def test_input_wrong_x_y():
"""Test the down sampling with wrong input types for x or/and y"""
x = 1
y = np.array([True] * ARRAY_SIZE, dtype=np.bool)
with pytest.raises(TypeError):
lttbc.downsample(x, y, THRESHOLD)
x = np.array([True] * ARRAY_SIZE, dtype=np.bool)
y = 4
with pytest.raises(TypeError):
lttbc.downsample(x, y, THRESHOLD)
x = "wrong"
y = np.array([True] * ARRAY_SIZE, dtype=np.bool)
with pytest.raises(TypeError):
lttbc.downsample(x, y, THRESHOLD)
x = np.array([True] * ARRAY_SIZE, dtype=np.bool)
y = "wrong"
with pytest.raises(TypeError):
lttbc.downsample(x, y, THRESHOLD)
x = 1
y = "wrong"
with pytest.raises(TypeError):
lttbc.downsample(x, y, THRESHOLD)
def test_single_dimension_validation():
"""Test that the downsample algorithm rejects arrays with multiple dims"""
x = np.array([[0., 0.], [1., 0.8], [0.9, 0.8], [0.9, 0.7], [0.9, 0.6],
[0.8, 0.5], [0.8, 0.5], [0.7, 0.5], [0.1, 0.], [0., 0.]],
dtype=np.double)
assert x.shape == (10, 2)
assert x.ndim == 2
y = np.array([True] * ARRAY_SIZE, dtype=np.bool)
with pytest.raises(ValueError):
lttbc.downsample(x, y, THRESHOLD)
def test_single_inf():
"""Test single 'inf' input for down sampling
XXX: Apparently infinite values provide a crappy result...
"""
x = np.arange(20, dtype='int32')
y = np.array([
0.0, 1.0, 2.0, np.inf, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0,
13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0
],
dtype=np.float)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, 10)
assert len(nx) == 10
assert len(ny) == 10
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array(
[0., 1., np.inf, np.inf, np.inf, np.inf, np.inf, np.inf, np.inf, 19.],
dtype=np.float)
np.testing.assert_array_almost_equal(ny, test_array)
def test_array_size():
"""Test the input failure for different dimensions of arrays"""
x = np.arange(ARRAY_SIZE)
y = np.random.randint(1000, size=ARRAY_SIZE - 1, dtype='uint64')
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
with pytest.raises(ValueError):
assert lttbc.downsample(x, y, ARRAY_SIZE)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
def test_downsample_uint64():
"""Test the base down sampling of the module"""
x = np.arange(ARRAY_SIZE, dtype='int32')
y = np.random.randint(1000, size=ARRAY_SIZE, dtype='uint64')
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, THRESHOLD)
assert len(nx) == THRESHOLD
assert len(ny) == THRESHOLD
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
def test_negative_threshold():
"""Test if a negative threshold provides problems"""
x = np.arange(ARRAY_SIZE, dtype='int32')
y = np.random.randint(1000, size=ARRAY_SIZE, dtype='uint64')
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, -THRESHOLD)
assert len(nx) == ARRAY_SIZE
assert len(ny) == ARRAY_SIZE
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
np.testing.assert_array_almost_equal(ny, y)
def test_nan():
"""Test the down sampling with NaN types"""
x = np.arange(ARRAY_SIZE, dtype='int32')
y = np.array([np.nan] * ARRAY_SIZE, dtype=np.float)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, THRESHOLD)
assert len(nx) == THRESHOLD
assert len(ny) == THRESHOLD
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array([0.0] * THRESHOLD, dtype=np.float)
np.testing.assert_array_almost_equal(ny, test_array)
def test_input_list():
"""Test the down sampling with lists types"""
x = [value for value in range(ARRAY_SIZE)]
y = [True] * ARRAY_SIZE
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, THRESHOLD)
assert len(nx) == THRESHOLD
assert len(ny) == THRESHOLD
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array([1.0] * THRESHOLD, dtype=np.float)
test_array_bool = np.array([1.0] * THRESHOLD, dtype=np.bool)
np.testing.assert_array_almost_equal(ny, test_array)
np.testing.assert_array_almost_equal(ny, test_array_bool)
def test_downsample_bool():
"""Test the down sampling with boolean types"""
x = np.arange(ARRAY_SIZE, dtype='int32')
y = np.array([True] * ARRAY_SIZE, dtype=np.bool)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, THRESHOLD)
assert len(nx) == THRESHOLD
assert len(ny) == THRESHOLD
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array([1.0] * THRESHOLD, dtype=np.float)
test_array_bool = np.array([1.0] * THRESHOLD, dtype=np.bool)
np.testing.assert_array_almost_equal(ny, test_array)
np.testing.assert_array_almost_equal(ny, test_array_bool)
def test_threshold_larger():
"""Test if a larger threshold provides problems"""
x = np.arange(ARRAY_SIZE, dtype='int32')
y = np.random.randint(1000, size=ARRAY_SIZE, dtype='uint64')
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
# Will return the arrays!
nx, ny = lttbc.downsample(x, y, ARRAY_SIZE + 1)
assert len(nx) == ARRAY_SIZE
assert len(ny) == ARRAY_SIZE
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
# NOTE: Known feature, we return double arrays ...
np.testing.assert_array_almost_equal(nx, x)
np.testing.assert_array_almost_equal(ny, y)
def test_single_nan():
"""Test single 'nan' input for down sampling"""
x = np.arange(20, dtype='int32')
y = np.array([
0.0, 1.0, 2.0, np.nan, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0,
13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0
],
dtype=np.float)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, 10)
assert len(nx) == 10
assert len(ny) == 10
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array([0., 0., 4., 5., 7., 10., 12., 14., 16., 19.],
dtype=np.float)
np.testing.assert_array_almost_equal(ny, test_array)
def lttb_downsample(x: np.ndarray,
y: np.ndarray,
max_pts: int = 5_000) -> Tuple[np.ndarray, np.ndarray]:
"""
Downsample x, y for visualisation
Parameters
----------
x : np.ndarray
x array
y : np.ndarray
y array
max_pts : int, optional
Maximum number of points after downsampling, by default 5000
Returns
-------
Tuple[np.ndarray, np.ndarray]
(new_x, new_y), the downsampled x and y arrays
Raises
------
ValueError
If the size of x does not match the size of y
"""
if x.size != y.size:
raise ValueError(f"x size {x.size} must equal y size {y.size}")
if max_pts >= x.size:
return x, y
x_dtype = x.dtype
y_dtype = y.dtype
nx, ny = lttbc.downsample(
x.astype(np.float32),
y.astype(np.float32),
max_pts,
)
return nx.astype(x_dtype), ny.astype(y_dtype)
def test_array_mix_inf_nan():
"""Test mix of problematic input 'inf' and 'nan'"""
x = np.arange(20, dtype='int32')
y = np.array([
0.0, 1.0, 2.0, np.nan, 4.0, 5.0, 6.0, np.nan, np.inf, np.inf, 10.0,
np.nan, 12.0, -np.inf, 14.0, 15.0, 16.0, 17.0, np.nan, 19.0
],
dtype=np.float)
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
nx, ny = lttbc.downsample(x, y, 10)
assert len(nx) == 10
assert len(ny) == 10
assert nx.dtype == np.double
assert ny.dtype == np.double
assert sys.getrefcount(x) == 2
assert sys.getrefcount(y) == 2
assert sys.getrefcount(nx) == 2
assert sys.getrefcount(ny) == 2
test_array = np.array(
[0., 0., 4., 4., 4., 10., -np.inf, -np.inf, -np.inf, 19.],
dtype=np.float)
np.testing.assert_array_almost_equal(ny, test_array)
def sample():
nx, ny = lttbc.downsample(x, y, LARGE_THRESHOLD)
return nx, ny