Public Member Functions | Protected Member Functions | Protected Attributes

CStringFileFeatures< ST > Class Template Reference


Detailed Description

template<class ST>
class shogun::CStringFileFeatures< ST >

File based string features.

StringFeatures that are file based. Underneath memory mapped files are used. Derived from CStringFeatures thus transparently enabling all of the StringFeature functionality.

Supported file format contains one string per line, lines of variable length are supported and must be separated by '
'.

Definition at line 34 of file StringFileFeatures.h.

Inheritance diagram for CStringFileFeatures< ST >:
Inheritance graph
[legend]

List of all members.

Public Member Functions

 CStringFileFeatures ()
 CStringFileFeatures (const char *fname, EAlphabet alpha)
virtual ~CStringFileFeatures ()

Protected Member Functions

ST * get_line (uint64_t &len, uint64_t &offs, int32_t &line_nr, uint64_t file_length)
virtual void cleanup ()
virtual void cleanup_feature_vector (int32_t num)
void fetch_meta_info_from_file (int32_t granularity=1048576)

Protected Attributes

CMemoryMappedFile< ST > * file

Constructor & Destructor Documentation

default constructor

Definition at line 41 of file StringFileFeatures.h.

CStringFileFeatures ( const char *  fname,
EAlphabet  alpha 
)

constructor

Parameters:
fname filename of the file containing line based features
alpha alphabet (type) to use for string features

Definition at line 50 of file StringFileFeatures.h.

virtual ~CStringFileFeatures (  )  [virtual]

default destructor

Definition at line 60 of file StringFileFeatures.h.


Member Function Documentation

virtual void cleanup (  )  [protected, virtual]

cleanup string features

Reimplemented from CStringFeatures< ST >.

Definition at line 112 of file StringFileFeatures.h.

virtual void cleanup_feature_vector ( int32_t  num  )  [protected, virtual]

cleanup a single feature vector

Reimplemented from CStringFeatures< ST >.

Definition at line 131 of file StringFileFeatures.h.

void fetch_meta_info_from_file ( int32_t  granularity = 1048576  )  [protected]

obtain meta information from file

i.e., determine number of strings and their lengths

Definition at line 141 of file StringFileFeatures.h.

ST* get_line ( uint64_t &  len,
uint64_t &  offs,
int32_t &  line_nr,
uint64_t  file_length 
) [protected]

get next line from file

The returned line may be modfied in case the file was opened read/write. It is otherwise read-only.

Parameters:
len length of line (returned via reference)
offs offset to be passed for reading next line, should be 0 initially (returned via reference)
line_nr used to indicate errors (returned as reference should be 0 initially)
file_length total length of the file (for error checking)
Returns:
line (NOT ZERO TERMINATED)

Definition at line 81 of file StringFileFeatures.h.


Member Data Documentation

CMemoryMappedFile<ST>* file [protected]

memory mapped file

Definition at line 190 of file StringFileFeatures.h.


The documentation for this class was generated from the following file:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines

SHOGUN Machine Learning Toolbox - Documentation