/// <summary>
/// if the token contains a defined abbreviation, returned token will
/// contain the term wich was abbreviated. Otherwise same token is
/// returned
/// </summary>
/// <param name="token"></param>
/// <returns></returns>
public Token Transform(Token token)
{
string replacement;
return Abbreviations.TryGetValue(token.Value, out replacement)
? new Token(replacement)
: token;
}
/// <summary>
/// If token has been seen returns the number of times it was seen in added documents
/// otherwise returns null
/// </summary>
/// <param name="token"></param>
/// <returns></returns>
public int? GetDocumentFrequency(Token token)
{
int frequency = 0;
return _dict.TryGetValue(token, out frequency)
? frequency
: (int?)null;
}
/// <summary>
/// REturns a number between 0.0 (entirely different) to 1.0 (exactly the same)
/// </summary>
/// <param name="source"></param>
/// <param name="target"></param>
/// <returns></returns>
public double Score(Token source, Token target)
{
return _distance.GetSimilarity(source.Value, target.Value);
}