MAINT: get rid of np.int

larsmans · larsmans · commit 9e678959a7b7 · 2015-07-25T15:22:34.000+02:00
np.int is a confusing synonym for __builtin__.int; it maps to C long, the size of which is platform-dependent. The NumPy devs are speaking of deprecating np.int, and all uses of it in SciPy have been replaced. See http://github.com/numpy/numpy#6103 for rationale.
diff --git a/benchmarks/bench_covertype.py b/benchmarks/bench_covertype.py
@@ -77,7 +77,7 @@ def load_data(dtype=np.float32, order='C', random_state=13):
     data = fetch_covtype(download_if_missing=True, shuffle=True,
                          random_state=random_state)
     X = check_array(data['data'], dtype=dtype, order=order)
-    y = (data['target'] != 1).astype(np.int)
+    y = (data['target'] != 1).astype(int)
 
     ## Create train-test split (as [Joachims, 2006])
     print("Creating train-test split...")
diff --git a/benchmarks/bench_lasso.py b/benchmarks/bench_lasso.py
@@ -64,7 +64,7 @@ def compute_bench(alpha, n_samples, n_features, precompute):
     alpha = 0.01  # regularization parameter
 
     n_features = 10
-    list_n_samples = np.linspace(100, 1000000, 5).astype(np.int)
+    list_n_samples = np.linspace(100, 1000000, 5).astype(int)
     lasso_results, lars_lasso_results = compute_bench(alpha, list_n_samples,
                                             [n_features], precompute=True)
 
@@ -81,7 +81,7 @@ def compute_bench(alpha, n_samples, n_features, precompute):
     pl.axis('tight')
 
     n_samples = 2000
-    list_n_features = np.linspace(500, 3000, 5).astype(np.int)
+    list_n_features = np.linspace(500, 3000, 5).astype(int)
     lasso_results, lars_lasso_results = compute_bench(alpha, [n_samples],
                                            list_n_features, precompute=False)
     pl.subplot(212)
diff --git a/benchmarks/bench_plot_fastkmeans.py b/benchmarks/bench_plot_fastkmeans.py
@@ -95,9 +95,9 @@ def compute_bench_2(chunks):
     from mpl_toolkits.mplot3d import axes3d  # register the 3d projection
     import matplotlib.pyplot as plt
 
-    samples_range = np.linspace(50, 150, 5).astype(np.int)
-    features_range = np.linspace(150, 50000, 5).astype(np.int)
-    chunks = np.linspace(500, 10000, 15).astype(np.int)
+    samples_range = np.linspace(50, 150, 5).astype(int)
+    features_range = np.linspace(150, 50000, 5).astype(int)
+    chunks = np.linspace(500, 10000, 15).astype(int)
 
     results = compute_bench(samples_range, features_range)
     results_2 = compute_bench_2(chunks)
diff --git a/benchmarks/bench_plot_nmf.py b/benchmarks/bench_plot_nmf.py
@@ -146,8 +146,8 @@ def benchmark(samples_range, features_range, rank=50, tolerance=1e-5):
     axes3d
     import matplotlib.pyplot as plt
 
-    samples_range = np.linspace(50, 500, 3).astype(np.int)
-    features_range = np.linspace(50, 500, 3).astype(np.int)
+    samples_range = np.linspace(50, 500, 3).astype(int)
+    features_range = np.linspace(50, 500, 3).astype(int)
     timeset, err = benchmark(samples_range, features_range)
 
     for i, results in enumerate((timeset, err)):
diff --git a/benchmarks/bench_plot_omp_lars.py b/benchmarks/bench_plot_omp_lars.py
@@ -100,8 +100,8 @@ def compute_bench(samples_range, features_range):
 
 
 if __name__ == '__main__':
-    samples_range = np.linspace(1000, 5000, 5).astype(np.int)
-    features_range = np.linspace(1000, 5000, 5).astype(np.int)
+    samples_range = np.linspace(1000, 5000, 5).astype(int)
+    features_range = np.linspace(1000, 5000, 5).astype(int)
     results = compute_bench(samples_range, features_range)
     max_time = max(np.max(t) for t in results.values())
 
diff --git a/benchmarks/bench_plot_svd.py b/benchmarks/bench_plot_svd.py
@@ -57,8 +57,8 @@ def compute_bench(samples_range, features_range, n_iter=3, rank=50):
     from mpl_toolkits.mplot3d import axes3d  # register the 3d projection
     import matplotlib.pyplot as plt
 
-    samples_range = np.linspace(2, 1000, 4).astype(np.int)
-    features_range = np.linspace(2, 1000, 4).astype(np.int)
+    samples_range = np.linspace(2, 1000, 4).astype(int)
+    features_range = np.linspace(2, 1000, 4).astype(int)
     results = compute_bench(samples_range, features_range)
 
     label = 'scikit-learn singular value decomposition benchmark results'
diff --git a/benchmarks/bench_plot_ward.py b/benchmarks/bench_plot_ward.py
@@ -35,9 +35,9 @@
 pl.imshow(np.log(ratio), aspect='auto', origin="lower")
 pl.colorbar()
 pl.contour(ratio, levels=[1, ], colors='k')
-pl.yticks(range(len(n_features)), n_features.astype(np.int))
+pl.yticks(range(len(n_features)), n_features.astype(int))
 pl.ylabel('N features')
-pl.xticks(range(len(n_samples)), n_samples.astype(np.int))
+pl.xticks(range(len(n_samples)), n_samples.astype(int))
 pl.xlabel('N samples')
 pl.title("Scikit's time, in units of scipy time (log)")
 pl.show()
diff --git a/benchmarks/bench_sample_without_replacement.py b/benchmarks/bench_sample_without_replacement.py
@@ -145,7 +145,7 @@ def bench_sample(sampling, n_population, n_samples):
     ###########################################################################
     time = {}
     n_samples = np.linspace(start=0, stop=opts.n_population,
-        num=opts.n_steps).astype(np.int)
+        num=opts.n_steps).astype(int)
 
     ratio = n_samples / opts.n_population
 
diff --git a/benchmarks/bench_sgd_regression.py b/benchmarks/bench_sgd_regression.py
@@ -22,7 +22,7 @@
 from sklearn.datasets.samples_generator import make_regression
 
 if __name__ == "__main__":
-    list_n_samples = np.linspace(100, 10000, 5).astype(np.int)
+    list_n_samples = np.linspace(100, 10000, 5).astype(int)
     list_n_features = [10, 100, 1000]
     n_test = 1000
     noise = 0.1
diff --git a/examples/applications/plot_tomography_l1_reconstruction.py b/examples/applications/plot_tomography_l1_reconstruction.py
@@ -104,7 +104,7 @@ def generate_synthetic_data():
     mask_outer = (x - l / 2) ** 2 + (y - l / 2) ** 2 < (l / 2) ** 2
     mask = np.zeros((l, l))
     points = l * rs.rand(2, n_pts)
-    mask[(points[0]).astype(np.int), (points[1]).astype(np.int)] = 1
+    mask[(points[0]).astype(int), (points[1]).astype(int)] = 1
     mask = ndimage.gaussian_filter(mask, sigma=l / n_pts)
     res = np.logical_and(mask > mask.mean(), mask_outer)
     return res - ndimage.binary_erosion(res)
diff --git a/examples/cluster/plot_adjusted_for_chance_measures.py b/examples/cluster/plot_adjusted_for_chance_measures.py
@@ -66,7 +66,7 @@ def uniform_labelings_scores(score_func, n_samples, n_clusters_range,
 # 2 independent random clusterings with equal cluster number
 
 n_samples = 100
-n_clusters_range = np.linspace(2, n_samples, 10).astype(np.int)
+n_clusters_range = np.linspace(2, n_samples, 10).astype(int)
 
 plt.figure(1)
 
@@ -95,7 +95,7 @@ def uniform_labelings_scores(score_func, n_samples, n_clusters_range,
 # with fixed number of clusters
 
 n_samples = 1000
-n_clusters_range = np.linspace(2, 100, 10).astype(np.int)
+n_clusters_range = np.linspace(2, 100, 10).astype(int)
 n_classes = 10
 
 plt.figure(2)
diff --git a/examples/cluster/plot_cluster_comparison.py b/examples/cluster/plot_cluster_comparison.py
@@ -97,7 +97,7 @@
         algorithm.fit(X)
         t1 = time.time()
         if hasattr(algorithm, 'labels_'):
-            y_pred = algorithm.labels_.astype(np.int)
+            y_pred = algorithm.labels_.astype(int)
         else:
             y_pred = algorithm.predict(X)
 
diff --git a/examples/linear_model/plot_logistic_l1_l2_sparsity.py b/examples/linear_model/plot_logistic_l1_l2_sparsity.py
@@ -33,7 +33,7 @@
 X = StandardScaler().fit_transform(X)
 
 # classify small against large digits
-y = (y > 4).astype(np.int)
+y = (y > 4).astype(int)
 
 
 # Set regularization parameter
diff --git a/sklearn/cluster/hierarchical.py b/sklearn/cluster/hierarchical.py
@@ -428,7 +428,7 @@ def linkage_tree(X, connectivity=None, n_components=None,
             i, j = np.triu_indices(X.shape[0], k=1)
             X = X[i, j]
         out = hierarchy.linkage(X, method=linkage, metric=affinity)
-        children_ = out[:, :2].astype(np.int)
+        children_ = out[:, :2].astype(int)
 
         if return_distance:
             distances = out[:, 2]
diff --git a/sklearn/cluster/k_means_.py b/sklearn/cluster/k_means_.py
@@ -469,7 +469,7 @@ def _labels_inertia_precompute_dense(X, x_squared_norms, centers, distances):
 
     Returns
     -------
-    labels : numpy array, dtype=np.int, shape (n_samples,)
+    labels : numpy array, dtype=np.int32, shape (n_samples,)
         Indices of clusters that samples are assigned to.
 
     inertia : float
@@ -523,7 +523,7 @@ def _labels_inertia(X, x_squared_norms, centers,
 
     Returns
     -------
-    labels: int array of shape(n)
+    labels: int32 array of shape(n)
         The resulting assignment
 
     inertia : float
diff --git a/sklearn/cluster/mean_shift_.py b/sklearn/cluster/mean_shift_.py
@@ -193,7 +193,7 @@ def mean_shift(X, bandwidth=None, seeds=None, bin_seeding=False,
 
     # ASSIGN LABELS: a point belongs to the cluster that it is closest to
     nbrs = NearestNeighbors(n_neighbors=1).fit(cluster_centers)
-    labels = np.zeros(n_samples, dtype=np.int)
+    labels = np.zeros(n_samples, dtype=int)
     distances, idxs = nbrs.kneighbors(X)
     if cluster_all:
         labels = idxs.flatten()
diff --git a/sklearn/cluster/tests/test_hierarchical.py b/sklearn/cluster/tests/test_hierarchical.py
@@ -251,7 +251,7 @@ def test_scikit_vs_scipy():
 
             out = hierarchy.linkage(X, method=linkage)
 
-            children_ = out[:, :2].astype(np.int)
+            children_ = out[:, :2].astype(int)
             children, _, n_leaves, _ = _TREE_BUILDERS[linkage](X, connectivity)
 
             cut = _hc_cut(k, children, n_leaves)
