/// <summary>
/// Traverse the suffix tree, following the longest path from the root that matches a prefix of words[wordNum].
/// This allows the caller to skip over these duplicate characters, and process only the part of the coming word.
/// </summary>
/// <param name="active">The current active suffix</param>
/// <param name="endIndex">The number of characters skipped</param>
/// <param name="wordNum">The index of the current word begin processed</param>
/// <seealso cref="http://www.cs.uku.fi/~kilpelai/BSA05/lectures/slides08.pdf">
/// The first 10 slides of this slideshow by Pekka Kilpeläinen
/// have useful tips on creating a generalized suffix tree.
/// </seealso>
/// <remarks>
/// TODO: Note: The following method is WORK IN PROGRESS, and does not yet work.
/// </remarks>
private void skipDuplicateInitialSubstring(ref GstSuffix active, ref int endIndex, int wordNum)
{
GstNode curNode = root;
GstEdge nextEdge = null;
GstEdge curEdge = null;
// Traverse matching edges
while (
(endIndex < wordDict[wordNum].Length) &&
((nextEdge = curNode.GetChildEdge(GetWordChar(wordNum, endIndex))) != null)
)
{
int strLen = nextEdge.Span(0) + 1;
// edgeStr = String in next edge
string edgeStr = nextEdge.GetText();
// wordStr = next segment of upcoming word that corresponds to edgeStr
string wordStr = wordDict[wordNum].Substring(endIndex, Math.Min(strLen, wordDict[wordNum].Length - endIndex));
bool foundMismatch = false;
int numCharsMatched = 0;
// Traverse matching characters within edge
for (int i = 0; i < strLen; i++)
{
if (edgeStr[i] == wordStr[i])
{
numCharsMatched++;
}
else
{
foundMismatch = true; break;
}
}
if (foundMismatch)
{
GstUtil.WriteLine(GstVerbosityLevel.Verbose, String.Format(
" skipDuplicateInitialSubstring: Word #{0:d} does not cover existing edge #{1:d}",
wordNum, nextEdge.Id));
active.OriginNode = nextEdge.ParentNode;
active.EndIndex = active.BeginIndex;
break;
}
else
{
nextEdge.SetBeginIndex(wordNum, endIndex);
nextEdge.SetEndIndex(wordNum, endIndex + strLen - 1);
GstUtil.WriteLine(GstVerbosityLevel.Verbose, String.Format(
" skipDuplicateInitialSubstring: Word #{0:d} covers existing edge #{1:d} ({2:s})",
wordNum, nextEdge.Id, nextEdge.ToString(wordNum)));
active.OriginNode = nextEdge.ChildNode;
active.BeginIndex += numCharsMatched;
active.EndIndex = active.BeginIndex;
}
endIndex += numCharsMatched;
// Set up next iteration of loop
curEdge = nextEdge;
curNode = curEdge.ChildNode;
}
}
public void MoveFromTo(GstNode oldParentNode, char oldFirstChar, GstNode newParentNode, char newFirstChar)
{
GstEdge self = oldParentNode.GetChildEdge(oldFirstChar);
if (self != this)
{
throw new ArgumentException("Error: MoveTo called with incorrect parent node and/or first char arguments");
}
oldParentNode.RemoveChildEdge(oldFirstChar);
newParentNode.AddChildEdge(newFirstChar, this);
}