private void load(string clusterDir)
{
DocumentProcessor docProcessor = new DocumentProcessor();
ArrayList docs = new ArrayList();
string[] clusterFiles = Directory.GetFiles(clusterDir, "*.txt", SearchOption.TopDirectoryOnly);
foreach (string filename in clusterFiles)
{
string fileText = File.ReadAllText(filename, Encoding.Default);
Document doc = docProcessor.process(fileText);
docs.Add(doc);
}
DocsStatistics docStats = DocsStatistics.generateStatistics(docs);
Hashtable centroid = new Hashtable();
foreach (string word in docStats.wordsCount.Keys)
{
//centroid[firstWord] = (((int)docStats.wordsCount[firstWord]) * idf(docStats, firstWord)) / docs.Count;
centroid[word] = (((int)docStats.wordsCount[word]) * IDF.getInstance().get(word) ) / docs.Count;
}
this.centroidWords = applyKeepWords(centroid, this.keepWords);
}
private DocsStatistics processFiles(string[] files)
{
DocumentProcessor docProcessor = new DocumentProcessor();
DocsStatistics docStats = new DocsStatistics();
foreach (string filename in files)
{
Document doc = docProcessor.process(filename);
docStats.addDocument(doc);
}
return (docStats);
}
public void preprocessLanguageModel(string[] documentFiles, string bigramFilePath)
{
// No need for Stop Words Removal.
StopWordsHandler.setInstance(new NullStopWordsHandler());
DocumentProcessor docProcessor = new DocumentProcessor();
BigramStatisticsModel bigramStats = new BigramStatisticsModel();
int i = 0;
foreach (string filename in documentFiles)
{
++i;
string fileContent = File.ReadAllText(filename, Encoding.Default);
Document doc = docProcessor.process(fileContent);
bigramStats.addDocument(doc);
}
bigramStats.toFile(bigramFilePath);
}
private static void processFile(DocsStatistics docStats, string filename)
{
DocumentProcessor docProcessor = new DocumentProcessor();
string fileContent = File.ReadAllText(filename, Encoding.Default);
using (Document doc = docProcessor.process(fileContent))
{
docStats.addDocument(doc);
}
}
public static IDF fromFiles(string[] files)
{
DocsStatistics docStats = new DocsStatistics();
DocumentProcessor docProcessor = new DocumentProcessor();
int i = 0;
foreach (string file in files)
{
++i;
//processFile(docStats, file);
//*
string fileContent = File.ReadAllText(file, Encoding.Default);
Document doc = docProcessor.process(fileContent);
docStats.addDocument(doc);
/*
if ((i % 1000) == 0)
{
System.GC.Collect();
Trace.write("Done for : " + i);
}
//*/
//*/
//doc = null;
}
IDF idf = new IDF();
foreach (string word in docStats.wordsCount.Keys)
{
//double wordRefCount = docStats.wordRefs[firstWord] == null ? 0 : ((HashSet<Document>)docStats.wordRefs[firstWord]).Count;
double wordRefCount = docStats.wordRefsCount[word] == null ? 0 : ((int)docStats.wordRefsCount[word]);
double wordIdf = Math.Log(docStats.docCount / (wordRefCount));
idf.idf[word] = wordIdf;
}
return (idf);
}
// Relative Paths
//public const string TRAINING_PATH = @"Training\";
//public const string TESTING_PATH = @"Testing\";
public static DocumentProcessor getDocumentProcessor()
{
if (docProcessor == null)
docProcessor = new DocumentProcessor();
return (docProcessor);
}