SHOGUN: HDF5File.h Source File

Go to the documentation of this file.
00001 /*
00002  * This program is free software; you can redistribute it and/or modify
00003  * it under the terms of the GNU General Public License as published by
00004  * the Free Software Foundation; either version 3 of the License, or
00005  * (at your option) any later version.
00006  *
00007  * Written (W) 2010 Soeren Sonnenburg
00008  * Copyright (C) 2010 Berlin Institute of Technology
00009  */
00010 #ifndef __HDF5_FILE_H__
00011 #define __HDF5_FILE_H__
00012 
00013 #include <shogun/lib/config.h>
00014 
00015 #ifdef HAVE_HDF5
00016 #include <shogun/lib/common.h>
00017 #include <shogun/io/SGIO.h>
00018 #include <shogun/io/File.h>
00019 #include <shogun/base/SGObject.h>
00020 #include <hdf5.h>
00021 
00022 
00023 namespace shogun
00024 {
00025 template <class ST> struct SGString;
00026 template <class ST> struct SGSparseVector;
00027 
00034 #define IGNORE_IN_CLASSLIST
00035 IGNORE_IN_CLASSLIST class CHDF5File : public CFile
00036 {
00037 public:
00039     CHDF5File(void);
00040 
00047     CHDF5File(char* fname, char rw='r', const char* name=NULL);
00048 
00050     virtual ~CHDF5File();
00051 
00059     virtual void get_vector(bool*& vector, int32_t& len);
00060     virtual void get_vector(uint8_t*& vector, int32_t& len);
00061     virtual void get_vector(char*& vector, int32_t& len);
00062     virtual void get_vector(int32_t*& vector, int32_t& len);
00063     virtual void get_vector(float64_t*& vector, int32_t& len);
00064     virtual void get_vector(float32_t*& vector, int32_t& len);
00065     virtual void get_vector(int16_t*& vector, int32_t& len);
00066     virtual void get_vector(uint16_t*& vector, int32_t& len);
00068 
00077     virtual void get_matrix(
00078             bool*& matrix, int32_t& num_feat, int32_t& num_vec);
00079     virtual void get_matrix(
00080             uint8_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00081     virtual void get_matrix(
00082             char*& matrix, int32_t& num_feat, int32_t& num_vec);
00083     virtual void get_matrix(
00084             int32_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00085     virtual void get_uint_matrix(
00086             uint32_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00087     virtual void get_long_matrix(
00088             int64_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00089     virtual void get_ulong_matrix(
00090             uint64_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00091     virtual void get_matrix(
00092             float32_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00093     virtual void get_matrix(
00094             float64_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00095     virtual void get_longreal_matrix(
00096             floatmax_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00097     virtual void get_matrix(
00098             int16_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00099     virtual void get_matrix(
00100             uint16_t*& matrix, int32_t& num_feat, int32_t& num_vec);
00102 
00111     virtual void get_ndarray(
00112             uint8_t*& array, int32_t*& dims, int32_t& num_dims);
00113     virtual void get_ndarray(
00114             char*& array, int32_t*& dims, int32_t& num_dims);
00115     virtual void get_ndarray(
00116             int32_t*& array, int32_t*& dims, int32_t& num_dims);
00117     virtual void get_ndarray(
00118             float32_t*& array, int32_t*& dims, int32_t& num_dims);
00119     virtual void get_ndarray(
00120             float64_t*& array, int32_t*& dims, int32_t& num_dims);
00121     virtual void get_ndarray(
00122             int16_t*& array, int32_t*& dims, int32_t& num_dims);
00123     virtual void get_ndarray(
00124             uint16_t*& array, int32_t*& dims, int32_t& num_dims);
00126 
00135     virtual void get_sparse_matrix(
00136             SGSparseVector<bool>*& matrix, int32_t& num_feat, int32_t& num_vec);
00137     virtual void get_sparse_matrix(
00138             SGSparseVector<uint8_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00139     virtual void get_sparse_matrix(
00140             SGSparseVector<char>*& matrix, int32_t& num_feat, int32_t& num_vec);
00141     virtual void get_sparse_matrix(
00142             SGSparseVector<int32_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00143     virtual void get_uint_sparsematrix(
00144             SGSparseVector<uint32_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00145     virtual void get_long_sparsematrix(
00146             SGSparseVector<int64_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00147     virtual void get_ulong_sparsematrix(
00148             SGSparseVector<uint64_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00149     virtual void get_sparse_matrix(
00150             SGSparseVector<int16_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00151     virtual void get_sparse_matrix(
00152             SGSparseVector<uint16_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00153     virtual void get_sparse_matrix(
00154             SGSparseVector<float32_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00155     virtual void get_sparse_matrix(
00156             SGSparseVector<float64_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00157     virtual void get_longreal_sparsematrix(
00158             SGSparseVector<floatmax_t>*& matrix, int32_t& num_feat, int32_t& num_vec);
00160 
00161 
00170     virtual void get_string_list(
00171             SGString<bool>*& strings, int32_t& num_str,
00172             int32_t& max_string_len);
00173     virtual void get_string_list(
00174             SGString<uint8_t>*& strings, int32_t& num_str,
00175             int32_t& max_string_len);
00176     virtual void get_string_list(
00177             SGString<char>*& strings, int32_t& num_str,
00178             int32_t& max_string_len);
00179     virtual void get_string_list(
00180             SGString<int32_t>*& strings, int32_t& num_str,
00181             int32_t& max_string_len);
00182     virtual void get_uint_string_list(
00183             SGString<uint32_t>*& strings, int32_t& num_str,
00184             int32_t& max_string_len);
00185     virtual void get_string_list(
00186             SGString<int16_t>*& strings, int32_t& num_str,
00187             int32_t& max_string_len);
00188     virtual void get_string_list(
00189             SGString<uint16_t>*& strings, int32_t& num_str,
00190             int32_t& max_string_len);
00191     virtual void get_long_string_list(
00192             SGString<int64_t>*& strings, int32_t& num_str,
00193             int32_t& max_string_len);
00194     virtual void get_ulong_string_list(
00195             SGString<uint64_t>*& strings, int32_t& num_str,
00196             int32_t& max_string_len);
00197     virtual void get_string_list(
00198             SGString<float32_t>*& strings, int32_t& num_str,
00199             int32_t& max_string_len);
00200     virtual void get_string_list(
00201             SGString<float64_t>*& strings, int32_t& num_str,
00202             int32_t& max_string_len);
00203     virtual void get_longreal_string_list(
00204             SGString<floatmax_t>*& strings, int32_t& num_str,
00205             int32_t& max_string_len);
00207 
00215     virtual void set_vector(const bool* vector, int32_t len);
00216     virtual void set_vector(const uint8_t* vector, int32_t len);
00217     virtual void set_vector(const char* vector, int32_t len);
00218     virtual void set_vector(const int32_t* vector, int32_t len);
00219     virtual void set_vector( const float32_t* vector, int32_t len);
00220     virtual void set_vector(const float64_t* vector, int32_t len);
00221     virtual void set_vector(const int16_t* vector, int32_t len);
00222     virtual void set_vector(const uint16_t* vector, int32_t len);
00224 
00225 
00233     virtual void set_matrix(
00234             const bool* matrix, int32_t num_feat, int32_t num_vec);
00235     virtual void set_matrix(
00236             const uint8_t* matrix, int32_t num_feat, int32_t num_vec);
00237     virtual void set_matrix(
00238             const char* matrix, int32_t num_feat, int32_t num_vec);
00239     virtual void set_matrix(
00240             const int32_t* matrix, int32_t num_feat, int32_t num_vec);
00241     virtual void set_uint_matrix(
00242             const uint32_t* matrix, int32_t num_feat, int32_t num_vec);
00243     virtual void set_long_matrix(
00244             const int64_t* matrix, int32_t num_feat, int32_t num_vec);
00245     virtual void set_ulong_matrix(
00246             const uint64_t* matrix, int32_t num_feat, int32_t num_vec);
00247     virtual void set_matrix(
00248             const float32_t* matrix, int32_t num_feat, int32_t num_vec);
00249     virtual void set_matrix(
00250             const float64_t* matrix, int32_t num_feat, int32_t num_vec);
00251     virtual void set_longreal_matrix(
00252             const floatmax_t* matrix, int32_t num_feat, int32_t num_vec);
00253     virtual void set_matrix(
00254             const int16_t* matrix, int32_t num_feat, int32_t num_vec);
00255     virtual void set_matrix(
00256             const uint16_t* matrix, int32_t num_feat, int32_t num_vec);
00258 
00266     virtual void set_sparse_matrix(
00267             const SGSparseVector<bool>* matrix, int32_t num_feat, int32_t num_vec);
00268     virtual void set_sparse_matrix(
00269             const SGSparseVector<uint8_t>* matrix, int32_t num_feat, int32_t num_vec);
00270     virtual void set_sparse_matrix(
00271             const SGSparseVector<char>* matrix, int32_t num_feat, int32_t num_vec);
00272     virtual void set_sparse_matrix(
00273             const SGSparseVector<int32_t>* matrix, int32_t num_feat, int32_t num_vec);
00274     virtual void set_uint_sparsematrix(
00275             const SGSparseVector<uint32_t>* matrix, int32_t num_feat, int32_t num_vec);
00276     virtual void set_long_sparsematrix(
00277             const SGSparseVector<int64_t>* matrix, int32_t num_feat, int32_t num_vec);
00278     virtual void set_ulong_sparsematrix(
00279             const SGSparseVector<uint64_t>* matrix, int32_t num_feat, int32_t num_vec);
00280     virtual void set_sparse_matrix(
00281             const SGSparseVector<int16_t>* matrix, int32_t num_feat, int32_t num_vec);
00282     virtual void set_sparse_matrix(
00283             const SGSparseVector<uint16_t>* matrix, int32_t num_feat, int32_t num_vec); 
00284     virtual void set_sparse_matrix(
00285             const SGSparseVector<float32_t>* matrix, int32_t num_feat, int32_t num_vec);
00286     virtual void set_sparse_matrix(
00287             const SGSparseVector<float64_t>* matrix, int32_t num_feat, int32_t num_vec);
00288     virtual void set_longreal_sparsematrix(
00289             const SGSparseVector<floatmax_t>* matrix, int32_t num_feat, int32_t num_vec);
00291 
00292 
00301     virtual void set_string_list(
00302             const SGString<bool>* strings, int32_t num_str);
00303     virtual void set_string_list(
00304             const SGString<uint8_t>* strings, int32_t num_str);
00305     virtual void set_string_list(
00306             const SGString<char>* strings, int32_t num_str);
00307     virtual void set_string_list(
00308             const SGString<int32_t>* strings, int32_t num_str);
00309     virtual void set_uint_string_list(
00310             const SGString<uint32_t>* strings, int32_t num_str);
00311     virtual void set_string_list(
00312             const SGString<int16_t>* strings, int32_t num_str);
00313     virtual void set_string_list(
00314             const SGString<uint16_t>* strings, int32_t num_str);
00315     virtual void set_long_string_list(
00316             const SGString<int64_t>* strings, int32_t num_str);
00317     virtual void set_ulong_string_list(
00318             const SGString<uint64_t>* strings, int32_t num_str);
00319     virtual void set_string_list(
00320             const SGString<float32_t>* strings, int32_t num_str);
00321     virtual void set_string_list(
00322             const SGString<float64_t>* strings, int32_t num_str);
00323     virtual void set_longreal_string_list(
00324             const SGString<floatmax_t>* strings, int32_t num_str);
00326 
00327     virtual void get_int8_matrix(int8_t*&, int32_t&, int32_t&)
00328     {
00329         SG_NOTIMPLEMENTED;
00330     }
00331     virtual void get_int8_sparsematrix(shogun::SGSparseVector<signed char>*&, int32_t&, int32_t&)
00332     {
00333         SG_NOTIMPLEMENTED;
00334     }
00335     virtual void get_int8_string_list(shogun::SGString<signed char>*&, int32_t&, int32_t&)
00336     {
00337         SG_NOTIMPLEMENTED;
00338     }
00339     virtual void set_int8_matrix(const int8_t*, int32_t, int32_t)
00340     {
00341         SG_NOTIMPLEMENTED;
00342     }
00343     virtual void set_int8_sparsematrix(const shogun::SGSparseVector<signed char>*, int32_t, int32_t)
00344     {
00345         SG_NOTIMPLEMENTED;
00346     }
00347     virtual void set_int8_string_list(const shogun::SGString<signed char>*, int32_t)
00348     {
00349         SG_NOTIMPLEMENTED;
00350     }
00351 
00352 
00354     inline virtual const char* get_name() const { return "HDF5File"; }
00355 
00356 protected:
00358     void get_boolean_type();
00359 
00368     hid_t get_compatible_type(H5T_class_t t_class,
00369                               const TSGDataType* datatype);
00370 
00378     void get_dims(hid_t dataset, int32_t*& dims, int32_t& ndims, int64_t& total_elements);
00379 
00381     void create_group_hierarchy();
00382 
00383 protected:
00385     hid_t h5file;
00387     hid_t boolean_type;
00388 };
00389 }
00390 #endif //  HAVE_HDF5
00391 #endif //__HDF5_FILE_H__
00392