void DoIndexOfSmallAmount(TextBuffer textBuffer)
{
//convention...
//data must me sorted (ascending) before use with the wordSpanList
for (int i = wordSpanList.Count - 1; i >= 0; --i)
{
WordSpan sp = wordSpanList[i];
#if DEBUG
//string dbugStr = sp.GetString(textBuffer);
#endif
if (sp.SameTextContent(this.prefixSpan, textBuffer))
{
this.PrefixIsWord = true;
break;
}
}
this._resultWordGroup = new WordGroup(
this.prefixSpan,
null,
this.wordSpanList.ToArray(),
this.PrefixIsWord);
}
public void GetWordList(char startWithChar, List <string> output)
{
if (startWithChar >= firstChar && startWithChar <= lastChar)
{
//in range
WordGroup found = this.wordGroups[TransformCharToIndex(startWithChar)];
if (found != null)
{//iterate and collect into
found.CollectAllWords(this.textBuffer, output);
}
}
}
void DoIndex(Dictionary <char, DevelopingWordGroup> wordGroups)
{
//1. expand word group
WordGroup[] newWordGroups = new WordGroup[this.lastChar - this.firstChar + 1];
foreach (var kp in wordGroups)
{
//for each dev word group
int index = TransformCharToIndex(kp.Key);
DevelopingWordGroup devWordGroup = kp.Value;
devWordGroup.DoIndex(this.textBuffer, this);
newWordGroups[index] = devWordGroup.ResultWordGroup;
}
this.wordGroups = newWordGroups;
}
internal WordGroup GetSubGroup(WordVisitor visitor, WordGroup wordGroup)
{
char c = visitor.Char;
if (!CanHandle(c))
{
//can't handle
//then no furtur sub group
visitor.State = VisitorState.OutOfRangeChar;
return(null);
}
//-----------------
//can handle
WordGroup[] subGroups = wordGroup.GetSubGroups();
if (subGroups != null)
{
return(subGroups[c - this.FirstUnicodeChar]);
}
return(null);
}
WordGroup _resultWordGroup;//after call DoIndex()
internal void DoIndex(TextBuffer textBuffer, CustomDic owner)
{
//recursive
if (this.PrefixLen > 7)
{
DoIndexOfSmallAmount(textBuffer);
#if DEBUG
dbugDataState = debugDataState.TooLongPrefix;
#endif
return;
}
//-----------------------------------------------
bool hasEvalPrefix = false;
if (subGroups == null)
{
subGroups = new DevelopingWordGroup[owner.LastChar - owner.FirstChar + 1];
}
//--------------------------------
int j = wordSpanList.Count;
int thisPrefixLen = this.PrefixLen;
int doSepAt = thisPrefixLen;
for (int i = 0; i < j; ++i)
{
WordSpan sp = wordSpanList[i];
if (sp.len > doSepAt)
{
char c = sp.GetChar(doSepAt, textBuffer);
int c_index = c - owner.FirstChar;
DevelopingWordGroup found = subGroups[c_index];
if (found == null)
{
//not found
found = new DevelopingWordGroup(new WordSpan(sp.startAt, (byte)(doSepAt + 1)));
subGroups[c_index] = found;
}
found.AddWordSpan(sp);
}
else
{
if (!hasEvalPrefix)
{
if (sp.SameTextContent(this.prefixSpan, textBuffer))
{
hasEvalPrefix = true;
this.PrefixIsWord = true;
}
}
}
}
#if DEBUG
this.dbugDataState = debugDataState.Indexed;
#endif
wordSpanList.Clear();
wordSpanList = null;
//--------------------------------
//do sup index
//foreach (WordGroup subgroup in this.wordGroups.Values)
bool hasSomeSubGroup = false;
foreach (DevelopingWordGroup subgroup in this.subGroups)
{
if (subgroup != null)
{
hasSomeSubGroup = true;
//****
//performance factor here,****
//in this current version
//if we not call DoIndex(),
//this subgroup need linear search-> so it slow
//so we call DoIndex until member count in the group <=3
//then it search faster,
//but dictionary-building time may increase.
if (subgroup.WordSpanListCount > 2)
{
subgroup.DoIndex(textBuffer, owner);
}
else
{
#if DEBUG
subgroup.dbugDataState = debugDataState.SmallAmountOfMembers;
#endif
subgroup.DoIndexOfSmallAmount(textBuffer);
}
}
}
//--------------------------------
#if DEBUG
this.dbugDataState = debugDataState.Indexed;
#endif
if (!hasSomeSubGroup)
{
//clear
subGroups = null;
}
//--------------------------------
WordGroup[] newsubGroups = null;
if (subGroups != null)
{
newsubGroups = new WordGroup[subGroups.Length];
for (int i = subGroups.Length - 1; i >= 0; --i)
{
DevelopingWordGroup subg = subGroups[i];
if (subg != null)
{
newsubGroups[i] = subg.ResultWordGroup;
}
}
}
//--------------------------------
this._resultWordGroup = new WordGroup(
this.prefixSpan,
newsubGroups,
null,
this.PrefixIsWord);
}
int FindInWordSpans(WordVisitor visitor, WordGroup wordGroup)
{
WordSpan[] wordSpans = wordGroup.GetWordSpans();
if (wordSpans == null)
{
throw new NotSupportedException();
}
//at this wordgroup
//no subground anymore
//so we should find the word one by one
//start at prefix
//and select the one that
int readLen = visitor.CurrentIndex - visitor.LatestBreakAt;
int nwords = wordSpans.Length;
//only 1 that match
TextBuffer currentTextBuffer = CurrentCustomDic.TextBuffer;
//we sort unindex string ***
//so we find from longest one( last) to begin
for (int i = nwords - 1; i >= 0; --i)
{
//loop test on each word
WordSpan w = wordSpans[i];
#if DEBUG
//string dbugstr = w.GetString(currentTextBuffer);
#endif
int savedIndex = visitor.CurrentIndex;
char c = visitor.Char;
int wordLen = w.len;
int matchCharCount = 0;
if (wordLen > readLen)
{
for (int p = readLen; p < wordLen; ++p)
{
char c2 = w.GetChar(p, currentTextBuffer);
if (c2 == c)
{
matchCharCount++;
//match
//read next
if (!visitor.IsEnd)
{
visitor.SetCurrentIndex(visitor.CurrentIndex + 1);
c = visitor.Char;
}
else
{
//no more data in visitor
break;
}
}
else
{
break;
}
}
}
//reset
if (readLen + matchCharCount == wordLen)
{
int newBreakAt = visitor.LatestBreakAt + wordLen;
visitor.SetCurrentIndex(newBreakAt);
//--------------------------------------------
if (visitor.State == VisitorState.End)
{
return newBreakAt;
}
//check next char can be the char of new word or not
//this depends on each lang
char canBeStartChar = visitor.Char;
if (CanHandle(canBeStartChar))
{
if (CanBeStartChar(canBeStartChar))
{
return newBreakAt;
}
else
{
//back to savedIndex
visitor.SetCurrentIndex(savedIndex);
return savedIndex;
}
}
else
{
visitor.State = VisitorState.OutOfRangeChar;
return newBreakAt;
}
}
visitor.SetCurrentIndex(savedIndex);
}
return 0;
}
internal WordGroup GetSubGroup(WordVisitor visitor, WordGroup wordGroup)
{
char c = visitor.Char;
if (!CanHandle(c))
{
//can't handle
//then no furtur sub group
visitor.State = VisitorState.OutOfRangeChar;
return null;
}
//-----------------
//can handle
WordGroup[] subGroups = wordGroup.GetSubGroups();
if (subGroups != null)
{
return subGroups[c - this.FirstUnicodeChar];
}
return null;
}
internal WordGroup(WordSpan prefixSpan, WordGroup[] subGroups, WordSpan[] wordSpanList, bool isPrefixIsWord)
{
this.prefixSpan = prefixSpan;
this.subGroups = subGroups;
this.wordSpans = wordSpanList;
this.prefixIsWord = isPrefixIsWord;
}
void DoIndexOfSmallAmount(TextBuffer textBuffer)
{
//convention...
//data must me sorted (ascending) before use with the wordSpanList
for (int i = wordSpanList.Count - 1; i >= 0; --i)
{
WordSpan sp = wordSpanList[i];
#if DEBUG
//string dbugStr = sp.GetString(textBuffer);
#endif
if (sp.SameTextContent(this.prefixSpan, textBuffer))
{
this.PrefixIsWord = true;
break;
}
}
this._resultWordGroup = new WordGroup(
this.prefixSpan,
null,
this.wordSpanList.ToArray(),
this.PrefixIsWord);
}
WordGroup _resultWordGroup;//after call DoIndex()
internal void DoIndex(TextBuffer textBuffer, CustomDic owner)
{
//recursive
if (this.PrefixLen > 7)
{
DoIndexOfSmallAmount(textBuffer);
#if DEBUG
dbugDataState = debugDataState.TooLongPrefix;
#endif
return;
}
//-----------------------------------------------
bool hasEvalPrefix = false;
if (subGroups == null)
{
subGroups = new DevelopingWordGroup[owner.LastChar - owner.FirstChar + 1];
}
//--------------------------------
int j = wordSpanList.Count;
int thisPrefixLen = this.PrefixLen;
int doSepAt = thisPrefixLen;
for (int i = 0; i < j; ++i)
{
WordSpan sp = wordSpanList[i];
if (sp.len > doSepAt)
{
char c = sp.GetChar(doSepAt, textBuffer);
int c_index = c - owner.FirstChar;
DevelopingWordGroup found = subGroups[c_index];
if (found == null)
{
//not found
found = new DevelopingWordGroup(new WordSpan(sp.startAt, (byte)(doSepAt + 1)));
subGroups[c_index] = found;
}
found.AddWordSpan(sp);
}
else
{
if (!hasEvalPrefix)
{
if (sp.SameTextContent(this.prefixSpan, textBuffer))
{
hasEvalPrefix = true;
this.PrefixIsWord = true;
}
}
}
}
#if DEBUG
this.dbugDataState = debugDataState.Indexed;
#endif
wordSpanList.Clear();
wordSpanList = null;
//--------------------------------
//do sup index
//foreach (WordGroup subgroup in this.wordGroups.Values)
bool hasSomeSubGroup = false;
foreach (DevelopingWordGroup subgroup in this.subGroups)
{
if (subgroup != null)
{
hasSomeSubGroup = true;
//****
//performance factor here,****
//in this current version
//if we not call DoIndex(),
//this subgroup need linear search-> so it slow
//so we call DoIndex until member count in the group <=3
//then it search faster,
//but dictionary-building time may increase.
if (subgroup.WordSpanListCount > 2)
{
subgroup.DoIndex(textBuffer, owner);
}
else
{
#if DEBUG
subgroup.dbugDataState = debugDataState.SmallAmountOfMembers;
#endif
subgroup.DoIndexOfSmallAmount(textBuffer);
}
}
}
//--------------------------------
#if DEBUG
this.dbugDataState = debugDataState.Indexed;
#endif
if (!hasSomeSubGroup)
{
//clear
subGroups = null;
}
//--------------------------------
WordGroup[] newsubGroups = null;
if (subGroups != null)
{
newsubGroups = new WordGroup[subGroups.Length];
for (int i = subGroups.Length - 1; i >= 0; --i)
{
DevelopingWordGroup subg = subGroups[i];
if (subg != null)
{
newsubGroups[i] = subg.ResultWordGroup;
}
}
}
//--------------------------------
this._resultWordGroup = new WordGroup(
this.prefixSpan,
newsubGroups,
null,
this.PrefixIsWord);
}
void DoIndex(Dictionary<char, DevelopingWordGroup> wordGroups)
{
//1. expand word group
WordGroup[] newWordGroups = new WordGroup[this.lastChar - this.firstChar + 1];
foreach (var kp in wordGroups)
{
//for each dev word group
int index = TransformCharToIndex(kp.Key);
DevelopingWordGroup devWordGroup = kp.Value;
devWordGroup.DoIndex(this.textBuffer, this);
newWordGroups[index] = devWordGroup.ResultWordGroup;
}
this.wordGroups = newWordGroups;
}
public override void BreakWord(WordVisitor visitor, char[] charBuff, int startAt, int len)
{
visitor.State = VisitorState.Parsing;
char c_first = this.FirstUnicodeChar;
char c_last = this.LastUnicodeChar;
int endAt = startAt + len;
Stack <int> candidate = visitor.GetTempCandidateBreaks();
for (int i = startAt; i < endAt;)
{
//find proper start words;
char c = charBuff[i];
//----------------------
//check if c is in our responsiblity
if (c < c_first || c > c_last)
{
//out of our range
//should return ?
visitor.State = VisitorState.OutOfRangeChar;
return;
}
//----------------------
WordGroup wordgroup = GetWordGroupForFirstChar(c);
if (wordgroup == null)
{
//continue next char
++i;
visitor.AddWordBreakAt(i);
visitor.SetCurrentIndex(visitor.LatestBreakAt);
}
else
{
//check if we can move next
if (visitor.IsEnd)
{
visitor.State = VisitorState.End;
return;
}
//---------------------
WordGroup c_wordgroup = wordgroup;
candidate.Clear();
int candidateLen = 1;
if (c_wordgroup.PrefixIsWord)
{
candidate.Push(candidateLen);
}
bool continueRead = true;
int savedIndex = visitor.CurrentIndex;
while (continueRead)
{
//not end
//then move next
candidateLen++;
visitor.SetCurrentIndex(i + 1);
WordGroup next = GetSubGroup(visitor, c_wordgroup);
//for debug
//string prefix = (next == null) ? "" : next.GetPrefix(CurrentCustomDic.TextBuffer);
if (next != null)
{
if (next.PrefixIsWord)
{
candidate.Push(candidateLen);
}
c_wordgroup = next;
i = visitor.CurrentIndex;
}
else
{
continueRead = false;
//no deeper group
//then check if
if (c_wordgroup.WordSpanListCount > 0)
{
int p1 = visitor.CurrentIndex;
//p2: suggest position
int p2 = FindInWordSpans(visitor, c_wordgroup);
if (p2 - p1 > 0)
{
visitor.AddWordBreakAt(p2);
visitor.SetCurrentIndex(p2);
}
else
{
//on the same pos
if (visitor.State == VisitorState.OutOfRangeChar)
{
return;
}
else
{
bool foundCandidate = false;
while (candidate.Count > 0)
{
int candi1 = candidate.Pop();
//try
visitor.SetCurrentIndex(visitor.LatestBreakAt + candi1);
//check if we can use this candidate
char next_char = visitor.Char;
if (CanBeStartChar(next_char))
{
//use this
//use this candidate if possible
visitor.AddWordBreakAt(visitor.CurrentIndex);
foundCandidate = true;
break;
}
}
if (!foundCandidate)
{
//no next word, no candidate
//skip this
char next_char = visitor.Char;
if (CanBeStartChar(next_char))
{
//use this
//use this candidate if possible
visitor.AddWordBreakAt(visitor.CurrentIndex);
foundCandidate = true;
break;
}
else
{
//TODO: review here
visitor.SetCurrentIndex(visitor.LatestBreakAt + 1);
visitor.AddWordBreakAt(visitor.CurrentIndex);
visitor.SetCurrentIndex(visitor.LatestBreakAt);
}
}
}
}
}
else
{
bool foundCandidate = false;
while (candidate.Count > 0)
{
int candi1 = candidate.Pop();
//try
visitor.SetCurrentIndex(visitor.LatestBreakAt + candi1);
if (visitor.State == VisitorState.End)
{
visitor.AddWordBreakAt(visitor.CurrentIndex);
return;
}
//check if we can use this candidate
char next_char = visitor.Char;
if (!CanHandle(next_char))
{
//use this
//use this candidate if possible
visitor.AddWordBreakAt(visitor.CurrentIndex);
foundCandidate = true;
break;
}
if (CanBeStartChar(next_char))
{
//use this
//use this candidate if possible
visitor.AddWordBreakAt(visitor.CurrentIndex);
foundCandidate = true;
break;
}
}
if (!foundCandidate)
{
if (candidateLen > 0)
{
//use that candidate len
visitor.AddWordBreakAt(visitor.CurrentIndex);
visitor.SetCurrentIndex(visitor.LatestBreakAt);
}
}
}
i = visitor.CurrentIndex;
}
}
}
}
//------
if (visitor.CurrentIndex >= len - 1)
{
//the last one
visitor.State = VisitorState.End;
}
}
int FindInWordSpans(WordVisitor visitor, WordGroup wordGroup)
{
WordSpan[] wordSpans = wordGroup.GetWordSpans();
if (wordSpans == null)
{
throw new NotSupportedException();
}
//at this wordgroup
//no subground anymore
//so we should find the word one by one
//start at prefix
//and select the one that
int readLen = visitor.CurrentIndex - visitor.LatestBreakAt;
int nwords = wordSpans.Length;
//only 1 that match
TextBuffer currentTextBuffer = CurrentCustomDic.TextBuffer;
//we sort unindex string ***
//so we find from longest one( last) to begin
for (int i = nwords - 1; i >= 0; --i)
{
//loop test on each word
WordSpan w = wordSpans[i];
#if DEBUG
//string dbugstr = w.GetString(currentTextBuffer);
#endif
int savedIndex = visitor.CurrentIndex;
char c = visitor.Char;
int wordLen = w.len;
int matchCharCount = 0;
if (wordLen > readLen)
{
for (int p = readLen; p < wordLen; ++p)
{
char c2 = w.GetChar(p, currentTextBuffer);
if (c2 == c)
{
matchCharCount++;
//match
//read next
if (!visitor.IsEnd)
{
visitor.SetCurrentIndex(visitor.CurrentIndex + 1);
c = visitor.Char;
}
else
{
//no more data in visitor
break;
}
}
else
{
break;
}
}
}
//reset
if (readLen + matchCharCount == wordLen)
{
int newBreakAt = visitor.LatestBreakAt + wordLen;
visitor.SetCurrentIndex(newBreakAt);
//--------------------------------------------
if (visitor.State == VisitorState.End)
{
return(newBreakAt);
}
//check next char can be the char of new word or not
//this depends on each lang
char canBeStartChar = visitor.Char;
if (CanHandle(canBeStartChar))
{
if (CanBeStartChar(canBeStartChar))
{
return(newBreakAt);
}
else
{
//back to savedIndex
visitor.SetCurrentIndex(savedIndex);
return(savedIndex);
}
}
else
{
visitor.State = VisitorState.OutOfRangeChar;
return(newBreakAt);
}
}
visitor.SetCurrentIndex(savedIndex);
}
return(0);
}
