def test_ensure_spacing_trivial(self):
for p in [1, 2, np.inf]:
for size in [30, 50, None]:
with self.subTest(p=p, size=size):
# --- Empty input
self.assertEqual(ensure_spacing([], p_norm=p), [])
# --- A unique point
coord = np.random.randn(1, 2)
np.testing.assert_array_equal(
coord,
ensure_spacing(coord, p_norm=p, min_split_size=size))
# --- Verified spacing
coord = np.random.randn(100, 2)
# --- 0 spacing
np.testing.assert_array_equal(
coord,
ensure_spacing(coord,
spacing=0,
p_norm=p,
min_split_size=size),
)
# Spacing is chosen to be half the minimum distance
spacing = pdist(coord, metric=minkowski, p=p).min() * 0.5
out = ensure_spacing(coord,
spacing=spacing,
p_norm=p,
min_split_size=size)
np.testing.assert_array_equal(coord, out)
def test_ensure_spacing_batch_processing(self):
for p in [1, 2, np.inf]:
for size in [50, 100, None]:
with self.subTest(p=p, size=size):
coord = np.random.randn(100, 2)
# --- Consider the average distance between the point as spacing
spacing = np.median(pdist(coord, metric=minkowski, p=p))
expected = ensure_spacing(coord, spacing=spacing, p_norm=p)
assert np.array_equal(
ensure_spacing(coord,
spacing=spacing,
p_norm=p,
min_split_size=size),
expected,
)
def test_ensure_spacing_nD(self):
for ndim in [1, 2, 3, 4, 5]:
for size in [2, 10, None]:
with self.subTest(ndim=ndim, size=size):
coord = np.ones((5, ndim))
expected = np.ones((1, ndim))
assert np.array_equal(
ensure_spacing(coord, min_split_size=size), expected)
def test_ensure_spacing_p_norm(self):
for p in [1, 2, np.inf]:
for size in [30, 50, None]:
with self.subTest(p=p, size=size):
coord = np.random.randn(100, 2)
# --- Consider the average distance between the point as spacing
spacing = np.median(pdist(coord, metric=minkowski, p=p))
out = ensure_spacing(coord,
spacing=spacing,
p_norm=p,
min_split_size=size)
self.assertGreater(
pdist(out, metric=minkowski, p=p).min(), spacing)
def test_multiple(self):
"""
`find_spot_positions` should find all spot positions in a generated
test image.
"""
n = 100
shape = (256, 256)
refractive_index = 1
numerical_aperture = 0.95
wavelength = 0.55 # [µm]
magnification = 40
pixel_size = 3.45 # [µm]
# Ensure that each time when the test case is run we have the same
# 'random' numbers.
numpy.random.seed(0)
sigma = (magnification / pixel_size) * synthetic.psf_sigma_wffm(
refractive_index, numerical_aperture, wavelength)
# Generate a test image containing at most 100 randomly distributed
# spots with a guaranteed minimal spacing.
spacing = int(round(10 * sigma))
border = numpy.asarray((spacing, spacing))
srange = numpy.asarray(shape) - 2 * border - numpy.array((1, 1))
loc = border + srange * numpy.random.random_sample((n, 2))
loc = ensure_spacing(loc, spacing=spacing, p_norm=2)
image = synthetic.psf_gaussian(shape, loc, sigma)
ji = spot.find_spot_positions(image, sigma)
# Map the found spot locations to the known spot locations.
tree = scipy.spatial.cKDTree(loc)
# NOTE: Starting SciPy v1.6.0 the `n_jobs` argument will be renamed `workers`
distances, indices = tree.query(ji, k=1, n_jobs=-1)
# Check that all spot positions have been found within the required
# accuracy.
numpy.testing.assert_array_equal(sorted(indices), range(len(loc)))
numpy.testing.assert_array_less(distances, 0.05)