Merge pull request #4198 from karlnapf/feature/factory

Feature/factory

doc/cookbook/source/examples/binary/kernel_support_vector_machine.rst

@@ -33,7 +33,7 @@ We then create a :sgclass:`CKernelMachine` instance, here :sgclass:`CLibSVM`, an
 
 .. sgexample:: kernel_support_vector_machine.sg:create_instance
 
-Then we train it on training data and apply it to test data, which here gives :sgclass:`CBinaryLabels`.
+Then we train it on training data and apply it to test data. This gives :sgclass:`CLabels`, which we can extract the label vector from.
 
 .. sgexample:: kernel_support_vector_machine.sg:train_and_apply
 

examples/meta/generator/targets/cpp.json

@@ -11,7 +11,7 @@
     "Comment": "//$comment\n",
     "Init": {
         "Construct": "auto $name = some<C$typeName>($arguments)$kwargs",
-        "Copy": "auto $name = $expr$kwargs",
+        "Copy": "auto $name = wrap($expr)$kwargs",
         "KeywordArguments": {
             "List": ";\n$elements",
             "Element": "$name->put(\"$keyword\", $expr)",

examples/meta/ruby/CMakeLists.txt

@@ -2,6 +2,7 @@ STRING(REGEX REPLACE "(.*)/narray.*$" "\\1" NARRAY_PATH ${NARRAY_LIB})
 
 LIST(APPEND EXCLUDED_RUBY_META_EXAMPLES
 			base_api-put_get
+      base_api-factory
 )
 
 # add test case for each generated example

examples/meta/src/base_api/factory.sg

@@ -0,0 +1,13 @@
+Machine lib_svm = machine("LibSVM")
+Machine lda = machine("LDA")
+Kernel kernel_gaussian = kernel("GaussianKernel")
+Kernel kernel_linear = kernel("LinearKernel")
+
+RealMatrix real_matrix(2,2)
+Features features_from_matrix = features(real_matrix)
+
+File features_file = csv_file("../../data/classifier_binary_2d_nonlinear_features_train.dat")
+Features features_from_file = features(features_file, enum EPrimitiveType.PT_FLOAT64)
+
+File labels_file = csv_file("../../data/classifier_binary_2d_nonlinear_labels_train.dat")
+Labels labels_from_file = labels(labels_file)

examples/meta/src/binary/kernel_support_vector_machine.sg

@@ -1,13 +1,13 @@
-CSVFile f_feats_train("../../data/classifier_binary_2d_nonlinear_features_train.dat")
-CSVFile f_feats_test("../../data/classifier_binary_2d_nonlinear_features_test.dat")
-CSVFile f_labels_train("../../data/classifier_binary_2d_nonlinear_labels_train.dat")
-CSVFile f_labels_test("../../data/classifier_binary_2d_nonlinear_labels_test.dat")
+File f_feats_train = csv_file("../../data/classifier_binary_2d_nonlinear_features_train.dat")
+File f_feats_test = csv_file("../../data/classifier_binary_2d_nonlinear_features_test.dat")
+File f_labels_train = csv_file("../../data/classifier_binary_2d_nonlinear_labels_train.dat")
+File f_labels_test = csv_file("../../data/classifier_binary_2d_nonlinear_labels_test.dat")
 
 #![create_features]
-RealFeatures features_train(f_feats_train)
-RealFeatures features_test(f_feats_test)
-BinaryLabels labels_train(f_labels_train)
-BinaryLabels labels_test(f_labels_test)
+Features features_train = features(f_feats_train)
+Features features_test = features(f_feats_test)
+Labels labels_train = labels(f_labels_train)
+Labels labels_test = labels(f_labels_test)
 #![create_features]
 
 #![set_parameters]
@@ -20,7 +20,8 @@ Machine svm = machine("LibSVM", C1=1.0, C2=1.0, kernel=k, labels=labels_train, e
 
 #![train_and_apply]
 svm.train(features_train)
-BinaryLabels labels_predict = svm.apply_binary(features_test)
+Labels labels_predict = svm.apply(features_test)
+RealVector labels = labels_predict.get_real_vector("labels")
 #![train_and_apply]
 
 #![extract_weights_bias]
@@ -33,5 +34,4 @@ AccuracyMeasure eval()
 real accuracy = eval.evaluate(labels_predict, labels_test)
 #![evaluate_accuracy]
 
-# additional integration testing variables
-RealVector output = labels_predict.get_real_vector("labels")
+

examples/meta/src/multiclass/gaussian_naive_bayes.sg

@@ -1,23 +1,20 @@
-CSVFile f_feats_train("../../data/classifier_4class_2d_linear_features_train.dat")
-CSVFile f_feats_test("../../data/classifier_4class_2d_linear_features_test.dat")
-CSVFile f_labels_train("../../data/classifier_4class_2d_linear_labels_train.dat")
+File f_feats_train = csv_file("../../data/classifier_4class_2d_linear_features_train.dat")
+File f_feats_test = csv_file("../../data/classifier_4class_2d_linear_features_test.dat")
+File f_labels_train = csv_file("../../data/classifier_4class_2d_linear_labels_train.dat")
 
 #![create_features]
-RealFeatures features_train(f_feats_train)
-RealFeatures features_test(f_feats_test)
-MulticlassLabels labels_train(f_labels_train)
+Features features_train = features(f_feats_train)
+Features features_test = features(f_feats_test)
+Labels labels_train = labels(f_labels_train)
 #![create_features]
 
 
 #![create_instance]
-GaussianNaiveBayes gnb(features_train, labels_train)
+Machine gnb = machine("GaussianNaiveBayes", features=features_train, labels=labels_train)
 #![create_instance]
 
 #![train_and_apply]
 gnb.train()
-MulticlassLabels labels_predict = gnb.apply_multiclass(features_test)
+Labels labels_predict = gnb.apply(features_test)
+RealVector labels = labels_predict.get_real_vector("labels")
 #![train_and_apply]
-
-
-# integration testing variables
-RealVector output = labels_predict.get_labels()

examples/meta/src/regression/support_vector_regression.sg

@@ -1,39 +1,36 @@
-CSVFile f_feats_train("../../data/regression_1d_sinc_features_train.dat")
-CSVFile f_feats_test("../../data/regression_1d_sinc_features_test.dat")
-CSVFile f_labels_train("../../data/regression_1d_sinc_labels_train.dat")
-CSVFile f_labels_test("../../data/regression_1d_sinc_labels_test.dat")
+File f_feats_train=csv_file("../../data/regression_1d_sinc_features_train.dat")
+File f_feats_test=csv_file("../../data/regression_1d_sinc_features_test.dat")
+File f_labels_train=csv_file("../../data/regression_1d_sinc_labels_train.dat")
+File f_labels_test=csv_file("../../data/regression_1d_sinc_labels_test.dat")
 
 #![create_features]
-RealFeatures features_train(f_feats_train)
-RealFeatures features_test(f_feats_test)
-RegressionLabels labels_train(f_labels_train)
-RegressionLabels labels_test(f_labels_test)
+Features features_train=features(f_feats_train)
+Features features_test=features(f_feats_test)
+Labels labels_train=labels(f_labels_train)
+Labels labels_test=labels(f_labels_test)
 #![create_features]
 
 #![create_appropriate_kernel]
-real width = 1.0
-GaussianKernel kernel(width)
+Kernel k=kernel("GaussianKernel", log_width=1.0)
 #![create_appropriate_kernel]
 
 #![create_instance]
-real svm_c = 1.0
-real svr_param = 0.1
-LibSVR svr(svm_c, svr_param, kernel, labels_train, enum LIBSVR_SOLVER_TYPE.LIBSVR_EPSILON_SVR)
+Machine svr=machine("LibSVR", C1=1.0, C2=1.0, tube_epsilon=0.1, kernel=k, labels=labels_train)
+# , solver_type=enum LIBSVR_SOLVER_TYPE.LIBSVR_EPSILON_SVR)
 #![create_instance]
 
 #![train_and_apply]
 svr.train(features_train)
-RegressionLabels labels_predict = svr.apply_regression(features_test)
+Labels labels_predict = svr.apply(features_test)
+RealVector output = labels_predict.get_real_vector("labels")
 #![train_and_apply]
 
 #![extract_alpha]
-RealVector alpha = svr.get_alphas()
+RealVector alpha = svr.get_real_vector("m_alpha")
 #![extract_alpha]
 
 #![evaluate_error]
 MeanSquaredError eval()
 real mse = eval.evaluate(labels_predict, labels_test)
 #![evaluate_error]
 
-# integration testing variables
-RealVector output = labels_test.get_labels()

examples/undocumented/libshogun/evaluation_cross_validation_classification.cpp

@@ -59,6 +59,7 @@ void test_cross_validation()
 		lab.vector[i]=i<num_vectors/2 ? -1.0 : 1.0;
 
 	CBinaryLabels* labels=new CBinaryLabels(lab);
+	SG_REF(labels);
 
 	/* gaussian kernel */
 	int32_t kernel_cache=100;
@@ -75,6 +76,7 @@ void test_cross_validation()
 	/* train and output */
 	svm->train(features);
 	CBinaryLabels* output=CLabelsFactory::to_binary(svm->apply(features));
+	SG_REF(output);
 	for (index_t i=0; i<num_vectors; ++i)
 		SG_SPRINT("i=%d, class=%f,\n", i, output->get_label(i));
 
@@ -113,6 +115,8 @@ void test_cross_validation()
 	result->print_result();
 
 	/* clean up */
+	SG_UNREF(labels);
+	SG_UNREF(output)
 	SG_UNREF(result);
 	SG_UNREF(cross);
 	SG_UNREF(features);
@@ -122,8 +126,6 @@ int main(int argc, char **argv)
 {
 	init_shogun(&print_message, &print_message, &print_message);
 
-	sg_io->set_loglevel(MSG_DEBUG);
-
 	test_cross_validation();
 
 	exit_shogun();

examples/undocumented/libshogun/evaluation_cross_validation_locked_comparison.cpp

@@ -58,6 +58,7 @@ void test_cross_validation()
 		lab.vector[i]=i<num_vectors/2 ? -1.0 : 1.0;
 
 	CBinaryLabels* labels=new CBinaryLabels(lab);
+	SG_REF(labels);
 
 	/* gaussian kernel */
 	CGaussianKernel* kernel=new CGaussianKernel();
@@ -74,6 +75,7 @@ void test_cross_validation()
 	SG_SPRINT("starting normal training\n");
 	svm->train(features);
 	CBinaryLabels* output=CLabelsFactory::to_binary(svm->apply(features));
+	SG_REF(output);
 
 	/* evaluation criterion */
 	CContingencyTableEvaluation* eval_crit=
@@ -147,6 +149,8 @@ void test_cross_validation()
 	SG_SPRINT("%f sec\n", time.cur_time_diff());
 
 	/* clean up */
+	SG_UNREF(labels);
+	SG_UNREF(output);
 	SG_UNREF(cross);
 	SG_UNREF(features);
 }

examples/undocumented/libshogun/evaluation_cross_validation_regression.cpp

@@ -64,6 +64,7 @@ void test_cross_validation()
 	/* train and output */
 	krr->train(features);
 	CRegressionLabels* output= CLabelsFactory::to_regression(krr->apply());
+	SG_REF(output);
 	for (index_t i=0; i<num_vectors; ++i)
 	{
 		SG_SPRINT("x=%f, train=%f, predict=%f\n", train_dat.matrix[i],

examples/undocumented/libshogun/kernel_machine_train_locked.cpp

@@ -90,6 +90,7 @@ void test()
 	SGVector<index_t>::display_vector(indices.vector, indices.vlen, "training indices");
 	svm->train_locked(indices);
 	CBinaryLabels* output=CLabelsFactory::to_binary(svm->apply());
+	SG_REF(output);
 	SGVector<float64_t>::display_vector(output->get_labels().vector, output->get_num_labels(), "apply() output");
 	SGVector<float64_t>::display_vector(labels->get_labels().vector, labels->get_labels().vlen, "training labels");
 	SG_SPRINT("accuracy: %f\n", eval->evaluate(output, labels));
@@ -103,6 +104,7 @@ void test()
 	indices.vector[2]=3;
 	SGVector<index_t>::display_vector(indices.vector, indices.vlen, "training indices");
 	output=CLabelsFactory::to_binary(svm->apply());
+	SG_REF(output);
 	SGVector<float64_t>::display_vector(output->get_labels().vector, output->get_num_labels(), "apply() output");
 	SGVector<float64_t>::display_vector(labels->get_labels().vector, labels->get_labels().vlen, "training labels");
 	SG_SPRINT("accuracy: %f\n", eval->evaluate(output, labels));
@@ -115,6 +117,7 @@ void test()
 	SGVector<index_t>::display_vector(indices.vector, indices.vlen, "training indices");
 	svm->train_locked(indices);
 	output=CLabelsFactory::to_binary(svm->apply());
+	SG_REF(output);
 	SGVector<float64_t>::display_vector(output->get_labels().vector, output->get_num_labels(), "apply() output");
 	SGVector<float64_t>::display_vector(labels->get_labels().vector, labels->get_labels().vlen, "training labels");
 	SG_SPRINT("accuracy: %f\n", eval->evaluate(output, labels));
@@ -125,6 +128,7 @@ void test()
 	svm->data_unlock();
 	svm->train();
 	output=CLabelsFactory::to_binary(svm->apply());
+	SG_REF(output);
 	ASSERT(eval->evaluate(output, labels)==1);
 	SGVector<float64_t>::display_vector(output->get_labels().vector, output->get_num_labels(), "output");
 	SGVector<float64_t>::display_vector(labels->get_labels().vector, labels->get_labels().vlen, "training labels");

examples/undocumented/libshogun/regression_libsvr.cpp

@@ -38,7 +38,9 @@ void test_libsvr()
 
 	/* shogun representation */
 	CLabels* labels_train=new CRegressionLabels(lab_train);
+	SG_REF(labels_train);
 	CLabels* labels_test=new CRegressionLabels(lab_test);
+	SG_REF(labels_test);
 	CDenseFeatures<float64_t>* features_train=new CDenseFeatures<float64_t>(
 			feat_train);
 	CDenseFeatures<float64_t>* features_test=new CDenseFeatures<float64_t>(
@@ -55,6 +57,7 @@ void test_libsvr()
 	/* predict */
 	CRegressionLabels* predicted_labels=CLabelsFactory::to_regression(
 			svm->apply(features_test));
+	SG_REF(predicted_labels);
 
 	/* evaluate */
 	CEvaluation* eval=new CMeanSquaredError();
@@ -66,6 +69,7 @@ void test_libsvr()
 	SG_UNREF(labels_test)
 	SG_UNREF(predicted_labels);
 	SG_UNREF(svm);
+	SG_UNREF(labels_train);
 }
 
 int main()

src/interfaces/swig/Kernel.i

@@ -8,8 +8,6 @@
  * Copyright (C) 2009 Fraunhofer Institute FIRST and Max-Planck-Society
  */
 
- %newobject kernel();
-
 #ifdef HAVE_PYTHON
 %feature("autodoc", "get_kernel_matrix(self) -> numpy 2dim array of float") get_kernel_matrix;
 %feature("autodoc", "get_POIM2(self) -> [] of float") get_POIM2;

src/interfaces/swig/Machine.i

@@ -23,7 +23,6 @@
 %newobject apply_structured(CFeatures* data);
 %newobject apply_latent();
 %newobject apply_latent(CFeatures* data);
-%newobject machine();
 
 #if defined(SWIGPYTHON) || defined(SWIGOCTAVE) || defined(SWIGRUBY) || defined(SWIGLUA) || defined(SWIGR)
 

src/interfaces/swig/factory.i

@@ -0,0 +1,6 @@
+%{
+    #include <shogun/util/factory.h>
+%}
+%include <shogun/util/factory.h>
+
+%template(features) shogun::features<float64_t>;

src/interfaces/swig/shogun.i

@@ -115,14 +115,18 @@
 %include "Boost.i"
 
 %include "ParameterObserver.i"
+%include "factory.i"
 
 #if defined(SWIGPERL)
 %include "abstract_types_extension.i"
 #endif
 
+%pragma(java) moduleimports=%{
+import org.jblas.*;
+%}
+
 namespace shogun
 {
-
 %extend CSGObject
 {
 	template <typename T, typename U= typename std::enable_if_t<std::is_arithmetic<T>::value>>

src/shogun/base/SGObject.h

@@ -492,36 +492,35 @@ class CSGObject
 	 */
 	std::vector<std::string> parameter_names() const;
 
-	/**
-	 * Utility method to specialize the feature to the required type.
+	/** Specializes a provided object to the specified type.
+	 * Throws exception if the object cannot be specialized.
 	 *
-	 * @param sgo CSGObject base type
-	 * @return The requested type if casting was successful.
+	 * @param sgo object of CSGObject base type
+	 * @return The requested type
 	 */
 	template<class T> static T* as(CSGObject* sgo)
 	{
 		REQUIRE(sgo, "No object provided!\n");
 		return sgo->as<T>();
 	}
 
-	/**
-	 * Utility method to specialize the feature to the required type.
+	/** Specializes the object to the specified type.
+	 * Throws exception if the object cannot be specialized.
 	 *
-	 * @param sgo CSGObject base type
-	 * @return The requested type if casting was successful, or throws exception.
+	 * @return The requested type
 	 */
 	template<class T> T* as()
 	{
-		T* c = dynamic_cast<T*>(this);
+		auto c = dynamic_cast<T*>(this);
 		if (c)
 			return c;
 
-		SG_SERROR("The object (%s) cannot be casted to the requested type %s!\n",
+		SG_SERROR(
+			"Object of type %s cannot be converted to type %s.\n",
 			demangled_type<std::remove_pointer_t<decltype(this)>>().c_str(),
 			demangled_type<T>().c_str());
 		return nullptr;
 	}
-
 #ifndef SWIG
 	/**
 	  * Get parameters observable