def test_find_peak_success():
data = np.zeros((21, 21))
y, x = np.indices(data.shape)
data = 100 * np.exp(-0.5 * ((x - 8)**2 + (y - 11)**2))
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=3)
assert fit_status == 'SUCCESS'
assert np.allclose(coord, (8, 11), rtol=0, atol=1e-6)
def test_find_peak_fail_lstsq():
data = np.zeros((11, 21))
data[6, 7] = 10
data[8, 7] = np.nan
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=7)
assert fit_status == 'WARNING:CENTER-OF-MASS'
assert np.allclose(coord, (7, 6), rtol=0, atol=1e-6)
def test_find_peak_fit_over_edge():
data = np.zeros((21, 21))
y, x = np.indices(data.shape)
data = 100 * np.exp(-0.5 * (x**2 + (y - 11)**2))
data[:, 0] = 0.0
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=7)
assert fit_status == 'WARNING:EDGE'
assert np.allclose(coord, (1, 11), rtol=0, atol=1e-6)
def test_find_peak_badfit():
data = np.zeros((21, 21))
y, x = np.indices(data.shape)
data = x + y
data[(x < 5) | (x > 11) | (y < 8) | (y > 14)] = 0
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=7)
assert fit_status == 'WARNING:BADFIT'
assert np.allclose(coord, (11, 14), rtol=0, atol=1e-6)
def test_find_peak_negative_peak():
data = np.zeros((11, 11))
i = random.choice(range(2, 9))
j = random.choice(range(2, 9))
data[i, j] = -1.0
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=2)
assert fit_status == 'ERROR:NODATA'
assert np.allclose(coord, (5, 5), rtol=0, atol=_ATOL)
def test_find_peak_nodata_after_fail():
data = np.zeros((21, 21))
i = random.choice(range(6, 14))
j = random.choice(range(6, 14))
data[i, j] = 1.0
data[i - 1, j - 1] = -1.0
data[i + 1, j + 1] = np.nan
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5)
assert fit_status == 'ERROR:NODATA'
assert np.allclose(coord, (10, 10), rtol=0, atol=_ATOL)
def test_find_peak_tiny_box_1pix():
data = np.zeros((4, 4))
mask = np.zeros((4, 4), dtype=np.bool_)
mask[2, 2] = True
data[2, 2] = 1.0
coord0 = (2, 2)
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5, mask=mask)
assert np.allclose(coord, coord0, rtol=0, atol=_ATOL)
assert fit_status == 'WARNING:CENTER-OF-MASS'
def test_find_peak_nodata_after_fail():
data = np.zeros((21, 21))
i = random.choice(range(6, 14))
j = random.choice(range(6, 14))
data[i, j] = 1.0
data[i - 1, j - 1] = -1.0
data[i + 1, j + 1] = np.nan
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5)
assert fit_status == 'SUCCESS'
shift = 0.53583061889252015
assert np.allclose(coord, (j + shift, i + shift), rtol=0, atol=20 * _ATOL)
def test_find_peak_few_data_center_of_mass():
data = np.zeros((20, 20))
mask = np.zeros((20, 20), dtype=np.bool_)
col = random.choice(range(5, 16))
data[9, col] = 1.0
mask[9, col] = True
coord0 = (col, 9)
coord, fit_status, fit_box = _find_peak(data, mask=mask)
assert fit_status == 'WARNING:CENTER-OF-MASS'
assert np.allclose(coord, coord0, rtol=0, atol=_ATOL)
def test_find_peak_few_data_for_center_of_mass():
data = np.zeros((21, 21))
mask = np.zeros((21, 21), dtype=np.bool_)
i = random.choice(range(5, 16))
j = random.choice(range(5, 16))
data[i, j] = 1.0
data[i - 1, j - 1] = -1.0
mask[i, j] = True
mask[i - 1, j - 1] = True
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=3, mask=mask)
assert fit_status == 'ERROR:NODATA'
assert np.allclose(coord, (10, 10), rtol=0, atol=_ATOL)
def test_find_peak_nodata_peak_is_invalid():
data = np.zeros((14, 17))
mask = np.zeros((14, 17), dtype=np.bool_)
col = random.choice(range(17))
mask[9, col] = True
data[9, 4] = 1.0
mask[9, 4] = False
coord0 = (8, 6.5)
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5, mask=mask)
assert fit_status == 'ERROR:NODATA'
assert np.allclose(coord, coord0, rtol=0, atol=_ATOL)
def test_find_peak_nodata_all_zeros(shape, mask):
data = np.zeros(shape)
if mask is not None:
if mask:
mask = np.random.choice([True, False], shape)
mask[0, 0] = True # make sure at least one element is non-zero
else:
mask = np.zeros(shape, dtype=np.bool_)
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5, mask=mask)
assert data[fit_box].shape == shape
assert np.allclose(coord, [0.5 * (x - 1) for x in shape[::-1]],
rtol=0, atol=_ATOL)
assert fit_status == 'ERROR:NODATA'
def test_find_peak_edge_1pix_valid_strip(along_row):
data = np.zeros((10, 20))
mask = np.zeros((10, 20), dtype=np.bool_)
if along_row:
row = random.choice(range(10))
data[row, :] = 1.0
mask[row, :] = True
coord0 = (0, row)
else:
col = random.choice(range(20))
data[:, col] = 1.0
mask[:, col] = True
coord0 = (col, 0)
coord, fit_status, fit_box = _find_peak(data, peak_fit_box=5, mask=mask)
assert fit_status == 'WARNING:EDGE'
assert np.allclose(coord, coord0, rtol=0, atol=_ATOL)
def test_find_peak_negative_box_size():
with pytest.raises(ValueError):
_find_peak(np.zeros((2, 2)), peak_fit_box=-1)
def test_find_peak_sparse_2dhist():
data = fits.getdata(get_pkg_data_filename('data/sparse-2dhist.fits'))
coord, fit_status, fit_box = _find_peak(data, mask=data > 0)
assert fit_status == 'WARNING:CENTER-OF-MASS'
assert np.allclose(coord, (35.046728971962615, 35.02803738317757),
rtol=0, atol=1e-14)
