Provides a detailed description of the attributes of a dataset. More...

#include <FileDesc.h>

Public Types
typedef FileDesc *const	FileDescConstPtr

typedef FileDesc *	FileDescPtr

Public Member Functions
	FileDesc ()

void	AddAAttribute (const KKB::KKStr &_name, KKMLL::AttributeType _type, bool &alreadyExists)

void	AddAAttribute (const KKMLL::Attribute &attribute)

void	AddANominalValue (kkint32 fieldNum, const KKStr &nominalValue, bool &alreadyExist, RunLog &log)

void	AddANominalValue (const KKStr &nominalValue, bool &alreadyExist, RunLog &log)

void	AddANominalValue (const KKStr &attributeName, const KKStr &nominalValue, bool &alreadyExist, RunLog &log)

void	AddAttributes (const KKMLL::AttributeList &attributes)

void	AddClasses (const MLClassList &classesToAdd)

bool	AllFieldsAreNumeric () const
	Allows the user to quickly determine if there are no nominal fields. More...

const KKMLL::AttributeList &	Attributes () const

const AttributeTypeVector &	AttributeVector () const

kkint32	Cardinality (kkint32 fieldNum) const

const VectorInt32 &	CardinalityVector () const

const MLClassList &	Classes () const

const KKStr &	ClassNameAttribute () const

void	ClassNameAttribute (const KKStr &_classNameAttribute)

const KKMLL::AttributePtr *	CreateAAttributeTable () const

AttributeTypeVector	CreateAttributeTypeTable () const

VectorInt32	CreateCardinalityTable () const

void	DisplayAttributeMappings ()

const KKStr &	FieldName (kkint32 fieldNum) const

const KKStr &	FileName () const

void	FileName (const KKStr &_fileName)

const KKMLL::Attribute &	GetAAttribute (kkint32 fieldNum) const

kkint32	GetFieldNumFromAttributeName (const KKStr &attributeName) const

MLClassPtr	GetMLClassPtr (const KKStr &className)

const KKStr &	GetNominalValue (kkint32 fieldNum, kkint32 code) const

const KKMLL::AttributePtr	LookUpByName (const KKStr &attributeName) const

MLClassPtr	LookUpMLClassByName (const KKStr &className)

kkint32	LookUpNominalCode (kkint32 fieldNum, const KKStr &nominalValue) const

MLClassPtr	LookUpUnKnownMLClass ()

kkint32	MemoryConsumedEstimated () const

kkuint32	NumOfFields () const

bool	operator!= (const FileDesc &rightSize) const
	Returns true if file description on the right size is NOT identical. More...

bool	operator== (const FileDesc &rightSize) const
	Returns true if file description on the right size is identical. More...

void	ReadXML (XmlStream &s, XmlTagConstPtr tag, VolConstBool &cancelFlag, RunLog &log)

bool	SameExceptForSymbolicData (const FileDesc &otherFd, RunLog &log) const

kkint32	SparseMinFeatureNum () const

void	SparseMinFeatureNum (kkint32 _sparseMinFeatureNum)

KKMLL::AttributeType	Type (kkint32 fieldNum) const

KKStr	TypeStr (kkint32 fieldNum) const

kkint16	Version () const

void	Version (kkint16 _version)

void	WriteXML (const KKStr &varName, std::ostream &o) const

Static Public Member Functions
static void	FinalCleanUp ()
	Clean up function, call just before exiting the application. More...

static bool	FinalCleanUpRanAlready ()

static FileDescPtr	GetExistingFileDesc (FileDescPtr fileDesc)
	Returns a pointer to an existing instance of 'fileDesc' if it exists, otherwise will use one being passed in. More...

static FileDescPtr	MergeSymbolicFields (const FileDesc &left, const FileDesc &right, RunLog &log)
	Merges the Symbolic fields of two different 'FileDesc' instances producing a new instance of 'FileDesc'. More...

static FileDescPtr	NewContinuousDataOnly (VectorKKStr &_fieldNames)
	Creates a simple FileDesc that consists of continuous data only. More...

Protected Member Functions
	~FileDesc ()

Friends
class	KKQueue< FileDesc >

Detailed Description

Provides a detailed description of the attributes of a dataset.

Author: Kurt Kramer

Describes the different fields of a dataset. There will be one instance of this class for each type of Dataset that you have in an application. From this object you can get information such as number of attributes, Attribute types, weather they are nominal, continuous. If nominal what are the accepted values. *

Never delete an instance of a FileDesc object.

Only one FileDesc object can exist for any Dataset. Example the Forest-Cover dataset. You can split the data into many files and manage them separately but you will only have one instance of a FileDesc object that they will all refer to. See "GetExistingFileDesc" method below. You would initially create an instance of FileDesc and then use "GetExistingFileDesc" to make sure that it is unique. This typically happens in the FeatureFileIO derived classes.

See also: GetExistingFileDesc; FeatureVectorList; FeatureFileIo

Definition at line 72 of file FileDesc.h.

Member Typedef Documentation

typedef FileDesc* const KKMLL::FileDesc::FileDescConstPtr

Definition at line 77 of file FileDesc.h.

typedef FileDesc* KKMLL::FileDesc::FileDescPtr

Definition at line 75 of file FileDesc.h.

Constructor & Destructor Documentation

FileDesc::FileDesc ( )

Definition at line 81 of file FileDesc.cpp.

References FileDesc().

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), KKMLL::FeatureEncoder::CreateEncodedFileDesc(), KKMLL::GrayScaleImagesFVProducer::DefineFileDescStatic(), FileDesc(), KKMLL::FeatureFileIODstWeb::GetFileDesc(), KKMLL::FeatureFileIOColumn::GetFileDesc(), KKMLL::FeatureFileIOC45::GetFileDesc(), KKMLL::FeatureFileIOUCI::GetFileDesc(), KKMLL::FeatureFileIOSparse::GetFileDesc(), MergeSymbolicFields(), NewContinuousDataOnly(), and KKMLL::XmlElementFileDesc::XmlElementFileDesc().

                    :
 
   attributes          (true),
   cardinalityVector   (),
   classes             (),
   curAttribute        (NULL),
   sparseMinFeatureNum (0),
   version             (0)
 {
 }

FileDesc::~FileDesc ( )

protected

Definition at line 96 of file FileDesc.cpp.

 {
   kkint32  ZED = 7887;
 }

Member Function Documentation

void FileDesc::AddAAttribute	(	const KKB::KKStr &	_name,
		KKMLL::AttributeType	_type,
		bool &	alreadyExists
	)

Definition at line 169 of file FileDesc.cpp.

References KKMLL::Numeric, and KKMLL::Attribute::Type().

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), KKMLL::FeatureEncoder::CreateEncodedFileDesc(), KKMLL::GrayScaleImagesFVProducer::DefineFileDescStatic(), KKMLL::FeatureFileIOColumn::GetFileDesc(), KKMLL::FeatureFileIOUCI::GetFileDesc(), KKMLL::FeatureFileIOSparse::GetFileDesc(), and NewContinuousDataOnly().

 {
   alreadyExists = false;
   AttributePtr existingAttribute = attributes.LookUpByName (_name);
   if  (existingAttribute)
   {
     // This is a very bad error, it should not be able to happen
     alreadyExists = true;
     return;
   }
 
   curAttribute = new Attribute (_name, _type, attributes.QueueSize ());
   attributes.PushOnBack (curAttribute);
   attributeVector.push_back (curAttribute->Type ());
 
   kkint32  card = 0;
   if  (curAttribute->Type () == AttributeType::Numeric)
      card = INT_MAX;
   cardinalityVector.push_back (card);
 }  /* AddAAttribute */

void FileDesc::AddAAttribute ( const KKMLL::Attribute & attribute )

Definition at line 140 of file FileDesc.cpp.

References KKMLL::Numeric, and KKMLL::Attribute::Type().

Referenced by MergeSymbolicFields().

 {
   attributes.PushOnBack (new Attribute (attribute));
   attributeVector.push_back (attribute.Type ());
 
   kkint32  card = 0;
   if  (attribute.Type () == AttributeType::Numeric)
      card = 999999999;
 
   cardinalityVector.push_back (card);
 }  /* AddAAttribute */

void FileDesc::AddANominalValue	(	kkint32	fieldNum,
		const KKStr &	nominalValue,
		bool &	alreadyExist,
		RunLog &	log
	)

Definition at line 220 of file FileDesc.cpp.

References KKMLL::Nominal, KKMLL::Symbolic, and Type().

Referenced by MergeSymbolicFields().

 {
   ValidateFieldNum (fieldNum, "AddANominalValue");
   AttributeType t = Type (fieldNum);
   if  ((t == AttributeType::Nominal)  ||
        (t == AttributeType::Symbolic)
       )
   {
     attributes[fieldNum].AddANominalValue (nominalValue, alreadyExist);
     if  (!alreadyExist)
       cardinalityVector[fieldNum]++;
   }
 }  /* AddANominalValue */

void FileDesc::AddANominalValue	(	const KKStr &	nominalValue,
		bool &	alreadyExist,
		RunLog &	log
	)

Definition at line 242 of file FileDesc.cpp.

References KKMLL::Attribute::AddANominalValue(), KKB::KKStr::Concat(), and KKB::KKException::KKException().

Referenced by AddANominalValue(), and KKMLL::FeatureFileIODstWeb::GetFileDesc().

 {
   if  (!curAttribute)
   {
     // This should never happen, means that there has not been a nominal feature added yet.
     KKStr  errMsg = "FileDesc::AddANominalValue    ***ERROR***    No Current Attribute Set.";
     log.Level (-1) << endl << errMsg << endl << endl;
     throw KKException (errMsg);
   }
   
   alreadyExist = false;
   curAttribute->AddANominalValue (nominalValue, alreadyExist);
   if  (!alreadyExist)
     cardinalityVector[curAttribute->FieldNum ()]++;
 }  /* AddANominalValue */

void FileDesc::AddANominalValue	(	const KKStr &	attributeName,
		const KKStr &	nominalValue,
		bool &	alreadyExist,
		RunLog &	log
	)

Definition at line 264 of file FileDesc.cpp.

References AddANominalValue(), KKB::KKStr::Concat(), KKB::KKException::KKException(), and KKB::KKStr::KKStr().

 {
   curAttribute = attributes.LookUpByName (attributeName);
   if  (!curAttribute)
   {
     KKStr  errMsg (128);
     errMsg << "FileDesc::AddANominalValue   ***ERROR***,   Invalid Attribute[" << attributeName << "].";
     log.Level (-1) << endl << errMsg << endl << endl;
     throw KKException (errMsg);
   }
 
   AddANominalValue (nominalValue, alreadyExist, log);
 }  /* AddANominalValue */

void FileDesc::AddAttributes ( const KKMLL::AttributeList & attributes )

Definition at line 156 of file FileDesc.cpp.

 {
   AttributeList::const_iterator  idx;
   for  (idx = attributes.begin ();  idx != attributes.end ();  ++idx)
   {
     AddAAttribute (**idx);
   }
 }

void FileDesc::AddClasses ( const MLClassList & classesToAdd )

Definition at line 196 of file FileDesc.cpp.

Referenced by KKMLL::FeatureFileIOC45::GetFileDesc(), and ReadXML().

 {
   MLClassList::const_iterator  idx;
   for  (idx = classesToAdd.begin ();  idx != classesToAdd.end ();  idx++)
   {
     MLClassPtr  ic = *idx;
     if  (classes.PtrToIdx (ic) < 0)
       classes.AddMLClass (ic);
   }
 }  /* AddClasses */

bool FileDesc::AllFieldsAreNumeric ( ) const

Allows the user to quickly determine if there are no nominal fields.

Example use is in CrossValidation application, by using this method it can quickly determine if it is worth while using encoding.

Definition at line 706 of file FileDesc.cpp.

References KKMLL::Ignore, KKMLL::Numeric, NumOfFields(), and Type().

Referenced by KKMLL::FeatureVectorList::AllFieldsAreNumeric().

 {
   for  (kkint32 fieldNum = 0;  fieldNum < (kkint32)NumOfFields ();  fieldNum++)
   {
     AttributeType  t = Type (fieldNum);
     if  ((t != AttributeType::Numeric)  &&  (t != AttributeType::Ignore))
       return false;
   }
 
   
   return  true;
 }  /* AllFieldsAreNumeric */

const KKMLL::AttributeList& KKMLL::FileDesc::Attributes ( ) const

inline

Definition at line 114 of file FileDesc.h.

114 {return attributes;};

const AttributeTypeVector& KKMLL::FileDesc::AttributeVector ( ) const

inline

Definition at line 115 of file FileDesc.h.

Referenced by KKMLL::FeatureNumList::AllFeatures(), and KKMLL::FeatureEncoder2::FeatureEncoder2().

115 {return attributeVector;};

kkint32 FileDesc::Cardinality ( kkint32 fieldNum ) const

Definition at line 341 of file FileDesc.cpp.

References KKMLL::Attribute::Cardinality(), KKB::KKStr::Concat(), KKMLL::Ignore, KKB::KKException::KKException(), KKMLL::Nominal, KKMLL::Numeric, KKB::operator<<(), KKMLL::Symbolic, and KKMLL::Attribute::Type().

Referenced by KKMLL::FeatureVectorList::FeatureCardinality(), MergeSymbolicFields(), KKMLL::ModelParam::NumOfFeaturesAfterEncoding(), KKMLL::SVMparam::NumOfFeaturesAfterEncoding(), and KKMLL::FeatureVectorList::ReSyncSymbolicData().

 {
   ValidateFieldNum (fieldNum, "Type");
 
   AttributePtr a = attributes.IdxToPtr (fieldNum);
 
   if  (!a)
   {
     KKStr  errMsg;
     errMsg << "FileDesc::Cardinality    ***ERROR***    Could not locate attribute[" << fieldNum << "]";
     cerr << errMsg;
     throw  KKException (errMsg);
   }
 
   switch  (a->Type ())
   {
   case  AttributeType::Ignore:   return  a->Cardinality ();
   case  AttributeType::Nominal:  return  a->Cardinality ();
   case  AttributeType::Numeric:  return  INT_MAX;
   case  AttributeType::Symbolic: return  a->Cardinality ();
   
   default: return  INT_MAX;
   }
 
   return  INT_MAX;
 }  /* Cardinality */

const VectorInt32& KKMLL::FileDesc::CardinalityVector ( ) const

inline

Definition at line 116 of file FileDesc.h.

Referenced by KKMLL::FeatureEncoder2::FeatureEncoder2().

116 {return cardinalityVector;}

const MLClassList& KKMLL::FileDesc::Classes ( ) const

inline

Definition at line 117 of file FileDesc.h.

117 {return classes;}

const KKStr& KKMLL::FileDesc::ClassNameAttribute ( ) const

inline

ClassNameAttribute added to support dstWeb data files.

Definition at line 118 of file FileDesc.h.

Referenced by KKMLL::FeatureFileIODstWeb::LoadFile().

void KKMLL::FileDesc::ClassNameAttribute ( const KKStr & _classNameAttribute )

inline

Definition at line 123 of file FileDesc.h.

References KKB::KKStr::operator=().

123 {classNameAttribute = _classNameAttribute;}

const AttributePtr * FileDesc::CreateAAttributeTable ( ) const

Caller will be responsible for deleting

Definition at line 408 of file FileDesc.cpp.

Referenced by KKMLL::FeatureFileIODstWeb::LoadFile(), KKMLL::FeatureFileIOC45::LoadFile(), KKMLL::FeatureFileIORoberts::SaveFile(), and KKMLL::FeatureFileIOC45::SaveFile().

 {
   AttributePtr*  table = new AttributePtr[attributes.QueueSize ()];
 
   for  (kkint32 x = 0;  x < attributes.QueueSize ();  x++)
     table[x] = attributes.IdxToPtr (x);
   
   return  table;
 }  /* CreateAAttributeTable */

AttributeTypeVector FileDesc::CreateAttributeTypeTable ( ) const

Definition at line 419 of file FileDesc.cpp.

Referenced by KKMLL::FeatureVectorList::CreateAttributeTypeTable(), KKMLL::NormalizationParms::NormalizationParms(), KKMLL::NormalizationParms::ReadXML(), KKMLL::SVMModel::ReadXML(), and KKMLL::SVMModel::SVMModel().

 {
   kkint32  x;
   AttributeTypeVector  attributeTypes (attributes.size (), AttributeType::NULLAttribute);
   for  (x = 0;  x < attributes.QueueSize ();  x++)
     attributeTypes[x] = attributes[x].Type ();
   return attributeTypes;
 }  /* CreateAttributeTypeTable () */

VectorInt32 FileDesc::CreateCardinalityTable ( ) const

Definition at line 430 of file FileDesc.cpp.

Referenced by KKMLL::FeatureVectorList::CreateCardinalityTable(), KKMLL::SVMModel::ReadXML(), and KKMLL::SVMModel::SVMModel().

 {
   kkint32  x;
   VectorInt32  cardinalityTable (attributes.QueueSize (), 0);
   for  (x = 0;  x < attributes.QueueSize ();  x++)
     cardinalityTable[x] = attributes[x].Cardinality ();
   return cardinalityTable;
 }  /* CreateCardinalityTable */

void FileDesc::DisplayAttributeMappings ( )

Definition at line 617 of file FileDesc.cpp.

References KKMLL::Attribute::Cardinality(), KKMLL::Attribute::GetNominalCode(), KKMLL::Attribute::GetNominalValue(), KKMLL::Ignore, KKMLL::Nominal, KKMLL::NULLAttribute, KKMLL::Numeric, NumOfFields(), KKMLL::Ordinal, KKMLL::Symbolic, and KKMLL::Attribute::Type().

 {
   kkuint32 i;
   kkint32  j;
   AttributePtr a;
 
   for (i = 0; i < NumOfFields(); i++)
   {
     a = attributes.IdxToPtr (i);
     cout << i << ": ";
 
     if  (a->Type() == AttributeType::Nominal)
     {
       for  (j = 0;  j<a->Cardinality ();  j++)
       {
         if (j != a->GetNominalCode( a->GetNominalValue(j) ))
         {
           cout << " Code does not match ";
         }
         cout << j << ":" << a->GetNominalValue(j) << "  ";
       }
     }
 
     else if  (a->Type() == AttributeType::Symbolic)
     {
       cout << "Symbolic (";
       for  (j = 0;  j<a->Cardinality ();  j++)
       {
         if (j != a->GetNominalCode ( a->GetNominalValue(j) ))
         {
           cout << " Code does not match ";
         }
         cout << j << ":" << a->GetNominalValue(j) << "  ";
       }
 
       cout << ")";
     }
 
     
     else if (a->Type() == AttributeType::Ignore)
     {
       cout << "ignore";
     }
 
     else if (a->Type() == AttributeType::Numeric)
     {
       cout << "numeric";
     }
 
     else if (a->Type() == AttributeType::Ordinal)
     {
       cout << "ordinal";
     }
 
     else if (a->Type() == AttributeType::NULLAttribute)
     {
       cout << "NULL";
     }
     cout << endl;
   }
 }  /* DisplayAttributeMappings */

const KKStr & FileDesc::FieldName ( kkint32 fieldNum ) const

Definition at line 387 of file FileDesc.cpp.

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), KKMLL::FeatureEncoder::CreateEncodedFileDesc(), KKMLL::FeatureVectorProducer::FeatureName(), KKMLL::FeatureVectorList::FieldName(), MergeSymbolicFields(), KKMLL::FeatureVectorList::ReSyncSymbolicData(), and SameExceptForSymbolicData().

 {
   ValidateFieldNum (fieldNum, "FieldName");
   return  attributes[fieldNum].Name ();
 } /* FieldName */

const KKStr& KKMLL::FileDesc::FileName ( ) const

inline

Definition at line 119 of file FileDesc.h.

Referenced by SameExceptForSymbolicData().

119 {return fileName;}

void KKMLL::FileDesc::FileName ( const KKStr & _fileName )

inline

Definition at line 124 of file FileDesc.h.

References KKB::KKStr::operator=().

124 {fileName = _fileName;}

void FileDesc::FinalCleanUp ( )

static

Clean up function, call just before exiting the application.

Before you terminate your application you should call this method. It will clean up the FileDesc objects that were created during the runtime of your application.

Definition at line 57 of file FileDesc.cpp.

References KKB::GoalKeeper::Destroy(), KKB::GoalKeeper::EndBlock(), and KKB::GoalKeeper::StartBlock().

Referenced by KKMLL::MLClass::FinalCleanUp().

 {
   if  (finalCleanUpRanAlready)
     return;
 
   FileDesc::CreateBlocker ();
   blocker->StartBlock ();
   if  (!finalCleanUpRanAlready)
   {
     if  (exisitingDescriptions)
     {
       delete  exisitingDescriptions;
       exisitingDescriptions = NULL;
     }
     finalCleanUpRanAlready = true;
   }
   blocker->EndBlock ();
 
   GoalKeeper::Destroy (blocker);  blocker = NULL;
 }  /* FinalCleanUp */

static bool KKMLL::FileDesc::FinalCleanUpRanAlready ( )

inlinestatic

Definition at line 88 of file FileDesc.h.

Referenced by KKMLL::MLClass::FinalCleanUp().

88 {return finalCleanUpRanAlready;}

const Attribute & FileDesc::GetAAttribute ( kkint32 fieldNum ) const

Definition at line 210 of file FileDesc.cpp.

Referenced by KKMLL::FeatureEncoder::FeatureEncoder(), and MergeSymbolicFields().

 {
   ValidateFieldNum (fieldNum, "GetAAttribute");
   const Attribute&  a = attributes [fieldNum];
   return  a;
 }  /* GetAAttribute */

FileDescPtr FileDesc::GetExistingFileDesc ( FileDescPtr fileDesc )

static

Returns a pointer to an existing instance of 'fileDesc' if it exists, otherwise will use one being passed in.

First looks to see if a the same FileDesc is already in the existing list

in that case will return back the same pointer.

> Second Will look for a existing FileDesc that is the same as the 'fileDesc'
  being passed in.
  if  one is found   them
    fileDesc is deleted
    exiting one will be returned.
  else
    fileDesc will be added to existinList (exisitingDescriptions) and returned.
    and returned.

Parameters

[in] fileDesc Pointer to a FileDesc object that you want to look and see if one that is identical already exists.

Returns: pointer to the 'FileDesc' instance that the caller should be using.

Definition at line 555 of file FileDesc.cpp.

References KKB::KKStr::Concat(), KKB::GoalKeeper::EndBlock(), KKMLL::FileDescList::FileDescList(), KKB::KKException::KKException(), and KKB::GoalKeeper::StartBlock().

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), KKMLL::GrayScaleImagesFVProducer::DefineFileDescStatic(), KKMLL::FeatureVectorProducer::FileDesc(), KKMLL::FeatureFileIOC45::LoadFeatureFile(), KKMLL::FeatureFileIO::LoadFeatureFile(), MergeSymbolicFields(), NewContinuousDataOnly(), and KKMLL::XmlElementFileDesc::XmlElementFileDesc().

 {
   if (fileDesc == NULL)
   {
     KKStr errMsg = "FileDesc::GetExistingFileDesc   ***ERROR***   (fileDesc == NULL).";
     cerr << endl << errMsg << endl << endl;
     throw KKException (errMsg);
   }
   FileDescPtr  result = NULL;
 
   CreateBlocker ();
 
   blocker->StartBlock ();
 
   if  (!exisitingDescriptions)
   {
     exisitingDescriptions = new FileDescList (true);
     exisitingDescriptions->PushOnBack (fileDesc);
     result = fileDesc;
     finalCleanUpRanAlready = false;
     atexit (FileDesc::FinalCleanUp);
   }
 
   else
   {
     for  (auto existingFileDesc: *exisitingDescriptions)  
     {
       if  (existingFileDesc == fileDesc)
       {
         result = existingFileDesc;
         break;
       }
       else if  (existingFileDesc == NULL)
       {
         continue;
       }
       else if  ((*existingFileDesc) == (*fileDesc))
       {
         // Looks like we already have a compatible "FileDesc" instance.
         // In this case this is the one the user will want.
         delete  fileDesc;
         result = existingFileDesc;
         break;
       }
     }
 
     if  (!result)
     {
       exisitingDescriptions->PushOnBack (fileDesc);
       result = fileDesc;
     }
   }
 
   blocker->EndBlock ();
 
   return  result;
 } /* GetExistingFileDesc */

kkint32 FileDesc::GetFieldNumFromAttributeName ( const KKStr & attributeName ) const

Definition at line 690 of file FileDesc.cpp.

References KKMLL::Attribute::FieldNum().

Referenced by KKMLL::FeatureFileIODstWeb::LoadFile().

 {
   AttributePtr  a = attributes.LookUpByName (attributeName);
   if  (!a)
     return -1;
   else
     return a->FieldNum ();
 }  /* GetFieldNumFromAttributeName */

MLClassPtr FileDesc::GetMLClassPtr ( const KKStr & className )

Definition at line 298 of file FileDesc.cpp.

 {
   return  classes.GetMLClassPtr (className);
 }

const KKStr & FileDesc::GetNominalValue	(	kkint32	fieldNum,
		kkint32	code
	)		const

Definition at line 395 of file FileDesc.cpp.

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), KKMLL::FeatureEncoder::CreateEncodedFileDesc(), MergeSymbolicFields(), and KKMLL::FeatureVectorList::ReSyncSymbolicData().

 {
   ValidateFieldNum (fieldNum, "GetNominalValue");
   return  attributes[fieldNum].GetNominalValue (code);
 }  /* GetNominalValue */

const AttributePtr FileDesc::LookUpByName ( const KKStr & attributeName ) const

Definition at line 682 of file FileDesc.cpp.

 {
   return  attributes.LookUpByName (attributeName);
 }  /* LookUpByName */

MLClassPtr FileDesc::LookUpMLClassByName ( const KKStr & className )

Definition at line 284 of file FileDesc.cpp.

Referenced by KKMLL::FeatureFileIOC45::LoadFile().

 {
   return  classes.LookUpByName (className);
 }

kkint32 FileDesc::LookUpNominalCode	(	kkint32	fieldNum,
		const KKStr &	nominalValue
	)		const

Definition at line 321 of file FileDesc.cpp.

References KKMLL::Attribute::GetNominalCode(), KKMLL::Nominal, KKMLL::Symbolic, and KKMLL::Attribute::Type().

Referenced by MergeSymbolicFields(), and KKMLL::FeatureVectorList::ReSyncSymbolicData().

 {
   ValidateFieldNum (fieldNum, "LookUpNominalCode");
   const Attribute& a = attributes[fieldNum];
   if  ((a.Type () != AttributeType::Nominal)  &&
        (a.Type () != AttributeType::Symbolic)
       )
   {
     return -1;
   }
 
   return  a.GetNominalCode (nominalValue);
 }  /* LookUpNominalCode */

MLClassPtr FileDesc::LookUpUnKnownMLClass ( )

Definition at line 291 of file FileDesc.cpp.

Referenced by KKMLL::FeatureFileIOC45::LoadFile().

 {
   return  classes.GetUnKnownClass ();
 }

kkint32 FileDesc::MemoryConsumedEstimated ( ) const

Definition at line 102 of file FileDesc.cpp.

 {
   kkint32  memoryConsumedEstimated = sizeof (FileDesc)         +
          attributes.MemoryConsumedEstimated ()               +
          sizeof (AttributeType) * attributeVector.size ()    +
          sizeof (kkint32)         * cardinalityVector.size ()  +
          classes.MemoryConsumedEstimated ()                  +
          classNameAttribute.MemoryConsumedEstimated ();
 
   return  memoryConsumedEstimated;
 }

FileDescPtr FileDesc::MergeSymbolicFields	(	const FileDesc &	left,
		const FileDesc &	right,
		RunLog &	log
	)

static

Merges the Symbolic fields of two different 'FileDesc' instances producing a new instance of 'FileDesc'.

This method will only work if both instances have the same number of fields, their names must be the same(NOT case sensitive), and each field in both instances must be the same type. If all these conditions are not 'true' will return NULL. The fields that are of 'Symbolic' will have their values merged together.

See also: KKMLL:Attribute

Definition at line 723 of file FileDesc.cpp.

References AddAAttribute(), AddANominalValue(), Cardinality(), KKB::KKStr::Concat(), FieldName(), FileDesc(), GetAAttribute(), GetExistingFileDesc(), GetNominalValue(), LookUpNominalCode(), NumOfFields(), KKB::KKStr::operator!=(), SameExceptForSymbolicData(), KKMLL::Symbolic, and Type().

Referenced by KKMLL::FeatureVectorList::SynchronizeSymbolicData().

 {
   if  (!left.SameExceptForSymbolicData (right, log))
   {
     log.Level (-1) << endl
                    << "FileDesc::MergeSymbolicFields   There are more differences between file descriptions besides symbolic data." << endl
                    << "          File[" << left.fileName << "] File[" << right.fileName << "]" << endl
                    << endl;
 
     return  NULL;
   }
 
   FileDescPtr  f = new FileDesc ();
 
   kkint32  numOfFields = left.NumOfFields ();
 
   kkint32  fieldNum = 0;
 
   for  (fieldNum = 0;  fieldNum < numOfFields;  fieldNum++)
   {
     const KKStr&  lName = left.FieldName (fieldNum);
     const KKStr&  rName = right.FieldName (fieldNum);
 
     if  (lName != rName)
     {
       log.Level (-1) << endl
                      << "FileDesc::MergeSymbolicFields   Field Name mis-match" << endl
                      << "          File[" << left.fileName << "] Name[" << lName << "]  File[" << right.fileName << "] Name[" << rName << "]" << endl
                      << endl;
 
       return NULL;
     }
 
     AttributeType  lType = left.Type (fieldNum);
     AttributeType  rType = right.Type (fieldNum);
 
     if  (lType != rType)
     {
       log.Level (-1) << endl
                      << "FileDesc::MergeSymbolicFields   Field Type  mis-match" << endl
                      << "          File[" << left.fileName << "] Name[" << lName << "]  File[" << right.fileName << "] Name[" << rName << "]" << endl
                      << endl;
 
       return  NULL;
     }
 
     f->AddAAttribute (left.GetAAttribute (fieldNum));
     if  (lType != AttributeType::Symbolic)
     {
       continue;
     }
 
     // We can merge in Nominal Values for this field.
 
     kkint32  z;
     for  (z = 0;  z < right.Cardinality (fieldNum);  z++)
     {
       const KKStr&  rightNomName = right.GetNominalValue (fieldNum, z);
       kkint32  lCode = f->LookUpNominalCode (fieldNum, rightNomName);
       if  (lCode < 0)
       {
         bool  alreadyExists = false;
         f->AddANominalValue (fieldNum, rightNomName, alreadyExists, log);
         lCode = f->LookUpNominalCode (fieldNum, rightNomName);
       }
     }
   }
 
   return  GetExistingFileDesc (f);
 }  /* MergeSymbolicFields */

FileDescPtr FileDesc::NewContinuousDataOnly ( VectorKKStr & _fieldNames )

static

Creates a simple FileDesc that consists of continuous data only.

Creates a file description that will consist of continuous fields only. The vector '_fieldNames' will provide the list of field names.

Parameters

[in]	_log	Logging file to use.
[in]	_fieldNames	Name of fields; one entry for each field.

Definition at line 116 of file FileDesc.cpp.

References AddAAttribute(), KKB::KKStr::Concat(), FileDesc(), GetExistingFileDesc(), KKMLL::Numeric, and KKB::StrFormatInt().

Referenced by KKMLL::FeatureEncoder::FeatureEncoder(), and KKMLL::FeatureEncoder2::FeatureEncoder2().

 {
   bool  alreadyExists = false;
   FileDescPtr  newFileDesc = new FileDesc ();
   for  (kkint32 fieldNum = 0;  fieldNum < (kkint32)_fieldNames.size ();  fieldNum++)
   {
     kkint32  seqNum = 0;
     do
     {
       KKStr  fieldName = _fieldNames[fieldNum];
       if  (seqNum > 0)
         fieldName << "_" << StrFormatInt (seqNum, "000");
 
       newFileDesc->AddAAttribute (fieldName, AttributeType::Numeric, alreadyExists);
       seqNum++;
     }
       while  (alreadyExists);
   }
 
   return  GetExistingFileDesc (newFileDesc);
 }  /* NewContinuousDataOnly */

kkuint32 KKMLL::FileDesc::NumOfFields ( ) const

inline

Definition at line 197 of file FileDesc.h.

Referenced by KKMLL::FeatureNumList::AllFeatures(), KKMLL::FeatureNumList::AllFeaturesSelected(), AllFieldsAreNumeric(), DisplayAttributeMappings(), KKMLL::FeatureVectorProducer::FeatureCount(), KKMLL::FeatureVectorProducer::FeatureName(), KKMLL::FeatureNumList::FeatureNumList(), KKMLL::FeatureVectorList::FeatureVectorList(), KKMLL::FeatureFileIODstWeb::LoadFile(), KKMLL::FeatureFileIOColumn::LoadFile(), KKMLL::FeatureFileIOUCI::LoadFile(), KKMLL::FeatureFileIOC45::LoadFile(), KKMLL::FeatureFileIOSparse::LoadFile(), MergeSymbolicFields(), KKMLL::FeatureEncoder2::NumEncodedFeatures(), KKMLL::FeatureVectorList::ResetFileDesc(), KKMLL::FeatureVectorList::ReSyncSymbolicData(), SameExceptForSymbolicData(), and KKMLL::FeatureNumList::ToHexString().

197 {return (kkuint32)attributes.size ();}

KKB::kkuint32

unsigned __int32 kkuint32

Definition: KKBaseTypes.h:89

bool FileDesc::operator!= ( const FileDesc & rightSize ) const

Returns true if file description on the right size is NOT identical.

If both FileDesc instances have different number of fields, or any one of the fields has a different name(NOT case sensitive), or one of the fields is of a different type.

Definition at line 456 of file FileDesc.cpp.

 {
   if  ((NumOfFields () != rightSide.NumOfFields ())  ||
        (Version ()     != rightSide.Version ())      ||
        (attributes     != rightSide.attributes)
       )
     return true;
 
   return  false;  
 } /* operator== */

bool FileDesc::operator== ( const FileDesc & rightSize ) const

Returns true if file description on the right size is identical.

Both FileDesc instances must have the same number of fields, the fields must have the same names(NOT case sensitive), and the fields must have matching types(ex numerical, ordinal, etc).

Definition at line 443 of file FileDesc.cpp.

Referenced by KKMLL::FeatureFileIO::FeatureDataReSink().

 {
   if  ((NumOfFields () != rightSide.NumOfFields ())  ||
        (Version ()     != rightSide.Version ())      ||
        (attributes     != rightSide.attributes)
       )
     return false;
 
   return  true;  
 } /* operator== */

void FileDesc::ReadXML	(	XmlStream &	s,
		XmlTagConstPtr	tag,
		VolConstBool &	cancelFlag,
		RunLog &	log
	)

Definition at line 800 of file FileDesc.cpp.

References AddClasses(), KKB::KKStr::Concat(), KKB::KKStr::EqualIgnoreCase(), KKB::XmlStream::GetNextToken(), KKB::XmlElement::SectionName(), SparseMinFeatureNum(), KKB::XmlToken::tokElement, KKB::XmlToken::TokenType(), KKMLL::XmlElementMLClassNameList::Value(), KKB::XmlElement::VarName(), and Version().

Referenced by KKMLL::XmlElementFileDesc::XmlElementFileDesc().

 {
   XmlTokenPtr  t = s.GetNextToken (cancelFlag, log);
   while  (t  &&  (!cancelFlag))
   {
     if  (t->TokenType () == XmlToken::TokenTypes::tokElement)
     {
       XmlElementPtr  e = dynamic_cast<XmlElementPtr> (t);
       const KKStr&  className = e->SectionName ();
       const KKStr&  varName = e->VarName ();
       if  (varName.EqualIgnoreCase ("FileName"))
       {
         XmlElementKKStrPtr  eKKStr = dynamic_cast<XmlElementKKStrPtr>(e);
         if  (eKKStr)
           FileName (*(eKKStr->Value ()));
       }
 
       else if  (varName.EqualIgnoreCase ("Attributes"))
       {
         XmlElementAttributeListPtr  eAttributes = dynamic_cast<XmlElementAttributeListPtr>(e);
         if  (eAttributes  &&  (eAttributes->Value ()))
           AddAttributes (*(eAttributes->Value ()));
       }
 
       else if  (varName.EqualIgnoreCase ("Classes"))
       {
         XmlElementMLClassNameListPtr  eCLasses = dynamic_cast<XmlElementMLClassNameListPtr>(e);
         if  (eCLasses  &&  (eCLasses->Value ()))
           AddClasses (*(eCLasses->Value ()));
       }
 
       else if  (varName.EqualIgnoreCase ("ClassNameAttribute"))
       {
         XmlElementKKStrPtr  eKKStr = dynamic_cast<XmlElementKKStrPtr>(e);
         if  (eKKStr)
           ClassNameAttribute (*(eKKStr->Value ()));
       }
 
       else if  (varName.EqualIgnoreCase ("Version"))
       {
         XmlElementInt32Ptr  eKKInt32 = dynamic_cast<XmlElementInt32Ptr>(e);
         if  (eKKInt32)
           Version ((kkint16)eKKInt32->Value ());
       }
 
       else if  (varName.EqualIgnoreCase ("SparseMinFeatureNum"))
       {
         XmlElementInt32Ptr  eKKInt32 = dynamic_cast<XmlElementInt32Ptr>(e);
         if  (eKKInt32)
           SparseMinFeatureNum (eKKInt32->Value ());
       }
       else
       {
         log.Level (-1) << endl
           << "XmlElementFileDesc   ***ERROR***   Unexpected Element <" << className << ", VarName=" << varName.QuotedStr () << ">" << endl
           << endl;
       }
     }
 
     delete t;
     t = s.GetNextToken (cancelFlag, log);
   }
 
   delete  t;
   t = NULL;
 }  /* ReadXML */

bool FileDesc::SameExceptForSymbolicData	(	const FileDesc &	otherFd,
		RunLog &	log
	)		const

Definition at line 472 of file FileDesc.cpp.

References KKB::KKStr::Concat(), FieldName(), FileName(), NumOfFields(), KKB::KKStr::operator!=(), and Type().

Referenced by MergeSymbolicFields(), KKMLL::FeatureVectorList::SameExceptForSymbolicData(), and KKMLL::FeatureVectorList::SynchronizeSymbolicData().

 {
   bool  same = true;
 
   if  (NumOfFields () != otherFd.NumOfFields ())
   {
     log.Level (-1) << endl
                    << "FileDesc::SameExceptForSymbolicData    Field count mis-match" << endl
                    << "          File[" << fileName << "] count[" << otherFd.NumOfFields () << "]  File[" << otherFd.fileName << "] Count[" << otherFd.NumOfFields () << "]" << endl
                    << endl;
 
     return  false;
   }
 
   kkint32  numOfFields = NumOfFields ();
   kkint32  fieldNum = 0;
 
   const KKStr&  rightFileName = otherFd.FileName ();
 
   for  (fieldNum = 0;  fieldNum < numOfFields;  fieldNum++)
   {
     const KKStr&  lName = FieldName (fieldNum);
     const KKStr&  rName = otherFd.FieldName (fieldNum);
 
     if  (lName != rName)
     {
       log.Level (-1) << endl
                      << "FileDesc::SameExceptForSymbolicData   Field Name mis-match" << endl
                      << "          File[" << fileName << "] Name[" << lName << "]  File[" << rightFileName << "] Name[" << rName << "]" << endl
                      << endl;
 
       same = false;
     }
 
     else
     {
       AttributeType  lType = Type (fieldNum);
       AttributeType  rType = otherFd.Type (fieldNum);
 
       if  (lType != rType)
       {
         log.Level (-1) << endl
                        << "FileDesc::SameExceptForSymbolicData   Field Type  mis-match" << endl
                        << "          File[" << fileName << "] Name[" << lName << "]  File[" << rightFileName << "] Name[" << rName << "]" << endl
                        << endl;
 
         same = false;
       }
     }
   }
 
   return  same;
 }  /* SameExceptForSymbolicData */

kkint32 KKMLL::FileDesc::SparseMinFeatureNum ( ) const

inline

Definition at line 120 of file FileDesc.h.

Referenced by KKMLL::FeatureFileIOSparse::LoadFile(), and KKMLL::FeatureFileIOSparse::SaveFile().

120 {return sparseMinFeatureNum;}

void KKMLL::FileDesc::SparseMinFeatureNum ( kkint32 _sparseMinFeatureNum )

inline

Definition at line 125 of file FileDesc.h.

Referenced by KKMLL::FeatureFileIOSparse::GetFileDesc(), and ReadXML().

125 {sparseMinFeatureNum = _sparseMinFeatureNum;}

AttributeType FileDesc::Type ( kkint32 fieldNum ) const

Definition at line 370 of file FileDesc.cpp.

Referenced by AddANominalValue(), AllFieldsAreNumeric(), KKMLL::FeatureVectorList::FeatureType(), KKMLL::FeatureVectorList::FeatureTypeStr(), MergeSymbolicFields(), KKMLL::ModelParam::NumOfFeaturesAfterEncoding(), KKMLL::SVMparam::NumOfFeaturesAfterEncoding(), KKMLL::FeatureVectorList::ReSyncSymbolicData(), SameExceptForSymbolicData(), and KKMLL::FeatureNumList::SetAllFeatures().

 {
   ValidateFieldNum (fieldNum, "Type");
   return  attributes[fieldNum].Type ();
 }  /* Type */

KKStr FileDesc::TypeStr ( kkint32 fieldNum ) const

Definition at line 378 of file FileDesc.cpp.

Referenced by KKMLL::FeatureEncoder2::CreateEncodedFileDesc(), and KKMLL::FeatureEncoder::CreateEncodedFileDesc().

 {
   ValidateFieldNum (fieldNum, "TypeStr");
   return  AttributeTypeToStr (attributes[fieldNum].Type ());
 }  /* TypeStr */

kkint16 KKMLL::FileDesc::Version ( ) const

inline

Definition at line 121 of file FileDesc.h.

121 {return version;}

void KKMLL::FileDesc::Version ( kkint16 _version )

inline

Definition at line 126 of file FileDesc.h.

Referenced by KKMLL::GrayScaleImagesFVProducer::DefineFileDescStatic(), and ReadXML().

126 {version = _version;}

void FileDesc::WriteXML	(	const KKStr &	varName,
		std::ostream &	o
	)		const

Definition at line 875 of file FileDesc.cpp.

References KKB::XmlTag::AddAtribute(), KKB::KKStr::Empty(), KKB::XmlTag::tagEnd, KKB::XmlTag::tagStart, KKB::XmlTag::WriteXML(), and KKB::XmlTag::XmlTag().

Referenced by KKMLL::Model::WriteModelXMLFields(), KKMLL::FeatureEncoder::WriteXML(), SVM289_MFS::Svm_Model::WriteXML(), KKMLL::SVMModel::WriteXML(), and KKMLL::XmlElementFileDesc::WriteXML().

 {
   XmlTag  startTag ("FileDesc", XmlTag::TagTypes::tagStart);
   if  (!varName.Empty ())
     startTag.AddAtribute ("VarName", varName);
   startTag.WriteXML (o);
   o << endl;
 
   if  (!fileName.Empty ())
     XmlElementKKStr::WriteXML (fileName, "FileName", o);
 
   XmlElementAttributeList::WriteXML (attributes, "Attributes", o);
 
   XmlElementMLClassNameList::WriteXML (classes, "Classes", o);
 
   if  (!ClassNameAttribute ().Empty ())
     XmlElementKKStr::WriteXML (ClassNameAttribute (), "ClassNameAttribute", o);
 
   XmlElementInt32::WriteXML (version, "Version", o);
 
   if  (sparseMinFeatureNum  != 0)
     XmlElementInt32::WriteXML (sparseMinFeatureNum, "SparseMinFeatureNum", o);
 
   XmlTag  endTag ("FileDesc", XmlTag::TagTypes::tagEnd);
   endTag.WriteXML (o);
   o << endl;
 }

Friends And Related Function Documentation

friend class KKQueue< FileDesc >

friend

Definition at line 107 of file FileDesc.h.

The documentation for this class was generated from the following files:

C:/Users/Kurt/GitHub/KSquareLibraries/KKMachineLearning/FileDesc.h
C:/Users/Kurt/GitHub/KSquareLibraries/KKMachineLearning/FileDesc.cpp

Public Types

Public Member Functions

Static Public Member Functions

Protected Member Functions

Friends

Detailed Description

Member Typedef Documentation

Constructor & Destructor Documentation

Member Function Documentation

Friends And Related Function Documentation