Merge pull request #4277 from shubham808/feature/refactor_labels_factory

revome methods from labels factory

src/shogun/labels/LabelsFactory.cpp

@@ -10,50 +10,6 @@
 
 using namespace shogun;
 
-CBinaryLabels* CLabelsFactory::to_binary(CLabels* base_labels)
-{
-	ASSERT(base_labels != NULL)
-	if (base_labels->get_label_type() == LT_BINARY)
-		return static_cast<CBinaryLabels*>(base_labels);
-	else
-		SG_SERROR("base_labels must be of dynamic type CBinaryLabels")
-
-	return NULL;
-}
-
-CLatentLabels* CLabelsFactory::to_latent(CLabels* base_labels)
-{
-	ASSERT(base_labels != NULL)
-	if (base_labels->get_label_type() == LT_LATENT)
-		return static_cast<CLatentLabels*>(base_labels);
-	else
-		SG_SERROR("base_labels must be of dynamic type CLatentLabels\n")
-
-	return NULL;
-}
-
-CMulticlassLabels* CLabelsFactory::to_multiclass(CLabels* base_labels)
-{
-	ASSERT(base_labels != NULL)
-	if (base_labels->get_label_type() == LT_MULTICLASS)
-		return static_cast<CMulticlassLabels*>(base_labels);
-	else
-		SG_SERROR("base_labels must be of dynamic type CMulticlassLabels\n")
-
-	return NULL;
-}
-
-CRegressionLabels* CLabelsFactory::to_regression(CLabels* base_labels)
-{
-	ASSERT(base_labels != NULL)
-	if (base_labels->get_label_type() == LT_REGRESSION)
-		return static_cast<CRegressionLabels*>(base_labels);
-	else
-		SG_SERROR("base_labels must be of dynamic type CRegressionLabels")
-
-	return NULL;
-}
-
 CStructuredLabels* CLabelsFactory::to_structured(CLabels* base_labels)
 {
 	ASSERT(base_labels != NULL)

src/shogun/labels/LabelsFactory.h

@@ -29,29 +29,6 @@ namespace shogun
 class CLabelsFactory : public CSGObject
 {
 public:
-	/** specialize a base class instance to CBinaryLabels
-	 *
-	 * @param base_labels its dynamic type must be CBinaryLabels
-	 */
-	static CBinaryLabels* to_binary(CLabels* base_labels);
-
-	/** specialize a base class instance to CLatentLabels
-	 *
-	 * @param base_labels its dynamic type must be CLatentLabels
-	 */
-	static CLatentLabels* to_latent(CLabels* base_labels);
-
-	/** specialize a base class instance to CMulticlassLabels
-	 *
-	 * @param base_labels its dynamic type must be CMulticlassLabels
-	 */
-	static CMulticlassLabels* to_multiclass(CLabels* base_labels);
-
-	/** specialize a base class instance to CRegressionLabels
-	 *
-	 * @param base_labels its dynamic type must be CRegressionLabels
-	 */
-	static CRegressionLabels* to_regression(CLabels* base_labels);
 
 	/** specialize a base class instance to CStructuredLabels
 	 *

src/shogun/latent/LatentModel.cpp

@@ -64,7 +64,7 @@ void CLatentModel::set_features(CLatentFeatures* feats)
 void CLatentModel::argmax_h(const SGVector<float64_t>& w)
 {
 	int32_t num = get_num_vectors();
-	CBinaryLabels* y = CLabelsFactory::to_binary(m_labels->get_labels());
+	CBinaryLabels* y = binary_labels(m_labels->get_labels());
 	ASSERT(num > 0)
 	ASSERT(num == m_labels->get_num_labels())
 

src/shogun/metric/LMNN.cpp

@@ -60,7 +60,7 @@ void CLMNN::train(SGMatrix<float64_t> init_transform)
 
 	// cast is safe, check_training_setup ensures features are dense
 	CDenseFeatures<float64_t>* x = static_cast<CDenseFeatures<float64_t>*>(m_features);
-	CMulticlassLabels* y = CLabelsFactory::to_multiclass(m_labels);
+	CMulticlassLabels* y = multiclass_labels(m_labels);
 	SG_DEBUG("%d input vectors with %d dimensions.\n", x->get_num_vectors(), x->get_num_features());
 
 	auto& L = init_transform;

src/shogun/metric/LMNNImpl.cpp

@@ -66,7 +66,7 @@ void CLMNNImpl::check_training_setup(
 
 void CLMNNImpl::check_maximum_k(CLabels* labels, int32_t k)
 {
-	CMulticlassLabels* y = CLabelsFactory::to_multiclass(labels);
+	CMulticlassLabels* y = multiclass_labels(labels);
 	SGVector<int32_t> int_labels = y->get_int_labels();
 
 	// back-up initial values because they will be overwritten by unique

src/shogun/multiclass/tree/CHAIDTree.cpp

@@ -100,14 +100,14 @@ CMulticlassLabels* CCHAIDTree::apply_multiclass(CFeatures* data)
 {
 	REQUIRE(data, "Data required for classification in apply_multiclass\n")
 
-	return CLabelsFactory::to_multiclass(apply_tree(data));
+	return apply_tree(data)->as<CMulticlassLabels>();
 }
 
 CRegressionLabels* CCHAIDTree::apply_regression(CFeatures* data)
 {
 	REQUIRE(data, "Data required for regression in apply_regression\n")
 
-	return CLabelsFactory::to_regression(apply_tree(data));
+	return apply_tree(data)->as<CRegressionLabels>();
 }
 
 void CCHAIDTree::set_weights(SGVector<float64_t> w)

tests/unit/base/Serialization_unittest.cc

@@ -99,7 +99,7 @@ TEST(Serialization, liblinear)
 	liblin->set_epsilon(1e-5);
 	liblin->train();
 
-	CBinaryLabels* pred = CLabelsFactory::to_binary(liblin->apply(test_feats));
+	CBinaryLabels* pred = liblin->apply(test_feats)->as<CBinaryLabels>();
 	for (int i = 0; i < num_samples; ++i)
 		EXPECT_EQ(ground_truth->get_int_label(i), pred->get_int_label(i));
 	SG_UNREF(pred);
@@ -119,7 +119,7 @@ TEST(Serialization, liblinear)
 	SG_UNREF(file);
 
 	/* classify with the deserialized model */
-	pred = CLabelsFactory::to_binary(liblin_loaded->apply(test_feats));
+	pred = liblin_loaded->apply(test_feats)->as<CBinaryLabels>();
 	for (int i = 0; i < num_samples; ++i)
 		EXPECT_EQ(ground_truth->get_int_label(i), pred->get_int_label(i));
 

tests/unit/clustering/kmeans_unittest.cc

@@ -43,7 +43,8 @@ TEST(KMeans, manual_center_initialization_test)
 	{
 		clustering->train(features);
 
-		CMulticlassLabels* result=CLabelsFactory::to_multiclass(clustering->apply());
+		CMulticlassLabels* result =
+		    clustering->apply()->as<CMulticlassLabels>();
 
 		EXPECT_EQ(0.000000, result->get_label(0));
 		EXPECT_EQ(0.000000, result->get_label(1));

tests/unit/multiclass/KNN_unittest.cc

@@ -39,23 +39,24 @@ TEST(KNN, brute_solver)
 	test.random(0, classes*num-1);
 
 	CMulticlassLabels* labels = new CMulticlassLabels(lab);
-	
+
 	CDenseFeatures< float64_t >* features = new CDenseFeatures< float64_t >(feat);
-	CFeatures* features_test = (CFeatures*) features->clone();	
+	CFeatures* features_test = (CFeatures*)features->clone();
 	CLabels* labels_test = (CLabels*) labels->clone();
 
 	int32_t k=4;
-	CEuclideanDistance* distance = new CEuclideanDistance();	
+	CEuclideanDistance* distance = new CEuclideanDistance();
 	CKNN* knn=new CKNN (k, distance, labels, KNN_BRUTE);
 	SG_REF(knn);
 
 	features->add_subset(train);
-	labels->add_subset(train);	
+	labels->add_subset(train);
 	knn->train(features);
 
 	features_test->add_subset(test);
 	labels_test->add_subset(test);
-	CMulticlassLabels* output=CLabelsFactory::to_multiclass(knn->apply(features_test));
+	CMulticlassLabels* output =
+	    knn->apply(features_test)->as<CMulticlassLabels>();
 	SG_REF(output);
 	features_test->remove_subset();
 
@@ -86,21 +87,22 @@ TEST(KNN, kdtree_solver)
 
 	CMulticlassLabels* labels = new CMulticlassLabels(lab);
 	CDenseFeatures< float64_t >* features = new CDenseFeatures< float64_t >(feat);
-	CFeatures* features_test = (CFeatures*) features->clone();	
+	CFeatures* features_test = (CFeatures*)features->clone();
 	CLabels* labels_test = (CLabels*) labels->clone();
 
 	int32_t k=4;
-	CEuclideanDistance* distance = new CEuclideanDistance();	
+	CEuclideanDistance* distance = new CEuclideanDistance();
 	CKNN* knn=new CKNN (k, distance, labels, KNN_KDTREE);
 	SG_REF(knn);
 
 	features->add_subset(train);
-	labels->add_subset(train);	
+	labels->add_subset(train);
 	knn->train(features);
 
 	features_test->add_subset(test);
 	labels_test->add_subset(test);
-	CMulticlassLabels* output=CLabelsFactory::to_multiclass(knn->apply(features_test));
+	CMulticlassLabels* output =
+	    knn->apply(features_test)->as<CMulticlassLabels>();
 	SG_REF(output);
 	features_test->remove_subset();
 
@@ -132,21 +134,22 @@ TEST(KNN, lsh_solver)
 
 	CMulticlassLabels* labels = new CMulticlassLabels(lab);
 	CDenseFeatures< float64_t >* features = new CDenseFeatures< float64_t >(feat);
-	CFeatures* features_test = (CFeatures*) features->clone();	
+	CFeatures* features_test = (CFeatures*)features->clone();
 	CLabels* labels_test = (CLabels*) labels->clone();
 
 	int32_t k=4;
-	CEuclideanDistance* distance = new CEuclideanDistance();	
+	CEuclideanDistance* distance = new CEuclideanDistance();
 	CKNN* knn=new CKNN (k, distance, labels, KNN_LSH);
 	SG_REF(knn);
 
 	features->add_subset(train);
-	labels->add_subset(train);	
+	labels->add_subset(train);
 	knn->train(features);
 
 	features_test->add_subset(test);
 	labels_test->add_subset(test);
-	CMulticlassLabels* output=CLabelsFactory::to_multiclass(knn->apply(features_test));
+	CMulticlassLabels* output =
+	    knn->apply(features_test)->as<CMulticlassLabels>();
 	SG_REF(output);
 	features_test->remove_subset();
 

tests/unit/multiclass/MCLDA_unittest.cc

@@ -28,7 +28,7 @@ TEST(MCLDA, train_and_apply)
 	SG_REF(lda);
 	lda->train();
 
-	CMulticlassLabels* output=CLabelsFactory::to_multiclass(lda->apply());
+	CMulticlassLabels* output = lda->apply()->as<CMulticlassLabels>();
 	SG_REF(output);
 
 	// Test

tests/unit/multiclass/QDA_unittest.cc

@@ -28,7 +28,7 @@ TEST(QDA, train_and_apply)
 	SG_REF(qda);
 	qda->train();
 
-	CMulticlassLabels* output=CLabelsFactory::to_multiclass(qda->apply());
+	CMulticlassLabels* output = qda->apply()->as<CMulticlassLabels>();
 	SG_REF(output);
 
 	// Test

tests/unit/regression/LibSVR_unittest.cc

@@ -58,8 +58,8 @@ TEST(LibSVR,epsilon_svr_apply)
 	svm->train();
 
 	/* predict */
-	CRegressionLabels* predicted_labels=CLabelsFactory::to_regression(
-			svm->apply(features_test));
+	CRegressionLabels* predicted_labels =
+	    svm->apply(features_test)->as<CRegressionLabels>();
 
 	/* LibSVM regression comparison (with easy.py script) */
 	EXPECT_NEAR(predicted_labels->get_labels()[0], 2.44343, 1E-5);
@@ -123,8 +123,8 @@ TEST(LibSVR,nu_svr_apply)
 	svm->train();
 
 	/* predict */
-	CRegressionLabels* predicted_labels=CLabelsFactory::to_regression(
-			svm->apply(features_test));
+	CRegressionLabels* predicted_labels =
+	    svm->apply(features_test)->as<CRegressionLabels>();
 
 	/* LibSVM regression comparison (with easy.py script) */
 	EXPECT_NEAR(predicted_labels->get_labels()[0], 2.18062, 1E-5);