public static Dictionary <string, int> GetWordFrequency(string path, PorterStemming porter)
{
Init();
var dic = new Dictionary <string, int>();
using (StreamReader reader = new StreamReader(path))
{
string word = string.Empty;
do
{
char c = (char)reader.Read();
if (Array.IndexOf(_delimiters, c) == -1 && !Char.IsWhiteSpace(c))
{
word += c;
continue;
}
word = word.ToLower();
if (_exceptionWords.Contains(word) || _stopWords.Contains(word) || word.Length < 3)
{
word = string.Empty;
continue;
}
word = porter.Stem(word);
if (dic.ContainsKey(word))
{
dic[word] = dic[word] + 1;
}
else
{
dic[word] = 1;
}
word = string.Empty;
} while (!reader.EndOfStream);
}
return(dic);
}
public VectorialSearch()
{
Porter = new PorterStemming();
}