Merge branch 'master' of git://github.com/karlnapf/shogun

examples/undocumented/libshogun/modelselection_apply_parameter_tree.cpp

@@ -29,7 +29,7 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* c=new CModelSelectionParameters("C1");
 	root->append_child(c);
-	c->build_values(1, 2, R_EXP);
+	c->build_values(1.0, 2.0, R_EXP);
 
 	CGaussianKernel* gaussian_kernel=new CGaussianKernel();
 	CModelSelectionParameters* param_gaussian_kernel=
@@ -39,7 +39,7 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* param_gaussian_kernel_width=
 			new CModelSelectionParameters("width");
-	param_gaussian_kernel_width->build_values(1, 2, R_EXP);
+	param_gaussian_kernel_width->build_values(1.0, 2.0, R_EXP);
 	param_gaussian_kernel->append_child(param_gaussian_kernel_width);
 
 	return root;

examples/undocumented/libshogun/modelselection_grid_search_kernel.cpp

@@ -36,18 +36,18 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* c1=new CModelSelectionParameters("C1");
 	root->append_child(c1);
-	c1->build_values(-5, 5, R_EXP);
+	c1->build_values(-5.0, 5.0, R_EXP);
 
 	CModelSelectionParameters* c2=new CModelSelectionParameters("C2");
 	root->append_child(c2);
-	c2->build_values(-5, 5, R_EXP);
+	c2->build_values(-5.0, 5.0, R_EXP);
 
 	CGaussianKernel* gaussian_kernel=new CGaussianKernel();
 	CModelSelectionParameters* param_gaussian_kernel=
 			new CModelSelectionParameters("kernel", gaussian_kernel);
 	CModelSelectionParameters* gaussian_kernel_width=
 			new CModelSelectionParameters("width");
-	gaussian_kernel_width->build_values(-5, 5, R_EXP, 1, 2);
+	gaussian_kernel_width->build_values(-5.0, 5.0, R_EXP, 1.0, 2.0);
 	param_gaussian_kernel->append_child(gaussian_kernel_width);
 	root->append_child(param_gaussian_kernel);
 
@@ -59,7 +59,7 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* param_power_kernel_degree=
 			new CModelSelectionParameters("degree");
-	param_power_kernel_degree->build_values(1, 1, R_EXP);
+	param_power_kernel_degree->build_values(1.0, 1.0, R_EXP);
 	param_power_kernel->append_child(param_power_kernel_degree);
 
 	CMinkowskiMetric* m_metric=new CMinkowskiMetric(10);
@@ -70,7 +70,7 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* param_power_kernel_metric1_k=
 			new CModelSelectionParameters("k");
-	param_power_kernel_metric1_k->build_values(1, 12, R_LINEAR);
+	param_power_kernel_metric1_k->build_values(1.0, 12.0, R_LINEAR);
 	param_power_kernel_metric1->append_child(param_power_kernel_metric1_k);
 
 	return root;

examples/undocumented/libshogun/modelselection_grid_search_linear.cpp

@@ -32,11 +32,11 @@ CModelSelectionParameters* create_param_tree()
 
 	CModelSelectionParameters* c1=new CModelSelectionParameters("C1");
 	root->append_child(c1);
-	c1->build_values(-15, 15, R_EXP);
+	c1->build_values(-15.0, 15.0, R_EXP);
 
 	CModelSelectionParameters* c2=new CModelSelectionParameters("C2");
 	root->append_child(c2);
-	c2->build_values(-15, 15, R_EXP);
+	c2->build_values(-15.0, 15.0, R_EXP);
 
 	return root;
 }

examples/undocumented/libshogun/modelselection_model_selection_parameters_test.cpp

@@ -29,7 +29,7 @@ CModelSelectionParameters* build_complex_example_tree()
 
 	CModelSelectionParameters* c=new CModelSelectionParameters("C");
 	root->append_child(c);
-	c->build_values(1, 1, R_EXP);
+	c->build_values(1.0, 1.0, R_EXP);
 
 	CPowerKernel* power_kernel=new CPowerKernel();
 	CModelSelectionParameters* param_power_kernel=
@@ -39,7 +39,7 @@ CModelSelectionParameters* build_complex_example_tree()
 
 	CModelSelectionParameters* param_power_kernel_degree=
 			new CModelSelectionParameters("degree");
-	param_power_kernel_degree->build_values(1, 1, R_EXP);
+	param_power_kernel_degree->build_values(1.0, 1.0, R_EXP);
 	param_power_kernel->append_child(param_power_kernel_degree);
 
 	CMinkowskiMetric* m_metric=new CMinkowskiMetric(10);
@@ -50,7 +50,7 @@ CModelSelectionParameters* build_complex_example_tree()
 
 	CModelSelectionParameters* param_power_kernel_metric1_k=
 			new CModelSelectionParameters("k");
-	param_power_kernel_metric1_k->build_values(1, 12, R_LINEAR);
+	param_power_kernel_metric1_k->build_values(1.0, 12.0, R_LINEAR);
 	param_power_kernel_metric1->append_child(param_power_kernel_metric1_k);
 
 	CGaussianKernel* gaussian_kernel=new CGaussianKernel();
@@ -61,7 +61,7 @@ CModelSelectionParameters* build_complex_example_tree()
 
 	CModelSelectionParameters* param_gaussian_kernel_width=
 			new CModelSelectionParameters("width");
-	param_gaussian_kernel_width->build_values(1, 2, R_EXP);
+	param_gaussian_kernel_width->build_values(1.0, 2.0, R_EXP);
 	param_gaussian_kernel->append_child(param_gaussian_kernel_width);
 
 	CDistantSegmentsKernel* ds_kernel=new CDistantSegmentsKernel();
@@ -72,12 +72,12 @@ CModelSelectionParameters* build_complex_example_tree()
 
 	CModelSelectionParameters* param_ds_kernel_delta=
 			new CModelSelectionParameters("delta");
-	param_ds_kernel_delta->build_values(1, 2, R_EXP);
+	param_ds_kernel_delta->build_values(1.0, 2.0, R_EXP);
 	param_ds_kernel->append_child(param_ds_kernel_delta);
 
 	CModelSelectionParameters* param_ds_kernel_theta=
 			new CModelSelectionParameters("theta");
-	param_ds_kernel_theta->build_values(1, 2, R_EXP);
+	param_ds_kernel_theta->build_values(1.0, 2.0, R_EXP);
 	param_ds_kernel->append_child(param_ds_kernel_theta);
 
 	return root;
@@ -95,7 +95,7 @@ CModelSelectionParameters* build_sgobject_no_childs_tree()
 CModelSelectionParameters* build_leaf_node_tree()
 {
 	CModelSelectionParameters* c_1=new CModelSelectionParameters("C1");
-	c_1->build_values(1, 1, R_EXP);
+	c_1->build_values(1.0, 1.0, R_EXP);
 
 	return c_1;
 }
@@ -111,11 +111,11 @@ CModelSelectionParameters* build_root_value_childs_tree()
 
 	CModelSelectionParameters* c_1=new CModelSelectionParameters("C1");
 	root->append_child(c_1);
-	c_1->build_values(1, 1, R_EXP);
+	c_1->build_values(1.0, 1.0, R_EXP);
 
 	CModelSelectionParameters* c_2=new CModelSelectionParameters("C2");
 	root->append_child(c_2);
-	c_2->build_values(1, 1, R_EXP);
+	c_2->build_values(1.0, 1.0, R_EXP);
 
 	return root;
 }
@@ -143,7 +143,7 @@ CModelSelectionParameters* build_root_sg_object_child_value_child_tree()
 
 	CModelSelectionParameters* c=new CModelSelectionParameters("C");
 	root->append_child(c);
-	c->build_values(1, 1, R_EXP);
+	c->build_values(1.0, 1.0, R_EXP);
 
 	root->append_child(param_power_kernel);
 

src/shogun/features/SimpleFeatures.h

@@ -969,6 +969,8 @@ template<class ST> class CSimpleFeatures: public CDotFeatures
 	/** Creates a new CFeatures instance containing copies of the elements
 	 * which are specified by the provided indices.
 	 *
+	 * possible with subset
+	 *
 	 * @param indices indices of feature elements to copy
 	 * @return new CFeatures instance with copies of feature data
 	 */

src/shogun/features/StringFeatures.h

@@ -2112,6 +2112,36 @@ template <class ST> class CStringFeatures : public CFeatures
 		}
 	*/
 
+		/** Creates a new CFeatures instance containing copies of the elements
+		 * which are specified by the provided indices.
+		 *
+		 * possible with subset
+		 *
+		 * @param indices indices of feature elements to copy
+		 * @return new CFeatures instance with copies of feature data
+		 */
+		virtual CFeatures* copy_subset(SGVector<index_t> indices) const
+		{
+			/* string list to create new CStringFeatures from */
+			SGStringList<ST> copy(indices.vlen, max_string_length);
+
+			/* copy all features */
+			for (index_t i=0; i<indices.vlen; ++i)
+			{
+				/* eventually check subset */
+				index_t real_idx=subset_idx_conversion(indices.vector[i]);
+				copy.strings[i]=features[real_idx];
+			}
+
+			/* create copy instance */
+			CStringFeatures* result=new CStringFeatures(copy, alphabet);
+
+			/* max string length may have changed */
+			result->determine_maximum_string_length();
+
+			return result;
+		}
+
 		/** @return object name */
 		inline virtual const char* get_name() const { return "StringFeatures"; }
 

src/shogun/lib/DataType.h

@@ -142,23 +142,78 @@ template<class T> class SGNDArray
 
 template<class T> struct SGString
 {
+public:
+	/** default constructor */
+	SGString() : string(NULL), length(0), do_free(false) { }
+
+	/** constructor for setting params */
+	SGString(T* s, index_t l, bool free_string=false)
+		: string(s), length(l), do_free(free_string) { }
+
+	/** constructor for setting params from a SGVector*/
+	SGString(SGVector<T> v)
+		: string(v.vector), length(v.vlen), do_free(v.do_free) { }
+
+	/** constructor to create new matrix in memory */
+	SGString(index_t len, bool free_string=false) :
+		length(len), do_free(free_string)
+	{
+		string=new T[len];
+	}
+
+	/** copy constructor */
+	SGString(const SGString &orig)
+		: string(orig.string), length(orig.length), do_free(orig.do_free) { }
+
+public:
 	/** string  */
 	T* string;
 	/** length of string  */
 	index_t length;
+	/** whether string needs to be freed */
+	bool do_free;
 };
 
 /** template class SGStringList */
 template <class T> struct SGStringList
 {
+public:
+	/** default constructor */
+	SGStringList() : strings(NULL), num_strings(0), max_string_length(0),
+		do_free(false) { }
+
+	/** constructor for setting params */
+	SGStringList(T* s, index_t num_s, index_t max_length, bool free_strings=false)
+		: strings(s), num_strings(num_s), max_string_length(max_length),
+		  do_free(free_strings) { }
+
+	/** constructor to create new matrix in memory */
+	SGStringList(index_t num_s, index_t max_length, bool free_strings=false)
+		: num_strings(num_s), max_string_length(max_length),
+		  do_free(free_strings)
+	{
+		strings=new SGString<T>[num_strings];
+	}
+
+	/** copy constructor */
+	SGStringList(const SGStringList &orig)
+		: num_strings(orig.num_strings),
+		  max_string_length(orig.max_string_length), strings(orig.strings),
+		  do_free(orig.do_free) { }
+
+
+public:
 	/* number of strings */
-	int32_t num_strings;
+	index_t num_strings;
 
 	/** length of longest string */
-	int32_t max_string_length;
+	index_t max_string_length;
 
 	/// this contains the array of features.
 	SGString<T>* strings;
+
+	/** whether vector needs to be freed */
+	bool do_free;
 };
 
 /** template class SGSparseVectorEntry */