def test_seq_dataset_shuffle():
dataset1 = ArrayDataset(X, y, sample_weight, seed=42)
dataset2 = CSRDataset(X_csr.data, X_csr.indptr, X_csr.indices,
y, sample_weight, seed=42)
# not shuffled
for i in range(5):
_, _, _, idx1 = dataset1._next_py()
_, _, _, idx2 = dataset2._next_py()
assert_equal(idx1, i)
assert_equal(idx2, i)
for i in range(5):
_, _, _, idx1 = dataset1._random_py()
_, _, _, idx2 = dataset2._random_py()
assert_equal(idx1, idx2)
seed = 77
dataset1._shuffle_py(seed)
dataset2._shuffle_py(seed)
for i in range(5):
_, _, _, idx1 = dataset1._next_py()
_, _, _, idx2 = dataset2._next_py()
assert_equal(idx1, idx2)
_, _, _, idx1 = dataset1._random_py()
_, _, _, idx2 = dataset2._random_py()
assert_equal(idx1, idx2)
def test_seq_dataset():
dataset1 = ArrayDataset(X, y, sample_weight, seed=42)
dataset2 = CSRDataset(X_csr.data,
X_csr.indptr,
X_csr.indices,
y,
sample_weight,
seed=42)
for dataset in (dataset1, dataset2):
for i in range(5):
# next sample
xi_, yi, swi, idx = dataset._next_py()
xi = sp.csr_matrix((xi_), shape=(1, X.shape[1]))
assert_array_equal(xi.data, X_csr[idx].data)
assert_array_equal(xi.indices, X_csr[idx].indices)
assert_array_equal(xi.indptr, X_csr[idx].indptr)
assert_equal(yi, y[idx])
assert_equal(swi, sample_weight[idx])
# random sample
xi_, yi, swi, idx = dataset._random_py()
xi = sp.csr_matrix((xi_), shape=(1, X.shape[1]))
assert_array_equal(xi.data, X_csr[idx].data)
assert_array_equal(xi.indices, X_csr[idx].indices)
assert_array_equal(xi.indptr, X_csr[idx].indptr)
assert_equal(yi, y[idx])
assert_equal(swi, sample_weight[idx])
def test_seq_dataset_shuffle():
dataset1 = ArrayDataset(X, y, sample_weight, seed=42)
dataset2 = CSRDataset(X_csr.data, X_csr.indptr, X_csr.indices,
y, sample_weight, seed=42)
# not shuffled
for i in range(5):
_, _, _, idx1 = dataset1._next_py()
_, _, _, idx2 = dataset2._next_py()
assert_equal(idx1, i)
assert_equal(idx2, i)
for i in range(5):
_, _, _, idx1 = dataset1._random_py()
_, _, _, idx2 = dataset2._random_py()
assert_equal(idx1, idx2)
seed = 77
dataset1._shuffle_py(seed)
dataset2._shuffle_py(seed)
for i in range(5):
_, _, _, idx1 = dataset1._next_py()
_, _, _, idx2 = dataset2._next_py()
assert_equal(idx1, idx2)
_, _, _, idx1 = dataset1._random_py()
_, _, _, idx2 = dataset2._random_py()
assert_equal(idx1, idx2)
def make_dataset(X, y, sample_weight, random_state=None):
"""Create ``Dataset`` abstraction for sparse and dense inputs.
This also returns the ``intercept_decay`` which is different
for sparse datasets.
"""
rng = check_random_state(random_state)
# seed should never be 0 in SequentialDataset
seed = rng.randint(1, np.iinfo(np.int32).max)
if sp.issparse(X):
dataset = CSRDataset(X.data,
X.indptr,
X.indices,
y,
sample_weight,
seed=seed)
intercept_decay = SPARSE_INTERCEPT_DECAY
else:
dataset = ArrayDataset(X, y, sample_weight, seed=seed)
intercept_decay = 1.0
return dataset, intercept_decay