/**
* Gets an array containing the indices of all string attributes.
*
* @param insts the Instances to scan for string attributes.
* @return an array containing the indices of string attributes in the input structure. Will be
* zero-length if there are no string attributes
*/
protected int[] getStringIndices(M5Instances insts) {
// Scan through getting the indices of String attributes
int[] index = new int[insts.numAttributes()];
int indexSize = 0;
for (int i = 0; i < insts.numAttributes(); i++) {
if (insts.attribute(i).type() == M5Attribute.STRING) {
index[indexSize++] = i;
}
}
int[] result = new int[indexSize];
System.arraycopy(index, 0, result, 0, indexSize);
return result;
}
/**
* Copies string values contained in the instance copied to a new dataset. The Instance must
* already be assigned to a dataset. This dataset and the destination dataset must have the same
* structure.
*
* @param instance the Instance containing the string values to copy.
* @param destDataset the destination set of Instances
* @param strAtts an array containing the indices of any string attributes in the dataset.
*/
private void copyStringValues(M5Instance inst, M5Instances destDataset, int[] strAtts) {
if (strAtts.length == 0) {
return;
}
if (inst.dataset() == null) {
throw new IllegalArgumentException("Instance has no dataset assigned!!");
} else if (inst.dataset().numAttributes() != destDataset.numAttributes()) {
throw new IllegalArgumentException("Src and Dest differ in # of attributes!!");
}
copyStringValues(inst, true, inst.dataset(), strAtts, destDataset, strAtts);
}
/**
* Method for testing filters ability to process multiple batches.
*
* @param options should contain the following arguments:<br>
* -i (first) input file <br>
* -o (first) output file <br>
* -r (second) input file <br>
* -s (second) output file <br>
* -c class_index <br>
* or -h for help on options
* @exception Exception if something goes wrong or the user requests help on command options
*/
public static void batchFilterFile(NominalToBinaryFilter filter, String[] options)
throws Exception {
M5Instances firstData = null;
M5Instances secondData = null;
Reader firstInput = null;
Reader secondInput = null;
PrintWriter firstOutput = null;
PrintWriter secondOutput = null;
boolean helpRequest;
try {
helpRequest = M5StaticUtils.getFlag('h', options);
String fileName = M5StaticUtils.getOption('i', options);
if (fileName.length() != 0) {
firstInput = new BufferedReader(new FileReader(fileName));
} else {
throw new Exception("No first input file given.\n");
}
fileName = M5StaticUtils.getOption('r', options);
if (fileName.length() != 0) {
secondInput = new BufferedReader(new FileReader(fileName));
} else {
throw new Exception("No second input file given.\n");
}
fileName = M5StaticUtils.getOption('o', options);
if (fileName.length() != 0) {
firstOutput = new PrintWriter(new FileOutputStream(fileName));
} else {
firstOutput = new PrintWriter(System.out);
}
fileName = M5StaticUtils.getOption('s', options);
if (fileName.length() != 0) {
secondOutput = new PrintWriter(new FileOutputStream(fileName));
} else {
secondOutput = new PrintWriter(System.out);
}
String classIndex = M5StaticUtils.getOption('c', options);
if (filter instanceof NominalToBinaryFilter) {
((NominalToBinaryFilter) filter).setOptions(options);
}
M5StaticUtils.checkForRemainingOptions(options);
if (helpRequest) {
throw new Exception("Help requested.\n");
}
firstData = new M5Instances(firstInput, 1);
secondData = new M5Instances(secondInput, 1);
if (!secondData.equalHeaders(firstData)) {
throw new Exception("Input file formats differ.\n");
}
if (classIndex.length() != 0) {
if (classIndex.equals("first")) {
firstData.setClassIndex(0);
secondData.setClassIndex(0);
} else if (classIndex.equals("last")) {
firstData.setClassIndex(firstData.numAttributes() - 1);
secondData.setClassIndex(secondData.numAttributes() - 1);
} else {
firstData.setClassIndex(Integer.parseInt(classIndex) - 1);
secondData.setClassIndex(Integer.parseInt(classIndex) - 1);
}
}
} catch (Exception ex) {
String filterOptions = "";
// Output the error and also the valid options
if (filter instanceof NominalToBinaryFilter) {
filterOptions += "\nFilter options:\n\n";
Enumeration enume = ((NominalToBinaryFilter) filter).listOptions();
while (enume.hasMoreElements()) {
Information option = (Information) enume.nextElement();
filterOptions += option.synopsis() + '\n' + option.description() + "\n";
}
}
String genericOptions =
"\nGeneral options:\n\n"
+ "-h\n"
+ "\tGet help on available options.\n"
+ "-i <filename>\n"
+ "\tThe file containing first input instances.\n"
+ "-o <filename>\n"
+ "\tThe file first output instances will be written to.\n"
+ "-r <filename>\n"
+ "\tThe file containing second input instances.\n"
+ "-s <filename>\n"
+ "\tThe file second output instances will be written to.\n"
+ "-c <class index>\n"
+ "\tThe number of the attribute to use as the class.\n"
+ "\t\"first\" and \"last\" are also valid entries.\n"
+ "\tIf not supplied then no class is assigned.\n";
throw new Exception('\n' + ex.getMessage() + filterOptions + genericOptions);
}
boolean printedHeader = false;
if (filter.setInputFormat(firstData)) {
firstOutput.println(filter.getOutputFormat().toString());
printedHeader = true;
}
// Pass all the instances to the filter
while (firstData.readInstance(firstInput)) {
if (filter.input(firstData.instance(0))) {
if (!printedHeader) {
throw new Error("Filter didn't return true from setInputFormat() " + "earlier!");
}
firstOutput.println(filter.output().toString());
}
firstData.delete(0);
}
// Say that input has finished, and print any pending output instances
if (filter.batchFinished()) {
if (!printedHeader) {
firstOutput.println(filter.getOutputFormat().toString());
}
while (filter.numPendingOutput() > 0) {
firstOutput.println(filter.output().toString());
}
}
if (firstOutput != null) {
firstOutput.close();
}
printedHeader = false;
if (filter.isOutputFormatDefined()) {
secondOutput.println(filter.getOutputFormat().toString());
printedHeader = true;
}
// Pass all the second instances to the filter
while (secondData.readInstance(secondInput)) {
if (filter.input(secondData.instance(0))) {
if (!printedHeader) {
throw new Error("Filter didn't return true from" + " isOutputFormatDefined() earlier!");
}
secondOutput.println(filter.output().toString());
}
secondData.delete(0);
}
// Say that input has finished, and print any pending output instances
if (filter.batchFinished()) {
if (!printedHeader) {
secondOutput.println(filter.getOutputFormat().toString());
}
while (filter.numPendingOutput() > 0) {
secondOutput.println(filter.output().toString());
}
}
if (secondOutput != null) {
secondOutput.close();
}
}
/**
* Method for testing filters.
*
* @param options should contain the following arguments: <br>
* -i input_file <br>
* -o output_file <br>
* -c class_index <br>
* or -h for help on options
* @exception Exception if something goes wrong or the user requests help on command options
*/
public static void filterFile(NominalToBinaryFilter filter, String[] options) throws Exception {
boolean debug = false;
M5Instances data = null;
Reader input = null;
PrintWriter output = null;
boolean helpRequest;
try {
helpRequest = M5StaticUtils.getFlag('h', options);
if (M5StaticUtils.getFlag('d', options)) {
debug = true;
}
String infileName = M5StaticUtils.getOption('i', options);
String outfileName = M5StaticUtils.getOption('o', options);
String classIndex = M5StaticUtils.getOption('c', options);
if (filter instanceof NominalToBinaryFilter) {
((NominalToBinaryFilter) filter).setOptions(options);
}
M5StaticUtils.checkForRemainingOptions(options);
if (helpRequest) {
throw new Exception("Help requested.\n");
}
if (infileName.length() != 0) {
input = new BufferedReader(new FileReader(infileName));
} else {
input = new BufferedReader(new InputStreamReader(System.in));
}
if (outfileName.length() != 0) {
output = new PrintWriter(new FileOutputStream(outfileName));
} else {
output = new PrintWriter(System.out);
}
data = new M5Instances(input, 1);
if (classIndex.length() != 0) {
if (classIndex.equals("first")) {
data.setClassIndex(0);
} else if (classIndex.equals("last")) {
data.setClassIndex(data.numAttributes() - 1);
} else {
data.setClassIndex(Integer.parseInt(classIndex) - 1);
}
}
} catch (Exception ex) {
String filterOptions = "";
// Output the error and also the valid options
if (filter instanceof NominalToBinaryFilter) {
filterOptions += "\nFilter options:\n\n";
Enumeration enuma = ((NominalToBinaryFilter) filter).listOptions();
while (enuma.hasMoreElements()) {
Information option = (Information) enuma.nextElement();
filterOptions += option.synopsis() + '\n' + option.description() + "\n";
}
}
String genericOptions =
"\nGeneral options:\n\n"
+ "-h\n"
+ "\tGet help on available options.\n"
+ "\t(use -b -h for help on batch mode.)\n"
+ "-i <file>\n"
+ "\tThe name of the file containing input instances.\n"
+ "\tIf not supplied then instances will be read from stdin.\n"
+ "-o <file>\n"
+ "\tThe name of the file output instances will be written to.\n"
+ "\tIf not supplied then instances will be written to stdout.\n"
+ "-c <class index>\n"
+ "\tThe number of the attribute to use as the class.\n"
+ "\t\"first\" and \"last\" are also valid entries.\n"
+ "\tIf not supplied then no class is assigned.\n";
throw new Exception('\n' + ex.getMessage() + filterOptions + genericOptions);
}
if (debug) {
System.err.println("Setting input format");
}
boolean printedHeader = false;
if (filter.setInputFormat(data)) {
if (debug) {
System.err.println("Getting output format");
}
output.println(filter.getOutputFormat().toString());
printedHeader = true;
}
// Pass all the instances to the filter
while (data.readInstance(input)) {
if (debug) {
System.err.println("Input instance to filter");
}
if (filter.input(data.instance(0))) {
if (debug) {
System.err.println("Filter said collect immediately");
}
if (!printedHeader) {
throw new Error("Filter didn't return true from setInputFormat() " + "earlier!");
}
if (debug) {
System.err.println("Getting output instance");
}
output.println(filter.output().toString());
}
data.delete(0);
}
// Say that input has finished, and print any pending output instances
if (debug) {
System.err.println("Setting end of batch");
}
if (filter.batchFinished()) {
if (debug) {
System.err.println("Filter said collect output");
}
if (!printedHeader) {
if (debug) {
System.err.println("Getting output format");
}
output.println(filter.getOutputFormat().toString());
}
if (debug) {
System.err.println("Getting output instance");
}
while (filter.numPendingOutput() > 0) {
output.println(filter.output().toString());
if (debug) {
System.err.println("Getting output instance");
}
}
}
if (debug) {
System.err.println("Done");
}
if (output != null) {
output.close();
}
}