3RNN/Lib/site-packages/scipy/sparse/tests/test_common1d.py
2024-05-26 19:49:15 +02:00

458 lines
16 KiB
Python

"""Test of 1D aspects of sparse array classes"""
import pytest
import numpy as np
import scipy as sp
from scipy.sparse import (
bsr_array, csc_array, dia_array, lil_array,
)
from scipy.sparse._sputils import supported_dtypes, matrix
from scipy._lib._util import ComplexWarning
sup_complex = np.testing.suppress_warnings()
sup_complex.filter(ComplexWarning)
spcreators = [sp.sparse.coo_array, sp.sparse.dok_array]
math_dtypes = [np.int64, np.float64, np.complex128]
@pytest.fixture
def dat1d():
return np.array([3, 0, 1, 0], 'd')
@pytest.fixture
def datsp_math_dtypes(dat1d):
dat_dtypes = {dtype: dat1d.astype(dtype) for dtype in math_dtypes}
return {
sp: [(dtype, dat, sp(dat)) for dtype, dat in dat_dtypes.items()]
for sp in spcreators
}
# Test init with 1D dense input
# sparrays which do not plan to support 1D
@pytest.mark.parametrize("spcreator", [bsr_array, csc_array, dia_array, lil_array])
def test_no_1d_support_in_init(spcreator):
with pytest.raises(ValueError, match="arrays don't support 1D input"):
spcreator([0, 1, 2, 3])
# Main tests class
@pytest.mark.parametrize("spcreator", spcreators)
class TestCommon1D:
"""test common functionality shared by 1D sparse formats"""
def test_create_empty(self, spcreator):
assert np.array_equal(spcreator((3,)).toarray(), np.zeros(3))
assert np.array_equal(spcreator((3,)).nnz, 0)
assert np.array_equal(spcreator((3,)).count_nonzero(), 0)
def test_invalid_shapes(self, spcreator):
with pytest.raises(ValueError, match='elements cannot be negative'):
spcreator((-3,))
def test_repr(self, spcreator, dat1d):
repr(spcreator(dat1d))
def test_str(self, spcreator, dat1d):
str(spcreator(dat1d))
def test_neg(self, spcreator):
A = np.array([-1, 0, 17, 0, -5, 0, 1, -4, 0, 0, 0, 0], 'd')
assert np.array_equal(-A, (-spcreator(A)).toarray())
def test_1d_supported_init(self, spcreator):
A = spcreator([0, 1, 2, 3])
assert A.ndim == 1
def test_reshape_1d_tofrom_row_or_column(self, spcreator):
# add a dimension 1d->2d
x = spcreator([1, 0, 7, 0, 0, 0, 0, -3, 0, 0, 0, 5])
y = x.reshape(1, 12)
desired = [[1, 0, 7, 0, 0, 0, 0, -3, 0, 0, 0, 5]]
assert np.array_equal(y.toarray(), desired)
# remove a size-1 dimension 2d->1d
x = spcreator(desired)
y = x.reshape(12)
assert np.array_equal(y.toarray(), desired[0])
y2 = x.reshape((12,))
assert y.shape == y2.shape
# make a 2d column into 1d. 2d->1d
y = x.T.reshape(12)
assert np.array_equal(y.toarray(), desired[0])
def test_reshape(self, spcreator):
x = spcreator([1, 0, 7, 0, 0, 0, 0, -3, 0, 0, 0, 5])
y = x.reshape((4, 3))
desired = [[1, 0, 7], [0, 0, 0], [0, -3, 0], [0, 0, 5]]
assert np.array_equal(y.toarray(), desired)
y = x.reshape((12,))
assert y is x
y = x.reshape(12)
assert np.array_equal(y.toarray(), x.toarray())
def test_sum(self, spcreator):
np.random.seed(1234)
dat_1 = np.array([0, 1, 2, 3, -4, 5, -6, 7, 9])
dat_2 = np.random.rand(5)
dat_3 = np.array([])
dat_4 = np.zeros((40,))
arrays = [dat_1, dat_2, dat_3, dat_4]
for dat in arrays:
datsp = spcreator(dat)
with np.errstate(over='ignore'):
assert np.isscalar(datsp.sum())
assert np.allclose(dat.sum(), datsp.sum())
assert np.allclose(dat.sum(axis=None), datsp.sum(axis=None))
assert np.allclose(dat.sum(axis=0), datsp.sum(axis=0))
assert np.allclose(dat.sum(axis=-1), datsp.sum(axis=-1))
# test `out` parameter
datsp.sum(axis=0, out=np.zeros(()))
def test_sum_invalid_params(self, spcreator):
out = np.zeros((3,)) # wrong size for out
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
with pytest.raises(ValueError, match='axis must be None, -1 or 0'):
datsp.sum(axis=1)
with pytest.raises(TypeError, match='Tuples are not accepted'):
datsp.sum(axis=(0, 1))
with pytest.raises(TypeError, match='axis must be an integer'):
datsp.sum(axis=1.5)
with pytest.raises(ValueError, match='dimensions do not match'):
datsp.sum(axis=0, out=out)
def test_numpy_sum(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
dat_sum = np.sum(dat)
datsp_sum = np.sum(datsp)
assert np.allclose(dat_sum, datsp_sum)
def test_mean(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
assert np.allclose(dat.mean(), datsp.mean())
assert np.isscalar(datsp.mean(axis=None))
assert np.allclose(dat.mean(axis=None), datsp.mean(axis=None))
assert np.allclose(dat.mean(axis=0), datsp.mean(axis=0))
assert np.allclose(dat.mean(axis=-1), datsp.mean(axis=-1))
with pytest.raises(ValueError, match='axis'):
datsp.mean(axis=1)
with pytest.raises(ValueError, match='axis'):
datsp.mean(axis=-2)
def test_mean_invalid_params(self, spcreator):
out = np.asarray(np.zeros((1, 3)))
dat = np.array([[0, 1, 2], [3, -4, 5], [-6, 7, 9]])
if spcreator._format == 'uni':
with pytest.raises(ValueError, match='zq'):
spcreator(dat)
return
datsp = spcreator(dat)
with pytest.raises(ValueError, match='axis out of range'):
datsp.mean(axis=3)
with pytest.raises(TypeError, match='Tuples are not accepted'):
datsp.mean(axis=(0, 1))
with pytest.raises(TypeError, match='axis must be an integer'):
datsp.mean(axis=1.5)
with pytest.raises(ValueError, match='dimensions do not match'):
datsp.mean(axis=1, out=out)
def test_sum_dtype(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
for dtype in supported_dtypes:
dat_sum = dat.sum(dtype=dtype)
datsp_sum = datsp.sum(dtype=dtype)
assert np.allclose(dat_sum, datsp_sum)
assert np.array_equal(dat_sum.dtype, datsp_sum.dtype)
def test_mean_dtype(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
for dtype in supported_dtypes:
dat_mean = dat.mean(dtype=dtype)
datsp_mean = datsp.mean(dtype=dtype)
assert np.allclose(dat_mean, datsp_mean)
assert np.array_equal(dat_mean.dtype, datsp_mean.dtype)
def test_mean_out(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
dat_out = np.array([0])
datsp_out = np.array([0])
dat.mean(out=dat_out, keepdims=True)
datsp.mean(out=datsp_out)
assert np.allclose(dat_out, datsp_out)
dat.mean(axis=0, out=dat_out, keepdims=True)
datsp.mean(axis=0, out=datsp_out)
assert np.allclose(dat_out, datsp_out)
def test_numpy_mean(self, spcreator):
dat = np.array([0, 1, 2])
datsp = spcreator(dat)
dat_mean = np.mean(dat)
datsp_mean = np.mean(datsp)
assert np.allclose(dat_mean, datsp_mean)
assert np.array_equal(dat_mean.dtype, datsp_mean.dtype)
@sup_complex
def test_from_array(self, spcreator):
A = np.array([2, 3, 4])
assert np.array_equal(spcreator(A).toarray(), A)
A = np.array([1.0 + 3j, 0, -1])
assert np.array_equal(spcreator(A).toarray(), A)
assert np.array_equal(spcreator(A, dtype='int16').toarray(), A.astype('int16'))
@sup_complex
def test_from_list(self, spcreator):
A = [2, 3, 4]
assert np.array_equal(spcreator(A).toarray(), A)
A = [1.0 + 3j, 0, -1]
assert np.array_equal(spcreator(A).toarray(), np.array(A))
assert np.array_equal(
spcreator(A, dtype='int16').toarray(), np.array(A).astype('int16')
)
@sup_complex
def test_from_sparse(self, spcreator):
D = np.array([1, 0, 0])
S = sp.sparse.coo_array(D)
assert np.array_equal(spcreator(S).toarray(), D)
S = spcreator(D)
assert np.array_equal(spcreator(S).toarray(), D)
D = np.array([1.0 + 3j, 0, -1])
S = sp.sparse.coo_array(D)
assert np.array_equal(spcreator(S).toarray(), D)
assert np.array_equal(spcreator(S, dtype='int16').toarray(), D.astype('int16'))
S = spcreator(D)
assert np.array_equal(spcreator(S).toarray(), D)
assert np.array_equal(spcreator(S, dtype='int16').toarray(), D.astype('int16'))
def test_toarray(self, spcreator, dat1d):
datsp = spcreator(dat1d)
# Check C- or F-contiguous (default).
chk = datsp.toarray()
assert np.array_equal(chk, dat1d)
assert chk.flags.c_contiguous == chk.flags.f_contiguous
# Check C-contiguous (with arg).
chk = datsp.toarray(order='C')
assert np.array_equal(chk, dat1d)
assert chk.flags.c_contiguous
assert chk.flags.f_contiguous
# Check F-contiguous (with arg).
chk = datsp.toarray(order='F')
assert np.array_equal(chk, dat1d)
assert chk.flags.c_contiguous
assert chk.flags.f_contiguous
# Check with output arg.
out = np.zeros(datsp.shape, dtype=datsp.dtype)
datsp.toarray(out=out)
assert np.array_equal(out, dat1d)
# Check that things are fine when we don't initialize with zeros.
out[...] = 1.0
datsp.toarray(out=out)
assert np.array_equal(out, dat1d)
# np.dot does not work with sparse matrices (unless scalars)
# so this is testing whether dat1d matches datsp.toarray()
a = np.array([1.0, 2.0, 3.0, 4.0])
dense_dot_dense = np.dot(a, dat1d)
check = np.dot(a, datsp.toarray())
assert np.array_equal(dense_dot_dense, check)
b = np.array([1.0, 2.0, 3.0, 4.0])
dense_dot_dense = np.dot(dat1d, b)
check = np.dot(datsp.toarray(), b)
assert np.array_equal(dense_dot_dense, check)
# Check bool data works.
spbool = spcreator(dat1d, dtype=bool)
arrbool = dat1d.astype(bool)
assert np.array_equal(spbool.toarray(), arrbool)
def test_add(self, spcreator, datsp_math_dtypes):
for dtype, dat, datsp in datsp_math_dtypes[spcreator]:
a = dat.copy()
a[0] = 2.0
b = datsp
c = b + a
assert np.array_equal(c, b.toarray() + a)
# test broadcasting
# Note: cant add nonzero scalar to sparray. Can add len 1 array
c = b + a[0:1]
assert np.array_equal(c, b.toarray() + a[0])
def test_radd(self, spcreator, datsp_math_dtypes):
for dtype, dat, datsp in datsp_math_dtypes[spcreator]:
a = dat.copy()
a[0] = 2.0
b = datsp
c = a + b
assert np.array_equal(c, a + b.toarray())
def test_rsub(self, spcreator, datsp_math_dtypes):
for dtype, dat, datsp in datsp_math_dtypes[spcreator]:
if dtype == np.dtype('bool'):
# boolean array subtraction deprecated in 1.9.0
continue
assert np.array_equal((dat - datsp), [0, 0, 0, 0])
assert np.array_equal((datsp - dat), [0, 0, 0, 0])
assert np.array_equal((0 - datsp).toarray(), -dat)
A = spcreator([1, -4, 0, 2], dtype='d')
assert np.array_equal((dat - A), dat - A.toarray())
assert np.array_equal((A - dat), A.toarray() - dat)
assert np.array_equal(A.toarray() - datsp, A.toarray() - dat)
assert np.array_equal(datsp - A.toarray(), dat - A.toarray())
# test broadcasting
assert np.array_equal(dat[:1] - datsp, dat[:1] - dat)
def test_matvec(self, spcreator):
A = np.array([2, 0, 3.0])
Asp = spcreator(A)
col = np.array([[1, 2, 3]]).T
assert np.allclose(Asp @ col, Asp.toarray() @ col)
assert (A @ np.array([1, 2, 3])).shape == ()
assert Asp @ np.array([1, 2, 3]) == 11
assert (Asp @ np.array([1, 2, 3])).shape == ()
assert (Asp @ np.array([[1], [2], [3]])).shape == ()
# check result type
assert isinstance(Asp @ matrix([[1, 2, 3]]).T, np.ndarray)
assert (Asp @ np.array([[1, 2, 3]]).T).shape == ()
# ensure exception is raised for improper dimensions
bad_vecs = [np.array([1, 2]), np.array([1, 2, 3, 4]), np.array([[1], [2]])]
for x in bad_vecs:
with pytest.raises(ValueError, match='dimension mismatch'):
Asp.__matmul__(x)
# The current relationship between sparse matrix products and array
# products is as follows:
dot_result = np.dot(Asp.toarray(), [1, 2, 3])
assert np.allclose(Asp @ np.array([1, 2, 3]), dot_result)
assert np.allclose(Asp @ [[1], [2], [3]], dot_result.T)
# Note that the result of Asp @ x is dense if x has a singleton dimension.
def test_rmatvec(self, spcreator, dat1d):
M = spcreator(dat1d)
assert np.allclose([1, 2, 3, 4] @ M, np.dot([1, 2, 3, 4], M.toarray()))
row = np.array([[1, 2, 3, 4]])
assert np.allclose(row @ M, row @ M.toarray())
def test_transpose(self, spcreator, dat1d):
for A in [dat1d, np.array([])]:
B = spcreator(A)
assert np.array_equal(B.toarray(), A)
assert np.array_equal(B.transpose().toarray(), A)
assert np.array_equal(B.dtype, A.dtype)
def test_add_dense_to_sparse(self, spcreator, datsp_math_dtypes):
for dtype, dat, datsp in datsp_math_dtypes[spcreator]:
sum1 = dat + datsp
assert np.array_equal(sum1, dat + dat)
sum2 = datsp + dat
assert np.array_equal(sum2, dat + dat)
def test_iterator(self, spcreator):
# test that __iter__ is compatible with NumPy
B = np.arange(5)
A = spcreator(B)
if A.format not in ['coo', 'dia', 'bsr']:
for x, y in zip(A, B):
assert np.array_equal(x, y)
def test_resize(self, spcreator):
# resize(shape) resizes the matrix in-place
D = np.array([1, 0, 3, 4])
S = spcreator(D)
assert S.resize((3,)) is None
assert np.array_equal(S.toarray(), [1, 0, 3])
S.resize((5,))
assert np.array_equal(S.toarray(), [1, 0, 3, 0, 0])
@pytest.mark.parametrize("spcreator", [sp.sparse.dok_array])
class TestGetSet1D:
def test_getelement(self, spcreator):
D = np.array([4, 3, 0])
A = spcreator(D)
N = D.shape[0]
for j in range(-N, N):
assert np.array_equal(A[j], D[j])
for ij in [3, -4]:
with pytest.raises(
(IndexError, TypeError), match='index value out of bounds'
):
A.__getitem__(ij)
# single element tuples unwrapped
assert A[(0,)] == 4
with pytest.raises(IndexError, match='index value out of bounds'):
A.__getitem__((4,))
def test_setelement(self, spcreator):
dtype = np.float64
A = spcreator((12,), dtype=dtype)
with np.testing.suppress_warnings() as sup:
sup.filter(
sp.sparse.SparseEfficiencyWarning,
"Changing the sparsity structure of a cs[cr]_matrix is expensive",
)
A[0] = dtype(0)
A[1] = dtype(3)
A[8] = dtype(9.0)
A[-2] = dtype(7)
A[5] = 9
A[-9,] = dtype(8)
A[1,] = dtype(5) # overwrite using 1-tuple index
for ij in [13, -14, (13,), (14,)]:
with pytest.raises(IndexError, match='index value out of bounds'):
A.__setitem__(ij, 123.0)