private void AddPatternWithCaseSensitivePrefix(string prefix, IPatternMatcher matcher, bool needSeparators, object[] tags)
{
char startChar = prefix[0];
int startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddPattern(prefix, true, needSeparators, matcher, tags);
}
private void AddCaseSensitiveToken(string text, bool needSeparators, object[] tags)
{
char startChar = text[0];
int startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddToken(text, true, needSeparators, tags);
}
public TokenTree()
{
nodes = new TokenTreeNode[65536];
Separators = ".,;:<>[](){}!\"#¤%&/=?*+-/\\ \t\n\r";
textLookup = new char[65536];
for (int i = 0; i < 65536; i++) {
textLookup[i] = (char) i;
}
textLookup['\t'] = ' ';
root = new TokenTreeNode();
}
public void AddPattern(string prefix, bool caseSensitive, bool needSeparators, IPatternMatcher matcher,
object[] tags)
{
if (string.IsNullOrEmpty(prefix))
{
throw new ArgumentNullException("prefix");
}
TokenTreeNode node = AddTokenInternal(prefix, caseSensitive);
var patternMatcherReference = new PatternMatchReference(matcher)
{
NextSibling = FirstExpression,
Tags = tags,
NeedSeparators = needSeparators
};
node.FirstExpression = patternMatcherReference;
}
private void AddPatternWithCaseInsensitivePrefix(string prefix, IPatternMatcher matcher, bool needSeparators, object[] tags)
{
//make a lowercase string and add it as a token
prefix = prefix.ToLower();
char startChar = prefix[0];
int startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddPattern(prefix, false, needSeparators, matcher, tags);
//make a lowercase string with a uppercase start char and add it as a token
prefix = char.ToUpper(startChar) + prefix.Substring(1);
startChar = prefix[0];
startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddPattern(prefix, false, needSeparators, matcher, tags);
}
private void AddCaseInsensitiveToken(string text, bool needSeparators, object[] tags)
{
//make a lowercase string and add it as a token
text = text.ToLower();
char startChar = text[0];
int startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddToken(text, false, needSeparators, tags);
//make a lowercase string with a uppercase start char and add it as a token
text = char.ToUpper(startChar) + text.Substring(1);
startChar = text[0];
startIndex = startChar;
if (nodes[startIndex] == null)
nodes[startIndex] = new TokenTreeNode();
nodes[startIndex].AddToken(text, false, needSeparators, tags);
}
public TokenTreeNode AddTokenInternal(string token, bool caseSensitive)
{
Char = token[0];
if (!caseSensitive)
{
ContainsCaseInsensitiveData = true;
}
if (token.Length == 1)
{
return(this);
}
string leftovers = token.Substring(1);
char childChar = leftovers[0];
int childIndex = childChar & 0xff;
//make a lookupindex (dont mind if unicode chars end up as siblings as ascii)
TokenTreeNode node = ChildNodes[childIndex];
TokenTreeNode res;
if (node == null)
{
var child = new TokenTreeNode();
ChildNodes[childIndex] = child;
res = child.AddTokenInternal(leftovers, caseSensitive);
MakeRepeatingWS(child);
}
else
{
node = GetMatchingNode(childChar, node);
res = node.AddTokenInternal(leftovers, caseSensitive);
}
return(res);
}
public TokenTreeNode AddTokenInternal(string token, bool caseSensitive)
{
Char = token[0];
if (!caseSensitive)
ContainsCaseInsensitiveData = true;
if (token.Length == 1)
return this;
string leftovers = token.Substring(1);
char childChar = leftovers[0];
int childIndex = childChar & 0xff;
//make a lookupindex (dont mind if unicode chars end up as siblings as ascii)
TokenTreeNode node = ChildNodes[childIndex];
TokenTreeNode res;
if (node == null)
{
var child = new TokenTreeNode();
ChildNodes[childIndex] = child;
res = child.AddTokenInternal(leftovers, caseSensitive);
MakeRepeatingWS(child);
}
else
{
node = GetMatchingNode(childChar, node);
res = node.AddTokenInternal(leftovers, caseSensitive);
}
return res;
}
private static void MakeRepeatingWS(TokenTreeNode child)
{
if (child.Char == ' ')
{
// if the node contains " " (whitespace)
// then add the node as a childnode of itself.
// thus allowing it to parse things like
// "end sub" even if the pattern is "end sub" // do not localize
child.ChildNodes[' '] = child;
}
}
private static TokenTreeNode GetMatchingNode(char childChar, TokenTreeNode node)
{
//find a bucket with the same childChar as we need
while (node.NextSibling != null && node.Char != childChar)
{
node = node.NextSibling;
}
if (node.Char != childChar)
{
var child = new TokenTreeNode();
node.NextSibling = child;
return child;
}
return node;
}
//this is wicked fast
//do not refactor extract methods from this if you want to keep the speed
public MatchResult Match(string text, int startIndex)
{
if (string.IsNullOrEmpty(text))
{
throw new ArgumentNullException(text);
}
var lastMatch = new MatchResult {
Text = text
};
int textLength = text.Length;
for (int currentIndex = startIndex; currentIndex < textLength; currentIndex++)
{
//call any prefixless patternmatchers
#region HasExpressions
if (root.FirstExpression != null)
{
//begin with the first expression of the _root node_
PatternMatchReference patternMatcherReference = root.FirstExpression;
while (patternMatcherReference != null)
{
int patternMatchIndex = patternMatcherReference.Matcher.Match(text, currentIndex);
if (patternMatchIndex > 0 && patternMatchIndex > lastMatch.Length)
{
bool leftIsSeparator = currentIndex == 0 ? true : separatorCharLookup[text[currentIndex - 1]];
bool rightIsSeparator = (currentIndex + patternMatchIndex) == textLength ? true : separatorCharLookup[text[currentIndex + patternMatchIndex]];
if (!patternMatcherReference.NeedSeparators || (leftIsSeparator && rightIsSeparator))
{
lastMatch.Index = currentIndex;
lastMatch.Length = patternMatchIndex;
lastMatch.Found = true;
lastMatch.Tags = patternMatcherReference.Tags;
}
}
patternMatcherReference = patternMatcherReference.NextSibling;
}
}
#endregion
//lookup the first token tree node
TokenTreeNode node = nodes[text[currentIndex]];
if (node == null)
{
if (lastMatch.Found)
{
break;
}
continue;
}
for (int matchIndex = currentIndex + 1; matchIndex <= textLength; matchIndex++)
{
//call patternmatchers for the current prefix
#region HasExpressions
if (node.FirstExpression != null)
{
//begin with the first expression of the _current node_
PatternMatchReference patternMatcherReference = node.FirstExpression;
while (patternMatcherReference != null)
{
int patternMatchIndex = patternMatcherReference.Matcher.Match(text, matchIndex);
if (patternMatchIndex > 0 && patternMatchIndex > lastMatch.Length)
{
bool leftIsSeparator = currentIndex == 0 ? true : separatorCharLookup[text[currentIndex - 1]];
bool rightIsSeparator = (currentIndex + patternMatchIndex + matchIndex) == textLength ? true : separatorCharLookup[text[currentIndex + patternMatchIndex + matchIndex]];
if (!patternMatcherReference.NeedSeparators || (leftIsSeparator && rightIsSeparator))
{
lastMatch.Index = currentIndex;
lastMatch.Length = patternMatchIndex + matchIndex - currentIndex;
lastMatch.Found = true;
lastMatch.Tags = patternMatcherReference.Tags;
}
}
patternMatcherReference = patternMatcherReference.NextSibling;
}
}
#endregion
#region IsEndNode
if (node.IsEnd && matchIndex - currentIndex >= lastMatch.Length)
{
bool leftIsSeparator = currentIndex == 0 ? true : separatorCharLookup[text[currentIndex - 1]];
bool rightIsSeparator = matchIndex == textLength ? true : separatorCharLookup[text[matchIndex]];
if (!node.NeedSeparators || (leftIsSeparator && rightIsSeparator))
{
lastMatch.Index = currentIndex;
lastMatch.Tags = node.Tags;
lastMatch.Found = true;
lastMatch.Length = matchIndex - currentIndex;
//TODO:perform case test here , case sensitive words might be matched even if they have incorrect case
if (currentIndex + lastMatch.Length == textLength)
{
break;
}
}
}
#endregion
//try fetch a node at this index
node = node.GetNextNode(textLookup[text[matchIndex]]);
//we found no node on the lookupindex or none of the siblingnodes at that index matched the current char
if (node == null)
{
break; // continue with the next character
}
}
//return last match
if (lastMatch.Found)
{
return(lastMatch);
}
}
if (lastMatch.Found)
{
return(lastMatch);
}
//no match was found
return(MatchResult.NoMatch);
}
