def test_electrode_contingencies_2_subset():
random_seed = np.random.seed(123)
noise = 0
gray = se.Brain(se.load('gray', vox_size=20))
# extract locations
gray_locs = gray.locs.iloc[:5]
mo_locs = gray_locs
c = se.create_cov(cov='random', n_elecs=5)
data = c[:, mo_locs.index][mo_locs.index, :]
model = se.Model(numerator=np.array(data),
denominator=np.ones(np.shape(data)),
locs=mo_locs,
n_subs=1)
# create brain object from the remaining locations - first find remaining locations
sub_locs = mo_locs.sample(2, random_state=random_seed).sort_values(
['x', 'y', 'z'])
# create a brain object with all gray locations
bo = se.simulate_bo(n_samples=5,
sample_rate=1000,
locs=gray_locs,
noise=noise,
random_seed=random_seed)
# parse brain object to create synthetic patient data
data = bo.data.iloc[:, sub_locs.index]
# put data and locations together in new sample brain object
bo_sample = se.Brain(data=data.as_matrix(),
locs=sub_locs,
sample_rate=1000)
# predict activity at all unknown locations
recon = model.predict(bo_sample, nearest_neighbor=False)
actual = bo.data.iloc[:, recon.locs.index]
corr_vals = _corr_column(actual.as_matrix(), recon.data.as_matrix())
#assert np.allclose(zscore(recon_2), recon.data, equal_nan=True)
assert 1 >= corr_vals.mean() >= -1
def test_model_get_slice():
mo = se.Model(data=data[1:3], locs=locs)
inds = [0, 1]
s = mo.get_slice(inds)
assert(type(s) == se.Model)
s_model = s.get_model()
assert s_model.shape[0] == s_model.shape[1]
assert s_model.shape[0] == len(inds)
assert s_model.shape[0] == len(inds)
mo.get_slice(inds, inplace=True)
assert(type(mo) == se.Model)
mo_model = mo.get_model()
assert mo_model.shape[0] == mo_model.shape[1]
assert mo_model.shape[0] == len(inds)
assert mo_model.shape[0] == len(inds)
def test_cpu_vs_gpu_single_subject():
cupy = pytest.importorskip("cupy")
locs = np.random.randn(25, 3)
bo1 = se.simulate_bo(n_samples=512*30, locs=locs, sample_rate=512)
bo2 = se.simulate_bo(n_samples=512*30, locs=locs, sample_rate=512)
mo_cpu = se.Model([bo1, bo2])
mo_gpu = se.Model([bo1, bo2], gpu=True)
assert mo_cpu.locs.equals(mo_gpu.locs)
assert np.allclose(mo_cpu.get_model(), mo_gpu.get_model())
mo_cpu = se.Model(bo1) + se.Model(bo2)
mo_gpu = se.Model(bo1, gpu=True) + se.Model(bo2, gpu=True)
assert mo_cpu.locs.equals(mo_gpu.locs)
assert np.allclose(mo_cpu.get_model(), mo_gpu.get_model())
def test_cpu_vs_gpu_mult_subject():
cupy = pytest.importorskip("cupy")
locs1 = np.random.randn(25, 3)
bo1a = se.simulate_bo(n_samples=512*30, locs=locs1, sample_rate=512)
bo1b = se.simulate_bo(n_samples=512*30, locs=locs1, sample_rate=512)
locs2 = np.random.randn(25, 3)
bo2a = se.simulate_bo(n_samples=512*30, locs=locs2, sample_rate=512)
bo2b = se.simulate_bo(n_samples=512*30, locs=locs2, sample_rate=512)
mo_cpu = se.Model([bo1a, bo1b, bo2a, bo2b])
mo_gpu = se.Model([bo1a, bo1b, bo2a, bo2b], gpu=True)
assert mo_cpu.locs.equals(mo_gpu.locs)
assert np.allclose(mo_cpu.get_model(), mo_gpu.get_model())
mo_cpu = se.Model([bo1a, bo1b]) + se.Model([bo2a, bo2b])
mo_gpu = se.Model([bo1a, bo1b], gpu=True) + se.Model([bo2a, bo2b], gpu=True)
assert mo_cpu.locs.equals(mo_gpu.locs)
assert np.allclose(mo_cpu.get_model(), mo_gpu.get_model())
[-21., 63., -3.],
[ -1., -37., 37.],
[ -1., 23., 17.],
[ 19., -57., -23.],
[ 19., 23., -3.],
[ 39., -57., 17.],
[ 39., 3., 37.],
[ 59., -17., 17.]])
n_samples = 10
n_subs = 3
n_elecs = 10
data = [se.simulate_model_bos(n_samples=10, sample_rate=10, locs=locs,
sample_locs = n_elecs) for x in range(n_subs)]
test_bo = data[0]
test_model = se.Model(data=data, locs=locs, rbf_width=20)
bo = se.load('example_data')
def test_load_example_data():
bo = se.load('example_data')
assert isinstance(bo, se.Brain)
def test_load_example_filter():
bo = se.load('example_filter')
assert isinstance(bo, se.Brain)
def test_load_example_model():
model = se.load('example_model')
assert isinstance(model, se.Model)
def test_load_nifti():
# simulate 100 locations
locs = se.simulate_locations(n_elecs=100)
# simulate brain object
bo = se.simulate_bo(n_samples=1000,
sample_rate=1000,
cov='toeplitz',
locs=locs,
noise=.3)
# sample 10 locations, and get indices
sub_locs = locs.sample(10, replace=False).sort_values(['x', 'y', 'z'])
R = se.create_cov(cov='random', n_elecs=len(sub_locs))
toe_model = se.Model(data=R, locs=sub_locs)
bo_s = toe_model.predict(bo, nearest_neighbor=False)
# sample 10 locations, and get indices
sub_locs = locs.sample(10).sort_values(['x', 'y', 'z']).index.values.tolist()
# index brain object to get sample patient
bo_sample = bo[:, sub_locs]
# plot sample patient locations
bo_sample.plot_locs()
# plot sample patient data
bo_sample.plot_data()
total_patients = len(files)
divided_patients = int(total_patients / 2)
all_idx = np.arange(total_patients)
n_samples = total_patients - divided_patients
corrs_rand = np.zeros((rand_iters, n_samples))
for r in np.arange(rand_iters):
half_1_idx = np.random.choice(total_patients, replace=False, size=divided_patients)
half_2_idx = np.delete(all_idx, half_1_idx)
half_1_mo = se.Model(list(np.array(files)[half_1_idx]), n_subs=len(half_1_idx))
#np.random.shuffle(half_2_idx)
for e, i in enumerate(half_2_idx):
if e == 0:
mo_2 = se.Model(np.array(files)[half_2_idx[e]], n_subs=1)
else:
mo_i = se.Model(np.array(files)[half_2_idx[e]], n_subs=1)
mo_2 = mo_2 + mo_i
noise=.1)
# sample 10 locations, and get indices
sub_locs = locs.sample(90,
replace=False).sort_values(['x', 'y',
'z']).index.values.tolist()
# index brain object to get sample patient
bo_sample = bo[:, sub_locs]
# plot sample patient locations
bo_sample.plot_locs()
# plot sample patient data
bo_sample.plot_data()
# make model from brain object
r_model = se.Model(data=bo, locs=locs)
# predict
bo_s = r_model.predict(bo_sample, nearest_neighbor=False)
# find indices for reconstructed locations
recon_labels = np.where(np.array(bo_s.label) != 'observed')
# find correlations between predicted and actual data
corrs = _corr_column(bo.get_data().as_matrix(), bo_s.get_data().as_matrix())
# index reconstructed correlations
corrs[recon_labels].mean()
def test_model_predict_nn_0():
model = se.Model(data=data[0:2], locs=locs)
bo_1 = model.predict(data[0], nearest_neighbor=True, match_threshold=0)
bo_2 = model.predict(data[0], nearest_neighbor=False)
assert isinstance(bo_1, se.Brain)
assert np.allclose(bo_1.get_data(), bo_2.get_data())
def test_model_predict_nn():
print(data[0].dur)
model = se.Model(data=data[0:2], locs=locs)
bo = model.predict(data[0], nearest_neighbor=True)
assert isinstance(bo, se.Brain)
def test_model_update_inplace():
mo = se.Model(data=data[1:3], locs=locs)
mo = mo.update(data[0])
assert mo is None
def test_create_model_model():
mo = se.Model(data=data[1:3], locs=locs)
model = se.Model(mo)
assert isinstance(model, se.Model)
[ 59., -17., 17.]])
# number of timeseries samples
n_samples = 10
# number of subjects
n_subs = 6
# number of electrodes
n_elecs = 5
# simulate correlation matrix
data = [se.simulate_model_bos(n_samples=10, sample_rate=10, locs=locs, sample_locs = n_elecs,
set_random_seed=123, noise=0) for x in range(n_subs)]
# test model to compare
test_model = se.Model(data=data[0:3], locs=locs, rbf_width=20, n_subs=3)
def test_create_model_1bo():
model = se.Model(data=data[0], locs=locs)
assert isinstance(model, se.Model)
def test_create_model_2bo():
model = se.Model(data=data[0:2], locs=locs)
assert isinstance(model, se.Model)
def test_create_model_superuser():
locs = np.random.multivariate_normal(np.zeros(3), np.eye(3), size=10)
numerator = scipy.linalg.toeplitz(np.linspace(0,10,len(locs))[::-1])
denominator = np.random.multivariate_normal(np.zeros(10), np.eye(10), size=10)
model = se.Model(numerator=numerator, denominator=denominator, locs=locs, n_subs=2)
locs = np.array([[-61., -77., -3.], [-41., -77., -23.], [-21., -97., 17.],
[-21., -37., 77.], [-21., 63., -3.], [-1., -37., 37.],
[-1., 23., 17.], [19., -57., -23.], [19., 23., -3.],
[39., -57., 17.], [39., 3., 37.], [59., -17., 17.]])
n_samples = 10
n_subs = 3
n_elecs = 10
data = [
se.simulate_model_bos(n_samples=10,
sample_rate=10,
locs=locs,
sample_locs=n_elecs) for x in range(n_subs)
]
test_bo = data[0]
test_model = se.Model(data=data, locs=locs)
bo = se.load('example_data')
def test_load_example_data():
bo = se.load('example_data')
assert isinstance(bo, se.Brain)
def test_load_example_filter():
bo = se.load('example_filter')
assert isinstance(bo, se.Brain)
def test_load_example_model():
model = se.load('example_model')
locs = se.simulate_locations(n_elecs=100)
# simulate correlation matrix
R = se.create_cov(cov='toeplitz', n_elecs=len(locs))
# simulate brain objects for the model that subsample n_elecs for each synthetic patient
model_bos = [
se.simulate_model_bos(n_samples=1000,
sample_rate=1000,
locs=locs,
sample_locs=10,
cov='toeplitz') for x in range(3)
]
# create the model object
model = se.Model(data=model_bos, locs=locs, n_subs=3)
model.plot_data()
# brain object locations subsetted
sub_locs = locs.sample(10).sort_values(['x', 'y', 'z'])
# simulate a new brain object using the same covariance matrix
bo = se.simulate_bo(n_samples=1000,
sample_rate=1000,
locs=sub_locs,
cov='toeplitz')
# update the model
new_model = model.update(bo, inplace=False)
# initialize subplots
except:
numtries += 1
time.sleep(5)
bo = se.load(sys.argv[1])
# load original brain object
og_fname = os.path.join(config['og_bodir'], fname.split('_' + freq)[0] + '.bo')
try:
og_bo = se.load(og_fname)
except:
og_bo = se.load(sys.argv[1])
og_bo.update_filter_inds()
# turn it into fancy ~BandBrain~
bo = BandBrain(bo, og_bo=og_bo, og_filter_inds=og_bo.filter_inds)
# filter
bo.apply_filter()
# turn it back into a vanilla Brain
bo = se.Brain(bo)
# make model
mo = se.Model(bo, locs=R)
# save model
mo.save(os.path.join(results_dir, fname))
else:
print('skipping model (not enough electrodes pass kurtosis threshold): ' + sys.argv[1])
# nii = se.load('gray', vox_size=3)
#
# locs = nii.get_locs()
locs_dir = config['locs_resultsdir']
mo_locs = np.load(os.path.join(locs_dir, 'mo_locs.npz'))
## for full locations
#locs = mo_locs['full_locs']
## for every 100th locations
locs = mo_locs['sub_locs']
fname = os.path.basename(os.path.splitext(fname)[0])
print('creating model object: ' + fname)
# load brain object
bo = se.load(sys.argv[1])
# filter
bo.filter = None
# make model
mo = se.Model(bo, locs=locs)
# save model
mo.save(os.path.join(results_dir, fname))
def test_create_model_2bo():
model = se.Model(data=data[0:2], locs=locs)
assert isinstance(model, se.Model)
def test_create_model_superuser():
locs = np.random.multivariate_normal(np.zeros(3), np.eye(3), size=10)
numerator = scipy.linalg.toeplitz(np.linspace(0,10,len(locs))[::-1])
denominator = np.random.multivariate_normal(np.zeros(10), np.eye(10), size=10)
model = se.Model(numerator=numerator, denominator=denominator, locs=locs, n_subs=2)
assert isinstance(model, se.Model)
def test_model_update_with_model():
mo = se.Model(data=data[1:3], locs=locs)
mo = mo.update(mo, inplace=False)
assert isinstance(mo, se.Model)
def test_model_predict_nn_thresh():
model = se.Model(data=data[0:2], locs=locs)
bo = model.predict(data[0], nearest_neighbor=True, match_threshold=30)
assert isinstance(bo, se.Brain)
def test_model_update_with_model_and_bo():
mo = se.Model(data=data[1:3], locs=locs)
mo = se.Model([mo, data[0]])
assert isinstance(mo, se.Model)
def test_create_model_str():
model = se.Model('example_data')
assert isinstance(model, se.Model)
def test_model_update_with_array():
mo = se.Model(data=data[1:3], locs=locs)
d = np.random.rand(*mo.numerator.shape)
mo = se.Model([mo, d], locs=mo.get_locs())
assert isinstance(mo, se.Model)
for p, m, n in param_grid:
d = []
for i in range(iter_val):
# create brain objects with m_patients and loop over the number of model locations and subset locations to build model
model_bos = [
se.simulate_model_bos(n_samples=1000,
sample_rate=100,
locs=locs,
sample_locs=m,
noise=.3) for x in range(p)
]
# create model from subsampled gray locations
model = se.Model(model_bos, locs=locs)
# brain object locations subsetted entirely from both model and gray locations
sub_locs = locs.sample(n).sort_values(['x', 'y', 'z'])
# simulate brain object
bo = se.simulate_bo(n_samples=1000,
sample_rate=100,
locs=locs,
noise=.3)
# parse brain object to create synthetic patient data
data = bo.data.iloc[:, sub_locs.index]
# create synthetic patient (will compare remaining activations to predictions)
bo_sample = se.Brain(data=data.as_matrix(),
def test_model_get_model():
mo = se.Model(data=data[1:3], locs=locs)
m = mo.get_model()
assert isinstance(m, np.ndarray)
# locs
locs = se.simulate_locations(n_elecs=100, set_random_seed=random_seed)
# create model locs from 75 locations
mo_locs = locs.sample(75,
random_state=random_seed).sort_values(['x', 'y', 'z'])
# create covariance matrix from random seed
c = se.create_cov(cov='random', n_elecs=100)
# pull out model from covariance matrix
data = c[:, mo_locs.index][mo_locs.index, :]
# create model from subsetted covariance matrix and locations
model = se.Model(data=data, locs=mo_locs, n_subs=1)
# create brain object from the remaining locations - first find remaining 25 locations
sub_locs = locs[~locs.index.isin(mo_locs.index)]
# create a brain object with all gray locations
bo = se.simulate_bo(n_samples=1000,
sample_rate=100,
locs=locs,
noise=noise,
random_seed=random_seed)
# parse brain object to create synthetic patient data
data = bo.data.iloc[:, sub_locs.index]
# put data and locations together in new sample brain object
def test_model_update_with_smaller_array_locs_specified():
mo = se.Model(data=data[1:3], locs=locs)
d = np.random.rand(3, 3)
mo = mo.update(d, inplace=False, locs=d)
assert isinstance(mo, se.Model)
freq = sys.argv[1]
model_dir = os.path.join(config['datadir'])
results_dir = config['resultsdir']
model_dir = config['datadir']
if os.path.exists(os.path.join(results_dir, 'ave_mat_' + freq)):
print('ave mat already exists')
exit()
try:
if not os.path.exists(results_dir):
os.makedirs(results_dir)
except OSError as err:
print(err)
mos = glob.glob(os.path.join(model_dir, '*' + freq + '.mo'))
if freq == 'raw':
freqnames = ['delta', 'theta', 'alpha', 'beta', 'lgamma', 'hgamma', 'broadband']
mos = set(glob.glob(os.path.join(model_dir, '*')))
for fre in freqnames:
mos -= set(glob.glob(os.path.join(model_dir, '*' + fre + '*')))
mos = list(mos)
print(len(mos))
mo = se.Model(mos, n_subs=len(mos))
mo.save(os.path.join(results_dir, 'ave_mat_' + freq))
def test_model_predict():
model = se.Model(data=data[0:2], locs=locs)
bo = model.predict(data[0], nearest_neighbor=False)
assert isinstance(bo, se.Brain)
