def _test_2d(self, matrix, out_file=None, sigma=None, **kwargs):
from elf.transformation import transform_subvolume_affine
shape = (512, 512)
x = np.random.rand(*shape)
exp = affine_transform(x, matrix, **kwargs)
if out_file is not None:
with open_file(out_file) as f:
x = f.create_dataset('tmp', data=x, chunks=(64, 64))
f = open_file(out_file, 'r')
x = f['tmp']
bbs = [
np.s_[:, :], np.s_[:256, :256], np.s_[37:115, 226:503],
np.s_[:200, :], np.s_[:, 10:115]
]
for bb in bbs:
bb, _ = normalize_index(bb, shape)
res = transform_subvolume_affine(x,
matrix,
bb,
sigma=sigma,
**kwargs)
exp_bb = exp[bb]
self.assertEqual(res.shape, exp_bb.shape)
if sigma is None:
self.assertTrue(np.allclose(res, exp_bb))
else:
self.assertTrue(~np.allclose(res, 0))
if out_file is not None:
f.close()
def _test_3d(self, matrix, out_file=None, **kwargs):
from elf.transformation import transform_subvolume_affine
shape = 3 * (64, )
x = np.random.rand(*shape)
exp = affine_transform(x, matrix, **kwargs)
if out_file is not None:
with open_file(out_file) as f:
x = f.create_dataset('tmp', data=x, chunks=3 * (16, ))
f = open_file(out_file, 'r')
x = f['tmp']
bbs = [
np.s_[:, :, :], np.s_[:32, :32, :32], np.s_[1:31, 5:27, 3:13],
np.s_[4:19, :, 22:], np.s_[1:29], np.s_[:, 15:27, :], np.s_[:, 1:3,
4:14]
]
for bb in bbs:
bb, _ = normalize_index(bb, shape)
res = transform_subvolume_affine(x, matrix, bb, **kwargs)
exp_bb = exp[bb]
self.assertEqual(res.shape, exp_bb.shape)
self.assertTrue(np.allclose(res, exp_bb))
if out_file is not None:
f.close()
def paths_from_ds(self, dataset):
if isinstance(dataset, ConcatDataset):
datasets = dataset.datasets
(n_samples, load_2d_from_3d, rois, raw_paths, raw_key, label_paths,
label_key) = self.paths_from_ds(datasets[0])
for ds in datasets[1:]:
ns, l2d3d, bb, rp, rk, lp, lk = self.paths_from_ds(ds)
assert rk == raw_key
assert lk == label_key
assert l2d3d == load_2d_from_3d
raw_paths.extend(rp)
label_paths.extend(lp)
rois.append(bb)
n_samples += ns
elif isinstance(dataset, ImageCollectionDataset):
raw_paths, label_paths = dataset.raw_images, dataset.label_images
raw_key, label_key = None, None
n_samples = len(raw_paths)
load_2d_from_3d = False
rois = [None] * n_samples
elif isinstance(dataset, SegmentationDataset):
raw_paths, label_paths = [dataset.raw_path], [dataset.label_path]
raw_key, label_key = dataset.raw_key, dataset.label_key
shape = open_file(raw_paths[0], 'r')[raw_key].shape
roi = getattr(dataset, 'roi', None)
if roi is not None:
roi = normalize_index(roi, shape)
shape = tuple(r.stop - r.start for r in roi)
rois = [roi]
if self.ndim == len(shape):
n_samples = len(raw_paths)
load_2d_from_3d = False
elif self.ndim == 2 and len(shape) == 3:
n_samples = shape[0]
load_2d_from_3d = True
else:
raise RuntimeError
else:
raise RuntimeError(
f"No support for dataset of type {type(dataset)}")
return (n_samples, load_2d_from_3d, rois, raw_paths, raw_key,
label_paths, label_key)
def test_normalize_index(self):
from elf.util import normalize_index, squeeze_singletons
shape = (128, ) * 3
x = np.random.rand(*shape)
# is something important missing?
indices = (
np.s_[10:25, 30:60, 100:103], # full index
np.s_[:], # everything
np.s_[..., 10:25], # ellipsis
np.s_[0, :, 10]) # singletons
for index in indices:
out1 = x[index]
index, to_squeeze = normalize_index(index, shape)
out2 = squeeze_singletons(x[index], to_squeeze)
self.assertEqual(out1.shape, out2.shape)
self.assertTrue(np.allclose(out1, out2))
def __getitem__(self, key):
mapped_key, to_squeeze = normalize_index(key, self.shape)
mapped_key = tuple(
slice(k.start + rs, k.stop + rs)
for k, rs in zip(mapped_key, self.roi_start))
return squeeze_singletons(self.source[mapped_key], to_squeeze)
