private void StripAnalysisText(List <string> rawAnalysisWords)
{
List <string> strippedList = new List <string>();
foreach (string s in rawAnalysisWords)
{
string p;
string q;
if (!string.IsNullOrEmpty(s))
{
if (Rgx.IsMatch(s))
{
p = Rgx.Replace(s, "").ToLower();
}
else
{
p = s.ToLower();
}
if (!StopWords.Contains(p.ToLower()) && !string.IsNullOrWhiteSpace(p))
{
if (Apos.IsMatch(p))
{
q = Apos.Replace(p, "").ToLower();
strippedList.Add(q);
}
else
{
strippedList.Add(p);
}
}
}
}
StrippedWords = strippedList;
}
public Dictionary <string, int> GetWordOccurancesFromText(string text)
{
if (string.IsNullOrEmpty(text))
{
throw new ArgumentNullException(nameof(text));
}
text = GetPlainText(text);
var result = new Dictionary <string, int>();
var matches = Regex.Matches(text, WordRegexPattern);
foreach (Match word in matches)
{
var key = word.Value.ToLower();
if (!StopWords.Contains(key, StringComparer.OrdinalIgnoreCase))
{
if (result.ContainsKey(key))
{
result[key]++;
}
else
{
result.Add(key, 1);
}
}
}
return(result);
}
private IDictionary <string, double> GetWordIfidf(string text, IEnumerable <string> allowPos)
{
IEnumerable <string> words = null;
if (allowPos.IsNotEmpty())
{
words = FilterCutByPos(text, allowPos);
}
else
{
words = Segmenter.Cut(text);
}
// Calculate TF
var freq = new Dictionary <string, double>();
foreach (var word in words)
{
var w = word;
if (string.IsNullOrEmpty(w) || w.Trim().Length < 2 || StopWords.Contains(w.ToLower()))
{
continue;
}
freq[w] = freq.GetDefault(w, 0.0) + 1.0;
}
var total = freq.Values.Sum();
foreach (var k in freq.Keys.ToList())
{
freq[k] *= IdfFreq.GetDefault(k, MedianIdf) / total;
}
return(freq);
}
public IEnumerable <Posting> Index(sqliteContext dbContext, List <string> words)
{
Dictionary <string, Posting> postings = new Dictionary <string, Posting>();
int index = 0;
foreach (var token in Tokens)
{
var word = token.ToString().ToLower();
index++;
if (StopWords.Contains(word))
{
continue;
}
if (words != null && !words.Contains(word))
{
continue;
}
if (dbContext?.IndexWord.Find(word) == null)
{
dbContext?.IndexWord.Add(new IndexWord()
{
Word = word
});
}
if (!postings.ContainsKey(word))
{
postings.Add(word, new Posting()
{
Word = word,
DocumentName = name,
Indexes = ""
});
}
var p = postings[word];
p.Frequency++;
if (p.Indexes != "")
{
p.Indexes += ",";
}
p.Indexes += index;
}
foreach (var posting in postings)
{
dbContext?.Posting.Add(posting.Value);
}
return(postings.Values);
}
/// <summary>
/// determines if the passed term is likely to be of interest in "more like" comparisons
/// </summary>
/// <param name="term"> The word being considered </param>
/// <returns> <c>true</c> if should be ignored, <c>false</c> if should be used in further analysis </returns>
private bool IsNoiseWord(string term)
{
int len = term.Length;
if (MinWordLen > 0 && len < MinWordLen)
{
return(true);
}
if (MaxWordLen > 0 && len > MaxWordLen)
{
return(true);
}
return(StopWords != null && StopWords.Contains(term));
}
public Boolean ShouldStop(ChatMessage chatMessage)
{
if (!chatMessage.IsModerator)
{
return(false);
}
if (chatMessage.Username.Equals("streamelements", StringComparison.OrdinalIgnoreCase))
{
return(false);
}
String[] messageWords = chatMessage.Message.Split(' ');
return(messageWords.Any(word => StopWords.Contains(word.ToLower()) || word.Equals(TwitchClientManager.Name.Value, StringComparison.OrdinalIgnoreCase)));
}
/// <summary>
/// Очищает текст используя базу (словарь) стоп-слов
/// </summary>
/// <param name="text">Неочищенный текст, который нужно канонизировать</param>
/// <returns>Коллекцию слов из текста, которые готовы к употреблению =)</returns>
private static List <string> TextPurify(string text)
{
//разделяем ввесь текст на отдельные слова
var rawTokens = text.Split(Separators).ToList();
//проходимся по этому списку слов в linq-выражении
var canonedTokens = rawTokens.Select(word => word.ToCharArray().Where(n => !char.IsDigit(n)).ToArray()).Select(purified => new string(purified)).ToList();
//из этой коллекции удаляем все пустые элементы и стоп-слова используя linq
canonedTokens.RemoveAll(item => StopWords.Contains(item.ToLower()) || string.IsNullOrWhiteSpace(item));
//также удаляются все стоп-символы из слов в коллекции
var purifiedTokens = (from item in canonedTokens let regex = new Regex("[0-9/|_!@#$%^&*()_+=?:;.,{}№><«»'\"`~" + @"\\[\]– -]*") select regex.Replace(item, "")).ToList();
//устанавливаются все слова в Lower Case
var purifiedLowerCaseTokens = purifiedTokens.Select(purifiedToken => purifiedToken.ToLower()).ToList();
var stemmedLowerCaseTokens = new List <string>();
var cyrillicStemmer = new RussianStemmer();
var latinStemmer = new EnglishStemmer();
foreach (var purifiedLowerCaseToken in purifiedLowerCaseTokens)
{
switch (Verifications.GetFontType(purifiedLowerCaseToken))
{
case FontType.Cyrillic:
stemmedLowerCaseTokens.Add(cyrillicStemmer.Stem(purifiedLowerCaseToken));
break;
case FontType.Latin:
stemmedLowerCaseTokens.Add(latinStemmer.Stem(purifiedLowerCaseToken));
break;
case FontType.Other:
case FontType.Numbers:
break;
default:
throw new ArgumentOutOfRangeException();
}
}
return(stemmedLowerCaseTokens);
}
public bool PairFilter(IEnumerable <string> allowPos, Pair wp)
{
return(allowPos.Contains(wp.Flag) &&
wp.Word.Trim().Length >= 2 &&
!StopWords.Contains(wp.Word.ToLower()));
}
public bool PairFilter(Pair wp)
{
return(DefaultPosFilter.Contains(wp.Flag) &&
wp.Word.Trim().Length >= 2 &&
!StopWords.Contains(wp.Word.ToLower()));
}
/// <summary>
/// Removes stop words and HTML from the specified string.
/// </summary>
/// <param name="content">
/// The content.
/// </param>
/// <param name="removeHtml">
/// The remove Html.
/// </param>
/// <returns>
/// The clean content.
/// </returns>
private static string CleanContent(string content, bool removeHtml)
{
if (removeHtml)
{
content = Utils.StripHtml(content);
}
content =
content.Replace("\\", string.Empty).Replace("|", string.Empty).Replace("(", string.Empty).Replace(
")", string.Empty).Replace("[", string.Empty).Replace("]", string.Empty).Replace("*", string.Empty).
Replace("?", string.Empty).Replace("}", string.Empty).Replace("{", string.Empty).Replace(
"^", string.Empty).Replace("+", string.Empty);
var words = content.Split(new[] { ' ', '\n', '\r' }, StringSplitOptions.RemoveEmptyEntries);
var sb = new StringBuilder();
foreach (var word in
words.Select(t => t.ToLowerInvariant().Trim()).Where(word => word.Length > 1 && !StopWords.Contains(word)))
{
sb.AppendFormat("{0} ", word);
}
return(sb.ToString());
}
