def test_dict_learning_online_positivity(transform_algorithm,
positive_code,
positive_dict):
rng = np.random.RandomState(0)
n_components = 8
dico = MiniBatchDictionaryLearning(
n_components, transform_algorithm=transform_algorithm, random_state=0,
positive_code=positive_code, positive_dict=positive_dict).fit(X)
code = dico.transform(X)
if positive_dict:
assert_true((dico.components_ >= 0).all())
else:
assert_true((dico.components_ < 0).any())
if positive_code:
assert_true((code >= 0).all())
else:
assert_true((code < 0).any())
code, dictionary = dict_learning_online(X, n_components=n_components,
alpha=1, random_state=rng,
positive_dict=positive_dict,
positive_code=positive_code)
if positive_dict:
assert_true((dictionary >= 0).all())
else:
assert_true((dictionary < 0).any())
if positive_code:
assert_true((code >= 0).all())
else:
assert_true((code < 0).any())
def test_dict_learning_online_numerical_consistency(method):
# verify numerically consistent among np.float32 and np.float64
rtol = 1e-4
n_components = 4
alpha = 1
U_64, V_64 = dict_learning_online(
X.astype(np.float64),
n_components=n_components,
alpha=alpha,
random_state=0,
method=method,
)
U_32, V_32 = dict_learning_online(
X.astype(np.float32),
n_components=n_components,
alpha=alpha,
random_state=0,
method=method,
)
# Optimal solution (U*, V*) is not unique.
# If (U*, V*) is optimal solution, (-U*,-V*) is also optimal,
# and (column permutated U*, row permutated V*) are also optional
# as long as holding UV.
# So here UV, ||U||_1,1 and sum(||V_k||_2) are verified
# instead of comparing directly U and V.
assert_allclose(np.matmul(U_64, V_64), np.matmul(U_32, V_32), rtol=rtol)
assert_allclose(np.sum(np.abs(U_64)), np.sum(np.abs(U_32)), rtol=rtol)
assert_allclose(np.sum(V_64 ** 2), np.sum(V_32 ** 2), rtol=rtol)
# verify an obtained solution is not degenerate
assert np.mean(U_64 != 0.0) > 0.05
assert np.count_nonzero(U_64 != 0.0) == np.count_nonzero(U_32 != 0.0)
def test_dict_learning_online(self):
iris = datasets.load_iris()
df = pdml.ModelFrame(iris)
result = df.decomposition.dict_learning_online(
random_state=self.random_state)
expected = decomposition.dict_learning_online(
iris.data, random_state=self.random_state)
self.assertEqual(len(result), 2)
self.assertIsInstance(result[0], pdml.ModelFrame)
tm.assert_index_equal(result[0].index, df.data.index)
self.assert_numpy_array_almost_equal(result[0].values, expected[0])
self.assertIsInstance(result[1], pdml.ModelFrame)
tm.assert_index_equal(result[1].columns, df.data.columns)
self.assert_numpy_array_almost_equal(result[1].values, expected[1])
result = df.decomposition.dict_learning_online(
return_code=False, random_state=self.random_state)
expected = decomposition.dict_learning_online(
iris.data, return_code=False, random_state=self.random_state)
self.assertIsInstance(result, pdml.ModelFrame)
tm.assert_index_equal(result.columns, df.data.columns)
self.assert_numpy_array_almost_equal(result.values, expected)
def test_dict_learning_online(self):
iris = datasets.load_iris()
df = pdml.ModelFrame(iris)
result = df.decomposition.dict_learning_online(random_state=self.random_state)
expected = decomposition.dict_learning_online(iris.data,
random_state=self.random_state)
self.assertEqual(len(result), 2)
self.assertIsInstance(result[0], pdml.ModelFrame)
tm.assert_index_equal(result[0].index, df.data.index)
self.assert_numpy_array_almost_equal(result[0].values, expected[0])
self.assertIsInstance(result[1], pdml.ModelFrame)
tm.assert_index_equal(result[1].columns, df.data.columns)
self.assert_numpy_array_almost_equal(result[1].values, expected[1])
result = df.decomposition.dict_learning_online(return_code=False,
random_state=self.random_state)
expected = decomposition.dict_learning_online(iris.data,
return_code=False,
random_state=self.random_state)
self.assertIsInstance(result, pdml.ModelFrame)
tm.assert_index_equal(result.columns, df.data.columns)
self.assert_numpy_array_almost_equal(result.values, expected)
def test_dict_learning_online_shapes():
rng = np.random.RandomState(0)
n_components = 8
code, dictionary = dict_learning_online(
X,
n_components=n_components,
batch_size=4,
max_iter=10,
random_state=rng,
return_code=True,
)
assert code.shape == (n_samples, n_components)
assert dictionary.shape == (n_components, n_features)
assert np.dot(code, dictionary).shape == X.shape
dictionary = dict_learning_online(
X,
n_components=n_components,
batch_size=4,
max_iter=10,
random_state=rng,
return_code=False,
)
assert dictionary.shape == (n_components, n_features)
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from io import StringIO
import sys
old_stdout = sys.stdout
try:
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components,
n_iter=20,
verbose=1,
random_state=0)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components,
n_iter=20,
verbose=2,
random_state=0)
dico.fit(X)
dict_learning_online(X,
n_components=n_components,
alpha=1,
verbose=1,
random_state=0)
dict_learning_online(X,
n_components=n_components,
alpha=1,
verbose=2,
random_state=0)
finally:
sys.stdout = old_stdout
assert dico.components_.shape == (n_components, n_features)
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from sklearn.externals.six.moves import cStringIO as StringIO
import sys
old_stdout = sys.stdout
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components,
n_iter=20,
verbose=1,
random_state=0)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components,
n_iter=20,
verbose=2,
random_state=0)
dico.fit(X)
dict_learning_online(X,
n_components=n_components,
alpha=1,
verbose=1,
random_state=0)
dict_learning_online(X,
n_components=n_components,
alpha=1,
verbose=2,
random_state=0)
sys.stdout = old_stdout
assert_true(dico.components_.shape == (n_components, n_features))
def test_batch_size_default_value_future_warning():
# Check that a FutureWarning is raised if batch_size is left to its default value.
# FIXME: remove in 1.3
msg = "The default value of batch_size will change"
with pytest.warns(FutureWarning, match=msg):
dict_learning_online(X, n_components=2, random_state=0)
with pytest.warns(FutureWarning, match=msg):
MiniBatchDictionaryLearning(n_components=2, random_state=0).fit(X)
def test_dict_learning_online_verbosity():
# test verbosity for better coverage
n_components = 5
from io import StringIO
import sys
old_stdout = sys.stdout
try:
sys.stdout = StringIO()
# convergence monitoring verbosity
dico = MiniBatchDictionaryLearning(n_components,
batch_size=4,
max_iter=5,
verbose=1,
tol=0.1,
random_state=0)
dico.fit(X)
dico = MiniBatchDictionaryLearning(
n_components,
batch_size=4,
max_iter=5,
verbose=1,
max_no_improvement=2,
random_state=0,
)
dico.fit(X)
# higher verbosity level
dico = MiniBatchDictionaryLearning(n_components,
batch_size=4,
max_iter=5,
verbose=2,
random_state=0)
dico.fit(X)
# function API verbosity
dict_learning_online(
X,
n_components=n_components,
batch_size=4,
alpha=1,
verbose=1,
random_state=0,
)
dict_learning_online(
X,
n_components=n_components,
batch_size=4,
alpha=1,
verbose=2,
random_state=0,
)
finally:
sys.stdout = old_stdout
assert dico.components_.shape == (n_components, n_features)
def test_dict_learning_online_deprecated_args(arg, val):
# check the deprecation warning for the deprecated args of
# dict_learning_online
# FIXME: remove in 1.3
depr_msg = (
f"'{arg}' is deprecated in version 1.1 and will be removed in version 1.3."
)
with pytest.warns(FutureWarning, match=depr_msg):
dict_learning_online(
X, n_components=2, batch_size=4, random_state=0, **{arg: val}
)
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from cStringIO import StringIO
import sys
old_stdout = sys.stdout
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=1)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=2)
dico.fit(X)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=1)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=2)
sys.stdout = old_stdout
assert_true(dico.components_.shape == (n_components, n_features))
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from cStringIO import StringIO
import sys
old_stdout = sys.stdout
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=1)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=2)
dico.fit(X)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=1)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=2)
sys.stdout = old_stdout
assert_true(dico.components_.shape == (n_components, n_features))
def test_dict_learning_online_shapes():
rng = np.random.RandomState(0)
n_components = 8
code, dictionary = dict_learning_online(X, n_components=n_components,
alpha=1, random_state=rng)
assert_equal(code.shape, (n_samples, n_components))
assert_equal(dictionary.shape, (n_components, n_features))
assert_equal(np.dot(code, dictionary).shape, X.shape)
def test_dict_learning_online_shapes():
rng = np.random.RandomState(0)
n_components = 8
code, dictionary = dict_learning_online(X, n_components=n_components,
alpha=1, random_state=rng)
assert code.shape == (n_samples, n_components)
assert dictionary.shape == (n_components, n_features)
assert np.dot(code, dictionary).shape == X.shape
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from sklearn.externals.six.moves import cStringIO as StringIO
import sys
old_stdout = sys.stdout
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=1,
random_state=0)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=2,
random_state=0)
dico.fit(X)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=1,
random_state=0)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=2,
random_state=0)
sys.stdout = old_stdout
assert_true(dico.components_.shape == (n_components, n_features))
def test_dict_learning_online_shapes():
# rng = np.random.RandomState(0)
# X = rng.randn(12, 10)
n_atoms = 8
code, dictionary = dict_learning_online(X,
n_atoms=n_atoms,
alpha=1,
random_state=rng)
assert_equal(code.shape, (n_samples, n_atoms))
assert_equal(dictionary.shape, (n_atoms, n_features))
assert_equal(np.dot(code, dictionary).shape, X.shape)
def make_dictionary(X,
n_components=20,
alpha=5.,
write_dir='/tmp/',
contrasts=[],
method='multitask',
l1_ratio=.5,
n_subjects=13):
"""Create dictionary + encoding"""
from sklearn.decomposition import dict_learning_online, sparse_encode
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import MultiTaskLasso, MultiTaskElasticNet
mem = Memory(write_dir, verbose=0)
dictionary = mem.cache(initial_dictionary)(n_components, X)
np.savez(os.path.join(write_dir, 'dictionary.npz'),
loadings=dictionary,
contrasts=contrasts)
if method == 'online':
components, dictionary = dict_learning_online(X.T,
n_components,
alpha=alpha,
dict_init=dictionary,
batch_size=200,
method='cd',
return_code=True,
shuffle=True,
n_jobs=1,
positive_code=True)
np.savez(os.path.join(write_dir, 'dictionary.npz'),
loadings=dictionary,
contrasts=contrasts)
elif method == 'sparse':
components = sparse_encode(X.T,
dictionary,
alpha=alpha,
max_iter=10,
n_jobs=1,
check_input=True,
verbose=0,
positive=True)
elif method == 'multitask':
# too many hard-typed parameters !!!
n_voxels = X.shape[1] // n_subjects
components = np.zeros((X.shape[1], n_components))
clf = MultiTaskLasso(alpha=alpha)
clf = MultiTaskElasticNet(alpha=alpha, l1_ratio=l1_ratio)
for i in range(n_voxels):
x = X[:, i:i + n_subjects * n_voxels:n_voxels]
components[i: i + n_subjects * n_voxels: n_voxels] =\
clf.fit(dictionary.T, x).coef_
return dictionary, components
def test_dict_learning_online_dtype_match(data_type, expected_type, method):
# Verify output matrix dtype
rng = np.random.RandomState(0)
n_components = 8
code, dictionary = dict_learning_online(
X.astype(data_type),
n_components=n_components,
alpha=1,
random_state=rng,
method=method,
)
assert code.dtype == expected_type
assert dictionary.dtype == expected_type
def test_dict_learning_online_verbosity():
n_components = 5
# test verbosity
from io import StringIO
import sys
old_stdout = sys.stdout
try:
sys.stdout = StringIO()
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=1,
random_state=0)
dico.fit(X)
dico = MiniBatchDictionaryLearning(n_components, n_iter=20, verbose=2,
random_state=0)
dico.fit(X)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=1,
random_state=0)
dict_learning_online(X, n_components=n_components, alpha=1, verbose=2,
random_state=0)
finally:
sys.stdout = old_stdout
assert dico.components_.shape == (n_components, n_features)
def main(output, image_files):
log.info('starting with {} image files'.format(len(image_files)))
images = (prep(i) for i in image_files)
log.info('starting online dictionary learning')
D = None
for image in images:
D = dict_learning_online(image,
dict_init=D,
n_components=2000,
verbose=True,
n_jobs=-1,
n_iter=N_ITER,
batch_size=BATCH_SIZE,
return_code=False)
output.write(pd.DataFrame(D).to_csv())
log.info('done')
def main(output, image_files):
log.info('starting with {} image files'.format(len(image_files)))
images = (prep(i) for i in image_files)
log.info('starting online dictionary learning')
D = None
for image in images:
D = dict_learning_online(
image,
dict_init=D,
n_components=2000,
verbose=True,
n_jobs=-1,
n_iter=N_ITER,
batch_size=BATCH_SIZE,
return_code=False)
output.write(pd.DataFrame(D).to_csv())
log.info('done')
def test_dict_learning_online_positivity(positive_code,
positive_dict):
rng = np.random.RandomState(0)
n_components = 8
code, dictionary = dict_learning_online(X, n_components=n_components,
method="cd",
alpha=1, random_state=rng,
positive_dict=positive_dict,
positive_code=positive_code)
if positive_dict:
assert (dictionary >= 0).all()
else:
assert (dictionary < 0).any()
if positive_code:
assert (code >= 0).all()
else:
assert (code < 0).any()
def fit(self, X, y=None):
"""Fit the model from data in X.
Parameters
----------
X: array-like, shape (n_samples, n_features)
Training vector, where n_samples in the number of samples
and n_features is the number of features.
Returns
-------
self: object
Returns the object itself
"""
from sklearn.decomposition import dict_learning_online
code_, dict_ = dict_learning_online(X,n_components=self.n_components,
alpha=self.alpha,
n_iter=self.n_iter)
self.code_ = code_
self.dict_ = dict_
return self
def test_dict_learning_online_lars_positive_parameter():
alpha = 1
err_msg = "Positive constraint not supported for 'lars' coding method."
with pytest.raises(ValueError, match=err_msg):
dict_learning_online(X, alpha=alpha, positive_code=True)
shape_data = np.load(
'/media/keyi/Data/Research/course_project/AdvancedCV_2020/data/COCO17/shape_val2017_128.npy'
)
else:
print('Not implemented, try n_vertices: 16, 32, 64')
exit()
out_dict = '/media/keyi/Data/Research/course_project/AdvancedCV_2020/data/COCO17/shape_codes/sparsity/dict_val2017_v{}_b{}_alpha{}.npy'.format(
n_vertices, n_atom, alpha)
out_code = '/media/keyi/Data/Research/course_project/AdvancedCV_2020/data/COCO17/shape_codes/sparsity/code_val2017_v{}_b{}_alpha{}.npy'.format(
n_vertices, n_atom, alpha)
n_dim, n_data = shape_data.shape
print('Shape data dims: ', shape_data.shape)
# dict_learner = DictionaryLearning(n_components=n_atom, alpha=1., max_iter=500)
learned_codes, learned_dict = dict_learning_online(np.transpose(shape_data),
n_components=n_atom,
alpha=alpha,
n_iter=2000,
batch_size=50,
return_code=True)
print("Learned dictionary dim: ", learned_dict.shape)
print("Learned codes dim: ", learned_codes.shape)
# calculate the reconstruction error
error = np.sum(
(np.matmul(learned_codes, learned_dict) - np.transpose(shape_data))**
2) / n_data
print('reconstruction error(frobenius): ', error)
np.save(out_dict, learned_dict)
np.save(out_code, learned_codes)
def test_dict_learning_online_lars_positive_parameter():
err_msg = "Positive constraint not supported for 'lars' coding method."
with pytest.raises(ValueError, match=err_msg):
dict_learning_online(X, batch_size=4, max_iter=10, positive_code=True)
def test_dict_learning_online_n_iter_deprecated():
# Check that an error is raised when a deprecated argument is set when max_iter
# is also set.
msg = "The following arguments are incompatible with 'max_iter'"
with pytest.raises(ValueError, match=msg):
dict_learning_online(X, max_iter=10, return_inner_stats=True)
