/// <summary>
/// Detect sentences in a string.
/// </summary>
/// <param name="input">
/// The string to be processed.
/// </param>
/// <param name="positions">
/// Start index and length of each detected sentence.
/// </param>
/// <returns>
/// A string array containing individual sentences as elements.
/// </returns>
public string[] SentenceDetect(string input, out Util.Pair <int, int>[] positions)
{
int[] startsList = SentencePositionDetect(input);
positions = new Util.Pair <int, int>[] { new Util.Pair <int, int>(0, input.Length) };
if (startsList.Length == 0)
{
return(new string[] { input });
}
bool isLeftover = startsList[startsList.Length - 1] != input.Length;
string[] sentences = new string[isLeftover ? startsList.Length + 1 : startsList.Length];
positions = new Util.Pair <int, int> [sentences.Length];
positions[0] = new Util.Pair <int, int>(0, (startsList[0]) - (0));
sentences[0] = input.Substring(positions[0].FirstValue, positions[0].SecondValue);
for (int currentStart = 1; currentStart < startsList.Length; currentStart++)
{
positions[currentStart] = new Util.Pair <int, int>(startsList[currentStart - 1], (startsList[currentStart]) - (startsList[currentStart - 1]));
sentences[currentStart] = input.Substring(positions[currentStart].FirstValue, positions[currentStart].SecondValue);
}
if (isLeftover)
{
positions[sentences.Length - 1] = new Util.Pair <int, int>(startsList[startsList.Length - 1], input.Length - startsList[startsList.Length - 1]);
sentences[sentences.Length - 1] = input.Substring(positions[sentences.Length - 1].FirstValue);
}
return(sentences);
}
/// <summary>
/// Generates a Parse structure from the specified tree-bank style parse string.
/// </summary>
/// <param name="parse">
/// A tree-bank style parse string.
/// </param>
/// <returns>
/// a Parse structure for the specified tree-bank style parse string.
/// </returns>
public static Parse FromParseString(string parse)
{
StringBuilder textBuffer = new StringBuilder();
int offset = 0;
Stack <Util.Pair <string, int> > parseStack = new Stack <Util.Pair <string, int> >();
List <Util.Pair <string, Util.Span> > consitutents = new List <Util.Pair <string, Util.Span> >();
for (int currentChar = 0, charCount = parse.Length; currentChar < charCount; currentChar++)
{
char c = parse[currentChar];
if (c == '(')
{
string rest = parse.Substring(currentChar + 1);
string type = GetType(rest);
if (type == null)
{
throw new ParseException("null type for: " + rest);
}
string token = GetToken(rest);
parseStack.Push(new Util.Pair <string, int>(type, offset));
if ((object)token != null && type != "-NONE-")
{
consitutents.Add(new Util.Pair <string, Util.Span>(MaximumEntropyParser.TokenNode, new Util.Span(offset, offset + token.Length)));
textBuffer.Append(token).Append(" ");
offset += token.Length + 1;
}
}
else if (c == ')')
{
Util.Pair <string, int> parts = parseStack.Pop();
string type = parts.FirstValue;
if (type != "-NONE-")
{
int start = parts.SecondValue;
consitutents.Add(new Util.Pair <string, Util.Span>(parts.FirstValue, new Util.Span(start, offset - 1)));
}
}
}
string text = textBuffer.ToString();
Parse rootParse = new Parse(text, new Util.Span(0, text.Length), MaximumEntropyParser.TopNode, 1);
for (int currentConstituent = 0, constituentCount = consitutents.Count; currentConstituent < constituentCount; currentConstituent++)
{
Util.Pair <string, Util.Span> parts = consitutents[currentConstituent];
string type = parts.FirstValue;
if (type != MaximumEntropyParser.TopNode)
{
Parse newConstituent = new Parse(text, parts.SecondValue, type, 1);
rootParse.Insert(newConstituent);
}
}
return(rootParse);
}
/// <summary>
/// Applies a nth-order Butterworth filter.
/// </summary>
/// <param name="factor"></param>
/// <param name="filterOrder"></param>
public DecimationFilter(int factor, int filterOrder, bool bidirectional)
{
this.factor = factor;
Util.Pair<double,double> cornerFreqs = new Util.Pair<double,double>(Math.PI / factor, 0.0);
FilterOrderSpec spec = FilterOrderSpec.CreateButterworthSpec(cornerFreqs, filterOrder, BandType.LowPass);
Filter antiAliasFilter = FilterFactory.CreateIirFilter(spec);
if(bidirectional)
antiAliasFilter = new BidirectionalFilter(antiAliasFilter);
Filter downSampleFilter = new DownSampleFilter(factor);
filter = new FilterChain(new Filter[] { antiAliasFilter, downSampleFilter });
return;
}
private void AddEvents(string line)
{
Util.Pair <ArrayList, ArrayList> linePair = ConvertAnnotatedString(line);
ArrayList tokens = linePair.FirstValue;
ArrayList outcomes = linePair.SecondValue;
List <string> tags = new List <string>();
for (int currentToken = 0; currentToken < tokens.Count; currentToken++)
{
string[] context = mContextGenerator.GetContext(currentToken, tokens.ToArray(), tags.ToArray(), null);
SharpEntropy.TrainingEvent posTrainingEvent = new SharpEntropy.TrainingEvent((string)outcomes[currentToken], context);
tags.Add((string)outcomes[currentToken]);
mEventList.Add(posTrainingEvent);
}
}
public static Util.Pair <ArrayList, ArrayList> ConvertAnnotatedString(string input)
{
ArrayList tokens = new ArrayList();
ArrayList outcomes = new ArrayList();
Util.StringTokenizer tokenizer = new Util.StringTokenizer(input);
string token = tokenizer.NextToken();
while (token != null)
{
Util.Pair <string, string> linePair = Split(token);
tokens.Add(linePair.FirstValue);
outcomes.Add(linePair.SecondValue);
token = tokenizer.NextToken();
}
return(new Util.Pair <ArrayList, ArrayList>(tokens, outcomes));
}
/// <summary>
/// Builds up the list of features based on the information in the object,
/// which is a pair containing a string and and integer which
/// indicates the index of the position we are investigating.
/// </summary>
public virtual string[] GetContext(Util.Pair <string, int> pair)
{
string data = pair.FirstValue;
int index = pair.SecondValue;
List <string> predicates = new List <string>();
predicates.Add("p=" + data.Substring(0, (index) - (0)));
predicates.Add("s=" + data.Substring(index));
if (index > 0)
{
AddCharPredicates("p1", data[index - 1], predicates);
if (index > 1)
{
AddCharPredicates("p2", data[index - 2], predicates);
predicates.Add("p21=" + data[index - 2] + data[index - 1]);
}
else
{
predicates.Add("p2=bok");
}
predicates.Add("p1f1=" + data[index - 1] + data[index]);
}
else
{
predicates.Add("p1=bok");
}
AddCharPredicates("f1", data[index], predicates);
if (index + 1 < data.Length)
{
AddCharPredicates("f2", data[index + 1], predicates);
predicates.Add("f12=" + data[index] + data[index + 1]);
}
else
{
predicates.Add("f2=bok");
}
if (data[0] == '&' && data[data.Length - 1] == ';')
{
predicates.Add("cc"); //character code
}
return(predicates.ToArray());
}
private void AddNewEvents(string token)
{
System.Text.StringBuilder buffer = mBuffer;
buffer.Append(token.Trim());
int sentenceEndPosition = buffer.Length - 1;
//add following word to stringbuilder
if (mNext != null && token.Length > 0)
{
int positionAfterFirstWordInNext = mNext.IndexOf(" ");
if (positionAfterFirstWordInNext != -1)
{
// should maybe changes this so that it usually adds a space
// before the next sentence, but sometimes leaves no space.
buffer.Append(" ");
buffer.Append(mNext.Substring(0, (positionAfterFirstWordInNext) - (0)));
}
}
for (System.Collections.IEnumerator iterator = mScanner.GetPositions(buffer).GetEnumerator(); iterator.MoveNext();)
{
int candidate = (int)iterator.Current;
Util.Pair <System.Text.StringBuilder, int> pair = new Util.Pair <System.Text.StringBuilder, int>(buffer, candidate);
string type = (candidate == sentenceEndPosition) ? "T" : "F";
SentenceDetectionEvent sentenceEvent = new SentenceDetectionEvent(type, mContextGenerator.GetContext(pair));
if (null != mTail)
{
mTail.NextEvent = sentenceEvent;
mTail = sentenceEvent;
}
else if (null == mHead)
{
mHead = sentenceEvent;
}
else if (null == mHead.NextEvent)
{
mHead.NextEvent = mTail = sentenceEvent;
}
}
buffer.Length = 0;
}
/// <summary>
/// Detect the position of the first words of sentences in a string.
/// </summary>
/// <param name="input">
/// The string to be processed.
/// </param>
/// <returns>
/// A integer array containing the positions of the end index of
/// every sentence
/// </returns>
public virtual int[] SentencePositionDetect(string input)
{
if (mUnicodeMapping)
{
input = Utils.MapUnicodeChars(input);
}
double sentenceProbability = 1;
mSentenceProbs.Clear();
System.Text.StringBuilder buffer = new System.Text.StringBuilder(input);
List <int> endersList = mScanner.GetPositions(input);
List <int> positions = new List <int>(endersList.Count);
for (int currentEnder = 0, enderCount = endersList.Count, index = 0; currentEnder < enderCount; currentEnder++)
{
int candidate = endersList[currentEnder];
int cInt = candidate;
// skip over the leading parts of non-token final delimiters
int firstWhiteSpace = GetFirstWhitespace(input, cInt + 1);
if (((currentEnder + 1) < enderCount) && ((endersList[currentEnder + 1]) < firstWhiteSpace))
{
continue;
}
Util.Pair <System.Text.StringBuilder, int> pair = new Util.Pair <System.Text.StringBuilder, int>(buffer, candidate);
double[] probabilities = mModel.Evaluate(mContextGenerator.GetContext(pair));
string bestOutcome = mModel.GetBestOutcome(probabilities);
sentenceProbability *= probabilities[mModel.GetOutcomeIndex(bestOutcome)];
if (bestOutcome.Equals("T") && IsAcceptableBreak(input, index, cInt))
{
if (index != cInt)
{
positions.Add(GetFirstNonWhitespace(input, GetFirstWhitespace(input, cInt + 1))); //moIntegerPool.GetInteger(GetFirstNonWhitespace(input, GetFirstWhitespace(input, cInt + 1))));
mSentenceProbs.Add(probabilities[mModel.GetOutcomeIndex(bestOutcome)]);
}
index = cInt + 1;
}
}
return(positions.ToArray());
}
public virtual void LocalEvaluate(SharpEntropy.IMaximumEntropyModel posModel, System.IO.StreamReader reader, out double accuracy, out double sentenceAccuracy)
{
mPosModel = posModel;
float total = 0, correct = 0, sentences = 0, sentencesCorrect = 0;
System.IO.StreamReader sentenceReader = new System.IO.StreamReader(reader.BaseStream, System.Text.Encoding.UTF7);
string line;
while ((object)(line = sentenceReader.ReadLine()) != null)
{
sentences++;
Util.Pair <ArrayList, ArrayList> annotatedPair = PosEventReader.ConvertAnnotatedString(line);
ArrayList words = annotatedPair.FirstValue;
ArrayList outcomes = annotatedPair.SecondValue;
ArrayList tags = new ArrayList(Beam.BestSequence(words, null).Outcomes);
int count = 0;
bool isSentenceOK = true;
for (System.Collections.IEnumerator tagIndex = tags.GetEnumerator(); tagIndex.MoveNext(); count++)
{
total++;
string tag = (string)tagIndex.Current;
if (tag == (string)outcomes[count])
{
correct++;
}
else
{
isSentenceOK = false;
}
}
if (isSentenceOK)
{
sentencesCorrect++;
}
}
accuracy = correct / total;
sentenceAccuracy = sentencesCorrect / sentences;
}
public virtual void CreatePdf(String dest)
{
//Initialize PDF document
PdfDocument pdf = new PdfDocument(new PdfWriter(dest));
// Initialize document
Document document = new Document(pdf);
PdfFont font = PdfFontFactory.CreateFont(StandardFonts.TIMES_ROMAN);
PdfFont bold = PdfFontFactory.CreateFont(StandardFonts.HELVETICA_BOLD);
document.SetTextAlignment(TextAlignment.JUSTIFIED).SetHyphenation(new HyphenationConfig("en", "uk", 3, 3))
.SetFont(font).SetFontSize(11);
StreamReader sr = File.OpenText(SRC);
String name;
String line;
Paragraph p;
bool title = true;
int counter = 0;
IList <Util.Pair <String, Util.Pair <String, int> > > toc = new List <Util.Pair
<String, Util.Pair <String, int> > >();
while ((line = sr.ReadLine()) != null)
{
p = new Paragraph(line);
p.SetKeepTogether(true);
if (title)
{
name = String.Format("title{0:00}", counter++);
Util.Pair <String, int> titlePage = new Util.Pair <string, int>(line, pdf.GetNumberOfPages());
p.SetFont(bold).SetFontSize(12).SetKeepWithNext(true).SetDestination(name).SetNextRenderer(new UpdatePageRenderer(p, titlePage));
title = false;
document.Add(p);
toc.Add(new Util.Pair <string, Util.Pair <string, int> >(name, titlePage));
}
else
{
p.SetFirstLineIndent(36);
if (String.IsNullOrEmpty(line))
{
p.SetMarginBottom(12);
title = true;
}
else
{
p.SetMarginBottom(0);
}
document.Add(p);
}
}
document.Add(new AreaBreak(AreaBreakType.NEXT_PAGE));
p = new Paragraph().SetFont(bold).Add("Table of Contents").SetDestination("toc");
document.Add(p);
toc.RemoveAt(0);
IList <TabStop> tabstops = new List <TabStop>();
tabstops.Add(new TabStop(580, TabAlignment.RIGHT, new DottedLine()));
foreach (Util.Pair <String, Util.Pair <String, int> > entry in toc)
{
Util.Pair <String, int> text = entry.Value;
p = new Paragraph().AddTabStops(tabstops).Add(text.Key).Add(new Tab()).Add(text.Value.ToString()).SetAction
(PdfAction.CreateGoTo(entry.Key));
document.Add(p);
}
//Close document
document.Close();
}
public UpdatePageRenderer(Paragraph modelElement, Util.Pair
<String, int> entry)
: base(modelElement)
{
this.entry = entry;
}
/// <summary>
/// Builds up the list of features, anchored around a position within the
/// StringBuilder.
/// </summary>
public virtual string[] GetContext(Util.Pair <System.Text.StringBuilder, int> pair)
{
string prefix; //string preceeding the eos character in the eos token.
string previousToken; //space delimited token preceding token containing eos character.
string suffix; //string following the eos character in the eos token.
string nextToken; //space delimited token following token containsing eos character.
System.Text.StringBuilder buffer = pair.FirstValue;
int position = pair.SecondValue; //character offset of eos character in
//if (first is string[])
//{
// string[] firstList = (string[])first;
// previousToken = firstList[0];
// string current = firstList[1];
// prefix = current.Substring(0, (position) - (0));
// suffix = current.Substring(position + 1);
// if (suffix.StartsWith(" "))
// {
// mCollectFeatures.Add("sn");
// }
// if (prefix.EndsWith(" "))
// {
// mCollectFeatures.Add("pn");
// }
// mCollectFeatures.Add("eos=" + current[position]);
// nextToken = firstList[2];
//}
//else
//{
// //compute previous, next, prefix and suffix strings and space previous, space next features and eos features.
// System.Text.StringBuilder buffer = (System.Text.StringBuilder)((Util.Pair)input).FirstValue;
int lastIndex = buffer.Length - 1;
// compute space previousToken and space next features.
if (position > 0 && buffer[position - 1] == ' ')
{
mCollectFeatures.Add("sp");
}
if (position < lastIndex && buffer[position + 1] == ' ')
{
mCollectFeatures.Add("sn");
}
mCollectFeatures.Add("eos=" + buffer[position]);
int prefixStart = PreviousSpaceIndex(buffer, position);
int currentPosition = position;
//assign prefix, stop if you run into a period though otherwise stop at space
while (--currentPosition > prefixStart)
{
for (int currentEndOfSentenceCharacter = 0, endOfSentenceCharactersLength = mEndOfSentenceCharacters.Length; currentEndOfSentenceCharacter < endOfSentenceCharactersLength; currentEndOfSentenceCharacter++)
{
if (buffer[currentPosition] == mEndOfSentenceCharacters[currentEndOfSentenceCharacter])
{
prefixStart = currentPosition;
currentPosition++; // this gets us out of while loop.
break;
}
}
}
prefix = buffer.ToString(prefixStart, position - prefixStart).Trim();
int previousStart = PreviousSpaceIndex(buffer, prefixStart);
previousToken = buffer.ToString(previousStart, prefixStart - previousStart).Trim();
int suffixEnd = NextSpaceIndex(buffer, position, lastIndex);
currentPosition = position;
while (++currentPosition < suffixEnd)
{
for (int currentEndOfSentenceCharacter = 0, endOfSentenceCharactersLength = mEndOfSentenceCharacters.Length; currentEndOfSentenceCharacter < endOfSentenceCharactersLength; currentEndOfSentenceCharacter++)
{
if (buffer[currentPosition] == mEndOfSentenceCharacters[currentEndOfSentenceCharacter])
{
suffixEnd = currentPosition;
currentPosition--; // this gets us out of while loop.
break;
}
}
}
int nextEnd = NextSpaceIndex(buffer, suffixEnd + 1, lastIndex + 1);
if (position == lastIndex)
{
suffix = "";
nextToken = "";
}
else
{
suffix = buffer.ToString(position + 1, suffixEnd - (position + 1)).Trim();
nextToken = buffer.ToString(suffixEnd + 1, nextEnd - (suffixEnd + 1)).Trim();
}
mBuffer.Append("x=");
mBuffer.Append(prefix);
mCollectFeatures.Add(mBuffer.ToString());
mBuffer.Length = 0;
if (prefix.Length > 0)
{
mCollectFeatures.Add(System.Convert.ToString(prefix.Length, System.Globalization.CultureInfo.InvariantCulture));
if (IsFirstUpper(prefix))
{
mCollectFeatures.Add("xcap");
}
if (mInducedAbbreviations.Contains(prefix))
{
mCollectFeatures.Add("xabbrev");
}
}
mBuffer.Append("v=");
mBuffer.Append(previousToken);
mCollectFeatures.Add(mBuffer.ToString());
mBuffer.Length = 0;
if (previousToken.Length > 0)
{
if (IsFirstUpper(previousToken))
{
mCollectFeatures.Add("vcap");
}
if (mInducedAbbreviations.Contains(previousToken))
{
mCollectFeatures.Add("vabbrev");
}
}
mBuffer.Append("s=");
mBuffer.Append(suffix);
mCollectFeatures.Add(mBuffer.ToString());
mBuffer.Length = 0;
if (suffix.Length > 0)
{
if (IsFirstUpper(suffix))
{
mCollectFeatures.Add("scap");
}
if (mInducedAbbreviations.Contains(suffix))
{
mCollectFeatures.Add("sabbrev");
}
}
mBuffer.Append("n=");
mBuffer.Append(nextToken);
mCollectFeatures.Add(mBuffer.ToString());
mBuffer.Length = 0;
if (nextToken.Length > 0)
{
if (IsFirstUpper(nextToken))
{
mCollectFeatures.Add("ncap");
}
if (mInducedAbbreviations.Contains(nextToken))
{
mCollectFeatures.Add("nabbrev");
}
}
string[] context = mCollectFeatures.ToArray();
mCollectFeatures.Clear();
return(context);
}
public virtual void CreatePdf(String dest)
{
PdfDocument pdf = new PdfDocument(new PdfWriter(dest));
pdf.GetCatalog().SetPageLayout(PdfName.TwoColumnRight);
pdf.GetCatalog().SetPageMode(PdfName.UseThumbs);
PdfPage page = pdf.AddNewPage();
page.SetPageLabel(PageLabelNumberingStyle.LOWERCASE_ROMAN_NUMERALS, null);
Document document = new Document(pdf);
document.Add(new Paragraph().Add("Page left blank intentionally"));
document.Add(new AreaBreak());
document.Add(new Paragraph().Add("Page left blank intentionally"));
document.Add(new AreaBreak());
document.Add(new Paragraph().Add("Page left blank intentionally"));
document.Add(new AreaBreak());
page = pdf.GetLastPage();
page.SetPageLabel(PageLabelNumberingStyle.DECIMAL_ARABIC_NUMERALS, null, 1);
PdfFont font = PdfFontFactory.CreateFont(StandardFonts.TIMES_ROMAN);
PdfFont bold = PdfFontFactory.CreateFont(StandardFonts.HELVETICA_BOLD);
document.SetTextAlignment(TextAlignment.JUSTIFIED).SetHyphenation(new HyphenationConfig("en", "uk", 3, 3))
.SetFont(font).SetFontSize(11);
StreamReader sr = File.OpenText(SRC);
String name;
String line;
Paragraph p;
bool title = true;
int counter = 0;
IList <Util.Pair <String, Util.Pair <String, int> > > toc = new List <Util.Pair
<String, Util.Pair <String, int> > >();
while ((line = sr.ReadLine()) != null)
{
p = new Paragraph(line);
p.SetKeepTogether(true);
if (title)
{
name = String.Format("title{0:00}", counter++);
p.SetFont(bold).SetFontSize(12).SetKeepWithNext(true).SetDestination(name);
title = false;
document.Add(p);
toc.Add(new Util.Pair <string, Util.Pair <string, int> >(name, new Util.Pair <string, int>(line, pdf.GetNumberOfPages())));
}
else
{
p.SetFirstLineIndent(36);
if (String.IsNullOrEmpty(line))
{
p.SetMarginBottom(12);
title = true;
}
else
{
p.SetMarginBottom(0);
}
document.Add(p);
}
}
document.Add(new AreaBreak(AreaBreakType.NEXT_PAGE));
p = new Paragraph().SetFont(bold).Add("Table of Contents").SetDestination("toc");
document.Add(p);
page = pdf.GetLastPage();
page.SetPageLabel(null, "TOC", 1);
toc.RemoveAt(0);
IList <TabStop> tabstops = new List <TabStop>();
tabstops.Add(new TabStop(580, TabAlignment.RIGHT, new DottedLine()));
foreach (Util.Pair <String, Util.Pair <String, int> > entry in toc)
{
Util.Pair <String, int> text = entry.Value;
p = new Paragraph().AddTabStops(tabstops).Add(text.Key).Add(new Tab()).Add(text.Value.ToString()).SetAction
(PdfAction.CreateGoTo(entry.Key));
document.Add(p);
}
document.Close();
}
/// <summary>
/// Detect the position of the first words of sentences in a string.
/// </summary>
/// <param name="input">
/// The string to be processed.
/// </param>
/// <returns>
/// A integer array containing the positions of the end index of
/// every sentence
/// </returns>
public virtual int[] SentencePositionDetect(string input)
{
if (mUnicodeMapping) { input = Utils.MapUnicodeChars(input); }
double sentenceProbability = 1;
mSentenceProbs.Clear();
System.Text.StringBuilder buffer = new System.Text.StringBuilder(input);
List<int> endersList = mScanner.GetPositions(input);
List<int> positions = new List<int>(endersList.Count);
for (int currentEnder = 0, enderCount = endersList.Count, index = 0; currentEnder < enderCount; currentEnder++)
{
int candidate = endersList[currentEnder];
int cInt = candidate;
// skip over the leading parts of non-token final delimiters
int firstWhiteSpace = GetFirstWhitespace(input, cInt + 1);
if (((currentEnder + 1) < enderCount) && ((endersList[currentEnder + 1]) < firstWhiteSpace))
{
continue;
}
Util.Pair<System.Text.StringBuilder, int> pair = new Util.Pair<System.Text.StringBuilder, int>(buffer, candidate);
double[] probabilities = mModel.Evaluate(mContextGenerator.GetContext(pair));
string bestOutcome = mModel.GetBestOutcome(probabilities);
sentenceProbability *= probabilities[mModel.GetOutcomeIndex(bestOutcome)];
if (bestOutcome.Equals("T") && IsAcceptableBreak(input, index, cInt))
{
if (index != cInt)
{
positions.Add(GetFirstNonWhitespace(input, GetFirstWhitespace(input, cInt + 1)));//moIntegerPool.GetInteger(GetFirstNonWhitespace(input, GetFirstWhitespace(input, cInt + 1))));
mSentenceProbs.Add(probabilities[mModel.GetOutcomeIndex(bestOutcome)]);
}
index = cInt + 1;
}
}
return positions.ToArray();
}
/// <summary>
/// Detect sentences in a string.
/// </summary>
/// <param name="input">
/// The string to be processed.
/// </param>
/// <param name="positions">
/// Start index and length of each detected sentence.
/// </param>
/// <returns>
/// A string array containing individual sentences as elements.
/// </returns>
public string[] SentenceDetect(string input, out Util.Pair<int, int>[] positions)
{
int[] startsList = SentencePositionDetect(input);
positions = new Util.Pair<int, int>[] { new Util.Pair<int, int>(0, input.Length) };
if (startsList.Length == 0)
{
return new string[] {input};
}
bool isLeftover = startsList[startsList.Length - 1] != input.Length;
string[] sentences = new string[isLeftover ? startsList.Length + 1 : startsList.Length];
positions = new Util.Pair<int, int>[sentences.Length];
positions[0] = new Util.Pair<int, int>(0, (startsList[0]) - (0));
sentences[0] = input.Substring(positions[0].FirstValue, positions[0].SecondValue);
for (int currentStart = 1; currentStart < startsList.Length; currentStart++)
{
positions[currentStart] = new Util.Pair<int, int>(startsList[currentStart - 1], (startsList[currentStart]) - (startsList[currentStart - 1]));
sentences[currentStart] = input.Substring(positions[currentStart].FirstValue, positions[currentStart].SecondValue);
}
if (isLeftover)
{
positions[sentences.Length - 1] = new Util.Pair<int, int>(startsList[startsList.Length - 1], input.Length - startsList[startsList.Length - 1]);
sentences[sentences.Length - 1] = input.Substring(positions[sentences.Length - 1].FirstValue);
}
return (sentences);
}
private void AddNewEvents(string token)
{
System.Text.StringBuilder buffer = mBuffer;
buffer.Append(token.Trim());
int sentenceEndPosition = buffer.Length - 1;
//add following word to stringbuilder
if (mNext != null && token.Length > 0)
{
int positionAfterFirstWordInNext = mNext.IndexOf(" ");
if (positionAfterFirstWordInNext != - 1)
{
// should maybe changes this so that it usually adds a space
// before the next sentence, but sometimes leaves no space.
buffer.Append(" ");
buffer.Append(mNext.Substring(0, (positionAfterFirstWordInNext) - (0)));
}
}
for (System.Collections.IEnumerator iterator = mScanner.GetPositions(buffer).GetEnumerator(); iterator.MoveNext(); )
{
int candidate = (int) iterator.Current;
Util.Pair<System.Text.StringBuilder, int> pair = new Util.Pair<System.Text.StringBuilder, int>(buffer, candidate);
string type = (candidate == sentenceEndPosition) ? "T" : "F";
SentenceDetectionEvent sentenceEvent = new SentenceDetectionEvent(type, mContextGenerator.GetContext(pair));
if (null != mTail)
{
mTail.NextEvent = sentenceEvent;
mTail = sentenceEvent;
}
else if (null == mHead)
{
mHead = sentenceEvent;
}
else if (null == mHead.NextEvent)
{
mHead.NextEvent = mTail = sentenceEvent;
}
}
buffer.Length = 0;
}