Adding Kmeans and Vector3Stats python interface and tests

Signed-off-by: Marcos Wagner <[email protected]>

src/python/CMakeLists.txt

@@ -70,13 +70,15 @@ if (PYTHONLIBS_FOUND)
     set(python_tests
       Angle_TEST
       GaussMarkovProcess_TEST
+      Kmeans_TEST
       Line2_TEST
       Line3_TEST
       python_TEST
       Rand_TEST
       SignalStats_TEST
       Vector2_TEST
       Vector3_TEST
+      Vector3Stats_TEST
       Vector4_TEST
       Temperature_TEST
     )

src/python/Kmeans.i

@@ -0,0 +1,66 @@
+/*
+ * Copyright (C) 2021 Open Source Robotics Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *
+*/
+
+%module kmeans
+%{
+#include <ignition/math/Helpers.hh>
+#include <ignition/math/Vector3.hh>
+#include <ignition/math/Kmeans.hh>
+%}
+
+%include "std_vector.i"
+%template(vector_vector3d) std::vector<ignition::math::Vector3<double>>;
+%template(vector_uint) std::vector<unsigned int>;
+
+%inline %{
+  struct ClusterOutput {
+    bool result;
+    std::vector<ignition::math::Vector3<double>> centroids;
+    std::vector<unsigned int> labels;
+  };
+%}
+
+namespace ignition
+{
+  namespace math
+  {
+    class Kmeans
+    {
+      %rename("%(undercase)s", %$isfunction, %$ismember, %$not %$isconstructor) "";
+      public: explicit Kmeans(const std::vector<Vector3<double>> &_obs);
+      public: virtual ~Kmeans();
+      public: std::vector<Vector3<double>> Observations() const;
+      public: bool Observations(const std::vector<Vector3<double>> &_obs);
+      public: bool AppendObservations(const std::vector<Vector3<double>> &_obs);
+
+      %pythoncode %{
+      def cluster(self, _k):
+        cluster_output = self._cluster(_k)
+        return [cluster_output.result,
+                vector_vector3d(cluster_output.centroids),
+                vector_uint(cluster_output.labels)]
+      %}
+    };
+    %extend Kmeans{
+      inline ClusterOutput _cluster(int _k) {
+        ClusterOutput output;
+        output.result = (*$self).Cluster(_k, output.centroids, output.labels);
+        return output;
+      }
+    }
+  }
+}

src/python/Kmeans_TEST.py

@@ -0,0 +1,173 @@
+# Copyright (C) 2021 Open Source Robotics Foundation
+#
+# Licensed under the Apache License, Version 2.0 (the "License")
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#       http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import unittest
+from ignition.math import Kmeans
+from ignition.math import Vector3d
+from ignition.math import vector_vector3d
+
+
+class TestKmeans(unittest.TestCase):
+
+    def test_kmeans_constructor(self):
+        # Create some observations.
+        obs = list([])
+        obs.append(Vector3d(1.0, 1.0, 0.0))
+        obs.append(Vector3d(1.1, 1.0, 0.0))
+        obs.append(Vector3d(1.2, 1.0, 0.0))
+        obs.append(Vector3d(1.3, 1.0, 0.0))
+        obs.append(Vector3d(1.4, 1.0, 0.0))
+        obs.append(Vector3d(5.0, 1.0, 0.0))
+        obs.append(Vector3d(5.1, 1.0, 0.0))
+        obs.append(Vector3d(5.2, 1.0, 0.0))
+        obs.append(Vector3d(5.3, 1.0, 0.0))
+        obs.append(Vector3d(5.4, 1.0, 0.0))
+
+        # Initialize Kmeans with two partitions.
+        kmeans = Kmeans(obs)
+
+        # ::GetObservations()
+        obs_copy = list(kmeans.observations()).copy()
+        for i in range(len(obs_copy)):
+            self.assertEqual(obs_copy[i], obs[i])
+
+        for idx, a in enumerate(obs_copy):
+            obs_copy[idx] = a + Vector3d(0.1, 0.2, 0.0)
+
+        self.assertTrue(kmeans.observations(obs_copy))
+
+        obs_copy = list(kmeans.observations()).copy()
+        for i in range(len(obs_copy)):
+            self.assertEqual(obs_copy[i], obs[i] + Vector3d(0.1, 0.2, 0.0))
+        self.assertTrue(kmeans.observations(obs))
+
+        # ::Cluster()
+        result, centroids, labels = kmeans.cluster(2)
+        self.assertTrue(result)
+
+        # Check that there are two centroids.
+        self.assertEqual(len(centroids), 2)
+
+        # Check that the observations are clustered properly.
+        self.assertEqual(labels[0], labels[1])
+        self.assertEqual(labels[1], labels[2])
+        self.assertEqual(labels[2], labels[3])
+        self.assertEqual(labels[3], labels[4])
+
+        self.assertNotEqual(labels[4], labels[5])
+
+        self.assertEqual(labels[5], labels[6])
+        self.assertEqual(labels[6], labels[7])
+        self.assertEqual(labels[7], labels[8])
+        self.assertEqual(labels[8], labels[9])
+
+        # Check that there are two centroids.
+        self.assertEqual(centroids.size(), 2)
+        # Check that the observations are clustered properly.
+        self.assertEqual(labels[0], labels[1])
+        self.assertEqual(labels[1], labels[2])
+        self.assertEqual(labels[2], labels[3])
+        self.assertEqual(labels[3], labels[4])
+
+        self.assertNotEqual(labels[4], labels[5])
+
+        self.assertEqual(labels[5], labels[6])
+        self.assertEqual(labels[6], labels[7])
+        self.assertEqual(labels[7], labels[8])
+        self.assertEqual(labels[8], labels[9])
+
+        # Check that there are two centroids.
+        self.assertEqual(centroids.size(), 2)
+
+        # Check that the observations are clustered properly.
+        self.assertEqual(labels[0], labels[1])
+        self.assertEqual(labels[1], labels[2])
+        self.assertEqual(labels[2], labels[3])
+        self.assertEqual(labels[3], labels[4])
+
+        self.assertNotEqual(labels[4], labels[5])
+
+        self.assertEqual(labels[5], labels[6])
+        self.assertEqual(labels[6], labels[7])
+        self.assertEqual(labels[7], labels[8])
+        self.assertEqual(labels[8], labels[9])
+
+        # Check the centroids.
+        expected_centroid1 = Vector3d(1.2, 1.0, 0.0)
+        expected_centroid2 = Vector3d(5.2, 1.0, 0.0)
+        if (centroids[0] == expected_centroid1):
+            self.assertEqual(centroids[1], expected_centroid2)
+        elif (centroids[0] == expected_centroid2):
+            self.assertEqual(centroids[1], expected_centroid1)
+        else:
+            self.fail()
+
+        # Try to use an empty observation vector.
+        obs_copy.clear()
+        self.assertFalse(kmeans.observations(obs_copy))
+
+        # Try to call 'Cluster()' with an empty vector.
+        kmeansEmpty = Kmeans(obs_copy)
+        result, centroids, labels = kmeansEmpty.cluster(2)
+        self.assertFalse(result)
+
+        # Try to use a non positive k.
+        result, centroids, labels = kmeans.cluster(0)
+        self.assertFalse(result)
+
+        # Try to use a k > num_observations.
+        result, centroids, labels = kmeans.cluster(len(obs) + 1)
+        self.assertFalse(result)
+
+    def test_kmeans_append(self):
+        # Create some observations.
+        obs = list([])
+        obs2 = list([])
+        obs_total = list([])
+
+        obs.append(Vector3d(1.0, 1.0, 0.0))
+        obs.append(Vector3d(1.1, 1.0, 0.0))
+        obs.append(Vector3d(1.2, 1.0, 0.0))
+        obs.append(Vector3d(1.3, 1.0, 0.0))
+        obs.append(Vector3d(1.4, 1.0, 0.0))
+
+        obs2.append(Vector3d(5.0, 1.0, 0.0))
+        obs2.append(Vector3d(5.1, 1.0, 0.0))
+        obs2.append(Vector3d(5.2, 1.0, 0.0))
+        obs2.append(Vector3d(5.3, 1.0, 0.0))
+        obs2.append(Vector3d(5.4, 1.0, 0.0))
+
+        for elem in obs:
+            obs_total.append(elem)
+
+        for elem in obs2:
+            obs_total.append(elem)
+
+        # Initialize Kmeans with two partitions.
+        kmeans = Kmeans(obs)
+
+        kmeans.append_observations(obs2)
+
+        obs_copy = vector_vector3d(kmeans.observations())
+
+        for i in range(obs_copy.size()):
+            self.assertEqual(obs_total[i], obs_copy[i])
+
+        # Append an empty vector.
+        emptyVector = vector_vector3d()
+        self.assertFalse(kmeans.append_observations(emptyVector))
+
+
+if __name__ == '__main__':
+    unittest.main()

src/python/Vector3Stats.i

@@ -0,0 +1,51 @@
+/*
+ * Copyright (C) 2021 Open Source Robotics Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *
+*/
+
+%module vector3stats
+%{
+#include <ignition/math/Helpers.hh>
+#include <ignition/math/SignalStats.hh>
+#include <ignition/math/Vector3.hh>
+#include <ignition/math/Vector3Stats.hh>
+%}
+
+%include "std_string.i"
+
+namespace ignition
+{
+  namespace math
+  {
+    class Vector3Stats
+    {
+      %rename("%(undercase)s", %$isfunction, %$ismember, %$not %$isconstructor) "";
+      public: Vector3Stats();
+      public: ~Vector3Stats();
+      public: void InsertData(const Vector3<double> &_data);
+      public: bool InsertStatistic(const std::string &_name);
+      public: bool InsertStatistics(const std::string &_names);
+      public: void Reset();
+      public: const SignalStats &X() const;
+      public: const SignalStats &Y() const;
+      public: const SignalStats &Z() const;
+      public: const SignalStats &Mag() const;
+      public: SignalStats &X();
+      public: SignalStats &Y();
+      public: SignalStats &Z();
+      public: SignalStats &Mag();
+    };
+  }
+}