COSMIT we no longer support Py 2.5

Also, avoid 'from foo import bar' or 'import as' as much as possible

sklearn/linear_model/omp.py

@@ -5,7 +5,7 @@
 #
 # License: BSD Style.
 
-from warnings import warn
+import warnings
 
 import numpy as np
 from scipy import linalg
@@ -75,7 +75,7 @@ def _cholesky_omp(X, y, n_nonzero_coefs, tol=None, copy_X=True):
         lam = np.argmax(np.abs(np.dot(X.T, residual)))
         if lam < n_active or alpha[lam] ** 2 < min_float:
             # atom already selected or inner product too small
-            warn(premature, RuntimeWarning, stacklevel=2)
+            warnings.warn(premature, RuntimeWarning, stacklevel=2)
             break
         if n_active > 0:
             # Updates the Cholesky decomposition of X' X
@@ -83,7 +83,7 @@ def _cholesky_omp(X, y, n_nonzero_coefs, tol=None, copy_X=True):
             solve_triangular(L[:n_active, :n_active], L[n_active, :n_active])
             v = nrm2(L[n_active, :n_active]) ** 2
             if 1 - v <= min_float:  # selected atoms are dependent
-                warn(premature, RuntimeWarning, stacklevel=2)
+                warnings.warn(premature, RuntimeWarning, stacklevel=2)
                 break
             L[n_active, n_active] = np.sqrt(1 - v)
         X.T[n_active], X.T[lam] = swap(X.T[n_active], X.T[lam])
@@ -169,14 +169,14 @@ def _gram_omp(Gram, Xy, n_nonzero_coefs, tol_0=None, tol=None,
         lam = np.argmax(np.abs(alpha))
         if lam < n_active or alpha[lam] ** 2 < min_float:
             # selected same atom twice, or inner product too small
-            warn(premature, RuntimeWarning, stacklevel=2)
+            warnings.warn(premature, RuntimeWarning, stacklevel=2)
             break
         if n_active > 0:
             L[n_active, :n_active] = Gram[lam, :n_active]
             solve_triangular(L[:n_active, :n_active], L[n_active, :n_active])
             v = nrm2(L[n_active, :n_active]) ** 2
             if 1 - v <= min_float:  # selected atoms are dependent
-                warn(premature, RuntimeWarning, stacklevel=2)
+                warnings.warn(premature, RuntimeWarning, stacklevel=2)
                 break
             L[n_active, n_active] = np.sqrt(1 - v)
         Gram[n_active], Gram[lam] = swap(Gram[n_active], Gram[lam])
@@ -520,11 +520,12 @@ class OrthogonalMatchingPursuit(LinearModel):
             self.n_nonzero_coefs = int(0.1 * n_features)
         if (Gram is not None or Xy is not None) and (self.fit_intercept is True
                                                  or self.normalize is True):
-            warn('Mean subtraction (fit_intercept) and normalization cannot '
-                 'be applied on precomputed Gram and Xy matrices. Your '
-                 'precomputed values are ignored and recomputed. To avoid '
-                 'this, do the scaling yourself and call with fit_intercept '
-                 'and normalize set to False.', RuntimeWarning, stacklevel=2)
+            warnings.warn('Mean subtraction (fit_intercept) and '
+                 'normalization cannot be applied on precomputed Gram '
+                 'and Xy matrices. Your precomputed values are ignored '
+                 'and recomputed. To avoid this, do the scaling yourself'
+                 'and call with fit_intercept and normalize set to False.',
+                 RuntimeWarning, stacklevel=2)
             Gram, Xy = None, None
 
         if Gram is not None:

sklearn/linear_model/tests/test_omp.py

@@ -2,11 +2,9 @@
 # License: BSD style
 
 import warnings
-from sys import version_info
 
 import numpy as np
 
-from nose import SkipTest
 from nose.tools import assert_raises, assert_true
 from numpy.testing import assert_equal, assert_array_almost_equal
 
@@ -23,12 +21,6 @@ G, Xy = np.dot(X.T, X), np.dot(X.T, y)
 # and y (n_samples, 3)
 
 
-def check_warnings():
-    if version_info < (2, 6):
-        raise SkipTest("Testing for warnings is not supported in versions \
-        older than Python 2.6")
-
-
 def test_correct_shapes():
     assert_equal(orthogonal_mp(X, y[:, 0], n_nonzero_coefs=5).shape,
                  (n_features,))
@@ -71,7 +63,6 @@ def test_with_without_gram_tol():
 
 
 def test_unreachable_accuracy():
-    check_warnings()  # Skip if unsupported Python version
     with warnings.catch_warnings(record=True) as w:
         warnings.simplefilter('always')
         assert_array_almost_equal(
@@ -133,7 +124,6 @@ def test_estimator():
 
 
 def test_scaling_with_gram():
-    check_warnings()  # Skip if unsupported Python version
     with warnings.catch_warnings(record=True) as w:
         warnings.simplefilter('always')
         # Use only 1 nonzero coef to be faster and to avoid warnings
@@ -157,7 +147,6 @@ def test_scaling_with_gram():
 
 
 def test_identical_regressors():
-    check_warnings()  # Skip if unsupported Python version
     newX = X.copy()
     newX[:, 1] = newX[:, 0]
     gamma = np.zeros(n_features)
@@ -185,7 +174,6 @@ def test_swapped_regressors():
 
 
 def test_no_atoms():
-    check_warnings()  # Skip if unsupported Python version
     y_empty = np.zeros_like(y)
     Xy_empty = np.dot(X.T, y_empty)
     with warnings.catch_warnings():