/// <summary> Loads a cost matrix in the old format from a reader. Adapted from code once sitting
/// in Instances.java
///
/// </summary>
/// <param name="reader">the reader to get the values from.
/// </param>
/// <exception cref="Exception">if the matrix cannot be read correctly.
/// </exception>
//UPGRADE_ISSUE: Class hierarchy differences between 'java.io.Reader' and 'System.IO.StreamReader' may cause compilation errors. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1186'"
public virtual void readOldFormat(System.IO.StreamReader reader)
{
StreamTokenizer tokenizer;
Token currentToken;
double firstIndex, secondIndex, weight;
tokenizer = new StreamTokenizer(reader);
initialize();
tokenizer.Settings.CommentChar('%');
tokenizer.Settings.GrabEol=true;
tokenizer.NextToken(out currentToken);
//while (SupportClass.StreamTokenizerSupport.TT_EOF != (currentToken = tokenizer.NextToken()))
while (!(currentToken is EofToken))
{
// Skip empty lines
//if (currentToken == SupportClass.StreamTokenizerSupport.TT_EOL)
if (currentToken is EolToken)
{
continue;
}
// Get index of first class.
//if (currentToken != SupportClass.StreamTokenizerSupport.TT_NUMBER)
if (!((currentToken is FloatToken)|| (currentToken is IntToken)))
{
throw new System.Exception("Only numbers and comments allowed " + "in cost file!");
}
//firstIndex = tokenizer.nval;
firstIndex = Convert.ToDouble(currentToken.StringValue);
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
if (!Utils.eq((double) firstIndex, firstIndex))
{
throw new System.Exception("First number in line has to be " + "index of a class!");
}
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
if ((int) firstIndex >= size())
{
throw new System.Exception("Class index out of range!");
}
// Get index of second class.
//if (SupportClass.StreamTokenizerSupport.TT_EOF == (currentToken = tokenizer.NextToken()))
tokenizer.NextToken(out currentToken);
if (currentToken is EofToken)
{
throw new System.Exception("Premature end of file!");
}
//if (currentToken == SupportClass.StreamTokenizerSupport.TT_EOL)
if (currentToken is EolToken)
{
throw new System.Exception("Premature end of line!");
}
//if (currentToken != SupportClass.StreamTokenizerSupport.TT_NUMBER)
if (!((currentToken is IntToken) || (currentToken is FloatToken)))
{
throw new System.Exception("Only numbers and comments allowed " + "in cost file!");
}
//secondIndex = tokenizer.nval;
secondIndex = Convert.ToDouble(currentToken.StringValue);
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
if (!Utils.eq((double) secondIndex, secondIndex))
{
throw new System.Exception("Second number in line has to be " + "index of a class!");
}
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
if ((int) secondIndex >= size())
{
throw new System.Exception("Class index out of range!");
}
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
if ((int) secondIndex == (int) firstIndex)
{
throw new System.Exception("Diagonal of cost matrix non-zero!");
}
// Get cost factor.
tokenizer.NextToken(out currentToken);
if (currentToken is EofToken)
//if (SupportClass.StreamTokenizerSupport.TT_EOF == (currentToken = tokenizer.NextToken()))
{
throw new System.Exception("Premature end of file!");
}
//if (currentToken == SupportClass.StreamTokenizerSupport.TT_EOL)
if (currentToken is EolToken)
{
throw new System.Exception("Premature end of line!");
}
//if (currentToken != SupportClass.StreamTokenizerSupport.TT_NUMBER)
if (!((currentToken is IntToken) || (currentToken is FloatToken)))
{
throw new System.Exception("Only numbers and comments allowed " + "in cost file!");
}
weight = Convert.ToDouble(currentToken.StringValue);
if (!Utils.gr(weight, 0))
{
throw new System.Exception("Only positive weights allowed!");
}
//UPGRADE_WARNING: Data types in Visual C# might be different. Verify the accuracy of narrowing conversions. "ms-help://MS.VSCC.v80/dv_commoner/local/redirect.htm?index='!DefaultContextWindowIndex'&keyword='jlca1042'"
setXmlElement((int) firstIndex, (int) secondIndex, weight);
tokenizer.NextToken(out currentToken);
}
}
/// <summary> Gets next token, checking for a premature and of line.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
/// <exception cref="IOException">if it finds a premature end of line
/// </exception>
protected internal virtual void getNextToken(StreamTokenizer tokenizer,out Token token)
{
tokenizer.NextToken(out token);
if (token is EolToken)
{
errms(tokenizer, "premature end of line");
}
if (token is EofToken)
{
errms(tokenizer, "premature end of file");
}
//else if ((tokenizer.ttype == '\'') || (tokenizer.ttype == '"'))
else if ((token is CharToken) && ((token.StringValue == "'") || (token.StringValue == "\"") ))
{
//tokenizer.ttype = SupportClass.StreamTokenizerSupport.TT_WORD;
}
//else if ((tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_WORD) && (tokenizer.sval.Equals("?")))
else if ((token is CharToken) && (token.StringValue == "?"))
{
//tokenizer.ttype = '?';
}
}
/// <summary> Reads and skips all tokens before next end of line token.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
protected internal virtual void readTillEOL(StreamTokenizer tokenizer)
{
Token token;
tokenizer.NextToken(out token);
while (!(token is EolToken))
{
tokenizer.NextToken(out token);
} ;
tokenizer.PushBack(token);
}
/// <summary> Gets token and checks if its end of line.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
/// <exception cref="IOException">if it doesn't find an end of line
/// </exception>
protected internal virtual void getLastToken(StreamTokenizer tokenizer,out Token token , bool endOfFileOk)
{
tokenizer.NextToken(out token);
if ( (!(token is EolToken)) && ( (!(token is EofToken)) || !endOfFileOk))
{
errms(tokenizer, "end of line expected");
}
}
/// <summary> Gets index, checking for a premature and of line.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
/// <exception cref="IOException">if it finds a premature end of line
/// </exception>
protected internal virtual void getIndex(StreamTokenizer tokenizer, out Token token)
{
tokenizer.NextToken(out token);
if (token is EolToken)
{
errms(tokenizer, "premature end of line");
}
if (token is EofToken)
{
errms(tokenizer, "premature end of file");
}
}
/// <summary> Gets next token, skipping empty lines.
///
/// </summary>
/// <param name="tokenizer">the stream tokenizer
/// </param>
/// <exception cref="IOException">if reading the next token fails
/// </exception>
protected internal virtual void getFirstToken(StreamTokenizer tokenizer,out Token token)
{
//while (tokenizer.NextToken() == SupportClass.StreamTokenizerSupport.TT_EOL)
tokenizer.NextToken(out token);
while(token is EolToken)
{
tokenizer.NextToken(out token);
} ;
//if ((tokenizer.ttype == '\'') || (tokenizer.ttype == '"'))
//if ((token.StringValue == "'") || (token.StringValue == "\"") )
//{
//tokenizer.ttype = SupportClass.StreamTokenizerSupport.TT_WORD;
//}
//else if ((tokenizer.ttype == SupportClass.StreamTokenizerSupport.TT_WORD) && (tokenizer.sval.Equals("?")))
//{
// tokenizer.ttype = '?';
//}
}
/// <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> Read a matrix from a stream. The format is the same the print method,
/// so printed matrices can be read back in (provided they were printed using
/// US Locale). XmlElements are separated by
/// whitespace, all the elements for each row appear on a single line,
/// the last row is followed by a blank line.
/// <p/>
/// Note: This format differs from the one that can be read via the
/// Matrix(Reader) constructor! For that format, the write(Writer) method
/// is used (from the original weka.core.Matrix class).
///
/// </summary>
/// <param name="input">the input stream.
/// </param>
/// <seealso cref="Matrix(Reader)">
/// </seealso>
/// <seealso cref="write(Writer)">
/// </seealso>
public static Matrix read(System.IO.StreamReader input)
{
StreamTokenizer tokenizer = new StreamTokenizer(input);
// Although StreamTokenizer will parse numbers, it doesn't recognize
// scientific notation (E or D); however, Double.valueOf does.
// The strategy here is to disable StreamTokenizer's number parsing.
// We'll only get whitespace delimited words, EOL's and EOF's.
// These words should all be numbers, for Double.valueOf to parse.
tokenizer.Settings.SetDefaults();//.ResetSyntax();
tokenizer.Settings.WordChars(0, 255);
tokenizer.Settings.WhitespaceChars(0,(int) ' ');// .WhitespaceChars(0, ' ');
tokenizer.Settings.GrabEol=true;
System.Collections.ArrayList v = System.Collections.ArrayList.Synchronized(new System.Collections.ArrayList(10));
// Ignore initial empty lines
Token token;
tokenizer.NextToken(out token);
while (token is EolToken)// == SupportClass.StreamTokenizerSupport.TT_EOL)
;
//if (token.ttype == SupportClass.StreamTokenizerSupport.TT_EOF)
if (token is EofToken)
throw new System.IO.IOException("Unexpected EOF on matrix read.");
do
{
// v.Add(System.Double.Parse(tokenizer.sval)); // Read & store 1st row.
v.Add(System.Double.Parse(token.StringValue)); // Read & store 1st row.
tokenizer.NextToken(out token);
}
while (token is WordToken);
//while (tokenizer.NextToken() == SupportClass.StreamTokenizerSupport.TT_WORD);
int n = v.Count; // Now we've got the number of columns!
double[] row = new double[n];
for (int j = 0; j < n; j++)
// extract the elements of the 1st row.
row[j] = ((System.Double) v[j]);
v.Clear();
v.Add(row); // Start storing rows instead of columns.
tokenizer.NextToken(out token);
while (token is WordToken)
//while (tokenizer.NextToken() == SupportClass.StreamTokenizerSupport.TT_WORD)
{
// While non-empty lines
v.Add(row = new double[n]);
int j = 0;
do
{
if (j >= n)
throw new System.IO.IOException("Row " + v.Count + " is too long.");
//row[j++] = System.Double.Parse(tokenizer.sval);
row[j++] = System.Double.Parse(token.StringValue);
tokenizer.NextToken(out token);
}
while (token is WordToken);
//while (tokenizer.NextToken() == SupportClass.StreamTokenizerSupport.TT_WORD);
if (j < n)
throw new System.IO.IOException("Row " + v.Count + " is too short.");
}
int m = v.Count; // Now we've got the number of rows.
double[][] A = new double[m][];
v.CopyTo(A); // copy the rows out of the vector
return new Matrix(A);
}
