/// <summary> Appends all elements of the supplied vector to this vector.
///
/// </summary>
/// <param name="toAppend">the FastVector containing elements to append.
/// </param>
public void appendXmlElements(FastVector toAppend)
{
Capacity = size() + toAppend.size();
Array.Copy(toAppend.m_Objects, 0, m_Objects, size(), toAppend.size());
m_Size = m_Objects.Length;
}
/// <summary> Split up a string containing options into an array of strings,
/// one for each option.
///
/// </summary>
/// <param name="optionString">the string containing the options
/// </param>
/// <returns> the array of options
/// </returns>
public static System.String[] splitOptions(System.String quotedOptionString)
{
FastVector optionsVec = new FastVector();
System.String str = new System.Text.StringBuilder(quotedOptionString).ToString();
int i;
while (true)
{
//trimLeft
i = 0;
while ((i < str.Length) && (System.Char.IsWhiteSpace(str[i])))
i++;
str = str.Substring(i);
//stop when str is empty
if (str.Length == 0)
break;
//if str start with a double quote
if (str[0] == '"')
{
//find the first not anti-slached double quote
i = 1;
while (i < str.Length)
{
if (str[i] == str[0])
break;
if (str[i] == '\\')
{
i += 1;
if (i >= str.Length)
throw new System.Exception("String should not finish with \\");
if (str[i] != '\\' && str[i] != '"')
throw new System.Exception("Unknow character \\" + str[i]);
}
i += 1;
}
if (i >= str.Length)
throw new System.Exception("Quote parse error.");
//add the founded string to the option vector (without quotes)
System.String optStr = str.Substring(1, (i) - (1));
optStr = unbackQuoteChars(optStr);
optionsVec.addElement(optStr);
str = str.Substring(i + 1);
}
else
{
//find first whiteSpace
i = 0;
while ((i < str.Length) && (!System.Char.IsWhiteSpace(str[i])))
i++;
//add the founded string to the option vector
System.String optStr = str.Substring(0, (i) - (0));
optionsVec.addElement(optStr);
str = str.Substring(i);
}
}
//convert optionsVec to an array of String
System.String[] options = new System.String[optionsVec.size()];
for (i = 0; i < optionsVec.size(); i++)
{
options[i] = ((System.String) optionsVec.elementAt(i));
}
return options;
}
/// <summary> Reads and stores header of an ARFF file.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
/// <exception cref="IOException">if the information is not read
/// successfully
/// </exception>
protected internal virtual void readHeader(StreamTokenizer tokenizer)
{
System.String attributeName;
FastVector attributeValues;
//int i;
Token token=null;
// Get name of relation.
getFirstToken(tokenizer, out token);
//if (tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_EOF)
if ((token != null) && (token is EofToken))
{
errms(tokenizer, "premature end of file");
}
if (ARFF_RELATION.ToUpper().Equals(token.StringValue.ToUpper()))
{
getNextToken(tokenizer,out token);
m_RelationName = token.StringValue;
getLastToken(tokenizer,out token, false);
}
else
{
errms(tokenizer, "keyword " + ARFF_RELATION + " expected");
}
// Create vectors to hold information temporarily.
m_Attributes = new FastVector();
// Get attribute declarations.
getFirstToken(tokenizer, out token);
//if (tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_EOF)
if ((token != null) && (token is EofToken))
{
errms(tokenizer, "premature end of file");
}
while (Attribute.ARFF_ATTRIBUTE.ToUpper().Equals(token.StringValue.ToUpper()))
{
// Get attribute name.
getNextToken(tokenizer,out token);
attributeName = token.StringValue;
getNextToken(tokenizer,out token);
// Check if attribute is nominal.
//if (tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_WORD)
if ((token != null) && (token is WordToken))
{
// Attribute is real, integer, or string.
if (token.StringValue.ToUpper().Equals(Attribute.ARFF_ATTRIBUTE_REAL.ToUpper()) || token.StringValue.ToUpper().Equals(Attribute.ARFF_ATTRIBUTE_INTEGER.ToUpper()) || token.StringValue.ToUpper().Equals(Attribute.ARFF_ATTRIBUTE_NUMERIC.ToUpper()))
{
m_Attributes.addElement(new Attribute(attributeName, numAttributes()));
readTillEOL(tokenizer);
}
else if (token.StringValue.ToUpper().Equals(Attribute.ARFF_ATTRIBUTE_STRING.ToUpper()))
{
m_Attributes.addElement(new Attribute(attributeName, (FastVector) null, numAttributes()));
readTillEOL(tokenizer);
}
else if (token.StringValue.ToUpper().Equals(Attribute.ARFF_ATTRIBUTE_DATE.ToUpper()))
{
System.String format = null;
tokenizer.NextToken(out token);
//if (tokenizer.NextToken() != SupportClass.StreamTokenizerSupport.TT_EOL)
if ((token != null) && (!(token is EofToken)))
{
//if ((tokenizer.ttype != SupportClass.StreamTokenizerSupport.TT_WORD) && (tokenizer.ttype != '\'') && (tokenizer.ttype != '\"'))
if ((token != null) && (!(token is WordToken)) && (token.StringValue!="'") && (token.StringValue!="\"") )
{
errms(tokenizer, "not a valid date format");
}
format = token.StringValue;
readTillEOL(tokenizer);
}
else
{
tokenizer.PushBack(token);
}
m_Attributes.addElement(new Attribute(attributeName, format, numAttributes()));
}
else
{
errms(tokenizer, "no valid attribute type or invalid " + "enumeration");
}
}
else
{
// Attribute is nominal.
attributeValues = new FastVector();
tokenizer.PushBack(token);
// Get values for nominal attribute.
tokenizer.NextToken(out token);
if ( token.StringValue != "{")
{
errms(tokenizer, "{ expected at beginning of enumeration");
}
tokenizer.NextToken(out token);
while ( token.StringValue != "}")
{
//if (tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_EOL)
if (token is EolToken)
{
errms(tokenizer, "} expected at end of enumeration");
}
else
{
attributeValues.addElement(token.StringValue);
}
tokenizer.NextToken(out token);
}
if (attributeValues.size() == 0)
{
errms(tokenizer, "no nominal values found");
}
m_Attributes.addElement(new Attribute(attributeName, attributeValues, numAttributes()));
}
getLastToken(tokenizer,out token, false);
getFirstToken(tokenizer,out token);
//if (tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_EOF)
if (token is EofToken)
errms(tokenizer, "premature end of file");
}
// Check if data part follows. We can't easily check for EOL.
if (!ARFF_DATA.ToUpper().Equals(token.StringValue.ToUpper()))
{
errms(tokenizer, "keyword " + ARFF_DATA + " expected");
}
// Check if any attributes have been declared.
if (m_Attributes.size() == 0)
{
errms(tokenizer, "no attributes declared");
}
// Allocate buffers in case sparse instances have to be read
m_ValueBuffer = new double[numAttributes()];
m_IndicesBuffer = new int[numAttributes()];
}
/// <summary> Help function needed for stratification of set.
///
/// </summary>
/// <param name="numFolds">the number of folds for the stratification
/// </param>
protected internal virtual void stratStep(int numFolds)
{
FastVector newVec = new FastVector(m_Instances.capacity());
int start = 0, j;
// create stratified batch
while (newVec.size() < numInstances())
{
j = start;
while (j < numInstances())
{
newVec.addElement(instance(j));
j = j + numFolds;
}
start++;
}
m_Instances = newVec;
}
/// <summary> Constructor for nominal attributes and string attributes, where
/// metadata is supplied. If a null vector of attribute values is passed
/// to the method, the attribute is assumed to be a string.
///
/// </summary>
/// <param name="attributeName">the name for the attribute
/// </param>
/// <param name="attributeValues">a vector of strings denoting the
/// attribute values. Null if the attribute is a string attribute.
/// </param>
/// <param name="metadata">the attribute's properties
/// </param>
//@ requires attributeName != null;
//@ requires metadata != null;
/*@ ensures m_Name == attributeName;
ensures m_Index == -1;
ensures attributeValues == null && m_Type == STRING
|| attributeValues != null && m_Type == NOMINAL
&& m_Values.size() == attributeValues.size();
signals (IllegalArgumentException ex)
(* if duplicate strings in attributeValues *);
*/
public Attribute(System.String attributeName, FastVector attributeValues, ProtectedProperties metadata)
{
m_Name = attributeName;
m_Index = - 1;
if (attributeValues == null)
{
m_Values = new FastVector();
m_Hashtable = System.Collections.Hashtable.Synchronized(new System.Collections.Hashtable());
m_Type = STRING;
}
else
{
m_Values = new FastVector(attributeValues.size());
//m_Hashtable = System.Collections.Hashtable.Synchronized(new System.Collections.Hashtable(attributeValues.size()));
m_Hashtable = new System.Collections.Hashtable(attributeValues.size());
for (int i = 0; i < attributeValues.size(); i++)
{
// No serialization
System.Object store = attributeValues.elementAt(i);
//if (((System.String) store).Length > STRING_COMPRESS_THRESHOLD)
//{
// try
// {
// store = new SerializedObject(attributeValues.elementAt(i), true);
// }
// catch (System.Exception ex)
// {
// System.Console.Error.WriteLine("Couldn't compress nominal attribute value -" + " storing uncompressed.");
// }
//}
if (m_Hashtable.ContainsKey(store))
{
//UPGRADE_TODO: The equivalent in .NET for method 'java.lang.Object.toString' may return a different value. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1043'"
throw new System.ArgumentException("A nominal attribute (" + attributeName + ") cannot" + " have duplicate labels (" + store + ").");
}
m_Values.addElement(store);
m_Hashtable.Add(store, (System.Int32) i);
}
m_Type = NOMINAL;
}
setMetadata(metadata);
}
