diff --git a/Makefile b/Makefile
index 428ca5de5f6ae62d0fdc540e85bdf100c439e54b..91260bb681c71f33fb077b44e0c472122eeffe3b 100644
--- a/Makefile
+++ b/Makefile
@@ -7,6 +7,14 @@ CYTHON ?= cython
NOSETESTS ?= nosetests
CTAGS ?= ctags
+# skip doctests on 32bit python
+BITS := $(shell python -c 'import struct; print(8 * struct.calcsize("P"))')
+
+ifeq ($(BITS),32)
+ NOSETESTS:=$(NOSETESTS) -c setup32.cfg
+endif
+
+
all: clean inplace test
clean-ctags:
@@ -28,9 +36,11 @@ test-code: in
test-sphinxext:
$(NOSETESTS) -s -v doc/sphinxext/
test-doc:
+ifeq ($(BITS),64)
$(NOSETESTS) -s -v doc/*.rst doc/modules/ doc/datasets/ \
doc/developers doc/tutorial/basic doc/tutorial/statistical_inference \
doc/tutorial/text_analytics
+endif
test-coverage:
rm -rf coverage .coverage
diff --git a/setup32.cfg b/setup32.cfg
new file mode 100644
index 0000000000000000000000000000000000000000..7294cb813fc8dc835f19c6d920e9b07354a67db3
--- /dev/null
+++ b/setup32.cfg
@@ -0,0 +1,22 @@
+# This config file is here to skip doctests on 32bit linux when running make or make test
+# For newer versions of nose, we can simply use "NOSE_IGNORE_CONFIG_FILES", which
+# we should do in the future.
+
+[aliases]
+# python2.7 has upgraded unittest and it is no longer compatible with some
+# of our tests, so we run all through nose
+test = nosetests
+
+[nosetests]
+# nosetests skips test files with the executable bit by default
+# which can silently hide failing tests.
+# There are no executable scripts within the scikit-learn project
+# so let's turn the --exe flag on to avoid skipping tests by
+# mistake.
+exe = 1
+cover-html = 1
+cover-html-dir = coverage
+cover-package = sklearn
+
+detailed-errors = 1
+with-doctest = 0
diff --git a/sklearn/ensemble/tests/test_forest.py b/sklearn/ensemble/tests/test_forest.py
index 631c726f0381e077ac6f04942ca27f645617b31a..eb6568b72be86cced84998f154d0016a6fa97978 100644
--- a/sklearn/ensemble/tests/test_forest.py
+++ b/sklearn/ensemble/tests/test_forest.py
@@ -30,6 +30,7 @@ from sklearn.utils.testing import assert_greater_equal
from sklearn.utils.testing import assert_raises
from sklearn.utils.testing import assert_warns
from sklearn.utils.testing import ignore_warnings
+from sklearn.utils.testing import skip_if_32bit
from sklearn import datasets
from sklearn.decomposition import TruncatedSVD
@@ -229,6 +230,7 @@ def check_importances(X, y, name, criterion):
assert_less(np.abs(importances - importances_bis).mean(), 0.001)
+@skip_if_32bit
def test_importances():
X, y = datasets.make_classification(n_samples=500, n_features=10,
n_informative=3, n_redundant=0,
diff --git a/sklearn/ensemble/tests/test_gradient_boosting.py b/sklearn/ensemble/tests/test_gradient_boosting.py
index 4f2329be5035818d754ebf33f3f0f17b45b87ba2..766505262bd762ef730b3a9fa7ca21bf67ddf33a 100644
--- a/sklearn/ensemble/tests/test_gradient_boosting.py
+++ b/sklearn/ensemble/tests/test_gradient_boosting.py
@@ -26,7 +26,7 @@ from sklearn.utils.testing import assert_less
from sklearn.utils.testing import assert_raises
from sklearn.utils.testing import assert_true
from sklearn.utils.testing import assert_warns
-from sklearn.utils.testing import ignore_warnings
+from sklearn.utils.testing import skip_if_32bit
from sklearn.utils.validation import DataConversionWarning
from sklearn.utils.validation import NotFittedError
@@ -186,9 +186,9 @@ def check_boston(presort, loss, subsample):
ones = np.ones(len(boston.target))
last_y_pred = None
for sample_weight in None, ones, 2*ones:
- clf = GradientBoostingRegressor(n_estimators=100,
+ clf = GradientBoostingRegressor(n_estimators=100,
loss=loss,
- max_depth=4,
+ max_depth=4,
subsample=subsample,
min_samples_split=1,
random_state=1,
@@ -199,10 +199,10 @@ def check_boston(presort, loss, subsample):
sample_weight=sample_weight)
leaves = clf.apply(boston.data)
assert_equal(leaves.shape, (506, 100))
-
+
y_pred = clf.predict(boston.data)
mse = mean_squared_error(boston.target, y_pred)
- assert_less( mse, 6.0 )
+ assert_less(mse, 6.0)
if last_y_pred is not None:
assert_array_almost_equal(last_y_pred, y_pred)
@@ -211,17 +211,17 @@ def check_boston(presort, loss, subsample):
def test_boston():
- for presort, loss, subsample in product(('auto', True, False),
- ('ls', 'lad', 'huber'),
+ for presort, loss, subsample in product(('auto', True, False),
+ ('ls', 'lad', 'huber'),
(1.0, 0.5)):
yield check_boston, presort, loss, subsample
def check_iris(presort, subsample, sample_weight):
# Check consistency on dataset iris.
- clf = GradientBoostingClassifier(n_estimators=100,
+ clf = GradientBoostingClassifier(n_estimators=100,
loss='deviance',
- random_state=1,
+ random_state=1,
subsample=subsample,
presort=presort)
clf.fit(iris.data, iris.target, sample_weight=sample_weight)
@@ -234,8 +234,8 @@ def check_iris(presort, subsample, sample_weight):
def test_iris():
ones = np.ones(len(iris.target))
- for presort, subsample, sample_weight in product(('auto', True, False),
- (1.0, 0.5),
+ for presort, subsample, sample_weight in product(('auto', True, False),
+ (1.0, 0.5),
(None, ones)):
yield check_iris, presort, subsample, sample_weight
@@ -250,7 +250,7 @@ def test_regression_synthetic():
# Friedman1
X, y = datasets.make_friedman1(n_samples=1200,
- random_state=random_state,
+ random_state=random_state,
noise=1.0)
X_train, y_train = X[:200], y[:200]
X_test, y_test = X[200:], y[200:]
@@ -1028,21 +1028,22 @@ def test_non_uniform_weights_toy_edge_case_clf():
gb.fit(X, y, sample_weight=sample_weight)
assert_array_equal(gb.predict([[1, 0]]), [1])
+
def check_sparse_input(EstimatorClass, X, X_sparse, y):
dense = EstimatorClass(n_estimators=10, random_state=0, max_depth=2).fit(X, y)
- sparse = EstimatorClass(n_estimators=10, random_state=0, max_depth=2,
- presort=False).fit(X_sparse, y)
- auto = EstimatorClass(n_estimators=10, random_state=0, max_depth=2,
- presort='auto').fit(X_sparse, y)
+ sparse = EstimatorClass(n_estimators=10, random_state=0, max_depth=2,
+ presort=False).fit(X_sparse, y)
+ auto = EstimatorClass(n_estimators=10, random_state=0, max_depth=2,
+ presort='auto').fit(X_sparse, y)
assert_array_almost_equal(sparse.apply(X), dense.apply(X))
assert_array_almost_equal(sparse.predict(X), dense.predict(X))
- assert_array_almost_equal(sparse.feature_importances_,
+ assert_array_almost_equal(sparse.feature_importances_,
dense.feature_importances_)
assert_array_almost_equal(sparse.apply(X), auto.apply(X))
assert_array_almost_equal(sparse.predict(X), auto.predict(X))
- assert_array_almost_equal(sparse.feature_importances_,
+ assert_array_almost_equal(sparse.feature_importances_,
auto.feature_importances_)
if isinstance(EstimatorClass, GradientBoostingClassifier):
@@ -1057,6 +1058,7 @@ def check_sparse_input(EstimatorClass, X, X_sparse, y):
auto.predict_log_proba(X))
+@skip_if_32bit
def test_sparse_input():
ests = (GradientBoostingClassifier, GradientBoostingRegressor)
sparse_matrices = (csr_matrix, csc_matrix, coo_matrix)
diff --git a/sklearn/feature_selection/tests/test_from_model.py b/sklearn/feature_selection/tests/test_from_model.py
index f28426d515196613a093a08aad351dfca4f0a002..06dc3d7f425c2815c9de1da207f843a0c49c5eb8 100644
--- a/sklearn/feature_selection/tests/test_from_model.py
+++ b/sklearn/feature_selection/tests/test_from_model.py
@@ -5,11 +5,11 @@ from nose.tools import assert_raises, assert_true
from sklearn.utils.testing import assert_less
from sklearn.utils.testing import assert_greater
-from sklearn.utils.testing import assert_equal
from sklearn.utils.testing import assert_array_almost_equal
from sklearn.utils.testing import assert_array_equal
from sklearn.utils.testing import assert_almost_equal
from sklearn.utils.testing import assert_warns
+from sklearn.utils.testing import skip_if_32bit
from sklearn import datasets
from sklearn.linear_model import LogisticRegression, SGDClassifier, Lasso
@@ -22,6 +22,7 @@ iris = datasets.load_iris()
data, y = iris.data, iris.target
rng = np.random.RandomState(0)
+
def test_transform_linear_model():
for clf in (LogisticRegression(C=0.1),
LinearSVC(C=0.01, dual=False),
@@ -62,6 +63,7 @@ def test_input_estimator_unchanged():
assert_true(transformer.estimator is est)
+@skip_if_32bit
def test_feature_importances():
X, y = datasets.make_classification(
n_samples=1000, n_features=10, n_informative=3, n_redundant=0,
@@ -88,7 +90,7 @@ def test_feature_importances():
transformer = SelectFromModel(estimator=est)
transformer.fit(X, y, sample_weight=sample_weight)
importances = transformer.estimator_.feature_importances_
- transformer.fit(X, y, sample_weight=3*sample_weight)
+ transformer.fit(X, y, sample_weight=3 * sample_weight)
importances_bis = transformer.estimator_.feature_importances_
assert_almost_equal(importances, importances_bis)
diff --git a/sklearn/preprocessing/tests/test_data.py b/sklearn/preprocessing/tests/test_data.py
index 7a91d46aeacd4451c113c505f2e8bbf2481d08bb..9c107469091ce5931063d9a8dcd337f695025878 100644
--- a/sklearn/preprocessing/tests/test_data.py
+++ b/sklearn/preprocessing/tests/test_data.py
@@ -29,6 +29,7 @@ from sklearn.utils.testing import assert_warns_message
from sklearn.utils.testing import assert_no_warnings
from sklearn.utils.testing import ignore_warnings
from sklearn.utils.testing import assert_allclose
+from sklearn.utils.testing import skip_if_32bit
from sklearn.utils.sparsefuncs import mean_variance_axis
from sklearn.preprocessing.data import _transform_selected
@@ -171,6 +172,7 @@ def test_scale_1d():
assert_array_equal(scale(X, with_mean=False, with_std=False), X)
+@skip_if_32bit
def test_standard_scaler_numerical_stability():
"""Test numerical stability of scaling"""
# np.log(1e-5) is taken because of its floating point representation
diff --git a/sklearn/utils/testing.py b/sklearn/utils/testing.py
index 9dcee223f11cd3ef1e1b36ccbbd6242ee8bee91d..0414f246369599855682670b00edd324badff29b 100644
--- a/sklearn/utils/testing.py
+++ b/sklearn/utils/testing.py
@@ -16,6 +16,7 @@ import warnings
import sys
import re
import platform
+import struct
import scipy as sp
import scipy.io
@@ -683,6 +684,18 @@ def if_matplotlib(func):
return run_test
+def skip_if_32bit(func):
+ """Test decorator that skips tests on 32bit platforms."""
+ @wraps(func)
+ def run_test(*args, **kwargs):
+ bits = 8 * struct.calcsize("P")
+ if bits == 32:
+ raise SkipTest('Test skipped on 32bit platforms.')
+ else:
+ return func(*args, **kwargs)
+ return run_test
+
+
def if_not_mac_os(versions=('10.7', '10.8', '10.9'),
message='Multi-process bug in Mac OS X >= 10.7 '
'(see issue #636)'):
diff --git a/sklearn/utils/tests/test_extmath.py b/sklearn/utils/tests/test_extmath.py
index 107f5b845e1d9eb5152fb62e29e4c2fe248b2214..0df11daed48e3d0e5b764444508380c881b30988 100644
--- a/sklearn/utils/tests/test_extmath.py
+++ b/sklearn/utils/tests/test_extmath.py
@@ -16,6 +16,7 @@ from sklearn.utils.testing import assert_array_almost_equal
from sklearn.utils.testing import assert_true
from sklearn.utils.testing import assert_greater
from sklearn.utils.testing import assert_raises
+from sklearn.utils.testing import skip_if_32bit
from sklearn.utils.extmath import density
from sklearn.utils.extmath import logsumexp
@@ -424,6 +425,7 @@ def test_incremental_variance_update_formulas():
assert_almost_equal(final_count, A.shape[0])
+@skip_if_32bit
def test_incremental_variance_numerical_stability():
# Test Youngs and Cramer incremental variance formulas.