def test_read_analogs(
usecols,
shape_val,
first_last_val,
mean_val,
median_val,
sum_val,
nans_val,
extension,
):
if extension == "csv":
data = Analogs.from_csv(**analogs_csv_kwargs, usecols=usecols)
decimal = 0
elif extension == "c3d":
data = Analogs.from_c3d(MARKERS_ANALOGS_C3D,
prefix_delimiter=".",
usecols=usecols)
decimal = 4
else:
raise ValueError("wrong extension provided")
if usecols and isinstance(usecols[0], str):
np.testing.assert_array_equal(x=data.channel, y=usecols)
is_expected_array(
data,
shape_val,
first_last_val,
mean_val,
median_val,
sum_val,
nans_val,
decimal=decimal,
)
def test_proc_center():
is_expected_array(MARKERS_DATA.meca.center(), **EXPECTED_VALUES[5])
is_expected_array(MARKERS_DATA.meca.center(MARKERS_DATA.isel(time=0)),
**EXPECTED_VALUES[6])
is_expected_array(ANALOGS_DATA.meca.center(), **EXPECTED_VALUES[7])
is_expected_array(ANALOGS_DATA.meca.center(mu=2), **EXPECTED_VALUES[8])
is_expected_array(ANALOGS_DATA.meca.center(ANALOGS_DATA.isel(time=0)),
**EXPECTED_VALUES[9])
def test_csv_without_header():
is_expected_array(
Analogs.from_csv(ANALOGS_CSV, first_row=5, first_column=2),
**EXPECTED_VALUES[59],
)
is_expected_array(
Markers.from_csv(MARKERS_CSV_WITHOUT_HEADER, first_column=2),
**EXPECTED_VALUES[58],
)
def test_read_xlsx():
is_expected_array(
Markers.from_excel(**{
**markers_csv_kwargs,
**dict(filename=MARKERS_XLSX)
}),
**EXPECTED_VALUES[60],
)
is_expected_array(
Analogs.from_excel(**{
**analogs_csv_kwargs,
**dict(filename=ANALOGS_XLSX)
}),
**EXPECTED_VALUES[61],
)
def test_proc_detect_outliers():
is_expected_array(MARKERS_DATA.meca.detect_outliers(threshold=3),
**EXPECTED_VALUES[51])
is_expected_array(MARKERS_DATA.meca.detect_outliers(threshold=1),
**EXPECTED_VALUES[52])
is_expected_array(ANALOGS_DATA.meca.detect_outliers(threshold=3),
**EXPECTED_VALUES[53])
is_expected_array(ANALOGS_DATA.meca.detect_outliers(threshold=1),
**EXPECTED_VALUES[54])
def test_proc_fft():
is_expected_array(ANALOGS_DATA.meca.fft(freq=ANALOGS_DATA.rate),
**EXPECTED_VALUES[40])
is_expected_array(
ANALOGS_DATA.meca.fft(freq=ANALOGS_DATA.rate, only_positive=False),
**EXPECTED_VALUES[41])
is_expected_array(MARKERS_DATA.meca.fft(freq=ANALOGS_DATA.rate),
**EXPECTED_VALUES[42])
is_expected_array(
MARKERS_DATA.meca.fft(freq=ANALOGS_DATA.rate, only_positive=False),
**EXPECTED_VALUES[43])
def test_proc_detect_onset():
m = MARKERS_DATA[0, 0, :]
r = xr.DataArray(m.meca.detect_onset(threshold=m.mean() + m.std()))
is_expected_array(r, **EXPECTED_VALUES[49])
r = xr.DataArray(
m.meca.detect_onset(threshold=m.mean(),
n_below=10,
threshold2=m.mean() + m.std()))
is_expected_array(r, **EXPECTED_VALUES[50])
np.testing.assert_array_equal(x=m.meca.detect_onset(threshold=m.mean() *
10),
y=0)
with pytest.raises(ValueError):
MARKERS_DATA[0, :, :].meca.detect_onset(threshold=0)
with pytest.raises(ValueError):
MARKERS_DATA[:, :, :].meca.detect_onset(threshold=0)
def test_proc_norm():
n_frames = 100
n_markers = 10
m = Markers(np.random.rand(3, n_markers, n_frames))
# norm by hand
expected_norm = np.linalg.norm(m[:3, ...], axis=0)
# norm with pyomeca
computed_norm = m.meca.norm(dim="axis")
np.testing.assert_almost_equal(computed_norm, expected_norm, decimal=10)
is_expected_array(MARKERS_DATA.meca.norm(dim="axis"),
**EXPECTED_VALUES[44])
is_expected_array(MARKERS_DATA.meca.norm(dim="channel"),
**EXPECTED_VALUES[45])
is_expected_array(MARKERS_DATA.meca.norm(dim="time"),
**EXPECTED_VALUES[46])
is_expected_array(ANALOGS_DATA.meca.norm(dim="channel"),
**EXPECTED_VALUES[47])
is_expected_array(ANALOGS_DATA.meca.norm(dim="time"),
**EXPECTED_VALUES[48])
def test_proc_time_normalize():
is_expected_array(MARKERS_DATA.meca.time_normalize(),
**EXPECTED_VALUES[26])
is_expected_array(MARKERS_DATA.meca.time_normalize(n_frames=1000),
**EXPECTED_VALUES[27])
time_vector = np.linspace(MARKERS_DATA.time[0], MARKERS_DATA.time[100],
100)
is_expected_array(
MARKERS_DATA.meca.time_normalize(time_vector=time_vector),
**EXPECTED_VALUES[28])
is_expected_array(
MARKERS_DATA.meca.time_normalize(norm_time=True).time,
**EXPECTED_VALUES[55])
is_expected_array(ANALOGS_DATA.meca.time_normalize(),
**EXPECTED_VALUES[29])
is_expected_array(ANALOGS_DATA.meca.time_normalize(n_frames=1000),
**EXPECTED_VALUES[30])
time_vector = np.linspace(ANALOGS_DATA.time[0], ANALOGS_DATA.time[100],
100)
is_expected_array(
ANALOGS_DATA.meca.time_normalize(time_vector=time_vector),
**EXPECTED_VALUES[31])
def test_proc_filters():
freq = ANALOGS_DATA.rate
order = 2
is_expected_array(
ANALOGS_DATA.meca.low_pass(order=order, cutoff=5, freq=freq),
**EXPECTED_VALUES[32],
)
is_expected_array(
ANALOGS_DATA.meca.low_pass(order=order, cutoff=5),
**EXPECTED_VALUES[32],
)
is_expected_array(
ANALOGS_DATA.meca.high_pass(order=order, cutoff=100),
**EXPECTED_VALUES[33],
)
is_expected_array(
ANALOGS_DATA.meca.band_pass(order=order, cutoff=[10, 200]),
**EXPECTED_VALUES[34],
)
is_expected_array(
ANALOGS_DATA.meca.band_stop(order=order, cutoff=[40, 60]),
**EXPECTED_VALUES[35],
)
freq = MARKERS_DATA.rate
is_expected_array(
MARKERS_DATA.meca.low_pass(freq=freq, order=order, cutoff=5),
**EXPECTED_VALUES[36],
)
is_expected_array(
MARKERS_DATA.meca.low_pass(order=order, cutoff=5),
**EXPECTED_VALUES[36],
)
is_expected_array(
MARKERS_DATA.meca.high_pass(order=order, cutoff=10),
**EXPECTED_VALUES[37],
)
is_expected_array(
MARKERS_DATA.meca.band_pass(order=order, cutoff=[1, 10]),
**EXPECTED_VALUES[38],
)
is_expected_array(
MARKERS_DATA.meca.band_stop(order=order, cutoff=[5, 6]),
**EXPECTED_VALUES[39],
)
with pytest.raises(ValueError):
Analogs.from_random_data().meca.band_stop(order=order, cutoff=[5, 6])
def test_proc_filters():
freq = ANALOGS_DATA.rate
order = 2
is_expected_array(
ANALOGS_DATA.meca.low_pass(freq=freq, order=order, cutoff=5),
**EXPECTED_VALUES[32],
)
is_expected_array(
ANALOGS_DATA.meca.high_pass(freq=freq, order=order, cutoff=100),
**EXPECTED_VALUES[33],
)
is_expected_array(
ANALOGS_DATA.meca.band_pass(freq=freq, order=order, cutoff=[10, 200]),
**EXPECTED_VALUES[34],
)
is_expected_array(
ANALOGS_DATA.meca.band_stop(freq=freq, order=order, cutoff=[40, 60]),
**EXPECTED_VALUES[35],
)
freq = MARKERS_DATA.rate
is_expected_array(
MARKERS_DATA.meca.low_pass(freq=freq, order=order, cutoff=5),
**EXPECTED_VALUES[36],
)
is_expected_array(
MARKERS_DATA.meca.high_pass(freq=freq, order=order, cutoff=10),
**EXPECTED_VALUES[37],
)
is_expected_array(
MARKERS_DATA.meca.band_pass(freq=freq, order=order, cutoff=[1, 10]),
**EXPECTED_VALUES[38],
)
is_expected_array(
MARKERS_DATA.meca.band_stop(freq=freq, order=order, cutoff=[5, 6]),
**EXPECTED_VALUES[39],
)
def test_proc_normalize():
is_expected_array(MARKERS_DATA.meca.normalize(), **EXPECTED_VALUES[20])
is_expected_array(MARKERS_DATA.meca.normalize(scale=1),
**EXPECTED_VALUES[21])
is_expected_array(
MARKERS_DATA.meca.normalize(ref=MARKERS_DATA.sel(time=5.76)),
**EXPECTED_VALUES[22])
is_expected_array(ANALOGS_DATA.meca.normalize(), **EXPECTED_VALUES[23])
is_expected_array(ANALOGS_DATA.meca.normalize(scale=1),
**EXPECTED_VALUES[24])
is_expected_array(
ANALOGS_DATA.meca.normalize(ref=ANALOGS_DATA.sel(time=5.76)),
**EXPECTED_VALUES[25])
def test_proc_abs():
is_expected_array(ANALOGS_DATA.meca.abs(), **EXPECTED_VALUES[1])
is_expected_array(MARKERS_DATA.meca.abs(), **EXPECTED_VALUES[2])
def test_proc_square_sqrt():
is_expected_array(MARKERS_DATA.meca.square().meca.sqrt(),
**EXPECTED_VALUES[3])
is_expected_array(ANALOGS_DATA.meca.square().meca.sqrt(),
**EXPECTED_VALUES[4])
