(new_cum_drift <= s_threshold)
check_dumping(sort_max)
check_dumping(sort_cum)
shutil.rmtree('test')
def test_thresh_noise_overlaps():
rec, sort = se.example_datasets.toy_example(dump_folder='test', dumpable=True, duration=10, num_channels=4, K=10,
seed=0)
noise_thresh = 0.3
noise_overlaps = compute_noise_overlaps(sort, rec, apply_filter=False, seed=0)
sort_noise = threshold_noise_overlaps(sort, rec, noise_thresh, 'less', apply_filter=False, seed=0)
original_ids = sort.get_unit_ids()
new_noise = []
for unit in sort_noise.get_unit_ids():
new_noise.append(noise_overlaps[original_ids.index(unit)])
new_noise = np.array(new_noise)
assert np.all(new_noise >= noise_thresh)
check_dumping(sort_noise)
shutil.rmtree('test')
def test_resample():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=2,
num_channels=4,
seed=0)
resample_rate_low = 0.1 * rec.get_sampling_frequency()
resample_rate_high = 2 * rec.get_sampling_frequency()
rec_rsl = resample(rec, resample_rate_low)
rec_rsh = resample(rec, resample_rate_high)
assert rec_rsl.get_num_frames() == int(rec.get_num_frames() * 0.1)
assert rec_rsh.get_num_frames() == int(rec.get_num_frames() * 2)
check_dumping(rec_rsl)
check_dumping(rec_rsh)
shutil.rmtree('test')
def test_bandpass_filter_with_cache():
rec, sort = se.example_datasets.toy_example(dump_folder='test',
dumpable=True,
duration=10,
num_channels=4,
seed=0)
rec_filtered = bandpass_filter(rec,
freq_min=5000,
freq_max=10000,
cache_to_file=True,
chunk_size=10000)
rec_filtered2 = bandpass_filter(rec,
freq_min=5000,
freq_max=10000,
cache_to_file=True,
chunk_size=None)
rec_filtered3 = bandpass_filter(rec,
freq_min=5000,
freq_max=10000,
cache_chunks=True,
chunk_size=10000)
rec_filtered3.get_traces()
assert rec_filtered3._filtered_cache_chunks.get('0') is not None
rec_filtered4 = bandpass_filter(rec,
freq_min=5000,
freq_max=10000,
cache_chunks=True,
chunk_size=None)
assert np.allclose(rec_filtered.get_traces(),
rec_filtered2.get_traces(),
rtol=1e-02,
atol=1e-02)
assert np.allclose(rec_filtered.get_traces(),
rec_filtered3.get_traces(),
rtol=1e-02,
atol=1e-02)
assert np.allclose(rec_filtered.get_traces(),
rec_filtered4.get_traces(),
rtol=1e-02,
atol=1e-02)
check_dumping(rec_filtered)
check_dumping(rec_filtered2)
check_dumping(rec_filtered3)
check_dumping(rec_filtered4)
shutil.rmtree('test')