public static List <String> getStringNGramSet(String item, Int32 N = 2, nGramsModeEnum mode = nGramsModeEnum.overlap)
{
var charSets = setAnalysisTools <Char> .getNGramSet <List <Char> >(item.ToCharArray().ToList(), N, mode);
List <String> ngrams = new List <string>();
foreach (List <char> charSet in charSets)
{
ngrams.Add(String.Concat(charSet));
}
return(ngrams);
}
/// <summary>
/// Gets descriptive line about n-grams deconstruction of the specified word
/// </summary>
/// <param name="word">The word to be splitted into n-grams</param>
/// <param name="N">Size of N-grams, e.g. for bigrams: N=2</param>
/// <param name="mode">The slicing mode</param>
/// <returns>Line used for debugging </returns>
public static String getNGramsDescriptiveLine(String word, Int32 N = 2, nGramsModeEnum mode = nGramsModeEnum.overlap)
{
List <String> ngrams = getNGrams(word, N, mode);
String line = "[" + word + "] (" + mode.ToString() + ", N=" + N + ") => ";
foreach (String ng in ngrams)
{
line = line.add(ng, ", ");
}
return(line);
}
/// <summary>
/// Breaks the specified word into <c>N</c>-grams
/// </summary>
/// <param name="word">The word to be splitted into n-grams</param>
/// <param name="N">Size of N-grams, e.g. for bigrams: N=2</param>
/// <param name="mode">The slicing mode</param>
/// <returns>Set of NGrams</returns>
public static List <String> getNGrams(String word, Int32 N = 2, nGramsModeEnum mode = nGramsModeEnum.overlap)
{
List <String> output = new List <string>();
Int32 step = 1;
Int32 remnant = 0;
switch (mode)
{
case nGramsModeEnum.overlap:
step = 1;
remnant = 1;
break;
case nGramsModeEnum.ordinal:
step = N;
remnant = 0;
break;
}
if (word.Length <= N)
{
output.Add(word);
return(output);
}
for (int i = 0; i < word.Length; i = i + step)
{
Int32 len = Math.Min(N, word.Length - i);
if (len > remnant)
{
output.Add(word.Substring(i, len));
}
}
return(output);
}
/// <summary>
/// Computes word similarity
/// </summary>
/// <param name="wordA">The word a.</param>
/// <param name="wordB">The word b.</param>
/// <param name="equationEnum">The equation enum.</param>
/// <param name="nGramSize">Size of the n gram.</param>
/// <param name="nGramMode">The n gram mode.</param>
/// <returns></returns>
public static Double GetSimilarity(String wordA, String wordB, nGramsSimilarityEquationEnum equationEnum = nGramsSimilarityEquationEnum.JaccardIndex, Int32 nGramSize = 2, nGramsModeEnum nGramMode = nGramsModeEnum.overlap)
{
var ngram_A = getNGrams(wordA, nGramSize, nGramMode);
var ngram_B = getNGrams(wordB, nGramSize, nGramMode);
switch (equationEnum)
{
default:
case nGramsSimilarityEquationEnum.JaccardIndex:
return(GetJaccardIndex(ngram_A, ngram_B));
break;
case nGramsSimilarityEquationEnum.DiceCoefficient:
return(GetDiceCoefficient(ngram_A, ngram_B));
break;
case nGramsSimilarityEquationEnum.continualOverlapRatio:
return(GetContinualOverlapRatio(ngram_A, ngram_B));
break;
}
return(0);
}
/// <summary>
/// Breaks the specified sequence into <c>N</c>-gram sub sequences
/// </summary>
/// <param name="sets">The sets.</param>
/// <param name="N">Size of N-grams, e.g. for bigrams: N=2</param>
/// <param name="mode">The slicing mode</param>
/// <returns>
/// Set of NGrams
/// </returns>
public List <List <List <T> > > getNGrams(List <List <T> > sets, Int32 N = 2, nGramsModeEnum mode = nGramsModeEnum.overlap)
{
List <List <List <T> > > output = new List <List <List <T> > >();
Int32 step = 1;
Int32 remnant = 0;
switch (mode)
{
case nGramsModeEnum.overlap:
{
step = 1;
remnant = 1;
break;
}
case nGramsModeEnum.ordinal:
{
step = N;
remnant = 0;
break;
}
default:
throw new Exception("Unexpected Case");
}
if (sets.Count <= N)
{
output.Add(sets);
return(output);
}
for (int i = 0; i < sets.Count; i = i + step)
{
Int32 len = Math.Min(N, sets.Count - i);
if (len > remnant)
{
output.Add(sets.GetRange(i, len));
}
}
return(output);
}
/// <summary>
/// Breaks the specified sequence of items into n-gram chunks
/// </summary>
/// <param name="items">The items.</param>
/// <param name="N">The n.</param>
/// <param name="mode">The mode.</param>
/// <returns></returns>
/// <exception cref="Exception">Unexpected Case</exception>
public static List <TNGram> getNGramSet <TNGram>(List <T> items, Int32 N = 2, nGramsModeEnum mode = nGramsModeEnum.overlap) where TNGram : List <T>, new()
{
List <TNGram> output = new List <TNGram>();
Int32 step = 1;
Int32 remnant = 0;
switch (mode)
{
case nGramsModeEnum.overlap:
{
step = 1;
remnant = 1;
break;
}
case nGramsModeEnum.ordinal:
{
step = N;
remnant = 0;
break;
}
default:
throw new Exception("Unexpected Case");
}
if (items.Count <= N)
{
TNGram nGram = new TNGram();
nGram.AddRange(items);
output.Add(nGram);
return(output);
}
for (int i = 0; i < items.Count; i = i + step)
{
Int32 len = Math.Min(N, items.Count - i);
if (len > remnant)
{
TNGram nGram = new TNGram();
nGram.AddRange(items.GetRange(i, len));
output.Add(nGram);
}
}
return(output);
}
