def lap_multiple_examples(moving_tractograms_dir, static_tractogram, ex_dir,
out_filename):
"""Code for LAP from multiple examples.
"""
moving_tractograms = os.listdir(moving_tractograms_dir)
moving_tractograms.sort()
examples = os.listdir(ex_dir)
examples.sort()
nt = len(moving_tractograms)
ne = len(examples)
if nt != ne:
print(
"Error: number of moving tractograms differs from number of example bundles."
)
sys.exit()
else:
result_lap = []
for i in range(nt):
moving_tractogram = '%s/%s' % (moving_tractograms_dir,
moving_tractograms[i])
example = '%s/%s' % (ex_dir, examples[i])
tmp = np.array([
lap_single_example(moving_tractogram, static_tractogram,
example)
])
result_lap.append(tmp)
result_lap = np.array(result_lap)
estimated_bundle_idx = np.hstack(result_lap[:, 0, 0])
min_cost_values = np.hstack(result_lap[:, 0, 1])
example_bundle_len_med = np.median(np.hstack(result_lap[:, 0, 2]))
print("Ranking the estimated streamlines...")
estimated_bundle_idx_ranked = ranking_schema(estimated_bundle_idx,
min_cost_values)
np.save('candidate_bundle_idx_ranked_lap.npy',
estimated_bundle_idx_ranked)
np.save('candidate_bundle_idx_lap.npy', estimated_bundle_idx)
np.save('min_cost_values_lap.npy', min_cost_values)
print("Extracting the estimated bundle...")
estimated_bundle_idx_ranked_med = estimated_bundle_idx_ranked[
0:int(example_bundle_len_med)]
np.save('estimated_bundle_idx_lap.npy',
estimated_bundle_idx_ranked_med)
with open('config.json') as f:
data = json.load(f)
step_size = data["step_size"]
save_bundle(estimated_bundle_idx_ranked_med, static_tractogram,
step_size, out_filename)
return result_lap
parser.add_argument('-lE',
nargs='?',
const=1,
default='',
help='Weight of the endpoint matrix')
parser.add_argument('-lR',
nargs='?',
const=1,
default='',
help='Weight of the waypoint matrix')
parser.add_argument('-out',
nargs='?',
const=1,
default='',
help='The output estimated bundle filename')
args = parser.parse_args()
result_lap = lap_single_example(args.moving, args.static, args.ex, args.lD,
args.lE, args.lR)
np.save('result_lap', result_lap)
if args.out:
estimated_bundle_idx = result_lap[0]
with open('config.json') as f:
data = json.load(f)
step_size = data["step_size"]
save_bundle(estimated_bundle_idx, args.static, step_size, args.out)
sys.exit()
def lap_multiple_examples(moving_tractograms_dir, static_tractogram, ex_dir,
lD, lE, lR, out_filename):
"""Code for LAP from multiple examples.
"""
moving_tractograms = os.listdir(moving_tractograms_dir)
moving_tractograms.sort()
examples = os.listdir(ex_dir)
examples.sort()
nt = len(moving_tractograms)
ne = len(examples)
#moving_tractograms = os.listdir(moving_tractograms_dir)
#moving_tractograms.sort()
#nt = len(moving_tractograms)
#moving_tractograms = ['%s/' %moving_tractograms_dir + moving_tractograms[i] for i in range(nt)]
#examples = os.listdir(ex_dir)
#examples.sort()
#ne = len(examples)
#examples = ['%s/' %ex_dir + examples[i] for i in range(ne)]
if nt != ne:
print(
"Error: number of moving tractograms differs from number of example bundles."
)
sys.exit()
else:
result_lap = []
for i in range(nt):
moving_tractogram = '%s/%s' % (moving_tractograms_dir,
moving_tractograms[i])
example = '%s/%s' % (ex_dir, examples[i])
tmp = np.array([
lap_single_example(moving_tractogram, static_tractogram,
example, lD, lE, lR)
])
result_lap.append(tmp)
result_lap = np.array(result_lap)
estimated_bundle_idx = np.hstack(result_lap[:, 0, 0])
min_cost_values = np.hstack(result_lap[:, 0, 1])
example_bundle_len_med = np.median(np.hstack(result_lap[:, 0, 2]))
#result_lap = np.array(Parallel(n_jobs=-1)(delayed(lap_single_example)(moving_tractograms[i], static_tractogram, examples[i], lD, lE, lR) for i in range(nt)))
#estimated_bundle_idx = np.hstack(result_lap[:,0])
#min_cost_values = np.hstack(result_lap[:,1])
#example_bundle_len_med = np.median(np.hstack(result_lap[:,2]))
print("Ranking the estimated streamlines...")
estimated_bundle_idx_ranked = ranking_schema(estimated_bundle_idx,
min_cost_values)
print("Extracting the estimated bundle...")
estimated_bundle_idx_ranked_med = estimated_bundle_idx_ranked[
0:int(example_bundle_len_med)]
with open('config.json') as f:
data = json.load(f)
step_size = data["step_size"]
save_bundle(estimated_bundle_idx_ranked_med, static_tractogram,
step_size, out_filename)
return result_lap