Draft for subset access of features

src/libshogun/features/Features.cpp

@@ -48,6 +48,9 @@ CFeatures::CFeatures(CFile* loader)
 
 CFeatures::~CFeatures()
 {
+	if (m_subset_idx)
+		delete[] m_subset_idx;
+
 	clean_preprocs();
 }
 
@@ -66,11 +69,16 @@ CFeatures::init(void)
 							 &num_preproc, "preprocessed",
 							 "Feature[i] is already preprocessed.");
 
+	m_parameters->add_vector(&m_subset_idx, &m_subset_len, "subset_idx", "Subset indices.");
+
+
 	properties = FP_NONE;
 	cache_size = 0;
 	preproc = NULL;
 	num_preproc = 0;
 	preprocessed = NULL;
+	m_subset_idx = NULL;
+	m_subset_len = 0;
 }
 
 /// set preprocessor
@@ -275,3 +283,34 @@ bool CFeatures::check_feature_compatibility(CFeatures* f)
 				(this->get_feature_type() == f->get_feature_type()));
 	return result;
 }
+
+void CFeatures::set_feature_subset(int32_t* subset_idx, int32_t subset_len)
+{
+	if (m_subset_idx)
+		delete[] m_subset_idx;
+
+	m_subset_idx=subset_idx;
+	m_subset_len=subset_len;
+}
+
+void CFeatures::remove_feature_subset()
+{
+	if (!m_subset_idx)
+	{
+		delete[] m_subset_idx;
+		m_subset_idx=NULL;
+		m_subset_len=0;
+	}
+}
+
+void CFeatures::get_feature_subset(int32_t** subset_idx, int32_t* subset_len)
+{
+	ASSERT(m_subset_idx);
+	int64_t length = sizeof(int32_t)*m_subset_len;
+
+	*subset_len=m_subset_len;
+	*subset_idx=(int32_t*)malloc(length);
+	if (!*subset_idx)
+		SG_ERROR("Allocating %ld bytes failes\n", length);
+	memcpy(*subset_idx, m_subset_idx, length);
+}

src/libshogun/features/Features.h

@@ -49,6 +49,9 @@ namespace shogun
  *   (dense matrices), CSparseFeatures (sparse matrices), CStringFeatures (a
  *   set of strings) from which all the specific features like CSimpleFeatures<float64_t>
  *   (dense real valued feature matrices) are derived.
+ *
+ *   It is possible to set/get/reset subset matrices.
+ *   These may be used by subclasses for access to subsets of features.
  */
 class CFeatures : public CSGObject
 {
@@ -149,13 +152,21 @@ class CFeatures : public CSGObject
 		 */
 		inline int32_t get_cache_size() { return cache_size; };
 
-		/** get number of examples/vectors
+		/** get number of examples/vectors, possibly corresponding to the current subset
 		 *
 		 * abstract base method
 		 *
-		 * @return number of examples/vectors
+		 * @return number of examples/vectors (possibly of subset, if implemented)
 		 */
-		virtual int32_t get_num_vectors()=0 ;
+		virtual int32_t get_num_vectors()=0;
+
+		/** get number of ALL examples/vectors (no subset here)
+		 *
+		 * abstract base method
+		 *
+		 * @ return number of examples/vectors (ALL, nosubset)
+		 */
+		//virtual int32_t get_num_vectors_all()=0;
 
 		/** in case there is a feature matrix allow for reshaping
 		 *
@@ -232,6 +243,45 @@ class CFeatures : public CSGObject
 			properties &= (properties | p) ^ p;
 		}
 
+		/** removes (and deletes) the current subset indices matrix */
+		virtual void remove_feature_subset();
+
+		/** getter for the subset indices
+		 *
+		 * @param m_subset_idx (copy of) subset indices matrix (returned)
+		 * @param m_subset_len (copy of) number ofsubset indices (returned)
+		 */
+		void get_feature_subset(int32_t** subset_idx, int32_t* subset_len);
+
+		/** getter for the subset indices
+		 *
+		 * @param m_subset_len reference to number of subset indices (returned)
+		 * @return subset indices array
+		 */
+		int32_t* get_feature_subset(int32_t& subset_len)
+		{
+			subset_len=m_subset_len;
+			return m_subset_idx;
+		}
+
+		/** sets the subset indices matrix which is afterwards used for feature access
+		 *
+		 * @param m_subset_idx index matrix
+		 * @param m_subset_len number of subset indices
+		 */
+		virtual void set_feature_subset(int32_t* subset_idx, int32_t subset_len);
+
+	protected:
+		/** returns the corresponding real index (in array) of a subset index
+		 * (if there is a subset)
+		 *
+		 * should/has to be be overwritten when subset support is
+		 * implemented in a subclass. for now, it just returns identity
+		 *
+		 * @ return array index of the provided subset index
+		 */
+		inline int32_t subset_idx_conversion(int32_t idx) { return idx; }
+
 	private:
 		/** feature properties */
 		uint64_t  properties;
@@ -247,6 +297,16 @@ class CFeatures : public CSGObject
 
 		/// i'th entry is true if features were already preprocessed with preproc i
 		bool* preprocessed;
+
+	protected:
+		/** (possibly) contains a matrix of indices that form a subset of the stored
+		 * features. If set, all vector access methods work on the specified subset.
+		 * If it is NULL, it is ignored.
+		 */
+		int32_t* m_subset_idx;
+
+		/** length of the subset */
+		int32_t m_subset_len;
 };
 }
 #endif