888 lines
33 KiB
Python
888 lines
33 KiB
Python
import copy
|
|
|
|
import numpy as np
|
|
from numpy.testing import (
|
|
assert_,
|
|
assert_equal,
|
|
assert_allclose,
|
|
assert_array_equal
|
|
)
|
|
import pytest
|
|
from pytest import raises, warns
|
|
|
|
from scipy.signal._peak_finding import (
|
|
argrelmax,
|
|
argrelmin,
|
|
peak_prominences,
|
|
peak_widths,
|
|
_unpack_condition_args,
|
|
find_peaks,
|
|
find_peaks_cwt,
|
|
_identify_ridge_lines
|
|
)
|
|
from scipy.signal.windows import gaussian
|
|
from scipy.signal._peak_finding_utils import _local_maxima_1d, PeakPropertyWarning
|
|
|
|
|
|
def _gen_gaussians(center_locs, sigmas, total_length):
|
|
xdata = np.arange(0, total_length).astype(float)
|
|
out_data = np.zeros(total_length, dtype=float)
|
|
for ind, sigma in enumerate(sigmas):
|
|
tmp = (xdata - center_locs[ind]) / sigma
|
|
out_data += np.exp(-(tmp**2))
|
|
return out_data
|
|
|
|
|
|
def _gen_gaussians_even(sigmas, total_length):
|
|
num_peaks = len(sigmas)
|
|
delta = total_length / (num_peaks + 1)
|
|
center_locs = np.linspace(delta, total_length - delta, num=num_peaks).astype(int)
|
|
out_data = _gen_gaussians(center_locs, sigmas, total_length)
|
|
return out_data, center_locs
|
|
|
|
|
|
def _gen_ridge_line(start_locs, max_locs, length, distances, gaps):
|
|
"""
|
|
Generate coordinates for a ridge line.
|
|
|
|
Will be a series of coordinates, starting a start_loc (length 2).
|
|
The maximum distance between any adjacent columns will be
|
|
`max_distance`, the max distance between adjacent rows
|
|
will be `map_gap'.
|
|
|
|
`max_locs` should be the size of the intended matrix. The
|
|
ending coordinates are guaranteed to be less than `max_locs`,
|
|
although they may not approach `max_locs` at all.
|
|
"""
|
|
|
|
def keep_bounds(num, max_val):
|
|
out = max(num, 0)
|
|
out = min(out, max_val)
|
|
return out
|
|
|
|
gaps = copy.deepcopy(gaps)
|
|
distances = copy.deepcopy(distances)
|
|
|
|
locs = np.zeros([length, 2], dtype=int)
|
|
locs[0, :] = start_locs
|
|
total_length = max_locs[0] - start_locs[0] - sum(gaps)
|
|
if total_length < length:
|
|
raise ValueError('Cannot generate ridge line according to constraints')
|
|
dist_int = length / len(distances) - 1
|
|
gap_int = length / len(gaps) - 1
|
|
for ind in range(1, length):
|
|
nextcol = locs[ind - 1, 1]
|
|
nextrow = locs[ind - 1, 0] + 1
|
|
if (ind % dist_int == 0) and (len(distances) > 0):
|
|
nextcol += ((-1)**ind)*distances.pop()
|
|
if (ind % gap_int == 0) and (len(gaps) > 0):
|
|
nextrow += gaps.pop()
|
|
nextrow = keep_bounds(nextrow, max_locs[0])
|
|
nextcol = keep_bounds(nextcol, max_locs[1])
|
|
locs[ind, :] = [nextrow, nextcol]
|
|
|
|
return [locs[:, 0], locs[:, 1]]
|
|
|
|
|
|
class TestLocalMaxima1d:
|
|
|
|
def test_empty(self):
|
|
"""Test with empty signal."""
|
|
x = np.array([], dtype=np.float64)
|
|
for array in _local_maxima_1d(x):
|
|
assert_equal(array, np.array([]))
|
|
assert_(array.base is None)
|
|
|
|
def test_linear(self):
|
|
"""Test with linear signal."""
|
|
x = np.linspace(0, 100)
|
|
for array in _local_maxima_1d(x):
|
|
assert_equal(array, np.array([]))
|
|
assert_(array.base is None)
|
|
|
|
def test_simple(self):
|
|
"""Test with simple signal."""
|
|
x = np.linspace(-10, 10, 50)
|
|
x[2::3] += 1
|
|
expected = np.arange(2, 50, 3)
|
|
for array in _local_maxima_1d(x):
|
|
# For plateaus of size 1, the edges are identical with the
|
|
# midpoints
|
|
assert_equal(array, expected)
|
|
assert_(array.base is None)
|
|
|
|
def test_flat_maxima(self):
|
|
"""Test if flat maxima are detected correctly."""
|
|
x = np.array([-1.3, 0, 1, 0, 2, 2, 0, 3, 3, 3, 2.99, 4, 4, 4, 4, -10,
|
|
-5, -5, -5, -5, -5, -10])
|
|
midpoints, left_edges, right_edges = _local_maxima_1d(x)
|
|
assert_equal(midpoints, np.array([2, 4, 8, 12, 18]))
|
|
assert_equal(left_edges, np.array([2, 4, 7, 11, 16]))
|
|
assert_equal(right_edges, np.array([2, 5, 9, 14, 20]))
|
|
|
|
@pytest.mark.parametrize('x', [
|
|
np.array([1., 0, 2]),
|
|
np.array([3., 3, 0, 4, 4]),
|
|
np.array([5., 5, 5, 0, 6, 6, 6]),
|
|
])
|
|
def test_signal_edges(self, x):
|
|
"""Test if behavior on signal edges is correct."""
|
|
for array in _local_maxima_1d(x):
|
|
assert_equal(array, np.array([]))
|
|
assert_(array.base is None)
|
|
|
|
def test_exceptions(self):
|
|
"""Test input validation and raised exceptions."""
|
|
with raises(ValueError, match="wrong number of dimensions"):
|
|
_local_maxima_1d(np.ones((1, 1)))
|
|
with raises(ValueError, match="expected 'const float64_t'"):
|
|
_local_maxima_1d(np.ones(1, dtype=int))
|
|
with raises(TypeError, match="list"):
|
|
_local_maxima_1d([1., 2.])
|
|
with raises(TypeError, match="'x' must not be None"):
|
|
_local_maxima_1d(None)
|
|
|
|
|
|
class TestRidgeLines:
|
|
|
|
def test_empty(self):
|
|
test_matr = np.zeros([20, 100])
|
|
lines = _identify_ridge_lines(test_matr, np.full(20, 2), 1)
|
|
assert_(len(lines) == 0)
|
|
|
|
def test_minimal(self):
|
|
test_matr = np.zeros([20, 100])
|
|
test_matr[0, 10] = 1
|
|
lines = _identify_ridge_lines(test_matr, np.full(20, 2), 1)
|
|
assert_(len(lines) == 1)
|
|
|
|
test_matr = np.zeros([20, 100])
|
|
test_matr[0:2, 10] = 1
|
|
lines = _identify_ridge_lines(test_matr, np.full(20, 2), 1)
|
|
assert_(len(lines) == 1)
|
|
|
|
def test_single_pass(self):
|
|
distances = [0, 1, 2, 5]
|
|
gaps = [0, 1, 2, 0, 1]
|
|
test_matr = np.zeros([20, 50]) + 1e-12
|
|
length = 12
|
|
line = _gen_ridge_line([0, 25], test_matr.shape, length, distances, gaps)
|
|
test_matr[line[0], line[1]] = 1
|
|
max_distances = np.full(20, max(distances))
|
|
identified_lines = _identify_ridge_lines(test_matr, max_distances, max(gaps) + 1)
|
|
assert_array_equal(identified_lines, [line])
|
|
|
|
def test_single_bigdist(self):
|
|
distances = [0, 1, 2, 5]
|
|
gaps = [0, 1, 2, 4]
|
|
test_matr = np.zeros([20, 50])
|
|
length = 12
|
|
line = _gen_ridge_line([0, 25], test_matr.shape, length, distances, gaps)
|
|
test_matr[line[0], line[1]] = 1
|
|
max_dist = 3
|
|
max_distances = np.full(20, max_dist)
|
|
#This should get 2 lines, since the distance is too large
|
|
identified_lines = _identify_ridge_lines(test_matr, max_distances, max(gaps) + 1)
|
|
assert_(len(identified_lines) == 2)
|
|
|
|
for iline in identified_lines:
|
|
adists = np.diff(iline[1])
|
|
np.testing.assert_array_less(np.abs(adists), max_dist)
|
|
|
|
agaps = np.diff(iline[0])
|
|
np.testing.assert_array_less(np.abs(agaps), max(gaps) + 0.1)
|
|
|
|
def test_single_biggap(self):
|
|
distances = [0, 1, 2, 5]
|
|
max_gap = 3
|
|
gaps = [0, 4, 2, 1]
|
|
test_matr = np.zeros([20, 50])
|
|
length = 12
|
|
line = _gen_ridge_line([0, 25], test_matr.shape, length, distances, gaps)
|
|
test_matr[line[0], line[1]] = 1
|
|
max_dist = 6
|
|
max_distances = np.full(20, max_dist)
|
|
#This should get 2 lines, since the gap is too large
|
|
identified_lines = _identify_ridge_lines(test_matr, max_distances, max_gap)
|
|
assert_(len(identified_lines) == 2)
|
|
|
|
for iline in identified_lines:
|
|
adists = np.diff(iline[1])
|
|
np.testing.assert_array_less(np.abs(adists), max_dist)
|
|
|
|
agaps = np.diff(iline[0])
|
|
np.testing.assert_array_less(np.abs(agaps), max(gaps) + 0.1)
|
|
|
|
def test_single_biggaps(self):
|
|
distances = [0]
|
|
max_gap = 1
|
|
gaps = [3, 6]
|
|
test_matr = np.zeros([50, 50])
|
|
length = 30
|
|
line = _gen_ridge_line([0, 25], test_matr.shape, length, distances, gaps)
|
|
test_matr[line[0], line[1]] = 1
|
|
max_dist = 1
|
|
max_distances = np.full(50, max_dist)
|
|
#This should get 3 lines, since the gaps are too large
|
|
identified_lines = _identify_ridge_lines(test_matr, max_distances, max_gap)
|
|
assert_(len(identified_lines) == 3)
|
|
|
|
for iline in identified_lines:
|
|
adists = np.diff(iline[1])
|
|
np.testing.assert_array_less(np.abs(adists), max_dist)
|
|
|
|
agaps = np.diff(iline[0])
|
|
np.testing.assert_array_less(np.abs(agaps), max(gaps) + 0.1)
|
|
|
|
|
|
class TestArgrel:
|
|
|
|
def test_empty(self):
|
|
# Regression test for gh-2832.
|
|
# When there are no relative extrema, make sure that
|
|
# the number of empty arrays returned matches the
|
|
# dimension of the input.
|
|
|
|
empty_array = np.array([], dtype=int)
|
|
|
|
z1 = np.zeros(5)
|
|
|
|
i = argrelmin(z1)
|
|
assert_equal(len(i), 1)
|
|
assert_array_equal(i[0], empty_array)
|
|
|
|
z2 = np.zeros((3,5))
|
|
|
|
row, col = argrelmin(z2, axis=0)
|
|
assert_array_equal(row, empty_array)
|
|
assert_array_equal(col, empty_array)
|
|
|
|
row, col = argrelmin(z2, axis=1)
|
|
assert_array_equal(row, empty_array)
|
|
assert_array_equal(col, empty_array)
|
|
|
|
def test_basic(self):
|
|
# Note: the docstrings for the argrel{min,max,extrema} functions
|
|
# do not give a guarantee of the order of the indices, so we'll
|
|
# sort them before testing.
|
|
|
|
x = np.array([[1, 2, 2, 3, 2],
|
|
[2, 1, 2, 2, 3],
|
|
[3, 2, 1, 2, 2],
|
|
[2, 3, 2, 1, 2],
|
|
[1, 2, 3, 2, 1]])
|
|
|
|
row, col = argrelmax(x, axis=0)
|
|
order = np.argsort(row)
|
|
assert_equal(row[order], [1, 2, 3])
|
|
assert_equal(col[order], [4, 0, 1])
|
|
|
|
row, col = argrelmax(x, axis=1)
|
|
order = np.argsort(row)
|
|
assert_equal(row[order], [0, 3, 4])
|
|
assert_equal(col[order], [3, 1, 2])
|
|
|
|
row, col = argrelmin(x, axis=0)
|
|
order = np.argsort(row)
|
|
assert_equal(row[order], [1, 2, 3])
|
|
assert_equal(col[order], [1, 2, 3])
|
|
|
|
row, col = argrelmin(x, axis=1)
|
|
order = np.argsort(row)
|
|
assert_equal(row[order], [1, 2, 3])
|
|
assert_equal(col[order], [1, 2, 3])
|
|
|
|
def test_highorder(self):
|
|
order = 2
|
|
sigmas = [1.0, 2.0, 10.0, 5.0, 15.0]
|
|
test_data, act_locs = _gen_gaussians_even(sigmas, 500)
|
|
test_data[act_locs + order] = test_data[act_locs]*0.99999
|
|
test_data[act_locs - order] = test_data[act_locs]*0.99999
|
|
rel_max_locs = argrelmax(test_data, order=order, mode='clip')[0]
|
|
|
|
assert_(len(rel_max_locs) == len(act_locs))
|
|
assert_((rel_max_locs == act_locs).all())
|
|
|
|
def test_2d_gaussians(self):
|
|
sigmas = [1.0, 2.0, 10.0]
|
|
test_data, act_locs = _gen_gaussians_even(sigmas, 100)
|
|
rot_factor = 20
|
|
rot_range = np.arange(0, len(test_data)) - rot_factor
|
|
test_data_2 = np.vstack([test_data, test_data[rot_range]])
|
|
rel_max_rows, rel_max_cols = argrelmax(test_data_2, axis=1, order=1)
|
|
|
|
for rw in range(0, test_data_2.shape[0]):
|
|
inds = (rel_max_rows == rw)
|
|
|
|
assert_(len(rel_max_cols[inds]) == len(act_locs))
|
|
assert_((act_locs == (rel_max_cols[inds] - rot_factor*rw)).all())
|
|
|
|
|
|
class TestPeakProminences:
|
|
|
|
def test_empty(self):
|
|
"""
|
|
Test if an empty array is returned if no peaks are provided.
|
|
"""
|
|
out = peak_prominences([1, 2, 3], [])
|
|
for arr, dtype in zip(out, [np.float64, np.intp, np.intp]):
|
|
assert_(arr.size == 0)
|
|
assert_(arr.dtype == dtype)
|
|
|
|
out = peak_prominences([], [])
|
|
for arr, dtype in zip(out, [np.float64, np.intp, np.intp]):
|
|
assert_(arr.size == 0)
|
|
assert_(arr.dtype == dtype)
|
|
|
|
def test_basic(self):
|
|
"""
|
|
Test if height of prominences is correctly calculated in signal with
|
|
rising baseline (peak widths are 1 sample).
|
|
"""
|
|
# Prepare basic signal
|
|
x = np.array([-1, 1.2, 1.2, 1, 3.2, 1.3, 2.88, 2.1])
|
|
peaks = np.array([1, 2, 4, 6])
|
|
lbases = np.array([0, 0, 0, 5])
|
|
rbases = np.array([3, 3, 5, 7])
|
|
proms = x[peaks] - np.max([x[lbases], x[rbases]], axis=0)
|
|
# Test if calculation matches handcrafted result
|
|
out = peak_prominences(x, peaks)
|
|
assert_equal(out[0], proms)
|
|
assert_equal(out[1], lbases)
|
|
assert_equal(out[2], rbases)
|
|
|
|
def test_edge_cases(self):
|
|
"""
|
|
Test edge cases.
|
|
"""
|
|
# Peaks have same height, prominence and bases
|
|
x = [0, 2, 1, 2, 1, 2, 0]
|
|
peaks = [1, 3, 5]
|
|
proms, lbases, rbases = peak_prominences(x, peaks)
|
|
assert_equal(proms, [2, 2, 2])
|
|
assert_equal(lbases, [0, 0, 0])
|
|
assert_equal(rbases, [6, 6, 6])
|
|
|
|
# Peaks have same height & prominence but different bases
|
|
x = [0, 1, 0, 1, 0, 1, 0]
|
|
peaks = np.array([1, 3, 5])
|
|
proms, lbases, rbases = peak_prominences(x, peaks)
|
|
assert_equal(proms, [1, 1, 1])
|
|
assert_equal(lbases, peaks - 1)
|
|
assert_equal(rbases, peaks + 1)
|
|
|
|
def test_non_contiguous(self):
|
|
"""
|
|
Test with non-C-contiguous input arrays.
|
|
"""
|
|
x = np.repeat([-9, 9, 9, 0, 3, 1], 2)
|
|
peaks = np.repeat([1, 2, 4], 2)
|
|
proms, lbases, rbases = peak_prominences(x[::2], peaks[::2])
|
|
assert_equal(proms, [9, 9, 2])
|
|
assert_equal(lbases, [0, 0, 3])
|
|
assert_equal(rbases, [3, 3, 5])
|
|
|
|
def test_wlen(self):
|
|
"""
|
|
Test if wlen actually shrinks the evaluation range correctly.
|
|
"""
|
|
x = [0, 1, 2, 3, 1, 0, -1]
|
|
peak = [3]
|
|
# Test rounding behavior of wlen
|
|
assert_equal(peak_prominences(x, peak), [3., 0, 6])
|
|
for wlen, i in [(8, 0), (7, 0), (6, 0), (5, 1), (3.2, 1), (3, 2), (1.1, 2)]:
|
|
assert_equal(peak_prominences(x, peak, wlen), [3. - i, 0 + i, 6 - i])
|
|
|
|
def test_exceptions(self):
|
|
"""
|
|
Verify that exceptions and warnings are raised.
|
|
"""
|
|
# x with dimension > 1
|
|
with raises(ValueError, match='1-D array'):
|
|
peak_prominences([[0, 1, 1, 0]], [1, 2])
|
|
# peaks with dimension > 1
|
|
with raises(ValueError, match='1-D array'):
|
|
peak_prominences([0, 1, 1, 0], [[1, 2]])
|
|
# x with dimension < 1
|
|
with raises(ValueError, match='1-D array'):
|
|
peak_prominences(3, [0,])
|
|
|
|
# empty x with supplied
|
|
with raises(ValueError, match='not a valid index'):
|
|
peak_prominences([], [0])
|
|
# invalid indices with non-empty x
|
|
for p in [-100, -1, 3, 1000]:
|
|
with raises(ValueError, match='not a valid index'):
|
|
peak_prominences([1, 0, 2], [p])
|
|
|
|
# peaks is not cast-able to np.intp
|
|
with raises(TypeError, match='cannot safely cast'):
|
|
peak_prominences([0, 1, 1, 0], [1.1, 2.3])
|
|
|
|
# wlen < 3
|
|
with raises(ValueError, match='wlen'):
|
|
peak_prominences(np.arange(10), [3, 5], wlen=1)
|
|
|
|
def test_warnings(self):
|
|
"""
|
|
Verify that appropriate warnings are raised.
|
|
"""
|
|
msg = "some peaks have a prominence of 0"
|
|
for p in [0, 1, 2]:
|
|
with warns(PeakPropertyWarning, match=msg):
|
|
peak_prominences([1, 0, 2], [p,])
|
|
with warns(PeakPropertyWarning, match=msg):
|
|
peak_prominences([0, 1, 1, 1, 0], [2], wlen=2)
|
|
|
|
|
|
class TestPeakWidths:
|
|
|
|
def test_empty(self):
|
|
"""
|
|
Test if an empty array is returned if no peaks are provided.
|
|
"""
|
|
widths = peak_widths([], [])[0]
|
|
assert_(isinstance(widths, np.ndarray))
|
|
assert_equal(widths.size, 0)
|
|
widths = peak_widths([1, 2, 3], [])[0]
|
|
assert_(isinstance(widths, np.ndarray))
|
|
assert_equal(widths.size, 0)
|
|
out = peak_widths([], [])
|
|
for arr in out:
|
|
assert_(isinstance(arr, np.ndarray))
|
|
assert_equal(arr.size, 0)
|
|
|
|
@pytest.mark.filterwarnings("ignore:some peaks have a width of 0")
|
|
def test_basic(self):
|
|
"""
|
|
Test a simple use case with easy to verify results at different relative
|
|
heights.
|
|
"""
|
|
x = np.array([1, 0, 1, 2, 1, 0, -1])
|
|
prominence = 2
|
|
for rel_height, width_true, lip_true, rip_true in [
|
|
(0., 0., 3., 3.), # raises warning
|
|
(0.25, 1., 2.5, 3.5),
|
|
(0.5, 2., 2., 4.),
|
|
(0.75, 3., 1.5, 4.5),
|
|
(1., 4., 1., 5.),
|
|
(2., 5., 1., 6.),
|
|
(3., 5., 1., 6.)
|
|
]:
|
|
width_calc, height, lip_calc, rip_calc = peak_widths(
|
|
x, [3], rel_height)
|
|
assert_allclose(width_calc, width_true)
|
|
assert_allclose(height, 2 - rel_height * prominence)
|
|
assert_allclose(lip_calc, lip_true)
|
|
assert_allclose(rip_calc, rip_true)
|
|
|
|
def test_non_contiguous(self):
|
|
"""
|
|
Test with non-C-contiguous input arrays.
|
|
"""
|
|
x = np.repeat([0, 100, 50], 4)
|
|
peaks = np.repeat([1], 3)
|
|
result = peak_widths(x[::4], peaks[::3])
|
|
assert_equal(result, [0.75, 75, 0.75, 1.5])
|
|
|
|
def test_exceptions(self):
|
|
"""
|
|
Verify that argument validation works as intended.
|
|
"""
|
|
with raises(ValueError, match='1-D array'):
|
|
# x with dimension > 1
|
|
peak_widths(np.zeros((3, 4)), np.ones(3))
|
|
with raises(ValueError, match='1-D array'):
|
|
# x with dimension < 1
|
|
peak_widths(3, [0])
|
|
with raises(ValueError, match='1-D array'):
|
|
# peaks with dimension > 1
|
|
peak_widths(np.arange(10), np.ones((3, 2), dtype=np.intp))
|
|
with raises(ValueError, match='1-D array'):
|
|
# peaks with dimension < 1
|
|
peak_widths(np.arange(10), 3)
|
|
with raises(ValueError, match='not a valid index'):
|
|
# peak pos exceeds x.size
|
|
peak_widths(np.arange(10), [8, 11])
|
|
with raises(ValueError, match='not a valid index'):
|
|
# empty x with peaks supplied
|
|
peak_widths([], [1, 2])
|
|
with raises(TypeError, match='cannot safely cast'):
|
|
# peak cannot be safely casted to intp
|
|
peak_widths(np.arange(10), [1.1, 2.3])
|
|
with raises(ValueError, match='rel_height'):
|
|
# rel_height is < 0
|
|
peak_widths([0, 1, 0, 1, 0], [1, 3], rel_height=-1)
|
|
with raises(TypeError, match='None'):
|
|
# prominence data contains None
|
|
peak_widths([1, 2, 1], [1], prominence_data=(None, None, None))
|
|
|
|
def test_warnings(self):
|
|
"""
|
|
Verify that appropriate warnings are raised.
|
|
"""
|
|
msg = "some peaks have a width of 0"
|
|
with warns(PeakPropertyWarning, match=msg):
|
|
# Case: rel_height is 0
|
|
peak_widths([0, 1, 0], [1], rel_height=0)
|
|
with warns(PeakPropertyWarning, match=msg):
|
|
# Case: prominence is 0 and bases are identical
|
|
peak_widths(
|
|
[0, 1, 1, 1, 0], [2],
|
|
prominence_data=(np.array([0.], np.float64),
|
|
np.array([2], np.intp),
|
|
np.array([2], np.intp))
|
|
)
|
|
|
|
def test_mismatching_prominence_data(self):
|
|
"""Test with mismatching peak and / or prominence data."""
|
|
x = [0, 1, 0]
|
|
peak = [1]
|
|
for i, (prominences, left_bases, right_bases) in enumerate([
|
|
((1.,), (-1,), (2,)), # left base not in x
|
|
((1.,), (0,), (3,)), # right base not in x
|
|
((1.,), (2,), (0,)), # swapped bases same as peak
|
|
((1., 1.), (0, 0), (2, 2)), # array shapes don't match peaks
|
|
((1., 1.), (0,), (2,)), # arrays with different shapes
|
|
((1.,), (0, 0), (2,)), # arrays with different shapes
|
|
((1.,), (0,), (2, 2)) # arrays with different shapes
|
|
]):
|
|
# Make sure input is matches output of signal.peak_prominences
|
|
prominence_data = (np.array(prominences, dtype=np.float64),
|
|
np.array(left_bases, dtype=np.intp),
|
|
np.array(right_bases, dtype=np.intp))
|
|
# Test for correct exception
|
|
if i < 3:
|
|
match = "prominence data is invalid for peak"
|
|
else:
|
|
match = "arrays in `prominence_data` must have the same shape"
|
|
with raises(ValueError, match=match):
|
|
peak_widths(x, peak, prominence_data=prominence_data)
|
|
|
|
@pytest.mark.filterwarnings("ignore:some peaks have a width of 0")
|
|
def test_intersection_rules(self):
|
|
"""Test if x == eval_height counts as an intersection."""
|
|
# Flatt peak with two possible intersection points if evaluated at 1
|
|
x = [0, 1, 2, 1, 3, 3, 3, 1, 2, 1, 0]
|
|
# relative height is 0 -> width is 0 as well, raises warning
|
|
assert_allclose(peak_widths(x, peaks=[5], rel_height=0),
|
|
[(0.,), (3.,), (5.,), (5.,)])
|
|
# width_height == x counts as intersection -> nearest 1 is chosen
|
|
assert_allclose(peak_widths(x, peaks=[5], rel_height=2/3),
|
|
[(4.,), (1.,), (3.,), (7.,)])
|
|
|
|
|
|
def test_unpack_condition_args():
|
|
"""
|
|
Verify parsing of condition arguments for `scipy.signal.find_peaks` function.
|
|
"""
|
|
x = np.arange(10)
|
|
amin_true = x
|
|
amax_true = amin_true + 10
|
|
peaks = amin_true[1::2]
|
|
|
|
# Test unpacking with None or interval
|
|
assert_((None, None) == _unpack_condition_args((None, None), x, peaks))
|
|
assert_((1, None) == _unpack_condition_args(1, x, peaks))
|
|
assert_((1, None) == _unpack_condition_args((1, None), x, peaks))
|
|
assert_((None, 2) == _unpack_condition_args((None, 2), x, peaks))
|
|
assert_((3., 4.5) == _unpack_condition_args((3., 4.5), x, peaks))
|
|
|
|
# Test if borders are correctly reduced with `peaks`
|
|
amin_calc, amax_calc = _unpack_condition_args((amin_true, amax_true), x, peaks)
|
|
assert_equal(amin_calc, amin_true[peaks])
|
|
assert_equal(amax_calc, amax_true[peaks])
|
|
|
|
# Test raises if array borders don't match x
|
|
with raises(ValueError, match="array size of lower"):
|
|
_unpack_condition_args(amin_true, np.arange(11), peaks)
|
|
with raises(ValueError, match="array size of upper"):
|
|
_unpack_condition_args((None, amin_true), np.arange(11), peaks)
|
|
|
|
|
|
class TestFindPeaks:
|
|
|
|
# Keys of optionally returned properties
|
|
property_keys = {'peak_heights', 'left_thresholds', 'right_thresholds',
|
|
'prominences', 'left_bases', 'right_bases', 'widths',
|
|
'width_heights', 'left_ips', 'right_ips'}
|
|
|
|
def test_constant(self):
|
|
"""
|
|
Test behavior for signal without local maxima.
|
|
"""
|
|
open_interval = (None, None)
|
|
peaks, props = find_peaks(np.ones(10),
|
|
height=open_interval, threshold=open_interval,
|
|
prominence=open_interval, width=open_interval)
|
|
assert_(peaks.size == 0)
|
|
for key in self.property_keys:
|
|
assert_(props[key].size == 0)
|
|
|
|
def test_plateau_size(self):
|
|
"""
|
|
Test plateau size condition for peaks.
|
|
"""
|
|
# Prepare signal with peaks with peak_height == plateau_size
|
|
plateau_sizes = np.array([1, 2, 3, 4, 8, 20, 111])
|
|
x = np.zeros(plateau_sizes.size * 2 + 1)
|
|
x[1::2] = plateau_sizes
|
|
repeats = np.ones(x.size, dtype=int)
|
|
repeats[1::2] = x[1::2]
|
|
x = np.repeat(x, repeats)
|
|
|
|
# Test full output
|
|
peaks, props = find_peaks(x, plateau_size=(None, None))
|
|
assert_equal(peaks, [1, 3, 7, 11, 18, 33, 100])
|
|
assert_equal(props["plateau_sizes"], plateau_sizes)
|
|
assert_equal(props["left_edges"], peaks - (plateau_sizes - 1) // 2)
|
|
assert_equal(props["right_edges"], peaks + plateau_sizes // 2)
|
|
|
|
# Test conditions
|
|
assert_equal(find_peaks(x, plateau_size=4)[0], [11, 18, 33, 100])
|
|
assert_equal(find_peaks(x, plateau_size=(None, 3.5))[0], [1, 3, 7])
|
|
assert_equal(find_peaks(x, plateau_size=(5, 50))[0], [18, 33])
|
|
|
|
def test_height_condition(self):
|
|
"""
|
|
Test height condition for peaks.
|
|
"""
|
|
x = (0., 1/3, 0., 2.5, 0, 4., 0)
|
|
peaks, props = find_peaks(x, height=(None, None))
|
|
assert_equal(peaks, np.array([1, 3, 5]))
|
|
assert_equal(props['peak_heights'], np.array([1/3, 2.5, 4.]))
|
|
assert_equal(find_peaks(x, height=0.5)[0], np.array([3, 5]))
|
|
assert_equal(find_peaks(x, height=(None, 3))[0], np.array([1, 3]))
|
|
assert_equal(find_peaks(x, height=(2, 3))[0], np.array([3]))
|
|
|
|
def test_threshold_condition(self):
|
|
"""
|
|
Test threshold condition for peaks.
|
|
"""
|
|
x = (0, 2, 1, 4, -1)
|
|
peaks, props = find_peaks(x, threshold=(None, None))
|
|
assert_equal(peaks, np.array([1, 3]))
|
|
assert_equal(props['left_thresholds'], np.array([2, 3]))
|
|
assert_equal(props['right_thresholds'], np.array([1, 5]))
|
|
assert_equal(find_peaks(x, threshold=2)[0], np.array([3]))
|
|
assert_equal(find_peaks(x, threshold=3.5)[0], np.array([]))
|
|
assert_equal(find_peaks(x, threshold=(None, 5))[0], np.array([1, 3]))
|
|
assert_equal(find_peaks(x, threshold=(None, 4))[0], np.array([1]))
|
|
assert_equal(find_peaks(x, threshold=(2, 4))[0], np.array([]))
|
|
|
|
def test_distance_condition(self):
|
|
"""
|
|
Test distance condition for peaks.
|
|
"""
|
|
# Peaks of different height with constant distance 3
|
|
peaks_all = np.arange(1, 21, 3)
|
|
x = np.zeros(21)
|
|
x[peaks_all] += np.linspace(1, 2, peaks_all.size)
|
|
|
|
# Test if peaks with "minimal" distance are still selected (distance = 3)
|
|
assert_equal(find_peaks(x, distance=3)[0], peaks_all)
|
|
|
|
# Select every second peak (distance > 3)
|
|
peaks_subset = find_peaks(x, distance=3.0001)[0]
|
|
# Test if peaks_subset is subset of peaks_all
|
|
assert_(
|
|
np.setdiff1d(peaks_subset, peaks_all, assume_unique=True).size == 0
|
|
)
|
|
# Test if every second peak was removed
|
|
assert_equal(np.diff(peaks_subset), 6)
|
|
|
|
# Test priority of peak removal
|
|
x = [-2, 1, -1, 0, -3]
|
|
peaks_subset = find_peaks(x, distance=10)[0] # use distance > x size
|
|
assert_(peaks_subset.size == 1 and peaks_subset[0] == 1)
|
|
|
|
def test_prominence_condition(self):
|
|
"""
|
|
Test prominence condition for peaks.
|
|
"""
|
|
x = np.linspace(0, 10, 100)
|
|
peaks_true = np.arange(1, 99, 2)
|
|
offset = np.linspace(1, 10, peaks_true.size)
|
|
x[peaks_true] += offset
|
|
prominences = x[peaks_true] - x[peaks_true + 1]
|
|
interval = (3, 9)
|
|
keep = np.nonzero(
|
|
(interval[0] <= prominences) & (prominences <= interval[1]))
|
|
|
|
peaks_calc, properties = find_peaks(x, prominence=interval)
|
|
assert_equal(peaks_calc, peaks_true[keep])
|
|
assert_equal(properties['prominences'], prominences[keep])
|
|
assert_equal(properties['left_bases'], 0)
|
|
assert_equal(properties['right_bases'], peaks_true[keep] + 1)
|
|
|
|
def test_width_condition(self):
|
|
"""
|
|
Test width condition for peaks.
|
|
"""
|
|
x = np.array([1, 0, 1, 2, 1, 0, -1, 4, 0])
|
|
peaks, props = find_peaks(x, width=(None, 2), rel_height=0.75)
|
|
assert_equal(peaks.size, 1)
|
|
assert_equal(peaks, 7)
|
|
assert_allclose(props['widths'], 1.35)
|
|
assert_allclose(props['width_heights'], 1.)
|
|
assert_allclose(props['left_ips'], 6.4)
|
|
assert_allclose(props['right_ips'], 7.75)
|
|
|
|
def test_properties(self):
|
|
"""
|
|
Test returned properties.
|
|
"""
|
|
open_interval = (None, None)
|
|
x = [0, 1, 0, 2, 1.5, 0, 3, 0, 5, 9]
|
|
peaks, props = find_peaks(x,
|
|
height=open_interval, threshold=open_interval,
|
|
prominence=open_interval, width=open_interval)
|
|
assert_(len(props) == len(self.property_keys))
|
|
for key in self.property_keys:
|
|
assert_(peaks.size == props[key].size)
|
|
|
|
def test_raises(self):
|
|
"""
|
|
Test exceptions raised by function.
|
|
"""
|
|
with raises(ValueError, match="1-D array"):
|
|
find_peaks(np.array(1))
|
|
with raises(ValueError, match="1-D array"):
|
|
find_peaks(np.ones((2, 2)))
|
|
with raises(ValueError, match="distance"):
|
|
find_peaks(np.arange(10), distance=-1)
|
|
|
|
@pytest.mark.filterwarnings("ignore:some peaks have a prominence of 0",
|
|
"ignore:some peaks have a width of 0")
|
|
def test_wlen_smaller_plateau(self):
|
|
"""
|
|
Test behavior of prominence and width calculation if the given window
|
|
length is smaller than a peak's plateau size.
|
|
|
|
Regression test for gh-9110.
|
|
"""
|
|
peaks, props = find_peaks([0, 1, 1, 1, 0], prominence=(None, None),
|
|
width=(None, None), wlen=2)
|
|
assert_equal(peaks, 2)
|
|
assert_equal(props["prominences"], 0)
|
|
assert_equal(props["widths"], 0)
|
|
assert_equal(props["width_heights"], 1)
|
|
for key in ("left_bases", "right_bases", "left_ips", "right_ips"):
|
|
assert_equal(props[key], peaks)
|
|
|
|
@pytest.mark.parametrize("kwargs", [
|
|
{},
|
|
{"distance": 3.0},
|
|
{"prominence": (None, None)},
|
|
{"width": (None, 2)},
|
|
|
|
])
|
|
def test_readonly_array(self, kwargs):
|
|
"""
|
|
Test readonly arrays are accepted.
|
|
"""
|
|
x = np.linspace(0, 10, 15)
|
|
x_readonly = x.copy()
|
|
x_readonly.flags.writeable = False
|
|
|
|
peaks, _ = find_peaks(x)
|
|
peaks_readonly, _ = find_peaks(x_readonly, **kwargs)
|
|
|
|
assert_allclose(peaks, peaks_readonly)
|
|
|
|
|
|
class TestFindPeaksCwt:
|
|
|
|
def test_find_peaks_exact(self):
|
|
"""
|
|
Generate a series of gaussians and attempt to find the peak locations.
|
|
"""
|
|
sigmas = [5.0, 3.0, 10.0, 20.0, 10.0, 50.0]
|
|
num_points = 500
|
|
test_data, act_locs = _gen_gaussians_even(sigmas, num_points)
|
|
widths = np.arange(0.1, max(sigmas))
|
|
found_locs = find_peaks_cwt(test_data, widths, gap_thresh=2, min_snr=0,
|
|
min_length=None)
|
|
np.testing.assert_array_equal(found_locs, act_locs,
|
|
"Found maximum locations did not equal those expected")
|
|
|
|
def test_find_peaks_withnoise(self):
|
|
"""
|
|
Verify that peak locations are (approximately) found
|
|
for a series of gaussians with added noise.
|
|
"""
|
|
sigmas = [5.0, 3.0, 10.0, 20.0, 10.0, 50.0]
|
|
num_points = 500
|
|
test_data, act_locs = _gen_gaussians_even(sigmas, num_points)
|
|
widths = np.arange(0.1, max(sigmas))
|
|
noise_amp = 0.07
|
|
np.random.seed(18181911)
|
|
test_data += (np.random.rand(num_points) - 0.5)*(2*noise_amp)
|
|
found_locs = find_peaks_cwt(test_data, widths, min_length=15,
|
|
gap_thresh=1, min_snr=noise_amp / 5)
|
|
|
|
np.testing.assert_equal(len(found_locs), len(act_locs), 'Different number' +
|
|
'of peaks found than expected')
|
|
diffs = np.abs(found_locs - act_locs)
|
|
max_diffs = np.array(sigmas) / 5
|
|
np.testing.assert_array_less(diffs, max_diffs, 'Maximum location differed' +
|
|
'by more than %s' % (max_diffs))
|
|
|
|
def test_find_peaks_nopeak(self):
|
|
"""
|
|
Verify that no peak is found in
|
|
data that's just noise.
|
|
"""
|
|
noise_amp = 1.0
|
|
num_points = 100
|
|
np.random.seed(181819141)
|
|
test_data = (np.random.rand(num_points) - 0.5)*(2*noise_amp)
|
|
widths = np.arange(10, 50)
|
|
found_locs = find_peaks_cwt(test_data, widths, min_snr=5, noise_perc=30)
|
|
np.testing.assert_equal(len(found_locs), 0)
|
|
|
|
def test_find_peaks_with_non_default_wavelets(self):
|
|
x = gaussian(200, 2)
|
|
widths = np.array([1, 2, 3, 4])
|
|
a = find_peaks_cwt(x, widths, wavelet=gaussian)
|
|
|
|
np.testing.assert_equal(np.array([100]), a)
|
|
|
|
def test_find_peaks_window_size(self):
|
|
"""
|
|
Verify that window_size is passed correctly to private function and
|
|
affects the result.
|
|
"""
|
|
sigmas = [2.0, 2.0]
|
|
num_points = 1000
|
|
test_data, act_locs = _gen_gaussians_even(sigmas, num_points)
|
|
widths = np.arange(0.1, max(sigmas), 0.2)
|
|
noise_amp = 0.05
|
|
np.random.seed(18181911)
|
|
test_data += (np.random.rand(num_points) - 0.5)*(2*noise_amp)
|
|
|
|
# Possibly contrived negative region to throw off peak finding
|
|
# when window_size is too large
|
|
test_data[250:320] -= 1
|
|
|
|
found_locs = find_peaks_cwt(test_data, widths, gap_thresh=2, min_snr=3,
|
|
min_length=None, window_size=None)
|
|
with pytest.raises(AssertionError):
|
|
assert found_locs.size == act_locs.size
|
|
|
|
found_locs = find_peaks_cwt(test_data, widths, gap_thresh=2, min_snr=3,
|
|
min_length=None, window_size=20)
|
|
assert found_locs.size == act_locs.size
|
|
|
|
def test_find_peaks_with_one_width(self):
|
|
"""
|
|
Verify that the `width` argument
|
|
in `find_peaks_cwt` can be a float
|
|
"""
|
|
xs = np.arange(0, np.pi, 0.05)
|
|
test_data = np.sin(xs)
|
|
widths = 1
|
|
found_locs = find_peaks_cwt(test_data, widths)
|
|
|
|
np.testing.assert_equal(found_locs, 32)
|