def test_shift_data_to_cycle():
ordinate = np.array([2.0, 1., 2., 3., 4., 5., 6.])
time = np.array([0.5, 1.0, 1.2, 1.35, 1.4, 1.5, 1.8])
arbitrary_cycle_start_time = 1.0
arbitrary_cycle_end_time = 1.5
new_cycle_start_time = 1.35
shifted_time, shifted_ordinate = pproc.shift_data_to_cycle(
arbitrary_cycle_start_time, arbitrary_cycle_end_time,
new_cycle_start_time,
time, ordinate)
testing.assert_allclose(shifted_ordinate, np.array([3, 4, np.nan, 1, 2]))
testing.assert_allclose(shifted_time,
np.array([0.0, 0.05, 0.15, 0.30, 0.5]))
testing.assert_equal(shifted_time.size, 5)
testing.assert_equal(shifted_ordinate.size, 5)
assert shifted_time[0] == 0.0
assert shifted_time[-1] == (arbitrary_cycle_end_time -
arbitrary_cycle_start_time)
# TODO add tests where the 3 specified times do not exactly match those
# in the time array.
arbitrary_cycle_start_time = 0.9
arbitrary_cycle_end_time = 1.6
new_cycle_start_time = 1.35
shifted_time, shifted_ordinate = pproc.shift_data_to_cycle(
arbitrary_cycle_start_time, arbitrary_cycle_end_time,
new_cycle_start_time,
time, ordinate)
testing.assert_allclose(shifted_ordinate, np.array([3, 4, np.nan, 1, 2]))
testing.assert_allclose(shifted_time,
np.array([0.0, 0.05, 0.25, 0.40, 0.7]))
testing.assert_equal(shifted_time.size, 5)
testing.assert_equal(shifted_ordinate.size, 5)
# Don't actually want this to be true.
testing.assert_equal(shifted_time[0], 0.0)
testing.assert_equal(shifted_time[-1] - shifted_time[0],
(arbitrary_cycle_end_time - arbitrary_cycle_start_time))
def test_shift_data_to_cycle():
ordinate = np.array([2.0, 1., 2., 3., 4., 5., 6.])
time = np.array([0.5, 1.0, 1.2, 1.35, 1.4, 1.5, 1.8])
arbitrary_cycle_start_time = 1.0
arbitrary_cycle_end_time = 1.5
new_cycle_start_time = 1.35
shifted_time, shifted_ordinate = pproc.shift_data_to_cycle(
arbitrary_cycle_start_time, arbitrary_cycle_end_time,
new_cycle_start_time, time, ordinate)
testing.assert_allclose(shifted_ordinate, np.array([3, 4, np.nan, 1, 2]))
testing.assert_allclose(shifted_time,
np.array([0.0, 0.05, 0.15, 0.30, 0.5]))
testing.assert_equal(shifted_time.size, 5)
testing.assert_equal(shifted_ordinate.size, 5)
assert shifted_time[0] == 0.0
assert shifted_time[-1] == (arbitrary_cycle_end_time -
arbitrary_cycle_start_time)
# TODO add tests where the 3 specified times do not exactly match those
# in the time array.
arbitrary_cycle_start_time = 0.9
arbitrary_cycle_end_time = 1.6
new_cycle_start_time = 1.35
shifted_time, shifted_ordinate = pproc.shift_data_to_cycle(
arbitrary_cycle_start_time, arbitrary_cycle_end_time,
new_cycle_start_time, time, ordinate)
testing.assert_allclose(shifted_ordinate, np.array([3, 4, np.nan, 1, 2]))
testing.assert_allclose(shifted_time,
np.array([0.0, 0.05, 0.25, 0.40, 0.7]))
testing.assert_equal(shifted_time.size, 5)
testing.assert_equal(shifted_ordinate.size, 5)
# Don't actually want this to be true.
testing.assert_equal(shifted_time[0], 0.0)
testing.assert_equal(
shifted_time[-1] - shifted_time[0],
(arbitrary_cycle_end_time - arbitrary_cycle_start_time))
def test_shift_data_to_cycle_for_less_than_full_cycle():
# When [arbitrary_cycle_start_time, arbitrary_cycle_end_time] is NOT a
# subset of the time interval of the given data. That is, we don't have
# data for the complete gait cycle.
import pylab as pl
arbitrary_cycle_start_time = 0.2
arbitrary_cycle_end_time = 1.3
duration = arbitrary_cycle_end_time - arbitrary_cycle_start_time
# Case 1: data ends early.
data_starts_at = 0.1
data_ends_at = 0.8
x1 = np.linspace(data_starts_at, data_ends_at, 81)
x10 = x1[pproc.nearest_index(x1, arbitrary_cycle_start_time)]
y1 = 2 * (x1 - x10)
pl.subplot(2, 2, 1)
pl.plot(x1, y1)
pl.xlim(xmin=arbitrary_cycle_start_time, xmax=arbitrary_cycle_end_time)
pl.ylim(0, y1[-1])
# Case 1a: don't shift, just represent as percent gait cycle.
xs1a, ys1a = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time,
arbitrary_cycle_start_time, x1, y1)
# Don't expect to satisfy this in this case; see docstring:
# testing.assert_equal(xs1a[-1], duration)
pl.plot(xs1a, ys1a)
testing.assert_almost_equal(
pproc.percent_duration(xs1a, 0, duration)[-1], 54.545454545454)
pl.subplot(2, 2, 2)
pl.plot(pproc.percent_duration(xs1a, 0, duration), ys1a, label='1a')
pl.xlim(0, 100)
#pl.ylim(0, y1[-1])
# Case 1b: shift to a new cycle start time.
xs1b, ys1b = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.4, x1,
y1)
# When the gap is in the middle, we DO want to satisfy this:
testing.assert_almost_equal(xs1b[-1], duration)
pl.plot(pproc.percent_duration(xs1b, 0, duration), ys1b, label='1b')
# Case 1c: the new cycle start time is in the missing-data portion.
xs1c, ys1c = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 1.0, x1,
y1)
testing.assert_almost_equal(xs1c[-1] + 0.2, duration)
testing.assert_almost_equal(xs1c[0], arbitrary_cycle_end_time - 1.0)
# Increasing (last time value is the maximum value).
assert np.all(np.diff(xs1c) > 0)
pl.plot(pproc.percent_duration(xs1c, 0, duration), ys1c, label='1c')
pl.legend()
# Case 2: data starts late.
data_starts_at = 0.4
data_ends_at = 1.5
x2 = np.linspace(data_starts_at, data_ends_at, 111)
x20 = x2[pproc.nearest_index(x2, arbitrary_cycle_start_time)]
pl.subplot(2, 2, 3)
y2 = 2 * (x2 - x20)
pl.plot(x2, y2)
pl.xlim(xmin=arbitrary_cycle_start_time, xmax=arbitrary_cycle_end_time)
pl.ylim(ymin=0, ymax=1.8)
# Case 2a: don't shift.
xs2a, ys2a = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time,
arbitrary_cycle_start_time, x2, y2)
pl.plot(xs2a, ys2a)
testing.assert_equal(xs2a[-1], duration)
pl.subplot(2, 2, 4)
pl.plot(pproc.percent_duration(xs2a, 0, duration), ys2a, label='2a')
pl.xlim(0, 100)
pl.ylim(ymin=0, ymax=1.8)
# Case 2b: shift to a new cycle start time.
xs2b, ys2b = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.8, x2,
y2)
testing.assert_equal(xs2b[-1], duration)
pl.plot(pproc.percent_duration(xs2b, 0, duration), ys2b, label='2b')
# Case 2c: the new cycle start time is in the missing-data portion.
xs2c, ys2c = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.3, x2,
y2)
# Can't satisfy this when the gap is not in the middle:
#testing.assert_equal(xs2c[-1] - xs2c[0], duration)
pl.plot(pproc.percent_duration(xs2c, 0, duration), ys2c, label='2c')
pl.legend()
def test_shift_data_to_cycle_for_less_than_full_cycle():
# When [arbitrary_cycle_start_time, arbitrary_cycle_end_time] is NOT a
# subset of the time interval of the given data. That is, we don't have
# data for the complete gait cycle.
import pylab as pl
arbitrary_cycle_start_time = 0.2
arbitrary_cycle_end_time = 1.3
duration = arbitrary_cycle_end_time - arbitrary_cycle_start_time
# Case 1: data ends early.
data_starts_at = 0.1
data_ends_at = 0.8
x1 = np.linspace(data_starts_at, data_ends_at, 81)
x10 = x1[pproc.nearest_index(x1, arbitrary_cycle_start_time)]
y1 = 2 * (x1 - x10)
pl.subplot(2, 2, 1)
pl.plot(x1, y1)
pl.xlim(xmin=arbitrary_cycle_start_time, xmax=arbitrary_cycle_end_time)
pl.ylim(0, y1[-1])
# Case 1a: don't shift, just represent as percent gait cycle.
xs1a, ys1a = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, arbitrary_cycle_start_time, x1, y1)
# Don't expect to satisfy this in this case; see docstring:
# testing.assert_equal(xs1a[-1], duration)
pl.plot(xs1a, ys1a)
testing.assert_almost_equal(pproc.percent_duration(xs1a, 0, duration)[-1],
54.545454545454)
pl.subplot(2, 2, 2)
pl.plot(pproc.percent_duration(xs1a, 0, duration), ys1a, label='1a')
pl.xlim(0, 100)
#pl.ylim(0, y1[-1])
# Case 1b: shift to a new cycle start time.
xs1b, ys1b = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.4, x1, y1)
# When the gap is in the middle, we DO want to satisfy this:
testing.assert_almost_equal(xs1b[-1], duration)
pl.plot(pproc.percent_duration(xs1b, 0, duration), ys1b, label='1b')
# Case 1c: the new cycle start time is in the missing-data portion.
xs1c, ys1c = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 1.0, x1, y1)
testing.assert_almost_equal(xs1c[-1] + 0.2, duration)
testing.assert_almost_equal(xs1c[0], arbitrary_cycle_end_time - 1.0)
# Increasing (last time value is the maximum value).
assert np.all(np.diff(xs1c) > 0)
pl.plot(pproc.percent_duration(xs1c, 0, duration), ys1c, label='1c')
pl.legend()
# Case 2: data starts late.
data_starts_at = 0.4
data_ends_at = 1.5
x2 = np.linspace(data_starts_at, data_ends_at, 111)
x20 = x2[pproc.nearest_index(x2, arbitrary_cycle_start_time)]
pl.subplot(2, 2, 3)
y2 = 2 * (x2 - x20)
pl.plot(x2, y2)
pl.xlim(xmin=arbitrary_cycle_start_time, xmax=arbitrary_cycle_end_time)
pl.ylim(ymin=0, ymax=1.8)
# Case 2a: don't shift.
xs2a, ys2a = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, arbitrary_cycle_start_time, x2, y2)
pl.plot(xs2a, ys2a)
testing.assert_equal(xs2a[-1], duration)
pl.subplot(2, 2, 4)
pl.plot(pproc.percent_duration(xs2a, 0, duration), ys2a, label='2a')
pl.xlim(0, 100)
pl.ylim(ymin=0, ymax=1.8)
# Case 2b: shift to a new cycle start time.
xs2b, ys2b = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.8, x2, y2)
testing.assert_equal(xs2b[-1], duration)
pl.plot(pproc.percent_duration(xs2b, 0, duration), ys2b, label='2b')
# Case 2c: the new cycle start time is in the missing-data portion.
xs2c, ys2c = pproc.shift_data_to_cycle(arbitrary_cycle_start_time,
arbitrary_cycle_end_time, 0.3, x2, y2)
# Can't satisfy this when the gap is not in the middle:
#testing.assert_equal(xs2c[-1] - xs2c[0], duration)
pl.plot(pproc.percent_duration(xs2c, 0, duration), ys2c, label='2c')
pl.legend()
