Merge pull request #40 from gmrukwa/develop

Release v2.3.18

.github/workflows/deploy.yml

@@ -12,7 +12,7 @@ on:
 env:
   MAJOR: ${{ 2 }}
   MINOR: ${{ 3 }}
-  FIXUP: ${{ 17 }}
+  FIXUP: ${{ 18 }}
   PACKAGE_INIT_FILE: ${{ 'divik/__init__.py' }}
   DOCKER_REPO: ${{ 'gmrukwa/divik' }}
   IS_ALPHA: ${{ github.event_name == 'pull_request' }}

README.md

@@ -40,7 +40,7 @@ docker pull gmrukwa/divik
 To install specific version, you can specify it in the command, e.g.:
 
 ```bash
-docker pull gmrukwa/divik:2.3.17
+docker pull gmrukwa/divik:2.3.18
 ```
 
 ## Python package
@@ -60,7 +60,7 @@ pip install divik
 or any stable tagged version, e.g.:
 
 ```bash
-pip install divik==2.3.17
+pip install divik==2.3.18
 ```
 
 If you want to have compatibility with

divik/__init__.py

@@ -1,4 +1,4 @@
-__version__ = '2.3.17'
+__version__ = '2.3.18'
 
 from divik import core
 from divik import feature_selection

divik/score/_sampled_gap.py

@@ -13,6 +13,11 @@
 _BIG_PRIME = 40013
 
 
+def _pool_initialize(initializers, *args):
+    for arg, sampler in zip(args, initializers):
+        sampler.initializer(*arg)
+
+
 @seeded(wrapped_requires_seed=True)
 def sampled_gap(data: Data, kmeans: KMeans,
                 sample_size: Union[int, float] = 1000,
@@ -28,9 +33,7 @@ def sampled_gap(data: Data, kmeans: KMeans,
     kmeans_ = clone(kmeans)
     seeds = list(seed + np.arange(n_trials) * _BIG_PRIME)
     with data_.parallel() as d, reference_.parallel() as r:
-        def initializer(*args):
-            for arg, sampler in zip(args, [d, r]):
-                sampler.initializer(*arg)
+        initializer = partial(_pool_initialize, [d, r])
         with maybe_pool(n_jobs, initializer=initializer,
                         initargs=(d.initargs, r.initargs)) as pool:
             compute_disp = partial(_dispersion, sampler=r, kmeans=kmeans_)

docs/instructions/installation.rst

@@ -14,7 +14,7 @@ To install latest stable version use::
 
 To install specific version, you can specify it in the command, e.g.::
 
-    docker pull gmrukwa/divik:2.3.17
+    docker pull gmrukwa/divik:2.3.18
 
 Python package
 --------------
@@ -31,7 +31,7 @@ package::
 
 or any stable tagged version, e.g.::
 
-    pip install divik==2.3.17
+    pip install divik==2.3.18
 
 If you want to have compatibility with
 `gin-config <https://github.com/google/gin-config>`_, you can install

test/cluster/divik/test_dunn_sklearn.py

@@ -51,6 +51,20 @@ def test_works_with_pool(self):
         npt.assert_array_equal(expected, labels1)
         npt.assert_array_equal(expected, labels2)
 
+    def test_works_with_pool_and_sampling(self):
+        X, _ = make_blobs(n_samples=1000, n_features=100, centers=5,
+                          random_state=42)
+        sequential = DunnDiviK(distance='euclidean', distance_percentile=.9,
+                               max_iter=10, sample_size=50, rejection_size=5)
+        parallel = DunnDiviK(distance='euclidean', distance_percentile=.9,
+                             n_jobs=-1, max_iter=10, sample_size=50,
+                             rejection_size=5)
+        expected = sequential.fit_predict(X)
+        labels1 = parallel.fit_predict(X)
+        labels2 = parallel.predict(X)
+        npt.assert_array_equal(expected, labels1)
+        npt.assert_array_equal(expected, labels2)
+
     def test_has_decent_performance_on_numerous_clusters(self):
         X, y = make_blobs(n_samples=200, n_features=100, centers=20,
                           random_state=42)

test/cluster/divik/test_gap_sklearn.py

@@ -57,6 +57,20 @@ def test_works_with_pool(self):
         npt.assert_array_equal(expected, labels1)
         npt.assert_array_equal(expected, labels2)
 
+    def test_works_with_pool_and_sampling(self):
+        X, _ = make_blobs(n_samples=1000, n_features=100, centers=5,
+                          random_state=42)
+        sequential = DiviK(distance='euclidean', distance_percentile=.9,
+                           max_iter=10, sample_size=50, rejection_size=5)
+        parallel = DiviK(distance='euclidean', distance_percentile=.9,
+                         n_jobs=-1, max_iter=10, sample_size=50,
+                         rejection_size=5)
+        expected = sequential.fit_predict(X)
+        labels1 = parallel.fit_predict(X)
+        labels2 = parallel.predict(X)
+        npt.assert_array_equal(expected, labels1)
+        npt.assert_array_equal(expected, labels2)
+
     def test_transforms_to_n_clusters_dimensions(self):
         X, _ = make_blobs(n_samples=600, n_features=100, centers=20,
                           random_state=42)